Hi Juha,
As later proofed in the reports, the actual problem was openser
un-related - the delay/bottleneck was not in openser - just because of
the integration.
On the other hand, I agree with Henning's statement, that comparing with
1.1, openser 1.3 is a bit slower, but probably the difference is small
and noticeable only is special cases - this comes form the overhead
added due intensive usage of PV in function parameters, adding extra
internal APIs, etc...
Regards,
Bogdan
Juha Heinanen wrote:
also i did some performance tests on openser trunk in
order to clear
doubt that has been expressed recently.
with only 4 child processes and several radius/mysql queries per invite,
i was able to achieve 50 new calls per second using sipp including
radius accounting to database of each call. this was on my two years
old laptop and everything was running on the same host. i didn't try
yet how many calls i could have achieved before retransmissions.
i'm satisfied with the result and don't think that there are any
performance issues with openser trunk.
-- juha
_______________________________________________
Users mailing list
Users(a)lists.openser.org
http://lists.openser.org/cgi-bin/mailman/listinfo/users