Hi Shiril,

I like Winston's thoughts about doing your own comparison and better
understanding what you would like to compare.  ECperf was meant to be run by
anyone on any late-model J2EE server, so theoretically you could run it
yourself and at least you would have some measurement data to look at and in
the process learn more about how the server(s) work.  You can download the
complete kit from the JCP site.  ECperf results are also just for that
particular hardware software configuration so for example you could use it
to evaluate whether you have more than enough hardware or not.  In either
case waiting for the vendor results may not be worth it as they will likely
not run ECperf on a similar hardware os configuration to yours (or each
other's :).

<Vendor>
If you have your own EJB(s) already you could use a tool like Bean-test to
compare them.  You can deploy the same EJB(s) on each server, and then make
performance measurements up to and through the load (number concurrent
clients) you are expecting for your application.  This is also a great way
to measure the performance in a controlled repeatable environment.  Like
with ECperf, you will then at least have some numeric result data for
comparison and analysis.
</Vendor>

Similar to what Winston eluded too, you should consider what kind of
concurrent load numbers you are expecting for your enterprise application
and evaluate whether JBoss will get you there.  Good luck with it.

Chris Thompson
<Vendor>Bean-test</Vendor> Developer
<Vendor> http://www.empirix.com </Vendor>

-----Original Message-----
From: Winston Gnananayagam [mailto:[EMAIL PROTECTED]]
Sent: Thursday, March 21, 2002 2:45 PM
Subject: Re: JBoss Server Performance


Shiril,
           To make a reasonable apple-apple comparison, currently ECperf is
the only standard available(I'd like to know any alternatives myself). For
JBoss to participate in it, first it has to be J2EE certified. Then, it will
have to shell out the huge amount of $$$ to have its ecperf results
published. So, its better to do your own custom comparison based on your
project needs, instead of waiting for all this to happen.
            While you are doing your own comparison, keep in mind JBoss 3.0
is due out in a month. As u might be aware, it has clustering capabilities.
Using v3.0 in your performance eval would make more sense when comparing to
an already cluster aware WL implementation. Otherwise it wont be fair to
both products.
             App support has always been the issue for any Open source
product. But, I believe there are strong initiatives from the JBoss group to
alleviate it. U might find more about it, if you contact them directly.
                I've heard from developers that their JMS implementation
JBossMQ is quite not stable. So, u might want to be a little careful with
that. Even then, JBoss is built on a very modular architecture(JMX). So, u
should be able to remove JBossMQ easily and add a different implementation,
as u wish.
HTH,
Winston


----- Original Message -----
From: "Shiril Lukose" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Thursday, March 21, 2002 1:49 PM
Subject: [EJB-INT] JBoss Server Performance


Hi Gurus

      We are planning to use JBoss Enterprise server for our Enterprise
Application...Can any one please tell me the disadvantages ( IF ANY ) of
using JBoss server... Can we compare its performance  with Bea Weblogic
Server 6.1 ? Please Help...

Shiril

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to