<html>
<font size=3>At 11:23 AM 5/29/2002 +0100, S.Booth wrote:<br><br>
<blockquote type=cite class=cite cite>I was wondering if it would make
sense to modify vic,rat,etc so that all<br>
RTP streams from an AG node use the same CNAME
attribute</font></blockquote><br>
It might indeed make sense to do that - it's something I've thought about
doing but haven't gotten around to. <br><br>
<br>
<blockquote type=cite class=cite cite><font size=3>At the moment CNAMEs
are constructed as <username>@<ip-address> <br>
so audio and video streams have different CNAMEs on a standard AG node
<br>
but my uderstanding of<br>
RFC-1889 is that CNAMEs are intended to associate streams from the
same<br>
participant for purposes like lip-sync and audio
switching.</font></blockquote><br>
There are a lot of SHOULD's in the section defining CNAME as user@host,
which means with good reason the recommendations there can be changed. I
think this situation definitely qualifies.<br><br>
(Latest RTP draft at
http://www.ietf.org/internet-drafts/draft-ietf-avt-rtp-new-11.txt)<br><br>
<br>
<blockquote type=cite class=cite cite><font size=3>I know that AG does
not need audio switching but this would make it easier<br>
to interface with other systems. </font></blockquote><br>
For instance, I suspect that VRVS in H323-bridging mode may get confused
with the audio/video separation (it does video-follows-voice
switching).<br><br>
<blockquote type=cite class=cite cite><font size=3>Even in an AG context
it would be nice to be able to get audio switching<br>
to work because we could re-use the mechanism to highlight the
windows<br>
from the speakers node as this would make it easier to locate people
<br>
in sessions with a large number of sites taking
part.</font></blockquote><br>
Agreed. <br><br>
I suspect I'll roll this into the new venue server/ node software work.
<br><br>
--bob</html>