Oracle is a tough one to predict.

OpenEJB is easier to predict, because it s open source. It appears to still be active, so it s likely to carry on for another year. Who can see further out than that?

The larger question is: What will the adoption rate of EJB3 be? Has the world passed the EJB model by?

Personally, I don t consider EJBs to be important anymore. I prefer Spring. As long as it s vibrant, I can t see any compelling reason to go back to EJBs.

I m not certain about OpenESB, because I m ambivalent about ESBs in general. I ve seen them as a piece of a "BIG SOA" selling strategy on the part of vendors that was big on promise and short on delivery. An ESB can be either a centralized mediator for all your web service traffic or a single point of failure bottleneck, depending on your point of view. I think that a lot of the functionality that is usually centralized in ESBs (e.g., transformation, routing, logging, auditing, etc.) could be done better and faster in hardware - think Data Power or smarter Cisco switches.

not a marketing trick but here is a white paper which provides a realworld insight on the future of OpenESB. It s past, Present and Future. http://www.logicoy.com/userfiles/openesb-past_present_future_news_letter.pdf

According Oracle, GlassfishESB/OpenESB will be supported (but I m not sure that it indicates that will be evolved and maintained). If you see the migrated glassfish esb / openesb site you can see that the product won t evolved since 2009 (it s a mess because it was a really integration product). Maybe the "forgerock.com" company (who is maintaining some famous and interesting Sun products like OpenSSO and OpenDs) could be interested in maintain and evolve this good integration product and continue the development of fuji (GlassfishESB 3). At this time you can compare the evolution of the product with another option (in ohloh net), and you ll see that unfortunately javaCAPS or OpenESB or GlassfishESB or fuji development has not evolved since 2009.02.

Oracle Strategy seems to be the soa suite evolution, that is a really good product, but have another license terms.

See this related links too:

• blogs.oracle.com/theaquarium/entry/openesb_new_beginnings
• openesb-users.794670.n2.nabble.com/News-update-on-OpenESB-and-Oracle-td4578158.html

NOTE that Oracle Service Bus (OSB) and BEA AquaLogic Service Bus were two different products and the first one was added to the soa suite implementing the SCA integration strategy.

This has been a very important concern shared by all those OpenESB/GlassfishESB customers. Although it is still not clear on the exact course it will take. Oracle has been largely promoting its Oracle Fusion middleware which is nowhere comparable to the flexibility and strengths of OpenESB/GlassfishESB.

Lately the OpenESB community has become active again and all credits to he noble work by Pymma , Logicoy and few others.

I personally liked OpenESB/GlassESB very much when compared to other available ESBs.

Oracle has a long and distinguished history of buying companies that developed java appservers, renaming that company s appserver as "Oracle AppServer", and quietly ditching the previous one. As a result, each major version ofOAS is completely different to the previous one.

First, they wrote their own (it was crap, noone used it), so they bought Orion and renamed OC4J to OAS. Then they bought BEA, and are in the process of turning Weblogic into OAS. Now they ve bought Sun, and Glassfish s technology is going to be better going forward than Weblogic (which, IMO is a bit of a dinosaur), and so expect Glassfish to suddenly be renamed OAS at some point in the future.