[OpenSIPS-Users] Dispatcher algorithm question

Rik Broers RBroers at motto.nl
Wed Mar 2 14:06:28 CET 2016


I’ve been using the hash over call-id for a long time, and distribution has been pretty even over all destinations.

Met vriendelijke groet,

Rik Broers
Voice Engineer
Van: users-bounces at lists.opensips.org [mailto:users-bounces at lists.opensips.org] Namens Gunjan Korlekar
Verzonden: woensdag 2 maart 2016 5:26
Aan: OpenSIPS users mailling list <users at lists.opensips.org>
Onderwerp: Re: [OpenSIPS-Users] Dispatcher algorithm question

Hi Olle,

Thanks for your reply. I will review my script for the re transmission of invites issue.

But in the meanwhile can someone answer my question regarding the 'hash over call-id' algorithm?

Thanks,
Gunjan

> On 29 Feb 2016, at 18:58, Gunjan Korlekar <gunjan at switch.co<mailto:gunjan at switch.co>> wrote:
>
> Hello,
>
> I have a dispatcher table with about 10 destinations and was using the round-robin algorithm to route to destinations since we wanted to have an even spread of traffic across the destinations. However I recently observed that with the round-robin, re-INVITEs with the same call id get routed to different destinations which we thought the dispatcher would handle. So we figured that if we want to circumvent this problem we need to use the hash over call-id algorithm.
>
> Does the 'hash over call-id' give a fair way of distributing the calls over different destinations in the dispatcher like the round robin so that the traffic is spread evenly over the destinations?
> Is there a way we can combine the two algorithms for one set of destinations?

Re-invites are in-dialog and should not be handled with dispatcher normally. Dispatcher is used to dialog-forming requests only. I would like to suggest you review your script.

/O

--
Thanks,
Gunjan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20160302/96a4acb6/attachment.htm>


More information about the Users mailing list