Si, I did it with a shell script which started the standalone-flashplayer with a swf out of a loop. Because of that, the lack of performance of the client computer is the greatest problem. Normally, a test with more machines would make more sense for a actual stress test.
Saludos, Micha Jonathan Hernández Velasco wrote: > El mié, 06-12-2006 a las 12:53 +0100, Michael Plehm escribió: >> Hi, >> >> I have done some testing. I tested with 100 clients connecting in a >> local network. Only connecting negotiation leads to high cpu usage. >> While streaming there is nearly no load, even less with mp3 streaming. >> I tested on a dual xeon with 4gb ram on linux kernel 2.6.17. >> Actually, a number of 100 concurrent users with pure streaming should >> not be a problem for any streaming server. In my opinion the hardware >> limits this ... same thing like on most java application servers ;-) > > Good news. And how did you make the connections? Did you make any script > or java class to do testing? > _______________________________________________ Red5 mailing list [email protected] http://osflash.org/mailman/listinfo/red5_osflash.org
