*PKI Labs Conference Call*
May 20, 2002

*Attendees*

Neal McBurnett (convener) - Internet2

Bob Brentrup - Dartmouth
Sean Smith - Dartmouth

Eric Norman - Wisconsin

Peter Honeyman - Michigan
Olga Kornievskaia - Michigan
Carl Ellison - Intel
Bob Morgan - Washington
Renee Frost - Internet2

Ben Chinowsky (scribe) - Internet2

*Discussion*

After discussing corrections to the April 8 minutes, the group noted the appearance of the second edition of *Network Security*, by Kaufman, Perlman, and Speciner. Sean said that the book nicely covers both academic and real-world aspects of PKI, but fails to address the John Wilson problem and non-X.509 alternatives. The group reviewed a single action item:

[8-April - Carl and Eric will whiteboard Eric's SDSI work at the PKI Research Workshop; thereafter Carl will harass Eric into making further progress on documenting this work.]
Done. Eric summed up the discussion: Carl confirmed that he might be on to something, but also pointed out that he needs to think about the fact that there might be too many privileges to give them all names, raising "a question of how you parameterize priviliges." Carl pointed out that when using named groups you can change dictionaries when you cross organizational boundaries, but there's no way to change authorization dictionaries; translating authorizations is more like translating a programming language. Bob Morgan noted that "in my part of the real world it seems like exchange of authorization information among multiple organizations is a requirement that many folks are realizing these days." XACML and many digital rights management languages are trying to standardize sets of authorization expressions. Carl suggested that both names and conditionals -- "parameterized named things" -- will be necessary, and Eric noted Lisp and Prolog as models.

Neal noted that VeriSign seems to be moving toward deploying DNSSEC; apparently they are looking to DNSSEC to help them cope with large zone files and with administrative issues arising from having to deal with everyone who manages a domain. Trials are planned for this autumn. Eric expressed skepticism about the value of DNSSEC, but there was general agreement that DNSSEC is well worth deploying: if DNS becomes secure, all the other stuff we build on top of it becomes secure too. There was also general agreement that there is reason for concern about DNS getting overloaded by too many such things being built. In particular, Bob Morgan noted that building on DNS to do key management had been suggested at the last IETF; he's not sure it would be wise to depend on DNS to bear the weight of all this additional trust.

Bob Brentrup noted that Elcomsoft (www.elcomsoft.com) now has a product that removes password security from Acrobat files; Peter described this as "a source of great fun for sure."

At Wisconsin, doctors and nurses are now using S/MIME to exchange patient information. Carl pointed out that S/MIME doesn't support restricting information to particular individuals within a medical office. Bob Morgan noted that most medical information security designs are driven by HIPAA, which insists on encryption on the network, but isn't concerned with what happens behind the firewall on either end -- in fact many designs don't even care about signatures, just encryption. Neal noted the contrast between this and HEPKI-TAG's S/MIME project, which has encountered much interest in signatures but little interest in encryption.

Eric reported that Ross Anderson spoke recently at Madison and that he had related a story about a student who supposedly found a bug in the 4758. Sean reiterated that the 4758 has not been cracked (see www.cs.dartmouth.edu/~pkilab/4758.shtml), but noted that two interesting hardware attacks were described last week at Open.

Bob Brentrup noted that the Dartmouth Lab's grant hires start next week. An X.509 PKI is being readied for testing with real users this summer. Sean gave a short update on the progress of his many students, noting that Alex has been written up in Wired (see www.wired.com/news/privacy/0,1848,51917,00.html).

The group discussed past and future PKI Research Workshops. Ben is working on a writeup for the final proceedings of the 1st Annual PKI Research Workshop. [AI] All will send Sean recommendations for papers from the 1st Annual PKI Research Workshop that should be included in a special journal issue Sean is planning with Ravi Sandhu. [AI] All who have suggestions on the membership of, or division of tasks within, the program committee for the 2nd Annual PKI Research Workshop, or who are interested in serving on it, will contact Sean. [AI] Sean will look into the sponsorship situation for the 2nd Annual PKI Research Workshop.

The next PKI Labs meeting will take place on June 10 at 4pm Eastern, per the normal schedule.

*Action Items*

[AI] 20-May - All will send Sean recommendations for papers from the 1st Annual PKI Research Workshop that should be included in a special journal issue Sean is planning with Ravi Sandhu.
[AI] 20-May - All who have suggestions on the membership of, or division of tasks within, the program committee for the 2nd Annual PKI Research Workshop, or who are interested in serving on it, will contact Sean.
[AI] 20-May - Sean will look into the sponsorship situation for the 2nd Annual PKI Research Workshop.
[AI] 18-December - Bob Moskowitz and Carl will further discuss ways of increasing the user-friendliness of using raw public keys to set up devices.
[AI] 13-August - Bob Moskowitz will forward the list email on PKI work at Fannie Mae.
[AI] 4-June - Bob Moskowitz will send the list information on Federal work related to attribute certs.