text stringlengths 234 589k | id stringlengths 47 47 | dump stringclasses 62 values | url stringlengths 16 734 | date stringlengths 20 20 ⌀ | file_path stringlengths 109 155 | language stringclasses 1 value | language_score float64 0.65 1 | token_count int64 57 124k | score float64 2.52 4.91 | int_score int64 3 5 |
|---|---|---|---|---|---|---|---|---|---|---|
At the National Institute of Standards and Technology, scientists work in a massive, purpose-built lab to understand the nature of fire and to develop ways to improve fire safety. Today's products and buildings are constructed with stronger fire codes developed from NIST's results.
The Summit supercomputer at Oak Ridge National Laboratory, the fastest supercomputer in the U.S., is about to be replaced by an even speedier supercomputer named Frontier, expected to be at least five times as fast — and possibly the fastest in the world.
Scientists working for the National Oceanic and Atmospheric Administration keep track of nature's events from the sky to the bottom of the ocean. Uncrewed aerial systems — UAS, or drones — have become a critical tool in this work. | <urn:uuid:7a5a1361-3c32-4b94-a62f-ea2306a73146> | CC-MAIN-2022-40 | https://fedtechmagazine.com/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338280.51/warc/CC-MAIN-20221007210452-20221008000452-00125.warc.gz | en | 0.945306 | 154 | 3.078125 | 3 |
LIMSPEC, HIMEM.sys and EMM386.sys. So that this history is not lost, here is more information on expanded memory managers than you ever wanted to know.
What is EMM386?
EMM386 is the DOS 5 implemented device driver to implement an EXPANDED memory manager in DOS using the services of the 80386 processor instead of dedicated hardware. Your next question: What is an EXPANDED memory manager and how does it differ from an EXTENDED memory manager? Good questions. This will take some text to answer, and this was all very important stuff in the late 1980s!
Nobody will ever need more than 640KB
With the Intel 8088 and the original IBM PC, there was addressability for 1MB of memory. People describe the original PC has having a memory limit of 640KB, but that is not really accurate. The machine had 20 bits of memory addressing on the ISA bus and (2**20) -1 is 1MB. By convention, the bottom 640K is dedicated to software and the high 360KB is dedicated to hardware – such as screen memory. Much of the hardware space was unused; reserved for possible new hardware, but things were scattered holes, monochrome screen memory being 4KB at B000:0000 for example and the graphics memory adapter living at B8000:0000. Write bits to those spaces and characters show up on the screen! Hardware devices are above 640KB and BIOS, DOS and applications are below. This became not absolute.
When Lotus 123 ruled the world – the 640KB RAM limit was exhausted with large spreadsheets as well as with other large applications. Application implemented paging storage to and from disk was a common work around – but it is slow, and hard to program. Think diskettes. Even spinning hard disk media of the day was also dreadfully slow.
The first solution was to buy memory cards that plugged in the ISA bus and filled memory space above 640KB inside the holes where hardware memory was not used. On some systems, it was possible to free the first 64KB of hardware space and add RAM memory to fill that hole and you could end up with a DOS computer with 704KB of usable RAM. This was interesting, but it was not expanded memory.
What was needed was an ability to get “more RAM” and page that RAM in when needed.
640KB is the DOS limit and there were a few interim 8088 usable expanded memory solutions that never took hold and diverged the industry. Each implemented “more memory” in an ISA bus card and provided some hardware means to copy memory to and from 8088 visible memory.
Each application that wanted to use this additional memory, had to code to the hardware of the individual hardware card. Applications coding to multiple hardware specs! This is untenable.
LIMSPEC – Lotus Intel Microsoft Specification
Eventually, Lotus, Intel and Microsoft got together and invented “LIM Spec” (Lotus Intel Microsoft SPEC) to standardize this behavior, providing an API that implemented the memory paging to and from memory outside of DOS visible space and this was officially titled “Expanded Memory Specification“, though the programmers continued to call it LIM Spec so they understood which API people were discussing.
Some of the other ones included “XMS from IBM”?? These were not the same.
The LIM spec hardware worked via hardware paging. The LIM spec memory adapter hardware could have LOTS of memory on it – Like 512KB or 1 MB or even more! I think 16MB was the theoretical limit, but nobody could afford that much memory.
By whacking a few registers on the hardware memory card via I/O bus INs and OUTs, the backing store behind the memory at particular “block” could be changed (was actually called a “page”, but it was a 16KB or 64KB page). So, if you have a large spreadsheet, instead of paging it to and from disk, you could page it to and from “invisible memory” by touching a couple I/O ports on the ISA bus, which meant you could page this memory in very fast! Only a few milliseconds!
This was blindly fast compared to other stuff like trying to use the 8088 CPU to copy 4 KB in and 4KB out of some other memory solution. LIM spec was used primarily for DATA. One could theoretically move program code to LIM Spec memory, but this wasn’t the normal need.
The important thing was that it was a standard that an application could code against, and then the hardware makers could produce compatible hardware and know it would work with the important applications of the day.
Even better, a device driver was written to do the talking to the hardware stuff, once implemented, expanded memory aware applications could code to the device driver LIMSPEC API and the applications were now insulated from how the expanded memory managers implemented paging.
LIM spec was good stuff – but you had a purchase a very expensive card to make it do things.
Intel releases the 80286
Now, there is 16MB of genuine addressable memory (24 bits) – but DOS still knows how to use only 1MB. RAM is also becoming cheaper but still nobody could afford 16MB, it was common at the time for machines to come with 2MB of RAM. If only DOS could see higher than 640KB, this memory could do something useful and there is lots of it. 2MB is a huge increase!
What is EXTENDED memory?
EXTENDED memory is the genuine 80286 visible system memory that is present in the machine, but isn’t usable because DOS can’t see that high. It is the stuff above 1MB. It is just memory, but it is called “Extended” so that you know it isn’t DOS normal memory.
Note that “extended memory” as a concept is incorrect, its just memory and all the memory present in the machine is addressable by the processor, the fact that DOS and applications under DOS understand only the memory below 1MB is just a statement to the architecture at the time.
Hardware ISA bus implemented LIM spec solutions still selling well at this time and the standard API exists, it now became possible to use the 80286 addressable extended memory to implement LIM Spec! In concept, use a couple GDT selectors of the 80286 and you can page memory to and from EXTENDED memory to accomplish LIM Spec, and usable by DOS and most importantly, cheap! No longer need a hardware memory card. But, the 80286 was never an important chip, so these techniques never took off, the hardware solutions for LIMSPEC kept selling.
VCPI – Virtual Control Program Interface
Virtual Control Program Interface (VCPI) became a standard for running applications in the protected memory space of the 80286. Here you could load PROGRAMS into the 80286 high memory and run that application code! Quarterdeck with QEMM were the players here, but I’m going from a long ago memory.
You had to “come back to DOS” for OS functions or I/O with device drivers, but you could put gobs of program and data up high. I once wrote application programs under VCPI; it was a small part of the DOS programming world that most people did not understand. Getting I/O to work to external hardware required writing code to “switch” to and from protected mode. The 80286 was a stepping-stone to the right solution. Yes, it would be possible to write device drivers in the high memory space, but interrupt delivery was space of the VCPI and this was not a common thing.
DPMI (80386) – DOS Protect Mode Interface
The market progresses and the Intel 80386 CPU comes out, a CPU with all the things one expects in the modern world. It has selectors rather than segments, just like the 80286, but it also adds page-table based memory addressing and MVDM (multiple virtual DOS machine) support. To be fair, the 80286 also had selector-based addressing, but it was the 80386 processor addition of page tables that really made it possible for solutions to high memory addressing.
The execution of applications in high memory under DOS on the 80386 was governed by DOS Protect Mode Interface.
80386 and DPMI
PharLap was another player on the high memory execution front, and the definition of DOS Protected Mode Interface. This provided a standard means to run programs in the high memory space, but it also implemented LIMSPEC in software which performed every bit as efficiently as LIMSPEC in hardware memory expansion cards. More efficiently actually because the modern 80386 machines and their RAM were hugely fast compared to ISA bus implemented RAM and the page tables of the processor could be quickly programmed to change backing store for LIM Spec paging.
Running application up high required a memory manager to accomplish, but the same techniques needed to manage the high memory, could also be used to implement LIMSPEC. During boot, you could load “DOS” device drivers which was really the extended memory manager, which put the machine into protected mode, and implemented DPMI and LIMSPEC.
DPMI provided a standardized method to get programs and data into high memory spaces.
The biggest runtime for executing applications in this high space was DOS4GW.exe distributed by Watcom. PharLap was the main player in the memory manager space, with the PharLap DOS extender. Exactly the relationship between PharLap and Watcom, I do not know.
DPMI was quite successful.
Microsoft releases DOS 5
DOS 5 included HIMEM.sys and EMM386! The functions of LIMSPEC and DPMI were brought to the market with DOS, and no need to purchase an expanded memory manager, and within 6 months, PharLap folded.
- HIMEM.sys puts the CPU into “protect mode” — Implements DPMI
- EMM386.sys uses the services of HIMEM.sys to implement “Expanded Memory Management (LIM Spec) using the 80386 CPU
I am partially vague on what is implemented in HIMEM.sys and what is implemented in EMM386.sys. The functions varied over time or at least my understanding of them and there were other drivers involved.
Who provides the DPMI manager?
One of the PRIMARY THINGS of DPMI is that all DPMI applications can INCLUDE a DPMI memory manager in their binary. The rule is that if a DPMI manager has not yet been loaded – you BECOME the DPMI manager. If one is already there, you use its services and do not load your memory manager. This also became not absolute over time.
If the DPMI manager already running is “old” – you can perform a coup.
You can also package the executable separate from the memory manager.
Thankfully for the market, once DOS 5 included the DPMI manager, the need for application programs to carry one along went away and the world became a more stable place.
Windows 3.1 to Windows 95
The most successful DPMI application ever was Windows. Starting with Windows 3.0 Enhanced mode along with Windows 3.1 and eventually Windows for Workgroups 3.11. Over time, more and more OS code was moved from DOS space to high memory space, and eventually Windows 95 was born. Windows 95 while celebrated as a great new operating system, was really a very large, DPMI application running on top of DOS. Andrew Schulman documented this well in his book Undocumented Windows.
Eventually, Windows NT took hold as a “real OS” and once the world got to Windows XP, the days of DOS were behind us. Except, its 2022, and yes, DOS is still running.
DOS 5, 6, 7
As DOS matured, the management of the hardware memory gaps above 640KB became well managed. By this time, much hardware allowed its hardware memory address to be configured in software, making it easier to produce large “gaps” in the memory above 640KB and below 1MB. Careful placement of device drivers during boot into these holes allowed maximizing of DOS memory below 640KB and the memory managers of DOS made this organization possible. MEM.exe included with DOS made it possible to get a clear map of all the memory being used and a bit of careful adjustment could produce a very cleanly running machine with RAM available for DOS and applications maximized.
There are places in this blog where the information is “good”, but also likely incomplete. If you have more details, please add in the comments – and send me an email at joe at this domain. The blog system is set to not post comments unless approved, and I’ll need a little kick to move them to published. | <urn:uuid:a1a76124-6b6b-4b3b-8bb8-4af5749f79c0> | CC-MAIN-2022-40 | https://www.joenord.com/blog/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338280.51/warc/CC-MAIN-20221007210452-20221008000452-00125.warc.gz | en | 0.964632 | 2,707 | 2.6875 | 3 |
The General Data Protection Regulation (GDPR) imposes security and privacy regulations that apply to businesses that store or process European Union residents’ personal data. It enacts a broad range of measures to give data subjects control over their data and protect them from unauthorized exposure.
Encryption is a vital aspect of obtaining GDPR compliance. Encryption protects your organization so that in the event that data is lost, stolen, or compromised, there is a line of defense. Adding encryption as a layer of protection for your data strengthens your organization’s ability to protect that data in a way that complies with the regulation and provides assurance to your clients. Businesses with EU users and customers need to know what GDPR encryption rules mean for their data security and privacy efforts.
What Does The GDPR Say About Encryption?
The GDPR does not mandate specific technologies or implementations, so no rule says, “you must encrypt personally identifiable data.” However, GDPR Article 32 (1) states that data controllers and processors must implement appropriate technological and organizational measures to secure personal data. Encryption is suggested as a measure that can help businesses to achieve their GDPR compliance objectives.
Encryption is the best way to protect data, provided it’s used as part of a secure system. Encryption is often built into infrastructure hosting platforms, and effective encryption technology is available to all businesses at a minimal cost.
1. Assess Which Data Falls Under the GDPR
The first step is to discover which personal data your business stores, processes, or transmits. That includes knowing which data is in scope for the GDPR, where it’s stored, and the privacy and security measures the business uses to protect it. Ignorance isn’t a defense; businesses often breach the GDPR by failing to protect information they don’t realize contains personal data.
A Data Protection Impact Assessment (DPIA) can help businesses discover whether encryption is appropriate. A DPIA assesses data processed by an organization to determine whether it poses a risk under the GDPR. It considers the data’s nature, the level of risk, and the measures that could be taken to mitigate risk, including encryption. GDPR provides a template that can guide your organization through this process.
2. Develop GDPR Encryption Policies
Encryption policies should clearly describe how and when data processed by your organization is to be encrypted. Encryption policies help avoid mistakes caused by ad-hoc and inconsistent implementation.
Encryption policies supported by the organization’s leadership have two main benefits:
- They provide a foundation on which specific procedures can be based, allowing the organization to develop consistent GDPR encryption practices to achieve compliance objectives while meeting the varied needs of different systems and data types.
- They can mandate training requirements for relevant staff to ensure they know encryption policies, procedures, and responsibilities. Many data breaches occur because employees fail to follow encryption best practices by, for example, downloading personal data to an unencrypted portable drive or uploading it to an improperly configured cloud storage service.
3. Encryption, GDPR, and Data in Transit
Data is said to be in transit when it is moved between systems or components of a system. For example, data in transit might be information submitted by a customer in a web browser or data delivered to a third-party processor by a business. Data in transit is at particular risk as it travels over open networks outside the influence of the data controller or processor. Standard encryption measures to protect data in transit include virtual private networks (VPNs) or HTTPS encryption using TLS certificates.
4. Encryption, GDPR, and Data At Rest
Data at rest is often considered a lower risk than data in transit because security measures should prevent an attacker from accessing internal storage devices. However, software vulnerabilities, insider threats, and phishing attacks may allow attackers to circumvent network border protections and steal unencrypted data. If data is encrypted at rest using securely managed keys, the attacker gets nothing of value. Encryption at rest is part of a layered approach to data protection and GDPR compliance.
5. Understand GDPR Encryption Requirements
There are many ways to encrypt data, but some are more effective than others. As computing power increases and cryptography advances, older standards and algorithms become easier to crack. To comply with the GDPR, use up-to-date, well-tested cryptographic tools that conform to reputable standards. While the GDPR doesn’t specify tools and standards, businesses typically rely on cryptographic security standards such as FIPS 140-2 and FIPS 197 in concert with broader information security standards such as ISO 27001 Annex A.10.1.
GDPR Compliance with KirkpatrickPrice
KirkpatrickPrice provides a range of services that can help your business comply with the GDPR and other information security regulations, including ISO 127001 audits, SOC 2 audits, and compliance audits for other regulations and standards. Businesses seeking to improve GDPR compliance also benefit from security awareness training, penetration testing, and remote access security testing. | <urn:uuid:2d624f82-4ff9-4ac4-87e4-0ade9bd6cb96> | CC-MAIN-2022-40 | https://kirkpatrickprice.com/category/blog/gdpr/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030331677.90/warc/CC-MAIN-20220924151538-20220924181538-00326.warc.gz | en | 0.915526 | 1,033 | 3.1875 | 3 |
A new report from Privacy International paints a picture of a largely opaque, unaccountable system of surveillance technology used by government agencies online. So-called “cloud extraction” tools are largely unknown to the public and subject to relatively low levels of oversight, but are used regularly to analyze and extract private data from all of the major cloud services.
This new application of surveillance technology is concerning in its scope, and in terms of issues of consent. A search of a device is no longer a one-time event when these tools are used; the party being searched may now also be having any or all of their cloud accounts accessed, and may be subject to ongoing monitoring and repeated searches.
What is cloud extraction?
Cloud extraction works by lifting authentication tokens off of a mobile device or computer that government agencies have physical access to, for example during a search of a suspect.
Authentication tokens for various cloud services can remain active for weeks at a time, and in some cases are permanent. If the investigating agency can extract these tokens, they do not need to coerce the subject into giving up login information; if they already have login information, they can maintain ongoing access even if the subject later changes their password. This also allows them to circumvent most two-factor authentication (2FA) measures.
Cloud extraction tools such as GTEX, Cellebrite, UFED and KeyScout will automatically comb computers and devices for login credentials saved to cookies, browsers and other applications. These tools can not only snap up existing tokens, but in some cases also create new tokens from what they find.
What kind of data is accessed?
Cloud extraction tools can open the doors to files stored with nearly all of the major cloud services: Google Drive, Dropbox, Facebook, Slack, various web-based email services and more. These tools can also mine data from voice assistants such as Google Home and Alexa, and fitness trackers such as FitBit. But this goes far beyond simple access to files.
As the report points out, getting access to a Google account in this way also allows investigators to potentially track searches, location history, browser history, instant messaging records and more. Anything visible to the account holder when logged in is also available to the investigator – including encrypted messages.
Some of these cloud extraction services have also added facial recognition and matching capabilities. At least one tool, from Oxygen Forensics, claims to offer emotion recognition as part of the services.
This type of surveillance technology allows investigators to not just access a huge amount of existing information, but also to surreptitiously track targets on an ongoing basis. They can repeatedly access the cloud accounts to check on location data, new calls, new messages, and videos and pictures that the subject takes among other possibilities. Most subjects will not be aware that they are being tracked in this manner after the initial physical search occurs.
Troubling surveillance technology
In a U.K. poll commissioned by Privacy International, about half of users of mobile phones in the country are not aware of where their cloud data is stored, and nearly half believe that their mobile apps are not generating data that is stored in the cloud. About the same amount feel that they do not have a good understanding of how cloud based systems work.
Most people are likely not aware that these agencies have this level of surveillance technology or that it is legal for them to use it in this way. If they consent to a search of their device, they may not be aware of the extent of the cloud data they are granting access to or that they are effectively giving permission to continual tracking of their private online activity.
Camilla Graham Wood, solicitor at Privacy International states:
“We are only just starting to gain a modicum of transparency around law enforcement use of mobile phone extraction, yet there are new concerning technologies on the horizon such as cloud extraction, about which very little is known.
“Cloud extraction technologies give law enforcement the ability to access eye-watering amounts of highly sensitive personal data, not only about individuals, but also their friends, colleagues and acquaintances. Concerningly, such technology also allows authorities to deploy facial recognition tech across people’s media as well as the ability to conduct continual monitoring of an individual’s social media without them ever knowing.
“Much of this data is uploaded to the cloud, often without our knowledge, by the big tech companies. This risks making our personal data more vulnerable, not more secure. There is an urgent need for the companies who we entrust with our data to ensure they protect it from the tech which can be operated by unskilled operatives at the push of a button.
It is a matter of urgency that law enforcement act with a greater degree of transparency in relation to the new forms of surveillance they are using, and that laws which are designed to protect against abuses are updated.”
The cloud extraction firms offer this surveillance technology internationally and are located all over the world: for example Oxygen Forensics is based in the United States, Cellebrite is based in Israel, and Elcomsoft is headquartered in Moscow.
Privacy International has requested that 17 of the largest tech companies that this surveillance technology claims to have access to, such as Google and Facebook, to take a public position on its use. The group is advocating for greater transparency and legal safeguards, urging UK citizens to contact their local law enforcement agencies and ask about their cloud extraction technology and policies. | <urn:uuid:5cd60f6d-dc16-4aa4-b388-374f3140d445> | CC-MAIN-2022-40 | https://www.cpomagazine.com/data-privacy/privacy-international-report-on-cloud-extraction-programs-sheds-light-on-far-reaching-government-surveillance-technology/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334992.20/warc/CC-MAIN-20220927064738-20220927094738-00326.warc.gz | en | 0.947456 | 1,110 | 2.65625 | 3 |
The cutting edge of internet privacy and anonymity today is Tor, The Onion Router. Developed at the U.S. Naval Research Laboratory in the mid-1990s as The Onion Routing program, Tor came about as a next-generation implementation of the original project.
Tor is the most widely used software in the world for anonymously accessing the internet. Tor is made up of close to 7,000 relays and close to 3,000 bridges at the time of writing, all of which are operated by volunteers.
When properly connected to Tor, your web history, online posts, instant messages, and other communication forms cannot be traced back to you.
How does Tor work?
Tor is a network made up of thousands of volunteer nodes, also called relays. A relay is a computer inside Tor, listed in the main directory, that receives internet signals from another relay and passes that signal on to the next relay in the path. For each connection request (e.g. website visited) the path is randomly generated. None of the relays keep records of these connections, so there is no way for any relay to report on the traffic that it has handled.
The Tor network (or simply “Tor”) is made up of close to 7,000 relays and 3,000 bridges at the time of writing.
When you connect to the Tor network, say, through the Tor browser, all the data you send and receive goes through this network, passing through a random selection of nodes. Tor encrypts all that data several times before it leaves your device, including the IP address of the next node in the sequence. One layer of encryption is removed each time the data reaches another node until it reaches the final exit node, a process called onion routing. This means no one, not even the people running the nodes, can see the contents of the data nor where it’s headed.
A bridge is a hidden relay, meaning it is not listed in the main Tor directory of relays. These are provided for people who are unable to access Tor with the normal setup. This can be because the network they are using has a proxy (a sort of intermediary between the user’s computer and the internet gateway) that has been configured to block Tor traffic.
The last relay in the path is the exit node. The exit node is the only part of the network that actually connects to the server that the user is trying to access and is, therefore, the only bit that the server sees and it can only log the IP address of that relay.
Anyone who intercepts the data won’t be able to trace it back to an individual. At best, they can determine the entry or exit node, but never both. This makes it impossible to track the user’s activity and browsing history. All of the relays and bridges are run, believe it or not, by volunteers–people donating some of their bandwidth and computing power to expand Tor’s capabilities.
Tor is setup this way to allow an internet user to surf the web anonymously by hiding their internet address (IP address) from the website and search engines that they access via Tor and by hiding their internet traffic from anyone monitoring their end of the connection. An observer will only see that the user is connected to Tor, and not see any other websites or online resources being sent to the user’s computer.
Also, as another part of the overall network, Tor offers certain hidden services in the form of .onion sites and an instant messaging server. The .onion sites are websites hosted on Tor servers and hidden by randomly generating paths to them from “introductory points” in the network. This allows users to access the sites, but not pinpoint the location of the servers hosting them.
How do I get started with Tor?
The Tor browser is the easiest and most popular way to use Tor. It’s based on Firefox and works like a bare bones version of any other web browser. It’s pretty much plug-and-play; there’s no special tweaks or configuration to start browsing anonymously after the initial setup.
Tor is unique in that it provides anonymous access to both the clear net and the DarkNet. The clear net is what the vast majority of people are most familiar with. Anything on the World Wide Web that you can access with your non-Tor browser is considered the “clear net”. When you access websites on the clear net using the Tor browser, that website has no default way of identifying you because all of your internet traffic already went through Tor’s anonymity network, meaning the website can only see that you are visiting from a Tor exit node.
Accessing the DarkNet
Then there’s the DarkNet. Actually, there’s a few DarkNets, but the one accessible to Tor users is by far the most popular. The DarkNet is made up of hidden services, mostly websites that are not accessible in your day-to-day browser. These Tor hidden services can be identified by their .onion top-level domains appended to the end of their URLs.
Only people connected to the Tor network can access Tor hidden services, they do so anonymously, and they must also know where to look—Google and other search engines do not index .onion websites. Additionally, the Tor network also protects the identity of people who create Tor hidden services, so it’s nearly impossible to know who owns or operates a given .onion site.
This is why the DarkNet is particularly well-suited to crime and has a reputation for being the seedy underbelly of the internet. Although Tor has many great legal and legitimate uses as discussed above, the DarkNet that it houses is also where you’ll find marketplaces for illicit goods and services as well as blogs and forums for extremist groups.
Web surfing inside of the Tor browser is completely anonymous, but other activities on your computer are not. If you want to connect other applications and services to the Tor network, things start getting more complicated. You can find a list of other projects further down this article.
And although the Tor network will anonymize your connection to the internet, you’re still responsible for keeping yourself anonymous. A single misstep can put your anonymity in jeopardy. Check out our guide on achieving complete anonymity while online.
Keep in mind that because your information is being routed through a series of dedicated relays, the speed will not be what you are used to. It will be considerably lower than what you’re used to, actually. That is because the data is taking a much more circuitous path than normal and will be hampered by:
- The speed of the internet connections of the various relays along that path
- The amount of traffic on those relays
- The overall congestion of Tor at that particular time
- All the normal internet congestion that fluctuates throughout the day
The people at the Tor Project strongly urge their users to refrain from torrenting while using Tor. Torrent traffic, even through Tor, it is not anonymous and will only serve to (a) slow down the network for everyone and (b) completely obliterate any anonymity you may have enjoyed previously. (Torrenting is a way of sharing large chunks of data between peers using special software called a torrent client.) You’re better off using a secure VPN for torrenting, than Tor.
Your internet provider or the company where you work may block the use of Tor on their network. This is fairly easy to achieve as the network administrator just needs to blacklist Tor relays, which are publicly listed. In which case you will need to learn a bit about using bridges (hidden relays) or a VPN to gain access.
Tor’s main weakness is that it is susceptible to traffic analysis attacks, but this isn’t exactly news. The original design documents even outlined this type of vulnerability. That being said, Tor was never meant to stand up to a large scale attack of poisoned nodes.
Combining Tor with a VPN
The best way to stay private is to use both Tor and a VPN together. A VPN, or Virtual Private Network, is an encrypted tunnel on the internet from your computer to a VPN server. Any web surfing you do over a VPN looks like it is coming from the VPN server, and not from your computer.
By way of example, a user in New York, NY can connect to a VPN server in Los Angeles and any websites that she accesses will see her as being a user in California, not New York. Her ISP in New York will not be able to see any of her web traffic. In fact, all that the ISP will see is that she has encrypted data going back and forth between her computer and the VPN server that she is connected to. Nothing else.
A VPN does not make the user anonymous (although most commercial VPN services add a layer of anonymity by having users share IP addresses). It only provides an encrypted tunnel between your computer and the VPN server that you are connected to. Whoever operates the VPN server can technically snoop on your online activity, though reputable VPN services claim not to do this in their privacy polices.
Tor is used to make your connection to any websites that you visit anonymous. Your computer connects to a Tor entry node, and all of your traffic then hops through an intermediary relay to an exit node before finally reaching the destination website. Your ISP can see that you are connecting to a Tor entry node, which is enough to raise eyebrows in many cases.
The solution seems simple enough. Connect to a VPN, then open the Tor browser and surf anonymously in private. Your ISP can’t see that you’re connected to Tor, and your VPN can’t read the contents of Tor traffic.
This solution appears to have created a bit of confusion, though, as people wonder if it is better to connect to the VPN first or Tor.
The easier option is to connect to your VPN and then launch the Tor browser. The more complicated option is to do it the other way round. This requires a remote secure shell connection to the VPN server or a VPN client that has this capability built in as a configuration option. You will most likely see something along the lines of using Tor as a “proxy” for the VPN client. That should initiate the secure connection through Tor.
Tor over VPN
In the first setup, Tor over VPN, your ISP sees that your computer is sending and receiving encrypted information with a specific server, and that is all. The VPN server only sees that it is connected to a Tor entry node and that all of your surfing is being handled through that network. All of the websites on the other end of Tor do not see any personally identifiable information. NordVPN operates servers with Tor over VPN pre-enabled so you don’t have to use Tor Browser, but bear in mind that Chrome and other common browsers might give websites other identifying information.
VPN over Tor
In the second setup, VPN over Tor, your ISP sees that your computer is communicating with a Tor entry node. The VPN server gets your login information from a Tor exit node, which can be considered personally identifiable information, thereby breaking Tor’s anonymity. The websites that you connect to only see the VPN server’s address. AirVPN and BolehVPN are just two VPN providers that allow users to connect using VPN over Tor.
If your VPN provider does keep logs of activity, any benefit you might have received from the Tor connection is lost.
Be part of the Tor network
If you find Tor useful and would like to support it, you can volunteer to create and maintain your own Tor node for others to use. Check our tutorial on how to build your own Tor relay, but be warned that doing so comes with a high probability that law enforcement and/or your internet service provider will take notice and approach you with questions. You’ll need a spare server or computer lying around to act as the node.
You can also set up your own Tor hidden service, which essentially means you create a website or service only available on the DarkNet. Again, this could draw undue attention from your ISP and law enforcement, but there’s nothing inherently illegal about it. Check out our guide on how to set up a Tor hidden service.
Note that operating a Tor relay or hidden service requires quite a bit more technical expertise than simply browsing the web with the Tor browser.
Is Tor legal?
Even though using Tor is completely legal, simply connecting to it is cause for suspicion in some areas of the world. No one has ever gone to prison or even been fined for using Tor. But there have been reports of people being questioned by their ISP and local law enforcement regarding their browsing habits simply because the ISP’s logs show that they are connecting to Tor.
Those who run a Tor relay are cautioned that they will eventually be contacted by either their ISP or local law enforcement, maybe both, regarding some illicit activity online that is linked to the IP address used by their relay. To date, in all instances, simply informing these entities that the IP address in question is assigned to a node of the Tor network and that no logs are kept on that server has been sufficient.
Who uses Tor?
The number of people accessing the internet through Tor changes almost every day. Between August 2012 and July 2013 the Oxford Internet Institute and the University of Oxford calculated that “over 126,000 people access the Internet through Tor every day from the United States.” This translates to somewhere between .025% and .05% of internet users in the United States.
True to their claim, there are no stats available from the folks at TorProject concerning what their userbase is doing online or what sort of content they are accessing. It’s a safe bet that most of the users of Tor are legitimately just trying to maintain their anonymity for obvious reasons.
- Journalists who are protecting their sources,
- Whistleblowers trying to keep their jobs,
- Law enforcement officials trying to track down criminals and not blow their cover in the process,
- Corporate IT professionals conducting security testing on their own networks,
- And a bunch of other users that need the anonymity that Tor provides just to do their jobs.
For a more concise list of possible legitimate uses of anonymity, check out Tor’s description of their userbase.
Alternatives to Tor
One thing that is obvious is that Tor is not finished. In spite of that, some of you are already asking yourselves “Ok. What else is out there?” Well, you’re in luck. As it so happens there are other tools out there for those who wish to avoid being tracked. Here are a few of the most popular that are currently available.
Similar to Tor, I2P, or the Invisible Internet Project, uses a distributed network database and peer selection for anonymous traffic. Also susceptible to traffic analysis attacks, I2P does have some benefits over Tor. The peers are selected through continuous profiling and ranking performance. It is also small enough that few, if any, active blocks are in place to prevent access.
Unlike Tor, Freenet does not rely on dedicated entry and exit points. Instead, users connect to Freenet preferably through their friends’ computers. If you don’t have any friends on Freenet, you do have the option of connecting through strangers’ computers, but that is considered less secure than connecting to the computers of trusted friends. Freenet also is a file distribution service where encrypted files are stored on computer hard drives throughout the network. Due to the encryption, it is unlikely that a user would be able to determine what that file actually is.
Another onion routing-type anonymizer for web surfing, JonDoFox is a profile for Mozilla Firefox or Firefox ESR. The user’s computer connects to a series of Mix operators that anonymize the user’s web traffic and wrap it in several layers of encryption. Just like Freenet, the network size is considerably smaller than Tor’s. This is primarily due to their certification process. In order for you to become a mix operator, you must go through their certification process. Theoretically, this could lower the chances of an attacker sneaking in modified servers, but such a claim would have to be tested in a simulated environment.
GNUnet is a peer-to-peer file sharing tool that relies on large groups to obfuscate the identities of those that are attached to the group. An individual in the group is virtually indistinguishable from any other user by anyone but the initiator of the group.
Tor alternatives still in development
The following projects are still in development, but are working toward creating even stronger anonymity networks, but for more specific applications. Tor was created as a sort of generic, one size fits all solution for anonymous web use. These projects are more focused on specific applications of web use.
Aqua is a file sharing network designed to be completely anonymous, while Herd is an anonymous Voice over IP network. The designers are working up a means of stripping the metadata from the network traffic, which is the primary way of tracing a client and the server that client is communicating with.
Alpenhorn is the second iteration of Vuvuzela, named after the horn normally used at soccer matches in Latin America and Africa. Alpenhorn is an anonymous, metadata free chat program that can be scaled to millions of users, in theory. Expect a public beta in the near future.
If anonymity is more important to you than latency, then Dissent offers some of the strongest available anonymity. Due to the higher latency and low bandwidth, dissent is best used for blogging, micro-blogging or even IRC type communications. The way Dissent works is rather simple, but bandwidth heavy. When one client transmits anything, all the other clients transmit a package of the same size. Instead of using onion routing, Dissent is based on DC-nets, a dining cryptographers algorithm. Combine that with a verifiable shuffle algorithm and you end up with the most anonymous design being looked at by researchers today.
Anonymous file sharing is becoming more and more sought after. Riffle is yet another attempt at providing an anonymous way for a user to share files of any size. However, it is not meant as a replacement for Tor, mainly because file sharing over Tor breaks anonymity. Riffle is meant to augment Tor by providing Tor users with a truly anonymous way to share files, without choking the Tor network. Inspired by Dissent, Riffle also uses a shuffle algorithm but drops the DC-net cryptographic algorithm.
Riposte was inspired by Dissent, but focused on micro-blogging. Micro-blogging is currently the realm of Twitter, Pinterest and other such services where users update their “blog” with small snippets of information like quotes from famous people or requests for feedback or even requests to join networks. Riffle is designed to allow a user to micro-blog anonymously at the expense of internet speed. Following in the footsteps of Dissent, Riposte also uses the DC-net type setup for hiding the original transmission in a storm of transmissions of random data bits of the same size.
Finally, as an added bonus, here is a list of all the other projects in the works over at TorProject, all with an interest in maintaining internet privacy for any and all who wish to make use of their products. Some of these are rather obvious and user friendly, while others are more behind-the-scenes. A couple of different programming libraries are available for software developers to allow their products to communicate with The Onion Network.
The Tor browser
This is what most people use to access Tor. It’s very simple to acquire and use. The browser is actually a customized version of Mozilla Firefox, and therefore looks and feels like any other web browser. The customization is designed to leave no trace of your web surfing on the computer. Simply download the compressed file for your operating system, be it Windows, MacOS or Linux, extract it to its own folder, run the executable file inside that folder and surf to your heart’s content in complete anonymity. When you close the browser, all traces of your browsing are cleared from memory. Only your bookmarks and downloads are left behind.
.onion web sites
These are websites that are only accessible within the Tor network, and by knowing where to go. There are special search engines like Onion.city and Onion.to, as well as a host of others. Keep in mind, though that there are hoaxes, scams, and honeypots strewn throughout the DarkNet. Be wary of what you click on. There are also some very disturbing images available in there. You have been warned.
You can access the Tor network on your Android device using Orbot. Orbot creates a Tor proxy on your device so that all internet traffic from your device goes through the Tor network. That means that all the apps on your phone or tablet will have their traffic routed through Tor as well. Of course, some apps are designed not to be anonymous and will break the anonymity provided by the Tor network. True anonymity requires just a few steps to make sure tattlers are disabled or, at the very least, not running while you’re tapping into Tor. Remember to disable auto-sync and shut down any apps that automatically log you into an account, like Gmail, Yahoo!, Facebook, Twitter and the like.
To go along with Orbot, there is also a browser for Android devices that allows you to surf the net using Tor. However, this only applies to web surfing in a browser. All the other apps on your Android device will be communicating through normal lines of traffic without the benefit of anonymity provided by the onion router.
This might be the ultimate usage of Tor. It is a “live operating system” that is run either from a CD or a USB thumb drive or memory stick. Put this in a computer right before you restart. If the computer’s BIO is setup correctly, it will load Tails instead of the OS that is loaded on the computer’s hard drive. Perfect for using a computer that does not belong to you for surfing the web anonymously and leaving no trace of your browsing anywhere on the computer. The computer’s internal hard drive is not touched while the computer is running Tails and the computer’s memory is erased with each reboot. Also, any cookies or temporary internet files that are loaded into Tails are not recorded to the CD or thumb drive while in use so those are also lost as soon as the computer is restarted.
You were first introduced to Arm at the end of the “How to build your own Tor relay or node” article. Arm is a command line-based monitor for a Tor relay. It displays real-time information for a relay or bridge in the Tor network. This helps you keep an eye on your relay by providing statistics, metrics and health reports. You can learn how many Tor users have accessed Tor through your relay or how much of your available bandwidth is being used in support of Tor.
Atlas is a web application that provides information on the current status of the Tor network’s relays. Type the name of a relay into the search box at the top of the site and get a basic overview of its current status. Click on the relay’s nickname to get a much more detailed report along with an explanation of all the flags that apply to that particular node.
Used to change the way your data stream appears. This is yet another way of keeping you connected to Tor. Some entities have started blocking Tor traffic based on the traffic itself, not the IP address of the relay or bridge that is being used to connect to the network. Pluggable Transports change the look and feel of Tor traffic to appear to be normal, un-Tor-like traffic to escape detection.
While Atlas is a site showing the status of the Tor network, OONI is the site showing the status of censorship in the world today. It does this by probing the internet using a known good result and comparing that result to an unprotected, unencrypted result. Any changes in the results are evidence of tampering or censorship.
Onionoo is a web-based protocol that gets information relating to the current status of The Onion Network. This information is not in a human readable format. It is meant to act as a service for other applications like Atlas or Tor2Web.
As the name implies, this is where you get metrics relating to the Tor network like available bandwidth and the estimated size of the current userbase. Any researcher that is interested in any specific, detailed statistics about the Tor network can find it here, or submit a request for the metric that they are looking for.
A simulation of a network using the real Tor browser. This is most useful in a lab type setup when you want to see how Tor can affect your network, without impacting your real network. Perfect for experimenting with Tor and various other programs before allowing or implementing them on your local area network.
Grants non-Tor browser users access to websites running in Tor hidden services. The idea is to allow internet users the option of sacrificing their anonymity while still granting them access to information hidden inside the Tor network, while at the same time not sacrificing the anonymity of the websites that they are accessing.
An instant messenger client that uses the Tor network for all of its transmissions. Secure by default with cross platform capabilities, it is an ideal chat program for anyone wanting to stay secure and anonymous.
This is a programmers library for writing Python based applications that talks to or launches a Tor program. It contains all the utilities for accessing Tor’s circuits, streams, logging functions and hidden-services. | <urn:uuid:a7f7d48a-808d-41b1-b1c4-a1d073c9264e> | CC-MAIN-2022-40 | https://www.comparitech.com/blog/vpn-privacy/ultimate-guide-to-tor/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335396.92/warc/CC-MAIN-20220929225326-20220930015326-00326.warc.gz | en | 0.946378 | 5,421 | 3.296875 | 3 |
The term “5G” frequently appears in the news as many communications companies have started setting it up in major cities. What is 5G, and what makes it so controversial? 5G is the latest generation of cellular network technology. Essentially, a higher number next to the G means the user will be able to do anything they’d typically do with a cell signal — phone calls, surfing the web, streaming videos — only faster.
5G is an investment for the next decade, and in previous mobile transitions, we've seen most of the big changes happening years after the first announcement. The first 4G phones in the US appeared in 2010, but the 4G applications that changed our world didn't appear until later.
Snapchat came in 2012, and Uber became widespread in 2013. Video calls over LTE networks also became big in the US around 2013. Because the 5G transition is so complicated, and because we've been having a pandemic for two years, this time the shift may take even longer. Scientists in Finland who helped develop 5G technology say that it may be 2027 before we see the robotics, smart objects, and augmented reality proposed
As 5G continues to replace 4G networks, it's crucial to know the 5G pros and cons if you're considering implementing the network in your business.
5G networks use a system of cell sites that divide their territory into sectors and send encoded data through radio waves. Each cell site must be connected to a network backbone, whether through a wired or wireless backhaul connection. 5G changes the way data is encoded and offers many more options to carriers in terms of airwaves to use.
5G networks use a type of encoding called OFDM, which is similar to the encoding that 4G LTE uses. The new system opens up "high-band," short-range airwaves that didn't work with 4G technology but 5G can run on any frequency, leading to three very different kinds of 5G experiences—low, middle, and high.
All wireless communications carry information through the air via radio frequencies or spectrums. 5G is no different, though it operates on higher radio frequencies called millimeter waves. The millimeter waves are less muddled with other wireless communication signals, enabling 5G networks to transmit data significantly faster.
5G will also use multiple antennae and smaller transmitters on buildings and various street features, like park benches, to increase capacity and signals. Doing so will allow the network to support more devices than 4G.
5G even has a feature called network slicing, which allows users to create subnetworks to specialize the connection for specific tasks. Network slicing reduces the risk of overloading a network so users can benefit from the high speeds. For example, a hospital may slice its 5G network by departments or floors to further maximize its capabilities.
Every few years, a new generation of cellular networks becomes available, bringing with it a slew of improvements to the speed at which we communicate. Unfortunately, there are also some growing pains as cell carriers and network distributors adapt to the new technology. As with previous generations of cellular networks and technology, there are advantages and disadvantages to the 5G network.
One of the most significant advantages of 5G network technology is blazing-fast speed. According to one study, 5G will save users about 23 hours every month in social media, gaming, music and movie downloads and streaming. This speed is possible because the increased bandwidth allows more network dedication to each smart device.
Latency is the interval between the moment when a device sends data to a receiver and the moment when the receiver can use the information. For example, you can measure the delay between a user clicking a link and the site loading in latency. 5G latency will be significantly less than that of 4G, allowing users to connect to each other and the internet in real-time.
A wireless network's bandwidth is essentially its user capacity. 5G networks can support and connect a significantly higher number of devices at once. This space for more devices also allows the network to provide faster speeds.
Higher speeds and increased bandwidth will likely open new technological doors for many industries. 5G is ideal for connecting smart devices beyond phones, like drones, AI technology and improved sensors in machinery. For example, driverless cars can become a reality, and storage clouds can become more powerful for securely storing sensitive data like medical records.
Of course, there's no such thing as the perfect technology.
As companies roll out 5G, it will not immediately be available everywhere. Big cities will see it first, if they haven't already, and rural areas may have to wait several years for full coverage. New technologies also will likely experience hiccups, which could further limit the coverage.
Unfortunately, trees and tall buildings obstruct the high-frequency radio waves that carry 5G connection, thus decreasing the network's broadcast distance. Ultimately, this means more cellular towers will be necessary to achieve the extensive coverage we expect with 5G networks. It's possible to place small transmitters on buildings to increase coverage, though this could take a while to implement.
As with any emerging technology, 5G networks will face new cybersecurity risks. The software used to manage the network might be more vulnerable to hackers. Additionally, with more devices connected, there are more opportunities for cybercriminals to attack.
It will likely take several years to achieve effective 5G coverage across all urban and rural areas. While the possibility of putting smaller transmitters on buildings and other city spaces could provide coverage for cities sooner, the relative lack of tall buildings in rural areas means it will take significantly longer for their residents to receive coverage. To compensate, companies must construct new cell towers to reach all rural areas of the world.
The most notable distinction between these two generations of cellular networks has to do with speed and capacity. While 4G was innovative when it launched in 2010, technology continues to advance and requires more efficiency, giving 5G advantages over 4G. 5G networks' speed and bandwidth capabilities set it apart from 4G. For example, while 4G connections have a standard latency of around 20 to 30 milliseconds, 5G networks can reach below 10 milliseconds of latency.
5G has the potential to be 100 times faster than 4G, with a top theoretical speed of around 20 Gbps and current, real-world speeds from 50 Mbps to 3 Gbps. The so-called low-band 5G is somewhat faster than 4G with a performance of around 50-250 Mbps. The fastest version of 5G, called high-band 5G, is the version that reaches 3 Gbps.
Additionally, there is a difference between 4G and 5G network architecture. 4G networks rely on large cell towers to provide coverage for users. 5G networks aim to use smaller, widely dispersed cell transmitters to provide a reliable connection to more users. While 5G will try to make up for the areas where 4G was lacking, it will likely take some time to work out all the kinks.
Our fiber department splices cables that are currently used in 4G networks.
Someday, they'll be splicing fibers to be used in 5G connections.
5G is already beginning to impact technology in some industries and applications like smart homes. As the 5G network evolves, it will play a critical role in the Fourth Industrial Revolution. Though these possibilities are still at least four or five years away, 5G will allow for effective augmented and virtual reality in the medical industry, smart factories and eventually smart cities and agriculture.
5G technology could help production operations in the manufacturing industry become more flexible and efficient while enhancing safety. This would enable manufacturers to build “smart factories” that rely on automation, augmented reality, and IoT. With 5G powering large amounts of IoT devices and sensors around the factory, artificial intelligence can be integrated more deeply with operations.
Overall, 5G's impact on society could provide value to citizens' well-being, enhanced infrastructure, innovation and sustainable manufacturing. While it could take several years to fully implement, and will almost surely have some downfalls, 5G will continue to impact future technology and technological innovation for years to come.
Like 3G and 4G before it, 5G will soon come to shape the way we interact with each other.
As you build a 5G network within your community, office or business, ensure you're getting quality products when you choose Multilink. A reliable network requires effective equipment. As a leader in the telecommunications industry, we have the products and expertise to get you started. From adapter panels to fiber assemblies, our specialists can help you find what you need or create a custom build for your unique situation. Contact us today to find out how we can help implement 5G in your company.Back to Multilog | <urn:uuid:74ac6e75-15a7-448e-bc39-2d64379dd27a> | CC-MAIN-2022-40 | https://www.gomultilink.com/blog/multilog/the-pros-cons-and-potentials-of-5g | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337446.8/warc/CC-MAIN-20221003231906-20221004021906-00326.warc.gz | en | 0.952037 | 1,806 | 3.296875 | 3 |
The key difference between a successful organization and a not-so-successful one could lie in the way they document and share their organizational knowledge. When documenting your existing procedures, you need to make sure your team members are all on the same page in order to bring consistency to your processes and meet all critical dependencies. This is where SOP documentation can play a vital role in the success of an organization.
In this blog, we’ll explore the significance of SOPs and how they can be used to overcome business obstacles.
What is an SOP document?
A standard operating procedure (SOP) refers to a written document that contains step-by-step instructions on how to perform certain processes. When there are no documented processes, employees tend to do things their own way, resulting in process inconsistency and leaving a lot of room for error.
SOPs bring standardization to complex tasks so that they remain consistent and error-free across the organization. For instance, regulatory demands are so complex in today’s world where security is the utmost priority. Meeting these demands consistently without SOP documentation is extremely difficult.
What are the goals of establishing SOPs?
The main goal of SOP documentation is to ensure organizational consistency in all processes and outcomes. By ensuring consistent outcomes, organizations can boost efficiency, adhere to regulatory demands, resolve issues, increase profits and more.
Organizations also establish SOPs for the following reasons:
- Business continuity: Businesses cannot afford to stop for any reason. The easy availability of standardized information ensures business continuity even if an employee departs or goes on vacation.
- Employee onboarding/offboarding: Standardized processes are vital when onboarding or offboarding an employee. SOPs can help establish them and bring more reliability to the process.
- Compliance adherence: Compliance regulations demand that critical processes are documented and periodically reviewed. This provides additional security to the company and prevents compliance penalties.
- Knowledge retention: When employees leave an organization, they take the acquired knowledge along with them. With standardized SOPs, you can retain this knowledge and ensure proper knowledge management.
- Higher quality: When people work on complicated tasks without knowing how to work, it leads to inconsistency and errors. With SOPs, information is readily available on how to perform tasks. This reduces errors in the process and improves quality.
Why are SOPs important?
An organization should be prepared to deal with all kinds of eventualities. They cannot do so without writing down their institutional knowledge. Companies should define everything including what they are doing, how they are doing it, who is supposed to be working on what, etc. Without documenting their critical processes, companies cannot grow their business in the desired way.
Here’s a list of consequences that are bound to happen without SOPs:
- Halt in operations: Although employees may come and go, your institutional knowledge should not change along with changes in your workforce. Without documentation, a company’s operations risk coming to a stop when key employees leave.
- Irrecoverable losses: From natural disasters to cyberattacks, tragedies may strike a business at any moment. SOPs have disaster recovery plans, standardized guidelines, business continuity information, etc., to ensure quick recovery from a disaster. Companies that don’t document any of these items will find it extremely difficult to recover.
- Lack of growth: Organizations innovate by documenting their growth at every step. Without documentation, they have to spend all their time and resources reinventing the wheel. Ultimately, this will compromise their growth.
What are the three types of SOP formats?
While there are no hard and fast rules preventing you from developing your own internal SOP document, there are three types of SOP documents that can make life easy for you. Most organizations choose one of the following types:
As the name implies, a step-by-step SOP breaks down a task or a process into a list of different steps with detailed instructions for each step. The steps detailed in the SOP should be simple and clear enough for anyone to follow.
The hierarchical SOP is similar to the step-by-step format but more detailed. It uses a top-down approach and breaks down complicated tasks into a list of steps. However, each step contains many more additional details as required by the task. In a step-by-step SOP, you can list out steps as 1, 2, 3, etc. In the hierarchical model, the steps contain additional steps, such as 1a, 1b, 1c, etc., depending on the complexity of the task.
This is mainly used for complicated tasks where detailed instructions are required to complete them successfully.
When your process needs a visual representation, it is better to use the flowchart SOP. The format is usually adopted when there are multiple outcomes in a process. You can assign conditional logic to illustrate different outcomes and organize your procedures accordingly.
What should an SOP document contain?
Considering the main goal of SOP documentation is to ensure consistency, it is important to develop a specific format with all the key elements. However, there are no rigid rules on what format to use and an organization may adjust these elements as per their requirements.
Proper SOP documentation should likely address each of the following in one way or another.
It is always better to start your SOP document with a title page. This should contain the following information about the document:
- Title of the task or process
- An identification number for your SOP
- Date of creation and revision
- Division or branch of the organization applicable to the SOP
- Name of the creator with approval signatures from relevant stakeholders
Table of contents
This can be skipped if your document is only a few pages long. For long documents, you must include a table of contents to locate relevant information easily.
Before you go about describing the entire process, you must include some preliminary information that covers the scope and purpose of your SOP. You can convey it by including the following details:
- Purpose & goal: There must be a reason behind creating an SOP document. By defining the purpose and goal of the document, you are providing a high-level overview of the process you are discussing in the document.
- Roles and responsibilities: This section provides straightforward information about all the stakeholders involved in the process. You can also use this section to define the specific roles of the individuals involved.
- Related documents: In this section, you define the documents related to the SOP for some additional information. This directs users to the right documentation if they need more information about a specific process or task.
- Resources: The successful completion of any task or process requires a variety of tools, technology and resources. By documenting this information, you can swiftly gather the materials required and get to work right away.
- Safety information: The SOP document should also include all the cautionary information about the use of resources and the steps involved in a process. All safety information must be provided clearly in simple language. You also need to mention who to contact if additional safety information is required.
Once you have documented all the preliminary information, you must outline the procedures involved in completing the process. This is probably the most important part of your SOP document since it provides vital information on the actual steps to be followed in completing a task. You can use sequential steps to describe the procedures with additional sub-steps when required.
Since this is the critical part of your SOP documentation, you need to make it as clear and detailed as possible. Some SOP documents even include diagrams and decision trees to illustrate the procedures clearly. The end user who reads the document must be able to grasp it quickly without much back and forth in clearing doubts.
This section concerns how your team members are performing with respect to the SOP document. You can do this on a case-by-case basis and see if your employees are adhering to the procedures mentioned in the SOP document. If you see any issues in the way things are done, you must rectify the issues and make the necessary adjustments going forward.
References and definitions
There should also be a section with references to all the source documents from which information is collected for this SOP document. Additionally, you can also create a glossary section with definitions of all important terms.
Finally, it is better to include a revision history with information about all the revisions made to the SOP document. This should include details such as who has made the revision, when it was made, why was it made, etc. You can see the different versions of the SOP document here and revert to an older version if required.
What are the benefits of SOP documentation?
SOPs can benefit any organization irrespective of its size. It helps companies stay organized, saves time and ensures the easy accomplishment of key tasks. Let’s look at some of the major benefits of SOP documentation.
Promotes consistency and reduces errors
The most important benefit of SOP documentation is its ability to ensure consistency in the process and minimize human errors. Since the processes are clearly outlined in a step-by-step manner, it removes guesswork from your tasks. As a result, all task outcomes remain consistent and error-free.
Aids in onboarding and employee training
SOP documents are also developed for training and onboarding new employees. This helps ensure that all your new hires get the same level of training for the same amount of time.
In organizations where employees work in hazardous situations, SOPs can be created to establish workplace guidelines on employee safety. Organizations can also use similar best practices to follow cybersecurity hygiene and meet compliance demands.
Preserves organizational knowledge
A report by Bridge Group has estimated that the average tenure of an employee in an organization has fallen to 1.8 years in 2021. If you have created SOPs and documented all key processes, you don’t have to worry about losing your institutional knowledge when an employee leaves.
Companies that fail to adhere to relevant compliance laws often end up paying huge fines and penalties. This can be avoided by creating SOPs with the exact procedures to be followed.
Who is responsible for writing SOPs?
Depending on the industry you operate in and the department for which the SOP is being created, there may be different people responsible for creating the SOPs. For instance, process engineers take up the role of SOP creation in manufacturing companies. Similarly, for SOP documentation concerning cybersecurity, IT managers or security consultants may do the writing tasks.
Irrespective of who writes the SOP document, it is better to get inputs from stakeholders in different areas. Also, the SOP document must be reviewed by all concerned stakeholders once the creation process is done.
What are SOP templates?
Organizations can use SOP templates to standardize their procedures. When you have standardized templates across the whole organization, you can get started with your SOPs quickly without worrying about fonts, margins, outlines and other trivial things. You can either create your own SOP templates or use the common ones available online.
SOP documentation with IT Glue
IT Glue is a robust documentation platform with powerful features. IT Glue enables documentation of SOPs alongside your key IT infrastructure information, including passwords, assets, licenses, etc., so everything is contextual and a click away. You can also have access to a library of pre-built SOP templates, many of which are based on ITIL and industry best practices. With these out-of-the-box templates, you can start documenting right away.
IT Glue’s SOC 2-compliant documentation platform features an immutable audit trail, multifactor authentication and next-generation password management engine. You can also integrate these features with your SOP documentation.
To learn more about how IT Glue can help with your SOP documentation, request a demo. | <urn:uuid:3dbdbeaf-d206-4e62-a2e5-fbe5992a7251> | CC-MAIN-2022-40 | https://www.itglue.com/blog/sop-documentation/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337446.8/warc/CC-MAIN-20221003231906-20221004021906-00326.warc.gz | en | 0.927359 | 2,503 | 2.71875 | 3 |
Kingston University’s foray into 5G technology for ‘smart’ farming would allow intelligent drones to monitor crops and livestock and could well help to thrust the agricultural world further into the ‘third green revolution’.
“It would involve using drones with on-board visual and infrared cameras, supported by a computer vision system allowing them to detect when an animal is sick, trapped, injured or missing and also to monitor, count and control crops and spot signs of disease or weeds,” explained Professor Paolo Remagnino from Kingston University’s Robot Vision team (RoViT).
5G goes green
At present, a drone would need massive amounts of processing power on board to fly over farmland for several hours at a time. And yet, this would be hugely impractical because of how much it would weigh. The project team will investigate how 5G could help to provide consistent, high-speed connectivity.
Using the speed of a 5G network, the team at Kingston will develop a way for all of that processing to be done, in real time, in a control station somewhere on the farm (such as a barn), or anywhere else, instead of embedded within the drone.
During the project, the RoViT team will employ a computer vision system that uses 5G enabled technology to provide a real time assessment of the visual data, using a form of machine learning called deep learning. This uses artificial neural networks to mimic the way the brain makes connections between pieces of information without being specifically programmed and would allow the drones to learn normal behavior and appearance of livestock and crops and identify irregularities.
A righteous revolution
In this new green revolution, farmers move will beyond the previous revolutions of breeding and genetics, and increasingly rely on the new digital era to guarantee food security. And, this move may well be more a necessity than a luxury. In a 2017 report [PDF], Beecham Research warns that the way farmers produce their food must radically change to feed the growing world population of the future.
Indeed, the United Nations Food and Agriculture Programme has noted that global food production must increase by 70% by 2050 to meet the demands of a growing population.
“This means that to optimize crop yields and reduce waste, the agriculture and farming industries will need to rely heavily on IoT and machine-to-machine (M2M) technologies moving forward,” explained report co-author and Beecham chief research officer Saverio Romeo. “GPS services, sensors and big data will all become essential farming tools in the coming years and are clearly set to revolutionize agriculture.”
While the report concedes that precision agriculture cannot solve all problems – unpredictable weather events, and a demanding and changing legislative environment to name but a few – it is optimistic that smart farming can help farmers control aspects of farming better and optimize results.
Beecham adds: “Sensor-based technologies and decision support systems play a vital role in the post farm-gate supply chain. This includes the detection of food fraud, identifying and dealing with bacterial and other contamination, mitigating spoilage and food waste, cold chain monitoring and meeting the growing need for traceability from farm to the consumer.”
The research by Kingston University is part of a £2.1 million ($2.7 million) government-funded project called the 5G Rural Integrated Testbed (5GRIT), a partnership of nine SMEs and universities developing and testing innovative uses of 5G. The project span a range of rural applications, from agriculture to tourism as well as connecting poorly-served communities, and will involve using TV white space technology to test the potential for shared spectrum radio to deliver 5G services to rural areas.
5GRIT is supported by Department for Culture Media and Sport through the 5G Testbeds Program.
Other proofs of concept
TM Forum’s Catalyst proof-of-concept program is also studying the use of drone technology in agriculture, looking specifically at how communications service providers can play a role in delivering drones as a service.
“There’s a very positive social benefit with precision farming,” says Dr. Lester Thomas, chief systems architect at Vodafone Group. “It’s eco-friendly; you’re reducing chemical usage; and you’re increasing the yield of crops.”
It’s also an innovative revenue opportunity for network operators. While some farmers are already using drone technology, Vodafone wants to take it step further, automating drone flights as a platform-based service that uses 5G network slicing to ensure redundancy and quality of service. Check out this article to lean more.
Written by Arti Mehta, editor of TM Forum Inform | Original story here
The TM Forum’s Digital Transformation Asia event is coming to Kuala Lumpur on November 13-15. Disruptive.Asia is an official media partner with the event. All the details you need are right here. | <urn:uuid:1f0d9a1a-749c-47e6-8d99-9815f03b6a54> | CC-MAIN-2022-40 | https://disruptive.asia/5g-smart-farming-green-revolution/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337853.66/warc/CC-MAIN-20221006155805-20221006185805-00326.warc.gz | en | 0.924055 | 1,019 | 3.375 | 3 |
Cyber attacks on space systems are escalating—just as space technology is becoming increasingly critical for national security, defense, and our way of life. Defense and intelligence agencies can accelerate cybersecurity with space IT solutions combining open systems architecture, model-based systems engineering (MBSE), and the power of a proactive approach.
Threats to Space Systems
The U.S. Government has identified some of the most critical threats to space systems:
- Rival nations have a sophisticated knowledge of satellite command and control and space distribution networks.
- Many active space systems were designed before space was considered a contested domain.
- More and more systems are connected to space assets, increasing the attack surface.
- Commercial space ventures create more players and complexity.
Space Policy Directive-5 mandates cyber protections for space systems, supporting the national security goal of maintaining U.S. leadership and freedom of action in space.
Operational Technology Meets IT
The White House’s executive order on improving cybersecurity for IT and operational technology (OT) underscores a critical aspect of space systems: They’re a hybrid of IT and OT.
“These systems are at the intersection of the cyber and physical worlds,” says Kevin Coggins, vice president and head of Booz Allen’s positioning, navigation, and timing (PNT) business. “You must protect the threat surface of the physical systems as well as their network connections.” | <urn:uuid:d415dc96-00df-4ef5-9196-a0e37479df9c> | CC-MAIN-2022-40 | https://www.boozallen.com/markets/space/space-cyber-defense-an-adaptive-proactive-approach.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337853.66/warc/CC-MAIN-20221006155805-20221006185805-00326.warc.gz | en | 0.926976 | 298 | 2.703125 | 3 |
What Is Decentralized Storage?
On the client’s end, decentralized distributed storage works precisely equivalent to conventional distributed storage choices like Amazon S3. However, they put your data away on many conveyed Nodes across the globe. Rather than your records getting stored on a larger server farm, helpless against blackouts and assaults.
How The Decentralized Cloud Works
Distributed storage comprises an enormous, appropriated network with many Nodes spread across the globe, autonomously claimed, and worked to store information.
Every piece of your information dwells on these nodes
A Node is just a hard drive or a capacity gadget somebody possesses secretly. Every Node Operator gets paid to store documents for customers and gets repaid for their transmission capacity.
Consider it like this: You have a 10 TB hard drive and are just using 1 TB00703;
You could join as a Node Operator and store bits of customers’ documents on your hard drive using your unused space. Contingent upon the number of records you keep and how often the information gets recovered, we’d repay you as needs be.
So, why decentralize it?
The fundamental problem with incorporated suppliers like Amazon S3 is that each information dwells in colossal server farms. On the off chance that a piece of Amazon’s organization goes down, you will not have the option to get to your information, best-case scenario.
Your info could be for all time lost or harmed.
Colossal server farms are additionally defenseless against programmers, as seen on different occasions. With decentralized distributed storage, start to finish encryption is standard on each document. Each record gets scrambled on a client’s PC before it’s transferred, broken into pieces, and afterward spread out to uncorrelated Nodes across our organization.
Encryption keys make it practically unthinkable for your information to get compromised or taken.
Besides, colossal server farms cost a massive load of cash and take plenty of assets to work. You don’t need to spend money working on a server farm, but you can use individual, exclusive gadgets. Reserve funds get passed on to clients.
But what about data loss or poor actors on the network?
Let’s quickly consider the Tardigrade network’s decentralized design. Tardigrade has 99.99999999% document strength, and it splits each record into 80 pieces. With 30 parts needed to reconstitute a data record, it would take 51 hubs getting disconnected simultaneously for your document to be lost. Complete records get recovered at lightning speed by downloading the quickest 30 of 80 pieces.
Perhaps you’re familiar with how torrents function? It’s a similar idea.
There’s no main issue of disappointment, guaranteeing your information is consistently accessible. Because each document transferred to Tardigrade gets parted into 80 pieces and encoded before getting put away, disconnecting one Node will not affect any records.
The real significance of the decentralized design lies in how a Node Operator doesn’t know what records get stored on their Node.
Whether a Node administrator needs to get to your records, they have a tiny shard or piece of that document. They would need to find, without a doubt, 30 different Nodes to reconstitute a record, and those documents additionally get encoded.
Is it secure?
For this question, “Storj” is what we like to call “trustless.” What does this mean?
It implies you don’t need to put your confidence in any single association, cycle, or framework to keep the organization running. You don’t have to stress about your information since we could not get to it regardless of whether we needed to.
Tardigrade is private and secure, and documents get encoded start to finish before transferring to organizations. They guarantee that nobody can get to information without approval.
A document on Tardigrade is exceedingly difficult to access without the legitimate keys or consent. Since everything gets scrambled locally, your information is in a proper sense in your grasp, and no other person’s. After records get encoded, they get parted into more modest pieces that are indistinct from one another.
A regular record gets parted into 80 pieces, of which 30 can reconstitute the document.
Every one of the 80 pieces is on an alternate drive, with various administrators, power supplies, organizations, topographies, and so on. For instance, there are at present 171 million documents on our Tardigrade administration.
To think twice about a single record, the programmer would initially need to find 30 of its pieces among the 171 million on the organization, making a highly elusive little thing. Then, at that point, they would need to decode the document, which is amazingly troublesome, if certainly feasible, without the encryption key.
Then, at that point, the programmer would need to rehash this to get to the document. | <urn:uuid:b69c31ee-ac62-41ba-b606-290a453e6612> | CC-MAIN-2022-40 | https://enteriscloud.com/decentralized-cloud-storage/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334644.42/warc/CC-MAIN-20220926020051-20220926050051-00526.warc.gz | en | 0.923806 | 1,023 | 2.5625 | 3 |
Optical satellite communication is otherwise called Free-space optical communication (FSO) that utilizes an optical communication innovation which utilizes light proliferating in free space to remotely transmit information for telecommunication or PC organizing. "Free space" indicates vacuum, space, air, or anything comparative. This stands out from utilizing solids, for example, optical fiber link or an optical transmission line. The innovation is helpful where the physical associations are ineffective because of staggering expenses or different contemplations. Mainly the global market hold by North America, optical satellite communication market principally determined by quick development in the telecommunication part. The multiplication of the Internet, just as telecommunication traffic around the world, is driving the market development around the world. The worldwide Internet traffic saw a CAGR of over 25% amid the period from 2013 to 2016. With rising Internet clients alongside telecommunication traffic over the world, a few private players have begun putting resources into optical satellite communication market looking for plausible approaches to oblige the rising demand for constant communication.
The global optical satellite communication market size is expected to reach at nearly USD 2.92 billion by 2026. The global optical satellite communication market is likely to expand at a CAGR of more than 31.1% from 2019 to 2026.
Governments of different nations are putting resources into various optical communication tasks to grow high information rate joins. For example, in recent past, NASA consented to an arrangement with BridgeSat Inc. to build up space optical communication framework for the office. Also, The European Space Agency works optical information hand-off administrations known as EDRS. The office propelled the Copernicus earth perception venture with five LEO satellites, which offer optical communication for information transmission. Likewise, different favorable circumstances, for example, colossal balance transmission capacity, tight shaft difference, less power and mass prerequisite, high directivity, unlicensed range, and high security make optical satellite communication a promising possibility for people to come (5G) arrange as it requires a high information rate (rapid and transfer speed limit) which drives the requirement for high limit backhaul joins.
Different favorable position of remote communication, for example, lower sending cost, simple development of system topology, and adaptable upkeep of working systems have expanded the number of portable clients all inclusive. For example, the quantity of cell phone clients internationally is required to achieve 6.7 billion by 2022 up from 4.3 billion every 2017. Consequently, this expanding number of cell phone clients has expanded the demand for rapid information transmission. In addition, expanding sight and sound administrations, for example, video-on-demand, sound on-demand, and shared have expanded the requirement for high system data transfer capacity. Furthermore, expanding reception of different propelled methods, for example, machine to machine communication and the Internet of Things likewise drives the demand for high information rate which has quickened the optical satellite communication market all around.
Developing demand for continuous communication from practically all end-client businesses putting gigantic weight on present system framework. This closes promising demand for optical satellite communication to empower transmission of high transfer speed traffic signals for a territorial or worldwide reason. Furthermore, the rising selection of optical satellite communication benefits in the media part is driving the general market development comprehensively. Developing demand for live communicating of amusement and games, business promotions, live broadcast of different continuous occasions and news is set to stimulate the further development of optical satellite communication market.
Based on the component segments, the market is portioned into the transmitter, beneficiary, modulator, demodulator, and others. The transmitter fragment held the predominant position in the market and is required to hold its strength all through the conjecture time frame. Expanding demand for fast information exchange from satellite-to-ground is one of the main considerations in charge of the extension of the fragment. High information rates have expanded the demand for consistent R&D in optical satellite communication components.
The transmitter segment commanded the global optical satellite communication market representing multiple/fourth of the worldwide market esteem. Currently, major of the modules in IoT driven world outfitted with area administrations. Furthermore, the expansion of associated vehicles exhibiting promising development of route administrations. Quick infiltration of route work in shrewd advanced hardware and associated vehicles among other area-based working frameworks, therefore driving the transmitter segment growth.
Based on application segments, the surveillance and security application segment held the prevailing position in the market and is projected to hold its strength all through the conjecture time frame. The high selection of satellite innovation by military and government security offices for surveillance and security have quickened the market in this application. For example, as per the Satellite Industry Association, in 2015, around 110 satellites were utilized for military reconnaissance tasks, up from 101 satellites in 2014. The prerequisite for continuous information and high information rates have quickened the demand for optical communication in observation and security applications.
The optical satellite communication market growth is additionally upheld by the expanding utilization of brilliant home machines, Internet of Things (IoT) applications and constant communication gadgets among others. More than 2 Bn gadgets other than cell phones anticipated to being associated utilizing cell innovation by 2023. This incorporates rural fields and hardware, apply autonomy, and coordination among others. Optical satellite communication multiplication in IoT applications accepted to altogether improve continuous availability, lessen operational expenses and upgrade mechanical practicality. With the Internet of Things (IoT) demonstrating its essential job crosswise over practically all legislature just as the business part, telecommunication application set to proceed with the pattern as the biggest application of the market.
North America market held an overwhelming position in the market and is projected to hold its predominance all through the estimated time frame. The U.S. also, Canada are the real driving motors of this locale. High spending on space ventures from government collections of nations in North America is a central point driving the market in the area. For example, as per Space Foundation, in 2017, the U.S. represented 57% of worldwide government space spending. In 2019, NASA's Laser Communication Relay Demonstration venture intends to dispatch an optical information transfer satellite framework into the geostationary circle.
Likewise, the huge nearness of market players in nations of North America is additionally a noteworthy explanation behind the predominant position of the district in the worldwide market. Further, Asia Pacific assessed to be the quickest developing market for optical satellite communication over the next years. India, China, and Japan are among the rewarding markets having promising penetration of optical satellite communication market. These nations hold enormous potential for market significantly credited to expanding computerized satellite telecom, Internet administration, satellite R&D studies and satellite propelling among a few others.
Market Segmentation by Component:
Market Segmentation by Application:
Market Segmentation by Region:
The market research report on “Optical Satellite Communication Market (By Component: Transmitter, Receiver, Modulator, Demodulator; By Application: Backhaul, Earth Observation, Enterprise Connectivity, Last Mile Access, Research and Space Exploration, Surveillance and Security, Telecommunication, Tracking and Monitoring) - Global Industry Analysis, Market Size, Opportunities and Forecast, 2019 - 2026” offers detailed insights on global optical satellite communicationmarket segments with market dynamics and their impact. The report provides insights on global optical satellite communicationmarket by component, application, and major geographic regions.
One of the major strategies embraced by worldwide players is an agreement with a small/medium scale organizes service providers of networks. This causes the organizations to swell their communication arrange achieve, along these lines obliging progressively a number of supporters. Key cases incorporate into January 2018, ATLAS Space Operations, Inc. also, Xenesis, Inc. declared the agreement to create progressed optical satellite communication arrange intended for high information stream from the space. Likewise, In July 2018, Maxer Technologies Ltd. declared the procurement of Neptec Design Group. This obtaining expected to help the Maxer Technologies Ltd. to convey mechanical frameworks and extended the arrangement of arrangements, fortifying the brand to catch the development in North America just as worldwide space investigation markets and quicken headway into growing space sections. Further, Ball Aerospace and Technologies Corporation declared the partnership with Honeywell Corporation to plan and create high unwavering quality and elite optical communication information joins. This partnership expected to empower Ball Aerospace and Technologies Corporation to convey high limit, reasonable and high data transfer capacity network to customers over the globe.
Major companies contributing the global optical satellite communication market are Analytical Space Inc., ATLAS Space Operations, Inc., Ball Aerospace & Technologies Corp., BridgeSat Inc., HisdesatServiciosEstrategicos S.A., Laser Light Communications Inc., Maxar Technologies Ltd., Mitsubishi Electric Corporation, Mynaric AG, and SITAEL S.p.A | <urn:uuid:fcaa2f38-fa41-4ab9-97e4-e3563923f097> | CC-MAIN-2022-40 | https://www.acumenresearchandconsulting.com/optical-satellite-communication-market | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334644.42/warc/CC-MAIN-20220926020051-20220926050051-00526.warc.gz | en | 0.891975 | 1,806 | 2.734375 | 3 |
What is Tokenization
Tokenization replaces a sensitive data element, for example, a bank account number, with a non-sensitive substitute, known as a token. The token is a randomized data string that has no essential or exploitable value or meaning. It is a unique identifier which retains all the pertinent information about the data without compromising its security.
A tokenization system links the original data to a token but does not provide any way to decipher the token and reveal the original data. This is in contrast to encryption systems, which allow data to be deciphered using a secret key.
How Data Tokenization Works
Tokenization, in relation to payment processing, demands the substitution of a credit card or account number with a token. The token has no use and is not connected to an account or individual.
The 16 digits primary account number (PAN) of the customer is substituted with a randomly-created, custom alphanumeric ID. The tokenization process removes any connection between the transaction and the sensitive data, which limits exposure to breaches, making it useful in credit card processing.
Tokenization of data safeguards credit card numbers and bank account numbers in a virtual vault, so organizations can transmit data via wireless networks safely. For tokenization to be effective, organizations must use a payment gateway to safely store sensitive data.
A payment gateway is a merchant service offered by an e-commerce application service provider that permits direct payments or credit card processing. This gateway stores credit card numbers securely and generates the random token.
Payment Tokenization Example
When a merchant processes the credit card of a customer, the PAN is substituted with a token. 1234-4321-8765-5678 is replaced with, for example, 6f7%gf38hfUa.
The merchant can apply the token ID to retain records of the customer, for example, 6f7%gf38hfUa is connected to John Smith. The token is then transferred to the payment processor who de-tokenizes the ID and confirms the payment. 6f7%gf38hfUa becomes 1234-4321-8765-5678.
The payment processor is the only party who can read the token; it is meaningless to anyone else. Furthermore, the token is useful only with that single merchant.
Tokenization vs Encryption
The main difference between tokenization and encryption is that tokenization uses a ‘token’ whereas encryption uses a ‘secret key’ to safeguard the data.
A core issue with data encryption is that it is reversible. Encrypted data is designed to be restored to its initial, unencrypted state. The safety of encryption is reliant on the algorithm used to protect the data. A more complex algorithm means safer encryption that is more challenging to decipher.
All encryption is, however, essentially breakable. The strength of your algorithm and the computational power available to the attacker will determine how easily an attacker can decipher the data. Encryption is thus better described as data obfuscation, rather than data protection. Encryption makes it more difficult to access the original information protected within the encrypted data, however not impossible.
The PCI Security Standards Council and similar compliance organizations treat encrypted data as sensitive data because it is reversible. Organizations are thus required to protect encrypted data.
Unlike encryption, tokenization of data cannot be reversed. Rather than using a breakable algorithm, a tokenization system substitutes sensitive data by mapping random data, thus the token cannot be decrypted. The token is a placeholder, with no essential value.
The true data is kept in a separate location, such as a secured offsite platform. The original data does not enter your IT environment. If an attacker penetrates your environment and accesses your tokens, they have gained nothing. Thus, tokens cannot be used for criminal undertakings.
The PCI and other security standards do not require organizations to safeguard tokenized data.
Benefits of Tokenization
Tokenization can provide several important benefits for securing sensitive customer data:
- Enhanced customer assurance—tokenization offers an additional layer of security for eCommerce websites, increasing consumer trust.
- Increased security and protection from breaches—by using tokenization, businesses do not have to capture sensitive information in their input terminals, keep it in internal databases, or transmit the data through their information systems. This safeguards businesses from security breaches.
- Data tokenization improves patient security—organizations can use tokenization solutions for scenarios covered under HIPAA. By substituting electronically protected health information (ePHI) and non-public personal information (NPPI) with a tokenized value, healthcare organizations can better comply with HIPAA regulations.
- Tokenization makes credit card payments more secure—the payment card industry needs to comply with extensive standards and regulations. Tokenization solutions provide a way to protect cardholder data, such as magnetic swipe data, primary account number, and cardholder information. Companies can comply with industry standards more easily, and better protect client information.
PCI Tokenization: Easing Compliance with Tokenization
The Payment Card Industry Data Security Standard (PCI DSS) ensures PAN data is protected by all organizations that accept, transmit, or store cardholder data. Failure to comply may result in fines and loss of brand authority.
Tokenization helps companies achieve PCI DSS compliance by reducing the amount of PAN data stored in-house. Instead of storing sensitive cardholder data, the organization only handles tokens, making for a smaller data footprint. Less sensitive data translates into fewer compliance requirements to comply with, which may lead to faster audits.
How Imperva Leverages Tokenization for Security and Compliance
Imperva’s security solution uses data masking and encryption to obfuscates core data, so it would be worthless to a threat actor, even if somehow obtained.
We offer a holistic security solution that protects your data wherever it lives—on-premises, in the cloud, and in hybrid environments. We help security and IT teams by providing visibility into how data is accessed, used, and moved across the organization.
Our security approach relies on multiple layers of protection, including:
- Database firewall—prevents SQL injection and similar threats, while assessing for known vulnerabilities.
- User rights management—tracks the data movements and access of privileged users to identify excessive and unused privileges.
- Data loss prevention (DLP)—monitors and tracks data in motion, at rest, in cloud storage, or on endpoint devices.
- User behavior analytics—creates a baseline of data access behavior and uses machine learning to isolate and alert on abnormal and potentially dangerous activity.
- Data discovery and classification—discloses the volume, location, and context of data on-premises and in the cloud.
- Database activity monitoring—monitors relational databases, data warehouses, big data, and mainframes to produce real-time alerts on violations of policy.
- Alert prioritization—Imperva uses AI and machine learning technology to examine the stream of security events and prioritize the most important events. | <urn:uuid:41435371-5d74-4057-9328-39a3fb7d2d34> | CC-MAIN-2022-40 | https://www.imperva.com/learn/data-security/tokenization/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334644.42/warc/CC-MAIN-20220926020051-20220926050051-00526.warc.gz | en | 0.888967 | 1,449 | 3.21875 | 3 |
Career Paths: Data Science
Data is everywhere in the digital age and is part of every piece of modern technology. It allows us to structure different websites, edit images, and prioritize our social media feed, whether we want it to or not. And it can be your career! We aren’t saying you need to work in structuring everyone’s social feed algorithms though.
We’ve been examining different career paths for the past two months, and now we’re on to our final segment. We’re bringing the microscope in close to see what Data Science career paths are available to you.
What do Data Science careers entail?
Data Science experts are in high demand! Think Google, Apple, and Microsoft. Every product they release has complex data systems built into the software. Even the simple iPod from the early 2000’s required data scientists to play a part in its creation.
The role of data scientists includes intense code development, design, and analysis. They also build out platforms and applications to assist with data analytics performance. Some of the most common skills needed to fulfill these responsibilities include knowledge of Python and Amazon Web Services (AWS).
What are some Data Science career options?
Today, organizations are relying more on big data and experts to swiftly collect, analyze, and leverage data for their operations and projects. LinkedIn alone has more than 120,000 data science jobs available. Not to mention every data science job comes with an enticing salary as well as opportunities for career development.
This role is responsible for using data and acquiring information on specific topics. Data acquisition often involves conducting surveys as well as interpreting and presenting data to corporate decision-makers. Key skills for this job include Python, SQL, Tableau Software, Data Modeling, and Data Mining. Mastery in these skills may even result in a salary increase.
Data Scientists are responsible for mining complex data and designing new ways to manage data using models. They work closely with statistical data and rely on skills like Data Modeling, Data Mining, Big Data analytics, and Python.
This role develops and translates computer algorithms into code. These engineers are also responsible for maintaining, organizing, and identifying trends in large data sets. Data engineers rely on knowledge in areas such as Apache Spark, AWS, Big Data analytics, Data Warehousing, and Python.
This profession involves the design, structure, and maintenance of data. Data management requires advanced skills with computers and computer language proficiency in SQL and XML. Other skills that are crucial to this role include Big Data analytics, Data Warehousing, and Data Modeling.
Senior Software Engineer
How does INE help?
We have Data Science training down to a science! We provide hands-on activities and training material that allow you or your team to practice working with classification algorithms and version control in GitHub in a risk-free setting. Our in-depth solutions give you a chance to try out your computing skills before you need to execute in business critical scenarios.
You can work through our Data Science Learning Paths in addition to virtual projects and quizzes. Learning Paths are designed to help you keep track of your course progress and see what is left on your journey to become a Data Science master. Courses cover everything from machine learning fundamentals to software development with python, and more. Regardless of the topics you choose to focus on, our experts have carefully designed the information to ensure clarity and an engaging experience.
Our activities, quizzes, and instructor-led training videos have been designed with your success in mind and will help you master the keys to web scraping, unit testing, as well as feature engineering techniques for a well-rounded training experience.
Learn more about INE Data Science training today! | <urn:uuid:7dcc48f1-2fa1-48a6-976e-ce54d13e6377> | CC-MAIN-2022-40 | https://ine.com/blog/career-paths-data-science | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335004.95/warc/CC-MAIN-20220927100008-20220927130008-00526.warc.gz | en | 0.931711 | 829 | 2.828125 | 3 |
Study Reveals Grammatical Structure Severely Undermines Password Strength
A new research study on the effectiveness of passwords may make you reevaluate the strength of your passwords.
A study conducted by researchers at Carnegie Mellon University’s Institute for Software Research shows that what we once thought of as strong passwords may not be so strong any more. The key, they have found, is grammar.
Results from the study found that passwords that incorporated grammatical structure, no matter how long, were easier to crack than comparatively shorter passwords with no apparent structure. To conduct their tests, the researchers developed what they describe as a grammar-aware password-cracking algorithm which uses separate dictionaries for each element of a sentence (i.e. one dictionary for verbs, one for nouns, etc.) to identify any structure within the password. With this cracking tool, they found that passwords with grammatical structure significantly reduced the time needed to crack a password due to the narrowing of possible word combinations and sequences.
A password such as “Th3r3 can b3 only #1!” was cracked in 22 minutes even though it used special symbols and letter substitutions. Compare to the password “Hammered asinine requirements” which took over three and a half hours for the researchers’ algorithm to crack, simply because it lacked a grammatical structure.
It’s studies like these that make us reconsider whether our supposed “strong” passwords are truly as strong as we think they are.
Original article by Jaikumar Vijayan.
Read the full story here. | <urn:uuid:1c055032-93b7-42b7-8867-1508ae075402> | CC-MAIN-2022-40 | https://my.infotex.com/study-reveals-grammatical-structure-severely-undermines-password-strength/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335004.95/warc/CC-MAIN-20220927100008-20220927130008-00526.warc.gz | en | 0.945409 | 327 | 2.9375 | 3 |
Mainframes continue to be viable for enterprise computing due to their inherent capabilities for supporting the most critical business functions. When it comes to supporting large volumes of data, high-speed transaction processing, and secure computing the mainframe continues to excel. The Fortune 1000 relies upon mainframes to run their business, and of course, much of those business applications are written in COBOL.
Indeed, a large number of mainframe applications are written in COBOL and new COBOL code is being written all the time. Various sources claim that there are between 100 billion to 800 billion lines of COBOL code in use today with over a billion new lines of COBOL being programmed each year.
Although it is wise to doubt the precise accuracy of such hyperbolic, large numbers, it is undoubtedly true that there is a lot of COBOL code running on mainframe systems. And that code powers important enterprise applications. In other words, COBOL is everywhere, so it should be a significant target for improvement.
Any pervasively implemented technology that has been in use for a long time, such as COBOL, should be reviewed over time for ways to improve and modernize it. One such option is to convert the COBOL code to a more modern language, such as Java. There are three primary reasons why you should consider converting some of your COBOL workload to Java.
• Cost Reduction
• On-going Support
The first COBOL programs can be traced back to 1960, which is a long time ago in the realm of computers and technology. By 1970, COBOL had become the most widely used programming language in the world. (source: Beyer, Kurt. Grace Hopper and the Invention of the Information Age. MIT Press. 2009. ISBN 978-0262013109.)
The current portfolio of existing COBOL programs spans a 60-year period. Technology changes rapidly, especially computer hardware and software, as well as development techniques and requirements. So, it stands to reason that a lot of COBOL code “out there” could benefit from a more modern approach.
For example, procedural, structured programming using waterfall design was the primary development paradigm for at least the first half of COBOL’s life. In the 1990s, object-oriented (OO) programming became the dominant form of development, along with the agile development methodology.
COBOL is a procedural programming language, whereas Java is object-oriented. More programmers today know and use OO techniques instead of structured techniques. So, it is not just that more programmers know and understand Java, it is that they know and understand a completely different form of application programming than is used with COBOL.
Optimizing ancient procedural COBOL code into the new OO, Java paradigm can make a lot of sense. Furthermore, consider the types of applications that are well-suited for COBOL. The term COBOL is an acronym for Common Business Oriented Language. The language was designed to support business functions like reporting, number crunching, and processing data. This does not mean that COBOL cannot perform other types of processing; it can. Just that some types of programs may be better developed using another language.
Java is an object-oriented programming language that is suitable for multi-purpose computing, with the added benefit of being portable across multiple hardware platforms. The ability run the same program on different computers (as long as a Java Virtual Machine is available for the platform) is one of the reasons that Java is one of the most popular languages for new development today.
So, optimization of your aging COBOL code into a more modern paradigm is one reason to consider converting some of your programs to Java.
Reducing the cost of their mainframe environment is a continual struggle for most large organizations. According to a recent BMC survey and white paper, IBM monthly license charge (MLC) costs, which are already greater than 30 percent of a typical company’s overall mainframe budget, are increasing annually by 5 percent to 11 percent.
Converting COBOL to Java can have a significant impact on your MLC cost. Java programs can run on the IBM zIIP specialty processor, whereas COBOL program code cannot. And work that runs on a zIIP instead of a general-purpose CP is not included in the MSU metrics used to calculate your monthly MLC software charges.
By converting existing COBOL programs to Java, much of that Java workload can be redirected to run on a zIIP. And that means it is not contributing to your rolling four hour average (R4HA) or your monthly software bill. Therefore, converting at least some of your COBOL programs to Java can contribute to cost savings.
The third reason it can make sense to convert COBOL programs to Java is to provide better on-going support for your application code. Support includes things like problem assessment, troubleshooting, code maintenance, code control, application integration, and other code-related duties.
It can be difficult to find skilled COBOL programmers. Most universities no longer teach COBOL and procedural coding. So many organizations hire programmers with other skills and train them on COBOL (if they can convince a prospect to accept a COBOL job offer).
On the other hand, it is easier to find programmers with Java skills. And there are more Java coders being trained every day in universities and trade schools everywhere.
Another consideration is the aging COBOL workforce. According to a study by zippia, most COBOL programmers are over 40 years old. As these programmers age and retire, COBOL will age with them, and unless something changes dramatically it will become even more difficult to hire skilled COBOL programmers.
On the other hand, Java is a newer, thriving language. It is object-oriented, taught in most college computer science programs, and one of the world’s most popular current programming languages.
The Bottom Line
By now you should be convinced that there is some merit in converting some of your COBOL applications to Java. Of course, no organization is going to immediately convert all their COBOL code even if there are significant potential gains to be had. The risk involved in converting all your mission-critical COBOL applications would likely be too large to undertake in one big step.
Nevertheless, strategically identifying and converting COBOL to Java is a tactic that can reap substantial benefits. Optimizing your development environment, reducing cost, and improving support are compelling reasons. Still manually converting from COBOL to Java can be tedious and time-consuming. Fortunately, CloudFrame provides several options for automating the conversion of COBOL to Java that can reduce the time, effort, and human error involved in conversions.
Have you considered the potential benefits of converting some of your older COBOL applications?
Find Out More
To learn more, check out our ebook,
The COBOL Modernization Opportunity. | <urn:uuid:8e3cd738-ff87-490b-916b-0881d2173293> | CC-MAIN-2022-40 | https://cloudframe.com/three-reasons-it-makes-sense-to-convert-some-of-your-cobol-to-java/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335276.85/warc/CC-MAIN-20220928180732-20220928210732-00526.warc.gz | en | 0.937334 | 1,460 | 2.53125 | 3 |
Most Commented Posts
A recent article published in Nature has revealed an entirely new phase of matter that has the potential to act as long-term quantum information storage.
Researchers at the Flatiron Institute’s Center for Computational Quantum Physics in New York ran an experiment that subjected a quantum computer’s qubits to “quasi-rhythmic laser pulses” based on the Fibonacci sequence, demonstrating a way of storing quantum information that is less prone to errors. A Fibonacci sequence is a series of numbers where the next value in the sequence is calculated by adding the two preceding numbers (for example, 0, 1, 1, 2, 3, 5).
By shining a laser pulse sequence inspired by the Fibonacci numbers at atoms inside a quantum computer, the physicists created a new phase of matter that has never been observed before. The phase has the benefits of two time dimensions.
The researchers said that information stored in the phase is far more protected against errors than with alternative setups currently used in quantum computers. As a result, the information can exist for much longer without getting garbled – an important milestone for making quantum computing viable, said study lead author Philipp Dumitrescu.
Dumitrescu spearheaded the study’s theoretical component with Andrew Potter of the University of British Columbia in Vancouver, Romain Vasseur of the University of Massachusetts, Amherst, and Ajesh Kumar of the University of Texas in Austin. The experiments were carried out on a quantum computer at Quantinuum in Broomfield, Colorado, by a team led by Brian Neyenhuis.
A typical crystal has a regular, repeating structure, like the hexagons in a honeycomb. A quasicrystal still has order, but its patterns never repeat. Quasicrystals are crystals from higher dimensions projected, or squished down, into lower dimensions. Those higher dimensions can even be beyond physical space’s three dimensions.
For the qubits, Dumitrescu, Vasseur and Potter proposed in 2018 the creation of a quasicrystal in time, rather than space. Whereas a periodic laser pulse would alternate (A, B, A, B, A, B, etc), the researchers created a quasi-periodic laser-pulse regimen based on the Fibonacci sequence. In such a sequence, each part of the sequence is the sum of the two previous parts (A, AB, ABA, ABAAB, ABAABABA, etc). This arrangement is ordered without repeating. It is also a 2D pattern squashed into a single dimension.
The researchers tested the theory using Quantinuum’s quantum computer, pulsing laser light at the computer’s qubits both periodically and using the sequence based on the Fibonacci numbers. The focus was on the qubits at either end of the 10-atom lineup. Dumitrescu said: “With this quasi-periodic sequence, there’s a complicated evolution that cancels out all the errors that live on the edge. Because of that, the edge stays quantum-mechanically coherent much, much longer than you’d expect.”
Toward error-free quantum computing
Meanwhile, in a recent blog post, IBM described its quantum error mitigation strategy as “the continuous path that will take us from today’s quantum hardware to tomorrow’s fault-tolerant quantum computers”.
Over the last few years, said IBM, its researchers have developed and implemented two general-purpose error mitigation methods, called zero noise extrapolation (ZNE) and probabilistic error cancellation (PEC). The ZNE method cancels subsequent orders of the noise affecting the expectation value of a noisy quantum circuit by extrapolating measurement outcomes at different noise strengths.
According to IBM, recent theoretical and experimental advances have shown that PEC can enable noise-free estimators of quantum circuits on noisy quantum computers. IBM has forecast that its approach to error mitigation – which is analogous to how early classical computers developed – will enable it to develop quantum computers with more circuits, which means greater power to solve hard problems.
One such hard problem is predicting the weather, which involves processing complex non-linear differential equations run on classical computer architectures.
The recent hot spell across Europe has shown everyone the importance of accurate weather forecasts. BASF has begun to explore how proprietary quantum algorithms developed by Pasqal could one day be used to predict weather patterns to support its digital farming business. By using parameters generated by weather models, BASF will be able to simulate crop yields and growth stages, as well as predict drift when applying crop protection products.
Advanced weather and climate modelling are usually run on classical computers using physics informed neutral networks (PINN). According to Hyperion Research, 5% of global high-performance computing (HPC) investments are focused on weather modelling.
Rather than rely on HPC, Pasqal aims to solve the underlying complex non-linear differential equations in what it calls “a novel and more efficient” way by implementing so-called quantum neural networks on its neutral atom quantum processors.
John Manobianco, senior weather modeller at BASF’s Agricultural Solutions division, said: “Leveraging Pasqal’s innovation for weather modelling validates quantum computing’s ability to go beyond what can be achieved with classical high-performance computing. Such transformational technology can help us prepare for climate change impacts and drive progress toward a more sustainable future.”
These algorithms will only be viable once researchers and quantum computing companies have improved error handling. However, some of the techniques used to solve problems can be run today on classical computing architectures.
For instance, in a recent podcast, Bloomberg CTO Shawn Edwards discussed why he believes mainstream quantum computing is many years away. Although a lot of progress has been made on the underlying science, Edwards said that some of the more useful things to come out of quantum computing are quantum computing-inspired algorithms. He said the quant teams at Bloomberg have been looking at improving certain algorithms based around quantum computing.
Such quantum-inspired algorithms may be the bridge that enables the mass adoption of quantum computing. Even if error correction is still years away, the research to improve error handling and the development of quantum-inspired code may encourage more IT heads to plan ahead and develop an IT strategy that incorporates quantum computing. | <urn:uuid:321bf259-6e2f-492f-bab0-179a58da3a4d> | CC-MAIN-2022-40 | https://ihowtoarticle.com/quantum-researchers-advance-error-handling/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335276.85/warc/CC-MAIN-20220928180732-20220928210732-00526.warc.gz | en | 0.917419 | 1,332 | 3.4375 | 3 |
Everyone’s looking for ways to stay safer now, and many are looking to technology for help. Facial recognition is one of the tools that government agencies and businesses are exploring—both to slow the spread of the new coronavirus and to protect data from cybercriminals trying to profit from pandemic-related disruption. However, it’s not always immediately clear what facial recognition can do reliably now, what it can’t do and what it can do when people are wearing masks, sunglasses and other items that obscure the face.
For example, facial recognition has become a biometric marker for some digital payment methods like Apple Pay to authenticate users. And many companies are looking for ways to use facial recognition to reduce the need to touch surfaces like payment terminals and locks with fingerprint scanners. But facial recognition systems that consistently recognize all kinds of people correctly in all kinds of settings have yet to be brought to market, let alone when those people are wearing many types of face masks to protect their health.
Before planning a safety or security program that includes facial recognition, it’s important to understand which applications are fully fledged, which are still developing and where the potential pitfalls are.
What facial recognition can do now
There are some situations where facial recognition can be a practical safety and security tool. One is contactless data and payment security for accountholders. For example, Apple users can unlock their phones and make purchases through Apple Pay with the company’s FaceID biometric tool. And FaceID appears to be moving beyond the Apple-only silo. Google Drive now recognizes Apple’s biometrics, including FaceID, for logins to its iOS app.
As account takeover fraud trends upward, facial recognition could be a useful tool to fight it, although it’s not without challenges and security issues. For example, what happens if someone’s image is used to commit fraud or another crime? Passwords can be changed. Faces, for the most part, cannot.
Although masks are required or recommended in many workplaces, it can be easy for workers who aren’t yet used to them to forget, especially if they’re not in a public-facing role. Facial recognition can help businesses maintain safety in areas where masks are required, such as in warehouses and on factory production lines. This capability requires mask-recognition software that’s written to alert managers when it detects an unmasked face.
Although it’s a controversial application, facial recognition can help police departments identify people who are wanted for crimes or have been reported missing. That’s the use case London’s Metropolitan Police department cited when it switched on live facial recognition systems in January. More widespread systems that were already in use in China and Russia are now helping authorities enforce regional lockdowns and quarantines of specific individuals. For example, in China, facial recognition cameras with thermal sensors can identify citizens who are running a fever. They can also spot people who aren’t wearing their required face mask.
What facial recognition can’t do yet
Because facial recognition maps many points on a face to get a match, masks, glasses and other items that cover part of the face can affect accuracy. For example, Apple is adding an automatic changeover to passcode entry for FaceID users when it detects a mask. This may seem like a step backward in terms of security, but it helps users maintain safety by leaving their mask in place instead of touching and possibly contaminating it. Especially in densely populated cities and high-risk settings like hospitals and nursing homes, the fewer times users have to touch their masks or faces, the safer they are.
Although some facial recognition providers—and some governments—say the tools they use work even when subjects are masked, it’s clear that’s not always the case. One Chinese facial recognition vendor claims a 95% accuracy rate on masked subjects, compared to 99.5% for unmasked people. However, the combination of a mask and sunglasses thwarts the system.
That lack of consistent performance has caused problems with everyday tasks for people in China who are wearing masks. Some have reported problems with everything from logging into their smartphones and bank accounts to getting into their apartment buildings. Clearly, even in countries that put a high premium on facial recognition accuracy, the tools aren’t perfect. That can lead to misidentification by employers and law enforcement, which can cause real problems for people.
Issues that facial recognition needs to address
In addition to accuracy concerns, researchers have found that facial recognition AI can be biased, skewed by data inputs that don’t reflect the full range of human identity. Studies have found that such systems are more likely to misidentify transgender and nonbinary people, women and people of color than white men.
Facial recognition technology also raises ethical quandaries related to surveillance and legal issues related to privacy. Privacy watchdog groups and ethicists are concerned that increased use of facial recognition now may lead to increased surveillance after the pandemic crisis has passed.
However, facial recognition may also face legal barriers to wide adoption, especially in Western countries. For example, some legal experts say that in the U.S., a patchwork of “inconsistent and complex” local and state laws about the use of biometrics could slow down its implementation at scale.
For now, it’s clear that facial recognition can help protect individual and public health in limited scenarios, like making a contactless payment, accessing buildings without touching a keypad or scanner and enforcing mask policies in the workplace. However, facial recognition right now is not a comprehensive solution for going contactless or for enforcing public health rules, especially when people are masked to protect themselves and others from Covid-19. Instead, facial recognition is one tool among many that can deliver safety benefits if it’s used responsibly. | <urn:uuid:ba20dc29-62d0-4189-b5c6-ce442b6157df> | CC-MAIN-2022-40 | https://www.cpomagazine.com/cyber-security/what-facial-recognition-technology-can-do-for-safety-right-now-and-what-it-cant-do-yet/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335424.32/warc/CC-MAIN-20220930020521-20220930050521-00526.warc.gz | en | 0.947938 | 1,215 | 2.84375 | 3 |
The data architecture defines the data along with the schemas, integration, transformations, storage and workflow required to enable the analytical requirements of the information architecture. A solid data architecture is a blueprint that helps align your company’s data with its business strategies. The data architecture guides how the data is collected, integrated, enhanced, stored, and delivered to business people who use it to do their jobs. It helps make data available, accurate and complete so it can be used for business decision-making.
Data architecture is important for many reasons, including that it:
- Helps you gain a better understanding of the data.
- Provides guidelines for managing data from its initial capture in source systems all the way to information consumption by business people
- Provides a structure upon which to develop and implement data governance.
- Helps with enforcement of security and privacy.
- Supports your BI and DW activities, particularly Big Data.
Data architecture has grown far more sophisticated than its hub-and-spoke roots. We recommend an analytical data architecture, which is a hybrid model.
How we've helped others with their data architecture needs:
- Revitalized the bottlenecked data architecture of a large financial firm, supporting their needs for trading across worldwide exchanges and over one billion trades per day.
- When configuring custom machinery, a heavy equipment manufacturer was forced to use paper documents because data could not be shared between the manufacturing plants, engineering departments and retail outlets. Our data architecture solution helped obtain, synchronize and update data from the various sources so they could develop the specifications, configure the products and deliver them to the customers more quickly, efficiently and cost-effectively. | <urn:uuid:8ceb6e8d-ab9a-4bf2-9b56-af4ac5fc4f76> | CC-MAIN-2022-40 | https://athena-solutions.com/services/data-architecture/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335609.53/warc/CC-MAIN-20221001101652-20221001131652-00526.warc.gz | en | 0.939426 | 339 | 2.53125 | 3 |
Both edge computing and artificial intelligence (AI) have continued to gather attention over the past few years as mobile and Internet of Things (IoT) technologies become increasingly adopted by a vast number of industries around the world.
These new technologies have also created new problems and challenges for those looking to implement and benefit from the advances and developments of the fourth industrial revolution.
Edge computing is a computing technique used to move decision making closer to the source of data (the edge) and artificial intelligence is an area of computer science that looks to create intelligent machines and also includes sub-fields such as machine learning.
Individually, these two technologies have been shown to work incredibly efficiently and both show huge potential for future development, however, combining the two for use within IoT systems could result in a match made in IoT heaven. In this article, we’ll explain how edge computing and artificial intelligence are being developed to work together and put forwards some possible use cases that could be applied to AI at the edge.
So, let’s jump straight in.
Edge Computing & Artificial Intelligence
There are various ways in which edge computing and artificial intelligence could be used together, however, certain limitations with each of these technologies will need to be overcome in order for this partnership to be effective.
Traditionally, artificial intelligence technologies require a vast amount of computational and processing power to run some of the more complex machine learning algorithms, for example, making them impractical for deployment in locations distant from the cloud servers they are mostly based in.
Conversely, most edge computing technologies are currently being designed to operate in low-power environments with little to no connectivity, making the integration of artificial intelligence systems into their architectures unpractical.
However, there are a selection of proposed methods that could finally see artificial intelligence brought to the edge. Top industry players such as Google and Microsoft, as well as organisations such as Movidius and iEx.ec are working on solutions to bring intelligence closer to the edge.
Two potential solutions showing promise are distributed computing using blockchain technologies and portable AI co-processors. Using blockchain and smart contracts technologies, computing resources can be shared among IoT devices and cooperate without the need for a central broker. This could then allow for the running of computation-heavy AI algorithms within the edge devices themselves.
Alternatively, portable AI coprocessors could be integrated into boards in order to create devices capable of deep learning. Movidius is one of a few companies that have been developing edge neural networks for a while and have even produced their Myriad 2 vision processing unit (VPU), which can be used to achieve computer vision and image signalling capabilities in low power edge environments.
Potential Use Cases for Edge AI
So, as we’ve seen, there are indeed ways in which edge computing and artificial intelligence are being developed so as to meet the needs of other technologies such as the Internet of Things. But how could they improve IoT systems? Let’s now take a look at a few potential use cases for AI at the edge.
Autonomous vehicles are looking extremely likely to be the next big step for intelligent transportation systems of the not-too-distant future. Self-driving cars and intelligent traffic management systems are already being tried and tested today and the integration of edge AI could very well be the way forward.
When it comes to autonomous systems, safety is paramount. Any delay, malfunction, or anomaly within the system can be devastating, and in the case of autonomous vehicles fatal. Using both edge computing and artificial intelligence capable of providing intelligent decision making within edge devices could ensure that these challenges are overcome.
While the robots many of us grew up seeing and reading about in science fiction aren’t quite here, robotics are no doubt a fundamental part of many industrial processes and, given the advances in wireless communications, automation, and IoT technologies, will likely grow in both their necessity and the scope of their applications.
And the excitement surrounding these new innovations within robotics is certainly growing. According to an interview with David Schatsky, a managing director at Deloitte who specializes in emerging technologies, “The new generation of robots can use AI technologies such as computer vision, speech recognition, and more sophisticated analytics of the sensors that they have.”
Maintenance, Monitoring & Security
Maintenance, monitoring and security are three areas in which IoT devices are already thriving. Their use in video surveillance, motion detectors, environmental sensors, and other monitoring and security devices could be greatly enhanced through the integration of edge AI and could allow for further automation of monitoring and security processes.
Predicative maintenance is also an area where edge AI could greatly improve the output of IoT systems. Rail is one of the industries that has started to look into solutions based on edge computing and AI to reduce maintenance costs and unplanned downtime. Machine learning powered monitoring systems are being retrofitted onto existing rail tracks in order to monitor the condition of the tracks and rolling stock. These systems claim to provide a reduction in maintenance costs of around 25%. For multi-billion dollar transit authorities, that’s a huge opportunity. | <urn:uuid:aee7c421-247e-4417-8094-68fc503590e3> | CC-MAIN-2022-40 | https://www.lanner-america.com/blog/edge-computing-artificial-intelligence-match-made-iot-heaven/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335609.53/warc/CC-MAIN-20221001101652-20221001131652-00526.warc.gz | en | 0.947757 | 1,043 | 3.296875 | 3 |
LAS VEGAS--Phishing has proven to be one of the more difficult security problems to solve, and new research from Google shows that it’s not just because humans are gullible. It turns out the attackers are pretty good at their jobs.
One of the most threadbare tropes in security, especially when it comes to phishing, is that human error is to blame for most of our problems. If only we could remove humans from the loop, everything would work perfectly. Of course, this is neither true (see: self-driving vehicles) nor helpful in crafting effective defense. Sure humans make mistakes and click on malicious links and fall for phishing scams, but it’s not always because of ignorance or indifference or carelessness. Quite often, it’s because the people crafting those phishing emails and creating the campaigns understand human psychology and know how to target their messages to create emotional responses.
In most cases, phishing campaigns are designed to separate victims from their money, but while the goal may be the same, tactics and techniques vary widely and evolve quickly. Google’s Gmail service blocks more than 100 million phishing emails each day, and 68 percent of those messages are ones that Google’s systems have never seen before. Phishing campaigns also are often quite short-lived, with the average length of a boutique campaign being just seven minutes, according to Google’s research. The rapid evolution of tactics makes phishing detection a serious challenge, even for sophisticated companies such as Google. It’s also quite difficult for individual people to detect, a fact that phishers know and use to their advantage.
A new study conducted by researchers at the University of Florida in cooperation with Google found that successful phishing campaigns use specific emotional triggers.
“Successful spear phishing emails apply psychological principles of influence – authority, commitment, liking, perceptual contrast, reciprocation, scarcity and social proof,” the new study says.
“These principles of influence exploit common human heuristics that are often beneficial in simplifying decision-making, but can also result in misrepresentation, and can lead to deception. The effectiveness of these weapons in spear phishing emails can be increased when the email places the weapon in a life domain context that is relevant for the user, such as the financial, health, ideological, legal, security, and social domains.”
The researchers presented their results at the Black Hat USA conference here, and in addition to the findings on emotional responses, they found that targeted phishing is more common and effective than bulk campaigns. The massive phishing spam runs pushing pharmaceuticals, lottery scams, and gift cards are still out there, but those emails rarely make it into users’ inboxes these days, thanks to better detection methods. The ones that present the clear and present danger to most people are the spear phishing or boutique phishing campaigns. Spear phishing targets a handful of individual people or organizations and boutique campaigns go after a few dozen companies or people. Google’s numbers show that enterprises are 4.8 times more likely to be targeted by phishing campaigns than any other group.
The University of Florida study involved 158 people of varying ages and found that older people, particularly women, were most susceptible to phishing. While highly targeted and well-crafted phishing campaigns can be effective, there are some equally effective defenses. Awareness of phishing is the first line of defense, and automated detection systems like those deployed by Google and other providers are key, as well. But the most effective defense is implementing two-factor authentication for individual accounts. Google’s statistics show that using a hardware security key is 100 percent effective against both spear phishing and boutique phishing, and device-prompt based 2FA is 99 percent effective against boutique phishing. | <urn:uuid:32005e98-ab1a-4412-8d6f-5fd3d9325a81> | CC-MAIN-2022-40 | https://duo.com/decipher/phishers-play-on-emotions-to-fool-victims | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337339.70/warc/CC-MAIN-20221002181356-20221002211356-00526.warc.gz | en | 0.952826 | 781 | 2.828125 | 3 |
Artificial intelligence analytics for video security system consists of software programs in different programming languages that analyze the audio and video from surveillance cameras in order to recognize certain human beings, vehicles, objects, events and things. The AI uses programming functions which is called machine vision. It is series of algorithms or mathematical procedures which works as reference images of humans.
The AI is capable of maintaining surveillance of 100 cameras altogether simultaneously. This type of AI for security is referred to as rule-based because a human programmer sets rules for things user wishes to be alerted. Now, behavioral analytics has been developed, here the software is self learning without any interference of humans. This type of technology is used in AI analytical cameras where through detectors; it can monitor various aspects of an individual and also environment.
Some of the aspects are as follows:
- Face recognition
- Mask detection
- Body Temperature measurement
- Industrial monitoring and surveillance
- Commercial monitoring
- Smart city surveillance | <urn:uuid:ce868627-7bfd-40c9-a08b-93c32b79945e> | CC-MAIN-2022-40 | https://www.adiance.com/cctv-camera-manufactured-with-ai-analytics-in-2021/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337339.70/warc/CC-MAIN-20221002181356-20221002211356-00526.warc.gz | en | 0.936051 | 192 | 2.90625 | 3 |
If you’re in the business of data, you’ll know that it’s a valuable asset that must be protected. You’ll also be acutely aware that wherever there is data, there is risk, and not just to your data.
Physical security – the protection of people, property and assets – should also be considered for their potential vulnerabilities.
While data centres are famously secure, ‘six layers deep’ in some cases, data theft still occurs. With a number of high-profile cases in the media, questions have been rightly raised over cybersecurity in the Internet of Things (IoT) and unfortunately, lighting and lighting control systems are not immune.
Data centre operators have come to expect that the products installed within their data hall meet certain criteria. Equipment should save energy, be sustainably sourced, but most of all, be safe and secure. However, technology is not without its vulnerabilities; we have all heard ‘that case’ with regards to ‘sub-standard’ data centres, security breaches and spying. As more things become connected, new levels of exposure are being discovered.
Considerations for a connected lighting system
It is important to note that connected (wired) lighting systems without an IP address only communicate within your building. They pose a relatively low-security risk because a person has to be in the facility to attack the system. For example, a conventional wired DALI lighting control system could only be breached if the attacker physically connected to the network.
Lighting and control systems in a wireless network communicate outside of the building. It is common practice to use encryption, which means only devices with the correct ‘key’ can communicate with your system. Correct commissioning is therefore vital.
We know for some businesses, the fear of the unknown makes them reluctant to embrace and invest in new technologies through the fear of being exposed to potential attacks. They instil a culture of ‘if it’s not broken, it doesn’t need to be fixed’, but with cyber-attacks increasing in sophistication, there is every reason to be more vigilant. After all, an ounce of prevention is worth a pound of cure.
As soon as systems get connected to the IoT (Cloud) proper protocols need to be in place. Potential forms of attack on connected lighting systems might include vectoring, Distributed Denial of Service (DDoS) and sniffing.
A Distributed Denial of Service attack is an attempt to make an online service unavailable to its users by temporarily or disrupting services indefinitely.
Occurs when there is a security breach that uses an unsecured system to gain access to other networked systems.
An attacker sees a packet (data) in transmission from one point to other systems that utilise protocols that are not encrypted. Because it’s not encrypted, the information can be modified i.e. to turn off the lights or CCTV.
How to mitigate risk
When it comes to the physical building infrastructure ecosystem, there are many different facets that need to be considered before you can be assured that the product meets your security criteria.
When considering the threats, we recommend starting at the beginning: with a rigorous procurement process, including developing trusted supply chain partnerships.
For example, when a luminaire or control system is specified, are you aware of every component that goes into that product?
Do you know if the manufacturer makes all components themselves? Or, do they rely on third-party suppliers? If so, you’re placing an enormous amount of trust in a potentially unknown supply chain: leaving systems open to security risks and significantly affecting quality control standards
So, what is the answer?
We’d recommend always working with a single-source supplier who can evidence where their components have been sourced and who offer full transparency of their supply chain partners.
As part of the product selection, thorough testing of both hardware and software used in any connected lighting and controls system is highly advisable.
Futureproofing for tomorrow
There is also another advantage of working with fewer trusted supply chain partners.
Not only does consolidating manufacturers into as few as possible make it easier to combat security vulnerabilities, it can also allow for future add-on services to be integrated at a later stage.
For example, it might be a lighting trunking system when installed, but it can also be a flexible infrastructure for future digital services.
A lighting track system can provide a backbone for adding future monitoring services that can grow with the data centre’s needs. It is simply a case of integrating sensors to accurately record the data a facility is interested in monitoring – for example, heat – to ensure the optimum operating temperature within the facility.
Alternatively, if a new sensor is required to measure other variables such as air quality, occupancy and motion, it is easy to remove the original sensor and add on the new one without reconfiguring the entire infrastructure. This naturally saves a significant amount of money in the long term, making it a fully flexible and future proof solution.
New connected lighting and control systems offer exciting improvements in energy and operational efficiencies, but care must be taken to ensure they are secure and not a chink in your data security armour.
We believe that it is crucial to focus on security from the very beginning of your product specification and selection process.
Data centre operators and their design teams should focus on working with supply chain partners who understand system security and who offer safe, strong and secure links to enable campus wide integration.
Mitigate risks by choosing a single source manufacturing partner who is able to offer full traceability and accountability of your lighting ecosystem and offer long term support through a range of services when required. | <urn:uuid:fcb67814-a05f-420b-ba5b-3d22ce08f4d1> | CC-MAIN-2022-40 | https://datacentrereview.com/2021/11/is-your-lighting-compromising-your-data-security/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337625.5/warc/CC-MAIN-20221005105356-20221005135356-00526.warc.gz | en | 0.945596 | 1,202 | 2.828125 | 3 |
APIs are languages that allow applications to exchange information. Today, there are many possible technologies that can be used to design and implement APIs. But before settling on one technology, it is also important to decide which API style to use. In this introduction, we have a brief look at the 5 major API styles.
- Tunnel Style: An API is a collection of functions that can be invoked remotely.
- Resource Style: An API is a collection of resources that allow various kinds of interactions.
- Hypermedia Style: An API is a collection of interlinked resources just like resources on the Web.
- Query Style: An API exposes a structured data model that can be queried and updated with a generic query language.
- Event-based Style: An API is a collection of events that are published by providers and can be subscribed to by consumers.
In this introduction, we have a brief look at these 5 styles and look at the differences in the main abstractions they build on. These 5 styles are the foundation of popular approaches and technologies such as REST, OpenAPI, HTTP, gRPC, GraphQL, and Kafka.
Picking an API style (and a technology that is a good fit for that style) is a design choice, and like any design choice, it should be based on design considerations such as API consumers, API producers, and the API scenario. Knowing the styles will help to better understand the decision to be made, and to make good choices when it comes to designing and implementing an API.
The most important lesson to learn about these 5 API styles is that there is no “best style”. It may be useful to have a default choice to fall back to all other things being equal, but it still is a good idea to have these styles in mind as general starting points for designing APIs.
API design should be approached as a process where a given problem can be solved in different ways, none of them being inherently worse or better than the other ones. The design choice of style and technology is a question of context (designing an API product for API consumers) as much as it is a question of the API itself.
Check out Erik’s YouTube channel for more “Getting APIs to Work” content. | <urn:uuid:0aadc7cf-efd0-4fe2-9ce5-1279dff3b90a> | CC-MAIN-2022-40 | https://blog.axway.com/learning-center/apis/api-management/api-styles-rest-openapi-http-grpc-graphql-and-kafka | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334802.16/warc/CC-MAIN-20220926051040-20220926081040-00726.warc.gz | en | 0.934979 | 463 | 2.84375 | 3 |
Machine learning (ML) technologies and solutions are expected to become a prominent feature of the information security landscape, as both attackers and defenders turn to artificial intelligence to achieve their goals.
“The advent of machine learning in security comes alongside the increased capability for collecting and analyzing massive datasets on user behavior, client characteristics, network communications, and more. As we have already witnessed in many other technological domains, I think machine learning will become the main driver for innovation in information security in the coming decade,” says security researcher Clarence Chio.
Machine learning is no silver bullet
But, he says in advance, machine learning is no silver bullet. And contrary to what some security marketing departments may claim, there are still many domains in which non-adaptive and non-learning methods perform better than machine learning techniques, and many reasons why someone might choose heuristics over machine learning to solve a problem.
“For instance, explainability of machine learning predictions is an important area of research that has to develop further before machine learning can see ubiquitous practical adoption. In the case of a typical web application firewall (WAF), it is often simple to explain why a particular request is blocked. However, in the case of a WAF powered by machine learning, it can sometimes be difficult to explain why a model classifies a request as an attack, especially if the model changes over time and has fuzzy decision boundaries,” he pointed out.
Another problem is auditing – good security systems should have a clear and comprehensive audit trail, and many security machine learning systems today don’t provide that.
Then comes the problem of security. “How susceptible are machine learning systems in a malicious environment? Is it possible to train models that are inherently resilient to adversarial samples or model poisoning? These and other questions have to be answered before we can deploy machine learning in mission-critical scenarios such as in infosec,” he says.
Slow introduction is a must
All the above mentioned issues are the reason while it is still not a good idea to ditch existing systems altogether in favor of machine learning. At the beginning, machine learning should be made to work alongside existing technologies.
“As the sophistication of attackers increase, early integration of such capabilities into your network is critical for more comprehensive attack prevention, detection, and remediation. Broadly speaking, machine learning solutions typically provide higher detection ratios at a cost of increased false positives. Combining this with more conservative rule-based systems can help to increase both the reliability and coverage of your defenses,” Chio notes.
“Enterprise CISOs should be liberal in adopting machine learning solutions. Dealing with machine learning systems and predictions is an important and inevitable skill that technologists and analysts need to have,” he advises.
Machine learning can be effectively used as a method for discovering facets of one’s data for generating better rules. When combined with human intelligence in security operations centers, it can be used to take the load of menial tasks such as incident triaging and log mining, allowing analysts to focus on aspects of the job that machine intelligence isn’t so good at.
The thing is, machine learning lacks the rich contextual, environmental, and experiential knowledge that humans have, and it’s generally bad at drawing correlations between vast, unrelated fields.
What it can be used for is to mine for patterns and discover latent trends in data. Also, machine learning security systems can be used to find flaws in existing technologies. For instance, Generative Adversarial Networks (GANs) have been used to find flaws and loopholes in a system’s security posture.
Machine learning education is very easy to obtain, says Chio. There are various massive open online courses (MOOCs) through which aspiring students can obtain a complete suite of skills necessary to get started in machine learning.
He himself is currently working on improving the quality of available material by co-authoring an O’Reilly book titled “Machine Learning and Security” that is scheduled to be released in late 2017. The book will focus on the many practical methods for using machine learning in security, as well as on concerns tied to that use.
But it’s good to note that learning about the theory behind support vector machines is very different from actually knowing how to use it in real projects.
Chio encourages all who are interested in getting started with machine learning to dive in and implement something.
“Build a spam detector or a malware classifier from scratch. Consider the different options that you have at each step of the way in building such systems, and the tradeoffs that you make choosing one model over another. Only then will one be able to get a wholesome understanding of machine learning,” he notes.
To help you on your path, there is a wide range of tools and frameworks that you can use to play with machine learning.
“Scikit-learn is a good and complete toolkit for general purpose machine learning,” Chio points out. “NLTK is the de facto natural language processing framework. TensorFlow (or Theano – a higher level of abstraction) is a good way to get started with deep learning. Finally, if you come from a statistics or scientific background, R sometimes provides a lot more complete and convenient ways for data and model manipulation compared to the alternatives.”
For more information about the Practical Machine Learning in InfoSecurity lab session at HITBSecConf go here. If you attend, you’ll get a quick introduction to ML concepts and will be up and running with the popular machine learning library, sci-kit learn.
Check out the complete agenda for the conference, it’s a highly recommended event Help Net Security will be attending as well. | <urn:uuid:f18d3edd-4048-48c4-8acd-427102c06a01> | CC-MAIN-2022-40 | https://www.helpnetsecurity.com/2017/03/07/infosec-machine-learning-getting-started/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030336674.94/warc/CC-MAIN-20221001132802-20221001162802-00726.warc.gz | en | 0.939782 | 1,195 | 2.515625 | 3 |
Over the past few decades the combination of Artificial Intelligence (AI) and advanced robotics usurping and eliminating the human race has become a popular science fiction trope. You only need marvel at Will Smith leaping across self-driving vehicles on a fully automated highway, battling out-of-control – or very much in control – machines in I, Robot to know that AI (theoretically) could decide that the only way to preserve human life – is to master it.
Despite the hyperbole, issues that titles such as I, Robot, Ex Machina, and Avengers; Age of Ultron all raise around AI ethics are all very real and represent the most monumental and imminent change ever to be recognised by humankind.
Differentiating between AI and AGI
Before we get buried into the ethics we need to understand what we mean by AI. Often, we confuse the term Artificial Intelligence (AI) with Artificial General Intelligence (AGI), with the latter being exaggerated across popular culture. AI functions are preprogramed beforehand, i.e. the decisions AI machine learning makes are based on empirical existing data, whereas AGI is sentient, self-aware and can understand its position in the world and the actions of others.
Today our most pressing problem centres on the use of AI rather than AGI as it becomes ever more invasive in our world without us even realising. Highly specialised AI’s are already used to control decision processes and reveal hidden data relationships, which is having a dramatic impact on society and shaping everyday decisions such as deciding what adverts we see to the level of healthcare we get.
AI is becoming responsible for our own personal view of the world, and what’s worrying is that those responsible for these decisions cannot guarantee this process is conducted in a fair and unbiased way. Yes, AI is being used for good: to discover new medicines, better treatments or crop yields. They are also being leveraged for commercial advantage, such as to provide threat detection within a firewall. That’s also fine, but what about more nefarious objectives, such as manipulating voters in a parliamentary election? Is this a competitive edge or is it illegal?
How far do we go?
Perhaps one of the most difficult questions to answer now is at what point do we allow AI to decide on our behalf? Today there are some obvious areas where as a society we would naturally draw the line, either through an understanding of the risk or driven by fear of the unknown, but there are plenty of examples where it is not that simple. Using science fiction again to demonstrate my point, take the decision of the machine in I, Robot that saves Will Smith’s character over a girl because his chance of survival is higher. Is this ethical? It is in the eyes of the creator.
Enterprises sit at the heart of the AI revolution and will be the decision makers of what is and is not ethical. The problem with this is that what is likely to happen is that the leaders will repeatedly leap at the wonders of using AI as a great problem solver or competitive edge without fully realising or considering the ethical dilemma first.
Driverless cars are a great example. We cannot influence the outcome for every eventuality and we do not have direct control over the AI decision process, so we must rely on something that can assess its environment and decide for itself. However, we need to put boundaries on those decisions and one such boundary is the decision to avoid an action that may endanger human life inside or outside the vehicle. This is a clear and extreme ethical problem to solve. Who would we blame for a fatal collision? The car company? Okay, but who is personally prosecuted, the CEO? The designer of the car? The AI? If we choose the latter, what are the consequences of attributing blame to something that isn’t sentient? Will this be used as a scapegoat for enterprises?
It’s taken over a year for US courts to determine that Uber will not face criminal charges for a fatal crash in Arizona involving one of its self-driving cars, and that in fact the back-up driver is liable and likely to be prosecuted. Indeed, it’s a lot easier to blame a person than an enterprise – but what happens when there’s no back-up driver? Are a few deaths here and there because of autonomous vehicles ‘ethical’ in the bigger picture when we consider just how many fatal road traffic accidents there are globally?
The more we delve into AI decision making the more we find ourselves faced with the same problem: is there a direct or indirect detrimental effect on a human being that could be considered unfair, unethical or uncivilised? Bigger still, how do we detect such a problem?
Laws and guidelines
The current overarching conundrum surrounding AI ethics is really in who decides what is ethical. AI is developing in a global economy, and there is a high likelihood of data exchange between multiple AI solutions. Without clear testing guidelines or even in most cases the ability to test, we can’t know if a system hasn’t been intentionally corrupted or simply built from a flawed set of principles. Take the Uber case. This incident is unprecedented. The state of Arizona has no legal guidelines whatsoever to work from – globally or locally – so appropriate punishment is determined solely by the principles of the judges in Arizona.
Some believe that if we make an ethical miscalculation in the use of AI we can always go back and try again in that area. This is incorrect. Consider market drivers and social dependency in the world today. Just look at our dependency on social media, which lacks many necessary controls and yet people still use it religiously. Also, consider the introduction of Internet of Things or “smart devices” which the consumer has lapped up despite having no clear regulation around security.
Although we are still some way off sophisticated AI solutions becoming commonplace, this much is clear. We need to establish the minimum global guidelines for AI security and ethics before enterprises leading the charge drive us down a route we cannot back out of. If we adopt the same blind faith in AI solutions and platforms as we have with companies like Facebook, perhaps science fiction won’t be fiction after all.
Image Credit: Geralt / Pixabay | <urn:uuid:1d797bb7-3383-4c77-b007-17c7a8ef3e84> | CC-MAIN-2022-40 | https://www.itproportal.com/features/avoiding-ais-darkest-future-is-ethical-ai-beyond-our-control/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337360.41/warc/CC-MAIN-20221002212623-20221003002623-00726.warc.gz | en | 0.956969 | 1,287 | 2.859375 | 3 |
High-Quality Copper Oxide Microcrystals Are Synthesized for Quantum Photonics
(LaserFocusWorld) Cuprous oxide (Cu2O) is a promising material for quantum photonics and optoelectronics, due to its ability to simulate highly excited Rydberg atomic states via the formation of solid-state excitons within the material. Now, a team of researchers from KTH Royal Institute of Technology (Stockholm, Sweden), Graz University of Technology (Graz, Austria), and CNRS, École Polytechnique (Palaiseau, France) has found a way to synthesize high-quality Cu2O microcrystals.
“The majority of quantum optics experiments with this material have been performed with geological samples found in mines—for instance, the Tsumeb mine in Namibia,” says Stephan Steinhauer, a researcher at KTH. “Our synthesis method is associated with very low-cost fabrication, suitable for mass production, and does not require gases or chemicals that are toxic or harmful for the environment.” He notes that the properties of Cu2O can lead to new schemes for quantum information processing with light in the solid state, which are difficult to realize with other materials. | <urn:uuid:ff42b039-a2e5-4d06-b45c-5c335d6d3cd3> | CC-MAIN-2022-40 | https://www.insidequantumtechnology.com/news-archive/high-quality-copper-oxide-microcrystals-are-synthesized-for-quantum-photonics/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337480.10/warc/CC-MAIN-20221004054641-20221004084641-00726.warc.gz | en | 0.911025 | 257 | 2.75 | 3 |
We would like to keep you up to date with the latest news from Digitalisation World by sending you push notifications.
Ransomware is one of the most prolific and dangerous types of cybercrime today. As a result of ransomware, millions of dollars and petabytes of data have been lost by organisations and individual users worldwide. No one is safe from a ransomware attack on their valuable data. However, it is possible to reduce the risks of data loss, minimise the negative impact of ransomware and be prepared for potential attacks. This article provides step-by-step recommendations on how to prevent ransomware attacks and what to do to enable recovery as quickly as possible, if you detect it.
The new ransomware
The increased rate of ransomware attacks in 2020 is the highest in history. New attacks are also becoming more sophisticated, with strikes on servers rising in popularity. By taking control of servers, attackers can infect more computers connected to an enterprise network and cause greater damage.
Ransomware of the late 2000s is a very different animal to ransomware today. When the COVID-19 epidemic hit in full force in 2020, a new wave of Netwalker attacks ensued. Attackers have been distributing Netwalker primarily via phishing links introduced as emails claiming important updates about the coronavirus in attachments. As most people are interested in finding out more about COVID-19, this increased the probability of users opening these malicious attachments. But data loss is a concern for companies and individual users alike. As, while ransomware news coverage indicates the current trend is that attacks mainly occur within large organisations, attacks on individual users also took place as companies shifted to remote work during the pandemic. This has drastically increased the vulnerability of corporate data.
It is vital that organisations be vigilant and move with the times, implementing security measures both within the organisation and at home to prevent being attacked by new ransomware.
How to protect data against new versions of ransomware
Online safety rules and company security policy should strictly be adhered to by enterprise users, even when working remotely. If users have received appropriate training and are made aware of the latest ransomware threats and infecting methods, they are less likely to be infected. Users should be educated about phishing, social engineering and other methods used to infect with ransomware.
Top 10 ransomware prevention tips:
1. Do not open suspicious email messages, attachments to these email messages and any other links. Distributing ransomware via email is one of the most common methods of infecting computers.
2. Configure filters on email servers to reject suspicious email messages that can turn out to be malicious. The best way to ensure that users won’t open malicious email messages is by rejecting these email messages on email
servers. Leading SaaS (software as a service) providers such as Google and Microsoft provide email filters for their email services to protect users.
3. Train users to detect suspicious messages and ensure they are familiar with social engineering methods. The more skilled users are, the less likely it is that their computers will be infected with ransomware.
4. Do not provide any personal information to unknown users even if they introduce themselves as partners, bank workers, social agencies, etc. via email, messengers, phone calls, social media or other communication tools.
5. Install antivirus/antimalware software on all computers. Antivirus must always be up to date. It is also vital to install all available updates and security patches on operating systems and other software installed on all computers.
6. Configure a router and firewall in your network properly. Close unused ports, allow access from trusted networks and IP addresses if possible. Change standard port numbers to custom port numbers for some protocols (SIP, RDP, SSH etc.). Often attackers scan standard ports to detect which of them are open. It is also worth considering configuring a firewall on user computers.
7. If you find a USB flash drive, flash card or other medium near your office or home, don’t rush to plug it into a computer. An attacker can drop an infected flash drive near your home or office to distribute ransomware. Notify users about this threat and tell them that they should report such cases to the system administrator. If the content needs to be checked on the found device, an isolated computer should be used, Linux should be booted and checked.
8. Configure an operating system on each computer to show hidden files, system files and extensions for all file types. Use strong and unique passwords for different accounts.
9. Choose a wired network over Wi-Fi. When using wired networks, physical access is required for connecting to a network. Attackers can crack the password to the wireless network. If a wireless network is in use, make sure to set a strong password. Keep in mind that attackers can steal a saved password from computers or mobile devices. If any user loses a computer or any other mobile device that was used for connecting to a Wi-Fi network, the user must notify the system administrator. It is highly recommended to change the Wi-Fi password in this case. When on a business trip, try not to connect to public and untrusted Wi-Fi networks. In addition, restricting permissions for users on their work computers is advisable if this is at all possible, as long as this doesn’t prevent users from doing their work.
10. Create regular backups of important data. Creating a backup is the most effective method of protection against ransomware attacks. According to Sophos white paper on the State of Ransomware 2020, more than 56% of victims (organisations) have restored their data from a backup. However, backing up files from internal disk drives to USB drives that are always plugged into the computer is not effective. If the computer is infected after a ransomware attack, ransomware encrypts files on all the attached disks, including USB drives. Burning backed up data to DVD discs, Blue-ray discs or tape drives is a reliable solution because ransomware cannot re-write data on these types of media. Follow the 3-2-1 backup rule and keep at least three copies of data, store copies on different media, and keep at least one copy offsite.
Ransomware detection and response
What should you do if you detect ransomware? Typically, ransomware will appear as a splash (lock) screen with a notification that files have been encrypted and will be irreversibly deleted, along with a countdown (designed to pressure the victim into making mistakes like paying the ransom).
Should you become compromised, it is recommended that you do not pay the ransom. There are ample cases where organisations never got their data back after paying the attackers. Psychological tricks are often used to inspire fear and panic. Paying the ransom would only reinforce this type of attack, encouraging cybercriminals to continue to stage more ransomware attacks in the future. In addition, even after paying the ransom, there is no guarantee that the files can be decrypted, or that the stolen data will not be sold to other criminals, or even competitors.
A ransomware attack can infect multiple computers in the network and irreversibly corrupt terabytes of files. Considering that cybercriminals are continuously coming up with new and more dangerous versions of ransomware, with attacks becoming increasingly sophisticated, it is vital that a set of protection measures be implemented for adequate ransomware protection. This should include proper configuration of firewall, installing security patches, educating users, using an up-to-date antivirus, and regular data backup. Regularly backing up data is one of the most effective methods to prevent data loss due to a ransomware attack. If organisations properly prepare themselves against any potential attackers in advance, they will be in a position to protect their valuable data, ensuring business continuity should the worst happen.
*Sophos white paper on the State of Ransomware 2020: https://secure2.sophos.com/en-us/content/state-of-ransomware.aspx
3-2-1 backup rule: https://www.nakivo.com/blog/3-2-1-backup-rule-efficient-data-protection-strategy/ | <urn:uuid:2d0fbb91-727e-4577-b5c6-742d90714c42> | CC-MAIN-2022-40 | https://digitalisationworld.com/blogs/56655/prevention-is-better-than-cure-best-practice-tips-for-new-ransomware | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334514.38/warc/CC-MAIN-20220925035541-20220925065541-00126.warc.gz | en | 0.931673 | 1,657 | 2.765625 | 3 |
We have an almost mystical faith in the ability of artificial intelligence (AI) to understand and solve problems. It’s being applied across many areas of our daily lives and, as a result, the hardware to enable this is starting to populate our data centers.
Data centers in themselves present an array of complex problems, including optimization and prediction. So, how about using this miracle technology to improve our facilities?
Turning the AI inwards
Machine learning, and especially deep learning, can examine a large set of data, and find patterns within it that do not depend on the model that humans would use to understand and predict that data. It can also predict patterns that will repeat in the future.
Data centers are already well-instrumented, with sensors that provide a lot of real-time and historical data on IT performance and environmental factors. In 2016, Google hit the headlines when it applied AI to that data, in order to improve efficiency.
Google used DeepMind, the AI technology it owns, to optimize the cooling in its data centers. In 2014, the company announced that data center engineer Jim Gao was using the AI tech to implement a recommendation engine.
In 2016, the project optimized cooling at Google's Singapore facility, using a set of neural networks which learned how to predict future temperatures and provide suggestions to respond proactively,
The results shaved 40 percent off the site's cooling bill, and 15 percent off its PUE (power utilization effectiveness), according to Richard Evans, a research engineer at DeepMind. In 2016, he promised: “Because the algorithm is a general-purpose framework to understand complex dynamics, we plan to apply this to other challenges in the data center environment and beyond.”
The next step, announced in 2018, was to move closer to a self-driving data center cooling system, where the AI tweaks the data center’s operational settings - under human supervision. To make sure the system operated safely, the team constrained its operation, so the automatic system “only” saves 30 percent on the cooling bill.
The system takes a snapshot of the data center cooling system with thousands of sensors every five minutes, and feeds it into an AI system in the cloud. This predicts how potential actions will affect future energy consumption and picks the best option. This is sent to the data center, verified by the local control system, and then implemented.
The project team reported that the system had started to produce optimizations that were unexpected. Dan Fuenffinger, one of Google’s data center operators who has worked extensively alongside the system, remarked: "It was amazing to see the AI learn to take advantage of winter conditions and produce colder than normal water, which reduces the energy required for cooling within the data center. Rules don’t get better over time, but AI does."
According to Gao, the big win here was proving that the system operates safely, as well as efficiently. Decisions are vetted against safety rules, and human operators can take over at any time.
At this stage, Google’s AI optimization has one customer: Google itself. But the idea has strong backing from academia.
Humans, and simple rule-based systems can respond to any steady-state situation, but when the environment changes, they react in a “choppy” way - and AI can do better, because it is able to predict changes, according to DCD keynote speaker Suvojit Ghosh, who heads up the Computing Infrastructure Research Centre (CIRC) at Ontario’s McMaster University.
“We know it's bad to run servers too hot.” said Ghosh. ”But it's apparently even worse if you have temperature fluctuations.” Simple rules take the data center quickly to the best steady state position, but in the process, they make sudden step changes in temperature, and it turns out that this wastes a lot of energy. If the conditions change often, then these energy losses can cancel out the gains.
“If you have an environment that goes from 70°F to 80°F (21-27°C) and back down, that really hurts," said Ghosh.
Companies in data center services are responding. Data center infrastructure management (DCIM) firms have added intelligence, and those already doing predictive analytics have added machine learning.
“The current machine learning aspects are at the initial data processing stage of the platform where raw data from sensors and meters is normalized, cleaned, validated and labeled prior to being fed into the predictive modeling engine,” said Zahl Limbuwala, co-founder of Romonet, an analytics company now owned by real estate firm CBRE.
The move for intelligence in power and cooling goes by different names. In China, Huawei’s bid to make power, cooling and DCIM smarter goes under the codenames iPower, iCooling and iManager.
Like Google and others, Huawei is starting with simple practical steps, like using pattern matching to control temperature and spot evidence of refrigerant leaks. In power systems, it’s working to identify and isolate faults using AI.
In its Langfang data center, with 1,540 racks, Huawei has reduced PUE substantially using iCooling, according to senior marketing manager Zou Xiaoteng. The facility operates at around 6kW per rack with a 43 percent IT load rate.
DCIM vendor Nlyte nailed its colors firmly to the DCIM mast in 2018, when it signed up to integrate its tools with one of the world’s highest profile AI projects, IBM’s Watson.
Launching the partnership at DCD>New York that year, Nlyte CEO Doug Sabella predicted that AI-enhanced DCIM would lead to great things: “The simple things are around preventive maintenance,” he told DCD. “But moving beyond predictive things, you’re really getting into workloads, and managing workloads. Think about it in terms of application performance management: today, you select where you’re going to place a workload based on a finite set of data. Do I put it in the public cloud, or in my private cloud? What are the attributes that help determine the location and infrastructure?
“There’s a whole set of critical information that’s not included in that determination, but from an AI standpoint, you can contribute into it to actually reduce your workloads and optimize your workloads and lower the risk of workload failure. There’s a whole set of AI play here that we see and our partner sees, that we’re working with on this, that is going to have a big impact.”
Amy Benett, North American marketing lead for IBM Watson IoT, saw another practical side: “Behold, a new member of the data center team, one that never takes a vacation or your lunch from the breakroom.”
DCD understands the partnership continues. The Watson brand has been somewhat tarnished by reports that it is not delivering as promised in more demanding areas such as healthcare. It's possible that this early brand leader has been oversold, but if so, data centers could be an arena to restore its good name. The vital system of a data center is much more simple than the human body.
The next stage
It's time for AI to reach for bigger problems, says Ghosh, echoing Sabella's point. After the initial hiccups, efforts to improve power and cooling efficiency will eventually reach a point of diminishing returns. At that point, AI can start moving the IT loads themselves:
“Using the cost of compute history to do intelligent load balancing or container orchestration, you can bring down the energy cost of a particular application,” Ghosh told his DCD audience. This could potentially save half the IT energy cost, “just by reshuffling the jobs [with AI] - and this does not take into account turning idle servers off or anything crazy like that.”
Beyond that, Ghosh is working on AI analysis of the sounds in a data center. “Experienced people can tell you something is wrong, because it sounds funny,” he said. CIRC has been creating sound profiles of data centers, and relating them to power consumption.
Huawei is doing this too: “If there is a problem in a transformer, the pattern of noise changes,” said Zou Xiaoteng. “By learning the noise pattern of the transformer, we can use the acoustic technology to monitor the status of the transformer.”
This sort of approach allows AI to extend beyond expert human knowledge and pick up “things that human cognition can never understand,” said Ghosh.
“In the next 10 years, we will be able to predict failures before they happen,” said Ghosh. “One of my dreams is to create an algorithm that will completely eliminate the need for preventative maintenance.”
Huawei’s Xiaoteng reckons there are less-tangible benefits too: AI can improve resource utilization by around 20 percent, he told DCD, while reducing human error.
Xiaoteng sees AI climbing a ladder from level zero, the completely manual data center. “On level one the basic function is to visualize the contents of the data center with sensors,and on level two, we have some assistance, and partially unattended operation,” where the data center will report conditions to the engineer, who will respond appropriately.
At level three, the data center begins to offer its own root cause analysis and virtual help to solve problems, he said. Huawei has reached this stage, he said: “In the future, I believe we can use AI to predict if there's any problem and use the AI to self-recover the data center.”
At this stage, DCIM systems may even benefit from specialized AI processors, he predicted. Huawei is already experimenting with using its Ascend series AI processors to work in partnership with its DCIM on both cloud and edge sides.
Right now, most users are still at the early stages compared with these ideas, but some clearly share this optimism: “Today we use [AI] for monitoring set points,” said Eric Fussenegger, a mission critical facility site manager at Wells Fargo, speaking at DCD>New York in 2019, adding to DCIM and “enhancing the single pane of glass.”
AI could get physical, further in the future, said Fussenegger, in a fascinating aside. “The ink is not even dry yet, maybe it hasn't even hit the paper.” he said, but intelligent devices could play a role in the day-to-day physical maintenance and operation of a data center.
One day, robots could take over "cleaning or racking equipment for us, so I don’t have to worry about personnel being in hot and cold aisle areas. There are grocery stores that are using AI to sweep.”
Even these extreme views are tempered, however. Said Fussenegger: “I think we’re always going to need humans in there as a backup.” | <urn:uuid:0b135fc9-afb2-46d5-8abb-aabd6e668df8> | CC-MAIN-2022-40 | https://www.datacenterdynamics.com/en/analysis/smartening-how-ai-and-machine-learning-can-help-data-centers/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334871.54/warc/CC-MAIN-20220926113251-20220926143251-00126.warc.gz | en | 0.945076 | 2,326 | 3.375 | 3 |
The year is 2020. The cyber attack of your worst nightmare has just become a reality. Are you at all prepared?
That’s the question — and the scenarios — that researchers and experts are undertaking to minimize any damage caused by extensive cyber attacks, the type that could shut down power grids and totally reshape internet users’ attitudes about the likelihood that their information will be stolen. By doing so, they hope to find the solutions and actions to fit different types of situations.
As part of its initiative to develop cybersecurity practices to handle future threats, the University of California, Berkeley Center for Long-Term Cybersecurity developed a study of worst-case scenarios involving attacks and solutions to address them.
In its report, the center noted that in the near future most people and things will be connected to digital networks, especially in the wake of an increasingly popular Internet of Things (IoT) environment. People will be using the internet to automatically open doors, close windows, operate refrigerators, monitor their homes, control temperatures and other tasks.
“For these reasons we believe the cybersecurity research and policy communities will soon confront a much more diverse set of problems and opportunities than they do today,” the researchers said in its report.
In another similar announcement, the North American Electric Reliability Corp. (NERC) said that in a worst-case scenario, a cyber attack on the electric grid could cause an outage that lasts one to two weeks. In making the announcement, the nonprofit organization said that a physical attack on electric substations could potentially be more damaging. A cyberattack in Ukraine left more than 250,000 without power in December 2015. There has been concern it could also happen in the United States.
To address the new issues that may come about in that future environment, the Berkley Center developed several worst-case scenarios. They include:
Scenario 1: After years of frequent data breaches, Internet users no longer will be assured about the security of their personal data. They will start assuming that it will be stolen — it’s just a matter of when. At that point, more individuals and institutions may decide to go offline or find ways to protect their own information.
Scenario 2: As data scientists will develop more powerful models that are able to predict and manipulate our behavior with a high level of accuracy, there is a risk for new security vulnerabilities that could lead to extensive damage, including financial.
Scenario 3: With advertising-driven business model for major Internet companies, such as Facebook, gathering massive amounts of data on Internet users, the researchers explore the possibility of that system falling apart. If the companies happen to become cash-strapped, for instance, the value of selling datasets.
Envisioning these scenarios can help institutions, including the government, develop strategies to help them become more resilient against these type of attacks and threats.
Even if the predictions turn out to be wrong, “the effort to anticipate is worth a lot, because it gives you the ability to react quickly,” said Herb Lin of Stanford’s Center for International Security and Cooperation. “ And if you can react quickly, you’ll be less screwed.”
Want to learn why EMP shielding, FedRAMP certification, and Rated-4 data centers are important?
Download our infographic series on EMP, FedRAMP, and Rated-4! | <urn:uuid:b2647acb-24b5-4be3-8ce0-004fc4e902ba> | CC-MAIN-2022-40 | https://lifelinedatacenters.com/data-center/cyber-attack-scenarios/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335469.40/warc/CC-MAIN-20220930113830-20220930143830-00126.warc.gz | en | 0.938767 | 692 | 3.078125 | 3 |
Just as recently as two to three years ago, environmentalists and industry leaders were expressing grave concerns that the demand for data centers would lead to a significant rise in energy consumption. There was reason for concern. With the emergence of the Internet of Things (IoT) to run businesses and households and the demand for Big Data to provide businesses with critical insights, the accelerated growth of data centers was inevitable.
However, recent studies are showing that those fears are unfounded as data centers have taken on the challenge to implement energy-saving measures.
A new report released by the U.S. government shows that water consumption, like energy, is revealing a slowing trend among data centers throughout the country. Electricity generation accounts for a large segment of water usage. Studies show that data centers were responsible for 2 percent of all electricity consumed in 2014 throughout the United States.
The report also revealed that water consumption trends between 2006 to 2020 has experienced a lower growth rate — similar to the way that energy consumption has been slowing. In 2014, data centers in the United States consumed 626 billion liters of water. Researchers predict it will reach 660 billion liters in 2020. Variances could be pointed to the fact that different power plants are adopting more efficient equipment and generation sources are using different amounts of water.
Innovations in data center operations have also contributed to an unanticipated decline in energy usage among data centers — all the while, the demand for data centers continues to grow.
According to a study by the U.S. Department of Energy, Carnegie Mellon University, Stanford University, and Northwestern University, America’s data centers consumed about 70 billion kilowatt-hours of electricity. That’s about 2 percent of the total energy consumption in the United States. The study also revealed that total data energy consumption grew by a rate of 4 percent from 2010 to 2014 — compared to 90 percent from 2005 to 2010.
That significantly slower rate of growth has been attributed to innovations and best practices that include cloud computing, efficient cooling systems, consolidation of servers, and reducing the number of idle systems.
Want to learn why EMP shielding, FedRAMP certification, and Rated-4 data centers are important?
Download our infographic series on EMP, FedRAMP, and Rated-4! | <urn:uuid:06392964-002f-4c15-bc1b-1610bb422ad0> | CC-MAIN-2022-40 | https://lifelinedatacenters.com/data-center/water-energy-usage/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335469.40/warc/CC-MAIN-20220930113830-20220930143830-00126.warc.gz | en | 0.958872 | 462 | 3.296875 | 3 |
Smishing – a word combination of SMS and fishing that most people know from phishing. The latter represents an attempt to “fish” for passwords. Smishing refers to the channel, namely SMS, as a means to an end. Since SMS is used for only a few purposes now, criminals focus on false package deliveries or notifications regarding an existing two-factor authentication.
More than 100 million false SMS messages in the Telekom network alone. The problem is large and affects many smartphones and users – regardless of the network. Even when network operators issue warnings, reliable mechanisms for early prevention are lacking. In order to reduce the risk of suffering damage, information, education & training in dealing with false SMS are indispensable – in private and business contexts.
Clicking on the link in an SMS can be particularly critical
Can a click on a link already be harmful? The answer to this question must be “yes”. This is just as true for a click in the email as it is for the SMS. The latter, however, is even more critical. This is because instead of receiving information about the alleged package delivery, the program prepares and carries out the sending of another numerous SMS messages to the contacts in the background. This approach works especially well because software on third-party sources are used. Unfortunately or fortunately, this is only a problem of Android-based smartphones.
But even the pleasant notification of a win should make you wonder. Especially if you have neither entered your mobile number nor actually participated in a sweepstakes. The Federal Office for Information Security has long been aware of the problem and regularly provides information about new fraud schemes in this context.
The tips and behaviors for dealing with dubious text messages are catchy:
- You should not click on any link.
- Downloading files from unknown source should be strictly avoided
- Delete the SMS
In addition, there are other tips that serve for prevention:
- Block the number of the sender
- Activate the third-party provider block to prevent unnecessarily high costs
What can I do if I clicked on a link in a smishing SMS?
The right time, the right occasion and the right person – when these aspects come together, criminals are often successful. But it is only statistics – with a number of more than 100 million SMS messages. But what can you do as an affected person after clicking on a malicious link in an SMS? The recommendations from the BSI are good advice and can be summarized as follows:
- Deactivate the mobile network. Activate flight mode as soon as possible. This prevents the sending of further SMS.
- Inform your mobile phone provider about the
- Be extra vigilant with your statements (account/cell phone bill)
- Making a report and resetting to factory settings are the last indications
Despite precautionary measures by the mobile network provider, such as anomaly and fraud detection, there is (still) no security on this communication channel. At regular intervals, everyone in possession of a mobile phone number should be aware of such and similar scams. | <urn:uuid:740dbb25-6443-4bc2-9f16-2c96eb2e9994> | CC-MAIN-2022-40 | https://aware7.com/blog/smishing-the-big-scam-with-the-fake-sms/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030336921.76/warc/CC-MAIN-20221001195125-20221001225125-00126.warc.gz | en | 0.927835 | 619 | 2.640625 | 3 |
A long-haul truck cruises by you on Interstate 10 in west Texas. You get a brief wave from the driver who seems unfocused on the road ahead. It’s hot and the road is bumpy. You didn’t realize it but you just encountered an autonomous driving truck in development, testing out the latest version of its Artificial Intelligence algorithms. The truck is loaded with unseen video cameras, lidar, radar and infer-red sensors. As it travels along it is seeing its environment and its on-board computers are making the thousands of little decisions that keep it moving safely to its prescribed destination. The driver is a safety layer that will be removed in future iterations of the design.
Meanwhile flying over a remote desert, a reconnaissance aircraft is gobbling up data at an amazing rate and using its artificial intelligence capability to build a sophisticated understanding of the environment it is operating in, both the natural and manmade, determining friend and foe. The plane is working in conjunction with a mobile command center below which is aggregating the intelligence from scores of sources and applying its own artificial intelligence to inform real time tactical decision making on the ground.
The truck is operating in extremes of hot and cold with constant shaking in an environment of dust and humidity with limited power available for computer equipment. In the plane the vibration is constant, with the occasional shock of unexpected turbulence, the altitude effects the ambient temperature, and space and weight are at a premium. The power on the plane is limited and unique. On the ground the mobile command center has its own environmental challenges with shock and vibration while in transit and the need to continue operations in extreme hot and cold temperatures. These are not environments friendly to sensitive electronic equipment.
These applications are like many similar in autonomous vehicles in mining, construction, or agriculture or in military applications on or under the sea. They share the common trait of needing to deploy the most sophisticated Artificial Intelligence but demanding that that capability be delivered in unforgiving harsh environments. These are environments where your typical computer systems cannot operate. This is the emerging challenge; how do you deploy the most capable AI platforms where the rubber meets the road.
The first thing to understand is that these applications will benefit from the very highest level of AI performance. The more performance available the more data can be utilized and the more insightful AI conclusions can be obtained. This means the AI platforms need to be no-compromise in the level of hardware technology they include; the latest and most power CPU, GPUs, co-processors, memory, and storage. Today GPUs are the workhorse of AI computation. High end GPUs utilize 100’s of watts of power and a system can require multiple 1000’s of watts. The electronics in these systems generate a significant amount of heat exasperating the need to cool the system in harsh conditions.
Unlike in the controlled environment of a datacenter with constant cool ambient temperatures, stable and abundant power and no need to accommodate for shock, vibration, humidity, or dirt these AI platforms must be built to tough standards.
The platforms must first be rugged. Their mechanical design and material selection needs to accommodate the instability of mobile deployment. They must be compact with form factors and mounting strategies that accommodate the unique space available on size optimized vehicles. They must include innovative cooling design that deals with extremes in ambient temperature while cooling powerful hot components. Strategic air-cooled topologies or the use of liquid cooling is required. Additionally, vehicles on land, sea or air will have different power sources available ranging from terrestrial vehicle providing 48V DC while some aircraft supply 3 phase 400Hz power.
With the proper design all of these requirements and constraints can be met but not with your standard off-the shelf HPC systems. What is required is a modular standard edge supercomputer that can deliver datacenter class AI performance with the rugged flexible design to accommodate harsh operating environments.
Now when you are passed by that long haul truck testing its autonomous capabilities you can think about the engineering challenges that need to bead dressed to provide for its ever-expanding intelligence requirements. And give that driver a wave, you might not see him next time. | <urn:uuid:fb99b201-ff24-47b7-8dd0-2997bdea947b> | CC-MAIN-2022-40 | https://enterpriseviewpoint.com/deploying-artificial-intelligence-where-the-rubber-meets-the-road/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334912.28/warc/CC-MAIN-20220926144455-20220926174455-00326.warc.gz | en | 0.935208 | 838 | 2.96875 | 3 |
As California’s forests glowed, lights at NERSC began to switch off.
For the second time in as many weeks, one of the world’s most powerful supercomputers was carefully being shut down.
Climate change, once an abstraction being dispassionately simulated on the 30 petaflops Cori system, had manifested into reality.
The machine, a pinnacle creation of mankind - a species that prioritized rapid progress over sustainable development - had been waylaid by the impact of that progress on the world.
Like the families that prepared to flee California’s suburbs, Cori could do nothing to stop the effects of anthropogenic climate change. It was truly powerless.
California on fire
In October, after an unusually long dry spell, remarkably high winds, and arid conditions, utility PG&E grew increasingly concerned a massive wildfire could be caused by a broken power line.
Partly responsible for the deadliest and most destructive wildfire in Californian history, 2018’s Camp Fire - which cost $16.5 billion and led to at least 85 civilian deaths - PG&E currently faces bankruptcy and a potential state takeover.
While climate change has exacerbated the causes of wildfires, and will increasingly make things worse, PG&E has been criticized for its lack of preparation, not clearing trees and brush from around power lines, and not having enough emergency staff.
This year, PG&E decided the best way to mitigate the risk of its grid sparking another fire was to switch off parts of that grid, preemptively shutting down power to nearly three million people in central and Northern California during this fire season.
Still, despite the precautions, some fires raged, including the Kincade Fire that burned 77,758 acres in Sonoma County.
Among those caught up in two separate multi-day power cuts was the Lawrence Berkeley National Laboratory (LBNL), home to the National Energy Research Scientific Computing Center (NERSC) and an unlucky Cori, the thirteenth most powerful supercomputer in the world.
“There was some warning,” Professor Katherine Yelick, associate laboratory director for computing sciences at LBNL, told DCD. “It was about five hours.”
It takes around two to three hours to shut down Cori, NERSC director Dr. Sudip Dosanjh said.
“If there's a sudden power outage, you can have issues with the system, maybe some parts fail, so just to be on the safe side, we decided both times to go ahead and bring the big system down.”
NERSC has uninterruptible power supplies and on-site generators, “but that's not enough to power Cori,” which consumes 3.9MW. “It's enough to power the network and some file systems,” Dosanjh said.
“So we kept the auxiliary services up during the entire outage, including the network, and something we call Spin,” which can be used to deploy websites and science gateways, workflow managers, databases and key-value stores.
Those systems could have stayed up indefinitely, as long as there is enough available fuel to refill the generators’ day tanks every six to eight hours.
The first shutdown, beginning on October 10, “was the first time that something like this had actually happened,” Yelick said. The lab had emergency procedures in place for similar events, but “there's a big difference between having a plan in place and then having to execute it,” she admitted. “We did have a plan, but it wasn't as though this was really expected.”
NERSC “certainly learned a lot during the first shutdown that helped with the second,” which began on October 26, she said. “We learned about communications and the generators and how each one works - those kinds of things.”
Another crucial lesson was how many people Emergency Operations Center (EOC) fielded to deal with LBNL’s power cut, which also took down other science resources, including the DNA sequencing lab, the Molecular Foundry, and the Advanced Light Source.
“[The first time] we didn't have a large number of people that were cycling through the EOC,” Yelick said. “And so I think they got pretty tired. We added some additional people the second time. [In future] we would want to make sure that there's enough people that are able to bring the systems up and are confident of doing it on their own, so that we don't overly fatigue a small group of people.”
The second time, NERSC was even able to do some maintenance, doing tests on the upcoming community file system ‘Storage 2020.’
Roughly 100 personnel were involved with the emergency operations at the lab, of which around 20 were actually on site. “We're trying to collect that list of exactly how many people were involved right now,” Yelick said. Cori itself only had a few staffers working on it during the shutdown and return, including employees from the supercomputer’s manufacturer, Cray.
The process of returning everything online after power came back took six to eight hours both times.
With various teams interacting, many of them working remotely, communications infrastructure was a key concern. Luckily, cell towers and Internet connectivity mainly stayed online during both outages.
“We were using cell phones,” Yelick said. “That's one of the things that we added in the second outage. And most people work pretty hard to find some way of communicating if they can't, even if it means driving someplace.”
“Long before the power went off, the emergency response teams were using email, text alert, their Slack channel, Twitter,” Computing Sciences Area communications manager Carol Pott said.
“They set up a website and other communications options for people to get the latest alerts. They were trying to cover as many bases as possible to communicate with people who might not have access to the Internet or had other limitations.”
Dosanjh added: “Now, if there were a broader outage - one that affected the entire East Bay, for example - that would be more problematic for all the staff just in terms of being able to get access to things.”
Communication travels both ways, and NERSC’s efforts to keep services online prompted an outpouring of encouragement from many of the 7,000 researchers that use its systems. “I was really pleasantly surprised at all the emails and support that we got from the community,” Dosanjh said.
“The staff worked very hard, they're very, very dedicated to the lab's mission, which is to further human knowledge of science.”
Preserving the mission
One of the many cruel ironies of the shutdown of Cori was that it is one of the tools necessary to fight the ravages of a planet off-kilter.
One workload on Cori may be simulating energy storage solutions that help us break free from our addiction to fossil fuels. Another may be studying the impact of our seemingly inevitable inability to escape our addictive nature.
Cori has been calculating how high the seas will rise, and how large the tornadoes could grow. Just one week before Cori’s first shutdown, it was actually simulating how the forests would burn.
"Results from the high-resolution model show counterintuitive feedbacks that occur following a wildfire and allow us to identify the regions most sensitive to wildfire conditions, as well as the hydrologic processes that are most affected," an October paper studying Camp Fire by LBNL researchers Erica R. Siirila-Woodburn and Fadji Zaouna Maina states.
The Department of Energy “does a lot of simulations of Earth systems,” Yelick said. “So, simulating climate change, as well as looking at alternative materials for solar panels, materials for batteries, and a lot of different aspects of energy solutions.”
Some of this work was delayed by the two outages, pushing back valuable research efforts. "At the end of the year, yes there was some lost time for sure," Yelick said, but she stressed that no data was lost, and that due to the normal backlog of jobs to run on NERSC systems, it "for the most part just changes the delay that people were expecting."
But NERSC does support some areas of scientific research where time is everything. “There's several where it's a major deal,” said Peter Nugent, LBNL scientist and an astronomy professor at Berkeley. “The ones that the Department of Energy is involved with a lot are at the Light Sources - these are very, very expensive machines that they run and scientists get a slot of time and it can be anywhere from a half a day to a few days. And that's it.
“If they don't have these capabilities there for them, they lose their run. That's a huge expense and a huge loss. But because of the nature of the detectors that they're running there, gathering more and more data, it's not possible for them to process it locally and do everything they want. They need to stream it to one of these HPC centers and get things done.”
Nugent’s work is also incredibly time-sensitive. “The research that I'm involved with right now uses supercomputers to search for the counterparts to the gravitational wave detections that the LIGO/Virgo collaboration is making,” he said.
Nugent - senior scientist, division deputy for science engagement, and department head for the computational science department in the computational research division at LBNL - crunches data from the Virgo interferometer in Italy when it spots gravitational wave events, and then tries to capture details on the four-meter Victor Blanco telescope in Chile.
There’s a problem, however: The gravitational wave discovery “usually comes with a large uncertainty in the sky as to where it would be,” so Nugent has to “start taking a bunch of images to follow these events up, then stream this data up to the NERSC supercomputers to process it,” and then take more images, as he hunts for signs of the event.
“Time is of the essence, these are transient events - they fade very rapidly in the course of 24 hours, so we have to get on them immediately. We have to do this search right away. It's a tremendous amount of data.”
When successful, the information gathered can yield important scientific insights. “These are very interesting new discoveries,” Nugent said. “This is the merger of black holes and neutron stars, the latter of which has led to the discovery of where all the elements that are very high on the periodic table - gold, platinum, silver - come from.
“So when somebody calls you up and tells you 'Oh, by the way, the computers are going to be all down.' You're like, 'Oh crap, what can we do?'”
Thankfully, just a few months before, Nugent’s team had already begun to prepare for Cori going down - although at the time, they were thinking of scheduled maintenance.
“We were like ‘what happens if an event goes off during those two days that they're down, what can we do?’ Nugent said. “And so we've looked at porting our entire pipeline to a cluster of computers that are run by the IT department at LBNL, known as Lawrencium.”
To pull this off, Nugent’s team had already put its code in Dockerized containers, making porting to different systems easier. “We did that earlier in the summer when NERSC was down for maintenance, and it worked out really well.
"But then this next thing came up, and we couldn’t use Lawrencium because it [would also go down] when PG&E shut off the power.”
The researchers turned to Amazon. “We applied for and received a special educational grant that gave us compute time over there,” Nugent said.
“And we were able to - with enough advance notice of when this is going to happen - push all of our data, our reference data and our new data to AWS.”
The process worked, but “was sort of last minute,” Nugent said. “It's a real pain, but we managed to get it done and keep it going.”
With more time now, Nugent’s team are looking at other cloud and cloud-like services. “We would love to run it at NERSC all the time, but now we have a backup plan for when this occurs and we’re looking at making it so that it naturally just turns over and goes from one service to another, depending upon the status.”
Commercial providers could form a part of the solution, but Nugent hopes to use government systems where possible. “The Department of Energy runs some smaller clusters, so we're going to talk with them about how we could set something like this up in the future,” he said.
“This is something that the DOE is certainly very invested in making happen, because sometimes there are bugs and they have to take systems down.
“Experimentalists come to rely on these HPC centers more and more for doing their data processing, so they will need to have the capability to shift from one place to another.”
He, like many in the HPC community, envisions the ‘super facility,’ a virtual supercomputer that takes advantage of the best aspects of different HPC deployments and services and combines them.
“It's this idea that you can use the networking and the streaming of the data to a different resource and process it where you want.”
The super supercomputer
That may take time, by which point NERSC could be home to another huge supercomputer, the 100 peak petaflops Perlmutter system, expected to draw more than 5MW when it launches in late 2020.
The system is named for Saul Perlmutter, who won the 2011 Nobel Prize in Physics for observations of supernovae which proved that the expansion of the Universe has been accelerating. “Saul Perlmutter was the guy who hired me out at the lab back in '96,” Nugent said.
Perlmutter - the person - is currently looking for more distant supernovae: “Now we have a computer named after somebody who is hunting for the same type of explosions in space that we started with some 20 or so years ago,” said Nugent. “It’s come full circle.”
By the time it comes online, or when the ~20MW exascale NERSC-10 system launches in 2024, it is not clear how regular grid power cuts will become.
"It is during a very limited period of time where this is an issue," NERSC's Dosanjh said.
"Almost all of our rain occurs between November and April, so it's really primarily an issue in October and November.
"It's not something we worry about every day, but there are certainly - as we've learned - occasions where you can be dry, six months after the rain, and there's high wind and high temperatures.”
PG&E has warned that it might use preemptive blackouts on its millions of customers for up to a decade, as it catches up on maintenance it should have undertaken years ago. Communities will have to prepare for sudden outages and fear potential fires.
But the danger with what happened in California is perhaps not just that of loss of life or property. It is a loss of perspective. It is the danger that this becomes the new normal.
“We don't want that,” Nugent said. “We really, really don't want that.”
His hope is that “necessity is the mother of invention,” and that the impact of climate change “will get us to do some interesting things because of this.”
Within Cori, and its successors, small slivers of replica worlds full of interesting things wink into existence. A weather system here, a turbine engine there. Perhaps in one there is a world where this works out, where a pathway out of destruction is found.
But Cori can’t get us there, it can’t change consumption habits, or craft policy proposals.
No matter the state of the grid, it can’t change the world beyond its racks. That is our world.
We have the power. | <urn:uuid:16ac6994-0295-4377-a743-16ca5b1e0ed6> | CC-MAIN-2022-40 | https://www.datacenterdynamics.com/en/analysis/how-californias-wildfires-took-down-supercomputer/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335059.31/warc/CC-MAIN-20220927225413-20220928015413-00326.warc.gz | en | 0.97547 | 3,550 | 3.375 | 3 |
Car-borne data storage is becoming a vital aspect of cars, trucks and lorries as they get smarter.
Intelligent vehicles will use and generate a lot of data that needs storage inside the vehicle. Although early days, we can begin to understand how much storage could be needed, and its characteristics.
Let’s start by defining what we mean by an intelligent vehicle. This an autonomous or near-autonomous road vehicle, using IT to manage vehicle component functions and assist or replace the driver.
Steps to autonomous vehicles
The Society of Automotive Engineers (SAE) defines five levels of autonomy.
- L0 – none as human driver controls everything, except stuff like an engine fuel injection system
- L1 – a single driver process is automated, like auto cruise control or lane-keep assist
- L2 – Multiple driver functions taken over by IT, such as cruise control, lane-centring and steering and braking
- L3 – Conditional automation with human driver on standby to take over
- L4 – fully autonomous cars with no human driver control.
The higher the autonomy level, the more IT is needed, and the more data that needs to be stored. The data is generated by a car’s sensors, which can be outward-looking, like cameras, radar and lidar instruments, and also inward-looking, such as logged engine output, exhaust emissions and suspension spring rates.
The data can also be preloaded in the car; firmware for embedded computer systems for example, and sent to it over-the-air, such as map data for navigation systems.
As yet, there is no consensus on just how much data intelligent vehicles will generate.
According to Mark Pastor, archive product marketing director, Quantum, autonomous test vehicles typically generate 5TB and 20TB of data per day. This figure is higher for test vehicles than that anticipated for normal operating mode.
Stan Dmitriev, an author at Tuxera which develops automotive storage systems, says autonomous cars will generate more than 300TB of data per year – less than 1TB/day.
Dmitriev has not published the number of hours per year but 300TB for 365×24 usage seems unrealistic. For the purpose of this article we will assume that this refers to 12 hour car usage per day.
He says cars at up to L2 autonomy can generate about 25GB/hour. That means 300GB in 12 hours – 0.3TB/day.
At CES this month Seagate highlighted that a single vehicle equipped with Renovo’s (AV software platform company) software can generate up to 32TB/day per vehicle. This is the headline number for data that will be collected and stored at the edge before being migrated to the cloud.
This is a wide, wide range, from 5TB/day to 32TB/day.
John Hayes, CEO of self-driving car technology company Ghost, said about the 32TB/day of data generation: “The headline data requirements are off by about a factor of 10,000.” He thinks Seagate and Renovo are talking tosh in other words and implies around 3.2GB/day is more plausible.
Keeping it safe
How the data is stored will depend upon the IT design in the car. This can basically be centralised, distributed or some combination of the two. A distributed design will need storage for each distributed computing element. A centralised scheme will have a central data storage facility, and a hybrid scheme will have a smaller central facility, and various storage elements in sub-systems around the car.
We’ll assume a hybrid scheme for now, with a main controlling computer for navigation and driving the car, complemented with subsidiary component ones for engine management, suspension and braking control, etc.
It is generally assumed an intelligent vehicle will be connected over-the-air to a remote host such as the manufacturer or a robo-taxi fleet operator, sending generated data to the host and receiving reference information such as traffic alerts.
The smart vehicle must make instant braking and steering decisions to avoid hazards and it needs to store information to carry out such real-time functions. The communications network cannot be assumed to be 100 per cent reliable and the car has to store generated data between uplink periods.
There could be minutes or hours of interrupted communications, or even days if the vehicle leaves mobile coverage areas for extended periods.
At Seagate’s high-end 32TB/day estimate of generated data, a 30 day storage period to cover extended non-uplink time would require 960TB capacity.
The car’s central data storage system will be mission critical for what is in effect a mobile edge computing data centre. If it fails, operations can be compromised, causing loss of functions and even failure of the vehicle or an accident.
It seems intuitively obvious that disk drive storage is too unreliable in a car’s vibrating environment and wide temperature ranges. Also it is possibly too slow to retrieve data for instantaneous decision making. These issues indicate flash storage will be needed.
That is more expensive than disk. The mixed read-write workload and a vehicle’s operating life will make read/write endurance important. We can also see that, at the extreme data generation and storage period above, a 960TB SSD would add tens of thousands of dollars to a car’s cost.
A 30TB Samsung PM643 SSD costs $12,09.99 on CDW. We would be looking at $30,000 or so for a 960TB drive; that seems a totally unrealistic cost. The moral here is that data generation and its storage are going to need careful consideration by manufacturers,
Selecting components that can operate inside an automotive environment and store data for a vehicle’s lifetime will be crucial. (I’m driving a 13-year old car and that would blow an SSD’s warranted working life out of the water.) Initial mistakes can come back to haunt a manufacturer, as has happened with Tesla.
Tesla‘s data storage problems
Tesla has been caught out with limited flash endurance. InsideEVs reports that Tesla fitted a Media Control Unit (MCU) to Tesla Model S and X cars from 2010 to 2018. This controls the car’s main touchscreen and has a flash card soldered to a motherboard housing the NVIDIA Tegra Arm-based CPU. It is an 8GB SK Hynix eMMC flash card and stores the firmware for the system.
At launch the MCUv1 firmware was 300MB in size but grew to take up 1GB.
Car logging information is written to the flash card, and this fills the card up. That means there are no extra cells (over-provisioning) for use by the card’s wear-levelling function and the MCU’s functionality becomes affected. The web browser may not work, for example, or system startup could take minutes or even fail. The car is still drivable though.
As the cars developed the fault Tesla replaced the entire MCU board as a fix under warranty. Out-of warranty cars need a fix costing $3,000 or more by a Tesla support shop. The firmware and card were upgraded in 2018 to handle the logging load, and this MCUv2 has a 32GB flash card.
Tesla owners have set up a petition to replace MCUv1 units with MCUv2 ones. A Tesla owners’ forum also discusses the issue.
Poster DallasModelS said in April last year: “Love the car but I’m afraid that service centers have no empathy and are just cold. Just can’t make a product, you have to stand behind it. How does a MCU which controls 99 per cent of the function of the car fail in under 3 years?”
Tesla has been contacted for a comment. | <urn:uuid:2c82c1b4-de46-4fff-86ce-eec7c5a41df7> | CC-MAIN-2022-40 | https://blocksandfiles.com/2020/01/17/connected-car-data-storage-estimates-vary-widely/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335326.48/warc/CC-MAIN-20220929065206-20220929095206-00326.warc.gz | en | 0.934496 | 1,643 | 3.046875 | 3 |
Implementing a Quantum Approximate Optimization Algorithm on a 53-Qubit NISQ Device
(Phys.org) A large team of researchers working with Google Inc. and affiliated with a host of institutions in the U.S., one in Germany and one in the Netherlands has implemented a quantum approximate optimization algorithm (QAOA) on a 53-qubit noisy intermediate-scale quantum (NISQ) device.
In this new effort, the researchers created a QAOA and ran it on Google’s state of the art NISQ computing platform. As Harvard’s Boaz Barak notes in Nature Physics, their QAOA worked as a combination of smaller algorithms that have been created to run simulations on a quantum computer, such as simulated annealing. Such algorithms begin by presenting a random answer and then seek to improve upon it using quantum operators. Using the algorithm, researchers learned more about ways to reduce noise or mitigate its effects. They also learned more about the use of hyperparameters and possible ways to map key problems onto a quantum architecture. | <urn:uuid:edc27fbf-c008-42ca-bbea-9bb3ca5c946b> | CC-MAIN-2022-40 | https://www.insidequantumtechnology.com/news-archive/implementing-a-quantum-approximate-optimization-algorithm-on-a-53-qubit-nisq-device/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335491.4/warc/CC-MAIN-20220930145518-20220930175518-00326.warc.gz | en | 0.919057 | 224 | 2.625 | 3 |
RPC or a remote procedure call, in the computer world, is required for a communication session of the inner processes. That will also let the computer programs to search out subroutines or else procedures for execution using any other address spaces. But all this can be done without a programmer’s coding facilitation in order to perform the remote interaction. Anyhow, a programmer can write the same code for the both local and remote subroutine.
Another point to be noted is that when at issue software is being used the object-oriented principals for the communication purpose then RPC will be either known as remote invocation or else remote routine invocation. But for the implementation of this concept, incompatible technologies usage is must.
Well! Message passing procedure involves the certain actions like when a remote procedure call, containing request message for the execution of specified procedures, is initiated from the side of client to a remote server that will further send a response in this course of action to the client so as to continue the process. But you can find out the variations and intricacies in different kinds of implementations which can lead to different and incompatible RPC protocols. But during the call processing by server, the client will remain blocked and on finishing the server’s processing, the client can launch an asynchronous request (XHTTP call) to the server.
But in case of the remote calls, you can face failure due to unpredictable and irregular network problems too. And all this can be happened without knowing that these remote procedures were actually called upon or not. On the other hand, procedures without added effects (known as idempotent procedure) can be easily handled if called excessively.
In addition to this, an event’s sequence in the course of a RPC can be as following:
- The client dubs to the end (client stub) by means of a local procedure call (along with parameters).
- The stub (client side) performs marshalling by packing parameters in the form of a message which is forwarded but after making a system call.
- The client message is forwarded to the server via a kernel (at the server) which duty is to transfer the packets on their reception to the server side stub end and as a result this stub provokes the server procedure.
Remote Procedure Call Protocol’s Structure
The RPC message protocol can be divided into two distinct constructions:
- Call message
- Reply message.
RPC Call Message:
Every remote procedure RPC call message is consisted on unsigned integer fields for the identification of the each remote procedure. These fields can be as: Program numeral, Program version number and Procedure number.
RPC Reply Message:
But for a reply message (a request), RPC protocol can be varied in nature and its makeup is depending upon the call message acceptance and rejection by the server. A reply message towards a request may contain the information in order to distinguish the following conditions such as: is RPC carried out the call message effectively and is remote program not accessible on the distant system etc. | <urn:uuid:ae0c8e95-3c45-44d1-b9a7-da77c6376074> | CC-MAIN-2022-40 | https://howdoesinternetwork.com/2012/rpc | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030336978.73/warc/CC-MAIN-20221001230322-20221002020322-00326.warc.gz | en | 0.903761 | 615 | 2.984375 | 3 |
Now a lot of you like me are watching Westworld and are likely thinking, whew, humanity is good for at least another year, but that’s not where I’m going with this. Google developed an AI to help with diagnosis at scale, and it was impressively accurate in the lab but not so much in actual practice Thailand; in fact, the deployment at times slowed the diagnosis and didn’t speed it up.
But this was a field test, and initial field tests often fail, but this failure wasn’t due to a problem with the AI. It was due to the nature of the deployment, and the problems are fixable and far easier to fix than it would be to create a new AI.
I think this also showcases why so many AI project fail, for what Thailand found, is what a lot of IT folks find, and that is that if you don’t understand the limitations and strengths of the AI you are deploying that lack of knowledge will cause the project to fail.
Let’s talk about deploying AIs correctly this week and why the learning Thailand, and Google, got from this trial will help assure the success of future deployments.
Understanding The Limited Power Of A Current Generation AI
While science fiction programs like Westworld imply AIs that can replace humans, we are between one and three decades away from that kind of capability. What we have today are focused AIs that are designed, and trained, to do one thing exceedingly well.
In the case of the Google project that one thing is being able to determine, from a high-quality image, whether a patent had diabetes while, the lab the AI, which was developed by Google Health, demonstrated it could identify signs of diabetic retinopathy in 10 minutes with 90% accuracy.
This capability is vast for Thailand, where their clinics are struggling to care for around 4.5 million patients with only 200 retinal specialists. In practice, 20% of the images were rejected, patent screening dropped to 5 per hour, and some people had to wait days for a diagnosis.
Now the causes of this were poor internet bandwidth. The AI tool is cloud-based, and the tests were often done in remote areas with poor internet connections, making uploads of high-quality images problematic. Also, overworked nurses were often unable to capture high enough quality images for the system to scan, which is what resulted in the extremely high image rejection rate.
This failure means that the AI required high-speed internet and a method to take high-quality pictures of eyes to function, both hardly a surprise given how the system was trained. But, for some reason, those requirements weren’t met, and that is what resulted in the system failure.
Now the system was designed to be used by people that weren’t well trained, but the apparent assumption was that the lack of training was in medicine, not in photography. It did require that nurses implementing the tool have adequate cameras and training to take the necessary high-quality pictures. It should have never been deployed in areas with limited bandwidth.
There was one nurse who was well trained and who had adequate bandwidth, and she screened 1,000 patients with great success. It is interesting to note that those patents were okay with a machine making the diagnosis.
Wrapping Up: AI Deployment Isn’t Multiple-Choice
This somewhat failed trail is a showcase for why you must understand the design parameters of an AI before deployment. In this case, if you can’t properly train the nurses or supply them with proper photographic equipment or you have limited bandwidth where you want the system to function, you need a different solution because this one will fail.
If you saw the Ford vs. Ferrari movie, you may recall that one of the reasons the Ford GT won the race was because the Ferrari driver pushed his car beyond the physical limits set for it and blew the engine. Or, if you were a fan of Clint Eastwood’s Dirty Harry, one of his sayings was “a man’s got to know his limitations.”
If you understand and stay within the design parameters of the AI, you are likely to be successful; if you don’t, you’ll probably feel a bit like that Ferrari driver and maybe even hear Dirty Harry’s voice mocking you. And, my friends, that wouldn’t be a good thing. | <urn:uuid:15b08535-f979-4615-b549-7c36fe46c9f3> | CC-MAIN-2022-40 | https://www.datamation.com/artificial-intelligence/what-we-learned-from-googles-healthcare-ai-mixed-results/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030336978.73/warc/CC-MAIN-20221001230322-20221002020322-00326.warc.gz | en | 0.978396 | 911 | 2.875 | 3 |
The COVID-19 pandemic is, by necessity, leading to a flurry of innovation. And now drones are taking their turn in the limelight.
In a recent announcement, the UK government said that unmanned aerial vehicles (UAVs) could be ferrying vital personal protective equipment (PPE) from the English mainland to NHS staff on the Isle of Wight as soon as early May. The Windracers Ultra fixed-wing drones, which can transport 100kg for over 600 miles, should be able to make the crossing in ten to 15 minutes.
This is just one example of the radical technological change that is being nodded through to help meet the unprecedented challenge of COVID-19.
The usefulness of drones to the medical sector has been clear for several years, and many well-funded start-ups have been trialling services globally.
In developed countries, where road infrastructure is good, airspace regulation is strict, and health services have well-established logistics networks, companies have struggled to get much traction. The company Matternet’s trial in Switzerland was a rare success – ferrying medical lab samples between hospitals in Lugano aboard multi-rotor drones – though even that has had teething troubles.
Instead, most of the innovation has taken place in developing countries – where the need is more acute, and the barriers more surmountable – with Zipline making progress in several African countries. The company’s service uses fixed-wing drones to drop packages, including blood, to rural areas with poor roads.
So this new trial marks a welcome departure for the UK.
At Nesta, we’ve been exploring the future of drones in the UK for the past two-and-a-half years. We’ve also looked into a number of hypothetical uses for drones – coincidentally, including the transport of medical items across the Solent, the body of water between mainland Britain and the Isle of Wight.
What we found suggests that this trial is likely to have positive implications that last beyond the coronavirus crisis. Not only will it perform a valuable public service now, but it will also help dismantle some of the barriers that lie in the path of wider drone adoption.
In a 2018 report, we found that there is already some enthusiasm for public service drone use. But there are three broad problems that need to be worked through before they become commonplace.
The first is technical, involving the development of safe, long-distance flight (or “beyond visual line of sight” in industry lingo), autonomous piloting and precision flight. These rest on developing and improving communications networks and low-altitude air traffic management.
The government’s industrial strategy project, the Future Flight Challenge, includes over £100 million of public money for research and development into drones and other aerospace technologies.
But long distance drone flights across the Solent will be a great test case that builds a track record, helps generate data and boosts confidence in the technology. The government’s recently announced funding for other coronavirus-related programmes, including for drones, should help generate further innovation in the field.
Collaboration is key
The second issue is the lack of alignment. There are many drone companies with innovative ideas. In our 2018 survey of the industry, we found over 700 in the UK – and there would be many more now. There are also organisations that are trying them out, construction companies and infrastructure owners, in particular.
But in our research, and in subsequent workshops we held with innovators and potential customers in 2019, we found that people who should be talking to each other often aren’t. Government and regulation have been moving slowly, without learning lessons from industry. Technology developers aren’t involving end users, such as local councils or NHS hospitals, in their research and development. And in turn, those potential users often aren’t taking the risk of paying for pilot drone programmes.
But there is movement in the right direction. The Civil Aviation Authority’s innovation team has transformed the regulator’s approach over the past two years. It has set up an initiative to help firms with innovative ideas navigate the complex rules around safety and gain the permissions they need.
And this trial drone service will be a welcome addition, creating tangible evidence of how a drone service can benefit the NHS and forge links between the NHS, drone companies and regulators.
And then there’s the public. Does the public actually want drones? For what purposes? And operated by whom?
When we investigated this, we found strong support for the public service use of drones, and more suspicion around commercial or hobbyist use. But public opinion is still forming, which is hardly surprising as drones are not widely used and most people have not yet had to think about them.
But that will change as drones become more widespread. And the questions that will arise are not ones with right or wrong answers; they’re about values and priorities. The drone industry needs to engage with the public on this, not in a superficial way but to learn from them as much as educate them about the potential benefits.
The NHS trial won’t hurt. In fact, having a real-world case study like this might make these conversations less abstract. But the extraordinary circumstances of the coronavirus pandemic – during which the public has enthusiastically embraced economic and public health measures more extreme than any in living history – isn’t representative of normal times. And drones operating in city centres, rather than over the sea, are far more likely to cause controversy.
For now, however, we need to move fast. Engaging the public and finding out what they really think will need to come later, when the fierce urgency of the crisis is over and we can all take a more dispassionate view of the future. | <urn:uuid:2d4ec803-ef79-430e-899b-5d91c887dc52> | CC-MAIN-2022-40 | https://news.networktigers.com/industry-news/medical-drones-could-help-beat-the-covid-19-crisis/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337404.30/warc/CC-MAIN-20221003070342-20221003100342-00326.warc.gz | en | 0.959073 | 1,205 | 2.6875 | 3 |
Since their initial development in the seminal AI research paper Attention Is All You Need, transformer-based architectures have completely redefined the field of Natural Language Processing (NLP) and set the state of the art for numerous AI benchmarks and tasks.
What are transformer models? They’re an advanced artificial intelligence model that has benefited from an “education” the likes of which some dozen humans might gain in a lifetime.
Transformer architectures are typically trained in a semi-supervised manner on a massive amount of text—think English Wikipedia, thousands of books, or even the entire Internet. By digesting these massive corpora of text, transformer-based architectures become powerful language models (LM) capable of accurately understanding and performing predictive analytics based on textual analysis.
In essence, this level of exhaustive training allows transformer models to approximate human text cognition – reading – at a remarkable level. That is, not merely simple comprehension but (at best) making upper level connections about the text.
Recently, it has been shown that these impressive learning models can also quickly be fine-tuned for upper level tasks such as sentiment analysis, duplicate question detection, and other text-based cognitive tasks. Additional model training on some separate dataset/task relative to what the model was originally trained on allows the parameters of the network to be slightly modified for the new task.
More often than not, this results in better performance and faster training than if the same model had been trained from scratch on the same dataset and task.
Also see: Top 10 Text Analysis Solutions
Benefits of Transformer Models
1) Great with Sequential Data
Transformer models are excellent at dealing with the challenges involved with sequential data. Because of this, they act as an encoder-decoder framework, where data is mapped to a representational space by the encoder. Then they are mapped to the output by way of the decoder. This makes them scale well to parallel processing hardware like GPUs – a processor that is super-charged to drive AI software forward.
2) Pre-Trained Transformers
Pre-trained transformers can be developed to quickly perform related tasks. This is because transformers already have a deep understanding of language, which allows training to focus on learning whatever goal you have in mind. For example, named-entity recognition, language generation, or conceptual focus. Their pre-training makes them particularly versatile and capable.
3) Gain Out-of-the-Box Functionality
By fine-tuning your pre-trained transformers, you can gain high performance out of the box, without enormous investment. In comparison, training from scratch would take longer, and use orders of magnitude more compute and energy just to reach the same performance metrics.
4) Sentiment Analysis Optimization
Transformer models enable you to take a large-scale LM (language model) trained on a massive amount of text (the complete works of Shakespeare), then update the model for a specific conceptual task, far beyond mere “reading,” such as sentiment analysis and even predictive analysis.
This tends to result in a significantly better performance because the pre-trained model already understands language really well, so it just has to learn the specific task, versus trying to learn both language and the task at the same time.
Looking Ahead: Redefining the Field of NLP
Since their early emergence, transformers have become the de facto standard for tasks like question answering, language generation, and named-entity generation. Though it’s hard to predict the future when it comes to AI, it’s reasonable to assume that transformer models bears close focus as a next-gen emerging technology.
Most significant, arguably, is their ability to allow machine learning models to not only approximate the nuance and comprehension of human reading, but to far surpass human cognition at many levels – far beyond mere quantity and speed improvements.
About the Author:
Dylan Fox is the CEO of AssemblyAI | <urn:uuid:ab3cd1a5-97d8-433e-972c-010980a31bbc> | CC-MAIN-2022-40 | https://www.eweek.com/big-data-and-analytics/reasons-transformer-models-are-optimal-for-handling-nlp-problems/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337516.13/warc/CC-MAIN-20221004152839-20221004182839-00326.warc.gz | en | 0.928073 | 806 | 2.828125 | 3 |
Cyber attacks are rising globally. The number of cyber security cases reported in the country has increased three times in the last year. Overall, there were 1.16 million cases reported in 2020, compared to the reported 55,000 cases in 2017. The number of cyber security cases reported in the country has increased three times in the last year.
In order to control cyber attacks, an online community named OWASP was set up by cybersecurity enthusiast Mark Curphey on September 9, 2001. OWASP stands for Open Web Application Security Project. Although the name only states security for web applications, the scope of OWASP is not limited to web applications.
To help the internet community fight against cyberattacks and vulnerabilities, OWASP procured a list of top 10 attacks named OWASP Top 10 for multiple technologies such as Web Application, Cloud, Mobile Security, etc. The process of testing the top 10 security risks mentioned in OWASP top 10 is known as OWASP Penetration Testing. Penetration testing framework is a guide to perform penetration testing and includes a set of security testing tools required to perform pentest.
What is OWASP Top 10?
The Open Web Application Security Project (OWASP) is a non-profit organization with a mission to make secure applications with free online educational content and community tools. Their mission is to make software security visible by providing all the tools, techniques, and mindsets to increase the application security of any software product. In addition, they provide a bunch of free educational content on security testing and web application security and a variety of other resources for developers and IT professionals.
OWASP Top 10 is a standard awareness document for developers and web application security. The OWASP Top 10 does not address all possible vulnerabilities, but it accurately identifies the most common and most critical web application security risks. It focuses on the most common types of security risks web applications face, not all possible security risks. The OWASP Top 10 is a good standard of security expectations for new applications and a helpful security checklist for more mature applications.
Recent Trends in OWASP Top 10
After more than four years of research, the Open Web Application Security Project (OWASP) has released its latest list of the top 10 application security risks. Below mentioned are the new security risks that have been added to the updated list.
1. Insecure Design: Insecure design can also occur when a design checked for security vulnerabilities is not updated to account for known security vulnerabilities. Insecure design can occur at different levels of a given product.
2. Software and Data Integrity Failures: Data integrity is the state of being whole, authentic, and unbroken. There are many ways that software or data can fail to uphold integrity. Insecure deserialization, untrusted CDN’s, insecure CI/CD pipelines are how software fails to maintain the integrity of the data.
3. Server-Side Request Forgery: SSRF (Server-Side Request Forgery) is a vulnerability when an application makes a request to an unauthenticated, remote host and does not validate the request correctly. Any website that allows external requests to be made to other hosts based on user input is vulnerable to SSRF.
|S No.||OWASP 2017||OWASP 2021|
|A01||Injection||Broken Access Control|
|A02||Broken Authentication||Cryptographic Failures|
|A03||Sensitive Data Exposure||Injection|
|A04||XML External Entities (XXE)||Insecure Design|
|A05||Broken Access Control||Security Misconfiguration|
|A06||Security Misconfiguration||Vulnerable and Outdated Components|
|A07||Cross-Site Scripting (XSS)||Identification and Authentication Failures|
|A08||Insecure Deserialization||Software and Data Integrity Failures|
|A09||Using Components with Known Vulnerabilities||Security Logging and Monitoring Failures|
|A10||Insufficient Logging & Monitoring||Server-Side Request Forgery|
What is OWASP Penetration Testing?
In order to undergo OWASP Penetration Testing, let’s first understand what penetration testing is. A penetration test is a method of evaluating a computer system or network to identify vulnerabilities that a malicious attacker could exploit. Security experts carry out penetration testing to identify vulnerabilities in the target software or system. Penetration testing is an integral part of a more extensive information security process to ensure proper risk management, compliance, and systems administration. A penetration test is an excellent way to determine the extent of the damage a hacker can cause. Penetration testing can be done at any point in time to find vulnerabilities in the system.
OWASP Penetration Testing is a specialized type of security testing that focuses on attack vectors and vulnerabilities listed in OWASP Top 10. An organization’s security landscape is complex, and thus it is essential to test the organization’s security measures to ensure that they are working correctly. Penetration testing can help to ensure that an organization’s security measures are working correctly.
Understanding OWASP Top 10 in-depth
As mentioned, OWASP updated the list of vulnerabilities recently (in 2021). Let’s understand them in-depth one by one:
1. Broken Access Control
Broken Access Control vulnerabilities are among the most common security bugs found in web applications. Broken Access Control vulnerabilities, also known as missing authorization, occur when an application does not correctly check for authorization or does not check for authorization at all. This allows users to view or edit protected data they should not be able to view or edit.
2. Cryptographic Failures
Encrypting sensitive data is a necessary step that needs to be performed in every development process. The aim of implementing cryptographic measures is to make sure that the data is protected. Some common issues that fall under cryptographic failures are:
- Using HTTP to transmit sensitive data
- Hard-Coded sensitive data such as API Keys
- Weak Encryption being used
- Improper key management
- Missing Key Rotation mechanisms
Injection attacks are a common type of attack. They happen when a user enters a malicious payload to a website’s input field. This payload is then processed by the website and executed as a malicious script written by the hacker. The malicious script is what performs the actual damage to the website’s server and its data. There are several different types of injection attacks. All of them involve a malicious payload. The payload is entered via a form field, URL, or API (Application Programming Interface).
4. Insecure Design
Insecure design refers to a variety of security issues and not a single issue. The cause of insecure design is the same as the cause of insecure coding, which is a lack of knowledge and awareness of security vulnerabilities. Most of the time, security is not taken as a serious issue and is not included in the list of requirements for software development.
5. Security Misconfiguration
Security Misconfiguration moved from #6 in OWASP Top 10 – 2013 to #5 in the updated OWASP Top 10 – 2021t. Security Misconfiguration is a broad range of vulnerabilities such as:
- Default Credentials being used
- Server version disclosure
- Missing security headers such as X-Frame-Options
6. Vulnerable and Outdated Components
The most common issue in most web applications is using components with known vulnerabilities and out-of-date components. Usually, the issue is not due to a lack of knowledge but due to difficulty in the version migration.
Components like jQuery, Bootstrap, Angular JS, etc., are the most vulnerable components. These are the most used components in web applications. They are also viral components for hackers; vulnerabilities are more common in them due to their popularity.
7. Identification and Authentication Failures
Identification and Authentication Failures were previously known as Broken Authentication and have moved from #2 to #7. Common Weakness Enumerations(CWE’s) have also been added to this. Some common vulnerabilities that fall under Identification and Authentication Failures are:
- Missing Brute Force protection
- Weak password policy
- Missing or weak multi-factor authentication
- Improper session management
8. Software and Data Integrity Failures
Data integrity failures are still widespread in software engineering. It isn’t easy to maintain data integrity of your data if you are not careful with your code. Data integrity failures lead to security flaws. Insecure deserialization, untrusted CDN’s, insecure CI/CD pipelines are how software fails to maintain the integrity of the data.
9. Security Logging and Monitoring Failures
Log monitoring is a crucial part of any security program. It is one of the most crucial areas of log management that helps companies detect and analyze security events in near real-time. Yet, in 40% of organizations, log management systems cannot detect and analyze security events in near real-time. The reason is that log monitoring is often insufficiently customized and managed.
10. Server-Side Request Forgery
Server-side request forgery (SSRF) is a vulnerability when an application makes a request to an unauthenticated, remote host and does not validate the request correctly. In essence, the application is tricked into requesting a host that it thinks is local but is remote. This type of vulnerability is dangerous because the software running on the remote host is not authenticated, meaning that it may have been compromised or configured to execute malicious code.
Why is OWASP Penetration Testing important?
OWASP penetration testing is an integral part of any security program. It helps the organization find and patch security risks before hackers exploit them. OWASP penetration testing is a systematic approach that identifies vulnerabilities in an application’s design, code, and supporting systems (such as network, hardware, and hosted services) and determines how an attacker can infiltrate the application.
Penetration testing and other security practices, such as risk analysis and source code review can help you build and maintain a secure enterprise.
Three benefits of performing OWASP Penetration Testing
- OWASP Top 10 is a list of the most common security vulnerabilities. OWASP penetration testing focuses on that list and helps companies uncover security risks.
- OWASP Penetration testing is also helpful in discovering and documenting vulnerabilities, which can help system administrators prioritize their efforts at securing the system.
- OWASP penetration testing can help you achieve common security standards such as HIPPA, PCI DSS, SOC2.
5 Best practices to avoid vulnerabilities
1. Implement Proper Multi-Factor Authentication
Multi-factor authentication is a security measure that requires you to provide more than one form of identification before accessing a system or service. A multi-factor authentication scheme can be a password, pin or fingerprint.
2. Validate user input
Input validation is the first and foremost step to securing your application from malicious attacks. It can be defined as a process of checking and testing an input provided by the user or the application to make sure that it does not pose any security or integrity issues. This can be done by validating the input data’s type, format, length, range, consistency, etc. Input Validation can be broadly classified into two categories:
- Client-side Validation
- Server-side Validation
3. Implement Cryptography
Cryptography is the practice and study of techniques for secure communication. Cryptography protects sensitive information, such as information used in credit card transactions from being accessed by an attacker.
Even though cryptography has been used for thousands of years, it is still a vital part of data security today. Cryptography is the art of creating codes and ciphers that make messages unreadable by anyone except those intended. It is a way to store and transmit information securely.
4. Manage Access Controls
Managing access controls properly is an integral part of an application. Privilege Escalation is a type of security risk in which a low-privileged user can access user resources with privileges that originate due to missing or improper access controls in the applications and managing these privileges properly can help prevent this risk.
5. Proper error handling
Error handling and logging are two of the most important things to remember when creating development tools. Error handling is the ability for code to anticipate error situations and provide the best possible outcome. Logging is the ability for error handling to keep track of these errors either through a log file or in memory. When creating development tools, error handling and logging are the most important things to remember.
The list of secure coding practices is not limited to the techniques mentioned above. Check out Astra’s Ultimate Secure Coding Checklist to get rid of security risks.
How much does an average Pentest cost?
The cost of a penetration test is a fixed value. The cost depends on a lot of factors such as:
- Size of Organization
- Approach to perform pentest
- Scope of pentest
- Experience of Pentest vendor
- Time to perform pentest
- And much more!
Penetration testing pricing varies on a case-by-case basis, but you can expect the cost to fall in the range of $2,000 to $20,000.
Why choose Astra for OWASP Penetration Testing?
If you search for an OWASP penetration testing service provider, you will find that the options seem endless. There are many reputable penetration testing companies that you can choose from, but choosing the right one for your needs may not be easy.
You need to have complete coverage of the code, the security of the network, and even of the organization’s employees. You need to have a full scope of services that tells you precisely what is being done, how it is being done, and why. With Astra, you get all of this. Astra offers various penetration testing solutions, including OWASP penetration testing at pocket-friendly pricing. All you need to do is get in touch with an Astra-naut, and we’ll take it from here.
OWASP penetration testing is to make sure that security flaws are not present in the applications. OWASP penetration testing will check for the vulnerabilities present in the application. At Astra, We believe that it is our job to help businesses create secure applications. When it comes to application security, companies need to know that their applications are secure and secure enough.
1. What is the timeline for OWASP based pentest?
The initial scan for OWASP penetration testing takes 7-10 days for web or mobile applications, and 4-5 days for cloud infrastructures. Vulnerabilities start showing up in Astra’s pentest dashboard from the second day of the scan. The time-line may vary slightly depending on the scope of the pentest.
2. How much does penetration testing cost?
The cost for penetration testing ranges between $99 and $399 per month for web apps.
3. Why choose Astra Pentest?
3000+ tests, adherence to OWASP top 10 testing methodology, intuitive dashboard with dynamic visualization of vulnerabilities and their severity, security audit with simultaneous remediation assistance, multiple rescans, these are the features that give Astra an edge over all competitors.
4. Do I also get rescans after a vulnerability is fixed?
Yes, you get 2-3 rescans depending on the plan you are on. You can use the rescans within a period of 30 days from initial scan completion even after a vulnerability is fixed. | <urn:uuid:a5a07286-9905-4286-900d-5ae9f4a77eae> | CC-MAIN-2022-40 | https://www.getastra.com/blog/security-audit/owasp-penetration-testing/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337516.13/warc/CC-MAIN-20221004152839-20221004182839-00326.warc.gz | en | 0.912676 | 3,250 | 3.015625 | 3 |
What about this course?
This course begins with a general introduction to Software Defined Networks, including discussion of control planes and data planes, APIs, logical layers, types of SDN networks, and more. Students will spend 15+ hours delving into real-world implementation and deployment of OpenFlow, an SDN standard that allows a remote controller to interact with the forwarding plane of a network switch or router over the network, making it easy to deploy innovative routing and switching protocols in a network. Learn how having complete transparency into setting up a flow gives you new insights as to how networking can change. Starting with basic tunneling, or traffic engineering, and then advancing to setting up multiple group tables, and tuning flow priorities and flow aging. You will gain hands on the tools and tricks needed to quickly deploy SDN.
Instructor for this course
This course is composed by the following modules
SDN Key Ideas
Separation of Control & Data Plane
Control Plane vs. Data Plane
Standard Northbound APIs for External Programming
Network Service Chaining
Evolution from Traditional to SDN Underlay Networks
Enabling Technologies for SDN
Logical Layers for SDN
Types of SDN Networks
Deployment View of Different SDN Networks
SDN Review :: Part 1
SDN Review :: Part 2
SDN Review :: Part 3
What is OpenFlow?
OpenFlow Key Ideas
Components of OpenFlow
Definition of an Abstraction
Queue & Tables
OpenFlow Data Model
Data Structure Introduction
Table Selection/Flow Selection
Egress / Group Processing
Normal Operations :: Configuration
Normal Operations :: Events
Normal Operations :: Statistics
OpenFlow Demonstration :: Part 1
OpenFlow Demonstration :: Part 2
OpenFlow Demonstration :: Part 3
OpenFlow Demonstration :: Part 4
OpenFlow Demonstration :: Part 5
OpenFlow Demonstration :: Part 6
OpenFlow Demonstration :: Part 7
Common Course Questions
If you have a question you don’t see on this list, please visit our Frequently Asked Questions page by clicking the button below.
If you’d prefer getting in touch with one of our experts, we encourage you to call one of the numbers above or fill out our contact form.
Do you offer training for all student levels?
Are the training videos downloadable?
I only want to purchase access to one training course, not all of them, is this possible?
Are there any fees or penalties if I want to cancel my subscription? | <urn:uuid:a2ae4ff8-5688-4736-9385-238aea1d5c4c> | CC-MAIN-2022-40 | https://ine.com/learning/courses/introduction-to-sdn-open-flow | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338001.99/warc/CC-MAIN-20221007080917-20221007110917-00326.warc.gz | en | 0.805723 | 593 | 2.703125 | 3 |
What is email phishing?
Email phishing is one of the biggest digital threats of the 21st century.
Simply put, a phishing event is when someone attempts to get sensitive information by pretending to be someone or something trustworthy through email. Often, phishing attacks come in via email in form of bulk messages from trusted sources.
Most of us see these all the time.
Some of the potential sources of phishing that we might see include:
- Credit card companies
- Insurance providers
- Email providers
- Spoofed emails of co-workers
- Healthcare Organizations
Forms of email phishing
There are multiple forms of email phishing.
Spear phishing – Unlike standard phishing which goes out to a broad audience hoping that some, spear phishing is about the individual receiving an email. Often, it’s your membership in a club or as a customer of a specific company.
Business Email Compromise – Also known as “The Man in the Email” attacks. There are many variations but what it commonly refers to is when an attacker spoofs a business executive or a creditor.
They will then in many cases do some research and send targeted emails to people in the accounts payable department. The end goal of these attacks is a wire transfer to a fraudulent account, which can never be recovered.
Clone phishing – This is where a recipient has received a legitimate email that had legitimate attachments or links. An attacker will then clone that email, insert a malicious attachment or link, and resend the email to the same recipient using a spoofed address of the original sender. They often will do it under the guise of a resend or an update.
Whaling – These are messages aimed specifically at a company’s upper management (catching the big guys, which is where the term whaling comes from). Very often, attacks are written as a legal subpoena, customer complaint, or executive issue. It asks the recipient to respond somehow, to act, and it will be that action that will make them fall victim. This is especially problematic as the user’s name has power and their computer has privileges meaning their PC likely has access to the company’s most sensitive data.
The main goals of most phishing attempts is to gain access to steal money through and passwords to bank accounts, and personal information like social security numbers.
How to Protect your business from email phishing
Here are five ways to avoid phishing scams. There are both technology-based solutions that revolve around increasing your email security posture and behavioral which focus on user actions.
Security Awareness Training
Train Users with Security Awareness Training – As you can see from the examples of phishing above, the one common denominator is that someone opened an email or clicked a link, or downloaded a document from a source that they couldn’t be certain was legitimate. Human error is by far the number one way that someone gets hacked and that data is stolen. In fact, direct hacks are extremely rare and exceedingly difficult, whereas the Dark Web is full of malicious software that you can use to attack someone’s systems.
User training is a concerted effort to teach everyone in a company what is legitimate and what isn’t. Using real examples, a person learns to spot possible problems before they even open an email. For example, seeing your own email address in your inbox is an automatic no-go. People can learn to avoid these traps and that can save companies billions of dollars and a great deal of embarrassment. We recommend a free phishing test by KnowBe4.
Detonate attachments – Because attachments are the top delivery method of ransomware, malware and things like fake purchase orders, it’s vital to know what they contain before they are available to be opened by users on the network. System emulation in ContentCatcher attachment defense opens all unknown email attachments to see what they do before delivery. The system operates completely autonomously and users will never notice that attachment defense is in place.
Scans URLs dynamically – The second most popular way to deliver an attack by email is to include a link to a malicious site. In the few seconds, it takes for someone to click off of the site, malicious code can be inserted into the computer and thus onto the network. By dynamically scanning all links in email, ContentCatcher URL defense will block these threats before they can cause any damage.
A URL scanning system will process an email before delivery to the end user and scan each of the URLs in the email. Once those have been scanned the system will deliver the email. Any emails containing malicious websites will be flagged and removed from the email. Links will also be scanned again dynamically when they are clicked within emails.
Check email domains
Check email domains – Spoofing of email addresses containing your domain is a very common threat. With the exception of marketing emails, there are few reasons you should ever receive emails with your domain from the internet. All internally sent emails should never tough the internet. Anti-spoofing built into your email security solution looks to prevent malicious emails spoofing your domain from coming in from the internet. ContentCatcher has spoofing protection built-in.
External email disclaimer
Add an external email disclaimer– One of the most effective ways to keep users from opening and falling victim to phishing attacks (especially spoofed internal email addresses) is to warn them to be vigilant when opening email from sources outside the company. Email phishing attacks (with the exception of some forms of business email compromise) all come from outside the company network. We recommend a disclaimer in your email similar to this: “Note: THIS IS AN EXTERNAL EMAIL. It did not originate at (your company name).” | <urn:uuid:8b758d77-1351-48c7-9f06-d56bc4142072> | CC-MAIN-2022-40 | https://www.clearnetwork.com/how-to-reduce-email-phishing-risk/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338001.99/warc/CC-MAIN-20221007080917-20221007110917-00326.warc.gz | en | 0.943193 | 1,188 | 2.984375 | 3 |
What Is Data Modeling?
Data modeling is the process of applying structures and methodologies to the data in order to convert it in a useful form for analysis and gaining insight. By preparing a model of the data involved in an information system, you optimize the database design and gain understanding of the data flow within the information system.
The data modeling process converts complex software design into a simple, easy-to-understand diagram of the data flow. Data modeling tools then help to create a database structure from these diagrams. In this article, we will be covering the following data modeling tools:
- Erwin Data Modeler
- Apache Spark
- SAP PowerDesigner
- Edraw Max’s Database Model Diagram
- Oracle SQL Developer Data Modeler
- MySQL Workbench
- Enterprise Architect
- IBM Infosphere Data Architect
What Is Data Modeling Software?
Data modeling is an important step before developing a database for an application. It’s rather hard to develop a database if you haven’t worked out its underlying structure first. A good data model is an abstract model of specifics in the database, such as how the data is captured, how the data flows within the system, how is data entered in individual tables, and what checks and constraints apply to the data before storing them in the databases.
Setting up the database can be tedious and a kind of scutwork programmers, database administrators, and data scientists don’t want to bother with. To make this tedious job simpler, or at least tolerable, data modeling tools are available for beginners and experts alike. Below are 10 of the most popular.
How to Select a Data Modeling Tool
- Usage and Needs: This is your first deciding factor. How are you going to use it, how extensively, and for what specific purpose? Not all data modeling tools are created equally and some have different areas of focus. Know your business requirements first to pick the right tool.
- Scalability: Your project might start small but grow in requirements. At that point you don’t want to be hamstrung by an incapable modeling tool. Give yourself a little room to grow.
- Features: Once you have a grasp of your business requirements, the next step is to determine the right tool. Not all tools are created equal, some are very niche or specific. Shop around.
- Integration: Be advised that some tools create a data model in a proprietary format and some use common or open file format, which is easily read into other tools. Make sure your modeling tool plays nice with your database.
- The User Community: Most every business tool has a user community which you can turn to for help. See what kind of community is around your tools of interest.
10 Data Modeling Tools
This is not a complete list of tools, but we did try to include the most popular and widely used. The list is in no particular order.
This data modeling tool is known for being powerful but less expensive tool for business than others and also complies with the governance database rules. It is considered one of the best data modeling tools, because it includes automated schema generation, cloud-based data solution and power to create hybrid architecture.
If it’s Apache, you know it works well with other Apache products, like Hadoop. Good at handling large databases and parallel tasks.
RapidMiner is ideal for those who have never used any data modeling tool before, because it is easy to use. It can easily be incorporated through any data source types, including Access, Teradata, Excel, Oracle, Microsoft SQL, Ingres, IBM SPSS, IBM DB2, MySQL. It uses visualization of pipelines to create analytics based on the settings of real-life data.
SAP Powerdesigner is capable of capturing, analyzing and presentation of business data. It follows industry best practices to give comprehensive coverage on the metadata storage and get an understanding of the input data. Supports a very wide range of databases.
Edraw Max’s Database Model Diagram is among the best of the free and open source tools to create a database model diagram. It uses a drag-and-drop interface to rapidly build tables and easily redesign them.
Made for the Oracle environment, Oracle SQL Developer Data Modeler is a data modeling tool which also supports physical database design. It covers capturing data, exploring data, managing data and getting insights from the data.
Ideal for beginners, this design tool is used to design SQL databases without having to write a single line of code. It allows you to manage large and small databases and data models seamlessly, import existing database schema and creating a physical model or ERD of your database.
Designed specifically for the MySQL database, the MySQL Workbench is a unified data modeling tool for database architects, developers and database admins. It provides tools for configuration, visual database design, administration, backup and deployment.
Enterprise Architect is ideal tool for both entry level modelers and advanced alike. It comes with a lot of functionalities and strategies for analyzing, visualizing, testing and maintaining all the data in any enterprise landscape. It uses diagram-based modeling, can pull in data from various domains and locations throughout the enterprise to create a single, unified version of the model.
Infosphere Data Architect is a data modeling tool from IBM built on the Eclipse Integrated Development Environment. InfoSphere is known for its ability to discover patterns within the data, model the data, find relations and also standardize the interfaces between various applications, servers and existing databases. | <urn:uuid:10b9e47a-c4f8-46c3-918c-682deb6fb4f0> | CC-MAIN-2022-40 | https://www.datamation.com/applications/top-data-modeling-tools/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338001.99/warc/CC-MAIN-20221007080917-20221007110917-00326.warc.gz | en | 0.890747 | 1,149 | 3.28125 | 3 |
We’re all relying on the internet more than ever — the last thing you need is for your connection to slow you down. But how much internet speed do you need? And is speed the only thing that keeps your internet hopping along?
Speed comes into play differently when a significant portion of an organization’s staff is working remotely. The company needs enough bandwidth and the appropriate routing protocols to handle the extra internet traffic to and from their servers, and employees working remotely need to have fast enough internet at their locations to stay productive. Here’s what you need to know:
Quick review of the lingo
Broadband: A broad (wide) band of frequencies that allow many different types of traffic to be sent or received at the same time — voice, video, data, and TV all on the same connection.
Packets: Small pieces of information (data) on the web that contain the text, images, and video you experience on your device. They travel on the information highway as small pieces of a puzzle. You need all the pieces on the receiving end to understand the message being sent. Packet loss is when some packets don’t get where they’re going so you get jittery or inconsistent performance.
Bandwidth: The maximum (width) or capacity of your connection (highway). The greater your bandwidth, the more packets you can send at one time.
Mbps: “Megabits per second” is the unit used for measuring how much data (packets) are being transferred per second. Higher Mbps means more data.
Broadband: High-speed internet with a minimum download speed or 25 Mbps and a minimum upload speed of 3 Mbps.
Latency: How long it takes your data to make it to its destination. This varies due to distance (geography), congestion, filters and other circumstances and is represented by the number of ms (milliseconds) when you test your Internet speed.
Peering: No one carrier owns the entire internet. The internet is comprised of many different companies all connecting to each other. Peering is when internet carriers buy connections from each other to route traffic to destinations or from destinations not directly connected to the network. Internet congestion typically happens at peering points (like cars on an interstate’s on- and off-ramps).
Routing: The process of selecting a path for traffic in a network, or between or across multiple networks.
ISP: Internet Service Provider. ISPs include companies like Comcast, AT&T, Google, Verizon, and smaller or local carriers.
Service Level Agreement (SLA): Contract between an ISP and an end-user that defines the expected level of service.
Routing protocol (warning — geek alert): Determines how routers communicate with one another. The most common protocols include OSPF (Open Shortest Path First, which finds the shortest route), BGP (Border Gateway Protocol, which dynamically updates routes), RIP (Routing Information Protocol, which shares information about shortest known routes), and Cisco’s older proprietary IGRP and EIGRP (Interior Gateway Routing Protocol and Enhanced IGRP).
Deciphering speed for your business network
STEP 1: Know if you’re too slow
Are your customers or employees complaining? If your company is using less than 80% of your available internet bandwidth, it’s probably good enough to keep everyone happy. If you’re routinely using more than 80% of the available bandwidth, you will be getting complaints. However, just because you have slow internet does not mean that you do not have enough bandwidth. The problem could be with bandwidth, latency, or a combination of the two (or even other reasons).
STEP 2: Measure the speed of your network
Ask your IT department to report on the throughput you’re actually getting. They’ll collect data from your firewall or router and can also contact your ISP for a report.
STEP 3: Compare what you’re using to what you’re paying for
If your ISP isn’t living up to your agreement, you need to contact your ISP. And it’s much easier to get your ISP to hold up their end of the bargain of the SLA when you have the data in hand. If they are making good on your SLA, then you need to do more troubleshooting.
STEP 4: Determine if your slow network is caused by latency or bandwidth issues
Bandwidth: If you’re routinely hitting your SLA’s max bandwidth, you don’t have enough capacity to cover the times you need to push more packets through the pipeline (bursts). For companies leveraging the cloud, as a rule of thumb, you need roughly 1.5 Mbps per employee. For example, a 50-user company using Office 365 and Salesforce should have an aggregate of 75 Mbps of internet bandwidth while companies whose employees use the internet for email and web browsing only need much less. Another rule of thumb is to look at your company’s average speed and then double it — you want 50% more speed available to handle bursts.
Latency: If your bandwidth looks good, then your IT team will look at the routes your network uses. There could be bottlenecks leaving your facility or traffic jams at peering locations or other points along the route. Once the problem areas are located, your IT team should work with your ISP to reroute your traffic, which will speed it up. Buying more speed from the ISP, even if your ISP wants you to, won’t fix a peering problem. Internet routing protocols are supposed to find the best routes automatically but sometimes they don’t. Think of it as GPS routing for a road trip. You can get stuck in traffic if your GPS doesn’t reroute you around an accident. The newest routing protocols look at response time in addition to the shortest distance.
Latency plus bandwidth: Sometimes, you have more than one problem going on. Start by solving the bandwidth issue by adding capacity then work on the latency issues by fixing the routing.
STEP 5: If you need to buy more capacity, consider buying it from a different carrier and/or different channel
Cover your bases by combining two different ISPs (and types of circuits) to arrive at your ideal Mbps. That way if there’s a problem with your main channel (or a problem with your carrier), the other can kick in. For most companies, it makes sense to use broadband for most of their internet use and add fiber, which is more expensive, for bursts or specific types of traffic. With the right tools, your IT team can classify traffic into different categories then prioritize the categories. And if your team is really geeky they can even prioritize certain websites.
STEP 6: Monitor your network
Your IT team can respond to problems faster if they continually monitor your network for problems, including security issues. Often, they’ll be able to reroute traffic with your ISP before your staff is affected very much.
Part of Leapfrog’s service is to continually monitor internet service and advocate on our client’s behalf with IT vendors when there’s a problem. We use tools to measure usage and pinpoint information about network congestion down to specific locations — it helps to have the data in hand when troubleshooting. We strongly recommend companies have more than one internet connection to reduce the likelihood of problems with speed.
Deciphering speed for your home network
Step 1: Test your current speed by visiting speedtest.net
Speedtest.net will give you three numbers: your ping, your download speed, and your upload speed. Try it at different times of day to see how the speeds fluctuate — your fastest times will probably be in the middle of the night when the web has less traffic. Once you know your numbers, you can see how they measure up to the recommended standards below. You can also see if your Internet Service Provider (ISP) is delivering what you’re paying for.
Step 2: Understand what each internet highway offers
- DSL, or Digital Subscriber Line, delivers internet over phone lines that were built for voice, not data. It maxes out at around 7 Mbps, which does not qualify as broadband.
- Cable delivers internet over lines that were built for TV, so it provides a wider highway than DSL and meets the 25 Mbps minimum for broadband. Speeds range from 20-100 Mbps depending on the provider and geographic location./li>
- Fiber optic cable is the fastest and can offer Gigabit speed, or speeds of 1000 Mbps or faster (that’s 1 million bits per second). It’s not available everywhere and you need equipment that can handle these speeds to take advantage of it.
Step 3: Know what your home needs:
First, determine what you use the internet for and how many devices are connected at the same time. Are one or more people in your household working (or schooling) from home? And are other people gaming or watching Netflix at the same time? For a basic guideline, see the FCC’s speed guide or Household Broadband Guide. You can add up the numbers on the FCC charts and use these download speed recommendations after adding up your totals:
- 5Mbps or less: Basic web surfing and email
- 5-10Mbps: Web surfing, email, occasional streaming and online gaming with few connected devices
- 10-25Mbps: Moderate HD streaming, online gaming and downloading with a moderate number of connected devices
- 25-40Mbps: Heavy HD streaming, online gaming and downloading with a lot of connected devices
- 40+Mbps: Hardcore streaming, gaming, and downloading with an extreme number of connected devices
Step 4: Choose your speed … for right now
Once you’ve answered the question of how much internet speed you need, compare it to your current plan. You may need to upgrade your plan along with your modem and router to get the speed you need. If you’re unsure your modem and router can handle faster internet speeds, just ask your ISP — they may offer you a good deal on new ones. Keep in mind buying them outright can be less expensive in the long run than renting them from your ISP.
Also, if you see ads touting Mbps that are ten or more times higher than those listed above — and those that are in the Gbps ranges — keep in mind these super-high speeds are maximum speeds under optimal conditions, which include four things:
1. Your highway infrastructure (DSL, cable, fiber)
2. Your modem
3. Your router
4. All of the servers along the highway to and from your home
If just one of these things can’t handle the speed, you won’t get the advertised lightning-fast speeds. But do you really need it? Now you know! | <urn:uuid:92ea3e82-ce85-4b98-a72d-cc9207b52808> | CC-MAIN-2022-40 | https://leapfrogservices.com/how-much-internet-speed-do-you-need-data-speeds-deciphered/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334528.24/warc/CC-MAIN-20220925101046-20220925131046-00526.warc.gz | en | 0.926675 | 2,249 | 3.046875 | 3 |
Projections in a new report from the National Oceanic and Atmospheric Administration provide valuable foresight that can help local governments act before the water rises.
As sea level rises, it can be easy to miss the subtlety of higher water. It’s much harder to overlook saltwater more frequently flooding streets, impeding daily life and making existing problems worse.
The frequency of high-tide flooding along the U.S. coasts has doubled since 2000, and it’s expected to increase five to 15 times more in the next 30 years, the National Oceanic and Atmospheric Administration warns in a new report released July 14, 2021.
I work with coastal communities in the northern Gulf of Mexico that are facing the risks of rising seas as they try to avoid preventable damages and costs, such as infrastructure failures and falling property values. Information like the NOAA report is critical to helping these communities succeed.
Last year, the U.S. averaged four days of high-tide flooding, but that number doesn’t tell the whole story – regionally, several areas saw far more. There were record-breaking numbers of high-tide flooding days in 2020 along the Gulf of Mexico and southeast Atlantic coasts. The city of Bay St. Louis, Mississippi, jumped from three days of high-tide flooding in 2000 to 22 days in 2020.
NOAA’s annual high-tide flooding report projects a national median of three to seven days of high-tide flooding this year, again with differences expected regionally. The western Gulf Coast, including Texas and Louisiana, is expected to see the most tidal flooding days, ranging between seven and 15 days. The northeast Atlantic is projected to have six to 11 days of high-tide flooding. The Pacific Coast is projected to be lower than the other regions.
Calling it ‘nuisance’ flooding overlooks damage
High-tide flooding impedes the use of roads and increases wear and tear on stormwater and wastewater systems. The impact can seem minor, but as the frequency increases, these seemingly inconvenient flood days can have long-lasting impacts.
Already, areas at risk from sea level rise have seen decreases in property values, particularly where cities and homeowners haven’t taken steps to increase flood resilience. Insurance premiums are beginning to increase to reflect actual risk, and bond ratings are increasingly being tied to the resilience efforts of communities.
Flooded roadways can create dangerous situations where first responders struggle to safely reach those in need. Businesses get fewer visitors and feel the loss in depressed revenues. The more often this happens, the more it ripples through coastal economies. It can affect tax revenues and erode community ties.
Sea level rise disproportionately impacts poorer, marginalized communities, and the impact of high-tide flooding has been no exception. People living in some of the most underserved coastal communities are facing increases in their insurance premiums because of the flood and storm risks, sometimes with more than 90% of the insurance policies in a single ZIP code projected to increase.
Ways to reduce the threat of high-tide flooding
NOAA’s projections provide valuable foresight that can help local governments, property owners and other coastal stakeholders to act before the water rises.
Communities can upgrade their infrastructure, such as raising roadways and installing backflow preventers in stormwater systems, and modify building standards such as increasing freeboard, the distance required between the first floor and base flood level, or designating base flood elevations outside of current FEMA flood zones to help prepare communities to withstand higher seas. Communities can also work with nature to preserve and restore coastal habitats that provide natural flood protection such as marshes and barrier islands.
Pensacola, Florida, is one example of a city being proactive. It recently completed a sea level rise vulnerability analysis to determine where high-tide flooding will begin to strain infrastructure, low-income neighborhoods, economic hot spots and critical facilities. The city was able to recommend where to prioritize action and what kind of action will be necessary to keep high-tide flooding from being costly or as damaging.
The message from the new report is clear: High-tide flooding and other more severe types of flooding have already increased with sea level rise and are expected to accelerate in the coming years. Communities have an opportunity to act now to reduce the impacts.
Residents in any coastal community can reach out to their local governments to encourage forward thinking. For more information on how to get involved in coastal resilience, almost all coastal and Great Lakes states have a coastal resilience specialist within their Sea Grant programs. Each regional NOAA Office for Coastal Management can offer guidance on how to get involved, as well.
This article was first posted on The Conversation.
NEXT STORY: Air Force plans 3D scans of facilities | <urn:uuid:200e984d-9aea-4593-903c-2d07bcee7b38> | CC-MAIN-2022-40 | https://gcn.com/data-analytics/2021/07/high-tide-flood-risk-is-accelerating-putting-coastal-economies-at-risk/315615/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335059.43/warc/CC-MAIN-20220928020513-20220928050513-00526.warc.gz | en | 0.946327 | 983 | 3.171875 | 3 |
Wednesday, September 28, 2022
Published 3 Months Ago on Monday, Jul 04 2022 By Adnan Kayyali
The link between games and mental health is not an obvious one, but many gamers can attest to the effectiveness of some types of video games when it comes to improving or aiding in their mental health struggles. There’s nothing like spending a few hours isolated from the outside world, just you and you’re in game characters, missions, goals, and gameplay.
Everyday life can get monotone, and there are few things as engaging and exciting as videogames that grip both our mind and body. Engaging activities such as video games can do wonders for your mental health, sometime just by giving you a break from intrusive thoughts, occupying your hands, granting a sense of fulfilment, and more.
Depression, general anxiety, post-traumatic stress disorder (PTSD), obsessive-compulsive disorder, and panic disorder are some of the most prevalent mental health conditions that the general public experiences and are diagnosed with. These conditions are surprisingly prevalent all over the world.
According to the World Health Organization, 264 million people worldwide suffer from depression, along with 45 million people who have bipolar disorder and 20 million people who have various psychoses. As of 2017, there was 13 percent of people worldwide had a recognized mental or substance use disorder.
Despite the fact that this is a relatively new area of research, some scientists believe video games can aid in patients’ acceptance of their mental health issues. In their opinion, if some restrictions are put in place before playing video games, they may benefit people who are having mental health issues.
Stress relief is the first and most obvious benefit. Players can engage in low-stakes problem-solving to challenge their minds and take their attention away from whatever is upsetting them.
Finding solutions, resolving problems, amassing accomplishments, and thinking about cause and effect is crucial for developing brains and people struggling with mental health conditions like depression and anxiety.
Working toward a goal or desired outcome, even if it’s “just” on a mobile phone or tablet screen, is, in the end, an upbeat and mood-enhancing activity that may benefit your mental health and state of mind. Games are available to help people in treatment for conditions like PTSD, ADHD, and trauma recovery on their path to wellness. Some video games even aid in the growth or restoration of social or caregiving abilities, some of the first things to suffer during trying mental-health episodes.
Depending on the game’s structure, emotional resilience—the ability to fail gracefully—and teamwork are other advantages of video games for mental health. Sports and physical activity participation is frequently the first piece of advice given by health professionals regarding mental health. But video games open up a world of opportunities for those who have mobility issues, offering them the same experience in team dynamics as physical sports.
We have heard it all before, but it is true; Video games can be detrimental to our mental health. However, it is not the games themselves that can be mentally damaging. What
One might wonder if this perception will change with the rise of GameFi and Play-to-Earn games coming with the advent of blockchain technology and the proliferation of Non-Fungible Tokens in the gaming space. Perhaps these engaging and highly compelling games can be a productive way to pass the time soon if done right.
But it remains a fact that not all games are good games for mental health. I’m willing to bet that Flappy Birds would make a Buddhist monk turn red. The brain chemistry involved with such games is more akin to an addiction-like stimulus more so than anything else. Games such as Cookie Clicker appeal to a player’s greed, and shooter games accommodate players’ desire to win and their hand-eye coordination skills.
There are some types of video games that appeal to our better nature or to a calmer aspect of our psyche. These video games have the potential to develop and nurture those parts of our brains as well.
Interactions with the world, both real and virtual, spark certain brain reactions, and while some come from a place of anger, hatred, spite, and negativity, there are some games that do the opposite.
Some video games foster the better parts of our human psyche and train the parts of our brain that handle problem solving, teamwork, and empathy.
The kinds of interactions you have in virtual caretaker games, which involve taking care of and cultivating plants, animals, or anything else, can activate and engage brain regions that are involved in controlling interpersonal interactions and learning how to take care of others or oneself.
Play or focused training can sometimes be all you need to improve your mood, reconnect with your inner voice, or regain control of your physical abilities. The games that follow offer countless opportunities to test your mental stamina and quickness of thought.
Puzzle games never go out of style, remaining among the most popular games out there, even in AAA contexts. There are countless brilliant and beautiful puzzle games available for tablet and smartphone owners to relax their minds and exercise their problem-solving skills.
More than others, some video games explicitly promote improving one’s mental state. If you’re experiencing anxiety about the present or worries concerning the future, this might be the genre for you.
It is not easy to place a label on video games and mental health effects unless you look at the science, but we do know that the brain is a muscle, and some games train certain muscles over others. It is up to you to choose which muscles you want to train.
The world of foldable phones keeps welcoming more additions to its roster. And it makes sense. The foldable phones are selling well even with their pricy asking point. Huawei’s latest foldable is the Huawei P50 Pocket. While it does many things right, it also has its shortcomings. We will take a deeper look at it. […]
Stay tuned with our weekly newsletter on all telecom and tech related news.
© Copyright 2022, All Rights Reserved | <urn:uuid:07e8171c-2817-4068-ade4-2c974f359751> | CC-MAIN-2022-40 | https://insidetelecom.com/the-link-between-video-games-and-mental-health/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335059.43/warc/CC-MAIN-20220928020513-20220928050513-00526.warc.gz | en | 0.952284 | 1,270 | 3.140625 | 3 |
AI made history on Saturday as neural networks defeated human world champions in a best-of-three contest at Dota 2, a popular and complex online strategy game. OpenAI Five, the AI agent developed by the namesake research lab, managed to perform the feat after losing in a match-up against professional Dota 2 players in a famous tournament in August.
Games have historically been one of the main arenas to test the progress of artificial intelligence algorithms. Initial efforts involved developing AI models that could play board games such as chess and checkers, and eventually the complicated Chinese game of Go. More recently, AI researchers have turned their attention toward video games, which are significantly more complicated and challenging.
In this regard, OpenAI Five’s victory marked a milestone achievement for the AI community.
In Dota 2, two teams of five players compete against each other. The ultimate goal of the game is to destroy the enemy’s tower, known as the “Ancient, while preventing the destruction of your own. Each player can play one of the many “hero” characters the game features. Mastering Dota 2 requires combat tactics, resource management, using special skills and developing long-term strategies. The game is played in real time (as opposed to turn-based), which makes it all the more difficult.
In a nutshell, Dota 2 is a game that is easy to learn, hard to master.
OpenAI has yet to release the technical details of its Dota 2–playing AI. But it has released some initial information about how the model has developed its ability to play the game. The new AI is an advanced version of the original model, which OpenAI introduced in June 2018.
OpenAI Five is a team of five neural networks (hence the name), one for each of the five hero characters in a team. Neural networks are software constructions that develop their behavior by analyzing large data sets and finding correlations and patterns.
OpenAI Five trains its neural networks by using reinforcement learning, a subset of AI where the model is given the rules of the environment and a reward to pursue. The AI is then left to its own devices to try different combinations and figure out successful sequences that can maximize the reward.
In the case of Dota 2, OpenAI’s neural networks must find combinations that will help them move toward the many small and large goals of the game, such as gathering resources, making their heroes stronger, destroying enemy heroes and destroying the enemy team’s Ancient.
Comparing OpenAI Five to human players
While at first glance, reinforcement learning roughly mimics the way humans learn to play games, beneath the surface it is very different. Neural networks take in huge amounts of data, much more than a human needs to master a game. They also need a lot of compute power.
OpenAI Five trained on 45,000 years’ worth of games in ten months, consuming 800 petaflops per second. To put that in perspective, an Intel Core i7–970, which is one of the most powerful PC processors, averages at 109 gigaflops per second.
The AI’s playing style is also different from that of human players. In the first game against the team of human champions, OpenAI Five used tactics that seemed peculiar. For instance, the AI used in-game currency to immediately revive dead heroes, even early in the game, something that professional players don’t usually do.
According to OpenAI CTO Greg Brockman, the AI favors strategies that yield short-term gains. This shows OpenAI’s shortcomings in long-term strategizing and planning, a characteristic that human players acquire with little training. However, these same short-term tactics helped OpenAI beat human champions.
We see this happen in test games all the time: the bots buy back, the humans laugh, and then the humans lose. Hard to know if it’ll happen here too…
— Greg Brockman (@gdb) April 13, 2019
But the entire comparison of AI and human intelligence is flawed, as some experts point out. “Humans are cheating to some degree as [Dota] was designed with an average human in mind. Humans have (quite inefficiently) evolved over a long period of time to be highly effective at many of the tasks demanded from a game of DotA,” says AI researcher Stephen Merity, in written comments to TechTalks.
Humans already have an understanding of the many concepts in Dota, such as combat, defense, resource planning, cooperation, and more. AI models start with a clean slate and with zero knowledge.
“Machine learning algorithms come in to this with relatively few preconceptions about the task. The 45,000 years of training here is obviously a great deal, but the model is learning a complex set of subgoals and sub-objectives that eventually result in the model winning or losing,” Merity says.
The limits of artificial intelligence in playing Dota 2
Despite the vast amount of resources OpenAI has at its disposal to train its neural network, it is still very hard to create an AI that play perfect Dota 2 with all its various parameters. That’s why OpenAI introduced limits to the game to make it a little easier for the AI.
Out of the 117 different characters available in the game, OpenAI limited the competition to 17 characters. Given that each game involves ten heroes, this reduces the number of possibilities from approx. 89 trillion (117 choose 10) to 19,448 (17 choose 10).
“The combinatorial possibilities are the main reason why DotA is fascinating to human players. Whilst OpenAI Five is certainly a success, the fact that the game would likely still fall to an amateur player when it had to play on ‘real ground’ (i.e. all character combinations) is still a major limit,” Merity says.
It’s worth nothing that our calculation has not taken into account the different strengths and weaknesses each character type has and how that would affect the training of the network. Current AI technologies are not good at learning abstract concepts and transferring knowledge to new situations. If we see something new in the game, like a character we haven’t seen before, we can quickly make decisions based on our previous experience and knowledge. For AI, a new character is almost like a new game that it has to learn from scratch. That’s why changes to the game’s parameters require a huge amount of training to raise the AI to professional level again.
According to OpenAI, they had to remove one of the characters from the competition because its abilities had changed in a recent update to the game.
“The models produced by OpenAI Five are still not flexible compared to standard human competition, where patches that change character abilities do sometimes come at very inconvenient times,” Merity observes.
The strengths of artificial intelligence in playing Dota 2
While we examine the weaknesses of AI’s gameplaying skills, it’s also important to underline its strengths. In 2018, OpenAI Five lost to human champions. It turned out that the AI needed more training.
“We were expecting to need sophisticated algorithmic ideas, such as hierarchical reinforcement learning, but we were surprised by what we found: the fundamental improvement we needed for this problem was scale,” OpenAI notes in its blog post.
So while AI can’t mimic humans’ abstract thinking and commonsense, it can perform its own type of “thinking” and “learning” at a very fast pace. Training OpenAI Five in super-fast forward for ten months brought it to the level of champions. According to OpenAI, the new model wins against the old AI in 99.9 percent of games.
“Like Deep Blue’s chess, this is a problem where you can throw substantial compute at generating different scenarios,” says Merity, referring to the AI that won against world chess champion Garry Kasparov in 1997. “Indeed no human could have manually curated or annotated those many centuries of gameplay!”
OpenAI Five’s evolution into a champion Dota 2–playing bot is a reminder that so far, successful AI methods are those that can scale as data and compute resources become increasingly available. While this is not an approach that works in all scenarios, it surely helps in areas like playing games, where the AI has to explore and compare a large number of different scenarios and combinations.
Cooperation between AI and humans
One of the interesting features of Saturday’s event was the cooperation between humans and AI. After the human vs AI competition, OpenAI set up a match in which each team was comprised of two human players and three bots.
“Our testers reported feeling supported by their bot teammates, that they learned from playing alongside these advanced systems, and that it was generally a fun experience overall,” OpenAI notes, further explaining that the experience “presents a compelling vision for the future of human-AI interaction, one where AI systems collaborate and enhance the human experience.”
What was more interesting was that OpenAI Five pulled the feat without any special configuration. The AI had been modeled to only work with copies of itself but managed to adapt to cooperating with human players without further training.
But the cooperation match also shows the challenges of bringing AI and humans together. As Merity notes, the match highlighted many flaws and shortcomings. “There was no clear way for humans to cooperate with the bots. They couldn’t coordinate strategy,” he says. This means the humans can’t predict or direct the moves made by the AI and can only hope that the bots find a way to blend in with their strategy.
Okay, there's some degree of cooperation, but this feels very much like a bot taking pity on a human, like when my cat puts a dead mouse on my pillow as he thinks I'm too thin and don't spend enough time hunting for myself 🤣https://t.co/blXjUzsv8R
— Smerity (@Smerity) April 13, 2019
This would sometime result in awkward situations. “The bots would give up on humans and leave them to fight their own battles whilst going off elsewhere. It seemed a relatively lonely cooperative match,” Merity notes.
The AI and human players also had no effective method to communicate.
OpenAI will be launching an event called “OpenAI Five Arena” on Thursday through Sunday, where they will let anyone play Dota 2 against and alongside the champion AI.
“Seeing how humans interact when they have the opportunity to play alongside the bots in the short OpenAI Five ‘live’ period will be interesting to see,” Merity says.
What OpenAI Five tells us about the future of AI
There’s no point in spending immense expensive resources to teach AI to play games if it doesn’t serve real-world use cases. OpenAI employed the same model used in first version of Five to teach a robotic hand to handle objects by itself using reinforcement learning. It will be interesting to see what the updated AI model will achieve.
But while games are good arenas to train AI models, the real world is much more complex. “Unfortunately a sufficiently complex and interesting simulator for real world events is still quite rare,” Merity says.
OpenAI was founded with the goal of creating artificial general intelligence, the kind of AI that can replicate the general-problem-solving functionalities of the human mind. So far, there’s still no technology that comes close to mimicking human-level intelligence.
“What OpenAI is trying to do is build general artificial intelligence and to share those benefits with the world and make sure it’s safe,” OpenAI CEO Sam Altman told The Verge after Saturday’s event. “We’re not here to beat video games, as fun as that is. We’re here to uncover secrets along the path the AGI.”
But it’s not clear whether teaching AI to master Dota 2 has moved us closer to this goal, partly because we still don’t know what artificial general intelligence is, and there are still many unanswered questions about the human brain.
“We have no clue what mastery of one game may give us in terms of intelligence. Does it take wisdom to play a game of chess well? Does it take wisdom to play a game of DotA or StarCraft well? We’re still waiting to see,” Merity says. | <urn:uuid:8e54b730-45fe-4c45-94bb-2182a65bcf5e> | CC-MAIN-2022-40 | https://bdtechtalks.com/2019/04/17/openai-five-neural-networks-dota-2/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337415.12/warc/CC-MAIN-20221003101805-20221003131805-00526.warc.gz | en | 0.96344 | 2,624 | 2.59375 | 3 |
Nine years ago today, 20 children – all ages six and seven – and 6 adults were killed at Sandy Hook Elementary School. And just last week there was yet another school shooting this time in Oxford, Michigan.
Why it’s possible to prevent school shootings
From our interviews of school shooters, our investigations into their school shootings, and the thousands of threat cases that we have worked on throughout the past 20 years, we know that it is quite possible to prevent school shootings.
- School shooters don’t “just snap” when they carry out their violent acts. Instead, they plan out their violence beforehand AND their planning behavior is often observed by their friends, classmates, and sometimes family members beforehand. In fact, their pre-attack behavior often follows what the Secret Service and FBI call a “pathway to violence”
- They develop some idea to do harm (often to try to solve a problem or handle a situation where they don’t see other options);
- They plan how they want to carry out the harm;
- They prepare for the violent act by accessing firearms, other weapons, and other gear they think they’ll need;
- Then they implement the plan for violence.
- School shooters almost always tell other people. Whether it be their friends, classmates or online followers, they often talk about their violent plans before they carry them out (the FBI calls these communications “leakage”).
- Most carry out their shootings because they are feeling desperate, despondent, or in many cases are suicidal. They often hope that police will kill them during the school shooting or plan to kill themselves at the end of their school shooting. Some even attempted suicide, but failed, and resorted to carrying out a school shooting instead with the hope that police would end their life.
So, we know that school shootings can be prevented because they often share their violent plans beforehand, they engage in observable behavior that shows they are thinking about or planning a violent act and they often feel they have to resort to violence because they see no other way to solve a problem or are otherwise feeling desperate or despondent.
Although the investigation is still ongoing, current reporting about the Oxford High School shootings shows a similar pattern of behavior by the accused student prior to the school shooting.
- His teacher found a note on his desk with drawings of a gun, someone being shot, blood, and the words “The thoughts won’t stop. Help me.”
- He posted photos and captions a few days before the shooting on Instagram showing his new 9mm gun and his mother posted on her social media that she and her son were practicing at the gun range with the new gun.
- From a different account, the student appeared to post a phrase about “becoming death,”, and “See you tomorrow, Oxford.”
- Several students at the school stayed home that day or went to a friend’s house instead of going to school, allegedly after hearing about the student’s plans for a school shooting that day.
- During the shooting, one of his classmates posted a TikTok from lockdown that said “He’s being true to his word,” again suggesting that the student shooter posted beforehand about (and/or told friends and classmates about) his plans for a school shooting.
- Law enforcement said they found a journal in the student’s backpack with details about his ideas and plans to shoot up his school;
- He created two videos on his cellphone with details about his intentions to shoot up his school.
I’m often asked why school shooters tell other people beforehand. The answer is that they are hoping someone will stop them. In case after case, students who engaged in violence and students who were stopped beforehand have told us they were uncomfortable with the violent thoughts they were having but didn’t know how to handle them. The same seems true in the Oxford shooting as well: the student left a note on his desk with the words “The thoughts won’t stop. Help me.”
Using threat assessment to stop school shootings
So, how can we use this information to stop a school shooting? In short order, we can stop school shootings by using a process called Behavioral Threat Assessment and Management (BTAM or “threat assessment” for short). A threat assessment is essentially a fact-finding process, that looks like this:
- In a threat assessment, we look for signs that a person is on a “pathway to violence,” including evidence of planning, leakage, (i.e., telling others about plans) and trying to access lethal weapons to carry it out.
- If we determine someone is on a pathway to violence – i.e. they are planning and preparing for a violent act – we then ask WHY. They may be despondent and see violence as a way out.
- When we know the WHY, we can figure out a plan to get them off the pathway to violence and keep them off the pathway. Resources we use to prevent suicide can also be used to prevent school shootings as well.
When my colleagues and I interviewed school shooters in prison, they often described feeling torn about their violent plans beforehand. A part of them felt they had to be violent but a part of them didn’t want to at the same time. So, when my colleagues and I work on threat cases, we always look for that ambivalence because it’s usually there. Even if the person we are assessing has moved very far down the pathway to violence, we look for the part of them that doesn’t want to do it.
When we can get them help to solve the underlying problems – including getting them into mental health treatment – we can get them off the pathway to violence and onto a better path.
My colleagues and I have provided threat assessment training to thousands of school, mental health, and law enforcement professionals around the United States, and continue to do so. And we hear back from training participants – sometimes even years later – that they were able to use threat assessment to stop a school shooting in their community and got help to a student who was struggling.
Threat assessment can be just as helpful in preventing violence in the workplace. How do you actually create a threat assessment program for your organization? Follow the steps in our checklist, Building a Workplace Threat Assessment Program, to help you get started. | <urn:uuid:ffcd8eea-3c26-4c3e-8b9e-4f13354a6de9> | CC-MAIN-2022-40 | https://ontic.co/blog/how-to-stop-a-school-shooting-using-pre-incident-indicators-to-know-when-to-act-and-what-to-do/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337723.23/warc/CC-MAIN-20221006025949-20221006055949-00526.warc.gz | en | 0.980802 | 1,334 | 3.25 | 3 |
What Is Cyber Insurance?
Cybersecurity insurance (cyber insurance) is a product that enables businesses to mitigate the risk of cyber crime activity like cyberattacks and data breaches. It protects organizations from the cost of internet-based threats affecting IT infrastructure, information governance, and information policy, which often are not covered by commercial liability policies and traditional insurance products.
Cyber insurance coverage works the same way as businesses would purchase insurance against physical risks and natural disasters. It covers the losses an enterprise may suffer as a result of a cyberattack.
Why Is Cyber Insurance Important?
Cyber insurance is increasingly becoming essential for all companies as the risk of cyberattacks against applications, devices, networks, and users grows. That is because the compromise, loss, or theft of data can significantly impact a business, from losing customers to the loss of reputation and revenue.
Enterprises may also be liable for the damage caused by the loss or theft of third-party data. A cyber insurance policy can protect the enterprise against cyber events, including acts of cyber terrorism, and help with the remediation of security incidents.
For example, hackers breached Sony’s PlayStation Network in 2011 and exposed the data of 77 million users. The attack also prevented PlayStation Network users from accessing the service for 23 days. Sony incurred costs of over $171 million that could have been covered by cyber insurance. However, it did not have a policy, so it had to shoulder the total costs of the cyber damage.
Working Process of Cyber Insurance
The cybersecurity insurance process works in a similar way to other forms of insurance. Policies are sold by many suppliers that provide other forms of business insurance, such as errors and omissions insurance, liability insurance, and property insurance. Cyber insurance policies will often include first-party coverage, which means losses that directly impact an enterprise, and third-party coverage, which means losses suffered by other enterprises due to having a business relationship with the affected organization.
A cyber insurance policy helps an organization pay for any financial losses they may incur in the event of a cyberattack or data breach. It also helps them cover any costs related to the remediation process, such as paying for the investigation, crisis communication, legal services, and refunds to customers.
What Risks Does Cyber Insurance Cover?
Insurance for cybersecurity typically includes first-party coverage of losses incurred through data destruction, hacking, data extortion, and data theft. The main areas that cyber insurance covers include:
- Customer notifications: Enterprises are usually required to notify their customers of a data breach, especially if it involves the loss or theft of personally identifiable information (PII). Cyber insurance often helps businesses cover the cost of this process.
- Recovering personal identities: Cybersecurity insurance coverage helps organizations restore the personal identities of their affected customers.
- Data recovery: A cyber liability insurance policy usually enables businesses to pay for the recovery of any data compromised by an attack.
- System damage repair: The cost of repairing computer systems damaged by a cyberattack will also be covered by a cyber insurance policy.
- Ransom demands: Ransomware attacks often see attackers demand a fee from their victims to unlock or retrieve compromised data. Cyber insurance coverage can help organizations cover the costs of meeting such extortion demands.
- Attack remediation: A cyber insurance policy will help an enterprise pay for legal fees incurred through violating various privacy policies or regulations. It will also help them hire security or computer forensic experts who will enable them to remediate the attack or recover compromised data.
Cyber Risks Excluded from Cyber Insurance Coverage
A cybersecurity insurance policy will often exclude issues that were preventable or caused by human error or negligence, such as:
- Poor security processes: If an attack occurred as a result of an organization having poor configuration management or ineffective security processes in place
- Prior breaches: Breaches or events that occurred before an organization purchased a policy
- Human error: Any cyberattack caused by human error by an organization’s employees
- Insider attacks: The loss or theft of data due to an insider attack, which means an employee was responsible for the incident
- Preexisting vulnerabilities: If an organization suffers a data breach as a result of failing to address or correct a previously known vulnerability
- Technology system improvements: Any costs related to improving technology systems, such as hardening applications and networks
Can Cyber Insurance Take the Place of Cyber Defense?
Cyber insurance should not be considered in place of effective and robust cyber risk management. All companies need to purchase cyber insurance but should only consider it to mitigate the damage caused by a potential cyberattack. Their cyber insurance policy needs to complement the security processes and technologies they implement as part of their risk management plan.
Cyber insurance suppliers analyze an organization’s cybersecurity posture in the process of issuing a policy. Having a solid security posture enables an enterprise to obtain better coverage. In contrast, a poor security posture makes it more difficult for an insurer to understand their approach, resulting in ineffective insurance purchases.
Furthermore, failing to invest in appropriate or effective cybersecurity solutions can result in enterprises either failing to qualify for cyber insurance or paying more for it.
Considerations for Selecting the Appropriate Cyber Insurance Policy
Pricing cyber risk will typically depend on an enterprise’s revenue and the industry they operate in. To qualify, they will likely need to allow an insurer to carry out a security audit or provide relevant documentation courtesy of an approved assessment tool. The information accrued from an audit will guide the type of insurance policy the provider can offer and the cost of any premiums.
Policies often vary between different providers. Therefore, it is best to review any details carefully to ensure the required protections and provisions are covered by the proposed policy. The policy also needs to provide protection against currently known and emerging cyber threat vectors and profiles.
Three Steps To Reduce Cyber Risk
Cyber risk is a significant concern for companies of all sizes and across all industries. Organizations need to take decisive action to strengthen their cyber defenses and manage their cyber risk through the combination of cyber insurance, secure devices, domain expertise, and technology.
- Step 1—Assess: The first step in reducing cyber risk is to assess cyber readiness with a respected professional services organization. This process includes carrying out a security audit before providing appropriate cyber insurance.
- Step 2—Implement: The next step is to implement technology that protects the elements an organization intends to take out cyber insurance against. This can include an anti-malware solution to protect the enterprise against the threat of malicious software.
- Step 3—Insurance: The first two steps enable an organization to prove they have the necessary processes and technologies in place to qualify for cyber insurance from a provider.
How Fortinet Can Help?
Fortinet provides industry-leading technology, such as enterprise-grade ransomware and phishing solutions, that help organizations strengthen their cybersecurity defenses and prove they have the processes in place to qualify for cyber insurance. Fortinet technology protects organizations from advanced cyber threats like malware and distributed denial-of-service (DDoS) attacks and prevents unauthorized access to their networks and systems by cyber criminals.
FortiGuard Labs' Derek Manky and Jim Richberg, Fortinet Field CISO for the Public Sector, offer their perspectives on current ransomware trends, cyber insurance, ransomware settlements, and how organizations can better defend against and recover from attacks.
What does a cyber insurance policy cover?
A cyber insurance policy protects organizations from the cost of internet-based threats affecting IT infrastructure, information governance, and information policy, which often are not covered by commercial liability policies and traditional insurance products.
What is cyber insurance and how does it work?
Cyber insurance works in a similar way to other forms of insurance. Policies are sold by many suppliers that provide other forms of business insurance, such as errors and omissions insurance, liability insurance, and property insurance.
What is not covered by cyber insurance?
A cybersecurity insurance policy will often exclude issues that were preventable or caused by human error or negligence. | <urn:uuid:ea09222b-128a-46be-95e8-f1f01d774325> | CC-MAIN-2022-40 | https://www.fortinet.com/ru/resources/cyberglossary/cyber-insurance | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337723.23/warc/CC-MAIN-20221006025949-20221006055949-00526.warc.gz | en | 0.945794 | 1,639 | 2.796875 | 3 |
Comparing Detection and Prevention Security Controls
It worth stressing the differences between detection and prevention security controls. A detective control can’t predict when an incident will occur and it can’t prevent it. In contrast, prevention security controls stop the incident from occurring at all. Consider cameras and guards.
- Video surveillance. A simple camera without recording capabilities can prevent incidents because it acts as a deterrent. Compare this to a CCTV system with recording abilities. It includes cameras, which can deter and prevent incidents but the full system is also a detection control because of the recording capabilities. Security professionals can review the recordings to detect incidents after they’ve occurred.
- Guards. Guards are primarily prevention security controls. They will deter many incidents just by their presence. If attackers try to circumvent a security system, such as trying to sneak into a secured area, guards can intervene and stop them.
Corrective Security Controls
Corrective controls attempt to reverse the impact of an incident or problem after it has occurred. Some examples of corrective controls are:
- Active IDS. Active intrusion detection systems (IDSs) attempt to detect attacks and then modify the environment to block the attack from continuing.
- Backups and system recovery. Backups ensure that personnel can recover data if it is lost or corrupted. Similarly, system recovery procedures ensure administrators can recover a system after a failure.
Deterrent Security Controls
Deterrent controls attempt to discourage a threat. Some deterrent controls attempt to discourage potential attackers from attacking, and others attempt to discourage employees from violating a security policy.
You can often describe many deterrent controls as preventive controls. For example, you can have a security guard control access to a restricted area of your building. This guard will deter most people from trying to sneak in. This deterrence prevents security incidents related to unauthorized access. Similarly, a social engineer might try to trick a building receptionist but if you require visitors to go through the security guard first, it will deter many social engineers and prevent unauthorized entry.
The following list identifies some physical security controls used to deter threats:
- Cable locks. Securing laptops to furniture with a cable lock deters thieves from stealing the laptops. Thieves can’t easily steal a laptop secured this way. If they try to remove the lock, they destroy the laptop. Admittedly, a thief could cut the cable with a large cable cutter. However, someone walking around with a four-foot cable cutter looks suspicious.
- Hardware locks. Other locks such as locked doors securing a wiring closet or a server room also deter attacks. Many server bay cabinets also include locking cabinet doors.
Compensating Security Controls
Compensating controls are alternative controls used instead of a primary control. As an example, an organization might require smart cards as part of a multi-factor authentication solution. However, it might take time for new employees to receive their smart card. To allow new employees to access the network and still maintain a high level of security, the organization might choose to implement a Time-based One-Time Password (TOTP) as a compensating control. The compensating control still provides multi-factor authentication.
Page 2 Security Control Goals continued (this page)
Pass the First Time with the
Other Security+ Study Resources
- Security+ blogs organized by categories
- Security+ blogs with free practice test questions
- Security+ blogs on new performance-based questions
- Mobile Apps: Apps for mobile devices running iOS or Android
- Audio Files: Learn by listening with over 6 hours of audio on Security+ topics
- Flashcards: 494 Security+ glossary flashcards, 222 Security+ acronyms flashcards and 223 Remember This slides
- Quality Practice Test Questions: Over 300 quality Security+ practice test questions with full explanations
- Full Security+ Study Packages: Quality practice test questions, audio, and Flashcards | <urn:uuid:d3664f04-4d1f-4447-961a-84cd38f36767> | CC-MAIN-2022-40 | https://blogs.getcertifiedgetahead.com/prevention-security-controls-goals-2-of-3/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335124.77/warc/CC-MAIN-20220928051515-20220928081515-00726.warc.gz | en | 0.917822 | 791 | 2.75 | 3 |
These creative professionals design visually stunning animations for a range of mediums, including film, television, video games, mobile apps and websites. As the title suggests, computer animators use computer software and hardware to ply their trade (as opposed to the hand-drawn designs of traditional animators). Recent advances in animation technology, along with the increased popularity and diversity of devices used to access entertainment have made computer animator one of the most exciting and lucrative career paths for technically-inclined artists. This career path is also ideal for those who want to be their own boss; 57% of multimedia animators are self-employed, according to the latest figures from the Bureau of Labor Statistics.
Specializing in one of the aforementioned mediums – or better yet a specific skill set within that medium – is one of the best ways to advance your career in computer animation. For example, an animator who works solely on video game animation (and concentrates on the software and techniques used therein) can raise his or her proficiency faster than one who works across multiple platforms, and by focusing on 3D level design in video games he will improve and advance even quicker – and find himself that much closer to reaping the rewards that come with being a true expert in his niche. Discovering your ideal niche (where your passion and talent meet industry demand) will likely occur during the computer animation training process, or at least upon joining the workforce and gaining exposure to different types of animation.
Accredited colleges and private design schools offer a range of training programs that will prepare you for a career in computer animation. Compare some of the top-reviewed computer animation courses and degrees online and in your area.
a.k.a. Motion Graphics Designer | Multimedia Animator | Visual Effects Supervisor | Special Effects Artist
Computer Animator Salary
Courses and Degrees
Computer Animation Jobs
Job Outlook for Animators
Frequently Asked Questions
Computer Animator Skills & Responsibilities
Computer animators need a range of creative, technical and communication skills to perform effectively in this role. Here are some common day-to-day duties and marketable computer animation skill sets. Computer animators:
- Are proficient in the leading animation software, such as Autodesk Maya, Mari, and a variety of programs from Adobe’s Creative Suite (notably Character Animator, Animate, Photoshop, and After Effects).
- Wield cutting-edge digital animation techniques, such as motion capture, character rigging, 3D modeling and texture development, to make their on-screen creations come to life.
- Benefit from a working knowledge of physics and anatomy, as a deeper understanding of how and why things move will help you develop the most realistic animations.
- Usually work as part of a team to create large-scale effects and productions. This team setting is another reason why career-minded animators should focus on a computer animation specialty.
- Utilize storyboarding to plan complex scenes and visualize sustained animation sequences.
- May be called upon to collaborate with talent in the entertainment industry, e.g., actors and directors.
- Possess strong knowledge of traditional art concepts, such as color, shading, composition and perspective.
- Can work 50+ hours per week and even pull all-nighters when deadlines are approaching or work is past due.
Computer Animator Salary
The median annual wage for computer animators in the U.S. is $78,000, according to the U.S. Bureau of Labor Statistics.
Salaries for popular job roles similar to computer animation:
- Computer Animation Instructor: $55,000
- Motion Graphics Designer: $58,000
- Flash Animator: $61,000
- Maya Animator: $62,000
- Advertising Animator: $66,000
- Web Animator: $68,000
- Computer Animator: $78,000
- Mobile Animator: $75,000
- Video Game Animator: $80,000
- Computer Animation Manager: $84,000
Highest paying American cities for computer animators:
- Los Angeles, California: $110,000
- San Francisco, California: $103,000
- San Jose, California: $101,000
- Portland, Oregon: $100,000
- NYC Metropolitan Area: $97,000
- Seattle, Washington: $91,000
Sources: Indeed.com | United States Bureau of Labor Statistics
Computer Animator Education Requirements
Positions in this space typically require applicants with a Bachelor’s degree in computer animation, multimedia design, fine arts, graphic design, or a related field of study. As is the case with many creative design positions, your demo reel and portfolio will have a huge impact on your job prospects and starting salary in computer animation. You should start building a reel as soon as possible and continue to grow it with your best work throughout the training process and your career. I've hired dozens of animators and design pros over the years, and can attest that a brilliant portfolio speaks volumes when comparing candidates, much more so than your educational background.
Sought-after creative and technical skills to look for in a computer animation degree program include popular animation programs (such as Maya, 3ds Max and Mari) as well as fluency in relevant Adobe programs (such as Flash, Photoshop and After Effects), graphic/multimedia design, motion capture, character rigging, 3D modeling, storyboarding, creative writing, physics, mathematics, and knowledge of fine arts concepts like color, composition, life drawing, texture and lighting. You also want a program that will teach you how to assemble and promote a professional demo reel/portfolio.
Marketable communication skills (a.k.a. soft skills) for computer animators include time management, effective verbal/written communication, active listening, and creative problem solving.
Computer animation training works very well in distance learning formats. Compare online training programs that align with computer animators' education requirements.
Computer Animation Training
Browse college degrees, certificates and self-paced online courses matching the computer animator career path.
BA in Graphic Design: 3D Modeling & Animation
- Program highlights:
- Learn the tools and techniques to pursue jobs in computer animation & multimedia design
- 3D animation and modeling using Autodesk products like Maya & Mudbox
- Fine art foundations inc. color and typography
- Graphic design using Adobe's Creative Suite
- Design & render photorealistic digital effects
CG Spectrum Online Animation Courses with Certificate
- Program highlights:
- 2D and 3D animation using Toon Boom Harmony & Autodesk Maya
- Storytelling, dialogue & emotional expression for animation
- Advanced 3D animation ft. character acting and body mechanics
- Build a strong portfolio to propel your computer animator career
- One-on-one mentorship from accomplished animation experts
Computer Animator Job Openings
Your training and experience in computer animation may qualify you for a range of positions, including:
- Computer Animator jobs
- 3D Animation jobs
- Multimedia Designer jobs
- Production Artist jobs
- Flash Developer jobs
- Motion Graphics jobs
Computer Animator Job Outlook
The U.S. Bureau of Labor Statistics predicts 16% job growth for computer animators from 2020 to 2030, much faster than the 8% average for all occupations. Despite increases in overall animation production and the diversity of delivery methods for animated content, such as web apps and mobile devices, domestic job growth in this area will be tempered by fierce global competition -- a great deal of the animation we consume is produced overseas, due to the lower costs of manpower and infrastructure.
Animators with an adventurous spirit can take the outsourcing trend as an opportunity to seek employment across the globe, but when it comes to improving your computer animation job prospects stateside, the name of the game is specialization. Meaning, for the best chance of getting hired (or promoted) you should focus your computer animation training, experience and portfolio projects on mastering a niche skill set, as employers will pay top-dollar for a true expert and thought-leader in the skills they need. For example, if you want to work in film, concentrate your learning plan on the animation tools and techniques used in movie studios. Or take specialization to the next level by focusing on a specific subset of film animation, such as character sculpting and texturing. Specializing allows you to hone your skill set in a way that’s simply not possible with generalization, thus making you more visible and valuable to employers seeking those skills.
Source: U.S. Bureau of Labor Statistics' Occupational Outlook Handbook
Frequently Asked Questions
Technology and design insiders answer common questions from people considering the Computer Animator career path.
Which computer course is best for animation?
The best courses in computer animation cover the platforms you need to succeed in this role, such as Autodesk Maya, Mari, Unreal Engine, and Adobe programs like Animate, After Effects & Character Animator, as well as multimedia skills like 3D animation, life art, game design, shading and lighting, and compositing.
Which fields use computer animation?
The top industries that employ computer animators include film and television, gaming, software production, and marketing & advertising.
Is computer animation a good career?
Computer animator is a great career path for the right person, i.e., someone who's creative, detail-oriented, patient, and works well on a team. Computer animators earn a median salary of $78,000 (much higher than the $42,000 average for all occupations) and job growth in this field is projected at 16% over the next decade (double the 8% average for all occupations).
Is computer animation hard to learn?
For those with the creative chops and a good learning plan, it’s not especially difficult to learn the basics of computer animation, but achieving a high level of proficiency in any niche of animation takes a great deal of time, effort and persistence. | <urn:uuid:5c7df355-e76a-459a-b04f-76d648c7c625> | CC-MAIN-2022-40 | https://www.itcareerfinder.com/it-careers/computer-animator.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335355.2/warc/CC-MAIN-20220929131813-20220929161813-00726.warc.gz | en | 0.919387 | 2,069 | 3.015625 | 3 |
In the Kubernetes world, there are two types of resources:
- Compute resources, such as CPU (units) and memory (bytes), which can be measured.
- API resources, such as pods, services, etc., which are simply K8s objects.
In this article, I’ll explain compute resources.
(This article is part of our Kubernetes Guide. Use the right-hand menu to navigate.)
Quality of Service (QoS)
By default, when you deploy a pod, the pod containers essentially consume whatever compute resources are available on that worker node instance. That’s great—we don’t have to worry about anything, then, right?
Well, not exactly.
What happens when one of the containers consumes more CPU and memory resources than others? What happens to the other pod containers running on the node? There is a high chance that some of the containers may fail due to lack of resources; if we are running an application with a poor codebase, that can be impactful to the business.
To solve this issue, Kubernetes has resource requests and limits which help control the amount of resources a pod container consumes, ensuring the highest level of quality of service (QoS). It is important to point out that QoS (requests and limits) can also apply to namespaces, not just pods. That way if you have multiple dev teams deploying to a cluster, you can control resource consumption depending on the priority of the service.
How resource request and limit works
Fundamentally, a worker node or instance only allocate a certain portion of its overall compute resource for pods to use so, that portion is what is shared between all pods scheduled on that node. Kubernetes scheduler makes sure that the sum of all the compute resource of containers running on the instance does not exceed the overall allocated resource for pods. It is important to know the difference between request and limit.
- Request is how much resource Kubernetes will guarantee a pod.
- Limit is the maximum amount of resource Kubernetes will allow a pod to use.
Keep in mind that pod scheduling is based on request, not limit, but a pod and its containers are not allowed to exceed the limit specified. When you deploy a pod, a capacity check determines whether the total amount of compute resource requested is not more than what the node can allocate.
How request and limits are enforced
How the request and limit are enforced is based on whether the resource is compressible or incompressible. A compressible resource can be throttled, but an impressible resource—not so much. For example, CPU is considered compressible and memory is incompressible.
Pods are guaranteed the amount of CPU requested; they may or may not get more CPU depending on other pods running and anything past the limit is throttled. Pod memory request are also guaranteed but when pods exceed memory limit, the process inside the container that is using the most memory will be killed. The pod will not be terminated but the process inside the container using the memory will be terminated.
There are times when a node is completely out of compute resources; the sum of all the running containers compute resources is greater than the compute resource on the node itself. In extreme cases like this, we have to start killing running containers on the node. Ideally, we first remove less critical containers to free up resources. So, how do we tell Kubernetes which pod needs to go first in the case when all resources are exhausted on the node? By using QoS classes.
We can designate three types of classes—best effort, guaranteed, and burstable—to tell the system which pods to terminate first.
- Best effort. Pods with no request or limits, which are terminated first in case of resource exhaustion.
Containers: name: foo Resources: name: bar resources:
- Guaranteed. Pods where both limit and (optionally) request are set for all resources (CPU and memory) and their values are the same. These pods are high priority, therefore they are terminated only if they are over the limit and there are no lower priority pods to terminate.
containers: name: pong resources: limits: cpu: 100m memory: 100Mi requests: cpu: 100m memory: 100Mi
- Burstable. Pods where both request and (optionally) limits are set for one or more resources (CPU and/or memory) and their values are NOT the same. These pods are terminated when there are no best effort pods to terminate.
containers: name: foo resources: limits: memory: 1Gi name: bar resources: limits: cpu: 100m
Because application uptime is part of business SLAs, it is important to apply adequate resources to these applications. Understanding how to properly plan these resources is crucial to business. Using requests and limits will help team properly allocate resources to critical applications and QoS will help control the lifecycle of the application in case of resource exhaustion on the worker node. Teams can also define these requests and limits at the namespace level, that way resources can be carved based on dev teams. | <urn:uuid:8a35c943-d3a3-4574-8bf8-2f93e3344d72> | CC-MAIN-2022-40 | https://blogs.bmc.com/kubernetes-compute-resources/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337287.87/warc/CC-MAIN-20221002052710-20221002082710-00726.warc.gz | en | 0.915197 | 1,071 | 2.75 | 3 |
Data centers require high-quality and reliable uninterrupted power 24/7. Any disruption in the power source leads to serious consequences, downtime and damage to equipment. Data can be corrupted, and servers can malfunction. This all results in lost revenue. Downtime can cost tens of thousands of dollar per hour.
Backup power typically comes from generators. Backup generators supply power when the main power source experiences disruptions. They rely on different types of diesel to function. Hence, generators are not very eco-friendly.
Today, sustainability is the movement for data centers. More and more alternative technologies are coming out. One of these is the hydrogen backup power system also known as hydrogen fuel cells. Fuel cells are now incorporated in data center operations. They can generate reliable and high-quality power.
The hydrogen backup power can help companies to improve their data centers by reworking and improving their power systems. Hydrogen backup power can also render diesel backup generators redundant along with UPS systems. This serves as a promising future for the backup power systems in the data center.
Types of Backup Power in Data Centers
When power outages happen, the backup power source takes over the facility’s power load. Typically, It facilities have an eight-hour backup power source. But facilities that are more prone to power outages need 24 to 72-hour backup power. There are three types of power sources used in IT facilities like data centers.
The most common source of backup power. Batteries can supply DC power. Lead-acid batteries get electricity from the grid power source. They are and stored electricity so that facilities may use them in case of a power outage. The number of power batteries can supply depends on how much electricity is stored in them.
These backup power can supply AC power. Generators can be used automatically or manually. They can be used in remote and off-grid locations to serve as general power sources. Generators rely on non-renewable energy to function.
- Hydrogen Backup Power
This is also known as fuel cell backup power. They are powered by hydrogen. They utilize the proton-electron membrane (PEM) technology to supply DC power. Hydrogen backup power can function even at low temperatures. They also have a quick warm-up period and are smaller than other fuel cells.
Microsoft’s Hydrogen Backup Power
In 2020, hyperscale company Microsoft started to explore the potential of hydrogen backup power. The company’s aim is to be carbon negative by 2030. Through research, they successfully powered a row of data center servers for 48 consecutive hours all by using hydrogen fuel cells.
Diesel generators are an absolute necessity to be able to have five-nines of availability at data centers. A lot of data centers keep 24-to-48 hours of fuel on-site.
Diesel generators are noisy and produce, not just carbon, but also particulates, including nitrous oxides. Not to mention that its getting harder to get a permit at the levels that some hyperscale companies are moving.
From then, Microsoft started to look for alternative power sources. One that would replace diesel technology but can maintain – or even improve – the availability and reliability of such power sources. Hydrogen fuel cells proved to be an effective subject for this technology.
Why Choose Hydrogen Backup Power
While there are many types of backup power sources, fuel cells are more environmentally friendly. As we head towards a sustainable data center industry, it is time to shift to alternative technologies as they offer more benefits to companies and to the environment.
Hydrogen backup power has many benefits compared to other types of backup power. For one, they generate power through chemical processes which does not include the burning of fuel or combustion. Pure hydrogen does not emit pollution. It only produces heat and water as by-products.
Hydrogen backup power can also be used as a primary source of power. While the local power grid serves as a backup if needed. Moreover, the excess heat produced by the fuel cells can aid in cooling for servers. Hydrogen backup power produces electricity at a lower cost. This is a big advantage for companies given the power costs needed in data centers.
Below are some other benefits of hydrogen backup power as stated by the NREL – Department of Energy.
- Cleaner Energy. Hydrogen backup power is cleaner and pollutes less than the typical generator backup.
- 50% more operating efficiencies. Fuel cells can power many types of equipment.
- Longer lifespan of the equipment. Hydrogen backup power can last for up to ten years with the same power quality and quantity.
- Scalability and Modularity.
- Wider scale for operating temperature. From -40°F up to 122°F.
Hydrogen Backup Power in the Future
In the near future, hydrogen generators will replace diesel generators. Not only as an emergency power source but as a primary power source as well.
There are many new technologies that can aid data center power systems all of which have their own benefits and disadvantages. Nowadays, there are many renewable energies that data centers can use for power supply. Energy storage technology is also largely employed anywhere. The use of hydrogen as an alternative power source aligns with the sustainability of renewable power sources. Hence, it provides a promising future for the data center industry.
Hyperscale companies have made promising technologies towards sustainability. They drive innovation in the data center industry. But these advanced technologies still cater to big companies. The next step is to localize clean power production and consumption. Considering that all types of data centers, big or small, may be able to afford the costs of implementing these designs.
On the other hand, batteries also face issues with their efficiency. Storing renewable energy such as wind and solar power is a major challenge. There is still a lot of improvement needed in this area. Batteries will likely work as a supplementary backup power source.
Technological advancements together with sustainability advocacies will foster the growth of hydrogen power sources. Fuel cells will no longer be a backup source but more so a primary source of power in the data centers.
State of the Art in Fuel Cell Monitoring
Permanent damage might occur if the battery temperature is not monitored and correctly maintained. At best, there will be some mechanical deformation or a change in the chemical composition, requiring a pricey battery replacement. In the worst-case scenario, the battery cell could break, explode, leak chemicals, or catch fire.
The negative terminal of the battery frequently shows signs of increased battery temperature. The temperature should not rise more than 3°C above the ambient temperature when normal operating circumstances are applied. Two temperature sensors can be used, one to monitor the ambient temperature and the other to monitor the negative terminal of the battery. The difference between the two sensors can then be utilized to detect potential battery health problems or circuit faults.
Battery Monitoring System
A simple manual battery temperature monitoring system would be to have someone physically check the battery string once or twice a week. An IR temperature gun can do this, and you can record the data in a spreadsheet. Remember you are looking for not more than a 3°C difference between the battery and ambient temperature.
In these days of automation, why not install a permanent 24/7 battery temperature monitoring system? The most basic is a temperature sensor installed on the negative terminal post of the battery. You will have a graph of the temperature, and with the addition of an ambient temperature sensor, the two can be plotted on the same graph and the ∆T shown. More advanced battery monitoring systems can monitor individual cell temperatures, charge current, voltage, and load current and voltage.
Data centers are facilities that operate 24/7. They need a reliable power source in case of unexpected issues such as power outages. Power outages can last for a few minutes. But even the shortest outage can cost billions of dollars lost for a company. Power outages can be caused by power quality issues like voltage sags or surges. This can damage sensitive equipment and disrupt data center operations. Hence, reliable and high-quality backup power is essential. Hydrogen backup power does not only offer this but more importantly, it is a sustainable practice for your data center. | <urn:uuid:48e96e49-2363-4de2-8dbc-3ef4e4b07e22> | CC-MAIN-2022-40 | https://www.akcp.com/articles/hydrogen-backup-power-for-data-centers/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337287.87/warc/CC-MAIN-20221002052710-20221002082710-00726.warc.gz | en | 0.92788 | 1,843 | 3.203125 | 3 |
A new network design could increase internet speeds by 100 times, by removing the need to convert optical signals to electrical ones, according to MIT boffins.
The group of researchers at the Massachusetts Institute of Technology have discovered (opens in new tab) that a new type of router which doesn't have to translate optical signals into electrical signals could speed the Internet up by at least 100 times, and theoretically even as much as a thousand times.
Professor Vincent Chan has demonstrated a new way of organising optical networks that, in most cases, would eliminate this inefficient conversion process, speeding up data transfers and reducing the amount of energy consumed.
According to Chan, optical data transmissions are so efficient because different wavelengths of light loaded with different information can travel over the same fibre. But problems arise when optical signals coming from different directions reach a router at the same time. Converting them to electrical signals allows the router to store them in memory until it can get to them.
"The wait may be a matter of milliseconds," he writes, "but there’s no cost-effective way to hold an optical signal still for even that short a time."
Chan’s approach, called 'flow switching', solves this problem in a different way. Between locations that exchange large volumes of data, flow switching would establish a dedicated path across the network. For certain wavelengths of light, routers along that path would accept signals coming in from only one direction and send them off in only one direction. Since there’s no possibility of signals arriving from multiple directions, there’s never a need to store them in memory.
With growing demand for higher bandwidth, the new technology could become a reality, but the major hurdle is economic rather than technical. Implementing the new scheme would mean replacing huge numbers of existing routers, and most Internet users don't need the headline speeds right now.
Give it a couple of years when everyone is demanding high resolution 3D video at 120Hz on demand and that will be another thing. | <urn:uuid:7c9e6fbb-6191-49c7-8784-9487af45a595> | CC-MAIN-2022-40 | https://www.itproportal.com/2010/06/30/mit-design-could-make-internet-100-times-faster/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337287.87/warc/CC-MAIN-20221002052710-20221002082710-00726.warc.gz | en | 0.945232 | 406 | 3.25 | 3 |
What ever happened to printed encyclopedias?
That is a question usually asked by older generations as many of the youth today have never really had a daily interaction with this printed media. Most people today are used to receiving information via their digital books or other paperless mediums that are better for the environment. We will look at a brief history of the once loved encyclopedias and were all the rage of yesteryear and how many people today rely on their various tablets and laptops.
Earlier this year, in March 2012, our beloved Encyclopedia Britannica made the inevitable announcement that many people were already predicting. They mentioned that they were going to discontinue printing out there book sets. Many people get nostalgic when they think about Encyclopedia Britannica and they are often reminded of trips to grandma and grandpa’s house where they had these books on the shelves. Unfortunately, this is as far as the thought goes because one look around their house will show that these books are nowhere to be found. If they don’t have them how can they expect their children to be well versed in looking for information in this matter?
While many were upset about this news, the president of the company quickly highlighted the simple fact that every company needs to evolve and they are going to evolve as well. For those that are still interested in gathering information via this medium, they can now access the digital versions. If many people have already successfully determined that printed newspapers and magazines will no longer be effective, it would be simple to assume that the Encyclopedia Britannica would follow the same route. On a related topic, when is the last time that you actually pulled out a map or at list to look for directions? There may be a few hard-core old-line traditionalists, but the average person will simply refer to the GPS that is either in their car or smart phone.
The paperless revolution has finally arrived! If you can imagine, the statements were really uttered in the mid-1980s with the invention of the fax machine. This machine was promoted to limit the excess paper that everyone used in their day-to-day personal and professional lives. However, many people use just as much if not more paper and it has taken 20+ years to slowly adapt to the electronic means.
Another thing to consider for the benefit of society is crowd sourcing. Crowd sourcing sites like Wikipedia has been a huge boon to anyone looking for up to date information. It is extremely difficult to imagine waiting for the new Britannica Encyclopedia to arrive so that you can absorb new information. Today, many people rely on sites like Wikipedia to retrieve information that is very current. If you were to make a comparison, Wikipedia today more than dwarfs all of the information that Britannica had compiled over the last 20 years.
In addition to being paperless, digital books, ebooks and other digital media has cut costs tremendously for many publishers. Long gone are the days of having to depend on manufacturing paper and destroying rain forests for public consumption. The digital age, while still in its infancy, will continue to mature and meet the demands of consumers that are always seeking to stay up to date with how they consume information.
Key words: digital books, ebooks, paperless | <urn:uuid:15167bd9-f08b-4f18-9405-029cb6afcaf3> | CC-MAIN-2022-40 | https://jorgep.com/blog/remember-encyclopedias/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337421.33/warc/CC-MAIN-20221003133425-20221003163425-00726.warc.gz | en | 0.979497 | 657 | 2.8125 | 3 |
In our last post we discussed the differences between MAC and IP addresses. In this post we’ll delve deeper into IP addresses. Specifically, this post is going to focus on how to translate IP addresses into their original binary numbers.
An IP address looks something like this to us: 188.8.131.52. Your computer doesn’t see it that way though. It sees a string of 1’s and 0’s. In fact, that’s how your computer sees pretty much everything, and that number system is called binary. Learning how to convert IP addresses into binary is one of the easiest ways to get some insight into binary math.
You see, at its most basic, computers are just a series of digital switches. When a computer reads binary, it’s reading which switches to turn on and off. 1 means “on” and 0 means “off.” But humans occasionally need to read this information, and looking at pages upon pages of 1’s and 0’s is just confusing, so it gets translated. The 1’s and 0’s get grouped together, assigned values, and added up to make bigger but shorter numbers, like IP addresses.
Let’s break down an IP address. IP addresses are composed of four numbers separated by periods. Each of these four numbers is called an “octet.” They are called octets because each is determined by eight possible values. The possible values, in order, are: 128, 64, 32, 16, 8, 4, 2, and 1. The binary string of 1’s and 0’s takes these values and tells you if they are included on/in the value of that octet or not.
Example of IP Binary Math
So let’s look at the IP address 184.108.40.206. Written out in binary, it would be:
That means that 198 = 11000000 in binary. So if we break down those 8 numbers:
1.) 128 = 1 = On/yes
2.) 64 = 1 = On/yes
3.) 32 = 0 = Off/no
4.) 16 = 0 = Off/no
5.) 8 = 0 = Off/no
6.) 4 = 0 = Off/no
7.) 2 = 0 = Off/no
8.) 1 = 0 = Off/no
Then take all the “yes” values and add them up.
128 + 64 = 198.
Hooray, that’s the match behind one of the octets for our example IP address! It’s pretty simple to figure out if you have the binary string. If you have just the IP and need to figure out the binary string, remember that you always have the same eight possible values, and you always start with the biggest.
Would you prefer focusing on running your business and allowing us to handle all the technical stuff? Contact us today to learn how we can leverage technology to propel your business to new heights. | <urn:uuid:50fb6866-37e2-41e2-8d5f-b2ae19f33ae7> | CC-MAIN-2022-40 | https://corenets.com/translating-ip-addresses-binary/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337731.82/warc/CC-MAIN-20221006061224-20221006091224-00726.warc.gz | en | 0.935991 | 650 | 3.734375 | 4 |
As the world is now aware, China is now home to the world’s most powerful supercomputer, toppling the previous reigning system, Tianhe-2, which is also located in the country.
In the wake of the news, we took an in-depth look at the architecture of the new Sunway TiahuLight machine, which will be useful background as we examine a few of the practical applications that have been ported to and are now running on the 10 million-core, 125 petaflop-capable supercomputer.
The sheer size and scale of the system is what initially grabbed headlines when we broke news about the system last week at the International Supercomputing Conference (full coverage listing of that event here). However, as details emerged, it became quickly apparent that was no stunt machine designed to garner headlines by gaming the Top 500 supercomputer benchmark. Rather, this system rolled out with full system specs backed by news that several real-world scientific applications were able to run on the machine, some of which could use well over 8 million cores—a stunning bit of news in a community where application scalability and real-world performance often is at dramatic odds with projected theoretical peak performance.
News about this new supercomputer, unlike the mystery about the practical value of Tianhe-2 when it was announced, had more credibility because of the number of Gordon Bell prize submissions that accompanied the formal launch. This prize is awarded to teams that can demonstrate remarkable scalability on massive machines, showing scientific/application value as well performance and efficiency. As one might imagine, in the supercomputing arena, this is a grand challenge.
Despite the availability of millions of compute cores, sometimes boosted by accelerators, getting real-world codes to scale to make full, efficient use of such resources is ongoing, pressing challenge. In fact, this is one of the great questions as the impetus builds for exaflop-capable systems—even with such power, how many codes will be able to scale to advantage of that capability?
In addition to the Gordon Bell prize submissions (more on those below), Dr. Haohuan Fu, Deputy Director of NSCC-Wuxi, where the Sunway TaihuLight supercomputer is housed, shared details and performance results for some key applications running on the new machine in a session at ISC 16. The Next Platform was on hand to gather some insight from this talk and share a few slides.
Deep Learning Libraries, Large-Scale Neural Networks
Although supercomputing applications are still just out of reach of the influence of deep learning (something we expect will shift in the next couple of years) the TaihuLight supercomputer is being harnessed for some interesting work on deep neural networks. What is fascinating here is that currently, the inference side of such workloads can scale to many processors, but the training side is often scale-limited hardware and software-wise.
Fu described an ongoing project on the Sunway TaihuLight machine to develop an open source deep neural network library and make the appropriate architectural optimization for both high performance and efficiency on both the training and inference parts of deep learning workloads. “Based on this architecture, we can provide support for both single and double precision as well as fixed point,” he explains. The real challenge, he says, is to understand why most existing neural network libraries cannot benefit much from running at large scale and looking at the basic elements there to get better training results over a very large number of cores.
Above are some noteworthy preliminary performance results for convolutional layers for double-precision. The efficiency isn’t outstanding (around 30%A) but Fu says they’re working on the library to bolster it and get equal or better performance than the GPU—the standard thus far for training.
Weather and Atmospheric Codes
Earth systems modeling, weather forecasting, and atmospheric simulations are a few key application areas where scientists using TaihuLight are scaling to an incredible number of cores. The Chinese-developed CAM weather model has been focal point for teams to scale and represents some of the challenges inherent to exploiting a new architecture.
According to Fu, “there is a lot of complexity in the legacy codebase with over a half million lines of code. We can’t do all of this manually, so we’re working on the tools to port them since the legacy codes were not designed for multicore and not for a manycore architecture like the Sunway processor.” The tools they are working on are targeting the right level of parallelism, code sizes, and memory footprint, but ultimately, he says, this leads to one of the greatest challenges—finding the right talent that can understand the underlying physics and the computational and software problems. “Even the climate scientists don’t understand the code well, it’s been added to over the course of three decades.”
Scalability and performance results for the CAM model can be seen above comparing both use with the management core and sub-cores and with just the management core. For some kernels that are compute intensive, the team saw a speedup of between 10-22X, but for others that were memory-bound, the speedup wasn’t high, just 2-3X. The results here show speedup for the entire model and if there is any takeaway here, this is scaling to quite impressive heights for code that’s still in process on a new architecture—1.5 million cores.
Fu says to get to this point, they had to divide CAM into two parts; the dynamic part, which was rewritten in the last decade (they ported and optimized manually), and the CAM physics component, which was the difficult part. “We’re relying on transformation tools here to expose the right level of parallelism and code sizes for the 260 cores on this architecture. We also developed our own memory footprint analysis tool for this part.”
Another earth systems application, a high-res atmospheric model is showing good results as well. This is an experimental project that differs from the porting and optimization requirements of the legacy code above. Here the team is taking a hardware and software co-design approach and applying a loosely coupled scheme to the scalable model. They have run experiments for 10 to 3 kilometer resolution—an impressive feat when one considers the current scalability and resolution capabilities for leading centers like ECMWF, among others.
In the example above, the team was able to use the entire system as was during this run—38 cabinets, which is still well over 8 million cores. Fu says he expects that when they continue research with this code they will be able to use the full machine—over 10 million cores.
Gordon Bell Submissions
The following slide highlights the five applications that were submitted with the three accepted submissions highlighted. The winners of this award will be announced in November, but given the breadth of systems on the Top 500 now and their core counts, it is unlikely any will scale beyond 8 million cores since, well, none of them have even close to that many to begin with (the #2 machine, Tianhe-2, “only” has a tick over 3 million).
In terms of the code work for the Sunway TaihuLight machine, the unique architecture obviously creates some barriers. Fu says they have a parallel OS environment and are using their own homegrown file system (Sunway GFS) which many guess is based on Lustre. The machine will support C, C++ and Fortran compilers and support for all basic software libraries. Fu says they are using a combination of OpenMP, OpenACC and MPI, but for many of the early stage applications demonstrated here, they are using a hybrid mode that balances OpenACC and MPI (for the different compute groups, one MPI process is allocated and OpenACC is used to execute parallel threads).
As an interesting final side note, this government-funded supercomputer is set to support the needs of manufacturing operations in the region, which includes large cities nearby, including Shanghai. One can expect that many of the solvers and other simulation workflows will go to support the regions automotive and other industries, which explains why the $270 million funding for the supercomputer came from a collection of sources, including the province and cities near the center. | <urn:uuid:f06e78fb-1a52-4a18-a1e5-aaa9eaa9fb41> | CC-MAIN-2022-40 | https://www.nextplatform.com/2016/06/30/inside-look-key-applications-chinas-new-top-supercomputer/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338213.55/warc/CC-MAIN-20221007143842-20221007173842-00726.warc.gz | en | 0.958619 | 1,708 | 2.578125 | 3 |
The Biden administration is looking at how the climate crisis will impact the federal budget and revenue.
Under the worst-case scenario, nearly 13,000 federal facilities may be at risk from rising sea levels due to the climate crisis.
That is one of several points included in various documents released by the Office of Management and Budget on Monday related to climate risks in the federal budget and federal revenue going forward. This follows an executive order President Biden issued in May 2021 ordering such calculations.
“At the upper end of that range, climate change could lead to an annual federal revenue loss at the end of the century of 7.1%, which in today’s dollars would equal $2 trillion per year,” Candace Vahlsing, associate director for climate, energy, environment, and science at OMB, and Danny Yagan, chief economist at OMB, wrote in a blog post. “Furthermore, the analyses found that the federal government could spend between an additional $25 billion to $128 billion annually on just six types of federal expenditure: coastal disaster relief, flood insurance, crop insurance, healthcare insurance, wildland fire suppression, and flooding at federal facilities.” This is the first time OMB is formally accounting for climate risks in the federal budget, they added.
While there isn’t a comprehensive dataset of all federal buildings and structures, the Federal Real Property Profile Management System is the closest thing to that and thus can provide some insight into the financial risks to the federal government as a result of climate change, said a white paper accompanying OMB’s assessment.
Using the system, OMB identified more than 40,000 federal buildings and structures that have a total replacement cost of about $81 billion (in 2020 dollars) that are in the current 100-year floodplain, which are areas that have at least a 1% chance of flooding annually and considered “high risk” by the Federal Emergency Management Agency.
Also, there are about 60,000 structures that have a total replacement cost of $493 billion that were identified as part of the current 500-year floodplain, which are areas that have at least a 0.2% chance of flooding annually.
“As global warming continues to exacerbate sea level rise and extreme weather, our nation’s floodplains are expected to grow by approximately 45% by century’s end,” the Natural Resources Defense Council, an environmental advocacy group, said in April 2019. “Among other things, FEMA’s floodplains determine how and where homes and other structures are built, as well as who is required to purchase flood insurance (coverage is mandatory if you live in a floodplain and have a federally backed mortgage).” It also noted that some of the flood maps have been found to be inaccurate, according to a Bloomberg story from 2017.
Additionally, OMB and the National Oceanic and Atmospheric Administration “identified 10,250 individual federal buildings and structures, with a combined replacement cost of $32.3 billion, that would be inundated or severely affected by typical high tide under an eight-foot sea level rise scenario,” the white paper said. “Under a ten-foot ‘worst case’ sea level rise scenario…12,195 individual federal buildings and structures would be inundated, with [a] total combined replacement cost of over $43.7 billion,” which was in 2020 dollars. This was also reviewed using the Federal Real Property Profile Management System.
The white paper noted that some of these federal facilities appeared to be outside of the 100-year floodplain “reinforcing the expectation that sea level rise will appreciably expand the number and value of federal facilities facing flood risk in the coming decades.”
There are other potential climate related risks for federal facilities that were not assessed here and there were also data limitations in assessing the full extent of flooding risks, said the paper.
In order to mitigate these effects, last year, Biden reinstated the Federal Flood Risk Management Standard, which President Obama established and President Trump revoked, to encourage agencies to consider and manage present and future risks of floods as well as established a working group to coordinate implementation of it. Additionally, in 2020 the General Services Administration did a high-level assessment of flood vulnerabilities for venues within its jurisdiction and has started integrating climate considerations into its decision-making processes, said the documents.
The assessment also lays out potential direct and indirect climate risks and costs to the federal government’s mortgage insurance programs and Medicare and Medicaid spending, wildfire response, infrastructure services and expenditures, and national security environment.
Biden’s budget proposal for fiscal 2023, which he released last Monday, lists several budgetary needs for the near term that will “both help reduce the federal government’s long-term fiscal exposure to climate-related financial risk and reduce future climate risks for all Americans,” OMB’s assessment said. “In total, the budget invests a historic $44.9 billion in discretionary funding to tackle the climate crisis,’ which is almost a 60% increase from the fiscal 2021 enacted level.” | <urn:uuid:37ed83b4-22b5-4c27-8749-25b871f9161a> | CC-MAIN-2022-40 | https://fcw.com/workforce/2022/04/worst-case-13000-federal-facilities-flooding/364093/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335254.72/warc/CC-MAIN-20220928113848-20220928143848-00126.warc.gz | en | 0.961599 | 1,058 | 2.765625 | 3 |
Basilica of Our Lady of Protection
Nôtre Dame de la Garde or Our Lady of Protection is at the highest point above Marseille, 162 meters above the Old Port to its south.
Charles II of Anjou ordered in 1302 that beacons were set up along the Mediterranean coast of Provence. This hill was the site of one of those beacons.
A chapel had been built on the hill in 1214. A fortification was built here in the early 1500s and a more significant church constructed within. It continued as a church within a fortress until the Revolution in the late 1700s. The religious buildings were shut down and worship stopped in November, 1793. The fortress became a prison for members of the royalty and nobility.
The church was reopened for worship in 1807. The number of worshipers increased significantly, and the fortress chapel was expanded.
In 1850, the priest asked the Ministry of War, which still owned and controlled the fortress, for permission to further enlarge the existing church. Alphonse Henri, comte d'Hautpoul, on his final day on the job as Minister of War, agreed in principle but said that the request was too vague. Come back with details.
The following year, in April 1851, a new request was submitted. This was for an entirely new church, much larger. This came with the support of General Adolphe Niel.
The project was approved. Construction started in 1853. There were problems, both financial and practical, as it was difficult to cut the foundations into the very hard rock at the peak.
The church was finally consecrated in June, 1864.
The views are fantastic. In the first picture we are looking west over Île d'If and the other islands in the Frioul Archipelago. Below, we look over the Old Port and beyond that to the new port complex continuing for kilometers up the coast.
In the second picture we have turned toward our right. The innermost third of the Old Port is barely visible at left, just above the nearby trees.
Below, the long red roof of Gare Saint-Charles is visible if you know what to look for (about 11 o'clock from the cross, to the right and slightly above a row of three large rectangular buildings). The blue and white tourist tram provides one way to get up and down the hill between the Old Port and the basilica.
The interior walls of the church are covered with devotional plaques, thanking Our Lady of Protection for saving the donors from storms, shipwrecks, and other troubles.
Ship models hang from the ceiling, representing ships saved by Mary's intervention and ships that were lost but someone survived.
Above the plaque listing the numerical characteristics of the hill and cathedral, we see a plaque thankful for protection against cholera, one of the recent plagues.
That's right, this is a plague plaque.
Reconnaissance à N.D. de la Garde
qui nous a préservées du choléra
nous et nos familles
Les Dames Télégraphistes
Recognition to Our Lady of Protection
who has saved us and our families
from the cholera
The Lady Telegraphists
There are several plaques commemorating individuals and groups involved in the final fight to eject the German occupiers. The Germans retreated to the basilica and its surroundings, and the battle to finally eject them raged around the peak of the hill.
Brave leader, killed on the field of honor on 7 September 1944 at the head of the 2nd Battalion of the 3rd Algerian Infantry Regiment, having fought for the liberation of Our Lady of Protection on 26 August 1944 with his brothers in arms of the 7th Algerian Infantry Regiment of the 9th Cavalry and of the Free French Forces.
Another honors the Marseille Volunteers of the African Commandos killed for the liberation of France.
From here we will leave Marseille by car. We will drive south and east along the coast to Cassis and La Ciotat. | <urn:uuid:872e031f-4203-4198-bc9a-09407b9bcc92> | CC-MAIN-2022-40 | https://cromwell-intl.com/travel/france/marseille/notre-dame-de-la-garde.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337322.29/warc/CC-MAIN-20221002115028-20221002145028-00126.warc.gz | en | 0.963847 | 862 | 2.53125 | 3 |
Artificial intelligence can worsen systemic racism.
DALLAS — When software engineer Bejoy Narayana was developing Bob.ai, an application to help automate Dallas-Fort Worth’s Section 8 voucher program, he stopped and asked himself, ‘‘Could this system be used to help some people more than others?”
Bob.ai uses artificial intelligence, known as AI, and automation to help voucher holders find rental units, property owners complete contracting and housing authorities conduct inspections. The software and mobile app were released in 2018 in partnership with the Dallas Housing Authority, which gave Narayana access to data from some 16,000 Section 8 voucher holders.
Artificial intelligence is used in a host of algorithms in medicine, banking and other major industries. But as it has proliferated, studies have shown that AI can be biased against people of color. In housing, AI has helped perpetuate segregation, redlining and other forms of racial discrimination against Black families, who disproportionately rely on vouchers.
Narayana worried that Bob.ai would do the same, so he tweaked his app so that tenants could search for apartments using their voucher number alone, without providing any other identifying information.
As an Indian immigrant overseeing a team largely made up of people of color, Narayana was especially sensitive to the threat of racial bias. But lawmakers in a growing number of states don’t want to rely on the goodwill of AI developers. Instead, as AI is adopted by more industries and government agencies, they want to strengthen and update laws to guard against racially discriminatory algorithms—especially in the absence of federal rules.
Since 2019, more than 100 bills related to artificial intelligence and automated decision systems have been introduced in nearly two dozen states, according to the National Conference of State Legislatures. This year, lawmakers in at least 16 states proposed creating panels to review AI’s impact, promote public and private investment in AI, or address transparency and fairness in AI development.
A bill in California would be the first to require developers to evaluate the privacy and security risks of their software, as well as assess their products’ potential to generate inaccurate, unfair, biased or discriminatory decisions. Under the proposed law, the California Department of Technology would have to approve software before it could be used in the public sector.
The bill, introduced by Assembly Member Ed Chau, a Democrat and chair of the Committee on Privacy and Consumer Protection, passed the California State Assembly earlier this month and was pending in the state Senate at publication time. Chau’s office did not respond to multiple requests for comment.
Vinhcent Le, a lawyer at the Greenlining Institute, an advocacy group focused on racial economic justice, helped write the California legislation. Le described algorithms such as Bob.ai as gatekeepers to opportunity that can either perpetuate segregation and redlining or help to end them.
“It’s great that the developers of Bob.ai decided to omit a person’s name, but we can’t rely on small groups of people making decisions that can essentially affect thousands,” Le said. “We need an agreed way to audit these systems to ensure they are integrating equity metrics in ways that don’t unfairly disadvantage people.”
According to an October report by the Massachusetts Institute of Technology, AI often has exacerbated racial bias in housing. A 2019 report from the University of California, Berkeley, showed that an AI-based mortgage lending system charged Black and Hispanic borrowers higher rates than White people for the same loans.
In 2019, U.S. Sen. Cory Booker, a New Jersey Democrat, introduced a bill like the one under consideration in California, but it died in committee and has not been reintroduced.
"Fifty years ago, my parents encountered a practice called 'real estate steering' where black couples were steered away from certain neighborhoods in New Jersey. With the help of local advocates and the backing of federal legislation, they prevailed,” Booker said in a news release introducing the bill.
“However, the discrimination that my family faced in 1969 can be significantly harder to detect in 2019: houses that you never know are for sale, job opportunities that never present themselves, and financing that you never become aware of—all due to biased algorithms."
Several states have struggled in recent years with problematic software.
Facebook overhauled its ad-targeting system to prevent discrimination in housing, credit and job ads in 2019 as part of a settlement to resolve legal challenges filed by the National Fair Housing Alliance, the American Civil Liberties Union, the Communications Workers of America and other advocacy groups.
In Michigan, an AI system that cost the state $47 million to build in 2013 falsely accused as many as 40,000 people of unemployment insurance fraud, forcing some people into bankruptcy, according to the Detroit Free Press.
In Pennsylvania, a child abuse prediction model unfairly targets low-income families because it relies on data that is collected only on families using public resources, according to Virginia Eubanks' 2018 book “Automating Inequality.”
“Automated decision-making shatters the social safety net, criminalizes the poor, intensifies discrimination, and compromises our deepest national values,” Eubanks wrote. “And while the most sweeping digital decision-making tools are tested in what could be called ‘low rights environments’ where there are few expectations of political accountability and transparency, systems first designed for the poor will eventually be used on everyone.”
The Sacramento Housing Redevelopment Agency began using Bob.ai in March. Laila Darby, assistant director of the housing voucher program, said the agency vetted Bob.ai before using it to make sure it didn’t raise privacy and discrimination concerns.
Narayana said he’s sure Bob.ai would pass any state-mandated test for algorithmic discrimination.
“We’re a company that is fighting discrimination and doing everything possible to expand housing for voucher holders,” Narayana said. “Vetting these systems is beneficial because discrimination and inequality is something everyone should be concerned about.”
Narayana worked as an engineer at IBM until he decided to start his own company with the mission of rethinking government functions. He founded BoodsKapper in 2016 and began developing Bob.ai out of a co-working space near the Dallas-Fort Worth airport.
Narayana’s creation has been a huge success—in Dallas and beyond. The Dallas Housing Authority has used Bob.ai to cut the average wait time for an apartment inspection from 15 days to one. Since the launch of Bob.ai, Dallas and more than a dozen other housing agencies have added some 20,000 Section 8 units from landlords who were not participating in the program because of the long inspection wait times.
“We partnered with [Narayana] to come up with some technology advancements to our workflows and automation so that we could more timely respond to our business partners so that they didn’t see this as a lost lead in terms of working with the voucher program,” said Troy Broussard, Dallas Housing Authority CEO.
Marian Russo, executive director of the Village of Patchogue Community Development Agency on Long Island, New York, said she hopes Bob.ai can help the agency reverse the area’s long history of redlining. The authority plans to begin using Bob.ai to manage its 173 housing vouchers later this year.
“We’re one of the most segregated parts of the country,” Russo said of Long Island. “We have 25 housing authorities, so if we could just have a central place with all the landlords who are renting through the program and all the individuals who are looking for housing in one place, that could be a part of equalizing the housing issues on Long Island.”
U.S. Rep. Bill Foster, an Illinois Democrat, has similar hopes for AI. In a May 7 hearing, members of the Task Force on Artificial Intelligence of the U.S. House Committee on Financial Services discussed how AI could expand lending, housing and other opportunities. But they also warned that historical data inputted into AI can create models that are racist or sexist. Foster’s office did not respond to multiple requests for comment.
“The real promise of AI in this space is that it may eventually produce greater fairness and equity in ways that we may not have contemplated ourselves,” said Foster, chair of the task force, in the hearing. “So, we want to make sure that the biases of the analog world are not repeated in the AI and machine-learning world.”
This article first appeared on Stateline, an initiative of The Pew Charitable Trusts. | <urn:uuid:365debe6-72d9-448b-8639-d99c55343b8b> | CC-MAIN-2022-40 | https://gcn.com/emerging-tech/2021/06/programmers-lawmakers-want-ai-to-eliminate-bias-not-promote-it/315575/?oref=gcn-next-story | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337322.29/warc/CC-MAIN-20221002115028-20221002145028-00126.warc.gz | en | 0.954993 | 1,811 | 2.796875 | 3 |
Understanding and combating issues of fairness in supervised learning.
“Being good is easy, what is difficult is being just.” ― Victor Hugo
“We need to defend the interests of those whom we’ve never met and never will.” ― Jeffrey D. Sachs
Note: This article is intended for a general audience to try and elucidate the complicated nature of unfairness in machine learning algorithms. As such, I have tried to explain concepts in an accessible way with minimal use of mathematics, in the hope that everyone can get something out of reading this.
Supervised machine learning algorithms are inherently discriminatory. They are discriminatory in the sense that they use information embedded in the features of data to separate instances into distinct categories — indeed, this is their designated purpose in life. This is reflected in the name for these algorithms which are often referred to as discriminative algorithms (splitting data into categories), in contrast to generative algorithms (generating data from a given category). When we use supervised machine learning, this “discrimination” is used as an aid to help us categorize our data into distinct categories within the data distribution, as illustrated below.
Whilst this occurs when we apply discriminative algorithms — such as support vector machines, forms of parametric regression (e.g. vanilla linear regression), and non-parametric regression (e.g. random forest, neural networks, boosting) — to any dataset, the outcomes may not necessarily have any moral implications. For example, using last week’s weather data to try and predict the weather tomorrow has no moral valence attached to it. However, when our dataset is based on information that describes people — individuals, either directly or indirectly, this can inadvertently result in discrimination on the basis of group affiliation.
Clearly then, supervised learning is a dual-use technology. It can be used to our benefits, such as for information (e.g. predicting the weather) and protection (e.g. analyzing computer networks to detect attacks and malware). On the other hand, it has the potential to be weaponized to discriminate at essentially any level. This is not to say that the algorithms are evil for doing this, they are merely learning the representations present in the data, which may themselves have embedded within them the manifestations of historical injustices, as well as individual biases and proclivities. A common adage in data science is “garbage in = garbage out” to refer to models being highly dependent on the quality of the data supplied to them. This can be stated analogously in the context of algorithmic fairness as “bias in = bias out”.
Some proponents believe in data fundamentalism, that is to say, that the data reflects the objective truth of the world through empirical observations.
“with enough data, the numbers speak for themselves.” — Former Wired editor-in-chief Chris Anderson (a data fundamentalist)
Data and data sets are not objective; they are creations of human design. We give numbers their voice, draw inferences from them, and define their meaning through our interpretations. Hidden biases in both the collection and analysis stages present considerable risks, and are as important to the big-data equation as the numbers themselves. — Kate Crawford, principal researcher at Microsoft Research Social Media Collective
Superficially, this seems like a reasonable hypothesis, but Kate Crawford provides a good counterargument in a Harvard Business Review article:
Boston has a problem with potholes, patching approximately 20,000 every year. To help allocate its resources efficiently, the City of Boston released the excellent StreetBump smartphone app, which draws on accelerometer and GPS data to help passively detect potholes, instantly reporting them to the city. While certainly a clever approach, StreetBump has a signal problem. People in lower income groups in the US are less likely to have smartphones, and this is particularly true of older residents, where smartphone penetration can be as low as 16%. For cities like Boston, this means that smartphone data sets are missing inputs from significant parts of the population — often those who have the fewest resources. — Kate Crawford, principal researcher at Microsoft Research
Essentially, the StreetBump app picked up a preponderance of data from wealthy neighborhoods and relatively little from poorer neighborhoods. Naturally, the first conclusion you might draw from this is that the wealthier neighborhoods had more potholes, but in reality, there was just a lack of data from poorer neighborhoods because these people were less likely to have smartphones and thus have downloaded the SmartBump app. Often, it is data that we do not have in our dataset that can have the biggest impact on our results. This example illustrates a subtle form of discrimination on the basis of income. As a result, we should be cautious when drawing conclusions such as these from data that may suffer from a ‘signal problem’. This signal problem is often characterized as sampling bias.
Another notable example is the “Correctional Offender Management Profiling for Alternative Sanctions” algorithm or COMPAS for short. This algorithm is used by a number of states across the United States to predict recidivism — the likelihood that a former criminal will re-offend. Analysis of this algorithm by ProPublica, an investigative journalism organization, sparked controversy when it seemed to suggest that the algorithm was discriminating on the basis of race — a protected class in the United States. To give us a better idea of what is going on, the algorithm used to predict recidivism looks something like this:
Recidivism Risk Score = (age*−w)+(age-at-first-arrest*−w)+(history of violence*w) + (vocation education * w) + (history of noncompliance * w)
It should be clear that race is not one of the variables used as a predictor. However, the data distribution between two given races may be significantly different for some of these variables, such as the ‘history of violence’ and ‘vocation education’ factors, based on historical injustices in the United States as well as demographic, social, and law enforcement statistics (which are often another target for criticism since they often use algorithms to determine which neighborhoods to patrol). The mismatch between these data distributions can be leveraged by an algorithm, leading to disparities between races and thus to some extent a result that is moderately biased towards or against certain races. These entrenched biases will then be operationalized by the algorithm and continue to persist as a result, leading to further injustices. This loop is essentially a self-fulfilling prophecy.
Historical Injustices → Training Data → Algorithmic Bias in Production
This leads to some difficult questions — do we remove these problematic variables? How do we determine whether a feature will lead to discriminatory results? Do we need to engineer a metric that provides a threshold for ‘discrimination’? One could take this to the extreme and remove almost all variables, but then the algorithm would be of no use. This paints a bleak picture, but fortunately, there are ways to tackle these issues that will be discussed later in this article.
These examples are not isolated incidents. Even breast cancer prediction algorithms show a level of unfair discrimination. Deep learning algorithms to predict breast cancer from mammograms are much less accurate for black women than white women. This is partly because the dataset used to train these algorithms is predominantly based on mammograms of white women, but also because the data distribution for breast cancer between black women and white women likely has substantial differences. According to the Center for Disease Control (CDC) “Black women and white women get breast cancer at about the same rate, but black women die from breast cancer at a higher rate than white women”.
These issues raise questions about the motives of algorithmic developers — did the individuals that designed these models do so knowingly? Do they have an agenda they are trying to push and trying to hide it inside gray box machine learning models?
Although these questions are impossible to answer with certainty, it is useful to consider Hanlon’s razor when asking such questions:
Never attribute to malice that which is adequately explained by stupidity — Robert J. Hanlon
In other words, there are not that many evil people in the world (thankfully), and there are certainly less evil people in the world than there are incompetent people. On average, we should assume that when things go wrong it is more likely attributable to incompetence, naivety, or oversight than to outright malice. Whilst there are likely some malicious actors who would like to push discriminative agendas, these are likely a minority.
Based on this assumption, what could have gone wrong? One could argue that statisticians, machine learning practitioners, data scientists, and computer scientists are not adequately taught how to develop supervised learning algorithms that control and correct for prejudicial proclivities.
Why is this the case?
In truth, techniques that achieve this do not exist. Machine learning fairness is a young subfield of machine learning that has been growing in popularity over the last few years in response to the rapid integration of machine learning into social realms. Computer scientists, unlike doctors, are not necessarily trained to consider the ethical implications of their actions. It is only relatively recently (one could argue since the advent of social media) that the designs or inventions of computer scientists were able to take on an ethical dimension.
This is demonstrated in the fact that most computer science journals do not require ethical statements or considerations for submitted manuscripts. If you take an image database full of millions of images of real people, this can without a doubt have ethical implications. By virtue of physical distance and the size of the dataset, computer scientists are so far removed from the data subjects that the implications on any one individual may be perceived as negligible and thus disregarded. In contrast, if a sociologist or psychologist performs a test on a small group of individuals, an entire ethical review board is set up to review and approve the experiment to ensure it does not transgress across any ethical boundaries.
On the bright side, this is slowly beginning to change. More data science and computer science programs are starting to require students to take classes on data ethics and critical thinking, and journals are beginning to recognize that ethical reviews through IRBs and ethical statements in manuscripts may be a necessary addition to the peer-review process. The rising interest in the topic of machine learning fairness is only strengthening this position.
Fairness in Machine Learning
As mentioned previously, widespread adoption of supervised machine learning algorithms has raised concerns about algorithmic fairness. The more these algorithms are adopted, and the increasing control they have on our lives will only exacerbate these concerns. The machine learning community is well aware of these challenges and algorithmic fairness is now a rapidly developing subfield of machine learning with many excellent researchers such as Moritz Hardt, Cynthia Dwork, Solon Barocas, and Michael Feldman.
That being said, there are still major hurdles to overcome before we can achieve truly fair algorithms. It is fairly easy to prevent disparate treatment in algorithms — the explicit differential treatment of one group over another, such as by removing variables that correspond to these attributes from the dataset (e.g. race, gender). However, it is much less easy to prevent disparate impact —implicit differential treatment of one group over another, usually caused by something called redundant encodings in the data.
A redundant encoding tells us information about a protected attribute, such as race or gender, based on features present in our dataset that correlate with these attributes. For example, buying certain products online (such as makeup) may be highly correlated with gender, and certain zip codes may have different racial demographics that an algorithm might pick up on.
Although an algorithm is not trying to discriminate along these lines, it is inevitable that data-driven algorithms that supersede human performance on pattern recognition tasks might pick up on these associations embedded within data, however small they may be. Additionally, if these associations were non-informative (i.e. they do not increase the accuracy of the algorithm) then the algorithm would ignore them, meaning that some information is clearly embedded in these protected attributes. This raises many challenges to researchers, such as:
- Is there a fundamental tradeoff between fairness and accuracy? Are we able to extract relevant information from protected features without them being used in a discriminatory way?
- What is the best statistical measure to embed the notion of ‘fairness’ within algorithms?
- How can we ensure that governments and companies produce algorithms that protect individual fairness?
- What biases are embedded in our training data and how can we mitigate their influence?
We will touch upon some of these questions in the remainder of the article.
The Problem with Data
In the last section, it was mentioned that redundant encodings can lead to features correlating with protected attributes. As our data set scales in size, the likelihood of the presence of these correlations scales accordingly. In the age of big data, this presents a big problem: the more data we have access to, the more information we have at our disposal to discriminate. This discrimination does not have to be purely race- or gender-based, it could manifest as discrimination against individuals with pink hair, against web developers, against Starbucks coffee drinkers, or a combination of all of these groups. In this section, several biases present in training data and algorithms are presented that complicate the creation of fair algorithms.
The Majority Bias
Algorithms have no affinity to any particular group, however, they do have a proclivity for the majority group due to their statistical basis. As outlined by Professor Moritz Hardt in a Medium article, classifiers generally improve with the number of data points used to train them since the error scales with the inverse square root of the number of samples, as shown below.
This leads to an unsettling reality that since there will, by definition, always be less data available about minorities, our models will tend to perform worse on those groups than on the majority. This assumption is only true if the majority and minority groups are drawn from separate distributions, if they are drawn from a single distribution then increasing sample size will be equally beneficial to both groups.
An example of this is the breast cancer detection algorithms we discussed previously. For this deep learning model, developed by researchers at MIT, of the 60,000 mammogram images in the dataset used to train the neural network, only 5% were mammograms of black women, who are 43% more likely to die from breast cancer. As a result of this, the algorithm performed more poorly when tested on black women, and minority groups in general. This could partially be accounted for because breast cancer often manifests at an earlier age among women of color, which indicates a disparate impact because the probability distribution of women of color was underrepresented.
This also presents another important question. Is accuracy a suitable proxy for fairness? In the above example, we assumed that a lower classification accuracy on a minority group corresponds to unfairness. However, due to the widely differing definitions and the somewhat ambiguous nature of fairness, it can sometimes be difficult to ensure that the variable we are measuring is a good proxy for fairness. For example, our algorithm may have 50% accuracy for both black and white women, but if there 30% false positives for white women and 30% false negatives for black women, this would also be indicative of disparate impact.
From this example, it seems almost intuitive that this is a form of discrimination since there is differential treatment on the basis of group affiliation. However, there are times when this group affiliation is informative to our prediction. For example, for an e-commerce website trying to decide what content to show its users, having an idea of the individual’s gender, age, or socioeconomic status is incredibly helpful. This implies that if we merely remove protected fields from our data, we will decrease the accuracy (or some other performance metric) of our model. Similarly, if we had sufficient data on both black and white women for the breast cancer model, we could develop an algorithm that used race as one of the inputs. Due to the differences in data distributions between the races, it is likely that the accuracy would have increased for both groups.
Thus, the ideal case would be to have an algorithm that contains these protected features and uses them to make algorithmic generalizations but is constrained by fairness metrics to prevent the algorithm from discriminating.
This is an idea proposed by Moritz Hardt and Eric Price in ‘Equality of Opportunity in Supervised Learning’. This has several advantages over other metrics, such as statistical parity and equalized odds, but we will discuss all three of these methods in the next section.
Definitions of Fairness
In this section we analyze some of the notions of fairness that have been proposed by machine learning fairness researchers. Namely, statistical parity, and then nuances of statistical parity such as equality of opportunity and equalized odds.
Statistical parity is the oldest and simplest method of enforcing fairness. It is expanded upon greatly in the arXiv article “Algorithmic decision making and the cost of fairness” The formula for statistical parity is shown below.
For statistical parity, the outcome will be independent of my group affiliation. What does this mean intuitively? It means that the same proportion of each group will be classified as positive or negative. For this reason, we can also describe statistical parity as demographic parity. For all demographic groups subsumed within p, statistical parity will be enforced.
For a dataset that has not had statistical parity applied, we can measure how far our predictions deviate from statistical parity by calculating the statistical parity distance shown below.
This distance can provide us with a metric for how fair or unfair a given dataset is based on the group affiliation p.
What are the tradeoffs of using statistical parity?
Statistical parity doesn’t ensure fairness.
As you may have noticed though, statistical parity says nothing about the accuracy of these predictions. One group may be much more likely to be predicted as positive than another, and hence we might obtain large disparities between the false positive and true positive rates for each group. This itself can cause a disparate impact as qualified individuals from one group (p=0) may be missed out in favor of unqualified individuals from another group (p=1). In this sense, statistical parity is more akin to equality of outcome.
The figures below illustrate this nicely. If we have two groups — one with 10 people (group A=1), and one with 5 people (group A=0) — and we determine that 8 people (80%) in group A=1 achieved a score of Y=1, then 4 people (80%) in group A=0 would also have to be given a score of Y=1, regardless of other factors.
Statistical parity reduces algorithmic accuracy
The second problem with statistical parity is that a protected class may provide some information that would be useful for a prediction, but we are unable to leverage that information because of the strict rule imposed by statistical parity. Gender might be very informative for making predictions about items that people might buy, but if we are prevented from using it, our model becomes weaker and accuracy is impacted. A better method would allow us to account for the differences between these groups without generating disparate impact. Clearly, statistical parity is misaligned with the fundamental goal of accuracy in machine learning — the perfect classifier may not ensure demographic parity.
For these reasons, statistical parity is no longer considered a credible option by several machine learning fairness researchers. However, statistical parity is a simple and useful starting point that other definitions of fairness have built upon.
There are slightly more nuanced versions of statistical parity, such as true positive parity, false positive parity, and positive rate parity.
True Positive Parity (Equality of Opportunity)
This is only possible for binary predictions and performs statistical parity on true positives (the prediction output was 1 and the true output was also 1).
It ensures that in both groups, of all those who qualified (Y=1), an equal proportion of individuals will be classified as qualified (C=1). This is useful when we are only interested in parity over the positive outcome.
False Positive Parity
This is also only applicable to binary predictions and focuses on false positives (the prediction output was 1 but the true output was 0). This is analogous to the true positive rate but provides parity across false positive results instead.
Positive Rate Parity (Equalized Odds)
This is a combination of statistical parity for true positives and false positives simultaneously and is also know as equalized odds.
Notice that for equal opportunity, we relax the condition of equalized odds that odds must be equal in the case that Y=0. Equalized odds and equality of opportunity are also more flexible and able to incorporate some of the information from the protected variable without resulting in disparate impact.
Notice that whilst all of these provide some form of a solution that can be argued to be fair, none of these are particularly satisfying. One reason for this is that there are many conflicting definitions of what fairness entails, and it is difficult to capture these in algorithmic form. These are good starting points but there is still much room for improvement.
Other Methods to Increase Fairness
Statistical parity, equalized odds, and equality of opportunity are all great starting points, but there are other things we can do to ensure that algorithms are not used to unduly discriminate individuals. Two such solutions which have been proposed are human-in-the-loop and algorithmic transparency.
This sounds like some kind of rollercoaster ride, but it merely refers to a paradigm whereby a human oversees the algorithmic process. Human-in-the-loop is often implemented in situations that have high risks if the algorithm makes a mistake. For example, missile detection systems that inform the military when a missile is detected allow individuals to review the situation and decide how to respond — the algorithm does not respond without human interaction. Just imagine the catastrophic consequences of running nuclear weapon systems with AI that had permission to fire when they detected a threat — one false positive and the entire world would be doomed.
Another example of this is the COMPAS system for recividism — the system does not categorize you as a recidivist and make a legal judgment. Instead, the judge reviews the COMPAS score and uses this as a factor in their evaluation of the circumstance. This raises new questions such as how humans interact with the algorithmic system. Studies using Amazon Mechanical Turk have shown that some individuals will follow the algorithm’s judgment wholeheartedly, as they perceive it to have greater knowledge than a human is likely to, other individuals take its output with a pinch of salt, and some ignore it completely. Research into human-in-the-loop is relatively novel but we are likely to see more of it as machine learning becomes more pervasive in our society.
Another important and similar concept is human-on-the-loop. This is similar to human-in-the-loop, but instead of the human being actively involved in the process, they are passively involved in the algorithm’s oversight. For example, a data analyst might be in charge of monitoring sections of an oil and gas pipeline to ensure that all of the sensors and processes are running appropriately and there are no concerning signals or errors. This analyst is in an oversight position but is not actively involved in the process. Human-on-the-loop is inherently more scalable than human-in-the-loop since it requires less manpower, but it may be untenable in certain circumstances — such as looking after those nuclear missiles!
The dominant position in the legal literature for fairness is through algorithmic interpretability and explainability via transparency. The argument is that if an algorithm is able to be viewed publicly and analyzed with scrutiny, then it can be ensured with a high level of confidence that there is no disparate impact built into the model. Whilst this is clearly desirable on many levels, there are some downsides to algorithmic transparency.
Proprietary algorithms by definition cannot be transparent.
From a commercial standpoint, this idea is untenable in most circumstances — trade secrets or proprietary information may be leaked if algorithms and business processes are provided for all to see. Imagine Facebook or Twitter being asked to release their algorithms to the world so they can be scrutinized to ensure there are no biasing issues. Most likely I could download their code and go and start my own version of Twitter or Facebook pretty easily. Full transparency is only really an option for algorithms used in public services, such as by the government (to some extent), healthcare, the legal system, etc. Since legal scholars are predominantly concerned with the legal system, it makes sense that this remains the consensus at the current time.
In the future, perhaps regulations on algorithmic fairness may be a more tenable solution than algorithmic transparency for private companies that have a vested interest to keep their algorithms from the public eye. Andrew Tutt discusses this idea in his paper “An FDA For Algorithms”, which focused on the development of a regulatory body similar to the FDA to regulate algorithms. Algorithms could be submitted to the regulatory body, or perhaps third party auditing services, and analyzed to ensure they are suitable to be used without resulting in disparate impact.
Clearly, such an idea would require large amounts of discussion, money, and expertise to implement, but this seems like a potentially workable solution from my perspective. There is still a long way to go to ensure our algorithms are free of both disparate treatment and disparate impact. With a combination of regulations, transparency, human-in-the-loop, human-on-the-loop, and new and improved variations of statistical parity, we are part of the way there, but this field is still young and there is much work to be done — watch this space.
In this article, we have discussed at length multiple biases present within training data due to the way in which it is collected and analyzed. We have also discussed several ways in which to mitigate the impact of these biases and to help ensure that algorithms remain non-discriminatory towards minority groups and protected classes.
Although machine learning, by its very nature, is always a form of statistical discrimination, the discrimination becomes objectionable when it places certain privileged groups at a systematic advantage and certain unprivileged groups at a systematic disadvantage. Biases in training data, due to either prejudice in labels or under-/over-sampling, yields models with unwanted bias.
Some might say that these decisions were made on less information and by humans, which can have many implicit and cognitive biases influencing their decision. Automating these decisions provides more accurate results and to a large degree limits the extent of these biases. The algorithms do not need to be perfect, just better than what previously existed. The arc of history curves towards justice.
Some might say that algorithms are being given free rein to allow inequalities to be systematically instantiated, or that data itself is inherently biased. That variables related to protected attributes should be removed from data to help mitigate these issues, and any variable correlated with the variables removed or restricted.
Both groups would be partially correct. However, we should not remain satisfied with unfair algorithms, there is also room for improvement. Similarly, we should not waste all of this data we have and remove all variables, as this would make systems perform much worse and would render them much less useful. That being said, at the end of the day, it is up to the creators of these algorithms and oversight bodies, as well as those in charge of collecting data, to try to ensure that these biases are handled appropriately.
Data collection and sampling procedures are often glazed over in statistics classes, and not understood well by the general public. Until such a time as a regulatory body appears, it is up to machine learning engineers, statisticians, and data scientists to ensure the equality of opportunity is embedded in our machine learning practices. We must be mindful of where our data comes from and what we do with it. Who knows who our decisions might impact in the future?
“The world isn’t fair, Calvin.”
“I know Dad, but why isn’t it ever unfair in my favor?”
― Bill Watterson, The Essential Calvin and Hobbes: A Calvin and Hobbes Treasury
Big Data: A Report on Algorithmic Systems, Opportunity, and Civil Rights. The White House. 2016.
Bias in computer systems. Batya Friedman, Helen Nissenbaum. 1996
The Hidden Biases in Big Data. Kate Crawford. 2013.
Big Data’s Disparate Impact. Solon Barocas, Andrew Selbst. 2014.
Blog post: How big data is unfair. Moritz Hardt. 2014
Semantics derived automatically from language corpora contain human-like biases. Aylin Caliskan, Joanna J. Bryson, Arvind Narayanan
Sex Bias in Graduate Admissions: Data from Berkeley. P. J. Bickel, E. A. Hammel, J. W. O’Connell. 1975.
Simpson’s paradox. Pearl (Chapter 6). Tech report
Certifying and removing disparate impact. Michael Feldman, Sorelle Friedler, John Moeller, Carlos Scheidegger, Suresh Venkatasubramanian
Equality of Opportunity in Supervised Learning. Moritz Hardt, Eric Price, Nathan Srebro. 2016.
Blog post: Approaching fairness in machine learning. Moritz Hardt. 2016.
COMPAS Risk Scales: Demonstrating Accuracy Equity and Predictive Parity. Northpointe Inc.
Fairness in Criminal Justice Risk Assessments: The State of the Art
Richard Berk, Hoda Heidari, Shahin Jabbari, Michael Kearns, Aaron Roth. 2017.
Limitations of mitigating judicial bias with machine learning. Kristian Lum. 2017.
Probabilistic Outputs for Support Vector Machines and Comparisons to Regularized Likelihood Methods. John C. Platt. 1999.
Inherent Trade-Offs in the Fair Determination of Risk Scores. Jon Kleinberg, Sendhil Mullainathan, Manish Raghavan. 2016.
Fair prediction with disparate impact: A study of bias in recidivism prediction instruments. Alexandra Chouldechova. 2016.
Attacking discrimination with smarter machine learning. An interactive visualization by Martin Wattenberg, Fernanda Viégas, and Moritz Hardt. 2016.
Algorithmic decision making and the cost of fairness. Sam Corbett-Davies, Emma Pierson, Avi Feller, Sharad Goel, Aziz Huq. 2017.
The problem of Infra-marginality in Outcome Tests for Discrimination. Camelia Simoiu, Sam Corbett-Davies, Sharad Goel. 2017.
Equality of Opportunity in Supervised Learning. Moritz Hardt, Eric Price, Nathan Srebro. 2016.
Elements of Causal Inference. Peters, Janzing, Schölkopf
On causal interpretation of race in regressions adjusting for confounding and mediating variables. Tyler J. VanderWeele and Whitney R. Robinson. 2014.
Counterfactual Fairness. Matt J. Kusner, Joshua R. Loftus, Chris Russell, Ricardo Silva. 2017.
Avoiding Discrimination through Causal Reasoning. Niki Kilbertus, Mateo Rojas-Carulla, Giambattista Parascandolo, Moritz Hardt, Dominik Janzing, Bernhard Schölkopf. 2017.
Fair Inference on Outcomes. Razieh Nabi, Ilya Shpitser
Fairness Through Awareness. Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, Rich Zemel. 2012.
On the (im)possibility of fairness. Sorelle A. Friedler, Carlos Scheidegger, Suresh Venkatasubramanian. 2016.
Why propensity scores should not be used. Gary King, Richard Nielson. 2016.
Raw Data is an Oxymoron. Edited by Lisa Gitelman. 2013.
Blog post: What’s the most important thing in Statistics that’s not in the textbooks. Andrew Gelman. 2015.
Deconstructing Statistical Questions. David J. Hand. 1994.
Statistics and the Theory of Measurement. David J. Hand. 1996.
Measurement Theory and Practice: The World Through Quantification. David J. Hand. 2010
Survey Methodology, 2nd Edition. Robert M. Groves, Floyd J. Fowler, Jr., Mick P. Couper, James M. Lepkowski, Eleanor Singer, Roger Tourangeau. 2009
Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings. Tolga Bolukbasi, Kai-Wei Chang, James Zou, Venkatesh Saligrama, Adam Kalai. 2016.
Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints. Jieyu Zhao, Tianlu Wang, Mark Yatskar, Vicente Ordonez, Kai-Wei Chang. 2017.
Big Data’s Disparate Impact. Solon Barocas, Andrew Selbst. 2014.
It’s Not Privacy, and It’s Not Fair. Cynthia Dwork, Deirdre K. Mulligan. 2013.
The Trouble with Algorithmic Decisions. Tal Zarsky. 2016.
How Copyright Law Can Fix Artificial Intelligence’s Implicit Bias Problem. Amanda Levendowski. 2017.
An FDA for Algorithms. Andrew Tutt. 2016 | <urn:uuid:c44d16a5-f4e8-4806-a2a9-4f9c6b18fcbc> | CC-MAIN-2022-40 | https://resources.experfy.com/ai-ml/programming-fairness-in-algorithms/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337322.29/warc/CC-MAIN-20221002115028-20221002145028-00126.warc.gz | en | 0.9373 | 7,123 | 3.078125 | 3 |
Blockchain analysis is involves transaction analysis and fund tracing. Cryptocurrency is any form of currency that exists digitally and uses cryptography to secure transactions. As a peer-to-peer system, no bank is involved in verifying transactions between the sender and the receiver. An online database (public ledger) records the specific transactions.
Also, no central or official institution, such as the Treasury Department’s Bureau of Engraving and Printing (BEP), issues cryptocurrencies. Instead, any person can “mine” cryptocurrencies using dedicated hardware and an internet connection. In order to ensure that only verified crypto miners can mine and validate transactions, a proof-of-work (PoW) consensus protocol must be followed. There are currently an estimated 12,000 different types of cryptocurrencies, most of which were coined for a specific purpose. Cryptocurrencies are also traded on a cryptocurrency exchange aka a digital currency exchange (DCE). There are almost 600 DCEs worldwide.
On a technical level, cryptocurrencies run on a blockchain, which describes the way transactions are recorded into “blocks” and time-stamped. As a distributed public ledger. It records of all updated transactions. The owners/users of cryptocurrencies keep their currencies in a digital wallet, which can only be accessed using a private key to prove ownership. There are two types: hot wallets that use online software to protect the private key, and cold (hardware) wallets that are offline electronic devices where the private is stored securely.
There are currently an estimated 300 million cryptocurrency users. The popularity of cryptocurrencies is rising; the projected market size for cryptocurrency is expected to reach $1087.7 million in 2026. Cryptocurrencies are popular since they enable cheaper and faster money transfers and decentralized systems that do not collapse at a single point of failure. The latter happened during the financial crisis of 2007 – 2008. For the underbanked, it allows them to access capital and make financial transactions quickly and conveniently, especially for cheaper, faster, and secure cross-border money remittances. There are, of course, also disadvantages. Cryptocurrencies are known for their price volatility, high energy consumption for mining activities, and limited acceptance for purchasing goods and services.
Due to its peer-to-peer nature that keeps translations obfuscated, threat actors use cryptocurrencies to buy and sell illicit goods and services on the dark web and to finance terrorist activities. Fraudsters abuse the popularity of cryptocurrencies to promote virtual Ponzi schemes and other scams to trick people into giving them money. Although cryptocurrencies are well-protected since the digital ledgers of cryptocurrency transactions are hard to tamper with by people without a specific private key. This did not stop threat actors from successfully breaching and stealing the equivalent of millions of US dollars. To illustrate, the Ronin Network was hacked in March 2022. The culprits stole around $625 million in cryptocurrency from Ronin Network’s blockchain, making it the largest crypto heist ever recorded. The legitimate owners of the breached crypto wallets might be left without any recourse. In many cases in the past, a victimized exchange or operator of the cryptocurrency did not have enough reserves (in contrast to traditional financial institutions) to compensate the victims.
To trace such culprits, a blockchain analysis is needed. Although peer-to-peer and encrypted, blockchain transactions can still be traced since they leave digital footprints behind. With a blockchain analysis, an investigator can detect, categorize, and model these digital footprints to get key information about the threat actors that performed the illicit transactions. Since blockchain analysis is often part of a dark web investigation, it is necessary to constantly identify the technical details of online activities, behavior, and other relevant information. With a tool such as the AI-powered WEBINT platform of Cobwebs, investigators can connect threat actors to illicit blockchain activities by uncovering hidden identities. The tool’s AI and smart algorithms adapt to the latest blockchain and cryptocurrency industry services and platforms, as well as digital wallets.
In the case of cryptocurrency-related dark web investigations, the platform also enables investigators to scan the dark web-based on specific search terms or phrases without the need to access the dark web themselves. This eliminates the need for them to personally explore the dark web by accessing dark web websites, forums, message boards, and other publicly accessible dark web sources. The platform uses AI-based dark web scanning technology to connect the dots in a visual graph and gives actionable insights in the form of automated reports that can be used for follow-up. | <urn:uuid:70db9660-bd6c-47d8-8785-8c26b9cad7ba> | CC-MAIN-2022-40 | https://cobwebs.com/blockchain-analysis-of-cryptocurrency-transactions-and-how-to-keep-them-safe/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338280.51/warc/CC-MAIN-20221007210452-20221008000452-00126.warc.gz | en | 0.932227 | 919 | 3.359375 | 3 |
What are the various kinds of email spoofing? What are the security controls on your email platform to sprevent email fraud and impersonation due to email spoofing? Read on.
Business v/s Consumer Email Accounts (M) 2013-2017
Source : radicati.com
Worldwide daily email traffic (B) 2013-2017
While the % of growth in the consumer and business email accounts is about the same year on year, it is observed that the the business email traffic will exhibit growth year on year (more business transacted on email), while consumer email traffic will reduce (possibly due to social media).
Source : radicati.com
Global Spam Statistics
Source : Statista
What is ’email spoofing’?
Spoofing, in a nutshell, is email fraud and deception. Spoofing is the most common form of the modern con game called ‘phishing’. The email spoofer is a spammer attempting to hide his true identity, impersonating someone else, and trying to “phish” or lure you to reveal your passwords, login names and other personal and financial information.
Why Would Someone Fraudulently “Spoof” an Email?
Email spoofing may occur in different forms, but all have a similar result: a user receives an email that appears to have originated from one source when it actually was sent from another source.
Email spoofing is often an attempt to trick the user into making a damaging statement, releasing sensitive information (such as passwords) and/or performing a financial transaction.
Examples of email spoofing
- An Email claiming to be from a system administrator (and also appearing so) requesting users to change their passwords to a specified string and threatening to suspend their account if they do not do this.
- A mail claiming to be from a person in authority requesting users to send them a copy of a password file or other sensitive information.
- Receiving an Email from a partner, requesting for funds transfer pertaining to a sale or transaction.
Email spoofing can be done by one of the valid users on a domain or users on external domains. Lets look at some common scenarios of how spoofing is done and how it can be prevented.
To explain the scenarios, lets take an example: Ravi and Smita are colleagues in the same organisation and Ravi sends a mail on behalf of Smita (Ravi has spoofed Smita’s email id)
Email spoofing type 1: Ravi hacks into Smita’s email box:
Ravi could do this since she has a weak password and the email system doesn’t have any password security in place e.g. password history, account lockout, password age, etc and was able to send a mail on Smita’s behalf. In this scenario the mail has gone from Smita’s account, but she is unaware of it.
This situation can be easily prevented if the following security policies are in place for each account:
- Strict Password Policies to ensure complex password, regular password rotation, automatic account lockout on several unsuccessful attempts and always fresh passwords by referring to the password history.
- Access control to define which services the user can access and the trusted network ranges from which the user can access the server.
- Mail Policies to control whereall each user can send mail and under what condition.
- Every mail send request requires authentication by the a valid sender in the network
Email spoofing type 2: Ravi impersonates Smita:
Ravi connects to the organisation mail server, authenticates using his own account but sends a mail containing Smita’s email id in the “From ID” header. When the recipient gets this mail, it appears to have been sent from Smita’s email id. This is possible with mail servers which have a weak authorisation system.
This can be prevented by applying the following security policies for each account
Email Spoof Check: This means that if Ravi is authenticating then the mail should also contain ONLY Ravi’s email id in the “From ID” field. If the authentication ID and the From ID don’t match, the mail will be rejected.
Domain Spoof check: This means that, this server will only allow mail FROM domains authorised to relay mail from here. In other words, for the server to acccept any request to send mail, the envelop Sender ID and the From ID field in the mail should have email ids only from one the authorised listed domains. This check prevents clients and other servers from sending mail FROM foreign domains (open relay) in case of any compromise.
Email spoofing type 3: Ravi impersonates Smita from an external mail system:
Using an Open relay
Ravi sends a mail to the recipient on Smita’s behalf but by using an external mail server or mail sending tool. In this method, Ravi sends a mail using a tool and using the services of an open relay server on the Internet or by creating his own server. He composes the mail and sets Smita’s email address in the mail header (MIME structure).
Modern mail servers and mail landing services, now easily detect this by
IP reputation of sender: The IP of the server from which the mail originates should belong to the sender’s domain as designated by the Sending organisation in the SPF record in the DNS. Hence, while Ravi will succeed in sending the mail, the recipient server will reject the mail in all likelihood due to bad IP reputation.
In case the recipient mail server doesn’t have strong policies, it may accept the mail and deliver it to the recipient and the mail may appear that it came from Smita. On closer inspection of the mail, it is possible for the recipient to determine that it is a spoof however, but this needs technical expertise of understanding mail headers.
Changing the display name on a public email platform
Another way Ravi can do this is to simply change the display name of his email id on whatever email solution he is using and send a mail to Smita. In this case, the email is a legitimate email from a valid email service, with a display name matching Ravi. So actually this is not a spoof mail at all. It’s simply a mail from a person whose name matches the name of one of the people Smita knows. This is perfectly legit and cannot be blocked at any level.
Email spoofing type 4: In transit modification of mail:
Normally mail can sniffed in transit (i.e. the contents can be read) and modified at hop points by mail administrators if they have the privilege. Under normal working conditions these are rare situations.
Ensure that the following practices are followed to bring the chance of this happening to near zero
SSL/TLS: All access to services happens over secure encrypted layer. This means that all the data flow from client to server and server to server is encrypted over SSL/TLS.
Use SPF, DKIM, and DMARC for mail SENT from your domain to prevent outbound email spoofing
Besides the mail security controls described above, there are 3 more powerful and effective ways to control impersonation, email spoofing and email fraud for your email domain as described below:
1. Sender Policy Framework (SPF)
The Sender Policy Framework (SPF) record enables a domain to publicly state which servers may send emails on its behalf.
You’ll need to add/modify DNS records of type ‘TXT’ via your domain registrar, or DNS provider, for your organisation’s domain to mention that the emails of your domain are being sent via Mithi SkyConnect servers.
Please refer to this topic to know what changes to make in your DNS records to enable SPF for your domain
This record will help protect your domain from attackers/spammers who send emails (on your behalf without your knowledge) with spoofed headers and will add more legitimacy to your sent emails.
2. Domain Keys Identified Mail (DKIM)
Domain Keys Identified Mail (DKIM) is a method of email authentication that cryptographically verifies if an email is sent by authorized servers and has not been modified during transit.
During onboarding, our team will request your approval to enable DKIM for your domain and provide you with the content of the TXT record to be added to your domain’s DNS zone.
Even if you don’t enable this during onboarding, you can do it any later point in time by simply reaching out to our support desk.
This record will help the recipient verify the integrity of the email and also confirm the source.
3. Domain-based Message Authentication, Reporting, and Conformance (DMARC)
Now after verifying the email’s source using SPF and DKIM, the question is what should a receiving server do if it gets an email which failed these checks.
This is where Domain-based Message Authentication, Reporting, and Conformance (DMARC) comes in by allowing the domain owner to specify what should happen with failed checks as well as get feedback/reports.
When you opt to enable DKIM for your domain, our team will give you the content for DMARC TXT record as well, which you will have to add to your domain’s DNS Zone.
Use SPF, DKIM and, DMARC for mail RECEIVED by your domain to check inbound email spoofing
The above section described how you can help your recipients identify mail of your domain accurately by setting up SPF, DKIM and, DMARC in your domain’s DNS zone.
This section describes how you as a recipient can enable tighter sender verification by enabling these checks on our inbound email gateways. Our gateway security engine at the periphery has options to check SPF, DKIM and DMARC records of all senders to your domain.
During onboarding, we only enable the SPF check and will ask your approval to enable DKIM and DMARC.
By default, if an inbound mail fails these checks, the action taken will be determined by the owner of the sender’s domain via their DMARC record. Typically, most owners configure this for quarantine, meaning mail failing these checks will be marked as spam and quarantined for your user’s review.
Data leak prevention with SkyConnect – Critical to your security strategy | <urn:uuid:dc8b584e-7f01-4651-8b00-3b09440c4c9c> | CC-MAIN-2022-40 | https://skyconnect.mithi.com/blogs/email-phishing-spoofing-and-protection-with-dkim-dmarc-spf-education/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338280.51/warc/CC-MAIN-20221007210452-20221008000452-00126.warc.gz | en | 0.907673 | 2,176 | 2.859375 | 3 |
Why Trust Matters in AI
We can all agree that AI has the potential to help businesses, organizations, and society solve real problems. But there are still many concerns about the consequences of using AI improperly. Things like ethics, privacy, bias, and security are top of mind. In order for AI projects to be fully embraced, companies must address these concerns, because people need to trust their AI in order for projects to be successful. In this installment of our blog post series, we will explore the question of why trust is an essential component to any conversation around AI.
Performance, Operations, and Ethics
Trust in AI is multidimensional. AI creators, operators, and consumers all have different needs and different factors that they consider when evaluating and determining if an AI application is trustworthy. For example, with a consumer-facing application, the requirements of trust for the business department, who created and owns the AI app, are very different from those of the consumer who interacts with it potentially on their own home devices. To satisfy the needs of different stakeholders, it can be helpful to organize trust in an AI system into three main categories:
- 1. Trust in the performance of an AI/machine learning model.
- 2. Trust in the operations of an AI system.
- 3. Trust in the ethics of the workflow, both to design the AI system and how it is used to inform a business process.
In each of these three categories, we identify dimensions of trust that help define them more tangibly. Combining each of the dimensions together holistically constitutes a system that can earn your trust. Another blog post in this series entitled, How to Build Trust in AI, does a deep dive into all 13 of the dimensions but, for this post, we’ll focus on answering the question of why trust in AI matters.
AI Success Hinges on Trust
As noted in a 2020 European Commission whitepaper, “The current and future sustainable economic growth and societal well being increasingly draws on value created by data.” The whitepaper goes on to say that “AI is simply a collection of technologies that combine data, algorithms and computing power.” This means AI has the potential to deliver many benefits to society, such as improved health care, fewer breakdowns of household machinery, safer and cleaner transport systems, and better public services. For business, it can help foster a new generation of products and services in areas like machinery, transport, cybersecurity, agriculture, the green economy, healthcare, and high value-added sectors like fashion and tourism. For public interest, it can help reduce the costs of providing services
by improving the sustainability of products and by equipping law enforcement authorities with the best tools to ensure the security of citizens.
Society has a lot to gain by embracing AI. But before it can reap those rewards, people have to trust AI. While lack of investment in AI and a skills gap are holding back full realization of AI’s potential, lack of trust is the main factor holding back a wider adoption of AI. Therein lies the answer to the question of why trust matters in AI and how it is the cornerstone of AI success. | <urn:uuid:e6c43689-0bbb-4cf2-af4b-cfbcf9436d0d> | CC-MAIN-2022-40 | https://www.datarobot.com/blog/why-trust-matters-in-ai/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338280.51/warc/CC-MAIN-20221007210452-20221008000452-00126.warc.gz | en | 0.94635 | 649 | 3.125 | 3 |
If we speak about what is networking and then try explain how does networking work or maybe how does internet work in some technical way, soon we will have to mention the routing protocols. We should mention routing protocols as the most important networking protocols when it comes to transfer packets from origin to destination points on the network. After all they are responsible for finding the best way across our multipath network.
|Check out some other stories about protocols:|
A routing protocol (either BGP or OSPF) is simply a set of rules, messages, and of routing algorithms. The key purpose of such internet protocols is to be taught the routes to some other networks by the routers which aren’t actually linked to them. Anyway, with a dynamic routing protocol, routers can discover virtual paths with dynamism as regards to the other distant networks. Moreover, with these protocol’s help such networks can be automatically added to a network’s particular routing table. But two most important routing protocols categories are in use today. And name of them are as: IGP (Interior Gateway Protocols) which is the majority suitable for lone independent system. Some examples of routing protocols that fall under this category are as follow: are IGRP, RIP, EIGRP, and OSPF.
But EGP (Exterior Gateway Protocols) routing protocol is used between the different systems routers. And BGP is the classification of EGP (a common routing protocol).
Facts about RIP (Routing Information Protocol)
RFC (Request for Comment, that is a description of a standard for networking protocols.) for the routing information protocol is 1058. A routing daemon (a program) can insert the routing course of action to the system. If the routes to the destination are many then the best one will be chosen for forwarding. Anyhow, the RIP message may contain: command, version (ver1 or ver2), family, IP address (32 bit), metrics (hop count) and information about around 25 possible routes. But RIP message set-up may contain: request, reply, asks over to throw the entire or part of routing table for the system, and poll entry etc. Routing table broadcasts are sent periodically by RIP to the neighboring routers. But due to the certain drawbacks of routing information protocol (RIP) like: having no subnet addressing information and consume lots of time for making a link stabilized, after its failure; RIP Ver2 is defined by the RFC (Request for Comment) 1388. Some added fields of this version is included a 32 bit subnet mask, an IP address of the next hop and a routing domain (daemon identifier) etc.
Facts about OSPF (Open Shortest Path First)
Open Shortest Path First as a link state protocol checks the link status with respect to each of its connecting neighbors. After that, obtained information is sent to them. It (OSPF) has the quality to stabilize a route or link after its failure in much more quicker way than those systems based on the distance vector protocol. Features of OSPF: support subnet Mask, allocate traffic regularly over the equivalent cost routes, can use multicasting and use IP directly.
Costs for specific hops can be set by the administrators. Adjacent routers swap information instead of broadcasting to all routers.
Facts about BGP (Border Gateway Protocol)
RFC 1267, 1268, in addition to 1497 have expressed BGP protocol that is running over the top of TCP (transmission control protocol) using the port No 179. Updates are triggered whenever required. Some other features of BGP protocol are as: distance vectoring usage, failures detections by sending periodic messages and information exchanges about accessible networks etc.
Facts about EIGRP (Enhanced Interior Gateway Routing Protocol)
Enhanced Interior Gateway Routing Protocol is an improved edition of the Cisco Interior Gateway Routing Protocol (IGRP). It is considered as both an interior gateway protocol (IGP) and an exterior gateway protocol, meant for the purpose of inter-domain routing. An EIGRP running router can store information about its neighbors’ routing tables. Fast convergence besides support for the variable-length subnet masking and for partial updates is possible with it.
EIGRP can use: bandwidth, delay reliability and load metrics traits while its components are including: neighbor discovery/recovery, and reliable transport protocol etc. From internet point of view, packet routing is generally divided into two sets, named as: interior and exterior routing. But the routing is performed with an algorithm help that is stored up in a router’s memory. Moreover, you can also observe two main categories of routing algorithms: distance vector and link-state.
Distance Vector Routing is chiefly determines the best paths without concerning its destination. Distance value is corresponded to by calculating more than one metric. Well, an IP distance vector routing protocols like RIP v1 (Routing Information Protocol v1), RIP v2 or Routing Information Protocol v2 and Interior Gateway Routing Protocols are still in use because these are simple as well as excellent at their job in the small networks with a little management, if it is required.
But Link State protocols are able to use sophisticated methods for taking into concern link variables like bandwidth and reliability.
Facts about IS-IS (Intermediate System to Intermediate System)
IS-IS as a routing protocol is able to forward information within a physically connected computers group with efficiency. With respect to a packet switched network, the task of routing is done by finding the best path for the datagrams.
The inter-domain routing protocol “IDRP” is consisted on the path-vector algorithm for routing. It provides the routing job for the OSI defined networking environments, which is more-or-less alike to BGP protocol in the TCP/IP networks. | <urn:uuid:c05daab4-3189-44ec-b87a-485fb291f881> | CC-MAIN-2022-40 | https://howdoesinternetwork.com/2012/dynamic-routing-protocols | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030331677.90/warc/CC-MAIN-20220924151538-20220924181538-00327.warc.gz | en | 0.917464 | 1,207 | 4.28125 | 4 |
HTTPS has become the protocol of choice for any serious website, but effectively enforcing the use of HTTPS instead of HTTP requires the HTTP Strict Transport Security header, or HSTS. By sending the HSTS header with suitable parameters, the server informs the visiting browser that only the HTTPS version of the requested site is available, and plain HTTP will not be served. To avoid redirects at the start of every visit to the site, the browser remembers this information for the duration specified in the response header. In this article, we will look at the history of HSTS, see how it works and how to set it up, and learn why using it can actually drive traffic to your website.
Just Serving HTTPS Is Not Enough
HTTPS was introduced to allow secure communication – something the plain-text HTTP protocol was never intended for. To ensure end-to-end security, all communication between the client and server must be encrypted, but prior to HSTS, there was no way for websites to enforce encrypted communication or indicate that only HTTPS is supported. This left users vulnerable to man-in-the-middle attacks with SSL stripping (as demonstrated in 2009 by Moxie Marlinspike), where the attacker intercepts the browser’s request for an HTTPS page and forwards it to the server as a request for plain HTTP. The user is then served plain-text HTTP content which the attacker can easily view and manipulate. Although the user could then see that only HTTP was being used, without HSTS there was no way to determine if the connection should have used HTTPS in the first place.
In 2012, the IETF published the HTTP Strict Transport Security (HSTS) specification as RFC 6797 to address this gap in the HTTPS encryption chain. Currently, all major web browsers support HSTS. The
Strict-Transport-Security HTTP response header allows servers to indicate that content from the requested domain will only be served over HTTPS. When this header is specified in web server responses, any attempts to fetch the plain HTTP version of the site are redirected to the HTTPS version, with no tolerance for certificate errors. To avoid redirects during future visits, the server also indicates how long the browser should remember the site as HTTPS-only. Let’s see how this works in practice.
How HTTP Strict Transport Security Works
Most users don’t specify the protocol in URLs typed into the address bar, so if you request www.netsparker.com (or just netsparker.com), the browser will assume the default HTTP protocol and send an HTTP request to http://www.netsparker.com. Because the Netsparker site uses HSTS to enforce HTTPS-only communication, it responds with a redirect to the HTTPS site (301 response code) and includes the
Strict-Transport-Security response header to indicate that only the HTTPS version of the site will be served.
Strict-Transport-Security header can specify three directives:
max-ageis the only mandatory directive and indicates how long the browser should remember that the site is HTTPS only. The
max-agevalue is given in seconds, so the typical expiry periods of 1 or 2 years correspond to 31536000 or 63072000. If
max-ageis 0, the browser will forget the site and treat it as a new one on the next connection attempt.
includesubdomainsis optional and indicates that HTTPS is also required for all subdomains of the specified domain. For maximum security, the HTTPS redirect with the
Strict-Transport-Securityheader should include the
includesubdomainsdirective and reference the base domain (e.g. netsparker.com), so all its subdomains (especially the www subdomain) are covered by HSTS.
preloadis also optional and indicates that the site meets requirements for HSTS preloading and is on the HSTS preload list or has applied for it – see below for more information on preloading.
As soon as the browser receives a response with
Strict-Transport-Security, it knows that it must use only HTTPS for that site during the entire max-age period. If HTTPS communication is unavailable, for example because HTTPS is not set up properly or a certificate has expired, the browser must terminate the connection. From now on, every time you visit this site, your browser will automatically send an HTTPS request, regardless of the protocol you specify – netsparker.com and http://www.netsparker.com will both direct you to the HTTPS site. Apart from increasing security, this also improves load times, since you don’t have to wait for the server’s redirect response every time you request a URL without entering the https:// prefix.
Preloading: a Remedy for the Bootstrap Vulnerability
Once the browser has set up HSTS for a site, all communication with that site will use HTTPS, which eliminates the SSL stripping threat and guarantees secure transmission. But the first time you visit an unknown site, the browser relies on the server’s responses to determine the correct protocol – this is the Trust On First Use principle (or TOFU for short). Attackers or malware could intercept this first visit to redirect you to an insecure website using a man-in-the-middle attack. This is not as rare as it seems – it just means that the site is not present in the browser’s HSTS database. While this can be the result of a time-based attack via NTP, it can also happen if the
max-age expiry period since the last visit has passed, the browser is a fresh install, the server is misconfigured (for example
max-age was left at 0 after testing) and so on.
To remedy this and to improve page load speeds, all major browsers (including Chrome, Firefox, Internet Explorer and Safari) contain hard-coded lists of known HTTPS sites. This is called preloading, and is based on the Chromium (Chrome) HSTS preload list. If a site is listed, the browser always connects to it using HTTPS, even on the first visit. This effectively eliminates the only significant vulnerability in HTTPS enforcement, ensuring end-to-end encryption. The only weak point is that the updated list is only distributed with new browser versions, so the update process is slow and if you added your site recently, there is no guarantee that any individual browser installation will already have your site preloaded.
To add your site to the list, first make sure that it meets all the requirements for preload submission:
- Present a valid certificate
- Redirect all HTTP requests on port 80 to HTTPS
- Include the HSTS header in any additional HTTPS redirects
- Serve all subdomains over HTTPS
- Specify all header directives:
max-agemust be at least 31536000 seconds (1 year), and
preloadmust be present
When you’re sure everything that everything works, has been thoroughly tested, and all requirements will continue to be met in the future, you can submit your site to the preload list and start including the preload directive in your HSTS response headers. Remember that by adding the site to the
preload list, you are telling browsers to only use HTTPS, so any misconfiguration or invalid SSL certificate can result in your site becoming completely unavailable to users.
With properly implemented preloading, HSTS can ensure that a browser initiates and uses only secure connections to HTTPS-serving sites. Of course, the number of secured websites continues to grow, so it’s clear that preloading is only a temporary solution – at some point the static list may become too large to maintain and distribute.
NTP Attacks on HSTS
For the present, HSTS is a fairly robust way of enforcing HTTPS connections. The only practical approach to compromising HSTS is based on attacks against the Network Time Protocol (NTP) that attempt to manipulate system time by faking time values from NTP servers. This allows attackers to fool the browser into expiring HSTS entries and allowing insecure HTTP connections. Note that this is not a problem intrinsic to HSTS – NTP vulnerabilities can also be used for attacks against other security technologies and protocols, including SSL/TLS, Kerberos and Active Directory.
Browser Tracking Using HSTS Supercookies
Implementing HSTS has also introduced a potential privacy issue related to browser fingerprinting. Because each browser stores a separate database of known HTTPS-only sites, attackers can use specially prepared sites to fingerprint the browser by checking its reaction to a suitably large number of requests to domains controlled by the attacker.
For example, a website might contain multiple single-pixel beacons, each requested over HTTP from a different subdomain controlled by the attacker. By specifying or omitting HSTS headers in specific replies, the attacker can store a potentially unique pattern in the browser’s HSTS database, in effect assigning a fingerprint. When the browser visits the site again, it will attempt to use HTTPS to load the beacons that initially had the HSTS header in replies, and plain HTTP for the remaining ones. By reading this pattern, the attacker can identify and track the returning browser. No cookies are used, and fingerprinting works across multiple sessions and regardless of incognito mode, so these identifiers have been dubbed “supercookies”.
Figure 1. Browser fingerprinting based on HSTS behaviors
How unique this fingerprinting is depends on the number of beacons used. Each beacon provides 1 bit of information, so with 5 beacons you can get 2^5 = 32 values, but increase this to 30 beacons and you have 2^30 – that’s over 1 billion unique identifiers. Attackers can also reset bits previously stored in the browser by sending
max-age=0 in the response header, giving them the ability to store multiple bits of information in your browser many times. While it seems that at present this vulnerability is largely theoretical and not used in the wild, it’s good to know about it, as there is no way around it without changing the HSTS spec.
Benefits and Caveats of Enabling HSTS
For users, the obvious benefit of deploying HSTS is improved security through true end-to-end encryption for HTTPS communication. HSTS headers are only valid over HTTPS connections, so using HSTS guarantees that no unencrypted HTTP traffic is sent. Combined with preloading, HSTS also improves page load times by eliminating server redirects from HTTP to HTTPS. Because page loading speed is a vital part of the user experience, especially with mobile devices, faster loads can improve a site’s SEO metrics and drive more traffic your way – so enabling HSTS and preloading makes good business sense.
On the flip side, any HSTS deployment must be carefully planned and considered, especially if you use preloading. Remember that after you enable HSTS, any attempts to connect to your site or web application using plain HTTP will be rejected, and any configuration issues on your web server (such as an expired certificate) might render the site completely inaccessible. You also have to use HTTPS for every single page in your domain and all its subdomains (if you specified
includesubdomains), which for complex sites might require additional work.
By enabling HSTS, you are committing yourself to maintaining HTTPS and meeting all preloading requirements continuously and for a potentially indefinite time. However, if you can ensure this and your business model supports HTTPS-only operation, the benefits from improved security and user experience can be significant. | <urn:uuid:92d1f1bf-7ffa-4b66-9991-3d1123d20530> | CC-MAIN-2022-40 | https://www.invicti.com/blog/web-security/http-strict-transport-security-hsts/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334620.49/warc/CC-MAIN-20220925225000-20220926015000-00327.warc.gz | en | 0.887303 | 2,431 | 3.5 | 4 |
There is now an official IPv6 prefix set aside for documentation purposes: 2001:0DB8::/32. (Leading zero courtesy of APNIC.)
The how and why is documented at a page at APNIC. Note that there is also a prefix set aside for documentation purposes in IPv4: 192.0.2.0/24. See RFC 3330 for more information and other special IPv4 prefixes.
At prik.net there is now a list of IPv6-enabled hosts or sites. I have no idea how complete the list is, but it has more than 3000 entries so it's better than the manually maintained stuff in some other places. If the link doesn't work, this is probably because your browser doesn't understand compressed content. In that case, use the uncompressed version. The compression ratio is about 1 : 6.Permalink - geplaatst 2004-01-04
The RIPE NCC has changed its policy regarding the initial allocation that new LIRs receive. The rule that efficient use for at least a /22 must be demonstrated is now off the table, and the minimum allocation is now a /21 rather than a /20. See the announcement. RIPE also maintains a list of minimum allocation and assignment sizes for their address blocks (linked from the announcement), but this is pretty much useless because filtering on allocation size is too restrictive while filtering on assignment size is isn't restrictive enough for many address blocks. So be very careful when implementing prefix length filtering.
Without the jargon, please!
Right. Most of us get our IP addresses from our ISPs, and ISPs usually have one or more blocks of IP address space of their own. Having their own address space is important for ISPs because this allows them to be independent from their ISPs by allowing them to change ISPs without having to change addresses. (Obviously this is useful to end-users as well, but this changed policy applies to ISPs.) Until now, ISPs that wanted to get address space of their own needed to show that they and/or their customers would start using 1024 addresses (a /22) immediately. In this case, they would get a block of 4096 addresses (a /20). The advantage of having such a large block is that everyone in the world is prepared to store a pointer to it in their routers, making the addresses globally usable without limitations.
Since some networks only accept routing information for the smallest address blocks that RIPE and the other Regional Internet Registries (ARIN, APNIC and LACNIC) give out to ISPs. Smaller address blocks aren't entirely useless, but they may not be globally reachable without having to depend on the ISP the addresses came from, which of course limits ISP independence.
Since RIPE is now giving out blocks of 2048 addresses (/21) from some of their address blocks, networks are expected (and pretty much forced) to accept these blocks. This is good news for small ISPs that want their own independent block: they no longer have to jump through hoops trying to show they need 1024 addresses, or make do with only semi-independent addresses.
Note that the other RIRs haven't changed their policies (or at least there are no announcements to be found). ARIN's policy for instance, is even more restrictive than the old RIPE policy: multihomed networks must show efficient use of a /21 to get a /20, single homed ISPs must even show efficient use of a full /20 to get a /20. So for now the good news only applies to ISPs in the RIPE region, which is roughly Europe, the Middle East, Africa north of the Sahara and the former Soviet Union. For more info, see the RIR policy comparison matrix.
Permalink - geplaatst 2004-01-10
As I wrote a few weeks ago in an article under the name "no ip unreachables", path MTU discovery doesn't work all that well across the internet in practice. Since then, I've noticed that people end up on this site looking for ways to clear the don't fragment bit in the IP header. So here is an example of how to do this on a Cisco router:
! route-map nodf permit 10 set ip df 0 ! interface FastEthernet2/0 ip policy route-map nodf !Note that the "ip policy route-map nodf" command must be applied on the interface receiving the packets for which the DF bit must be cleared, and not the interface with the reduced MTU itself, where the packets are subsequently transmitted. See a page at Cisco for additional strategies. Permalink - geplaatst 2004-01-12
IPv6-enabled operating systems such as Windows, Linux and FreeBSD all come with a web browser that also supports IPv6 and prefers IPv6 when both IPv4 is available. Things are slightly different with Safari, Apple's browser application. Initially, Safari only supported literal IPv6 addresses and some corner case DNS names. In the current version, Safari will do IPv6 if no IPv4 address is available, but it won't prefer IPv6 over IPv4 or fall back to IPv6 when IPv4 doesn't work.
However, Nicholas Humfrey has come up with a trick. By enabling Safari's debug mode and switch off one of the two HTTP loaders that are normally used, Safari will prefer IPv6 when it's available. See the Mac OS X hints article that Nick posted for the details.Permalink - geplaatst 2004-01-28
It has been a while since the last news posting. My apologies for that. Here is something to hold you over until I can find some real news:
When perusing the HTTP referrer log, I noticed that a lot of people are finding this site in search of "bgp+2500" or something similar. So... is it possible to run BGP on a Cisco 2500 router?
The short answer is "yes". The IOS images for the 2500 support BGP, including BGP for IPv6. (They do not, however, support OSPF for IPv6 even though OSPF for IPv4 is supported and the "ipv6 router ospf" command may exist. Same thing for IS-IS: the command exists, but the protocol isn't present in any of the 2500 images.)
The slightly longer answer is that a 2500 is of limited use for inter-domain routing. Actually way back when I got started with BGP I used a 2514 with 16 MB RAM and it could hold the entire 35000 or so entry global routing table. From two upstreams even, if I remember correctly. However, in the mean time the global routing table has gotten four times as big, and the 2500's memory limit is still 16 MB. The fact that the 2500 series sports a 68030 CPU doesn't really help either. All of this means that you can only run BGP on a 2500 if you don't send it more than a few thousand routes. You also shouldn't send it a full feed and have the 2500 filter out the unwanted routes, as this will tax the CPU too much and make for many-minute convergence times.
Note that even 35000 routes wouldn't work anymore today, as modern IOS images need a lot more memory for their internal house keeping. On bigger routers it gets even worse because unlike the 2500, those can't run their software from flash, so it must be copied to RAM. Additionally, the switching path of choice is CEF these days, which takes a lot of memory. (Without CEF you'll be using fast switching which uses just as much memory but only when needed, so it's not only the CPU that melts down but you also run out of memory when a slammer-like worm hits.)
So you may be able to get away with a full table on a Cisco with 128 MB RAM (or you may not), but 256 MB gives you much more elbow room. Unfortunately, Cisco still makes boxes that can run BGP but won't take enough memory to do so properly. A good example are the 3550 series multilayer switches.Permalink - geplaatst 2004-03-31
At the NANOG 26 meeting in october 2002, Dave Meyer presented a very simple proposal to protect BGP sessions against attacks: set the TTL to 255 on outgoing packets, and check whether the TTL in received packets is equal to 255. Since routers always lower the Time To Live (or Hop Limit in IPv6) when forwarding a packet, and routers discard packets with a TTL of 0, there is no way for anyone who isn't attached to the subnet in question to inject packets with a TTL of 255 into a subnet. RFC3682 was published in february and describes the details of the "Generalized TTL Security Mechanism (GTSM)".
Cisco has now included GTSM into IOS release 12.3(7)T, as explained in the feature guide. It seems there are some interesting caveats. First of all, Cisco states that enabling the feature using the neighbor ... ttl-security command will only enable the check for incoming packets and not change any behavior as to outgoing packets. So this must mean they always use a TTL of 255 for outgoing packets now. However, older IOS versions set the TTL for BGP packets to 1 (in the absence of any ebgp-multihop settings). If this is the case, then detecting whether a neighbor is directly connected won't be very reliable right now. Then again, Cisco says the feature must be configured on both ends of an eBGP session (no support for iBGP as of yet) which seems to contradict this.
Another thing is that they look for a TTL of 254 or higher. This suggests that for incoming packets with a TTL of 255, they first decrease the TTL and then go on to process the TCP segment. Again, this is not how things work in older IOS versions, as it's perfectly possible to set the TTL to 0 and still interact with a Cisco router on the local subnet. So unless something changed in this regard as well, accepting a TTL of 254 means that there can still be a router in between!
Note that this mechanism only offers protection against attacks on port 179 of a router from "far away". Anyone on the local subnets still gets to do whatever they please and the content of the BGP sessions isn't protected any better than before.Permalink - geplaatst 2004-04-08
Rumors have been floating around for days, as the referrer log for this site shows large numbers of people looking for "BGP hack" and "BGP MD5". But the cat is out of the bag now:
Please see the page linked above for detailed information. The short version is that TCP sequence numbers turn out much easier to guess than assumed until now, which makes long-lived TCP sessions vulnerable to reset attacks. Since BGP sessions can remain for days, weeks or even months, and other pertinent information is relatively easy to find, BGP is the protocol most affected by this vulnerability.
Fortunately, the BGP TCP MD5 option protects against exactly this problem. Enable it if at all possible. Most, if not all, routers support it. The option is enabled on Cisco routers as follows:
! router bgp 12345 neighbor 192.168.0.1 password use-upto-80-characters !
However, this will break any running BGP sessions so coordinate the change closely with the remote AS.
Since this mechanism operates at the TCP level, host-based routers such as Zebra or Quagga running on BSD or Linux typically don't support this option. However, there is some rudimentary support in both OSes, see the SANS advisory.
Note: The "BGP TTL hack" or GTSM (see below or above) also offers protection against the TCP vulnerability, without adding the MD5 crypto overhead. And good anti-spoofing filters do the same, but the problem there is that the other AS also needs to implement them, something that can't be assumed.
It seems the actual risks aren't as bad as the reports seem to indicate at first glance. I'll post a more detailed analysis later, but from discussions on NANOG it seems the only new aspect is that previously people didn't realize that the RST packet could have any sequence number that falls inside the receive window on the potential victim, which is often around 16k. This means the attacker only has to guess the first 18 bits of the sequence number rather than the full 32 bits. However, she also needs to guess both port numbers, which makes the number of possible combinations an attacker must try around a billion, which amounts to a DoS attack of 10000 packets per second for more than a day.Permalink - geplaatst 2004-04-20
In an article in Wired (Flaw Could Cripple Entire Net) Paul Watson is said to claim that he can reset TCP sessions " with as few as four attempts". I have a hard time believing this, but we'll have to see thursday, when all will be revealed. An attacker still needs to know the IP addresses for both sides and the appropriate TTL (simple) and the port numbers used on both sides. The latter may or may not be trivial: routers typically start using ephemeral ports at a fixed number after booting, so for the first few BGP sessions this should be relatively easy to guess. However, a router that has been running for a while and has lots of BGP sessions (which is common on internet exchanges), these port numbers are well randomized within the range used by the system, which is typically at least 4000 ports.
So if Paul Watson is correct it may be possible to reset sessions with less than a hundred to a few thousand packets. This takes only moments. If he isn't, but the router has few BGP sessions, port numbers are easily guess-able and the default window size of a little less than 16k makes it possible to reset sessions with about 250 thousand packets per port combination, which is in line with reports that people were able to do this in the lab within about half an hour. This is short enough to incur flap dampening difficulties if it happens repeatedly.
So what can we do?
The MD5 option is also a double edged sword because it opens the door to CPU exhaustion based denial of service attacks. In theory the crypto should only be done when the packet passes all regular TCP checks, but in reality this isn't the case so making the CPU burn cycles on MD5 hashing should be easier to do for an attacker than sending a successful RST. Based on the information I have right now, I believe the upsides of having MD5 on peering sessions with relatively small peers over exchanges don't outweigh the downsides, as the work and the MD5 DoS risks are the same for small peers, but the damage when a session breaks is fairly negligible.
For very large peers and especially transit connections, the situation is different: the instability caused by session resets can be significant, so MD5 is a good idea here.
! interface gigabit3/0 hold-queue 2048 in ! router bgp 12345 timers bgp 1 180 !
Lines in a Cisco access list for filtering BGP TCP RSTs look like this:
access-list 123 deny tcp any any eq bgp rst log-input access-list 123 deny tcp any eq bgp any rst log-input(Note that some legitimate hits are possible if a BGP session is only configured on one router.)
These must be applied on input on interfaces that may received spoofed RSTs (i.e., external connections, but also customer facing connections if those don't have proper anti-spoofing filters). The log-input keyword makes sure the interface and sometimes the MAC address of the system that sent the offending packet are logged. This is very useful on shared/switched media interfaces such as internet exchanges. Don't worry about overwhelming the router with logging information too much, as this is rate limited. (However, having log-input in place when a full-fledged DoS attack is in progress isn't advisable either.)
The referrer log for this site can be interesting reading at times. It seems several people have landed here when looking for information for the BGP TCP MD5 option and IPv6. Despite the fact that RFC 2385 doesn't mention IPv6, it's possible to have an MD5 password on IPv6 BGP sessions. At least, it is in recent IOS versions. I have a very old one that allows this to configured but it doesn't compute the checksum correctly. I assume this is particular to this specific version (from 1999), though.
Someone else seemed interested in BGP and Apple. Good news: under Panther, Zebra 0.94 compiles without trouble, so it's possible to run BGP on a Mac. Juguar/Zebra 0.93b didn't work for me.Permalink - geplaatst 2004-04-21
It is becoming clear that there are indeed systems that are vulnerable to having TCP sessions reset within only four tries, assuming IP addresses and port numbers are known. Unfortunately, there is little information about which systems have this vulnerability. However, judging from the secrecy at Cisco and Juniper, it is far from inconceivable that they are vulnerable. (If my suspicions are correct, the hole was fixed in FreeBSD in 1998 (!!!), though.)
The details will probably be public on thursday, and we can expect exploits very soon after that. Since the required number of packets to take advantage of the vulnerability is very low, having MD5 in place on BGP sessions is almost certainly a good idea, as it is unlikely a router will receive so many packets that the CPU is overloaded. Also, filtering BGP RSTs as outlined below/above where possible will make sure your routers won't terminate TCP sessions. However, your sessions may still be vulnerable depending on the status of your BGP neighbor's router.
So set up MD5 passwords on important BGP sessions (such as the ones to transit networks) as soon as possible.
The Cisco advisory on just the problem with RSTs and the window. If I interpret this correctly, fixed IOS versions are already available, even to Cisco users without support contracts. (Note that non-IOS products are also affected, and it's a good idea to upgrade anyway as per Cisco's recently uncovered SNMP vulnerability.)Permalink - geplaatst 2004-04-22
I hate to admit it, but I got infected by the hype surrounding the TCP "vulnerability". As it turns out, all of this was pretty much yesterday's news from the start. In a news.com article we can read Paul Watson complain that "it's crazy". No argument there. So here is the real story.
RFC 793 (TCP) clearly indicates that established TCP sessions must be torn down when an RST packet is received with a sequence number that falls within the current window.
Quick detour: TCP is responsible for making sure that all data from the sending application is received once, only once and in the correct order by the receiving application. In order to do this, it numbers every byte of data using a sequence number. At any time, TCP has a specific "window" in the sequence number space for bytes that it is prepared to receive. Bytes that fall before the window have already been received, so if those come in again they are ignored. Bytes that fall beyond the window are too far in the future and are also ignored. A packet or segment of data that starts with the first byte of the current window is processed immediately, and any data that falls further within the window is buffered and will be processed later as this data is received out of order.
RST packets are supposed to be generated in order to reset stale sessions that can for instance occur after one side reboots. So when the other side sends a packet, the system that just booted doesn't know this session and sends back an RST packet in reply to the data packet, copying the sequence number in the process. The sender of the original packet now receives the RST with a sequence number that obviously falls within the window, so the session is torn down.
The important part here is that according to RFC 793 the sequence number doesn't have to be an exact match: as long as it's within the window, it'll be accepted. For some reason, many people assumed there would have to be an exact match. Since the sequence number is 32 bits in size, this means 4.3 billion possibilities. So if an attacker wants to reset a TCP session, assuming he already knows the correct IP addresses and port numbers, he would have to send up to 4.3 billion packets. But in reality the number of packets necessary to reset a session must be divided by the window size.
So what would be the window size for a typical TCP session, or, more importantly, a TCP session used for BGP? Well, if a system implements the RFC 1323 TCP high performance extensions, the window size can be almost a gigabyte. This is where Paul Watson's claim that resetting a TCP session can be done with "as few as four packets" comes in. However, this is pure nonsense as such huge windows are never necessary. Even when moving data across the globe at 10 Gbps a window less than half that size is more than sufficient. BGP isn't exactly in the business of moving data across the globe at high speed over TCP, so it only requires a very modest window size. These are some of the initial packets for two BGP sessions between a Cisco router and a FreeBSD box running Zebra:
09:42:06.537772 IP 18.104.22.168.11164 > 22.214.171.124.179: S 3863598700:3863598
700(0) win 16384 <mss 1460>
09:44:35.984407 IP 126.96.36.199.54660 > 188.8.131.52.179: S 1733311468:1733311 468(0) win 65535 <mss 1460,nop,wscale 0,nop,nop,timestamp 3272949461 0>
09:44:36.003219 IP 184.108.40.206.179 > 220.127.116.11.54660: S 947956194:94795619 4(0) ack 1733311469 win 16384 <mss 1460>
The packet from the Cisco router doesn't even have the window scale option. The FreeBSD machine has the option but doesn't bother to actually use it. This means that the maximum window for the subsequent session is limited to 65535 bytes. However, routers almost universally use around 16000 bytes. Bottom line: in order to reset a BGP session using a TCP RST, it is necessary to send between 65 and 268 thousand packets. This takes several minutes at DSL speeds. So if IP addresses and port numbers are known, resetting a BGP session isn't too hard for an attacker. However, if the attacker must also guess the ephemeral port, then the time it takes to reset a session becomes too long to make this an interesting attack vector.
In theory it's even better than this, at least on Cisco routers, because those rate limit the handling of RST packets. Unfortunately, RFC 793 also suggests that TCP sessions should be terminated if an in-window SYN rather than a RST packet is received. SYN packets are normally used to open sessions, so the rationale for terminating sessions when unexpected SYNs come along is murky at best.
Last but not least, there was a bug in several TCP implementations that allows any "left of window" RSTs to be acted upon
without further checks. This bug was fixed in 1998 in FreeBSD. From the FreeBSD 4.9 /usr/src/sys/netinet/tcp_input.c file:
* First check the RST flag and sequence number since reset segments * are exempt from the timestamp and connection count tests. This * fixes a bug introduced by the Stevens, vol. 2, p. 960 bugfix * below which allowed reset segments in half the sequence space * to fall though and be processed (which gives forged reset * segments with a random sequence number a 50 percent chance of * killing a connection).
This bug was still present in NetBSD until a few days ago. This led me to believe that the real issue here was that this same bug was also present in the TCP code from one or more major router vendors, and that this was what Paul Watson was talking about. But that didn't turn out to be the case, so essentially the whole story boils down to "router vendors implement TCP according to RFC 793". Big news indeed.
Permalink - geplaatst 2004-04-24
Canon PowerShot A40 f/8, 1/160, ISO 50, 5.4 mm (2004:06:13 21:58:19)
Image link - posted 2004-06-13
Since the announcement was made on the 35th anniversary of the first moon landing, I think it's appropriate to modify Neil Armstrong's famous quote slightly:
"One small step for mankind... but a huge step for ICANN."
Yesterday, ICANN (the Internet Corporation for Assigned Names and Numbers) announced that IPv6 nameservers for .jp and .kr were added to the root zone. Now obviously this is a necessary step for making it possible to resolve domain names using IPv6, but it's really not a big deal as IPv6 delegations have been around lower down the delegation chain for a very long time, and no reasonable DNS implementation has a problem with this. Also, ns.ripe.net has had an IPv6 address for some time now, and this nameserver is secondary for country domains such as .nl and .it, not to mention the reverse mapping of huge slabs of IP address space. The difference between that and the new .jp and .kr delegations is that those also have an AAAA glue record, while ns.ripe.net only has an A glue record.
But the real issue here is having IPv6 addresses for the root servers themselves. There has been considerable experimenting with actually running root service over IPv6 transport. However, in order for an IPv6-only host to be able to get at the root servers, those IPv6 addresses must be published in two places. Number one is the named.root "hints" file, which DNS servers use to find the root servers on startup. It's fairly trivial to add IPv6 addresses to this file. This would allow a DNS server to perform the first query over IPv6. Unfortunately, that first query is for the full set of root servers. Since there are 13 of those, the response to this query comes fairly close to the maximum response size of 512 bytes. Adding IPv6 addresses for all root servers isn't possible without going over this limit, which has the potential to cause several problems. I'm very interested to see how ICANN is going to handle this, and especially when they're going to handle this.
As a bonus, at long last E.F.F.3.IP6.ARPA has been delegated so that 6bone addresses (3ffe::/16) can now enjoy proper reverse service in the DNS.Permalink - geplaatst 2004-07-21
Update: this is what it looks like.
So apparently Japanese for "gazelle" (as on the cover of the English version) is "mule"...Permalink - geplaatst 2004-08-21
Vanaf 16 oktober 2004 gingen de trams die voorheen over de Grote Markstraat in Den Haag reden via de tramtunnel eronderdoor rijden. Dat waren toen lijnen 2, 3 en 6: Randstadrail zou pas een paar jaar later komen. De HTM pakte groots uit met een tram-schouw elders in de stad en daarna vrij reizen om het allemaal te bekijken.Lees het artikel - geplaatst 2004-10-16 | <urn:uuid:8ba964e2-1b99-49f5-8126-59653c0f40af> | CC-MAIN-2022-40 | http://alles.iljitsch.com/2004/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335257.60/warc/CC-MAIN-20220928145118-20220928175118-00327.warc.gz | en | 0.944301 | 5,928 | 2.71875 | 3 |
Through the decades, I have been asked many times to explain the difference between data and information — in lay terms. The reason is that today, data is the reference for everything.
Let’s begin by looking at the definition of data. According to Merriam-Webster, data is “factual information (such as measurements or statistics) used as a basis for reasoning, discussion or calculation.” Note here the reference to “information” and the assumption that we know the context in which this information, or data, is collected and presented.
Let’s take it a step further and look at raw data — without context. For example: 42. If I were to provide you with this number — 42 — you have no sense of its meaning or purpose. When left for you to determine, and dependent upon your point of reference, you may think 42 is the number baseball player Jackie Robinson wore, or the answer to life, the Universe and everything as presented in Douglas Adams’ novel The Hitchhiker’s Guide to the Galaxy.
Which Is Right?
In each case you would be correct because I provided no context to the number 42 for you to reference, and therefore, you are not able to correctly understand its purpose. As a data point, this would be near impossible to use since we don’t have the context. Now, if I tell you that 42 represents a monetary value, as in 42 dollars, you now have a better understanding that it is currency. If I expand this to represent the price I paid for a shirt, you now have information, not raw data, to use in analysis and other actions that will provide value.
Simplifying this more, data without context has no meaning or purpose. Data with context becomes information that we can understand. Information pulled together for a purpose, now transforms into content. So, data with context is information and the way we provide context is through the use of metadata that describes what the raw data represents.
In My View
This is where I may have a few folks throw rocks, but data without context is useless. 42, 53, car, elevator and any other number or word by itself, other than representing what it is, has no purpose or value. To know something is an elevator is good as a point of reference as to the item but serves no analytical value on its own.
Metadata provides the context by which elevator becomes useable information that can be analyzed, provided results and enable decision making. The better the metadata, the clearer the context and as such, the stronger our analytical capabilities in leveraging data to enhance our Business Intelligence initiatives.
When you strategize about how information will be captured and managed, look at the whole organization and develop a metadata strategy that will bring benefit not only for findability, but useability. 42. I leave it up to you to decide. | <urn:uuid:fd35a3e5-87b6-41c1-91f5-d9953b102e8e> | CC-MAIN-2022-40 | https://documentmedia.com/article-3168-Context-Is-Key-to-Understanding.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335573.50/warc/CC-MAIN-20221001070422-20221001100422-00327.warc.gz | en | 0.943784 | 593 | 3.203125 | 3 |
Penetration Testing, also referred to as pentesting or ethical hacking, is a service whereby a security professional is hired to "hack" into your network. In this article, we'll discuss what pentesting is, when you should consider pentesting, and how to use pentesting as part of your security regimen.
What is Penetration Testing?
Pentesting involves hiring a security professional to hack into your network. It's important to understand that pentesting is as much of an art as a science. As such, there are differences in approach and quality in penetration testers. Some penetration testers rely on social engineering to trick employees, while others focus on testing web applications to identify weaknesses in your webapps. Specifically defining – or scoping – your pentest is critical to getting the results you need.
Pentesting is NOT a vulnerability assessment - It is not the job of the pentester to find all the possible weaknesses in your systems. Rather, the pentester's job is to get into your network by any means possible (within scope) and show you how they do it. It's important to remember that a pentest is successful when they find one way in – not every way in.
Types of penetration tests:
Full-Scope (red team testing)
Network penetration test
Social engineering (e.g., phishing, smsishing, phone call attacks)
Web Application Pentest (specific for a web application)
Physical Pentest (Breaking into facilities)
There are also variations on each of these types. For example, it's common to do a full-scope without a physical pentest.
When should I consider a pentest?
Pentesting is a highly-skilled professional service that comes with a price tag. Ensuring that you get the most out of a pentest is critical and is something often overlooked.
Pentesting is used to test the results of your existing security program. If you don't have an existing security program, don't waste your money on a pentest (yet). First, we recommend a security risk assessment, combined with vulnerability assessments, to understand your current security posture. Then, after fixing (or otherwise addressing) the vulnerabilities found, you can consider hiring a penetration testing firm.
Prematurely hiring a pentester can help prove to management that security is necessary, but it will provide very little in terms of sustainable cybersecurity improvements.
How do I use pentesting for long-term security?
Pentesting is part of the overall security solution, not the complete security solution. We recommend scheduling a pentest once a year, following an annual security risk assessment. When reviewing the results of a pentest, it's tempting to simply fix the flaw that allowed the pentester to "get in," however, it's also wise to look at the root causes of any breaches. Common questions to ask include: Was it a lack of email security? Network security? Should this be in your risk assessment? Do you need to invest more?
In summary, a penetration test can be used to enhance your security program when used properly. Resist the temptation to hire a pentest company before doing basic cybersecurity first. Then, after the test, make sure to use the results and the findings to enhance your understanding of your overall cybersecurity posture.
For more information on this topic, please reach out to us; we're here to help! | <urn:uuid:60fbf7bc-c511-4575-8bdf-dee0491f97df> | CC-MAIN-2022-40 | https://help.coalitioninc.com/en/articles/3749881-using-penetration-testing-to-improve-security | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335573.50/warc/CC-MAIN-20221001070422-20221001100422-00327.warc.gz | en | 0.945434 | 693 | 2.671875 | 3 |
Spam from botnets accounted for 88.2% of all spam and the average spam levels in 2010 increased by 1.4% and reached 89.1%, compared with spam levels in 2009, according to a report from Symantec.
The spam rates increased at 92.2% in August 2010 when the Rustock botnet was aggressively seeded by new malware variants and was put to use, lending to an overall increase in spam activity for the year.
The report, MessageLabs Intelligence 2010 Security Report, revealed that by the end of 2010 a reduction in the contribution of botnets to spam, to 77% of spam, resulting from the closure of spam affiliate, Spamit, in early October 2010.
However, by the end of 2010, the number of active bots remained same as at the end of 2009, increasing 6% in the latter half of 2010.
The analysis also revealed that the total number of botnets worldwide is between 3.5 and 5.4 million.
The report predicted that in 2011 botnet controllers will resort to employing steganography techniques to control their computers, by hiding their commands in plain view perhaps within images or music files distributed through file sharing or social networking webites.
This approach will allow criminals to surreptitiously issue instructions to their botnets without relying on an ISP to host their infrastructure thus minimising the chances of discovery.
Although 2010 has experienced fluctuation in the number of botnets and their associated output, the top three botnets have not changed in the latter half of 2010.
Rustock remains the most dominant botnet, with its spam output having more than doubled since last year to over 44 billion spam emails per day and more than one million bots under its control while Grum and Cutwail are the second and third largest respectively.
Cutwail and Grum have also been responsible for an increase in the volume of malware being sent in spam from botnets.
The report revealed that one noteworthy security threat in 2010 was the ‘Here You Have’ virus which on September 9, 2010 used old mass-mailer techniques to send malicious emails, peaking at 2,000 emails blocked per minute.
For 2010, the average number of new malicious websites blocked each day rose to 3,066 compared to 2,465 for 2009, an increase of 24.3%.
The annual average global spam rate was 89.1%, an increase of 1.4% on the 2009.
Further, in August, the global spam rate peaked at 92.2% when the proportion of spam sent from botnets rose to 95% as a new variant of the Rustock botnet was seeded and quickly put to use.
The average rate for malware contained in email traffic in 2010 was 1 in 284.2 emails (0.352%) almost unchanged when compared with 1 in 286.4 (0.349%) for 2009.
There were 339.673 different malware strains identified in the malicious emails blocked, which represents more than a hundred-fold increase over 2009 and is due to growth in polymorphic malware variants.
In 2010, the average ratio of email traffic blocked as phishing attacks was 1 in 444.5 (0.23%), compared with 1 in 325.2 (0.31%) in 2009 and approximately 95.1 billion phishing emails were projected to be in circulation in 2010. | <urn:uuid:d7447e9b-2fc5-4313-929b-f7cf73bc9621> | CC-MAIN-2022-40 | https://techmonitor.ai/technology/software/spam-from-botnets-were-882-of-all-spam-in-2010-symantec-081210 | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335573.50/warc/CC-MAIN-20221001070422-20221001100422-00327.warc.gz | en | 0.965716 | 691 | 2.53125 | 3 |
What is a Duress Alarm?
A duress alarm is a kind of mobile device used to keep track of people. These alarms are not targeted towards buildings or objects, but rather focus on the safety of an individual. They can be used to track people’s whereabouts, check up their well-being, or used as an alert when in dangerous situations.
Duress alarm systems consist of wireless mobile devices that keep people in touch with others. These devices generally have a very simple design to serve a straightforward purpose. While different alarms feature unique buttons or layouts, the main idea is to have an emergency button to press in stressful situations.
This will then alert whoever is on the other end of the alarm and they can get help. Duress alarms can also be used to track the whereabouts of people and allow them to check in regularly. Different workers, and people who are often in potentially risky situations, will carry around a duress alarm to help keep track of their whereabouts for safety purposes.
What are the Different Types of Duress Alarms
Duress alarms vary according to the kind of situation they will be used in. In the most common and simple form, a duress alarm features an emergency button. This button can be hit in times of need to alert responders on the other end of the alarm. Responders can then use the duress alarm to locate the person in danger and then help them.
Other duress alarms feature a two-way system. This means that those with the alarm can check in with people on the other end, and visa-versa. This is often used in situations where workers will need to check in with headquarters while out on the job.
Two other features found on duress alarms are a “no response” feature and a “man down” button. No response alerts will occur when the person equipped with the alarm doesn’t respond in a certain period of time. This will then alert the headquarters of their absence. A man down feature will alert the other end of the alarm if someone is unconscious or incapacitated. This will signal for help.
How Does a Duress Button Work
A duress alarm will generally be activated via a button. This button will be linked to a monitoring center that will provide a duress alert when someone presses the button. These buttons vary for each device.
You can activate the alarm manually, automatically (common in a man down alert), silently, or audibly. As soon as the button is activated, it sends a signal to the monitoring center or headquarters. This notifies them of the location of the alarm and what kind of situation it is.
When Should You Use a Duress Alarm
Duress alarms are utilized across many different industries and scenarios. They are most commonly found as panic or emergency buttons. Workers can use a duress alarm in any environment that poses potential threats or dangers.
Duress alarms are commonly found in hospitals, nursing homes, and medical facilities. This is to ensure that patients can get immediate help if something were to happen to them while by themselves.
You will also find these systems in industrial and factory environments, municipal buildings, and government facilities. Security guards that patrol regular routes may also use duress alarms to keep people in touch with their location and alert their fellow guards if they get into a bad situation.
Duress alarms are very useful to have in any situation where there is potential harm to an individual as they can alert other people and get help as soon as possible. | <urn:uuid:4b47aa3f-1174-4050-9a65-854eb28776fb> | CC-MAIN-2022-40 | https://www.getkisi.com/guides/duress-alarm | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337853.66/warc/CC-MAIN-20221006155805-20221006185805-00327.warc.gz | en | 0.950551 | 718 | 2.75 | 3 |
According to the CDC, 90 percent of the United States’ $3.8 trillion annual healthcare expenditures are for people with chronic and mental health conditions. Where prevention isn’t possible, then managing chronic diseases to prevent hospital intervention has the potential to reduce healthcare costs.
Examining Medicare claims alone, the National Center for Biotechnology Information estimates that in-patient surgical care accounts for nearly 50 percent of hospital expenditures and 30 percent total overall healthcare costs.
The Internet of Things (IoT) has the potential to reduce healthcare costs by essentially decentralizing healthcare and giving patients the devices and monitoring to recover and/or manage healthcare conditions wherever they may be, which can lessen the strain on healthcare costs and capacity. This is enabled through remote patient monitoring.
Remote patient monitoring (RPM) is an IoT-enabled technology that pairs medical-grade devices with secure communications and platforms to connect medical providers with patients. The medical data collected by the patient is distributed to the cloud and stored on a patient’s electronic medical record, which can be viewed by medical professionals.
Remote patient monitoring is growing in areas particularly related to chronic disease care including:
Congestive heart failure: Provides patients with pre-programmed devices to record patient vitals. This is applicable in both disease and post-operative care to help reduce time in and trips to the hospital.
Diabetes: Whether through a continuous glucose monitor or digital devices that read finger prick results, patient self-advocacy and autonomy in diabetes management is a growing trend enabled through IoT technology.
Chronic Obstructive Pulmonary Disease (COPD): Management of COPD through remote patient monitoring enables patients to measure weight, blood pressure, respirations, and oximetry through connected devices. This technology is also available for recovering pneumonia patients for care outside of a traditional hospital setting.
Hypertension: Through both medication adherence and blood pressure readings, management of hypertension can be more proactive and comprehensive through remote patient monitoring.
The benefits to the patient in remote patient monitoring are having greater access to healthcare, more autonomy and self-advocacy, and an approach to comprehensive care from the home.
For the provider, revenue generation is possible by decentralizing practices. Logistical issues associated with in-office visits and appointments can be lessened by monitoring patient readings from a unified platform. This can also help practices scale. While the need for more practitioners may arise from increasing patient load, the need for physical infrastructure will be less if a percentage of patients can be monitored remotely.
Digital healthcare is on the rise, whether it’s remote patient monitoring or telehealth for RPM, it’s an area on a significant upswing. According to a report from Mercom Capital Group, funding activity for digital health companies was up by 138 percent in the first half of 2021. This is a fast-moving and competitive market space.
But it’s also one with many hurdles and complexities. One of which is secure communications between patients and the provider. KORE Connected Health Telemetry Solutions is a bundled suite of services that helps connect medical devices to the cloud and back securely using encryption keys.
Learn more about how KORE Connected Health makes launching remote patient monitoring solutions simpler.
Check out our recent eBook, "Remote Patient Monitoring: A Guide to Connected Health Telemetry".
KORE keeps you up to date on all things IoT.
Stay up to date on all things IoT by signing up for email notifications. | <urn:uuid:4eeb4d7e-50ef-4bc7-b63d-a5d869ea47e5> | CC-MAIN-2022-40 | https://www.korewireless.com/news/from-chronic-to-post-operative-care-iot-can-help-reduce-hospital-costs | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337853.66/warc/CC-MAIN-20221006155805-20221006185805-00327.warc.gz | en | 0.925409 | 712 | 2.734375 | 3 |
Starlink is a satellite Internet constellation constructed by SpaceX to provide satellite Internet access. The constellation will consist of thousands of mass-produced small satellites in low Earth orbit, which communicate with designated ground receivers. Starlink is headed by tech wizard Elon Musk, aimed to deliver “high-speed broadband internet to locations where access has been unreliable or completely unavailable”. It currently only operates for homes/apartments and isn’t mobile, but that’s expected to come in the near future.
Starlink has 60 satellites currently in orbit, with SpaceX indicating that it needs 420 satellites in the constellation to achieve minor broadband coverage of Earth, 780 to provide moderate coverage, and 1,600 to provide solid coverage all across the globe.
What does this mean for an SMB?
While Starlink is a novel idea and may or may not be available in your area just yet, there are other options that have been around for some time including Hughes Net satellite Internet services.
The primary disadvantage to these space-based solutions is the time it takes for packets to be sent and received. These are not designed for Gaming that requires low latency communications or high-frequency trading. On the other hand, they provide a robust method of access to the Internet for remote locations.
Start building your robust, defense-in-depth cybersecurity plan today with CyberHoot.
To learn more about Starlink, watch this short 5-minute video:
CyberHoot does have some other resources available for your use. Below are links to all of our resources, feel free to check them out whenever you like:
- Cybrary (Cyber Library)
- Press Releases
- Instructional Videos (HowTo) – very helpful for our SuperUsers!
Note: If you’d like to subscribe to our newsletter, visit any link above (besides infographics) and enter your email address on the right-hand side of the page, and click ‘Send Me Newsletters’. | <urn:uuid:735db1a5-58e5-4672-b6f2-8e2f9669d108> | CC-MAIN-2022-40 | https://cyberhoot.com/cybrary/starlink/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335004.95/warc/CC-MAIN-20220927100008-20220927130008-00527.warc.gz | en | 0.933564 | 422 | 2.9375 | 3 |
Hospitals are filled with machines connected to the internet. With a combination of both wired and wireless connectivity, knowing and managing which devices are connected has become more complicated and, consequently, the institutions’ attack surface has expanded.
When did these devices get smart?
A brief timeline shows the FDA didn’t start regulating the connectivity of devices until 2005, but medical devices started to leverage software back in the ‘80s. Clinical capabilities have benefited greatly from this digitalization, bringing features, data collection and analytic computing to clinical care. Some devices that have been digitized include pacemakers, infusion pumps, ventilators, CT and MRI scanners, all of which (as a result) contain patient information and have some level of connectivity. Walk into a healthcare conference today and you’ll be hard-pressed to find devices that don’t offer connectivity via wires, Bluetooth, and/or wirelessly.
Wearable devices and at-home medical devices are also becoming increasingly common. The ability for a device to transmit vital sign data from a patient’s home to hospital staff has encouraged the expansion of the telehealth industry. It has also made possible for the emergency medical community to respond to device alerts. In some cases, health insurance companies use data from fitness trackers to incentivize medical expense management and “wellness” promotions.
Information exposure and theft
There was a time when medical devices relied on physical security to limit who could update a device. But to enhance clinical experience, many of these devices have since been retrofitted so they can be networked and managed remotely by both provider and vendor.
These devices often carry patient personal information (such as Social Security numbers), health insurance information, contact information and information about health conditions. Connectivity and inevitable software vulnerabilities mean that this data can potentially be exposed.
There are some predictable schemes for obtaining a person’s SSN – insurance claims, tax filings, rebate claims, bank loan documents. More healthcare-specific is the idea of a deceased patient’s SSN being used to run a scheme, as there tends to be less monitoring of financial activity after someone has died. There are also those who use insurance and contact information to claim prescriptions or run phishing schemes on an aging population. The combination of patient health factors and geographic location can sometimes also allow scammers to pinpoint a person’s identity and discover other personal information that can be of use.
Understanding devices in the field
When a connected medical device is procured by an HDO, the terms for ongoing support are a critical component of the negotiation. This often includes medical device manufacturers (MDMs) supporting device bug resolution, patching for known vulnerabilities, and enhancements to security over the warrantied lifetime of the device.
However, there is no mandate to remove a device that’s past vendor warranty from operation. With payers influencing HDO procurement strategies, devices that “still work” can be difficult to throw out, especially when a cybersecurity vulnerability is “theoretical.”
Imagine a vulnerability is identified on a single device that is no longer under warranty. This means the vendor no longer provides software patches. This vulnerability may be exploited to other installations of the same device. Devices that are no longer receiving updates for known vulnerabilities are an exponential threat for hackers looking for an entry point into critical healthcare data.
An additional consideration is the development practices for medical devices. Many MDMs develop their software on commercial operating systems such as Windows. Software is phased out all the time – it’s part of the development life cycle. But, for example, the end of Windows 7 support in 2020 means medical devices in the field that run Windows 7 will become more vulnerable with each passing day. Every virus or malware attempt will no longer face Microsoft’s security capabilities and improvements. These devices and HDOs will have to fend for themselves.
Setting aside the data available on a device, there is also the possibility of attackers using devices as a gateway into an HDO’s network. Due to budgeting decisions and the organizations’ preference for clinical investments, hospitals IT departments often work with limited resources. In some instances, the limited allocation of resources towards recovery procedures has made HDOs especially susceptible to ransomware attacks.
Some have suggested that a hospital should revert to emergency protocols (i.e. pencil and paper mode) to operate during a cyber attack, as occurred when parts of the NHS were shut down due to WannaCry. This can limit the impact of attacks on elective procedures, but what about patients with urgent needs?
Research shows a 13.3% higher mortality rate for cardiac arrest patients who experienced a four-minutes delay in care. And a delay in care due to a network takeover by hackers is likely to be more than four minutes.
Even in the wake of multiple HDOs implementing better security practices after an attack, there is evidence of negative outcomes for patients in facilities with a historic breach. The 0.04% increase in mortality rate observed is the equivalent of the 0.04% increase in positive outcomes for patients based on enhanced treatments.
What happens next?
The FDA draft premarket cybersecurity guidance from October 2018 recommends incorporating the NIST Cybersecurity Framework (NIST-CSF). NIST-CSF includes a combination of both technical and procedural interventions into both the design and support of devices. While there is no risk rating associated with the NIST-CSF sub-categories, the technical sub-categories tend to require more effort and technical sophistication to implement.
However, there is no need for healthcare to go at it alone – we can learn from other industries. We have seen the financial services industry, often perceived as a cybersecurity leader, manage cyber threats through leveraging tools to implement and maintain security over time. The migration away from building personalized data centers to using commercially available cloud-based service providers is a prime example of this. There have been numerous case studies showing how cloud hosting enhances security responses (especially redundancy & availability), expedites product development and reduces maintenance cost over the lifetime of a product.
As medical device manufacturers develop new products and update products currently in the field, using relevant tools to address the FDA premarket guidance and incorporating industry leading best practices is surely the most sustainable and scalable approach. | <urn:uuid:64da0120-ea8e-4c75-a3a5-d6f6263305e7> | CC-MAIN-2022-40 | https://www.helpnetsecurity.com/2019/04/30/cybersecurity-for-medical-devices/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335004.95/warc/CC-MAIN-20220927100008-20220927130008-00527.warc.gz | en | 0.94448 | 1,294 | 3.203125 | 3 |
Energy consumption accounts for a large portion of datacenter operating budgets. The added environmental impact compounded with capital costs has motivated operators to increase the efficiency and lower the cost of computing. Last year, Facebook spun up the Open Compute Project, with the goal of improving datacenter designs to meet those goals.
So far, the project has produced a number of components including a server chassis, a battery cabinet and two x86 motherboards. All of the designs are freely available to download from their website.
To rate efficiency, datacenters calculate their power usage effectiveness (PUE) ratio. The rating compares how much incoming power to energy used for computational processes. One estimate puts the average datacenter PUE at 1.8. Larger Internet players have used some unique solutions to bring that number closer to the 1.0 mark, with Google currently averaging a 1.14 PUE.
Prineville, Oregon is home to one of the world’s most efficient datacenters. Built by Facebook, the center utilizes designs from the Open Compute Project as well as innovative LED lighting and gray water facilities to receive a 1.07 PUE rating at full load.
Ken Pratchett, Manager of the Prineville data center, discussed the cost and efficiency of the Open Compute Servers, “These machines are 38 percent more efficient than another machine that you could find on the open market,” he said. “In fact, they cost 24 percent less to create.”
The combination of green technologies implemented at the Oregon datacenter earned it a Gold Certification from the US Green Building Council’s Leadership in Energy and Environmental Design (LEED). Compared to other datacenters built to code, it consumed 52 percent less energy and 72 percent less water for occupant use. It also recycles captured water for landscape irrigation.
Next month, the Open Compute Project will hold a summit in San Antonio. The two-day event will include the following workshops:
- Mechanical design and modular power distribution
- Defining open systems management for the enterprise
- New initiatives; building for different geographies
- New hardware and building for the 100 year standard
- Pushing the limits of connectivity, software and hardware modularity
The consumer-driven demand for computer technology has spurred the need for more and larger datacenters, resulting in increased energy consumption and higher operational costs. Examples like the one in Prineville not only lead to innovation in the field of green datacenter technology, but also provide a framework for cost-effective operations as well. | <urn:uuid:3b325b4d-d366-44a4-bd4f-27a18e8e1b58> | CC-MAIN-2022-40 | https://www.hpcwire.com/2012/04/26/facebook_showcases_green_datacenter/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335609.53/warc/CC-MAIN-20221001101652-20221001131652-00527.warc.gz | en | 0.926628 | 533 | 2.71875 | 3 |
There are various situations where you would want to optimize the performance of your route-reflector in BGP network. By default BGP downloads all the routes in RIB. BGP Table-Map is one of the methods to achieve BGP selective route download particularly used on the Route-Reflectors.
BGP Table-map is primary used to suppress the unnecessary downloading of certain BGP routes to the RIB or Forwarding Information Base (FIB) on a dedicated route reflector, which propagates BGP updates without carrying transit traffic.
Suppressing/Filtering the routes in RIB/FIB on the dedicated route-reflectors provides essential saving on the memory and CPU resources on the router.
Note that the Selective Route Download feature is not applicable to Multiprotocol Label Switching (MPLS) Layer 3 VPN because the route download is already automatically suppressed on a route reflector.
The table-map command is configured under route bgp process and is applied on the router as a whole and not on per neighbor basis as route-maps.
It is important to understand the use of the filter keyword in the table-map command:
- When the table-map command is used without the filter keyword, the route map referenced in the table-map command is used to set certain properties (such as the traffic index) of the routes for installation into the RIB. The route is always downloaded, regardless of whether it is permitted or denied by the route map.
- When the table-map command is used with the filter keyword, the route map referenced is also used to control whether a BGP route is to be downloaded to the RIB (hence the filter). A BGP route is not downloaded to the RIB if it is denied by the route map.
We can create a route-map to permit or deny the routes to be downloaded in the RIB or FIB and then use this route-map with the table-map command.
Related- BGP Suppress MAP and Attribute MAP
LET’S USE THE BELOW TOPOLOGY TO UNDERSTAND THE USE OF TABLE-MAPS –
In the topology diagram above R1 is advertising a route of its loopback 1 i.e. 220.127.116.11 to Route-Reflector. We have applied a route-map on Route-Reflector to deny everything and then this route-map is called in the table-map command so avoid the installation of the route in the RIB and FIB on the RR.
Though RR doesn’t install the route in its RIB & FIB it still reflects the route to R2.
SCENARIO 1: TABLE-MAP ISN’T APPLIED
As of now we are able to see the route of 18.104.22.168 on R3 in RIB and FIB.
R3#sh ip route
Route for 22.214.171.124 is also received on R2 from R3.
R2#sh ip route 126.96.36.199
SCENARIO 2: TABLE-MAP IS APPLIED UNDER ROUTER BGP
Now we have configured table-map to deny the suppress route for 188.8.131.52 in RIB and FIB.
R1 sends the route to R3.
R1#sh ip bgp neighbors 184.108.40.206 advertised-routes
R3 doesn’t install the route in its RIB & FIB as the route-map tmap is configured to deny all the routes.
Note that under the route-map we can allow or deny the routes if required.
R3#sh ip route 220.127.116.11
R3#sh ip cef 18.104.22.168
R2#sh ip route 22.214.171.124 Related- Non-Exist-Map & BGP Advertise-Map
R2#sh ip route 126.96.36.199
Related- Non-Exist-Map & BGP Advertise-Map | <urn:uuid:2dcecb2f-7244-4700-945c-b9f83e6a5056> | CC-MAIN-2022-40 | https://ipwithease.com/understand-and-configure-bgp-table-map/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337339.70/warc/CC-MAIN-20221002181356-20221002211356-00527.warc.gz | en | 0.878662 | 881 | 2.640625 | 3 |
A virus is a software program designed to make its way onto your computer with the intention of causing potentially irreparable damage to its hard drive and files. Viruses are not the only danger to your computer but they are the most commonly known form. Viruses are a type of malware, or malicious software which is designed specifically to download onto your device, whether it is your desktop PC, smartphone, or tablet in order to manipulate or damage them.
Furthermore, malware which has been downloaded onto your device can also record and steal your information such as credit card account details. Many people are familiar with the most common forms of malware including phishing scams, adware, spyware, viruses and worms. But here we will explore other hacking methods that are less well known but still result in users obtaining viruses, malware or the risk of being hacked.
Social Media Scams and Malware
Something we have all experienced on social media is clickbait and this has become a popular business in recent years. Clickbait is defined as “content whose main purpose is to attract attention and encourage visitors to click on a link to a particular web page”.
When someone posts a link on social media such as Facebook or Twitter, they will include a headline to entice users to click on the link to see more, without really telling them what site they will be taken to.
Here users are often asked to enter their personal contact information which could be collected and sold to cybercriminals. The most obvious scenario is when new TV shows become the focal point of social media, various platforms will become heavily populated with posts asking you to click a link to get a preview of the latest episode. This is a classic example of a ploy which hackers use to persuade individuals to click on strange links and thus, downloading unknown files.
Exploit kits are a type of toolkit that is used with the intent to exploiting any security holes found in software applications, simply for the purpose of distributing malware. These kits target users by running insecure or outdated software applications on their computers and in 2015, Yahoo fell victim to this as they hosted malicious ads that redirected users to websites hosting these kits.
Ransomware is a topic most people associate with laptops and computers but with the increase in the number of smartphones that people use today, this has given cyber criminals a window of opportunity to attack users whilst on the go.
Again, mobile ransomware heavily exploits the use of social media sites to trick people into downloading what they think is innocent content. Once downloaded, the ransomware then encrypts files or locks the device and demands a ransom payment to unencrypt the files or to unlock the device. People who fall victim to this are likely to respond by making the payment as many of us store content that is essential to our personal and professional lives.
Online Gaming Malware Attacks
Online gaming is becoming an increasingly popular market for cybercriminals. Dedicated gamers tend to spend quite some time building up characters and in the world of gaming, these can be sold at a very attractive price. Gaming and gamers are therefore the perfect target and are very vulnerable to concept of malware. If a user’s account is hacked,cybercriminals can steal player credentials and then gain access to their gaming account. This then allows them to transfer in-game assets to other accounts, and sell those assets on the grey market, which is an unauthorised, but not necessarily illegal place that is used to sell virtual items and currency for real money.
Browser Extension Adware and Malware
If you find your web browser is randomly being redirected to the “Ads by Browser Extension” advertisements when using the Internet, then it is possible that your computer has been infected with an adware program.
As a result, downloading free software “Ads by Browser Extension” is a form of adware which may appear as a result of the download and could install itself onto your computer without you wanting it there or – even without your knowledge. Common symptoms of adware being installed include an increase in advertising banners and pop ups being displayed, trying to trick the user into downloading fake updates or other software.
What can we learn from these methods?
Cybercriminals are always coming up with new and inventive ways to scam you and their methods can come in all shapes and sizes. Although systems to tackle these threats are being developed, cybercrime will always be a step ahead therefore its vital that we have the best understanding of the threats so we can put our knowledge into practice so we can stay safe and secure on a daily basis.
Request a demo of our phishing simulation software MetaPhish and target phishing and ransomware attacks directly to your staff to increase your employees’ sensitivity to these fraudulent emails. You can also contact us about our eLearning module on Essential Phishing Awareness to help your staff understand what to look out for in a phishing email. | <urn:uuid:9396d25a-779d-4496-b89b-df6f1748eb97> | CC-MAIN-2022-40 | https://www.metacompliance.com/fr/blog/security-awareness-training/5-hacking-methods-you-werent-even-aware-of | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337339.70/warc/CC-MAIN-20221002181356-20221002211356-00527.warc.gz | en | 0.948747 | 1,007 | 3.125 | 3 |
You may not realize it, but your car probably already has some self-driving technologies—even basic ones. For example, many of the newest cars have lane assist or park assist, which can help you avoid unintended lane violations or better ease into parking spots.
But those car assistance technologies depend on the internet and computers, and those are at risk of hacking. In fact, that’s one of the biggest concerns with the continued development of self-driving cars: how can thieves access them, and what will they do with the data? Vound Software compiled a detailed infographic to get you all the details.
This infographic originally appeared on Vound Software’s Intella Blog | <urn:uuid:ec70d3f7-81a0-4bb8-901c-39935cb30e6d> | CC-MAIN-2022-40 | https://dataconomy.com/2017/01/infographic-car-hacking-not-thing-future/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337473.26/warc/CC-MAIN-20221004023206-20221004053206-00527.warc.gz | en | 0.951306 | 139 | 2.6875 | 3 |
Last week in our special report series, we explored three external heat rejection systems. This week, in our final article in the series on sustainable data center cooling challenges, we’ll discuss three more external heat rejection systems.
Carbon Usage Effectiveness
Carbon Usage Effectiveness (CUE) is another essential sustainability metric directly tied to the source of energy generation.
It is expressed as CUE Source – Based on the annualized carbon emitted by the energy source (power generation), expressed as kilograms of carbon emitted per kWh of IT Energy generated. (kg/IT kWh).
The EPA’s Emissions & Generation Resource Integrated Database (eGRID) website has a tool that allows users to enter their zip codes (or select a region) to view their power profiles.
Free cooling is typically referred to as any type of cooling system which reduces or eliminates the need for mechanical cooling (i.e., compressor based). This can include direct or indirect air-side economizers, as well as water-side economizers. The water-side economizer is generally based on a heat exchanger which allows cooling tower water to reduce the load on water-cooled chillers during cooler weather. All of these systems generally save a percentage of mechanical cooling energy during cooler weather. Overall, free cooling is more effective in colder climates, for both air-side and water- side economizers. There are many variations and combinations of methodologies that can be combined to maximize energy effectiveness and optimize water usage over the seasons.
While water in itself is not energy, it requires energy to process and deliver clean water. This is often overlooked or ignored. However, as water shortages increase, this is being addressed in the California Energy Code: Codes and Standards Enhancement (CASE) Initiative 2022 – “Title 24 2022 Nonresidential Computer Room Efficiency” (CA-Title 24), as “Embedded Electricity in Water.”
Alternately, waste heat can also be rejected into bodies of water, such as lakes and rivers or even the ocean. In many cases, the temperature ranges of those bodies of water are such that they can be used all year round to cool data centers without the need for mechanical cooling, a significant energy savings. These savings can also reduce or eliminate consumption of source water, as well as reducing fossil fuels used in non-renewable power generation.
Energy Reuse Effectiveness
The total waste heat generated by an air-cooled or a liquid-cooled data center is virtually the same for a given amount of total energy consumed. Clearly, a lower PUE will reduce the total energy consumed for a given IT load. While end-to-end thermal management typically refers to “chip-to-atmosphere” heat rejection, it is still just waste heat. One of the long-term sustainability goals is being able to recover and reuse waste heat.
The Green Grid introduced the energy reuse effectiveness (ERE) metric in 2011, energy reuse factor (ERF), with relatively little impact. This was primarily because waste heat from air-cooled ITE being rejected is more difficult to recover effectively.
Liquid Cooling provides a pathway to more effective opportunities for energy reuse and recovery. Its higher fluid operating and return temperatures improve the ability to recover a portion of the waste heat energy. However the challenges are multi- dimensional, but as time progresses the interest, technology improvements and cost effectiveness will continue to drive this initiative. While higher power densities can create cooling challenges, it also offers multiple benefits and opportunities to improve overall energy efficiency, IT performance, and sustainability. However, the challenges are multi-dimensional, but the interest, technology improvements, and cost-effectiveness will continue to drive this initiative as time progresses.
Download the full report, “Sustainably Meeting High Density Cooling Challenges: When, Where, and How,” courtesy of Nautilus Data Technologies for exclusive look at the Nautilus water-cooled data center. | <urn:uuid:635de887-8f5c-4fcb-acc0-d58bd1ee18cf> | CC-MAIN-2022-40 | https://datacenterfrontier.com/sustainable-data-center-cooling-cue-free-cooling-and-ere/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337855.83/warc/CC-MAIN-20221006191305-20221006221305-00527.warc.gz | en | 0.930642 | 819 | 3.328125 | 3 |
How Do You Defend the DNS Infrastructure?
What is DNS?
The DNS (Domain Name System) protocol is a critical part of the Internet’s control plane, providing name resolution functionality and supporting other capabilities that are taken for granted such as load-balancing and internet traffic management. The availability of DNS is key for anyone providing services or content across the Internet. If the DNS infrastructure is unavailable or slow, the user experience would be impacted to the point of receiving a “no internet service” error message. | <urn:uuid:43c86c66-b73c-4c23-a6ff-14ca19dc0808> | CC-MAIN-2022-40 | https://www.netscout.com/video/how-do-you-defend-the-dns-infrastructure | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337855.83/warc/CC-MAIN-20221006191305-20221006221305-00527.warc.gz | en | 0.881433 | 107 | 3.0625 | 3 |
How Smart Grid Technology Is Driving Renewable Energy
In April 2019, the amount of energy generated from renewable resources such as wind, solar and hydropower for the first time exceeded the total energy churned out by coal-powered electric plants, according to the Electric Power Monthly, published by the U.S. Energy Information Administration. It was a landmark event, although renewable’s superiority was expected to be short-lived once seasonal adjustments such as the increase in hydropower dip. Still, it’s hard to overlook the ascendancy of renewable electric power generation.
Renewable energy, whether produced by regional utilities or self-contained enterprises, is thriving, and IoT infrastructure and related technologies can claim a considerable amount of credit for its rise. IoT enables effective management of these volatile and often widely dispersed electricity-producing resources through smart grid technology.
“With IoT, utilities can collect data and better manage their grid by detecting and restoring service outages quickly, and even create the concept of a ‘self-healing’ grid,” wrote Warren Westrup, director of business development, utilities, for Sierra Wireless, in a recent blog post, Powering up the Energy IoT with LTE Wireless Connectivity.
IoT Facilitates Renewable Energy Management and Growth
The Internet of Things can be even more than merely a good management enabler, according to Olivier Amiot, director of marketing, energy for Sierra Wireless. In a blog post, How the IoT Can Speed the World’s Transition to a Clean Energy Economy, he noted that “digital technologies – in particular, the IoT – can help us address the climate change challenge by accelerating the transition to a more sustainable, renewable-energy-powered economy.”
IoT infrastructure’s role as an indispensable component of a renewable energy grid is true for both privately maintained microgrid environments and those supported by regional utilities. This fact is especially true for wind farms, which rely on complex mechanical gear that requires each node to support dozens of or more sensors. Those sensors not only measure the performance of various node-level components but also take stock of current climate and environmental conditions to provide ample context for data analysis.
In many ways, the ability to collect so much data so quickly is breaking new ground in energy management.
“You’ve got this story of this invisible, dangerous commodity that travels at the speed of light that we call electricity and for the last hundred-plus years most people could interact with it in only the most rudimentary ways,” noted Mark Feasel, vice president of smart grid for Schneider Electric. Feasel added that with the current array of sensing devices “we’re capable now of [deploying] some devices that can capture information on electricity 60,000 times a second.”
Still Some Missing Links
The amount of electricity produced privately by both large businesses and other institutions such as universities and medical centers presents new challenges for both of those camps. These distributed energy resources — or DERs — must interconnect to ensure that power is produced efficiently and delivered promptly. The onus is clearly on the utilities, which absorb excess energy from DERs and provide a backstop for these institutions when their resources are not cranking at full throttle. | <urn:uuid:4415675a-a43e-4969-8225-5adbc5204561> | CC-MAIN-2022-40 | https://www.iotworldtoday.com/2019/08/29/how-smart-grid-technology-is-driving-renewable-energy/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030333541.98/warc/CC-MAIN-20220924213650-20220925003650-00727.warc.gz | en | 0.93619 | 673 | 2.875 | 3 |
Published: 29 August 2017
Negative publicity and societal debate around the intended and unintended ethical consequences of using digital technology requires CIOs and their IT leaders, chief data and digital officers, board members, risk officers and investors to make digital ethics an integral part of any digital business.
Included in Full Research
- What Is Digital Ethics?
- Why Are Digital Ethics Important?
- What Makes Digital Ethics Special?
- Difference From General Ethics
- Difference From Business Ethics
- Who Is Responsible for Defining Digital Ethics?
- How Does "Care Ethics" Support the Digital World?
- What Is the Scope of Digital Ethics?
- Social Interaction
- Free Will
- Society and Economy | <urn:uuid:a30cf36f-20c0-479d-a9c3-83585f82a60f> | CC-MAIN-2022-40 | https://www.gartner.com/en/documents/3794164 | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334802.16/warc/CC-MAIN-20220926051040-20220926081040-00727.warc.gz | en | 0.862106 | 163 | 2.578125 | 3 |
With the advent of artificial intelligence, big data and 5G era, face recognition technology has gradually penetrated into various fields of production and life. Bring great convenience to people's lives.
What is a face recognition system
Face recognition is a kind of biometrics recognition technology based on human facial feature information. A series of related technologies that use cameras or cameras to collect images or video streams containing human faces, and automatically detect and track human faces in the images, and then recognize the detected faces, are also commonly called face recognition and facial recognition.
Face recognition is mainly used for identification. Due to the rapid popularization of video surveillance, many video surveillance applications urgently need a rapid identification technology in a long-distance, non-cooperative state of users, in order to quickly confirm the identity of personnel at a long distance and realize intelligent early warning. Face recognition technology
It is the best choice. The fast face detection technology can be used to find faces in real-time from the surveillance video and perform real-time comparison with the face database to achieve fast identity recognition.
Face recognition system advantages
Non-contact: The process of face verification does not need to be in contact with the device or terminal, which is completely different from fingerprints and punch-in methods. The experience is strong, the device should not be touched frequently, with high stability and low failure rate. Naturalness: The first perception of the characteristics of human recognition objects is to pass back to the brain through the senses of the eyes. Face recognition technology is the process of using the computer to deeply learn the human sensory back to the brain. And experience is more natural, and does not require the characteristics of other objects, and there will be no problems such as loss and replication.
Fast recognition speed and high accuracy: Due to the development of computer technology, face recognition speed can be increased to 0.3 seconds per person under the fixed sample library, and the behavior of changing the face at different angles, wearing glasses and makeup can also be adjusted Face recognition algorithm to correct the comparison data.
Faces have become an important data source in the era of big data: through face recognition and comparison in a large number of regions, using big data analysis, mining and other technologies, you can fully grasp the behavior trajectories of specific objects or the range of activities of specific groups of people. Manual intervention is completely handled by cloud computing and big data systems.
From the point of view of the payment industry, as a new thing, "face-swapping" verification and "face-swapping" payment not only bring convenience to people's lives, but also enhance security. As a simple example, the password can be changed, but the face and fingerprints cannot be changed, so it is relatively safer.
At present, more than 40 cities have opened "face-washing government affairs" to check provident funds, pay traffic tickets, declare a tax, receive ID cards, electronic social security cards and other electronic documents. People no longer need to use a pile of materials to prove that I am me . It turned out that I had to be present, but now I just brush my face and blink.
Face recognition technology application
Access control system: The application of face recognition access control, changing the previous access mode, effectively preventing strangers from entering and leaving the community at will, reducing the frequency of community security accidents as much as possible, improving the residents' life experience, promoting the technology of community services, and creating new properties Management mode to achieve intelligent management.
Office system: As smart buildings become a hot topic in the industry, smart offices are also heating up. With the rapid development of Internet technology, the networking of intelligent office attendance system has become an inevitable trend. Intelligent office uses face recognition cloud platform software, face device mobile terminal, face recognition visitor management system, face recognition access control attendance system, office control attendance system, face recognition employee recognition attendance system and other recognition systems to achieve intelligence Office and intelligent management, professionally create a more scientific space and humanized environment, thereby improving office efficiency and office quality, improving office conditions, reducing labor intensity, and realizing scientific management and decision-making, thereby reducing working hours, Provides work efficiency.
Surveillance system: Face recognition video surveillance has profoundly transformed the entire security industry in the form of intelligent hardware, big data mining and the Internet of Things. From the original "high-definition, clear and visible" of the high-definition surveillance system to the "find fast and accurate" based on 5G technology big data. In a smart city, the rapid growth of video image data and the constantly emerging user demand indicate that the demand for video big data is becoming stronger and stronger. Video big data will build smarter systems for users and provide more valuable services.
Payment system: Compared with QR code payment, face-to-face payment is more convenient, eliminating the need for consumers to take out the mobile phone scan code when paying, and only need to face the camera to brush their face when paying. Many merchants said that “face-paying” not only improves employee efficiency, but also saves queue time
With the development of real-name authentication and security technology, face-to-face technology also makes real-name authentication easier. Alipay's face recognition technology has reached financial-level accuracy and security. Combined with multi-factor verification such as eye patterns, the accuracy rate can reach 99.99%.
Public security system: Continuously collect video through the front end of the video surveillance, compare the "concerned" personnel with the face recognition technology as the core in the back end, and notify the on-site police officers to carry out target arrest through the alarm method. For some personnel-intensive areas, such as key entrances and exits and passageways of stations, subway stations, airports, and communities, the personnel bayonet is arranged, and the back-end blacklists the key personnel concerned and the prevention and control personnel through real-time video logistics. Matching control blacklist to realize face matching recognition.
Huifan Technology is a technology-based enterprise dedicated to the construction, service and operation of smart communities. It provides customers with the best Internet solutions through independent innovation and comprehensive integration services. Face recognition technology is also applied to smart access control in the community It brings great convenience to the residents of the community, facilitates property management, and improves the quality of life of the owners. At present, it has reached cooperation with many property companies to actively promote the construction of smart communities. | <urn:uuid:bb99e765-2cc1-4a0c-b838-28a660d3dec4> | CC-MAIN-2022-40 | https://www.hfteco.com/news/Hardcore-science-face-recognition-knowledge.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334802.16/warc/CC-MAIN-20220926051040-20220926081040-00727.warc.gz | en | 0.924073 | 1,326 | 3.140625 | 3 |
1. Robots Will Replace Humans
The first and most significant misconception is that RPA exists to eliminate the jobs of humans. RPA is excellent in handling simple, repetitive low-value tasks. If the work you do consists of these simple repetitive tasks, then your concerns may be realistic, but they are not simply due to the existence of RPA. Low-value repetitive tasks can equally be replaced by another software tool, by an intern, or by a low-cost outsourcing company. The fact is that the world is changing rapidly, and most organizations are required to focus on higher value-added tasks. To draw a simple analogy, the existence of printers, calculators, machines, the internet, etc. did not destroy any job environment, but the introduction of these technologies transformed the way people executed their jobs.
The rapid rise of RPA technology will likely mean that many jobs will be partially, or fully automated, freeing human employees from repetitive and monotonous tasks, so they can invest their time and energy in tasks that require creative thinking, intellectual judgment, or social skills (e.g. sales, marketing, strategy, solutions). In fact, the World Economic Forum Report predicts the “robot revolution” will actually create 58 million new jobs, with an emphasis on reskilling and upskilling employees. Moreover, according to a Forbes Insights report, 92% of companies saw improvement in employees’ satisfaction after deploying RPA to automate repetitive tasks.
Ultimately, the existence of RPA is transforming, rather than replacing roles. Human employees will be supplemented by the introduction of RPA technologies. The help of software robots in their daily task will enable them to be more productive, so more work can be done within the same period.
2. RPA Software Robots Are 100% Accurate
RPA software robots are only as accurate as how the robot developer programs them to be. RPA software robots are not equipped with cognitive capabilities, and they are designed to only execute a set of pre-configured instructions. If there are flaws in the set of instructions provided to the RPA software robot, they will still execute the instructions as configured, and replicate the same error until someone spots it.
Therefore, it is important for any robot developer to ensure processes are tested and optimized before the automation will go into production. Afterwards, the robot owner needs to keep monitoring the software robot to ensure no errors are coming up.
3. RPA Will Not Work In My Industry
Often there seems to be a belief that RPA is only applicable for certain industries, like finance and banking, for example. Yet the fact is that back-office activities must be carried out in all industries, which makes RPA relevant to most, if not all industries. Whenever there are repetitive, standardised, and numerous menial tasks to complete, RPA has the potential place to fill. RPA can be deployed for simple tasks like data entry, to complicated tasks like insurance claims processing, accounts payable/receivable etc.
Every organization, regardless of industry, will have back-office processes that need to be executed. Invoices need to be booked, payrolls need to be prepared, and reports need to be generated. In most organizations, there are still large groups of people that manually copy information from different places and consolidate this information in an Excel sheet. For any of these activities, RPA can provide an adequate solution.
4. RPA Is Expensive And Not Worth The Investment
RPA, like any other business initiative, requires an initial upfront investment in license fees and implementation costs. From the provisioning of IT infrastructure, the design of automated workflows and the time investment to get the organization up to speed, investments are required.
From our experience, investing in RPA is only expensive for organizations who only use limited use of technology. An organization that uses RPA to automate only one process, or only use RPA data entry, will struggle to see a good return on investment. On the other hand, when an RPA solution is properly set up and structured, it can improve efficiency, productivity and lower operational costs, making the return on investment far exceed the implementation costs.
Studies often find that RPA implementation provides superior productivity and efficiency, significant error reduction, operational cost reduction, service delivery improvement and employee satisfaction in a short period of time. There are plenty of case studies mentioned that organizations achieved a significant amount of ROI within the first few months of implementation.
5. RPA Is Mainly About Cost Saving
A last common misconception that we frequently hear is that RPA is solutions focus mostly on cost savings. Implementation of RPA into an organization brings much more benefits than they should have focused on, and cost saving is only the by-product after RPA is deployed. Let’s look into the benefits RPA could bring:
1. Increased operational efficiency
Software robots work faster than humans, and they can operate 24/7/365 without break. Therefore, they eliminate delays by eliminating the complacency creates by repetitive and monotonous tasks.
2. Increased employee productivity
Software robots eliminate the unnecessary workload on employees, enabling them to focus on their expertise and provide a value-added workforce to their organizations.
3. Elimination of errors
When humans process tasks that are tedious, repetitive and monotonous, the rate of making mistakes increases. While software robots are able to run continuously as per designed without error.
4. Improved Service Delivery
Whenever you are serving internal or external customers, their level of satisfaction is proportional to the accuracy, quality, and efficiency of services. Therefore, even when you automate a small part of the service you provide to your customers with RPA, you will instantly observe the increases in service quality and customer satisfaction.
5. Raised Employees Satisfaction
When RPA is deployed to process repetitive and monotonous tasks, and human employees focus on more satisfying high-value tasks, this raises employees’ satisfaction with their jobs and increases employee retention.
6. Increased compliance
With the availability of automated reports that document all software robot’s activities, you don’t have to worry about complying with internal or external regulations.
7. RPA Scalability
The scalability of RPA is tremendously critical to an organization, where they can scale up or down depending on the business needs with either minimal or no extra cost, while expecting the absolute consistency in performance.
This list shows that the benefits from RPA solutions go beyond just the financial motive, and that there are many additional reasons why RPA is a good idea. So next time you hear a common RPA misconception again, you will have the knowledge to address these points. | <urn:uuid:311073e9-1cec-4543-afc9-eac8fec0faeb> | CC-MAIN-2022-40 | https://www.cybiant.com/resources/robotic-process-automation-myths-and-misconceptions/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335444.58/warc/CC-MAIN-20220930051717-20220930081717-00727.warc.gz | en | 0.931936 | 1,370 | 2.65625 | 3 |
AI-Driven Testing: A New Era of Test Automation
Test automation as we know it today is nothing more than an illusion, created in the absence of technological software testing innovation. If you don’t believe me just look up the definition of automation in the dictionary, or compare the automation practices in software testing with those of other disciplines. Automation is the process of making a machine or system work without being directly controlled by a human. When something is automated, it becomes capable of starting, operating, and completing independently. In manufacturing, as far back as the mid 1950’s, ideas around automation led to the notion of lights out manufacturing — production work being done with the factory lights out because it is not reliant on continual human presence. Unfortunately, our current view of automation in software testing is far from being a lights out philosophy.
“Automation as it is defined in software testing is neither very automated, nor has much to do with testing.”
When testers talk about automation, they are referring to manually encoding a predefined set of program input actions and output verification steps into a script, which can in turn be executed by a machine. Upon execution, a log of the results is generated, stored, and associated with the run. Clearly, the only aspects of this process that are truly automated are test execution and logging. Human testers are needed to define testing goals, acquire the knowledge necessary to adequately test the software, design and specify detailed test scenarios, write the test automation scripts, execute scenarios that could not be automated, and analyze the test results to determine any threats to the project. Not only does this activity require too much manual effort to be called automated, but it is too granular to be classified holistically as testing. Software testing expert Dr. Cem Kaner once described a test as a question you ask a program to gain information about it. The narrower the scope of the question, the more limited the investigation around it, and the less you are likely to learn from its answer. Along similar lines, James Bach and Michael Bolton distinguish between testing and checking.
• Testing is evaluating a product by learning about it through experimentation, which includes questioning, studying, modeling, observing and making inferences.
• Checking is making evaluations by applying algorithmic decision rules to specific observations of a product.
Simply put, a test is an experiment, and although checking specific facts about a program may be a part of that experiment—testing is a lot more than fact checking. Whether or not you agree with the distinction between testing and checking, current test automation practices are limited. Automated test scripts tend to only target functional, structural and performance issues. Testing concerns like usability and accessibility are often deemed too difficult to automate. On the bright side, we are entering a new era where artificial intelligence (AI) and machine learning (ML) are redefining the meaning of automated testing.
“If testing is about evaluating a product by learning about it through experimentation, then automated testing is about having machines perform those activities instead of humans.”
Traditionally machines have been programmed to follow explicit instructions. Humans, on the other hand, learn a lot through observation and experience. AI and ML allow computers to learn like humans by representing past observations and experiences as data. Instead of hard-coding knowledge and task-specific instructions, learning algorithms are trained using concrete examples of the concepts the machine needs to recognize. Modern applications of AI include image and voice recognition, e-mail spam filtering, credit card fraud detection, medical diagnosis, gaming, and self-driving vehicles. Researchers and practitioners have recognized the potential for advances in AI to help bridge the gap between human and machine-driven testing. As a result, a new wave of AI-based test automation tools is already being developed. Such tools leverage autonomous and intelligent agents, commonly referred to as bots, to automatically drive the testing process.
“Test bots are designed to mimic how human testers observe, explore, model, reason, and learn about a software product.”
So what are some of the ways in which AI can imitate human testers? Here are seven AI-driven testing capabilities that exist today:
1. Discovering Application Structure: Like humans, test bots can perceive the different screens and widgets in an application and classify them correctly, even if details are missing or it is the first time the bot is examining the application under test.
2. Exploring Application Behavior: Test bots generate actions such as filling out a form on a screen, clicking submit, and checking for an appropriate response. The likelihood of a bot performing a given exploratory action may be based on historical data collected from human-present software testing sessions.
3. Modeling and Reasoning: To perform their functions, test bots construct models of the application under test or of the different testing activities. They use these models to make decisions in the presence of uncertainty, or reason about the quality of their own actions and observations.
4. Detecting Failures and System Changes: Test bots leverage image recognition and other techniques to determine when a failure occurs, or to detect legitimate changes in the current version of the application. AI has enabled visual UI test automation.
5. Learning from Tests or User Traces: Artifacts such as test scripts and execution traces contain concrete examples of interesting paths that human testers and end users cover when exploring a given application. Test bots trained on these real-world examples can generalize them to new applications.
6. Declarative or Goal-Based Testing: An impressive feature of AI-driven testing is the ability to specify a testing goal and have the bots automatically figure out how to achieve that goal. Intent may be specified in natural language or using an abstract testing language.
7. Adapting Testing: Test bots can modify their behavior at runtime based on feedback. This is generally achieved through an ML technique known as reinforcement learning. In reinforcement learning, positive outcomes are rewarded, and negative outcomes are punished, allowing the bots to improve over time.
It is important to mention that AI-driven testing is still in the infant stages with much work to be done. Testing challenges like the oracle, input and state explosion, test data selection, among others, still remain open research problems. Furthermore, most of the current AI-driven testing tools address software testing from an external, black-box perspective. However, a significant amount of software testing today happens from the inside using white-box approaches. Nonetheless, despite its infancy, challenges, and limitations, AI-driven testing is bringing some much-needed innovation and buzz to the industry. The age of the testing robot is upon us — the question is: Are you ready for the rise of the testing machines? | <urn:uuid:92e07362-3949-4b4b-a496-9e3734421710> | CC-MAIN-2022-40 | https://softwaretesting.cioreview.com/cxoinsight/aidriven-testing-a-new-era-of-test-automation-nid-27744-cid-112.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337360.41/warc/CC-MAIN-20221002212623-20221003002623-00727.warc.gz | en | 0.930504 | 1,368 | 2.9375 | 3 |
The Basics Of GSA Contract Schedule – President Harry S. Truman founded the General Services Administration, or GSA, in 1949 to streamline the federal government’s administrative duties. The initial aim of the GSA, which was to dispose of war surplus items, manage and preserve government archives, oversee emergency preparedness, and stockpile strategic supplies for wartime, has given way to a considerably broader purpose and function.
GSA now functions as the federal government’s purchasing arm, facilitating the acquisition of products and services for tens of thousands of federal agencies and bureaus through Contract Schedules.
What is a GSA Schedule?
A GSA Contract Schedule is also known as MAS or Federal Supply Schedule. This schedule is a long-term, government-wide contract with commercial enterprises that gives federal, state, and local government buyers access to over 11 million commercial items and services at volume discount prices.
GSA provides federal agencies with a wide range of products and services that they require to serve the general public. The GSA Contract Schedules initiative gives the government and commercial partners reduced lead times and more transparency. The Federal Acquisition Service (FAS) provides a wide range of products and services to the federal government at the best possible price, bringing in about $33 billion per year.
One significant distinction between selling your products and services under a GSA contract and selling commercially is that when you sell under a GSA contract, the terms and conditions have already gotten agreed upon throughout the GSA Schedule procurement process. The purchase procedure for the many federal departments in the government is made easier by having agreed-upon terms and conditions.
Click here to calculate ROI!
A GSA Contract Schedule opens up a whole new world of possibilities in the federal market for the vendor. The General Services Administration developed the Contract Schedule. The schedule became established with the sole objective of assisting federal agencies in their purchases of goods and services from suppliers. Thus, government agencies can buy from you because you have a GSA Contract Schedule, which means your rates and company have gotten thoroughly verified.
How long is a GSA Contract Schedule?
As established, GSA supplies federal agencies with a wide range of products and services that they need to serve the general public. The GSA Contract Schedules effort reduces lead times and increases transparency for the government and commercial partners. The Federal Acquisition Service (FAS) provides the federal government with a wide range of products and services at the best available price.
During the process, pricing, terms, and conditions are all discussed. In this sense, a GSA consultant can assist you in positioning your contract for success throughout the contract’s anticipated 20-year duration, with an initial 5-year Contract Schedule.
The federal government’s 5-year, renewable GSA Contract Schedule explains why the GSA proposal process necessitates so much documentation, assessment, and evaluation; GSA wants to make sure that its contractors can fulfill orders for customers in the long run. This process also tells GSA clients that contractors awarded a GSA Schedule are determined to be stable and sustainable for a minimum of five years to provide allowed items or services to government customers, enticing government customers to the Schedules program for procurement.
How are GSA Contract Schedules Awarded?
Although any company that fulfills certain minimum qualifications can do business with the government through GSA, strict vetting procedures are in place to keep financially irresponsible or unscrupulous vendors out of this favored purchasing scheme. Suppose you want to do business with the government through GSA. In that case, you must first determine which Contract Schedule your goods or services fall under, create a System for Awards Management (SAM) profile, and electronically submit a proposal with all necessary supporting documentation of the GSA’s e-Offer system.
While gathering information about doing business with the GSA can be quite challenging, obtaining a Contract Schedule process is easier. On the other hand, completing the following stages and submitting a proposal to the GSA that they will accept can be one of the most arduous tasks a company will ever undertake.
After you submit your offer to GSA, it will get reviewed by a contracting officer. You will receive a GSA contract if all standards get met, the pricing is “fair and reasonable,” and the offer is in the government’s best interests.
You have finally arrived at the starting line after receiving your GSA Contract. You read that correctly: the starting line. Now, it is time to get down to business: GSA Management. Tired? Do not get discouraged just yet. Consider receiving your GSA Contract as Basic Training, and now you are ready to go into combat. Preparation and discipline are required to win a GSA contract. GSA management is difficult but necessary for success.
When developing a successful marketing plan to sell to the federal government, preparation is essential. This preparation will entail completing your set-asides, obtaining your GSA contract, and attending networking events, among other things. This development also entails looking for and tracking bid opportunities, as well as pursuing them with proposals. It necessitates extensive planning in terms of personnel and procedures.
Moreover, GSA Contract Maintenance and GSA Contract Compliance are where discipline comes into play. Call it what it is: GSA management is not pleasant, and it will not win you any awards, but it will surely keep you out of business troubles.
While obtaining a GSA MAS contract does not guarantee government sales, it can be a useful contract vehicle for businesses to tap into the federal market and increase sales. GSA MAS contracts necessitate a lot of effort and upkeep, both in terms of keeping your contract valid and actively marketing your products or services to create revenue. Many businesses find this process intimidating, but the numerous benefits that await are well worth the effort.
With MAS contracts accounting for $48 billion in annual federal procurements, it is worth your time to secure one if you can. When you realize that you can get a five-year MAS contract with the option to renew for up to 20 years, it is a wonder you have not already applied! | <urn:uuid:bf0cc33c-9bf6-496e-b4ca-0f7e7b2ac4b9> | CC-MAIN-2022-40 | https://www.gsascheduleservices.com/blog/basics-of-gsa-contract-schedule/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337480.10/warc/CC-MAIN-20221004054641-20221004084641-00727.warc.gz | en | 0.945187 | 1,256 | 2.53125 | 3 |
Global pandemics such as the scale of Covid-19 or the Spanish Flu have an annual occurrence probability that varies between 0.27% and 1.9%. And while organizations with robust enterprise risk functions had identified pandemics as one of their risks, the low probability meant that few had put in place measures to mitigate against the potential occurrence.
Safe to say, we have all been schooled at the moment.
From cyberattacks to air crashes, third party compromise to regulatory changes, employee unrest to economic downturn, the business environment is rife with uncertainties. Having an approach to anticipate and limit the impact should such materialize is critical for any enterprise that wants to remain.
As an organization defines its strategic goals and objectives, a realistic look at threats to success can go a long way in enabling the enterprise to remain on track. Investing in a risk management approach is the mark of mature companies who are well aware that the path to their vision is not always straightforward.
Let’s look at some of the key aspects define risk management.
What is risk?
The ISO 31000 standard for risk management guidelines defines a risk as:
The effect of uncertainty on objectives.
The outcome of the uncertainty can swing in either a positive or negative manner. If the risk is negative, then the uncertain outcome results in harm or loss for instance lost customers, regulatory penalties being imposed or reduced business revenue. On the other hand, if the risk is positive, the uncertain outcome can result in benefits if exploited e.g., regulation changes can be favorable in terms of new business opportunities.
Elements of risk
To fully express a risk, one has to consider the following elements:
- Risk source. An element which, alone or in combination, has the potential to give rise to risk. Examples here include weather conditions, government agencies, disgruntled employees, etc.
- Risk event. The potential occurrence or change of a particular set of circumstances. For example: a cyberattack, flooding of a data center, mass resignation, adverse regulation, etc.
- Risk consequence. The outcome of an event affecting objectives. For instance lost revenue, penalties from a regulator, disrupted operations, corrupted data, etc.
- Risk likelihood. The chance of something happening—for instance, low or high probability which can be objectively or subjectively computed.
Responding to risk
In order to effectively respond to risks, an approach is required. That’s where risk management comes into play.
Defining risk management
ISO 31000 defines risk management as
Coordinated activities to direct and control an organization with regard to risk.
Principles for effective risk management, as outlined in ISO 31000 include, ensuring that your risk management practice:
- Creates and protects value.
- Is made an integral part of all organizational processes.
- Is made part of decision making.
- Explicitly addresses uncertainty.
- Is systematic, structured, and timely.
- Is based on the best available information.
- Is tailored.
- Takes human and cultural factors into account.
- Is transparent and inclusive.
- Is dynamic, iterative, and responsive to change.
- Facilitates continual improvement of the organization.
Risk management steps
Let’s look at a couple well-known frameworks.
Management of Risk framework
At a high level, the risk management process can be broken down into five iterative steps as outlined by Axelos’ Management of Risk (M_o_R) framework:
The organization identifies its strategic and operational context, and then identifies the risks based on that context. The context leads to a determination of the organization’s capacity and tolerance to risks should they materialize. Risks identified are documented in a risk log or register.
The risks identified are then assessed to determine the likelihood and consequence. This then leads to an evaluation of the assessment to rank the risks from a priority perspective, where risks with higher consequence and likelihood are prioritized higher. A risk heat map is a tool that can be used to visualize risk prioritization.
Planning involves identifying and evaluating the appropriate risk response to remove or reduce threats, and to maximize opportunities. Responses can be categorized as follows:
- Avoid: Making the uncertainty void by not proceeding with the plan of action where the risk would materialize. For example, not hosting your data on the cloud due to risk of transfer of personal data outside local jurisdiction.
- Reduce: Identify actions to reduce the probability and/or consequence should the risk materialize by putting in place mitigation controls. For example, putting policies to prevent senior officials from travelling on same flight or vehicle.
- Transfer: Identify a third-party who is willing to take up the risk on behalf of the organization. This option is usually tagged to insurance covers.
- Share: Identify a third-party who is willing to take up part of the risk with the organization. This option is usually applied to customers, partners or suppliers.
- Accept: Live with the uncertainty and take no action to forestall it.
Here the planned risk responses will be actioned, their effectiveness monitored and corrective action taken where responses do not match expectations.
This is a standalone step that occurs concurrent to the previous four. Risk information and treatment status is reported to key stakeholders based on agreed channels. This step is also very relevant whenever an identified risk materializes.
NIST risk management framework
The NIST risk management framework (RMF) provides a comprehensive, flexible, risk-based process that integrates security, privacy, and cyber supply chain risk management activities into the system development life cycle through 7 steps outlined below:
- Prepare. Carry out essential activities to help prepare all levels of the organization to manage its security and privacy risks.
- Categorize. Determine the adverse impact with respect to the loss of confidentiality, integrity, and availability of systems and the information processed, stored, and transmitted by those systems.
- Select. Select, tailor, and document the controls necessary to protect the system and organization commensurate with risk.
- Implement. Implement the controls in the security and privacy plans for the system and organization.
- Assess. Determine if the controls are implemented correctly, operating as intended, and producing the desired outcome with respect to meeting the security and privacy requirements for the system and the organization.
- Authorize. Provide accountability by requiring a senior official to determine if the security and privacy risk based on the operation of a system or the use of common controls, is acceptable.
- Monitor. Maintain ongoing situational awareness about the security and privacy posture of the system and organization to support risk management decisions.
Risk Management Roles
Now that we understand the purpose and steps in any risk management practices, let’s look at the people involved. Key roles required for effective risk management in an organization include:
- Risk Committee. This is a subset of the organization’s board whose mandate is the oversight and approval of the enterprise risk management framework. This includes defining risk tolerance and appetite, providing resources for risk mitigation, setting governance policies, and evaluating performance of the implemented risk mitigation.
- Risk Manager. This role is responsible for coordinating the implementation of the enterprise risk management framework including guiding the rest of the organization in identifying, assessing, mitigating, and monitoring risks. The role will provide reports on the status of the risk management framework and can be elevated to Chief Risk Officer or Head of Risk depending on the size of the organization.
- Risk Officer. This role reports to the risk manager and carries out the basic risk management activities and maintains documentation on the same.
- Risk Owner. This role is responsible for the management, monitoring, and control of all aspects of a particular risk assigned to them, including the implementation of the selected responses to address the threats or to maximize the opportunities.
- Risk Actionee. This role is responsible for implementation of selected risk responses. It can be carried out by the Risk Owner or be outsourced to a third party.
Success factors in risk management
Success in risk management is a chance in itself—that’s because you can never plan perfectly (unless you can see the future). However, having a robust yet flexible framework can be the difference between successfully navigating through a challenging risk or seeing your enterprise going under.
Key elements required in successful risk management according to the ITIL 4 practice guide include:
- Establishing governance of risk management
- Nurturing a risk management culture and identifying risks
- Analyzing and evaluating risks
- Treating, monitoring, and reviewing risks
- BMC Service Management Blog
- What Is GRC? Governance, Risk, and Compliance Explained
- IT Security Vulnerability vs Threat vs Risk: What are the Differences?
- Risk Assessment vs Vulnerability Assessment: How To Use Both
- Error Budgets Explained: Risk & Reliability in One Metric
- Structured vs Unstructured Data: A Shift in Privacy | <urn:uuid:4f03234b-0b51-4b54-bf33-0cffd381ccdd> | CC-MAIN-2022-40 | https://www.bmc.com/blogs/risk-management/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337631.84/warc/CC-MAIN-20221005140739-20221005170739-00727.warc.gz | en | 0.912676 | 1,870 | 2.84375 | 3 |
- What Are Cookies
- Third-Party Cookies Explained
- Browsers vs. FLoC
Cookies – A word you encounter almost every time you open a web page, surf Facebook, or even chat using an online platform. These files are created the moment you visit a website, and their sole purpose is to store information about a user’s interactions with a certain website. Some are good, some are bad. However, the worst of them all is definitely those we call Third-Party Cookies. In fact, this particular type is why cookies have such a bad reputation among internet users.
Cookies – Inside the Crumbs of Data
First, let’s talk about the origin of the name. The word cookie is kind of weird to describe bits of data, and to date, there’s no straight answer on why it’s called that.
A theory states that the word Cookie came from Hansel and Gretel, two characters of our childhood bedtime stories who navigated the forest using Cookie crumbs.
Another theory states that the name came from “Magic Cookie,” which, back in 1979, was a term that reflects a packet of data.
Finally, there’s one reflecting the Andy Williams Show, where Cookie Bear – a character in the show always asks for a cookie, much like what your browser does almost every time.
Thanks to the EU’s GDPR law, all users must agree to cookies that store personal data. Cookies give convenience and efficiency.
By collecting all sorts of information about you when you visit a certain website (browser type, preferences, location, and more), the website you visit another time will enhance your experience.
There are 7 notable types of cookies:
- First-Party Cookies
- Session Cookies
- Secure Cookes
- HTTP-Only Cookies
- Flash Cookies
- Zombie Cookies
Yeah, we did mention six above because the main focus today goes to the ones that give all of them the bad reputation Cookies have nowadays – Third-Party Cookies.
The Bad Guys – Third-Party Cookies
Third-Party Cookies exist in the internet world to track users, period. All they do is collect information about everything you do online, including your online behavior, browsing history, demographics, spending habits, and more.
Such information is very valuable, not only to the user but to advertising networks as well. Using these cookies has become their favorite way to drive up their sales and pageviews.
Third-Party Cookies can be quite frustrating, especially when it comes to privacy. Not to mention that they’re the main reason you get bombarded with unwanted ads all the time.
Now, most browsers, if not all, offer a way for users to block third-party cookies. It’s always recommended that you take the necessary steps in the browser of choice.
Both Safari and Firefox took the necessary actions and started blocking these cookies by default. And now, Chrome is eliminating them altogether.
Google’s New Policy – Enter FLoC
Google will now be targeting ads based on the users’ general interest. It’ll do that using a new AI system that goes by the name of FLoC (Federated Learning of Cohorts), which takes the users’ web history, among other things, and puts them into a certain group based on their interests.
According to Bennett Cyphers, a technologist at civil liberties group the Electronic Frontier Foundation:
“They’re going to get rid of the infrastructure that allows individualized tracking and profiling on the web. They’re going to replace it with something that still allows targeted advertising – just doing it a different way.”
With the new policy, everything will change in the ad industry, especially when we speak of small businesses. The fact is that the new FLoC system will study how users browse and then group them into ‘cohorts.’
As a result, users won’t get ads based on their individual data. Instead, they’ll be served with ones according to what the “Group/Cohort” is interested in as a whole. In other words, advertisers will still target them with ads, but in a more anonymous matter.
When Will It Be Applied?
According to Google, the new policy won’t be applied until 2022. However, they did state that some users will be able to test it out in select countries such as Australia and New Zealand. Here’s the full statement:
“The initial testing of FLoC is taking place with a small percentage of users in Australia, Brazil, Canada, India, Indonesia, Japan, Mexico, New Zealand, Philippines, and the U.S. We’ll expand to other regions as the Privacy Sandbox expands globally. In April, we’ll introduce a control in Chrome Settings that you can use to opt-out of inclusion in FLoC and other Privacy Sandbox proposals. In the meantime, if you’ve chosen to block third-party cookies, you won’t be included in these origin trials.”
For the time being, Google is providing an “origin trial” that’s designed to help websites learn how FLoC works. Also, some users are even testing it on their own.
No Shift in Power
So, why is this challenging for smaller companies? First, many advertisers will have no other choice but to use closed platforms such as Facebook, TikTok, and the likes to target their ads.
Second, with this new mechanism, the powerful players in ad tech could become even more entrenched. That’s because they have all the means to cope with FLoC and what ads to target against them.
In other words, smaller advertisers won’t get their turn to act until efficiency and effectiveness benchmarking by larger companies are released.
Google says that it’s going to give this a try and see how it all turns out. Regardless of the outcome, if ad companies adopt the new policy, it should protect users from tracking attempts and mechanisms that are more intrusive than third-party cookies. We’re referring to the likes of fingerprinting.
Browsers vs. FLoC – A FLoC-ing Complication
While Google shows its best intentions with applying FLoC later on in 2022, no other browser vendor has signaled its intention to support the new policy.
In fact, many privacy-focused browsers such as Brave, DuckDuckGo, and Vivaldi have warned users against Google’s new policy and informed them that they are going to block it on their end. Here’s how each one stands:
The widely popular engine has enhanced tracker blocking on their own Chrome extension in order to block all FLoC interactions on websites. They’re really focusing on fighting the upcoming policy, and they’re not hiding it at all. According to DuckDuckGo:
“It groups you based on your interests and demographics, derived from your browsing history, to enable creepy advertising and other content targeting without third-party cookies. After a short trial period, Google decided not to make this new tracking method a user choice and instead started automatically including millions in the scheme. If you’re reading this in Chrome while logged in to a Google account, yes, that likely means you too, and if not now, then eventually.”
Apparently, DuckDuckGo does not believe that this won’t do users any good. Instead, it’ll bombard them with ads that might slightly relate to them but not at all what they wish to see. Remember, the ads you get are based on what the group you get sorted in likes.
The open-source web browser has also shown how much they’re against FLoC and how they will block it whenever it gets implemented.
So far, Brave disabled FLoC on the Nightly version of its application, be it on Android or desktop. Below, you can find Brave’s full statement:
“The worst aspect of FLoC is that it materially harms user privacy, under the guise of being privacy-friendly. Others have already detailed many of the ways FLoC is privacy harming. We note here just three aspects of FLoC that are particularly harmful and concerning. Google’s approach to determining whether a FLoC cohort is sensitive requires (in most cases) Google to record and collect that sensitive cohort in the first place! A system that determines whether a cohort is “sensitive” by recording how many people are in that sensitive cohort doesn’t pass the laugh test.”
So, according to Brave’s CEO, Google is not being very transparent when it comes to how it’s marketing its new policy. He did state that FLoC is privacy harming, regardless of Google stating that it’s privacy-friendly.
We don’t know how this shift in Google’s policy will affect us yet. However, we do know that if you’re a regular user, nothing will change when it comes to your experience.
Google’s push to eliminate third-party cookies seems quite weird, considering it alone will have the power over the advertisement. Is it a way to make more money?
Does it really care about protecting the users’ privacy? Only time can tell. What we do know, though, is that it’s going to change online advertising as we know it. If you have any thoughts on Google’s new Cookies Policy, please share them in the comment section below. | <urn:uuid:4f25231f-6231-49fc-b443-fe122741f399> | CC-MAIN-2022-40 | https://anonymania.com/new-chrome-cookies-policy-floc/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337889.44/warc/CC-MAIN-20221006222634-20221007012634-00727.warc.gz | en | 0.940727 | 2,122 | 3.03125 | 3 |
Miles Tayler, senior fellow at the R Street Institute, recently moderated a fascinating panel on quantum computing.
Several others were with him on the panel: Chris Fall, Ph.D., senior adviser, president’s office, Center for Strategic and International Studies; Scott Friedman, senior policy adviser, House Homeland Security Committee; Allison Schwartz, global government relations and public affairs leader, D-Wave Systems; and Kate Weber, Ph.D., policy lead for quantum, robotics, and fundamental research, Google.
They spoke about the practical applications of quantum computing, how the U.S. was falling behind several companies on this technology, and why that could be a terrible thing.
Let’s talk about quantum computing and how it will be a game-changer for everything from simulation and modeling to artificial intelligence (AI). I’m pretty sure we aren’t ready for a quantum AI:
Quantum capability is closer than we think
While we are still at least a decade from when we get to a point when quantum computers can even begin to reach their full potential, quantum emulators and current generation quantum computers are beginning to do some real work.
We are learning that quantum computers are naturally better when emulating and interacting with reality, because nature doesn’t consist of just 1s and Os. The more complex a problem, the better able a quantum computer can deal with it.
Initial applications are increasingly focused on logistic types of problems with many moving parts. For instance, it is being used heavily by foreign governments that use it for logistics management, such as Australia’s military using autonomous vehicles that don’t put soldiers or contractors at risk. Current generation computers were not powerful enough, and since the military lives or dies on logistics, having a far more effective tool using quantum computing could.
Australia would have an increasing advantage over time if it came to open conflict due to a quantum tool to aid in logistics. In addition, applied to emerging intelligent weapons, like drone swarms, quantum computing should position the drones collectively for maximum impact more effectively. Together, these two implementations would give a military unit with this capability a significant advantage over those that do not.
Other examples are emergency response modeling and execution. The quantum computer is first asked to create a plan for a major disaster. It then assists in getting the available resources to those that most need them and mapping out and dynamically changing the plan for evacuations depending on then-current conditions.
Quantum + AI
But things get interesting when we add quantum computing capability to an AI. Quantum computing can provide AIs with the ability to emulate emotions and act like they are feeling them. While this alone wouldn’t represent sentience, it would be hard to tell the difference. And a quantum AI would be better able to respond to complex signals, like expressions, eye movement, and body language that traditional computers find challenging.
An AI with a quantum back end could perform the role of therapist, say on a submarine or space exploration vessel that couldn’t justify a human therapist. And it would be far less likely to be biased, assuming it was properly trained.
A quantum AI would have a significant advantage as a 100% audit function. It would look at every transaction and see whether it was likely fraudulent or in violation of policy from the related metadata. Current human-driven audit organizations don’t have the bandwidth for 100% audits and miss a lot of actual crimes because they have to operate from far smaller samples.
Another area where quantum AI would make a considerable difference is government, as it could almost immediately identify graft and bribery. The relevant AI would be able to distribute the limited day-to-day resources daily more effectively, especially during a catastrophe, and assess liability for complex decisions far more accurately.
Quantum will be a major industrial advantage
While quantum computing is still far from its potential, with only a tiny fraction of the number of qubits needed to demonstrate that potential, it is already showing viability in several areas.
Those areas are military, such as in logistics and weapons, smart cities, government, emergency response, modeling, and simulation, where the complex problems are beyond conventional computers’ capabilities.
These capabilities will be a competitive game-changer for the armed forces, governments, and companies that effectively use this technology first. There will be such a significant advantage in any highly complex market, like stock trading, that those that don’t have access to this technology could quickly be eclipsed and made redundant by those that do. Fortunately, there is still time but that time is running out. Should a critical mass of governments, companies, or individuals get access to this technology, they’ll have massive advantages over those that don’t.
We are not ready for quantum AI, and we are running out of time. | <urn:uuid:f3cb068d-736b-431a-a5fc-f0f07c145de7> | CC-MAIN-2022-40 | https://www.datamation.com/artificial-intelligence/quantum-ai-are-we-ready/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337398.52/warc/CC-MAIN-20221003035124-20221003065124-00127.warc.gz | en | 0.948935 | 993 | 2.625 | 3 |
Technology is enabling the creation of open learning environments that will eventually replace the idea of a traditional classroom, and this will be significantly enhanced as virtual and augmented reality become more pervasive. Educators are largely embracing these changes, which bring higher levels of engagement and richer learning experiences to students. Learning — including vital digital skills such as coding — should be regarded as an enjoyable pursuit that is expected to continue throughout adult life, not as something we’re “finished with” as soon as we leave school.
Those were some of the key takeaways from a recent Technology in Education round table hosted by Dell at the iconic Gherkin building in London. The event brought together technology and education experts and was moderated by Matt Britland, ICT Director at Lady Eleanor Holles School, who led the discussion around how we can best incorporate existing and emerging technologies into the curriculum, supporting teachers and preparing students for a world where having a dynamic digital skill set will become even more essential.
“It’s not just about using technology, but understanding the reason why you’re using a particular technology. Once you get buy-in from parents and teachers as well as students, that’s when you have a successful integration of technology in the learning process,” says Britland drawing from his own personal experience of rethinking the systems in his school to ensure they could cope with being “bombarded” with over 4000 devices connected to the network, for example.
“This type of digital infrastructure needs to be robust, scalable, flexible, and, above all, invisible,” says John Bailey, an education expert from Dell. “You don’t think about the fact you have running water, electricity or heating. It’s ubiquitous, a part of life. In the same way, we shouldn’t’ need to think about the infrastructure we use to access digital resources.”
Universally implementing such infrastructure is, of course, an enormous challenge, as not all schools have the internal expertise and resources to devise and deploy a comprehensive digital strategy. So there is a role for commercial partners in supporting this transition, but also a need for governments to prioritize and enable investment in this area at a broader policy level, so that we don’t end up creating a two-tier system where some schools can afford to become much more digitally advanced than others.
The flipside of having intuitive technologies and invisible digital infrastructure is that the next generation might be proficient in using technology, yet lack the core digital skills required to understand what makes that technology work.
“We’re teaching HTML and Python and a website called Code Adventures where they earn points, badges, etc.” says Britland. “In parents’ evenings they want to know what technology the students are using, they want to know that their daughter is coding.”
There are also many successful start-ups coming up with creative hackable toys that teach children principles of electronics, robotics and coding through play and gaming such as Kitronik, Sphero, Raspberry Pi, Technology Will Save Us and many others. And the advent of Virtual Reality will only bring with it more opportunities for teachers and students to experience technology in new ways, and to “get their hands dirty,” so to speak.
“When we talk about VR, we talk about those experiences,” agrees Jaime Donally.
Instructional Technology Coordinator and founder of edcamp Global. “And I love that.” However, once you have a class set of cardboards in the classroom, there is a lot more that can be done with that: “The next step, of course, is our kids being the creators of those experiences”
Tools like CoSpaces promise to make that process of generating content in Virtual Reality easy by offering a 3D visualization tool that allows users to select environments from a library, adding and adjusting various elements to create a personalized experience.
It all boils down to building engagement, and in practice this translates into blurring the boundaries between work and play, making full use of tolls such as gamification so that students are constantly empowered to take responsibility and pride in their own learning, and to see it as a continuous process that doesn’t necessarily begin or end at the school gates. As David Whelan recently said when unveiling his Engage VR Education platform, we are much more likely to retain information if we create an emotional connection with that learning experience.
“That’s sometimes a difficult shift for schools to make, because we’re used to looking at things a certain way,” says Matt Britland, “but we want to instill a passion for learning in our students, and that doesn’t just come from learning what we tell them to learn, it’s about them going out and exploring, finding out things for themselves.”
This is, in fact, the direction that futurist Isaac Asimov predicted we would take. I find it fascinating to listen to old interviews — dating from a time way before the internet, when the idea of having a computer in one’s home seemed as outlandish as the plot of one his science fiction books — as he accurately predicts trends such as personalized learning, and how the symbiosis between technology and education will develop and democratize learning for the masses.
“Nowadays what people call learning is forced on you and everyone is forced to learn the same thing on the same day at the same speed, in class. For some it goes too fast, for some too slow, for some in the wrong direction. It seems to me that it’s through this machine that for the first time we’ll be able to have a 1:1 relationship between information source and information consumer.”
In the old days, he explains to his bewildered interviewer, you had tutors teaching children, and they would adapt their teaching to the tastes and abilities of each student. Few could afford this, however, so most children remained uneducated. When educating the masses later became a recognized necessity, the only way this could be achieved was through having one teacher for many students, and a curriculum to organize their learning. So you ended up with either a one-to-one relationship for the few, or a one-to-many relationship for the many. Through technology, however, we can finally enable a one-to-one relationship in the form of access to the gathered knowledge of the human species. And this, he emphasized, would not just benefit the young.
“People don’t stop what they enjoy doing just because they reach a certain age. They don’t stop having sex just because they turn 40, they keep it up as long as they can if they enjoy it. The trouble with learning is that most people don’t enjoy it because of the circumstances. Make it possible for people to enjoy learning and they’ll keep it up.”
I find that his outlook chimed with that of the panel, where the picture that emerges for the future of education is one where technology enables learning for everyone, regardless of their age, wealth, location, gender, race or any other factor. And one where everyone has quite a bit of fun in the process. | <urn:uuid:422d0eb5-eb66-42d6-aa2d-a8ef0ba99c3e> | CC-MAIN-2022-40 | https://www.cio.com/article/236284/imagining-the-future-of-edtech.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337971.74/warc/CC-MAIN-20221007045521-20221007075521-00127.warc.gz | en | 0.967292 | 1,509 | 3 | 3 |
Published On April 28, 2020Healthcare supply chain management has been severely impacted during COVID-19. Here’s a look at the FDA’s response.
Personal protective equipment and ventilator shortages have dominated headlines, but COVID-19 has also caused shortages across a range of drugs — many of which are critical for maintenance of unrelated conditions. At the same time, actions taken by the Food and Drug Administration (FDA) have had a significant impact on healthcare supply chain management in intercepting challenges and mitigating disruptions in access to drugs.
At Risk of Falling Short
In addition to a shortfall in drugs vital to the maintenance of chronic illnesses, anesthetics and medications to reduce pain are in danger of running low, as well.
Chloroquine and Hydroxychloroquine
Lupus and arthritis patients have been vocal about their challenges obtaining hydroxychloroquine ever since rumors began to spread that the drugs could be effective in treating COVID-19. Prescriptions of hydroxychloroquine (brand name Plaquenil) have spiked, putting the drug in short supply since March 9, according to the American Society of Health System Pharmacists. Currently, manufacturers are working to increase production to respond to demand.
To minimize the spread of COVID-19 within hospitals, doctors in many facilities began avoiding nebulizers for patients with respiratory issues, instead shifting to albuterol asthma inhalers. The result has been an inhaler shortage in both hospitals and retail pharmacies, largely caused by dependence on bulk ingredients and medicines produced offshore.
To mitigate the issue, the FDA recently approved the first generic albuterol inhaler to support the public’s needs, reports U.S. News.
While efforts to ramp up ventilator production are helpful, they only resolve part of the shortage issue — patients on ventilators require sedatives, anesthetics and paralytic agents to address respiratory failure.
For example, New York has seen a spike in demand for fentanyl, used as a sedative, and the FDA placed midazolam on the official drug shortage list in March, according to the Regulatory Affairs Professionals Society. These increases in use run headlong into challenges produced by the opioid epidemic; in response, the Drug Enforcement Administration proposed reducing quotas of fentanyl and midazolam back in September.
Supply Chain Dependence on China
As far back as January, experts expected interruptions to the supply of active pharmaceutical ingredients (APIs). A report by API wholesaler Kemiex found that 85% of supply professionals expected low to high impacts, which, at the time, was largely based on the spread of COVID-19 through China, the effects of an extended Chinese New Year and delayed start of production. As the disease continues to spread across the country, more professionals are questioning reliance on China for production of critical drugs.
Today, it’s unclear just how much we depend on China. So far, the FDA has reported that the number of Chinese facilities licensed to produce APIs is 13% (compared to 28% domestically), but that number provides no insight into whether they’re producing the drugs, how much they’re producing or where the drugs are being distributed, notes FiercePharma.
The agency has suggested advanced manufacturing technologies (such as continuous manufacturing and 3D printing) as an approach to improve U.S.-based production, but they can’t force drug manufacturers to move away from the century-old method of batch production toward more modern methods.
FDA Taking Action
Since the end of January, the FDA has been in contact with over 180 drug manufacturers to remind them of their legal obligation to notify the agency of anticipated disruptions, but also to urge them to evaluate their entire supply chain, including main ingredients that are manufactured in China. The agency also has identified about 20 drugs whose active ingredients or finished drug products are sourced directly from China, and has been in contact with those firms to assess COVID-related shortage risks.
In addition, the president of the United States’ budget includes three, drug-related proposals to support the agency in addressing product shortages:
lengthening expiration dates on critical drugs to the longest possible date that is scientifically justifiable
improving critical infrastructure through risk management plans that would identify supply chain vulnerabilities and enable the agency to strengthen the supply chain by creating contingencies for emergency situations
improving critical infrastructure by empowering the FDA to require access to information needed to assess infrastructure, and manufacturing quality and capacity
The FDA is keeping updated information about actions being taken around drug shortages, healthcare supply chain management and more on the U.S. FDA COVID-19 information page. | <urn:uuid:18bb4479-65c9-4705-8861-1d4066124b1f> | CC-MAIN-2022-40 | https://www.ironmountain.com/blogs/2020/drug-shortages-and-healthcare-supply-chain-management-during-covid-19 | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334515.14/warc/CC-MAIN-20220925070216-20220925100216-00327.warc.gz | en | 0.944408 | 983 | 2.515625 | 3 |
Excel System Variables
You can use Excel System Variables to manage Excel automation processes. These variables are specific to the scope of the Excel command.
Excel relevant system variables
- Excel Cell - This will indicate the column name and position of the cell from the header, when 'Contains Header' is enabled for a session in Open Spreadsheet of the Excel Command. It is applicable to Excel operations like Get Cells, Set Cells, Go to Cell, Delete Cells and Find/Replace.
- Excel Cell Column - This will return the column of the Excel on which the operation has been performed.
- Excel Cell Row - This will return the cell of the Excel on which the operation has been performed.
To know more about the Excel operation, refer Excel command - Contains Header | <urn:uuid:bb18be2a-e125-4918-bcdb-0289bd5eec3c> | CC-MAIN-2022-40 | https://docs.automationanywhere.com/bundle/enterprise-v11.3/page/enterprise/topics/aae-client/bot-creator/using-variables/excel-system-variables.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334912.28/warc/CC-MAIN-20220926144455-20220926174455-00327.warc.gz | en | 0.870774 | 161 | 2.921875 | 3 |
ASIC stands for application-specific integrated circuit. ASICs are silicon chips designed for a very specific purpose, created to perform a repeated function very effectively – as opposed to general-purpose chips which can perform an endless variety of functions, but less efficiently (such as GPUs or CPUs). ASICs are used in private data centers, public clouds, and connected devices around the world.
Here are a few examples of how ASICs are powering the future of IT today:
- Machine learning: Google’s Tensor Processing Units (TPU) are a type of ASIC designed to run key deep learning algorithms as part of the TensorFlow machine learning framework. Google originally used GPUs and CPUs to train machine learning models, but has since developed a new generation of TPU intended to both train and run the models. TensorFlow is the Google-developed opensource machine learning library which runs best on TPU’s but also runs on both CPUs and GPUs.
- Blockchain: It is the nature of many cryptocurrencies that blocks must be discovered by running hashing algorithms and the difficulty of these discoveries goes up over time as more blocks are found. The increasing difficulty leads to an arms-race of computing power and often results in ASICs overtaking CPUs and GPUs. Bitcoin, for instance, was originally mined on CPUs and GPUs, but around 2013 the first Bitcoin ASICs were produced which could run the SHA-256 hashing algorithm used by Bitcoin far faster and more efficiently than general purpose chips, making CPUs and GPUs obsolete for this function. Today Bitmain is the worldwide leader in blockchain ASIC design, production, and hardware deployment with revenue likely exceeding that of Nvidia in 2017. The market has become so hot that even the world’s largest chip seller, Samsung, is manufacturing ASIC chips for cryptocurrency mining. Bitmain doesn’t only design and produce the hardware, though. The company operates some of the largest datacenters in the world, filled with their own ASICs which they use to mine cryptocurrencies before selling the hardware to re-sellers and other miners. Bitmain is now turning their ASIC expertise to artificial intelligence and appear poised to enter the Machine Learning as a Service (MLaaS) market, to compete with the offerings of companies like AWS and Google.
- IoT “edge” devices: Powering the digital revolution is the circuitry baked into smart devices. IoT devices themselves often use custom-built ASICs to reduce physical space on the chip and function under low energy demands. Additionally, there are IoT kits that connect with cloud platforms like AWS IoT Core, TensorFlow, or Google Cloud – which themselves may run ASICs. In this way, IoT devices use ASICs to gather data with sensors, push that data into existing algorithmic models run on cloud-based ASICs, and send alerts or other outcomes from the model back to the end user or merely feed the model to better predict future outcomes.
- Multi-cloud: Enterprise IT, which powers everything from social media to sporting events to ATMs, must be viewed holistically as a multi-cloud environment. Digital businesses today rely on a mix of public cloud, private cloud, and on-premises hardware. As part of this environment, ASICs can sit either in on-premises or in a cloud environment. ASICs are already available in the multi-cloud through MLaaS and many organizations are already using this technology.
Why are organizations turning to ASICs?
As enterprises embrace the technologies that rely on ASICs, like machine learning and blockchain, ASICs provide benefits including speed and energy efficiency – both of which result in opex cost savings and often contribute to improved innovation.
When CPUs and GPUs can’t cut it at scale, organizations are hiring teams to create custom-purposed circuits. As Doug Burger, distinguished engineer in Microsoft Research’s New Experiences and Technologies (NExT) group, explained to IDG: “I think for applications, the big breakthrough at scale is going to come from non-CPU technologies.” It is ASIC or related tech that will replace CPUs.
KnuEdge is a company that produces military-grade voice recognition and authentication technology. After the company discovered they could not achieve the performance needed with general purpose hardware, KnuEdge formed a new team devoted to building ASICs. The result is the KNUPATH LambdaFabric processor – designed specifically for fast, efficient, and accurate voice recognition.
To better understand the value of ASIC technology, we can look to a recent study from UC San Diego which found Total Cost of Ownership (TCO) for ASIC clouds greatly outperforms GPUs and CPUs for applications such as deep learning, video transcoding, and cryptocurrency mining. From the study: “ASIC Clouds outperform CPU Clouds’ TCO per operations per second (ops/s) by 6,270, 704, and 8,695 times for Bitcoin, Litecoin, and video transcoding, respectively. ASIC Clouds outperform GPU Clouds’ TCO per ops/s by 1,057, 155, and 199 times for Bitcoin, Litecoin, and deep learning, respectively.”
Risks with developing or owning ASIC hardware
While ASICs are great at what they do, they are only great at that one thing. This can make purchasing or building ASICs risky in the event that the single purpose becomes obsolete in the future. To mitigate this risk, some companies are turning to FPGAs (field-programmable gate arrays) which are similar to ASICs but customizable – meaning they gain many of the efficiencies of ASICs without as much commitment to the underlying logic and function. Microsoft’s Bing, for instance, conducted a test to deploy FPGAs and ASICs in one of their data centers in order to improve speed and efficiency for the search engine. The test was a huge success, with 2x improvement in throughput and significant reduction in network latency with the FPGAs. Microsoft found better power-efficiency at scale with ASICs but ended up choosing FPGAs instead because they have the flexibility of being re-programmed later to handle other tasks.
Luckily, choosing between ASIC and FPGA for a capital expense – or creating a team of people to work specifically on a custom solutions – isn’t necessary thanks to cloud technologies enabling rented space on other organization’s machines. Though, as one might expect, the long-term cost savings may be mitigated by relying exclusively on a cloud provider for this technology. For now, the safest move for companies looking to get involved in ASIC-reliant technology is to start on the cloud and consider moving the capability in-house after it is proven to have staying power.
The future of ASICs is in the multi-cloud
ASICs are powering digital transformation and starting to play a pivotal role in data centers, whether private or public. For savvy IT leaders today, the question isn’t if they should use ASICs (or FPGAs), but how best to integrate this technology with conventional CPUs and GPUs within the multi-cloud environment and how best to manage the costs through the software development and production deployment lifecycle.
Taking risks such as that of technological obsolescence is often the burden of those companies that disrupt in the digital economy. Developing custom ASIC chips might only be possible for well-funded projects that rely on cutting-edge technology – and might be the only option for digital leaders to stay at the forefront of their markets. | <urn:uuid:f606a2c4-2d46-4636-9692-945d78bb49da> | CC-MAIN-2022-40 | https://www.cio.com/article/228545/how-application-specific-integrated-circuits-are-powering-the-future-of-it-today.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030334912.28/warc/CC-MAIN-20220926144455-20220926174455-00327.warc.gz | en | 0.942532 | 1,534 | 3.515625 | 4 |
University of Calgary Researchers Advance Biometric Security
The University of Calgary’s Biometric Technologies Laboratory has developed a technology that will allow security systems for merging biometric data such as facial scans, fingerprints, and even the shape of the face, to come up with a learning system that acts like the human mind to make decisions.
By comparing and studying different biometric patterns to make a decision, the system is more accurate and the recognition process is greatly improved. PhysOrg reports that the technology is still algorithm-based but it can learn new patterns along the way. For example, when an error occurs due to a bad sample, the system can instead switch to behavioral patterns or facial recognition to verify a person’s identity.
“Our goal is to improve accuracy and as a result improve the recognition process,” says Gavrilova, a professor in the Faculty of Science. “We looked at it not just as a mathematical algorithm, but as an intelligent decision making process and the way a person will make a decision.”
The algorithm learns new biometric patterns and associates information from different data sets, allowing combined information, such as fingerprint, voice, gait or facial features, instead of relying on a single set of measurements.
Biometric information is becoming more available and common nowadays. It is no longer unusual to find one or two sets of biometric data being implanted on drivers’ licenses, passports and other forms of identification. The new security system being developed can interpret biometric data taken from different databases and use that to formulate an intelligent decision.
When it comes to security measures and needs, an advanced system which is able to think on its own and make judgment calls, becomes flexible especially in environments that frequently undergo a lot of changes. The advanced security system will also be able to train itself to highlight important features and then store these for future use. | <urn:uuid:973af8ed-1948-4de6-a87c-42317e4f7969> | CC-MAIN-2022-40 | https://www.biometricupdate.com/201206/university-of-calgary-researches-advance-biometric-security | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335059.31/warc/CC-MAIN-20220927225413-20220928015413-00327.warc.gz | en | 0.946168 | 383 | 2.703125 | 3 |
The world of web hosting has looked pretty simple for most of the time within the last two decades. Most of the terms and niche hosting services were easy to explain even to someone who had very limited technology awareness. Most of the IT Hosting and computing services used to be delivered from physical, bare-metal servers or from virtualized environments (virtual hosts) that worked on top of a stand-alone physical servers. Compute clusters and other form of networking models of computer systems have been around for long time, but were immature as technologies, lacked automation and most of them were infrastructure models build for internet use. All this changed since 2009.
The Good Old Dedicated Server Hosting
10 years ago when any businesses needed to run a resource demanding website or computer program, they used to buy powerful physical servers. If a project required a huge amount of computing resources, companies used any form of clustering (also known as “Grid computing”) of physical servers (or on virtual machines wherever possible). Some enterprise computer systems also used external storage.
Most processing work and storage service have always been delivered from stand-alone physical servers. Therefore terms like “Dedicated Hosting”, “Dedicated Server”, “Dedicated Hosting Services”, “Dedicated Server Hosting” or any other related phrases have always meant a physical server used by one client. Terms like “Bare-Metal Server” or even “physical server” were not in use or were not popular in the industry.
Over the course of the massive transformation of the traditional economy, as we had for for decades, into economy based on computing technologies (Digital Economy), and with the introduction of an Internet in which appliances, and even objects, require network connectivity to function and exchange data (“Internet of things”), the concept of building and operating IT Infrastructures has changed. Computer virtualization technologies have beed transformed into Cloud Computing (Network model of computer systems, in which the infrastructure used for processing operations is physically separated from the data storage appliances).
Since 2004 the use of physical Dedicated Servers for hosting websites or any kind of software programs has declined sharply. The graph featured below represent the drop in Google searches for Dedicated Hosting since 2004.
While the number of physical servers used for web hosting has increased significantly within the last decade, the number of physical dedicated servers used by companies and individuals decreased as percentage of the total IT installations. Since 2004, the industry of provisioning dedicated (guaranteed) computing resources has been transformed from “Dedicated Hosting” (Dedicated Servers) to “Virtual Dedicated Servers”.
The Rise of Cloud Computing?
The Cloud computing software and infrastructure automation platforms have been maturing since 2004, the use of physical dedicated servers for hosting applications and provisioning IT services dropped even further. The launch of major compute infrastructure clouds has totally revamped the IT Hosting industry.
Amazon Elastic Compute Cloud (EC2)
EC2 was one of the fist enterprise infrastructure service providers to launch compute Cloud service available for public use. Amazon launched its cloud services in 2006. At the beginning EC2 was just a large physical compute infrastructure used for provisioning of Virtual Machines (something what is called today Virtual Dedicated Servers) created with Xen virtualization technology.
A user was able to create, launch, and terminate server-instances as needed, paying by the hour for active servers, where the term “Elastic” came from. At the end of 2010, Amazon transformed EC2 into Amazon Web Services (AWS). However, EC2 and later AWS used to deal with flexibility. They allowed users to scale up and down their computing instances and used to offer flexible contractual terms. EC2 was not created as a computing infrastructure to offer genuine Cloud computing services such as High Availability, Failover or Load-balancing.
VMware vCloud Air
In 2009 VMware was one of the fast major virtualization software provider that created a Infrastructure Cloud service, named vCloud Air. vCloud Air is a public cloud computing service built on Vmsware’s virtualization technology vSphere. It has been designed as “Infrastructure-as-a-Service” (IaaS) subscription IT hosting service. It offers “Dedicated Cloud”, “Virtual Private Cloud”, “Cloud Disaster Recovery”. It offered a pay-as-you-go service named “Virtual Private Cloud OnDemand”.
Another major Compute Cloud is Microsoft Azure. Microsoft arrived late on the Cloud. The company has been long struggling with Cloud computing. Azure was first announced in October 2008. It took Microsoft 14 months to launch it. Azure started in February 2010 as Windows Azure and was subsequently renamed to Microsoft Azure.
It is a cloud Computing platform and infrastructure services for building, deploying, and managing Windows OS applications and services through a network of Microsoft-managed data centers. Azure provides Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS) and Infrastructure-as-a-Service (IaaS) services and supports various tools and programming frameworks.
Find our more about how the dedicated hosting and web hosting terminology changed in article “Dedicated Hosting And Cloud Computing Terms, A Knowledge Boost?“ | <urn:uuid:44219b58-b450-4408-89ef-79faa95075bd> | CC-MAIN-2022-40 | https://www.dawhb.com/dedicated-server-hosting-change-cloud/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030336978.73/warc/CC-MAIN-20221001230322-20221002020322-00327.warc.gz | en | 0.94647 | 1,119 | 3.125 | 3 |
This handbook introduces the basic principles and fundamentals of cyber security towards establishing an understanding of how to protect computers from hackers and adversaries. The highly informative subject matter of this handbook, includes various concepts, models, and terminologies along with examples and illustrations to demonstrate substantial technical details of the field. It motivates the readers to exercise better protection and defense mechanisms to deal with attackers and mitigate the situation. This handbook also outlines some of the exciting areas of future research where the existing approaches can be implemented.
Exponential increase in the use of computers as a means of storing and retrieving security-intensive information requires the placement of adequate security measures to safeguard the entire computing and communication scenario. With the advent of the Internet and its underlying technologies, information security aspects are becoming a prime concern towards protecting the networks and the cyber ecosystem from a variety of threats, which is illustrated in this handbook.
This handbook primarily targets professionals in security, privacy, and trust to use and improve the reliability of businesses in a distributed manner, as well as computer scientists and software developers, who are seeking to carry out research and develop software in information and cyber security. Researchers and advanced-level students in computer science will also benefit from this reference.
- Book Title - Handbook of Computer Networks and Cyber Security
- Book Subtitle - Principles and Paradigms
- Editors - Brij B. Gupta, Gregorio Martinez Perez, Dharma P. Agrawal, Deepak Gupta
- DOI - https://doi.org/10.1007/978-3-030-22277-2
- Copyright Information - Springer Nature Switzerland AG 2020
- Publisher Name - Springer, Cham
- eBook Packages - Computer Science Computer Science (R0)
- Hardcover ISBN - 978-3-030-22276-5
- Softcover ISBN - 978-3-030-22279-6
- eBook ISBN - 978-3-030-22277-2
- Edition Number - 1
- Number of Pages - XX, 959
- Number of Illustrations - 59 b/w illustrations, 197 illustrations in color
- Topics - Security Computer Communication Networks Information Systems and Communication Service Artificial Intelligence
Read and Download the book | <urn:uuid:281822e8-a1e2-4611-91af-3ac45e03a978> | CC-MAIN-2022-40 | https://guptadeepak.com/handbook-of-computer-networks-and-cyber-security/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337404.30/warc/CC-MAIN-20221003070342-20221003100342-00327.warc.gz | en | 0.817548 | 458 | 2.875 | 3 |
30 9月 Edge AI – A More Local and Secure Approach to AI
The ability to access and analyze real-time data is critical for any business enterprise. Advancements in AI-enabled data collection and analysis have provided significant computing benefits which fuel improvements in operational performance.
For traditional AI solutions, the processing and storage of data is executed via a cloud computing architecture. All data is processed in a centralized data center and managed and maintained by cloud service providers. Connections to the cloud environment can be fixed (ethernet, cable, or fiber optic) or by way of a radio frequency (RF) connection that can come in many configurations including 2G, 3G, 4G, LTE, 5G, WiFi, Bluetooth or new low powered wide area network (LPWA) like LoRaWAN.
The cloud-based approach provided the baseline for operations looking to leverage the power of AI and machine learning (ML). However, as operational demands have increased in scope and complexity it has become clear that cloud-based AI has its limitations. There is a critical need to move AI much closer to the assets and processes that are being managed. This advanced technology is edge computing.
What is Edge Computing and how does it relate to AI?
Edge computing refers to the methodology of bringing the data processing closer to the data source, either on the device itself or a server close to the device. Rather than sending a device’s data to a data center or cloud, the edge architecture allows the data to be processed at or near the source.
Edge computing incorporates sensors attached to the targeted hardware or device. The sensors collect the data dictated by the AI algorithms and process that data on the spot. Processing and doing computational analysis on data in real-time enables the support and advancement of real-time applications, especially applications that require instant response time. According to Gartner, in 2018, only 10% of business data was produced and processed “at the edge.” By 2025, that number is estimated to reach 75%.
So, what is エッジAI?
Edge Artificial Intelligence is the same fundamental concept as edge computing. Edge AI technology refers to when the AI’s algorithms are generated and processed locally on the individual devices or machines. This revolutionary method of performing AI on the edge resolves the inability to process data and carry out AI models in real-time. Implementing AI on the edge is crucial for industries that require instant response times; namely, automotive, energy, manufacturing, telecom, and medical.
Here are the main elements of how this trailblazing approach of performing AI on the edge helps increase business performance and potential breakthroughs in technology.
Business Benefits of Edge AI
Security: Edge AI addresses the security issues that arise within a typical cloud platform. The reason is straightforward; AI edge computing operations are executed locally on or near a device, which means prominent data can be processed on the edge instead of sending it to the cloud. Fewer data in the cloud decreases the chances of an online cyber-attack, translating to better security for assets and the data they produce. Recently, many IoT (Internet of Things) systems have implemented successful hybrid solutions, utilizing the cloud for considerably large-sized applications while leveraging edge computing for processing the most salient information locally. Therefore, sensitive information doesn’t need to go through a network, foregoing the risk of being breached.
AI on the edge improves the connectivity of the data. Because the critical data is processed locally, the security and connectivity of the information are more reliable. While a step forward, the drawback of the hybrid solution method is that AI models are developed in the cloud and then pushed to the edge devices. Every edge device receives the same model, thus there is no device personalization to allow for contextualization that reflects the devices unique environment.
Latency/Speed: Edge computing improves the latency of AI applications. Latency is defined as the time it takes to transfer data between its original source and its destination. In cloud AI, data is sent from the source to the cloud and back to the source, which requires large amounts of bandwidth and cloud storage. This method can cause a delay in a device’s response time. Conversely, Edge AI brings computing to the device itself, enabling accelerated information processing and real-time analytics. For example, a slight difference in latency delaying the response time of an autonomous vehicle could be the difference between life or death.
Operational Excellence: Operational excellence refers to the ability to collect massive volumes of raw data about a manufacturing environment at the edge, analyzing it, and figuring out where improvements or changes are needed. Edge AI enables operational excellence by making real-time changes to improve the output in the manufacturing environment. Embedded AI at the edge can identify the bottleneck that is causing the entire facility to slow down.
Edge AI deployment increases the OEE (overall equipment effectiveness) via embedding and training the AI models required to support an effective predictive maintenance program. Edge AI technology reduces costs and increases the productivity of an operation through the advanced data analysis and ML algorithms that support continuous improvement solutions, for example, reduction in variability within the performance of an asset or a process.
Predictive Maintenance: Predictive maintenance is the process of using data analysis to identify anomalies in processes and potential faults in equipment so that future problems can be identified and a solved ahead of time, thereby preventing an operational disruption. Predictive capabilities require advanced data processing due to the extensive number of sensors collecting data from various sources.
Edge AI technology empowers predictive maintenance by analyzing and collecting the data locally instead of sending it to the cloud. Then, after the data has been analyzed on the edge, the analysis results are sent to the cloud. By processing and analyzing the data locally, companies save money on the cost of cloud computing and increase the speed of their audits.
Predictive maintenance brings tangible operational advantages to any enterprise that is asset and process intensive. Those advantages would include:
- Maximizing the lifespan of equipment: When sensors detect anomalies, Edge AI can fix the problem before it becomes irreversible while at the same time reducing MTTR (mean time to recover or restore).
- Improved production and product quality: The improvement of a business’ operational efficiency directly increases the productivity and grade of its work.
- Maximizing ROI: Predictive analytics enables improvement of operations, so businesses don’t leave any value left on the table for what they incur in their operating costs. According to data from the U.S. Department of Energy, deploying a predictive maintenance solution could produce an estimated 30% reduction in maintenance costs, a 75% decrease in breakdowns, and a 45% reduction in downtime. | <urn:uuid:b78542b8-bb4c-4c92-8719-89610c68a75f> | CC-MAIN-2022-40 | https://www.micro.ai/ja/blog/edge-ai-a-more-local-and-secure-approach-to-ai | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337404.30/warc/CC-MAIN-20221003070342-20221003100342-00327.warc.gz | en | 0.919338 | 1,365 | 3.109375 | 3 |
Steganos has discovered that 64% of people would be more worried about someone else having access to all their personal data than they would about the cost of the hardware if their computer was lost or stolen. Despite this, while nearly 70% said that their PC hardware was insured against theft, just 12% used encryption to keep their personal data secure even if files fell into the wrong hands.
Survey participants were asked what data from their hard disk they would be most worried about appearing on the internet: 27% said digital photographs, 27% said personal emails, 19% were worried about their personal documents and just 16% said they had nothing on their hard drive that would bother them. Things that people said they would be embarrassed about being made public include: emails to girlfriend, arguments with boyfriend, photos of drunken nights out with friends, business data, badly written poems (“the people I write about would be offended”), journals, adult browsing history, wills, and family tree research.
Despite the high level of concern about data falling into the wrong hands, few users took adequate precautions. Only 12% of respondents said that they used encryption software to ensure that nobody else could access their personal information, even if the medium it was stored on was lost or stolen. This compares with 73% who use antivirus software, 66% who password protect data, 41% who use a firewall, and 30% who use antispyware.
The primary reason for not using encryption was, according to 41%, simply that they had never considered the possibility, whereas 31% said they wouldn’t know where to start. Only 7% of people said that they thought encryption software would be too complicated to use and 10% thought it would be expensive. As many as 20% of people conceded that they knew they should use encryption, but they hadn’t got around to it yet. (Users were allowed to pick up to three reasons).
Aston Fallen, managing director for Steganos said: “This research clearly demonstrates that people understand that the data stored on their computers is far more valuable and important than the actual computer itself. It highlights the need for affordable, easy to use encryption software, such as Steganos Safe 8, that enables everybody to ensure that no matter what happens to their computer, their private data remains secure and hidden from prying eyes.” | <urn:uuid:94829384-1edb-4835-97c5-3e30ce264b4e> | CC-MAIN-2022-40 | https://it-observer.com/home-pc-users-fear-loss-their-personal-data.html | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030337680.35/warc/CC-MAIN-20221005234659-20221006024659-00327.warc.gz | en | 0.983774 | 484 | 2.546875 | 3 |
Interconnect Devices: Repeater, Bridge, Switch, Router
This page contains the following divisions:
- “Making The Connection”
This is a poem by Joseph Bardwell that expresses a general thought about the confusion in the interconnect marketplace.
- The Interconnect Box
This discussion explains some of the general history of interconnect technology and lays the groundwork for a perspective on repeaters, bridges, switches, routers, and gateways.
- Interconnect Device Topics
Here is the listing of the sub-topics in the section
A poem by Joseph Bardwell
Sometimes it amazes me
that routers work at Layer 3
when switches very will could do
the job at simply Layer 2
But switches work at Layer 3
Oh, how confusing this can be
When bridges work at Layer 2
and routers can be bridges too!
And when you hope there’d be no more
you find a switch at Layer 4
So Layer 4, and 2, and 3
imply OSI conformity
But these are simply building blocks
in what we’ll call an “Interconnect Box”
Years ago, in the early days of desktop computing, an engineer would have learned about interconnect products with some general differentiation’s like these:
…but, bridges became more sophisticated. They could translate between Ethernet and Token-Ring networks and they supported multiple ports; not just two connections. Bridges were able to filter traffic on a selective basis through new configuration options.
…routers became more sophisticated. They supported much more than simple IP routing and they also had the ability to filter traffic on a selective basis. Something had to be done to clarify the definition of bridging versus routing. In the late 1980’s and early 1990’s we learned differentiation’s like this:
|The distinction between devices is based on their function in the OSI Model.
And now let’s introduce the concept of switches! The switch takes the functions of the repeater and the bridge and combine them in clever ways to create a multi-port interconnect box that provides wonderful interconnectivity but challenges network protocol analyzer engineers. And let’s complicate things by, essentially, merging the bridging and routing functions into a single box from Cisco, Bay Networks, 3Com, Thomflex, or other vendors.
That’s why we can think of that ‘box’ in the wiring closet as, simply, “The Interconnect Box”. Does the vendor call it a Router? Is it a “Layer Four Switch”? How about a “Brouter”; and what about the frame forwarding functionality that’s inherent in a file server or a Unix box running the “routed” routing daemon?
This section of the compendium discusses the various forms and functions of many different types of interconnect box technologies that you may encounter. This section is by no means comprehensive – there’s just too much information to present. We are covering the core technologies and we’ll continue to update this section as we discover additional interesting ways that vendors discover to make forwarding decisions.
This section of the compendium includes the following sub-topics that discuss various elements of interconnection technology: | <urn:uuid:cce1b21e-f2b1-4851-8eb8-9713437d766e> | CC-MAIN-2022-40 | https://www.liveaction.com/resources/glossary/interconnect-devices-repeater-bridge-switch-router/ | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030338001.99/warc/CC-MAIN-20221007080917-20221007110917-00327.warc.gz | en | 0.914815 | 686 | 2.90625 | 3 |
This article is a good technical overview of DNS that can help you prevent spoofing. This is a cross-post from the EasyDMARC blog, a new KnowBe4 Ventures portfolio company.
What is a DNS record?
A Domain Name System record is a database record used to translate domain names to IP addresses. Also known as a resource record or a DNS query type, it consists of text files stored on DNS servers, which helps users connect their website to the internet.
Well, when you enter a domain name in your browser, a DNS query is sent from your device to a DNS server to confirm if the domain name has an IP address.
DNS records also encompass several syntaxes and commands telling the server how to handle a client request. By knowing the common types of DNS records, you can better understand your network activity.
So, how many types of DNS records are there? Officially, there are about 90 unique types, each corresponding to a different task or request. If these are misconfigured or used incorrectly, it can negatively affect your website’s performance or even indicate DNS spoofing.
Read on to discover the top 8 most common types of DNS record.
An A record is one of the most common types of DNS records. During an IP address lookup, an A record uses the domain name to locate the IPv4 address of the computer hosting the domain name on the internet. The “A” in this record stands for “Address.” When you visit a site like easyDMARC.com, an A record points to an IP address (Version 4).
This implies that a request from your browser to easyDMARC.com is directed to the corresponding IPV4 address. But an A record can do more than link a domain name to an IP address.
Using multiple A records for the same domain provides fallbacks and redundancy. In this case, each domain would have an A record directing users to the same IP address.
An A record conforms with the standard top-level format defined in RFC 1035. Below is an example of an A record format.
An AAAA record is another common DNS record type, and it’s quite similar to an A record. However, an AAAA record points to the IPv6 address of the DNS server rather than the IPv4.
This “Quad A” record allows the DNS client to learn about the IP address of a domain name and then connect to the website. Although less common, it’s experiencing increased popularity due to the wide global adoption of IPv6 addresses.
IPv6 is the latest version of the Internet Protocol address, and it’s longer than version 4. Like an A record, multiple AAAA records can also provide redundancy when used for the same domain.
A Canonical Name or CNAME record is a DNS record that points an alias domain name (a subdomain or different domain) to the canonical or main domain name. A CNAME record is often used to map an alias domain name to the main domain carrying the A or AAAA record.
For instance, a Canonical Name record can direct the web address www.easyDMARC.ca to the main website for the domain, www.easyDMARC.com, provided both domains are owned by the same company or individual
A CNAME record is ideal when your website has multiple subdomains. Each subdomain points to the root domain containing the A or AAAA record.
If your IP address changes, there’s no need to update the CNAME record of your subdomains. Since they all point to the same root domain, only the AAAA or A record for the root domain must be changed.
Below are a few restrictions of using a CNAME record.
- You can’t place the CNAME record in the root domain.
- A CNAME record must always point to another domain name and not an IP address.
- Pointing A CNAME record to another CNAME record is possible but not recommended.
- NS and MX records should never point to a CNAME record.
- A CNAME record should have no other resource record with the same name (A, MX, etc.) except for DNSSEC records like NSEC and RRSIG.
DNS PTR Record
The Pointer or PTR record specifies the domain name associated with a specific IP address. It’s the opposite of an A record, and it’s used in reverse DNS lookup.
A reverse DNS lookup is a process that starts with the IP address and returns with the associated domain name. PTR records store IPv4 addresses with segments in reverse order and reversed order of hexadecimal digits for IPv6 addresses.
A DNS PTR record typically acts as a security and anti-spam tool.
When you send an email, the receiving email server uses the PTR record in the message to check if the sending mail server matches the IP address it claims, thereby verifying the host.
An NS or Nameserver record is a DNS record type that specifies the authoritative DNS server of a given domain or subdomain. It can also indicate which DNS server houses all of the actual zone files or DNS records of a specific domain.
Generally, NS records inform the internet of which particular nameserver or DNS server has the IP address of the requested domain. You won’t be able to load your website without a properly configured NS record. Using multiple nameservers can also increase reliability.
In this case, there’s one primary nameserver and multiple secondary nameservers carrying similar DNS records as the primary server. So when the primary nameserver is down, one of the secondary servers can attend to DNS queries. An NS record can never point to an alias domain or CNAME record.
Here is an example of an NS record:
A Mail Exchange or MX record is a type of DNS record used for email servers. It indicates the email server of an email address domain via the SMTP protocol. Without configuring the DNS MX record, you won’t be able to receive mail from your domain email address.
While some mail providers only have one server, others can have multiple servers. In this case, each server is assigned a priority value to tell the Domain Name System which sequence to contact the servers.
The email server with the lowest value has the highest priority and will be the first point of contact. Servers with a higher value are only contacted if the others are down. However, the DNS balances the workload between email servers with the same priority number. Like NS records, MX records can never point to a CNAME record or alias domain.
The Start of Authority or SOA record is a common type of DNS record that stores crucial information about your DNS zone or domain. It’s used to oversee traffic between primary and secondary nameservers.
An SOA record is an essential element of zone transfers—the process of sharing DNS records between nameservers—and a DNS zone file is invalid without it.
DNS zone files prevent failures when mirrored to secondary servers. During a zone transfer, the DNS relies on the SOA record to identify the source of the zone files (AKA primary nameserver) and for instructions on how the transfer must proceed.
As such, an SOA record has additional information fields, including:
- MNAME – Primary nameserver of the domain or zone.
- RNAME – Nameserver administrator’s email address.
- REFRESH – DNS zone file refresh interval.
- SERIAL – Nameserver’s or zone’s serial number.
- RETRY – Refresh retry interval.
- EXPIRE – No response timeout.
A textual or TXT record is one of the common types of DNS records that contains descriptive, human-readable information. It’s often utilized together with other DNS record types to provide additional information.
A single domain can have multiple TXT records. Some use cases of TXT records are found in services for Domain-based Message Authentication, Reporting, and Conformance (DMARC), Sender Policy Framework (SPF), and DomainKeys Identified Email (DKIM)purposes. Overall, a TXT record can be used to verify domain ownership and prevent spam.
DNS servers use DNS records to map a domain name to its IP address. Although these processes happen in the background, DNS records are essential to the smooth running of a domain’s website or email server.
The common types of DNS records all serve a unique purpose but collectively, they help users keep their websites online without performance issues. If you want to know precisely what DNS records your domain uses, check out the EasyDMARC DNS Records Lookup tool. | <urn:uuid:1d2489dd-4d52-4be7-a42f-a90ba5d5f049> | CC-MAIN-2022-40 | https://blog.knowbe4.com/the-top-8-most-common-types-of-dns-records | null | s3://commoncrawl/crawl-data/CC-MAIN-2022-40/segments/1664030335350.36/warc/CC-MAIN-20220929100506-20220929130506-00527.warc.gz | en | 0.867967 | 1,965 | 3.171875 | 3 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.