text stringlengths 0 1.99k |
|---|
card that supports DES and extended the Kerberos library to use that card |
to perform client-side cryptography and key management. |
To support smart card authentication on Windows, we extended the Windows |
Graphical Identification and Authentication (GINA) to support the Pluggable |
Authentication Method (PAM) and developed a smart card PAM module. |
We built a VFS layer for the smart card's ISO 7816 hierarchical file |
system, allowing smart card files and directories to be navigated and |
updated through a POSIX interface. |
We implemented ISO 7816 IP tunnel endpoints on UNIX and JavaCard, |
implemented TCP and IP on JavaCard, and implemented a web server on top of |
that. |
We designed and implemented a middleware layer that supports remote access |
to physically secure smart cards and extended Kerberos V5 and SSH to use |
that infrastructure for client-side key management and cryptography. |
We built handsome hardware that allows a Palm Pilot to connect to a smart |
card and developed Palm Pilot software to use the smart card as a key |
store. |
We designed and implemented personal secure boot, an extension to GRUB that |
stores software component hashes on a smart card and uses the hashes to |
ensure component integrity at boot time. |
We open-sourced all the software that we developed under a very permissive |
license. |
We also used the funds to support travel and to evangelize the smart card |
way. I joined IFIP WG 8.8 as Co-Vice Chair to help organize smart card |
conferences in Louvain-la-Neuve, Chicago, Bristol, San Jose, Toulouse, and |
Tarragona; ran smart card demos at CCC in a field outside Berlin; and |
taught a graduate seminar in smart cards at Michigan, flying in European |
experts as guest speakers. CITI also made a major showing at HAL 2001 (Me, |
Jim Rees, Charles Antonelli, Andy Adamson, Dros Adamson, Dug Song, Niels |
Provos, Olga Kornievskaia; some CITI sons and lovers tagged along, too) and |
organized a well-attended smart card hacking workshop there. |
--[ 6 - Aftermath (and airing of grievances) |
CITI's burst of smart card research lasted only a few years, for reasons, |
some of them the usual ones, like money drying up and students graduating. |
We had good relationships with our industry partners in the Dallas suburbs, |
but the center of gravity for smart card R&D was in France -- which we |
loved because we were forced to travel to beautiful French cities for |
meetings -- but it was hard to make an impact in that crowd, which was |
excited about explosive growth in SIMs and not that interested in a smart |
card that communicated with TCP/IP and web protocols. (We honestly could |
not understand that.) |
Also, none of us is fluent in French, alas. When we tried to get some |
attention by nominating ourselves for the CARTES '2000 Innovation Award, |
well, the nomination had to be in French. We submitted it in English. You |
can imagine how the French jury responded. (They didn't.) Adding insult to |
injury, when CARTES got around to recognizing an Internet smart card, the |
award went to a group that did its TCP and IP processing off-card, merely |
tunneling application layer payloads through ISO 7816. >:( |
But it wasn't just our big egos. Hardware limitations were a big reason. |
CITI had a vision of the smart card as the trust anchor of a personal |
computing environment and built stepping stones toward that vision: a |
Kerberos single sign-on for UNIX and Windows you could carry in your |
wallet, a UNIX file system for secure storage of personal records, TCP/IP |
for reliable end-to-end communication, even a web server, for what it's |
worth! When we started, conventional wisdom held that all this -- on an |
eight-bit microprocessor running at 3.57 MHz with 200 bytes of RAM -- was |
impossible. |
At that point, it wasn't clear what to do next to achieve our vision: we |
needed more cycles, more storage, more bandwidth. So we waited for smart |
cards to get better but unlike every other computer technology smart cards |
got no faster, no more capable. They never did, really. |
Nonetheless, smart cards did succeed -- wildly, and essentially |
contemporaneously: mobile phone technologies exploded in the new |
millennium, with smart card SIMs the mechanism for low-level mobile phone |
identification and authentication. And it turns out that is all that a |
smart card really needs to do, with high-performance trust management now |
built directly into smartphone silicon. Soon, eSIMs and NFC payments will |
obsolete smart cards altogether. |
--[ 7 - Dodging the Feds |
So back to that tense meeting with the Secret Service. I told them all we |
had learned about the MCard and its vulnerabilities. Our guests were |
businesslike and intelligent, and the briefing was over in an hour. |
Then it was time for a lab tour, the high point of which was the lunchroom. |
As we approached the cowering, trembling vending machines, I asked the head |
fed if he would care for a demo. His enthusiasm was unvarnished: he would |
love a demo! I pulled a white card out of my pocket and offered it with |
outstretched arm. |
He paused. Smiled. |
No demonstration would be necessary. He left us with some Secret Service |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.