WebRTC DSCP Control API

[dontcallmedom]

Hello TAG!

I'm requesting a TAG review of WebRTC DSCP Control API - following up to #325 who had been closed for lack of an explainer.

An API to control the network priority of media streams sent over WebRTC connections, using DSCP marking and congestion control.

• Explainer (minimally containing user needs and example code): https://github.com/w3c/webrtc-dscp-exp/blob/master/explainer.md
• Specification URL: https://w3c.github.io/webrtc-priority/
• Tests: N/A
• Security and Privacy self-review²: Security and Privacy Questionnaire w3c/webrtc-priority#3
• GitHub repo (if you prefer feedback filed there): https://github.com/w3c/webrtc-dscp-exp/
• Primary contacts (and their relationship to the specification):
  • Harald Alvestrand @alvestrand , Google (editor)
• Organization(s)/project(s) driving the specification: WebRTC WG
• Key pieces of existing multi-stakeholder review or discussion of this specification: WG discussions
• External status/issue trackers for this specification (publicly visible, e.g. Chrome Status):

Further details:

• I have reviewed the TAG's API Design Principles
• Relevant time constraints or deadlines: N/A
• The group where the work on this specification is being done: WebRTC WG
• Major unresolved issues with or opposition to this specification:
• This work is being funded by:

We'd prefer the TAG provide feedback as
🐛 open issues in our GitHub repo for each point of feedback

[kenchris]

The spec URL returns a 404 for me

[kenchris]

The explainer is not very good at explaining the use-cases (very short and not use-case focused) and mostly lists examples. Please have a look at https://w3ctag.github.io/explainers for how to write a good and useful explainer

[dontcallmedom]

(updated the request with updated URL of the spec https://w3c.github.io/webrtc-priority/)

[ylafon]

I was wondering if letting a web app setting network priorities was a good thing to do. I'd rather have hints on relative priorities and let the system decide how to apply them (or not).

[cynthia]

@ylafon and I looked at this again during the Wellington F2F. Aside from the comments above (which we do consider substantial) we don't have further feedback.

[alvestrand]

Note to @ylafon - the spec contains two parts - the local prioritization and the setting of DSCP code points.
I take it the local prioritization (which is prioritization among the client's own flows) is uncontroversial.

DSCP code points are usually the subject of network manager rule-setting; if the network is not configured to act on them, they will have no effect. So they can be considered "hints on relative priorities" sent from the host to the network.

The idea that such hints should be settable by the app has been accepted in the IETF for a very long time (years); I'm not eager to reopen that issue there. The relevant IOCTLs are available to all apps running natively on the host, so if we add another level of policing to their use, we're increasing the capability distance between native apps and Web apps. Still, a permissioned model might be a possibility here.

Apologies for the slow response.