Jonathan Whitehead wrote:
> Could you explain this statement please (and the claim made elsewhere
> that this will improve performance)? I haven't yet used EJB with the
> webserver/browser applet architecture so maybe I don't understand
> something. But the way I see it, a java applet client can use RMI or
> JEREMIE to invoke EJB container class methods directly, so the location
> of the webserver is irrelevant,
Yes the location is irrelevant in term of programming visibility, but not in term of
performances, if the
RMI is optimized for local calls (calls within the same JVM) as in the case of
Jeremie. Jeremie local calls
are hundred times faster than remote calls. Of course everything cannot be in the same
JVM, and to define
the good distributed configuration for a given application is not trivial. What is
important is that this
is not done by programming, but only by configaration and that can be reconfigured
according to the needs
of the application. For simple applications a configuration with servlet (web
container) and ejb in the
same JVM can be the most efficient.
G�rard
G�rard Vandome BullSoftware
1, rue de Provence B.P. 208 - 38432 ECHIROLLES CEDEX - France
http://www.openmaster.com/ejb [EMAIL PROTECTED]
33 4 76 29.75.67 mobile 33 6 85 22 34 95 fax 33 4 76.29.76.00
----
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".
