CoRE Virtual interim - 2023-06-21 - 14:00-15:30 UTC

Chairs:

Remote instructions

Material:
https://datatracker.ietf.org/meeting/interim-2023-core-10/session/core
Meetecho:
https://meetings.conf.meetecho.com/interim/?short=1e4a7427-d659-483c-b8bb-ea463992572f

Notes: https://notes.ietf.org/notes-ietf-interim-2023-core-10-core
Zulip: https://zulip.ietf.org/#narrow/stream/21-core

Minute takers: Marco Tiloca, Rikard Höglund
Chat Monitor: Marco Tiloca

Agenda

Note Well

Remember that the note well applies, for IPR but also for WG
processes
and code of conduct. Please be nice to each other.

Jabber & Minutes / Agenda bashing

CORECONF (Carsten Bormann)

Objective: look at status we'll achieve for IETF 117 hallway meetings

Presented slides:
https://datatracker.ietf.org/meeting/interim-2023-core-10/materials/slides-interim-2023-core-10-sessa-coreconf-slides-00.pdf

CB presenting

CB (p1): summarizing the state of the 4 CORECONF documents.

Starting with -core-sid

CB (p2): Recent/ongoing updates on SID allocation triggered by Rob
Wilton's DISCUSS.
CB (p3): The main remaining thing is to run a patched up PYANG to
generate the SID file example. All relevant SIDs should be covered, but
it is not catastrophic if more need to be added manually later on.
CB (p4): Examples of SIDs for RPCs/actions from version -20.
CB (p5): -core-sid defines a process to be used by default. An addition
will clarify that alternative processes are acceptable, see, e.g.,
draft-toutain-lpwan-sid-allocation
CB (p6): Presenting timeline for planned updates. A new WGLC can happen
around week 27, right after the next interim meeting.
MT: The timeline looks good. The original plan didn't include another
WGLC, but it looks appropriate.
CB: The acting responsible AD decided that it's better to fully push
back to the WG, and doing another WGLC is reasonable.
MT: Good to have one more. We just have to move further back the status
in the Datatracker, as it now suggests that an actually last WGLC
happened. I would like to confirm with Jaime who is Shepherd for this
document.
CB: The WGLC period can be 1 or 2 weeks, 2 weeks may be feasible and
better.
MT: 2 weeks sounds good.

Moving on to -core-comi

CB (p7): Discussed during the IETF meeting in London. The conclusion was
that the draft is too complicated. Christian had ideas for significant
simplifications. We have now converged on the non-radical solution (keep
GET/PUT/DELETE for the whole-datastore).
CB (p8): The just posted version -13 can be discussed on the mailing
list. Open question: Do the PYANG changes for core-sid impact this work?
The plan is to have a WGLC during week 27, right after the next interim
meeting or earlier.
MT: Good to run the two WGLCs in parallel. The shepherd is Michael
Richardson.

CB: The next steps can be about finishing -yang-library. Another point
is that the actual YANG data is in a text-based format. It would be good
to be able to switch a YANG model to a more efficient representation
(design effort needed).

CoAP: Non-traditional response forms (Carsten Bormann)

Objectives:

MT: As I replied on the mailing list (see above), I think this is
useful.

CB: Unifying approach seems like a good thing, based on feedback.
CB: Writing up a new big picture/architecture is one point to do,
covering these forms of responses and presence of one-to-many
communication not fully imagined when originally working on CoAP. This
part about the architecture can be informative.
CB: This document also defines multiple CoAP options. Not aware of any
implementations of those. These options illustrate possible uses of the
architecture above.
CB: In the best case, both parts will be done quick, so we can point
from this document to the normative ones (with the options). Or we add
the illustrative options with a loud caveat that they are to be finally
defined in other documents.

CB: Next steps is to discuss things further. Including the topic of
terminology.

RH: I think this unifying approach makes sense, I'll catch up for the
discussion.

MT: I don't find the terminology confusing, for instance multicast
responses is clear to me. It means that the response is sent over
multicast, irrespective of the kind of request and how the request was
sent.
MT: I think the overall strategy makes sense. The architecture can be
informative, while the options are a different thing and can be one/more
normative documents.
MT: Christian had an idea for a further new option, similar in scope to
"Multicast-Timeout" from -groupcomm-proxy: a kind of flag for the client
to indicate to a proxy that the proxy should forward back multiple
responses until further notice (instead of for a specified amount of
time).
RH: I understood the same, but some signaling is also needed for when to
stop.
CB: One idea is to use the Observe option for the termination signaling.

MT: Christian thought about using the Observe option but found some
issue with that, hence the idea for the new option.
MT: A motivating use case is about: a client using Group OSCORE
end-to-end with a group of servers; the client using OSCORE with a proxy
P1 deployed between itself and the servers (see
-core-oscore-capable-proxies); and a proxy P2 deployed between the
client and P1. The client would use Multicast-Timeout with P1 and it
would be encrypted between the client and P1. But then, for things to
work, also P2 has to relay back multiple responses to the client, but,
in the interest of privacy, the client would prefer not to reveal to P2
exactly for how long. Hence, a new option for P2 rather than having
Multicast-Timeout also as outer option to be consumed by P2.

MT: We can continue on the list and have this topic also during the next
interim meeting.

AOB

Tentative items for the next CoRE interim meeting

Activities during IETF 117

CB: Shall we schedule a side-meeting during IETF 117?
MT: Sounds good
RH: I would join
CB: I will have a look for booking a room
MT: Same for me