[asterisk-dev] Mantis Issue Review Meeting Notes

[Olle E Johansson]

> -Issue 4825, "new codec negotiation algorithm" - Dwayne will review  
> the patch and summarize during the next couple of weeks

It's important to review the videocaps alternative too. It's  
something that is tested and works very well for video and
could easily be extended to audio too.

We need to move forward in this area and have many proposals to look at.

Writing a requirements document for a new system would be a good  
start, then review the various solutions and
come up with a platform for the future based on parts from the  
various proposals.

A related issue, that is handled in videocaps and turned out to be  
the killer of Mark's direct RTP setup solution
is the "back channel".

When we initiate a call, we take the incoming channel, invite the  
outbound channel. The outbound channel,
in the case of SIP, returns a series of properties for the call that  
need to be the base of the reply to the
incoming channel. Currently, there's no way to send that data back,  
so the call fails. This is an architectural
flaw we need to solve in the current development cycle, possibly by  
adding an attachment to the ANSWER
control message.

But what's the format of the attachment? Back to codec negotiation  
again. I would call it "call properties negotiation"
because it is more than codecs - it will soon also become security  
properties.

Just some random thoughts before returning to Good Friday Gardening :-)

/O