New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Priorities and Pooling #102
Comments
Let's discuss at 116. This could be a new feature in the drafts or it could be tackled by a future extension. |
FWIW, RFC 9218 sort of recommends sharing already for conventional CONNECT; https://www.rfc-editor.org/rfc/rfc9218.html#name-scheduling-and-the-connect- I keep hearing that people want strict ordering guarantees among streams. Are you saying that you'd want a fair share of bandwidth but strict ordering inside that fair share? |
I think I want to apply the HTTP prioritization per WT session (so one session could entirely block another, if they have different urgency, eg) but additional ability to prioritize streams within that session, which could be strict ordering, incremental or some combination. |
Discussed the following at IETF 116 Option 1: Nothing, you get what you get and don’t throw a fit The sense I got from the room was that folks preferred Option 4. |
Chair: discussed in editor's meeting. Plan is to have @afrind write up a PR for option 4 and we'll confirm consensus on the list before merging. |
I made this comment on a PR related to the Warp draft for moq: moq-wg/moq-transport#98
How are streams across different WebTransport sessions prioritized? The WebTransport CONNECT request can have an HTTP Priority header -- one logical interpretation is that all streams belonging to that session be prioritized "under" that top level prioritization. eg: if I have two sessions at the same urgency and incremental on, session A has 3 streams and session B has 1, then I might expect the scheduler to round-robin at the session level (rather than the QUIC stream level), splitting the bandwidth 50/50 between A and B. Is that correct? For WebTransport over H2, that is the most natural (or only) interpretation.
The text was updated successfully, but these errors were encountered: