[asterisk-dev] Recommendations for using a SIP stack with Asterisk

[Gunnar Hellström]

On 2012-11-10 01:26, Russell Bryant wrote:
> > With all of these in mind, what do you suggest the Asterisk project does
> > in order to use a third-party SIP stack?
Before getting that far, carefully select the candidates and verify 
their multimedia capabilities.

So many SIP stacks have begun as audio-only SIP stacks, and have got 
multimedia squeezed in late, resulting in a lot of unwanted behavior. 
Good behavior and good structure is needed from the beginning. This 
cannot be read in specs, it must be tested.

Assuming that it says it has at least audio and video support, do at 
least the following:

1. Enable audio and video in it, but call it with video only, and verify 
that a proper call with video is achieved.

2. Call it with audio and a media type that it does not support ( e.g. 
m=message  or m=text if some of these are not supported), and verify 
that the m-line is still there in the answer, but with 0 for the port.

3. Call it with video and audio and a complex combination of bandwidth 
settings for the whole session as well as for video, and check at what 
bandwidth it selects to transmit, and verify against what it should.

4. Call out from it with video and audio, and answer with video only ( 
properly with m=audio still in the sdp but wotj port 0 ).
Verify that the call is properly answered and that video works.

-----------

There are of course a lot of other situations that would be good to 
check, but the above cases are basic for revealing if it is a real 
multimedia stack or an amended audio stack.

Some verification about how prepared the structure is for adding support 
for new media would also be important in the evaluation. E.g. the PJSIP 
at the moment has only audio and video. Does it have a good structure 
for adding e.g. text and message media?

/Gunnar Hellstrom