Master at: https://notes.ietf.org/notes-ietf-124-anima
IETF124 is a hybrid IETF meeting.
Please familiarize yourself with the new&changed procedures for
participating in-person or remote:
https://www.ietf.org/how/meetings/preparation/
ANIMA will have one 2 hour session,
chaired by Sheng Jiang and Toerless Eckert, Wednesday, Session I, "Notre
Dame", 09:30 - 11:30 Meeting Time, 14:30 - 16:30 UTC
Number of participants on meetecho: 24 In the room: 11 (8 during the
later half of the meeting)
Notetakers: Toerless, Sheng, ... , Esko, collaborative editing by WG
members.
Presenter: Sheng Jiang (remote), Toerless Eckert (local)
Time: 10 minutes
(see slides)
RFC Editor Cluster C528
Updates on WG documents without own slots
Action: Lets tell AD to push BRSKI-PRM into the RFC-editor...
Draft: draft-ietf-anima-constrained-voucher-29 (was -28 at IETF123)
Draft: draft-ietf-anima-constrained-join-proxy-18 (was -17 at IETF123)
Presenter: Esko Dijk (ED)
Time: 20 min
Toerless also supporting ask to make dependency against new CORE-WG
.well-known path shortening as standard dependency. BRSKI does not
depend on it a lot, but given how the new work was done because of BRSKI
as trigger and does not hold up othrer work, it would be good if BRSKI
also uses it. And it does not delay other work in ANIMA.
Toerless: Q: Any specific thoughts about where beyond BRSKI the JPY
proxy could be useful ?
Esko: Not quite sure, the Thread (www.threadgroup.org) proxy is also
quite generic. SRP update protocol is an example of a protocol that can
run over the relay (over UDP).
Michael: Abandon JPY and use what Thread does ?
Esko: No, that's too thread-specific. But implemented in legacy Thread
devices so that's a reason for keeping using it.
Thread spec is a not-public license, but once you accept Thread license,
you can read the Thread spec. Thread relay/join-proxy is only stateless
and uses CoAP.
In ANIMA we stopped using CoAP, and use JPY protocol instead, because of
certain benefits.
Michael: Coap is just overhead in this case, and risks misuse of
flags/options/...
"me.arpa" issue discuss: Michael: does it work through proxy ?
Answer: yes if the intention is to point the client towards the proxy to
access the resource, instead of towards the origin server.
Discussion was not quite successful in identifying exactly in what
circumstances this may be or may not be working. me.arpa will be
discussed when possible in a next core-wg interim.
Draft: draft-ietf-anima-brski-discovery-09 (was -08 at IETF123)
Presenter: Toerless Eckert
Time: 10 min
Action Toerless: Send note about the RFC 6690 update to core-wg.
Esko: concerns about the document complexity for implementers - they
only needs variation string.
Toerless thoughts: write section for implementors to give explicit
guidance.
so going to write implementer considerations.
TBD - will work on that... (toerless)
Agreed to look if we can present the table with the strings (the main
result) first. IANA details follow later, in separate sections, that
developers don't have to read/understand.
Toerless+Esko: note that a person defining a new protocol variation (and
testing it) does have to understand all the details.
But an implementer building a Pledge or Join Proxy or Registrar does not
have to know the full IANA background / registries.
Draft: draft-ietf-anima-rfc8366bis-16 (was -13 at IETF123)
Presenter: Michael Richardson
Time: 5 Min
Discussion about tooling issues to generate SID values (for cBRSKI) for
BRSKI - likely to be part of generating those SID for all existing IETF
yang models. But in result, rfc8366bis might be the first RFC to come
out with pre-allocated SID - making adoption a lot easier.
Current concern with broken tools is that these SID allocations may need
to be done by personnel that is not technical experts, and the broken
tools may flag errors where there are none. And this would hold up the
process of allocating SIDs.
While contributors like Michael etc. are good developers, they do not
want to take official responsibility to fix problems in tools used then
by all of IETF for this process..
IANA allocating blocks works fine. Allocating individual leaves to nodes
is a tooling issue.
rfc9254 - iana SID files, not quite a registry.
Next steps on this tooling issue: Hope for Mahesh to bring it up at some
appropriate place. Not clear yet, who has ownership.
Mahesh: IANA will likely bring the issue, but unclear.
Esko: Would very much like to have an automated tool, no more
hand-generation of SID files.
Esko also will submit a final PR on the -17 version to be published soon
by Michael.
Draft: draft-ietf-anima-registrar-considerations-00
Draft: draft-ietf-anima-masa-considerations-00
Presenter: Michael Richardson
Time: 15 min
(see slides - no discussion in the meeting)
Draft: draft-ietf-anima-constrained-grasp-01
Presenter: Longwei Zhu (remote)
Time: 10 Min
Time line plan (Toerless), answer: wanting to have draft update and
demonstratable code for hackathon by IETF125
Esko: there is an RFC allowing to add status of implementations to draft
to help tracking the implementations
This is: https://www.rfc-editor.org/rfc/rfc7942
Draft: draft-mzsg-rtgwg-agent-cross-device-comm-framework
Presenter: Jianwei Mao (Remote)
Time: 15 minutes
Toerless: WOuld be useful to write a draft summarizing the value of ANI
for AI-agents in the network.
Should make it easier to then think of the next steps.
Mahesh: redirect agent2agent@ietf.org mailing list - so we have a good
dispatch.
Sheng: promote re-use of grasp.
Bing: WOuld prefer for agent2agent discussion to happen in anima.
TBD Toerless registrar considerations: Check if the registrar PKI
recommendations are consisten with 8994 discussions about PKI and point
to it for the relevant details.