Thanks for sharing your migration experience. It is good to know that
you found it easier to migrate your apps.
I recommend you stick to Java EE platform for future enhancements, which
means using the modular techniques that Java EE platform provides, e.g.,
separating business logic to an appropriate EJB module, using JPA for
data access, and war module for presentation. Even if we base GlassFish
V3 on OSGi, your application would continue to work. We shall try to
provide extension points so that you can take advantage of the OSGi
environment, but the app would still be a pure Java EE application.
Thanks,
Sahoo
Kristian Rink wrote:
> Folks;
>
> as some on these list might remember, I have been into migrating most
> of our tomcat/Spring based war contexts to a glassfish v2 environment,
> and so far, it seems we have done it. Though minor issues still remain
> unsolved (most notably logging configuration), the experience in
> general is pleasing...
>
> ... which, by now, makes me ponder a serious question of strategy: What
> next?
>
> So far, most of our applications are packaged as .war applications
> simply because, talking about tomcat, there is no "better" way of doing
> so. In terms of glassfish, my next step would be actually extracting
> non-web-relevant logic to, say, EJB modules and seperately dump them to
> the container. This would work I think, I even have a pretty good clue
> now how to do so using maven2. Nevertheless, I am not sure anymore
> whether this is the "way to go"... : A lot of fuzz has been around just
> as of recently, involving the Spring Application Platform and Glassfish
> V3's likely-to-be-OSGi foundation. Things seem to change rather fast.
> And I am into making a long-term decision which is thought of as to be
> the foundation for migrating not just the bunch of .war apps we have
> done so far but a system a whole lot more complex, so far implemented
> in a proprietary backend DMS environment. No matter which decision I am
> likely to make now, it is likely to follow us at least for the next
> five to seven years.
>
> So, my question to those deeper involved into glassfish development:
> What would be the "sane" way to go here? Rely upon Java EE 5 and go for
> an EJB 3.x approach knowing that its successor technology is already
> looming beyond the horizon, yet knowing that Java EE 5 so far is the
> "stable state-of-the-art"? Go for the new, unstable technologies and
> have an environment which is likely to be stable more or less the same
> time we will finish most of our migration work? Go for a middle ground,
> trying to be as technology-independent as somewhat possible knowing
> that no decision in this field of view really will have a "strategic"
> quality?
>
> What are your thoughts?
> Best regards,
> Kristian
>
>