Re: [MMUSIC] Unified Plan for SDP Handling - multiple media sources

[Martin Thomson <[e-mail]>]

On 17 July 2013 14:44, Adam Roach <[e-mail]> wrote:
> On 7/17/13 16:40, Jonathan Lennox wrote:
>>
>> That said, what I still don't see is, if you have the correlator, why you
>> then actually need to send the a=ssrc in the SDP at all? The correlator
>> should be sufficient for disambiguation, and it lets the sender switch the
>> ssrc it's sending for an m-line without needing a fresh offer/answer.
>
>
> That might be worth exploring. I honestly haven't given it a whole lot of
> thought. Perhaps Justin or Martin have a more well-formed opinion on the
> topic?

I did consider this, but there are several reasons that I prefer to
have signaling.  The main one being that it is the only guaranteed way
to get this information through.  In the happy cases, the RTP header
is enough, but if you are losing packets at session startup (a real
possibility in some configurations), or RTP header extensions are
being dropped, or you need to communicate about an in progress stream
and can't guarantee that you can cause the sender to re-send a few
packets with the header extension, or ...

Basically, there are too many ways for the RTP extension to break.
Not enough that it isn't hugely valuable for dealing with the race,
but it can't be the only mechanism that you rely on.