Re: [AVT] Comments on draft-ietf-avt-ulp-09.txt

[Magnus Westerlund <[e-mail]>]

Hi,

I have one further comment in regards to ULP.

Currently the draft and RFC 2733 specifies that in the case one uses RTP 
session multiplexing of FEC packets and the original streams, the FEC 
RTP session is in SDP signalled within the a=fmtp line. I think this has 
some problems associated with them.

1. If prevents from using another RTP profile, transport etc., then the 
original session.

2. It requires one to define a RTP payload type that looks like it will 
be transmitted in a session where it will not at all be used.

3. Any other mechanism that is defined in the m= media block concept is 
not possible to apply to the new session. One such example showing the 
importance of these problems are the "a=crypto" attribute defined by 
draft-ietf-mmusic-sdescriptions. Where it currently is impossible to 
specify a security context for this FEC session.

My proposal to resolve all these problems is that the FEC sessions are 
instead defined using normal m= blocks and one utilize the grouping of 
media line (RFC 3388) to bind original and FEC sessions together. This 
is what is done in the RTP retransmission work. As we anyway are 
redefining the FEC packets in a non backwards comptible ways, changing 
signalling at the same time does not appear to have any interoperability 
impact.


Cheers

Magnus Westerlund

Multimedia Technologies, Ericsson Research EAB/TVA/A
----------------------------------------------------------------------
Ericsson AB                | Phone [phone]
Torshamsgatan 23           | Fax   [phone]
S-164 80 Stockholm, Sweden | mailto: [e-mail]


_______________________________________________
Audio/Video Transport Working Group
[e-mail]
https://www1.ietf.org/mailman/listinfo/avt