[Freeswitch-users] Performance testing: FS and own App?

[Ken Rice]

Freeswitch can handle a large volume of call... I suggest you review your
configs to make sure you don't have any of the default or arbitrary other
limits in there... We routinely run > 1500 concurrent calls on dual quad
core hardware at call rates far above what you tested at

Ken


> From: Dennis <odermann at googlemail.com>
> Reply-To: <freeswitch-users at lists.freeswitch.org>
> Date: Tue, 23 Dec 2008 09:49:38 +0100
> To: <Freeswitch-users at lists.freeswitch.org>
> Subject: [Freeswitch-users] Performance testing: FS and own App?
> 
> hi,
> 
> i am quite new to freeswitch and now i finally have fs up and running
> as i want with my php scripts to handle the calls.
> 
> now that i want to start the service in the near future, i would like
> to test the performance of the whole system and the reliability.
> 
> what are your experiences and ideas, how i could do it the best way?
> and what do i have to look for?
> 
> i startet to test with sipp, which seems to be a useful tool, although
> i am not sure, how to read the results.
> 
> in the switch.conf i set "max-sessions" to 11000 and
> "sessions-per-second" to 2000.
> 
> in the console i entered: sipp -sn uac xx.xx.xx.xx -s 123456 -r 20 -l
> 1000 -d 4000 -rtp_echo
> 
> the results are as follow:
> 
> 20.0(4000 ms)/1.000s   5061      51.18 s         1023  xx.xx.xx.xx:5060(UDP)
> 
>   20 new calls during 1.001 s period     9 ms scheduler resolution
>   807 calls (limit 1000)                 Peak was 807 calls, after 51 s
>   0 Running, 807 Paused, 0 Woken up
>   0 out-of-call msg (discarded)
>   1 open sockets
>   75989 Total echo RTP pckts 1st stream  481.979 last period RTP rate (kB/s)
>   0 Total echo RTP pckts 2nd stream      0.000 last period RTP rate (kB/s)
> 
>                                  Messages  Retrans   Timeout   Unexpected-Msg
>       INVITE ---------->         1023      0         0
>          100 <----------         1023      0                   0
>          180 <----------         0         0                   0
>          183 <----------         249       0                   0
>          200 <----------  E-RTD1 249       0                   0
>          ACK ---------->         249       0
>        Pause [      0ms]         249                           0
>          BYE ---------->         216       0         0
>          200 <----------         216       0                   0
> 
> 
> 
>   Counter Name           | Periodic value            | Cumulative value
> -------------------------+---------------------------+------------------------
> --
>   Elapsed Time           | 00:00:00:072              | 00:08:30:106
>   Call Rate              |    0.000 cps              |    4.166 cps
> -------------------------+---------------------------+------------------------
> --
>   Incoming call created  |        0                  |        0
>   OutGoing call created  |        0                  |     2125
>   Total Call created     |                           |     2125
>   Current Call           |        0                  |
> -------------------------+---------------------------+------------------------
> --
>   Successful call        |        9                  |     2125
>   Failed call            |        0                  |        0
> -------------------------+---------------------------+------------------------
> --
>   Response Time 1        | 00:00:00:000              | 00:02:43:140
>   Call Length            | 00:04:04:169              | 00:02:47:149
> 
> 
> 
> i do not get any errors. but i think that it is strange, that there
> are so little calls running at the same time. when i try to call the
> same number as sipp does, i have to wait a very long time, till the
> call is beeing answered.
> 
> why are there so many paused calls waiting for fs to take care of?
> 
> is this normal or are there some problems, which i have to take care of?
> 
> i want to add, that i am using socket outbound with a number of
> php-scripts, which do a lot of call- and dialplan-handling. i am not
> using any xml. therefore i am sure, that the results are worse, that
> they would be with normal xml dialplans.
> 
> 
> thanks for your help and tipps.
> dannis
> 
> _______________________________________________
> Freeswitch-users mailing list
> Freeswitch-users at lists.freeswitch.org
> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
> http://www.freeswitch.org