Hi Nisan, Nisan Bloch wrote: > > Hi Aarno > At 04:52 PM 4/19/02 +0300, Aarno Syvänen wrote: > >Hi List, > > > >There have been some high stress Kannel tests (continous 350 msg/s, peak > >560 > >msg/s) here at Wiral. Kannel will crash, because incoming messages queue > > what hardware are you running this on?
There were four processors, two Kannels (with cimd simulators) running in 850 MHz Celerons and an application protocol running in a two-processor machine. Is this using Mysql or mem based dlr lists? No dlrs at all. This is because application protocol on the top of Kannel is of request - response type. So, one could say, dlrs are build into the protocol itself. Yes, it would be nice to test results with dlrs - this being more normal Kannel usage. What type of SMSCs are u connecting to? SMSCs were CIMD simulators running in same machine than Kannels. You must remember Kannel queue problem I mentioned. But this shows itself more in crashes than in performance. I report again when the queue problem is fixed. Aarno
