On Sat, Nov 28, 2009 at 9:47 AM, Robert B <[email protected]> wrote:
> I preface this with the fact that I am coming from the Asterisk world
> where things are "different"... ;)
>
> So I am getting my head about all these various components and one thing
> sort of puzzles me, and that is the media path.
>
> Let's say I have an sipXecs server sitting in a datacenter somewhere.
> Across town I have a handful of Polycom handsets in an office who has a
> business grade cable modem of somesuch. The Polycoms are configured to
> register with their sipXecs PBX in my datacenter over the public Internet.
>
> Now, in the Asterisk world, all the RTP media would get relayed through
> Asterisk and sent to the remote office. I understand this is different
> in the sipXecs world (e.g. the real world) where the SIP signaling would
> occur "out of band" with the RTP stream.
>
> Inside the office call:
>
> My assumption is that based on name resolution of the URI
> ([email protected], [email protected]) the phones know who/what to send the
> media to. (e.g. in the office, on the same LAN, it would get an internal
> IP and send the call from [email protected] to [email protected]). This
> is why internal DNS is critical.
>
> But how does this work when placing/receiving calls from an ITSP?
>
> Call comes in to [email protected] from the ITSP to sipXecs, the
> appropriate device is notified of an incoming call and the user answers...
>
> Here is where my understanding just falls apart. At this stage, I have
> no idea how the RTP stream "finds its way" to the device... sipxrelay?
> Doesn't that sort of introduce a single point of failure? Or maybe I
> just am completely wrong about all my above assumptions... ?
>
> Kudos to whomever can help me decipher this...


When signaling an ITSP, you have to send your call through sipxbridge
( assuming that is what you are using as your sbc ).

SipXbridge will re-write various parts of the request including the following :

- via header
- Contact
- SDP connection field.

In doing that, it effectively inserts a media relay in the media path
and itself in the signaling path.

It is correct that sipxrelay is a "point of failure" but it is not a
"single point of failure". You can have a redundant system where each
server runs its own relay. If one fails the calls get routed to the
other server.

There is a detailed design document included with the code. Please get
it from svn if interested. See sipXbridge/design/design.tar.gz

Ranga.


>
> -- Robert
>
> _______________________________________________
> sipx-users mailing list [email protected]
> List Archive: http://list.sipfoundry.org/archive/sipx-users
> Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-users
> sipXecs IP PBX -- http://www.sipfoundry.org/
>



-- 
M. Ranganathan
_______________________________________________
sipx-users mailing list [email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-users
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-users
sipXecs IP PBX -- http://www.sipfoundry.org/

Reply via email to