IETF MOQ WG: Knowing the start of a Subgroup

Message Snapshot

• List: moq@ietf.org
• Thread subject: Knowing the start of a Subgroup
• Message subject: [Moq] Re: Knowing the start of a Subgroup
• Author: Luke Curley
• Published: 2026-04-30T16:23:49.000Z
• Source tag: ietf-mailing-list
• Message URL: Open in IETF Mail Archive

Summary

Luke Curley posted to the MOQ working group mailing list on April 30, 2026. I missed the interim, what's the use-case? The original issue is about saving a byte or two when Object ID == Subgroup ID. That seems like a si…

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 missed the interim, what's the use-case? The original issue is about saving a byte or two when Object ID == Subgroup ID. That seems like a simple encoding flag that doesn't require changing

Plain-Text Body

I missed the interim, what's the use-case?

The original issue is about saving a byte or two when Object ID == Subgroup ID. That seems like a simple encoding flag that doesn't require changing the API.

I would not require that Subgroup ID equal the first Object ID. If you use sub-groups for SVC, you *should *have a catalog entry indicating that sub-group N is an enhancement layer with parameters X (even just for filtering). But without tight control of the encoder, there's no way to enforce that frame N is the first frame of sub-group N. You'd have to +100 each sub-group ID (or something) just to maintain this contract for the first frame of a sub-group... and for what purpose?

A flag indicating an object is the first within its sub-group would be easy. However, that information is already implicit via FETCH or SUBSCRIBE object arrival order. I'm not sure *why *we suddenly need an explicit flag, but whatever; it's just a bit.

A flag indicating an object is the last within its sub-group would be very tough. Encoders don't emit this information, nor would an app know ahead of time that there will be no more deltas. You only truly know that a group has ended when a new group begins and a stream FIN accomplishes this just fine.

On Thu, Apr 30, 2026, 8:11 AM Ian Swett <ianswett= 40google.com@dmarc.ietf.org> wrote: