JQ Tan wrote:
>
> I am trying to compare the performance between JOnAS and Weblogic EJB
> servers. The client application does simple lookup of about 200 records
> from 3 CMP entity beans. Generally it takes JOnAS using JEREMIE about 60%
> more time than Weblogic to finish the task. The improvement of using
> JEREMIE over RMI is not significant, it's about 10%.
>
> What's the experience of other people who have used both EJB servers? How
> to fine tune the JOnAS server, for examole the cache size for entity beans
> stored in cache, the pool size? Where would someone be able to find such
> information? How to use JEREMIE to get the most performance out of it than
> using the generic RMI, or what kind of environment is JEREMIE is designed to
> be best at?
Concerning Jeremie, communications are optimized for the case where
remote objects (e.g., beans) are colocated in the same JVM, getting
about 4O times better performance than generic RMI for that
particular case. It's true though that when communications are
truly distant, then the speed-up with respect to RMI is not as
significant (for the moment).
Kathleen Milsted
France Telecom R&D
Distributed Systems Architecture department (DTL/ASR)
28 chemin du Vieux Chene
BP 98, 38243 Meylan Cedex
FRANCE
E-mail: [EMAIL PROTECTED]
Tel: +33 4 76 76 45 94
Fax: +33 4 76 76 45 57
----
To unsubscribe, send email to [EMAIL PROTECTED] and
include in the body of the message "unsubscribe jonas-users".
For general help, send email to [EMAIL PROTECTED] and
include in the body of the message "help".
