RE: [Sipping] Transcoding

Gonzalo,

Some comments inline.

-----Original Message-----
From: sipping-bounces@ietf.org [mailto:sipping-bounces@ietf.org] On Behalf
Of Gonzalo Camarillo
Sent: Sunday, February 13, 2005 3:46 AM
To: sipping
Cc: Rohan Mahy; Janne Peisa (JO/LMF); Fredrik Aberg (JO/LMF)
Subject: [Sipping] Transcoding

Folks,

I guess it is time to re-start the discussions we had some time ago on 
transcoding invocation. As you know, the transcoding framework defines 
two transcoding models.

http://standards.ericsson.net/gonzalo/papers/draft-ietf-sipping-transc-frame
work-00.txt

[MB] The draft indicates that either one of the caller or callee could
invoke transcoding, but the determination of that is too weak on the caller
side (it may not be able to determine the media capabilities).

In the first one, referred to as the 3pcc model, the user agent invoking 
the transcoding service has a signalling relationship with the 
transcoder and with the remote user agent. This model has already been 
approved.

http://www.ietf.org/internet-drafts/draft-ietf-sipping-transc-3pcc-02.txt

In the second model, the transcoder receives SIP messages from a user 
agent and forwards them to the other user agent. It is a B2BUA.

The open issue is how to invoke this transcoder. There are currently two 
proposals:

1) Treat the transcoder as a simple conference server and use the 
URI-list service for INVITEs to establish the sessions. That is, the UAC 
sends an INVITE to the transcoder with the URI of the UAS in a body.

INVITE transcoder@domain.com
[SDP]
[UserB@domain.org]

This approach was documented in the following document:

http://standards.ericsson.net/gonzalo/papers/draft-camarillo-sipping-transc-
b2bua-01.txt

2) User Route headers as if the transcoder was a proxy. This is 
basically what SBCs do. The transcoder would rewrite the SDP.

INVITE UserB@domain.org
Route: transcoder@domain.com
[SDP]

Approach 1 requires user agent to generate one-element XML resource 
lists, and approach 2 has a number of problems associated to it (these 
problems have been discussed in the context of SBCs).

[MB] I don't see the difference between Approaches 1&2 from a semantic point
of view. Both require SDP rewrite it seems, which means any SDP encryption
must be hop-by-hop. What are the other problems ? At the media stream level,
it seems no different from the 3pcc model. 

Alternative proposals include defining a new header field similar to 
Route but which is used to include B2BUAs instead of Proxies...

Comments are welcome,

Gonzalo

_______________________________________________
Sipping mailing list  https://www1.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors@cs.columbia.edu for questions on current sip
Use sip@ietf.org for new developments of core SIP

_______________________________________________
Sipping mailing list  https://www1.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-implementors@cs.columbia.edu for questions on current sip
Use sip@ietf.org for new developments of core SIP