MMUSIC Working Group at IETF 96

The MMUSIC working group met at IETF #96 in Berlin, Germany on Friday July 22, 2016 from 12:20 to 13:20.

 

The meeting was chaired by Flemming Andreasen and Bo Burman.

 

Magnus Westerlund, Harald Alvestrand and the chairs took notes, and Olle Johansson acted as Jabber relay.

 

The meeting was broadcast live and recorded by the Meetecho team. The recording is available at: https://www.youtube.com/watch?v=Fy0caR4w-yo

Agenda review and status update

Slides

On RTSP 2.0 and RTSP NAT drafts:

·         Magnus believes it will probably be November timeframe before it's completely done (in RFC Ed queue)

·         Colin Perkins: Circuit breaker reference was taken out because it wasn't ready. Believe circuit-breaker is now getting ready, so we may consider adding it back as normative reference if that was the only reason it was taken out

RTCWeb dependencies

Drafts in ICE WG

·         One of the ICE WG chairs, Ari Keränen, said that 5245bis (ICEbis) and trickle-ICE have WGLC target date mid-September.

·         Martin Thomson reported on a discussion regarding ICE transport pacing that appears reasonable to resolve on the ICE list within a month.

·         ICEbis got three volunteering reviewers in the RTCWEB session

·         Olle Johansson commented that also STUNbis needs review (although that is in TRAM WG). Cullen Jennings is volunteering to review that.

 

Ben Campbell, one of the ART Area Directors, commented that the WG chairs should consider the order of drafts being submitted to the IESG for publication, to possibly allow more even distribution of the IESG and RFC Editor processing load.

Bundle

·         Jonathan Lennox commented that he didn't intend for the discussion of multiple MIDs for a single RTP stream to be blocking BUNDLE, it should be an extension. Colin Perkins agreed and commented that this topic changes RTCP and that requires AVTCORE discussion.

·         MID needs to be registered according to the SDES RTP Header Extension draft. Jonathan, Magnus and Christer Holmberg will sit down after the meeting.

·         Discussion around overall readability of the document, where Eric Rescorla noted that if editorial comments are not valued, it will not be worth reviewing. Ben Campbell stated that the authors should take the feedback and then decide what to do with it.

·         Reviewers volunteering: Magnus Westerlund, Cullen Jennings, Eric Rescorla.

 

sctp-sdp

·         Christer: Do not review current version; there will be changes.

·         Christer: Not sure anybody is interested in DTLS over SCTP.

·         Bo asked if anybody in the room is interested in DTLS over SCTP. Nobody is. Christer will send e-mail to list asking as well. If nobody speaks up, it will be removed.

·         Reviewers: Will ask for reviewers on the list, but this document has low priority.

 

ice-sip-sdp

·         Reviewers: Christer Holmberg and Adam Roach volunteered (send e-mail to remind them). Follow-up with Thomas and Roman on e-mail to see if they will sign up to review.

trickle-ice-sip

·         Reviewers: Adam Roach, Jonathan Lennox, Suhas Nandakumar (volunteered by Cullen Jennings), and Cullen Jennings. Send e-mail to remind/confirm. Paul Kyzivat and Martin Thomson declined.

sdp-simulcast

·         Reviewers: Jonathan Lennox, Peter Thatcher, Magnus Westerlund, Bo Burman.

rid

·         Reviewers: Jonathan Lennox, Peter Thatcher, Magnus Westerlund, Bo Burman.

·         Adam Roach presented a few open issues and proposed how to resolve them.

·         Suggest to leave declarative SDP use undefined in this document. Magnus agrees and nobody in the room disagrees. Declarative use will be left undefined, subject to mailing list verification.

·         Propose to restrict the RID SDES value to alphanumeric, rather than UTF-8. Colin Perkins commented that 4566bis may have an ABNF rule for UTF-8. Cullen Jennings: You get into complicated comparison rules with UTF-8, so that should be avoided. ASCII is a sub-set of UTF-8. Colin noted that the RID RTCP SDES item is the one to be restricted, but that restriction should then be made in the related AVTEXT draft. The conclusion is to restrict RID RTCP SDES items to alpha-numeric, as is currently defined in the draft. Adam will verify on the list and with AVTEXT.

 

SDP lang Attribute Issue

Bernard Aboba presented for Gunnar Hellström. Slides.

 

Colin noted that the “lang” text in SDP is there since 1995 and the MBONE tools. Bernard said that it would be helpful if people that have implemented this already could tell us what they actually implemented.

RTCWeb SDP Issues

JSEP and BUNDLE don't match in all cases. Slides (from RTCWEB WG).

 

a=rtcp usage

·         JSEP should change to match BUNDLE.

·         A few people agree explicitly, nobody disagrees.

 

a=fingerprint

·         JSEP puts fingerprint in all lines, BUNDLE only in the first line.

·         BUNDLE is right; fix JSEP.

 

Christer Holmberg volunteering to do review of JSEP for BUNDLE consistency.

 

In general, agree to align with BUNDLE, which means we don't repeat transport mux category parameters across bundled “m=” lines (unless you don't know whether the other side supports bundle).

Any Other Business

Magnus informed that he has received questions around simulcast and how the RTP Streams and their markings with MID and RID appear downstream for the various types of RTP mixers. Magnus and Bo volunteered to write an overview document that clarifies some of that, either in AVTCORE or MMUSIC.