Re: Do we really need TCP/DTLS/SCTP proto field?

As it was mentioned before, this is not only the re-INVITE caused by ICE
completion, but also any re-INVITE after ICE nomination completes. So, any
re-INVITE sent, for instance to update the codecs, which does not initiate
ICE restart will have one ICE candidate and will use the appropriate
transport in the m= line. We have a current ICE specification and this is
how it works. I agree we can update ICE SDP specification to define ICE
transport tag, which will allow us to stop all this nonsense. Until this is
done we need transport tags for both UDP and TCP ICE candidates. The
discussed draft (draft-ietf-mmusic-sctp-sdp) is not the right place to
update ICE specifications and we do want to finish it before all the ICE
updates.

Regards,
_____________
Roman Shpount

On Sun, Feb 19, 2017 at 7:19 PM, Iñaki Baz Castillo <[email protected]> wrote:

> 2017-02-16 18:38 GMT+01:00 Eric Rescorla <[email protected]>:
> > I also think the re-INVITE is unnecessary.
>
> Such a re-INVITE is like "hey, we have agreed on a candidate pair by
> using ICE procedures and, hence, now I send you a message confirming
> it to you, which is totally useless because you already know that".
>
>
> --
> Iñaki Baz Castillo
> <[email protected]>
>

_______________________________________________
mmusic mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/mmusic