axis-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Frank Cohen <fco...@pushtotest.com>
Subject Re: Performance
Date Fri, 07 Mar 2003 16:45:55 GMT
I've seen similar concurrency problems in Axis while I was doing a 
performance project for a customer. This was last Summer when Axis was 
still very, very new. I haven't had a chance to dig into the code but I 
suspect there is a problem with the way Axis handles threads and 
globals.

If you need to compare Axis performance to other SOAP implementations 
then take a look at my open-source performance kits at:
http://www.pushtotest.com/ptt/kits/index.html

-Frank

--
Frank Cohen, Founder, PushToTest, http://www.PushToTest.com, phone: 408 
374 7426
Come to PushToTest for free open-source test automation solutions that 
test and monitor
Web-enabled applications, especially Web Services for scalability and 
reliability.




On Wednesday, March 5, 2003, at 09:46 AM, tim@calidris.force9.co.uk 
wrote:

>
> I'm seeing dreadful performance using Axis with multiple concurrent 
> calls.
>
> I have a simple test app which runs some PL/SQL.  In single user mode, 
> it works quickly across an Axis call (Axis on both client and server), 
> taking about a second to complete.
>
> I've ramped up the app and have 10 concurrent threads doing the same 
> call.  The PL/SQL is now taking between 140ms and 1372ms, which is 
> fair enough, and the actual call after the Axis bit (ie inside the 
> method that is exposed using Axis) varies between 2173ms and 2424ms.
>
> These figures I can live with, and I'm sure there are further 
> optimizations I can put into my architecture to reduce this time.
>
> However, the total call time varies between 7331ms and 10445ms.  I 
> appear to be losing between 5 and 8 seconds simply doing Axis comms.
>
> Anyone got any tips on how to speed this up, or ideas of why it's so 
> bad?
>
> Axis 1.1rc1, JBoss 3.0.4 app server, jdk 1.4.0, windows NT.
>
> Tim.
>
>




Mime
View raw message