Re: [rtcweb] Security Architecture: IdP for RTP and RTCP

[Bernard Aboba <[e-mail]>]

Inaki said: "OK, a media endpoint could be internally divided into different
components, one of then handling RTP and the other RTCP, but form the
point of view of media transmission it should behave as a single
endpoint."

[BA] It seems logical for the browser to utilize the same identity for both
RTP and RTCP, regardless of whether they are multiplexed or not.

However, if both the RTP and RTCP peer assertions are validated, shouldn't
it be up to the application decide whether to care?

A given application could compare the RTP and RTCP identities and consider
it a fatal error if the identities are not the same, OR it could decide to
ignore the RTCP identity.

BTW, the "compare" operation is potentially non-trivial in the case of
internationalized identities.  None of the specifications currently
describe how the identities are to be normalized in preparation for the
comparison, so I can imagine that some "fun" could be had there.




On Tue, Jul 8, 2014 at 12:12 PM, Iñaki Baz Castillo <[e-mail]> wrote:

> 2014-07-08 20:09 GMT+02:00 Martin Thomson <[e-mail]>:
> > I think that the way that we manage identity in a multi-party
> > situation probably needs something different to that.  I don't see any
> > particular value in terminating RTCP when you aren't also terminating
> > RTP, the two are far too tightly coupled.  They shouldn't really have
> > been given different names in the first place.
>
> I agree. The fact that RTP and RTCP seem two different protocols and,
> even more, the fact that they can be carried over different
> transports, does not mean that we should imagine exotic scenarios in
> which a media endpoint generates just RTP or just RTCP.
>
> OK, a media endpoint could be internally divided into different
> components, one of then handling RTP and the other RTCP, but form the
> point of view of media transmission it should behave as a single
> endpoint.
>
>
> --
> Iñaki Baz Castillo
> <[e-mail]>
>