Would you then ignore the entity bean feature in EJB totally, and simply
call these mapped object from your session beans? What effect does this have
on transactions?. Surely the OR mapping tools would be as slow as the CMP
entity beans - unless they are explicitly told to save,restore and delete
themselves from the database.
Another thought would be to have a caching facility and some sort of option
in jonas that told the CMP Entity beans to only synchronize themselves with
the database when told to do so. When I was introduced to EJB, using an app
server called Apptivity from Progress, I was lead to believe that when an
entity bean was located and 'read', all the persistent data was cached in
memory. Does Jonas cache anything at all?
Finally I've just had a quick look at this posting on the JonasDev mailing
list:
http://192.44.60.125/messages/JonasDev/0176.html
Is this performance issue anything to do with RMI and not the quantity of
SQL statements issued?
Does anyone from Bull want to comment on this subject?
Regards
Robert Hargreaves
-----Original Message-----
From: Muly Oved & Netvision [mailto:[EMAIL PROTECTED]]
Sent: 10 October 1999 13:06
To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
Subject: Re: jonas-users: Entity bean performance
But I feel that coding the business logic directly with JDBC is a problem
from coding style point of view. I will preffer to model my database as a
collection of Objects with relation between them. There some OR mapping
tools available for that, But I am not sure which way to go myself.
Regards,
Muly.
