[asterisk-dev] Overlapped dialling on SIP trunks for outgoing calls

Andreas Brodmann andreas.brodmann at gmail.com
Sat Apr 26 02:03:31 CDT 2008


2008/4/25 Johansson Olle E <oej at edvina.net>:

>
> 25 apr 2008 kl. 17.34 skrev Tzafrir Cohen:
>
> > On Fri, Apr 25, 2008 at 05:25:33PM +0200, Andreas Brodmann wrote:
> >
> >> I read the docs before writing here and I used allowoverlap=yes.
> >>
> >> What I would like to know from you is wheter this is supported for
> >> phones only (as the example below shows) or wheter it is supposed
> >> to work on sip trunks for 'outgoing' calls:
> >>
> >> 1) in sip.conf: [general] allowoverlap=yes
> >> 2) asterisk sends an INVITE to a carrier
> >> 3) carrier sends 484 back
> >> 4) asterisk sends congestion msg to phone.
> >
> > Can you emulate it in the dialplan?
>
> Reading the source, if we get 484 we end up here:
>
> Case 484:       /* Address incomplete */
>                         return AST_CAUSE_INVALID_NUMBER_FORMAT
>
> And that's an error returned to the dialplan and the dialplan will
> have to try again.
>

 Oej,

I just went throught rfc3578 section 3.5.

This section may explain our problem. Currently
chan_sip implementation at once replies to an incomplete INVITE.
rfc3578 section 3.5. explains that we must note immediately reply
but wait either for the subsequent INVITE and then reply to the first
INVITE or wait for interdigit timeout and then reply to the INVITE.

If implemented so, we could use exten => 000.,1,... in the dialplan
as done so in chan_zap. This way asterisk can have overlap enabled
and still send outgoing calls over sip channels without having to
use overlap itself.

I've found a developer who is willing to modify chan_sip in the next months
to be compliant with rfc3578 section 3.5.

Would that be ok for you or do you prefer to look into it yourselfs?

-Andreas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.digium.com/pipermail/asterisk-dev/attachments/20080426/faa4e90f/attachment.htm 


More information about the asterisk-dev mailing list