[ http://jira.jboss.com/jira/browse/JBAS-1180?page=history ]
Scott M Stark updated JBAS-1180:
--------------------------------
Assign To: Francisco Reverbel (was: Scott M Stark)
Description:
SourceForge Submitter: marklittle .
I registered the following issue
(http://www.jacorb.org/cgi-bin/bugzilla/show_bug.cgi?
id=526) with JacORB bugzilla a couple of weeks ago, but
no response so far. I think it's sufficiently important to
register here too in case you are able to do something
for JBoss.
In the current implementation of JacORB (2.2), if I
register an interceptor with the ORB, then
it appears that all invocations to all objects go through
that interceptor,
even if they are co-located with the invoker. I
understand that rational given
the PI specification and the original intention behind
that work: to make
local invocations and remote invocations similar.
However, most ORBs provide a way of doing a local, no-
interceptor
optimization. The JDK 1.4 ORB is one example (though
that takes the other
extreme and won't invoke interceptors on any local
object).
Orbix 2000, for example, supports multiple ORBs and
POAs within the same VM,
and interceptors are registered on a per-ORB basis, not
a per VM basis as
appears to be the case with JacORB. You can also turn
on local optimizations
for the individual ORBs.
The reason for asking for this optimization is
performance. With a very simple
test, where the interceptor does nothing, we see a 66%
degredation in
performance for local calls. For example, with no
interceptors we can get > 3000 invocations, whereas
with a "null" interceptor this is around 1000.
was:
SourceForge Submitter: marklittle .
I registered the following issue
(http://www.jacorb.org/cgi-bin/bugzilla/show_bug.cgi?
id=526) with JacORB bugzilla a couple of weeks ago, but
no response so far. I think it's sufficiently important to
register here too in case you are able to do something
for JBoss.
In the current implementation of JacORB (2.2), if I
register an interceptor with the ORB, then
it appears that all invocations to all objects go through
that interceptor,
even if they are co-located with the invoker. I
understand that rational given
the PI specification and the original intention behind
that work: to make
local invocations and remote invocations similar.
However, most ORBs provide a way of doing a local, no-
interceptor
optimization. The JDK 1.4 ORB is one example (though
that takes the other
extreme and won't invoke interceptors on any local
object).
Orbix 2000, for example, supports multiple ORBs and
POAs within the same VM,
and interceptors are registered on a per-ORB basis, not
a per VM basis as
appears to be the case with JacORB. You can also turn
on local optimizations
for the individual ORBs.
The reason for asking for this optimization is
performance. With a very simple
test, where the interceptor does nothing, we see a 66%
degredation in
performance for local calls. For example, with no
interceptors we can get > 3000 invocations, whereas
with a "null" interceptor this is around 1000.
Environment:
Priority: Major (was: Critical)
> bi-pass JacORB interceptors for purely local invocations
> --------------------------------------------------------
>
> Key: JBAS-1180
> URL: http://jira.jboss.com/jira/browse/JBAS-1180
> Project: JBoss Application Server
> Type: Bug
> Components: IIOP service
> Versions: JBossAS-3.2.6 Final
> Reporter: SourceForge User
> Assignee: Francisco Reverbel
>
>
> SourceForge Submitter: marklittle .
> I registered the following issue
> (http://www.jacorb.org/cgi-bin/bugzilla/show_bug.cgi?
> id=526) with JacORB bugzilla a couple of weeks ago, but
> no response so far. I think it's sufficiently important to
> register here too in case you are able to do something
> for JBoss.
> In the current implementation of JacORB (2.2), if I
> register an interceptor with the ORB, then
> it appears that all invocations to all objects go through
> that interceptor,
> even if they are co-located with the invoker. I
> understand that rational given
> the PI specification and the original intention behind
> that work: to make
> local invocations and remote invocations similar.
> However, most ORBs provide a way of doing a local, no-
> interceptor
> optimization. The JDK 1.4 ORB is one example (though
> that takes the other
> extreme and won't invoke interceptors on any local
> object).
> Orbix 2000, for example, supports multiple ORBs and
> POAs within the same VM,
> and interceptors are registered on a per-ORB basis, not
> a per VM basis as
> appears to be the case with JacORB. You can also turn
> on local optimizations
> for the individual ORBs.
> The reason for asking for this optimization is
> performance. With a very simple
> test, where the interceptor does nothing, we see a 66%
> degredation in
> performance for local calls. For example, with no
> interceptors we can get > 3000 invocations, whereas
> with a "null" interceptor this is around 1000.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://jira.jboss.com/jira/secure/Administrators.jspa
-
If you want more information on JIRA, or have a bug to report see:
http://www.atlassian.com/software/jira
-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
JBoss-Development mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jboss-development
