Anthony Minessale napsal(a):
What does it look like if you serve the directory from the static xml file out of curiosity.
Good question :-) I have never thing about it, becasue I need "dynamic" users. But it should show up very impressive number :-) I'll try it tommorow (here is midnight) and let you know. BTW: I try to find some another server in colocation with higher performace. With mentioned P4 I'm affraid have no chance to stress FS with static xml directory... Thank you for your interest, Anthony! Best regards, kokoska.rokoska
On Tue, Feb 3, 2009 at 4:11 PM, kokoska.rokoska <kokoska.rokoska@post.cz <mailto:kokoska.rokoska@post.cz>> wrote:
Ken Rice napsal(a): ...
> On Registrations we have experienced Registration/second rates exceeding 150 > registrations per second using mod_xml_curl to feed the users directory. I > suspect, this number can be greatly increased if we were to feed directory > with something that cut out the apache and php over head >
If someone interested I have few numbers on Registrar performance:
DB server: 2x Quad core E5345 @ 2.33GHz, 16 GiB RAM Centos 5 x86_64, MySQL 5.0
Registrar server: 2x Quad core E5345 @ 2.33GHz, 16 GiB RAM Centos 5 x86_64
Tested using sipp with 10.000 and 30.000 "users".
FreeSWITCH as registrar - current trunk: 1. FreeSwitch si simply modified (code doing NAT-ping is commented out :-) 2. Directory is served through lighttpd and simple "C" binary doing one trivial select. Lighttpd runs on the same machine as FS. When I move lighhtpd to another machine, I cannot see any significat performance boost.
Result: I can go up to the 470-500 reg/s. and FS is heavy overloaded and retransmissions occurs.
Kamailio as registrar - 1.4.3. no TLS: 1. Kamailio runs with usrloc db_mode 3 (no caching)
Result: I can go up to the 3500-3700 reg/s. and Kamailio server is at 0.3 load and all 8 cores are bellow 15 %. Without retransmissions. The limit is DB throughput. Just for "curiosity" I switched userloc to db_mode 2 (write back) and at 5000 regs/s I stopped the sipp test, because I saw the bottle neck becomes the server runnig sipp (very old P4 box).
Conclusion: While I see amazing FreeSWITCH performance on INVITEs per seconds and concurrent calls (another galaxy from * point of view :-), if you have to handle lots of registrations per second, it is IMO better to use Kamailio/OpenSIPS/SER as separate registrar and "propagate" users to FS through SQL view.
Hope this helps someone...
Best regards,
kokoska.rokoska
_______________________________________________ Freeswitch-users mailing list Freeswitch-users@lists.freeswitch.org <mailto:Freeswitch-users@lists.freeswitch.org> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org
-- Anthony Minessale II
FreeSWITCH http://www.freeswitch.org/ ClueCon http://www.cluecon.com/
AIM: anthm MSN:anthony_minessale@hotmail.com <mailto:MSN%3Aanthony_minessale@hotmail.com> GTALK/JABBER/PAYPAL:anthony.minessale@gmail.com <mailto:PAYPAL%3Aanthony.minessale@gmail.com> IRC: irc.freenode.net <http://irc.freenode.net> #freeswitch
FreeSWITCH Developer Conference sip:888@conference.freeswitch.org <mailto:sip%3A888@conference.freeswitch.org> iax:guest@conference.freeswitch.org/888 <http://iax:guest@conference.freeswitch.org/888> googletalk:conf+888@conference.freeswitch.org <mailto:googletalk%3Aconf%2B888@conference.freeswitch.org> pstn:213-799-1400
------------------------------------------------------------------------
_______________________________________________ Freeswitch-users mailing list Freeswitch-users@lists.freeswitch.org http://lists.freeswitch.org/mailman/listinfo/freeswitch-users UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users http://www.freeswitch.org