[rmcat] IETF 102 RMCAT minutes

[Colin Perkins <[e-mail]>]

Draft minutes from RMCAT at IETF 102 in Montreal follow, and have been uploaded to the data tracker. Please send any comments or corrections to me.

Colin (as WG co-chair)







            RTP Media Congestion Control (RMCAT) Working Group

                     IETF 102, Montreal, 20 July 2018

Reported by Colin Perkins

The RMCAT working group met once at IETF 102 in Montreal. The meeting
was chaired by Anna Brunstrom and Colin Perkins.

Anna Brunstrom welcomed attendees, introduced the meeting, and reviewed
the document status: 
- Shared bottleneck detection (draft-ietf-rmcat-sbd-11) has been published
  as RFC 8382. 
- Coupled congestion control is in the RFC Editor queue.
- The NADA draft has completed working group last call, and is awaiting a
  minor update to address document shepherd review comments.
- The authors of the Google Congestion control algorithm have stated that
  they will not progress the draft at this time.

The group has four evaluation drafts. The eval-test, eval-criteria, and
video-traffic-model are in WG last call.  These were due to close today, 
but we have insufficient reviews to progress the drafts. Xiaoqing Zhu
volunteered to review eval-criteria, and the chairs solicited reviews
of the other drafts. At least one additional review of each is needed
before the drafts can progress.

Xiaoqing Zhu and her and colleagues working on integrating their code
with ns3 for the wireless-tests. The draft is ready for WG last call,
and this integration work shouldn't affect that. Accordingly, the 
chairs plan to issue a WG last call for the wireless-tests once the
other evaluation drafts have completed their WG last call.

The rtp-cc-feedback has been updated to match the latest version of the
feedback message draft.

Colin Perkins presented the feedback message draft. This was previously
discussed in AVTCORE earlier in the week, and has been updated to address
comments from Sergio Mena on the mailing list, and from previous meetings.
Sergio noted that he is happy with the changes made, and the proposal to
replace the end_seq+1 field with a length field (although he'd prefer a 14
bit length field, to match the recommendations around in-window checks for
1/4 the sequence number range). Jonathan Lennox asked that the draft give
guidance on when to send reports saying "nothing received". The draft will
be updated shortly, to finalise the packet format, provide this guidance,
and to more clearly discuss the design rationale.

To move forward, the chairs asked if there was interest in pursuing the
framework and cc-codec-interactions drafts. Zahed Sarker noted that there
is little energy in the working group and it may be hard to get reviews for
these draft; accordingly the authors have little incentive to proceed
(although he does see benefit in the work).  Mirja Kuehlewind noted that
the framework also defines terminology, which is important to get right if
we advance candidates to proposed standard. Conclusion was that we should
not progress these drafts at this time, but would consider doing so if we
come to advance the candidate congestion control algorithms to proposed
standard.  Xiaoqing Zhu asked do we want feedback on these draft from the
broader transport area?  Anna Brunstrom agreed that this would be useful
when we move the framework forward, but there is no point doing so now.

The chairs asked if is there any interest in a hackathon around the common
RTCP congestion control feedback format and candidate congestion control
algorithms at the Bangkok meeting? Zahed Sarker, Sergio Mena, and Jonathan
Lennox all supported the idea, possibly with remote attendance. Since
there seemed to be support, the chairs will explore whether this can be
set-up.

                                   - + -