Re: [rtcweb] A problem with both A and B

In cases where each m= line has a different port, when a new SSRC is
received usually it is either discarded (when latched to a pre exisiting
SSRC stream that does not stop) or is treated as replacing the previous
SSRC.

When multiplexing is being used it is not obvious whether the new SSRC is
replacing a previous one, or is another separate stream.

I think that in the multiplex case we need something like a virtual
destination port in each RTP packet, such as a RTP header extension or even
UDP over UDP!

On 21 May 2013 08:45, Emil Ivov <emcho@jitsi.org> wrote:

>
>
> On 21.05.13, 10:38, Christer Holmberg wrote:
> > Hi,
> >
> >>> If adding a new stream means adding a new SSRC, within an
> >>> existing m-, then it is also an SDP O/A issue - bundle or no
> >>> bundle.
> >>
> >> Why is that?
> >
> > Because, even WITHOUT bundle, you have to answer whether adding a new
> > SSRC (or, whatever mechanism we use to identify a stream) within an
> > existing m- line requires an SDP O/A transaction or not.
>
> OK, I guess I took your statement to mean that adding and removing SSRCs
> to m= lines is a matter that cannot be handled without SDP O/A which I
> think it isn't (and which is what this thread started with).
>
> I obviously agree that it is not a bundle-specific issue.
>
> Emil
>
> --
> https://jitsi.org
> _______________________________________________
> rtcweb mailing list
> rtcweb@ietf.org
> https://www.ietf.org/mailman/listinfo/rtcweb
>