With the merger of SUN and Oracle: What is the future of Open ESB a.k.a. GlassFish ESB? Is this a product which will be discontinued as Oracle has Oracle Service Bus (was BEA AquaLogic Service Bus)?
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.