Minutes IETF99: saag
minutes-99-saag-00

SAAG Minutes


*** WG reports.

Barry: proposing to close OpenPGP WG. DKG: will be happy to work with
authors if drafts pop up.

Lionel discusses Diameter e2e security needs. Asking for expert help
to revive work on AVP encryption/protection with a JOSE and possibly
COSE-based approach.

Jean Mahoney: please help! The WG cannot do it by itself.

PHB: needs something similar for his Mesh work. Could pull it
out. Willing to help.

Karen: the NTP Security work going into WGLC, please review.

Leif: UTA is closing unless people come up with proposals.

Hannes describes work on FUD (firmware update) and invites people to
join the list.

Sam Weiler: W3C is looking for security experts for help on security
and privacy. If interested, contact him.


*** Kenny Paterson on Post-Quantum Crypto

Richard Barnes: agrees that CFRG should work on combining classical
and PQ algorithms. Could IETF or CFRG support the NIST process? Kenny:
maybe attend the workshops and provide requirements.

Paul Hoffman: keys expected to be at least 10x larger, e.g. 15k
bits. Mentions his classical-to-PQ draft at CFRG. We are going to want
to know when to make the transition.

PHB: mentions Lamport signatures, we could bake them in now. We need
to have a deployment/migration plan for PQC.

Steven F.: we could avoid long-term encrypted things (identifiers?) so
they are not vulnerable in the future. What is the IPR situation?

Kenny: not sure.

Quynh (NIST): we require IPR to be declared, but not given up. But
practically will prefer unencumbered.

Nick Sullivan: there are trade-offs between key sizes and computation.

Tobias: the time-lines are vague. Devices and protocols have lifetimes
of 10 years or more. We need to build in ability to change
algorithms. This is a requirement we need to consider now.



*** Volker Birk presents PEP.

?? invites the team to join UTA.

Volker invites the team to attend their bar BOF this evening.

Paul Wouters: is this an effort to strengthen protocols? Federated
identity? I am confused.

Volker: this is not centralized. They do not manage identities, they
only map identities between different identity providers.

Martin Thomson: draft is hard to read. Did not see an
architecture. The presentation was helpful to understand. Please talk
to many people and ask for help. People would like to hear concrete
goals, what the architecture looks like, what are problems with
existing protocols. The project is very massive. There is difference
between shipping code and achieving interoperability.

PHB: I like the activity. There are lots of secure e2e protocols out
there. All are siloed (e.g. Signal). Also, all are sold as solutions
around government coercion. However even if crypto is e2e-secure, the
application itself is a point of coercion. Must have open source that
can be audited.



*** Dmitry Belyavskiy presents the Certificate Limitation Profile.

Martin Thomson: chairs, please include a link to the draft. This has a
lot of overlap with existing mechanisms around pinning, CT.

Richard Barnes: there are app vendors that have their own CRL sets,
those are proprietary but overlap with this proposal. What is the
value in standardizing it?

Dmitry: we are trying to formalize the syntax.

DKG: mentions P11glue, which has similar functionality. There is merit
in this kind of work because subtlety in the rules, but some people
are already doing it.

Rich Salz: would be interested in seeing a common data format.

Martin: they would be willing to participate.

Richard: need a notion of who is trusted to provide this information.

Eric Rescorla: two problems to solve. Policy: which anchors do you
trust. Distribution: need compression to deal with numerous
revocations. Hardcoding is a problem, it makes it hard for a new CA to
enter the market.

Kathleen: will update the agenda with the draft link.


*** Open Mic

Steven: querying the ADs re: their position re: rechartering TLS WG or
new potential proposals.

Kathleen: maybe opportunity for something that's data center
specific. The WG decides.

Steven: is this within the current charter.

Kathleen: needs to reread the charter.

EKR as individual: does not believe draft-green would be in charter.

Yaron: are quantum security models (QROM) usable/reliable and a good
base for NIST's process?

Kenny: yes.