csperkins.org

RTP Congestion Control Feedback: Updates from Bangkok Hackathon

I've just updated our draft on RTP Congestion Control Feedback to address feedback from Jonathan Lennox based on implementation experience during the hackathon at the Bangkok IETF meeting.

A small group comprising Jonathan Lennox, Nils Ohlmeier, and Sergio Mena spent some time during the IETF 103 Hackathon in Bangkok trying to implement the RTP Congestion Control Feedback draft in libwebrtc, specifically as used in Firefox; and in Vidyo’s proprietary codebase. Perhaps not unexpectedly, they found some issues with the draft that Jonathan reported to the AVTCORE mailing list, and that were than discussed during the IETF meeting.

Reassuringly, the issues reported were all relatively minor. There were no real technical changes required to the specification in order to address them. Rather, the feedback highlighted why we need implementation experience: corner cases that are easy to miss, or that are hidden in vague text, must be clarified when implementing a specification. Specifications written in English prose can be imprecise, but code needs precision. A useful reminder, perhaps, of the importance of both rough consensus and running code in the standards process.

Opinions expressed are my own, and do not represent those of my employers or the organisations that fund my research.