IETF MOQ WG: PR #1770 - changes to PUBLISHER_PRIORITY
Cullen Fluffy Jennings posted to the MOQ working group mailing list on August 24, 2026. I’d like to talk about this a bit more. I think this question is confusing two things and was not the key question we were trying t…
Message Snapshot
- List:
moq@ietf.org - Thread subject: PR #1770 - changes to PUBLISHER_PRIORITY
- Message subject: [Moq] Re: PR #1770 - changes to PUBLISHER_PRIORITY
- Author: Cullen Fluffy Jennings
- Published: 2026-08-24T17:29:43.000Z
- Source tag:
ietf-mailing-list - Message URL: Open in IETF Mail Archive
Summary
Cullen Fluffy Jennings posted to the MOQ working group mailing list on August 24, 2026. I’d like to talk about this a bit more. I think this question is confusing two things and was not the key question we were trying t…
Why It Matters
This message came from the public MOQ working group archive and captures active discussion, meeting coordination, or implementation feedback around Media over QUIC.
Message Excerpt
I’d like to talk about this a bit more. I think this question is confusing two things and was not the key question we were trying to sort out. I see the two things as: 1) now allowing publisher priority to change
Plain-Text Body
I’d like to talk about this a bit more. I think this question is confusing two things and was not the key question we were trying to sort out.
I see the two things as:
-
now allowing publisher priority to change
-
making it object perfect with no race conditions
My understanding was pretty clear uses cases for changing a publisher priority.
For example, a conferencing bridge that was publishing 3 tracks for 3 speakers but wanted to keep the video for the active speaker higher than the other video tracks. I would like to keep this.
The second issue was resolving the race condition of a relay might cause this change to happen one object later than the publisher intended. Mo had made the point relay priorities were far from exact to start with on what would happen. I think bulk of people don’t want to solve this. I agree it does not need to be solved.
I would like to add that 2 or 3 bytes of compression in an audio datagram is a huge amount if it is in all the objects.
I was not at interim were this was discussed but I did review the minutes.