Re: [rtcweb] Support of video with different resolutions

[Stefan Håkansson LK <[e-mail]>]

On 2013-01-02 19:32, Ted Hardie wrote:
> On Wed, Jan 2, 2013 at 9:15 AM, Stefan Håkansson LK
> <[e-mail]
> <mailto:[e-mail]>> wrote:
>
>
>     Another way to look at this is that we already have the following
>     requirement:
>
>     "F11             The browser MUST be able to transmit streams and
>                         data to several peers concurrently.",
>
>     and presumably there would be independent rate control for these
>     streams (since they may experience different network conditions),
>
>     simulcast is in  sense just a special case (the two other peer's are
>     the same endpoint).
>
>
> While I agree that you could model it this way, I think it may be more
> appropriate to treat simulcast as its own case, rather than as a special
> case of the multiple peer.

I agree fully. I was just trying to say that we already have 
requirements covering a large part of what is needed for simulcast, so 
simulcast would not require much more in terms of implementation than 
what we already have agreement on.

But it should be its own case. Apart from the point you make about rate 
control, there is also a tighter requirement on synchronization between 
two versions of the same stream in a simulcast case than what would 
(likely) be needed in a multiple peer case.

>
> Consider this pair of scenarios:
>
> Case one:
> There is one endpoint, Zelda, which is sending media via Turn service to
> two hosts behind the same NAT (let's call them Scott and Edouard).
> There is independent rate control for the streams sending this media.
>
> Case two:
> There is an endpoint, Zelda, which is sending simulcast media to Scott,
> which is using a Turn service and is behind a NAT.
>
> Isn't managing fairness going to be potentially very different between
> case one and case two?
>
> regards,
>
> Ted