I'm agree with Guillaume.
It's clear that Karaf is the premium container and most of the people
will certainly use it.
But there are some people which have:
1/ existing IT policies and hardware. For instance, in financial
companies, they have some kind of "print" to use only this kind of
application server (WebSphere, WebLogic, JBoss, etc).
2/ an ESB is never alone: a company would like to use an ESB in an
existing ecosystem. It means that applications are already in place and
the ESB has to use it. It means that some resources could be shared (for
instance, a transaction manager).
I'm afraid that without providing a WAR (which can be deployed in a JEE
server), we won't be able to address really enterprise users.
More over, from a technical perspective, as Guillaume said, it's better
to keep things unrelated to OSGi. OSGi specific components should be
clearly identify.
Regards
JB
On 06/29/2011 03:50 PM, Guillaume Nodet wrote:
What I really mean when talking about a Tomcat based deployement is
really to bring the value we'll work on (i.e. message analysis,
auditing, console and such) which are completely unrelated to OSGi, to
people that want to deploy camel routes using web apps in Tomcat. If
you look at the Camel survey
(http://camel.apache.org/camel-30-roadmap.data/camel-survey-2010.pdf)
a lot of Camel users don't really use OSGi, and there's no reason why
not trying to help those people when using Camel to build an ESB.
I'm not sure how it can be done technically, and I don't really want
to focus on this area specifically, but if it can be done without too
much pain, I think it would help. That said, I think Gert's points
are fully in line with my mindset in which Karaf is a much better
container, but not everybody is ready to pay the initial cost of
learning OSGi.
On Wed, Jun 29, 2011 at 10:45, Ioannis Canellos<[email protected]> wrote:
I am really happy to see the rebooting of Service Mix 5.
I agree with most of the points mentioned but I am skeptic about the tomcat
deployment.
I will skip talking about the advantages of such deployment option, since I
consider them obvious to all. I am worried on what it will mean to the pure
OSGi deployments (e.g. restrict the usage of OSGi APIs? Use of a framework
like pojosr?).
Any thoughts?
--
*Ioannis Canellos*
*
http://iocanel.blogspot.com
Apache Karaf<http://karaf.apache.org/> Committer& PMC
Apache ServiceMix<http://servicemix.apache.org/> Committer
*
