this sounds more like your webserver is not able to handle the load. Kannel can handle 2000 sms per seconds. I have seen that working.
You should look deeper into your logfiles to see what's causing the delay. On 05.04.2010, at 18:30, andros agena wrote: > We are in production with Kannel with the below version: > > At heavy traffic (30,000 messages or so for 15 mnts) We are facing delays of > 30 to 45 mnts for MO received from SMSC to be sent to HTTP trigger. We see > messages in Kannel log but we onyl get the MOs to http trigger servlet after > 30 to 40 mnts of delay. This is happeneing at high loads - say 20,000 > messages in 15 mnts span > When we see the queue status it shows 0 in queue. We think it is not doing > parallel processing but one by one message handover to http trigger. > > Please help us. Is there any configuration we need to set for multi > processing of messages > ? > --------------- > Kannel bearerbox version `cvs-20090729'. > Build `Jul 29 2009 20:44:55', compiler `3.4.6 20060404 (Red Hat 3.4.6-10)'. > System Linux, release 2.6.9-78.ELlargesmp, version #1 SMP Wed Jul 9 16:03:59 > EDT 2008, machine x86_64. > Hostname abc.yyyy.com, IP xx.xx.xx.xx > Libxml version 2.6.16. > Using OpenSSL 0.9.7a Feb 19 2003. > Compiled with MySQL 5.1.36, using MySQL 5.1.22-rc. > Using native malloc. > Status: running, uptime 30d 18h 7m 18s > WDP: received 0 (0 queued), sent 0 (0 queued) > SMS: received 174291 (0 queued), sent 395217 (0 queued), store size -1 > SMS: inbound (0.62,0.78,0.07) msg/sec, outbound (1.77,1.61,0.15) msg/sec > DLR: 0 queued, using internal storage > Box connections: > smsbox:LSMPP, IP xx.xx.xx.xx (0 queued), (on-line 30d 18h 7m 12s) > SMSC connections: > PARSMSC[PARSMSC] SMPP:xxx.xxx.xxx.xxxx:xxxx/xxxx:acdefg:SMPP (online > 29343s, rcvd 173533, sent 395217, failed 5147, queued 0 msgs) >
