csperkins.org

RTP Congestion Control Feedback: Update from IETF 104 (Prague) Hackathon

I've recently updated our draft on RTP Congestion Control Feedback to address feedback based on implementation experience from the hackathon at IETF 104, held in Prague earlier this year.

The updates were minor this time, comprising a series of clarifications around signalling, what packets should be included in congestion feedback reports, and on the implications of congestion feedback for retransmission. None of these changes the packet format or signalling, but should hopefully improve interoperability by addressing common misinterpretations.

The only remaining open issue is to add some discussion on how this format relates to the proprietary congestion control feedback used by Google Chrome, in particular on the costs and benefits of using per-flow feedback rather than adding a single congestion control sequence number as a header extension in RTP packets.

I'll give talks about these changes in the AVTCORE and RMCAT working group meetings at IETF 105 in Montreal next week. My hope is that the draft will be ready for working group last call soon thereafter.