Re: [rmcat] call for comments on draft-ietf-rmcat-app-interaction-00

[Varun Singh <[e-mail]>]

Hi all,

Comments inline

On Wed, Oct 22, 2014 at 12:05 PM, Karen Elisabeth Egede Nielsen
<[e-mail]> wrote:
> Hi,
>
> In order to progress the wg document on application interactions we and
> the authors of the doc would like to call for people, CC designers in
> particular, to
> read the document and provide comments with focus on the following
> aspects:
>
>   1. Scope of document: Which interfaces are we looking to have described
> in this document - especially should the document only describe the
> interface in between the RMCAT Congestion controller and the application
> or should the document provide an more extensive description of the
> various components and their interrelations - still likely with a special
> emphasis on  the RMCAT Congestion controller and the application/codec
> interface - ?
>
>  This was discussed already at IETF89 RMCAT session (see minutes) - this
> is an attempt to recap this discussion in order to establish consensus.
>

My personal thoughts on this:

I am waiting for the Constraints API for WebRTC/Media Source API
[1,2]. The API allows the web application to provide hints to the
underlying WebRTC stack on what it expects (or at least that is what I
think it does). For example through the API the application can
configure the range of expected video resolutions, video frame rate,
etc. This will be discussed and maybe finalized at next weeks W3C
meeting.

I don’t know, if we have to strictly follow the variables manipulated
by the Constraints API, but it will surely help the discussion in the
application interactions draft. Or even provide feedback to the W3C,
if there are variables that the congestion control proposal’s need.
But honestly, I am not sure I have figured out all these interfaces
yet.

Cheers,
Varun

[1]: http://w3c.github.io/mediacapture-main/#media-track-constraints
[2]: http://w3c.github.io/mediacapture-main/#constrainable-interface


>   2. What (given the scope assumed for 1.) are the top 3-4 interactions
> (e.g., allowed rate)   that matter the most and which are needed to be
> described in the document ?
>
> Thanks.
>
> BR, Karen/Chairs
>



-- 
http://www.netlab.tkk.fi/~varun/