I had mentioned to Ben last night that it was interesting that the large consultancies were missing from the presenter list. The drought was broken in the Spring Batch presentation when Rod Johnson from Interface 21 was joined by a couple of Accenture guys. It appears that Accenture is sponsoring the development, which is the first time I have heard of them being involved an open source project.
At a high level, Spring Batch aims to bring a lot of the Spring concepts to batch processing in Java. I am pre-disposed to like this, and after the session I have seen nothing to dissuade me from that.
The session covered a fair bit of detail of the sort of functionality that Spring Batch supports and the type of business scenarios that it covers. I certainly felt like it offered the sort of functionality that Shine could use and would gain great benefit from. For instance, they will have functionality in the framework to help manage scenarios such as a single failure not stopping the whole batch, restarting a batch run and monitoring batch processing statistics through a JMX console.
In addition, the aim of the framework is to separate concerns between POJOs that encapsulate business logic and the architecture that defines how the batch process works. This means that batch functionality should be able to be built independently of the techniques required to meet the batch processing requirements such as commit scope, parallelisation, etc.
Interestingly, a lot of the concepts match the batch processing framework that we built for our Shine products NBV and NBM. Batch, job, step, input source, output source, data access. Nice to see that the pattern we used was a good one!
However. The project has yet to release any code, and won’t do so until June at the earliest. Even then it is Beta (although interestingly they are going straight to 1.0, so I guess it is production ready? Really?). Based on this, I think we have to mark this as a technology that we should investigate, however anything talked about in the presentation is yet to be backed up by either industry adoption or any level of maturity.
Personally, this was probably the most interesting ‘new’ technology to come out of JavaOne that might have a real and immediate impact on how Shine builds solutions for our clients.