{ p } pruna.com.ar

Open source BPEL engines review: Conclusion

General

All evaluated projects are supported by several of the major Open Source foundations, that gives them a good background.

ODE is the only pure project. OpenESB and jBPM handle BPEL engines as components of a bigger architecture. This gives them a bigger community, but they can lost a bit the focus in this bigger place. Orchestra is in same way in the middle. It's an independent project, but at the same time is a kind of little brother for Bonita. It faces the biggest difficult for gaining community participation, but management and commercial people behind the project is really enthusiastic. I think that some guidelines for the use of collaboration software in their development team, that becomes available in the web, could be really helpful to show more activity, and then get this involvement from other people.

Engine and sharing efforts

PVM is for me a great idea, as a shared common engine. It doesn't mean that this engine is then better than others or not, I really don't know, but this synergy is great. And it'll be great if this becomes the base for other shared initiatives. An administration console (and it's API), and events API could be great choices. I also think that if PVM gives us an easy way to run different types of process with the same infrastructure (only changing the runtime), it's a great gain considering all kind of workflow that are really hard (or simply impossible) to implement in BPEL.

Achieve a complete standard support is clearly a hard task, that even biggest vendors couldn't get yet. So missing functionality is not necessarily a problem, but you should be sure that specific limitations doesn't become a problem for your needs.