[Standards-JIG] Defining IAX transport over jingle

Mikael Magnusson mikaelmagnusson at glocalnet.net
Wed Feb 22 19:52:51 UTC 2006


Antonio Cano damas wrote:
> Peter Saint-Andre wrote:
> 
> 
>>Simon Guindon wrote:
>>
>>
>>>I would also like to be included in the IAX discussions. I think IAX
>>>should be pretty easy to implement but at the same time a bit confusing.
>>
>>>I've thought about it quite a lot in the last while going through the
>>>DTMF thread and I'm a bit confused as to how we are going to separate
>>>IAX calling peers, and IAX calling Asterisk.
>>
>>Does it matter? All Jingle says is "here is how to negotiate a session
>>with another party" -- the parties could be two clients, a client and a
>>server, two servers, etc.
>>
> 
> I thought this doesn't matter. A peer is client and server at the same
> time, we don't mind if is a phone, jabber client or asterisk pbx.
> 
> I'll try to explain, let me use Asterisk terms,  for calling using IAX
> we have the next URI: 'user:password at ip:port/extension'. Now we are
> going the two cases:

The libiax2 library supports dial strings using the following format:

   [<user>[:<secret>]@]<peer>[:<portno>][/<exten>[@<context>]]

It's only the peer that is needed to establish an IAX call, and the
other values except port number are sent in Information Elements (IE:s) 
during call setup and authentication. The default (UDP) port is 4569.

   Informational Elements
   ----------------------
   0x01 CALLED NUMBER
   0x05 CALLED CONTEXT
   0x06 USERNAME
   0x07 (plaintext) PASSWORD
   0x10 MD5 result
   0x11 RSA result

For more information:
http://www.ietf.org/internet-drafts/draft-guy-iax-00.txt

>     1. Calling from pc 2 pc.  In that case we don't need the user and
> password because is the other user who accept or reject the call, and in
> that case don't mind wich extension to use... for example we could use
> the 's' that is default extension of asterisk. The URI is like this:
> ip:port/extension.

Why send an extension in a CALLED NUMBER IE at all? Is it really necessary?

/Mikael




More information about the Standards mailing list