text stringlengths 0 2.67k |
|---|
**Autumn Nash:** You probably made it worse. |
**Tim Banks:** We may have. |
**Autumn Nash:** Okay, so what was maybe not just like technical-wise, but what was your coolest job? Because you've had some -- like, what is being a cloud economist? You've got so many layers. Like, we heard about your DevOps career, but what about -- I feel like you've... |
**Tim Banks:** Being a cloud economist is being a principal level SRE, but the thing you're solving for is cost. |
**Autumn Nash:** How do you even get good at that? |
**Tim Banks:** Be the one in charge of saving the company money sometimes before it actually -- before FinOps was an actual practice, right? And again, that's something we did to some extent at Object Rocket, but we owned all our own hardware... Because you should... At Elastic we used AWS and GCP and Azure, to some ex... |
And then also, I really got to see how the sausage was made at AWS, working there as a TAM, and some of the things that they were doing, not only to optimize costs, but to make sure they were still making their money on the backside. And so then when you go and then you compete against AWS, in that aspect, you're kind ... |
Cost optimization - I beat this over and over \[unintelligible 00:33:53.29\] about pretty much any problem. Cost optimization is done on the whiteboard, first and foremost. It's done on planning, it's done in your architecture. You cannot have a cost-inefficient architecture and make up for it with AI, or some like ML ... |
**Autumn Nash:** Because I think that's where a lot of people are right now. You know what I mean? |
**Tim Banks:** Oh, yeah. Because it's easy. |
**Autumn Nash:** People are not sure if they want to go back to on-prem, if they want to own their own hardware, if they want to be in the cloud, they're not sure which cloud... So I feel like your experience is really relevant right now. What was it like when you had three different clouds? Because people are selling ... |
**Tim Banks:** Oh yeah, yeah. Money's always in the data. Always money is in the data, and that's all observability is. That's why AWS charges so much for... |
**Justin Garrison:** CloudWatch. |
**Tim Banks:** CloudWatch. Thank you. You use it because you have to, but nobody really wants to. |
**Justin Garrison:** It's terrible. |
**Tim Banks:** It is the Zookeeper of observability. Oh, that's going to hurt... I'm so sorry, CloudWatch team. That was unfair. That was unfair. |
**Justin Garrison:** Now, you've said something in the past that I wanted to get some background on. I know you've said that the cloud has made us irresponsible. |
**Tim Banks:** Oh, absolutely. |
**Justin Garrison:** I would love to hear more about that. |
**Tim Banks:** So the reason that cloud has made us reckless is because in the before times - and back then when it would take you a couple of weeks to spin up a VM, which consequently is what a lot of these startups still think is how it runs... And it's not like that. But you still have to do forecasting, right? You ... |
**Justin Garrison:** Right. A budget. |
**Tim Banks:** ...before you could get a VM approved. Before they would launch anything for you. That was the thing. There was forms for it, there was a whole way of predicting and forecasting. And when the cloud says "You don't have to do that anymore. You can just spin up, and we'll bill you for it later." "Great." O... |
\[36:21\] There are people that have money... And then you know, when they have so much money and the money don't really mean nothing to them, they just go and do anything with the money, and they get a little reckless with it. Cloud is kind of like that. You have, in theory, or a promise of unlimited resources... So w... |
**Justin Garrison:** Well, and you hide your bad architecture with bigger bills. You're just like "Oh, that was a bad choice" and "We'll just pay for it now." |
**Tim Banks:** Now we're just going to pay more for it, right. And what people say is like "Oh, when you look--" And I've done this before when I've done contract negotiation. I'll look at your bill, I'm like, your bill is hundreds of thousands of dollars a month. And it's going to keep being hundreds of thousands of d... |
**Justin Garrison:** Without the benefits. Yeah. |
**Tim Banks:** Yeah. But instead, you're leasing it and you can't even claim it as depreciating assets on your taxes. |
**Autumn Nash:** You know what's wild? For a while, we had legacy tech where people would make these huge commitments and they would get stuck, and then they'd take 20 years to move off of COBOL. And then it got to the point where now people are lift and shifting all these things, and they keep going back and forth bet... |
**Tim Banks:** That's exactly how you should be doing it. |
**Autumn Nash:** But there's so many different options, and I feel like people are -- it's almost like nobody knows what the right answer is. |
**Tim Banks:** So my take on it is, I think more people are going to buy compute than are going to lease compute. If you have the money to do it, it is always more cost economic to buy your compute, especially if capacity is a concern. |
**Autumn Nash:** What about the fact that you can scale in the cloud, and you don't have to wait for the hardware, you don't have to... |
**Tim Banks:** Have you ever tried to provision instances in Azure? |
**Autumn Nash:** No. We don't play with that here... |
**Tim Banks:** You want to roll those dice of whether it's going to be there when you need it? |
**Autumn Nash:** Oh, that's another thing I wanted to ask you about... What was your experience like when you were using three different clouds? Because I think that people talk about multi-cloud like it's the answer to everything... What was doing that in prod like? |
**Tim Banks:** It sucks. And here's why. So in theory, it's like, oh, I could just write one Terraform codebase and replicate it across. You can do that more now, but you couldn't back then, and \[unintelligible 00:39:21.19\] The observability stacks were all different before OTel existed, before the... Oh God, don't e... |
**Autumn Nash:** That's what I'm wondering... How do you get real observability and like context into what you're even doing with three different companies? |
**Tim Banks:** You have to use a third party tool. The one that I really liked - this is not a paid endorsement, but the one I really liked was CloudHealth. CloudAbility to some extent, but CloudHealth was really good about ingesting this stuff and being able to visualize it from a resource usage and a cost standpoint.... |
**Autumn Nash:** You know what I think is wild? |
**Tim Banks:** What's that? |
**Autumn Nash:** \[40:11\] With all of the multi-cloud and all that good stuff, and then AI, I think the people that are really going to make money in the next couple of years are going to be the people that make really good tools for multi-cloud, or things to lift and shift... Because that's what people seem to be doi... |
**Tim Banks:** Sold the shovels. Yeah, yeah. |
**Autumn Nash:** So I wonder if it's going to be the people that are making it easy to lift and shift from on-prem to like cloud, and the people that sell the chips. |
**Tim Banks:** I think that ship has already sailed. I think it is now the default that someone is going to run -- when you reach a certain size, right? Whatever your certain size of operations is... That you are going to have a presence in more than one provider. Economically, it makes the most sense to buy compute, v... |
And then I will get some AWS essentially equivalent savings plan for any kind of elastic, or unpredictable scale. |
My dev environments, I'm going to buy. My steady state compute batch jobs, all those kinds of things that I know I need every day, day in, day out, I'm going to buy. I'm going to buy 90% of my GPU fleet. You know why? Because I can sell it when I'm done with it. |
But also, not every GPU needs to be newer, faster, whatever like that. The vast majority of the work we do, the GPUs are just -- they have more memory, they have some abilities, but for the most part, you can still get by with 90% of the stuff you're doing with whatever you buy today. Or whatever you bought last year. |
**Autumn Nash:** What happens when all this stuff goes on sale though? You know what I mean? |
**Justin Garrison:** I'm so excited for all the GPUs to hit the secondhand market. It's going to be so great. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.