AVTCORE

Tuesday, 13:00-14:00

Note taker: Bo Burman

Jabber room: Jonathan Lennox

AVTCore Status Update

AD and chairs will have a chat about how to progress the aria draft.

AD Ben C has reviewed multi-stream drafts and will go into further processing.

The authors of the multiplexing guidelines intend to update the document before next IETF meeting.

(EKT milestone is still in AVTCORE, but may move soon).

MPRTP has open source implementation.

Circuit Breakers for Unicast RTP Sessions

Colin Perkins

draft-ietf-avtcore-rtp-circuit-breakers-11

No objections from the room to the changes in -11 regarding media timeout flow breaker. A question from Jonathan Lennox on what you do if there is no defined media framing interval. There are some words in the draft that may have to be checked again, e.g. for timed text media.

Congestion circuit -breaker was also significantly updated.

Colin plans to put the updated algorithms through the same set of simulations as was done with the previous ones.

Michael Ramalho: would like to have a better loss probability measure than what can be provided from RTCP report loss information, like instead estimated from time between loss events, which shows better characteristics when running against TCP. Colin commented that the circuit breaker can’t rely on metrics that are not present in regular RTCP SR/RRs. Jonathan: could we maybe add text saying that if you have something better, use it? Magnus: Then the circuit-breaker will not be consistent.

ACTION item: The chairs will start a 6 week WGLC.

AD Ben C asked if it can be expected that someone verifies the text by implementation during that time, which would be good.

A General Mechanism for RTP Header Extensions

Roni Even

draft-even-avtcore-rfc5285-bis-01

Jonathan sent mail to list just before this session with a few proposals that were discussed and seems agreeable to the WG.

Actions: update  the text about intermediaries removing the RTP headers to be less restrictive since RTP headers are important.

Update security section adding reference to RTP header encryption.

Recommend using 2 bytes header extensions for variable length RTP header that may be longer than 15 bytes.

 

Hum about WG adoption. Fair amount pro, none con. Confirm on the mailing list.

Would like to send to publication before next meeting.