Re: [Moq] Scope of MOQ and ULL Streaming

[Spencer Dawkins at IETF <[e-mail]>]

Hi, Bernard,

I agree with much of what's been said downthread, but wanted to poke at one
point that I haven't seen taken up.

On Wed, Feb 8, 2023 at 3:02 PM Bernard Aboba <[e-mail]>
wrote:

> [BA] To have a shot at wide deployment, MoQ needs a “killer application”
> that can’t be delivered nearly as well by modifying HLS or DASH.  T*he
> requirements document doesn’t really focus on what that is, nor does it
> fully enumerate all the requirements MoQ would need to satisfy to fully
> differentiate itself in that usage. *
>

Improving the requirements document was our goal in presenting open issues
at last week's interim. But, beyond that, we are talking about questions
like deployment incentives that I haven't heard raised before - did I miss
that?

Where I think MOQ is, is that we agreed to charter MOQ for a minimum of
three use cases ("use cases including live streaming, gaming, and media
conferencing <https://datatracker.ietf.org/wg/moq/about/>") as a starting
point, we haven't written down the differences between those three use
cases, and we've been steadily removing attributes in the use cases section
of the requirements draft (so, no longer distinguishing among the use cases
based on latency, in the editor's version
<https://fiestajetsam.github.io/draft-gruessing-moq-requirements/draft-gruessing-moq-requirements.html#name-use-cases-informing-this-pr>
).

My pre-interim plan (and also James Gruessing's plan, I think) was to focus
on high-level requirements that people working on WARP are uncovering, and
revisit the use cases later, when we could say something about additional
use cases that MOQ could support. We did get good feedback on requirements
during the interim (some during requirements draft discussions, and some
during WARP draft discussions), and I expect the design teams to also
provide helpful input as they do their work.

But I'm not sure the requirements draft would be the right place to put
discussion of deployment incentives, or Martin Duke's question, about how
someone knows that they should look at MOQ for their application, or keep
looking, because MOQ doesn't provide what they need. And there are probably
other useful questions to write down, and answer. (IMO, of course)

Does all that make sense?

Best,

Spencer