Attending
Members: Rob Carter, Chris Phillips, Les LaCroix, Marina Krenz, Kevin Hickey, Michael Grady, John Bradley, Chris Phillips, Erik Scott
With: Nicole Roy, Steven Premeau (TAC representative to CACTI), Richard Frovarp (CTAB representative to CACTI), Steve Zoppi
Regrets: Licia Florio, Barry Johnson, Kevin Morooney, Ann West
- Agenda bash/administrivia
- Announcements
We have a bit of a backlog of approvals needed, if you have not approved or requested modification the minutes from June 21, July 19 and August 2, please do so (Nicole)
Rob, Chris, Nicole meeting with Max Pala from CableLabs re: Post-quantum crypto next week, more to come
OpenID Federation discussions on the eduGAIN Slack instance, take a look (Chris)
Italian government has built a federation based on OIDC-Federation spec. So has France. Giuseppe wants to turn this into the EIDAS 2.0 spec.
Request for people to take a look at this over the course of the next month.
In the #oidcfed slack channel, in the following public Slack tenant: edugain.slack.com
- Main Business
- Post-quantum cryptography (Chris)
Quantum cryptanalysis risk/threat to pre-quantum crypto
Chris has a couple powerpoint slides on this
Federations (and IAM more generally) rely on cryptography
There is a lot of “FUD” out there on this stuff
Preparation in IAM/T&I space likely needed in order to be prepared to respond in time when cryptographically-relevant quantum computing arrives
Lots of US federal and other governments’ “three-letter agencies” concerned about this - US Department of Homeland Security PQC playbook, etc.
Potential risk to encrypted stuff at-rest
Operational risk for things like authentication, single sign-on, federations, etc.
Chris recommends thinking about this in the context of business continuity planning
Quantum deadline fuzzy, out in the distance, do not know how soon it is right now
OpenSSH project already supporting some PQC. Good info on openssh.com. my additional link: https://www.ssh.com/academy/cryptography/what-is-quantum-safe-cryptography
Table-topping a quantum-supremacy scenario with cryptographic relevance: What actions would we have to take?
Even if we take actions to be quantum-cryptanalysis-resistant, it may be insufficient. Then what?
A lot of the quantum-resistant algorithms are not drop-in replacements for things like RSA
XML digital signature - risk - unlikely we will get rid of XML before Q-day. Is it possible to update XML to be able to accept these new algos?
Dual signatures as a transition path?
Getting the infrastructure ready to plug in these algorithms is probably a good first step
Other challenges: Hardware optimization for algorithms falls down on legacy devices
Key exchange challenges
What obligations do software architects and federation operators have to prepare? What does that prep look like?
If we can’t do anything now because we don’t know what algos are going to win bake-offs/etc? We can start enumerating our risks at our institutions.
What are the to-dos on this that are actionable? Threat seems clear, but not hearing a clear signal from DHS, NIST, IBM, etc. what counter-measures should be baked into our plan that we don’t have in our plan today? Part of the problem are the “unknown unknowns”? Transition to “known unknowns”.
What questions should we be asking that are meaningful and not going to be responded to in vague / hand-wavey terms?
This is a community awareness piece of the puzzle: Risk-awareness, risk-acceptance.
Actual answer is “we hope there will be technical responses we can deploy when this happens” - the question then is, when those viable technical responses come about? What can we do so that we can get on that bandwagon fast enough? Identifying what you can do to make a response time closer to 12 months rather than 36 months seems like about the only thing you can do now, and we should be advising the community to be prepared in the same way. We have webinars/convening power as a community to help start the hive mind discussion. Creating a forum for the community out of this topic is something we can do much sooner than 12 months from now.
Pluggable algorithms, pluggable key-rotation, pluggable strategy - possible?
Next Meeting: Tuesday, August 30, 2022