IETF MOQ WG: PR #1770 - changes to PUBLISHER_PRIORITY

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: Alan Frindell
• Published: 2026-08-27T15:00:15.000Z
• Source tag: ietf-mailing-list
• Message URL: Open in IETF Mail Archive

Summary

Alan Frindell posted to the MOQ working group mailing list on August 27, 2026. In this case REQUEST_UPDATE is better than PUBLISH_STATE_NOTIFY because the arrival of the REQUEST_OK gives the publisher a signal of when i…

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

In this case REQUEST_UPDATE is better than PUBLISH_STATE_NOTIFY because the arrival of the REQUEST_OK gives the publisher a signal of when it is unequivocally safe to remove explicit priority from published objects and

Plain-Text Body

In this case REQUEST_UPDATE is better than PUBLISH_STATE_NOTIFY because the arrival of the REQUEST_OK gives the publisher a signal of when it is unequivocally safe to remove explicit priority from published objects and go back to compressing. Assuming priority changes happen on timescales of many RTTs, and publishers publish objects with increasing locations, including LARGEST_OBJECT when changing DEFAULT_PUBLISHER_PRIORITY gives the receiver relatively high chance of getting it right at the cost of holding only the Location and previous priority value.

Thanks

-Alan