
Last week, David Linthicum’s SOA podcast continued a theme he’s been on lately, a discussion of BPEL’s “fallings” [sic]. I think he meant failings…but in any event, he mentions several times in the podcast that a post he’d previously written on this topic had generated quite a discussion (it did) and feedback from unnamed “BPEL vendors” (that’d be us; I can’t imagine why he didn’t name us. (-: )
Anyway, today after I heard the podcast, I asked Chris Keller, our founder and vp of development and one of the most knowledgeable people on BPEL in the world for his feedback. Chris has not only written the BPEL engine that’s at the core of our visual orchestration system (a VOS is a whole lot more than a BPEL engine), he’s active on the OASIS committees that are furthering the standards.
Chris gave me a lot of food for thought, and being in a playful mood, I thought it might be fun to that feedback into a Q&A. Sorta like a game show, with Mr. Linthicum as the contestant. The prize, for correct answers, is a free ActiveVOS license. Let’s see how Mr. Linthicum does…
Question 1: In the podcast, David says that a major problem with BPEL is that it’s synchronous.
Did David get it right? Click the arrow to find out… Then click here to read the correct answer
Sorry, no. Chris points out that the exact opposite is true. BPEL is asynchronous by definition.
Question 2: David says BPEL has a few programmer-level issues including limitations around request/reply exchanges in a heterogeneous architecture.
Did David get it right? Click the arrow to find out…Then click here to read the correct answer
Whooppss…wrong again. Chris points out that this isn’t a problem with BPEL as much as it is a problem with vendor support for standards that make request-response asynchronous, like reliable messaging. It’s one prime reasons for the WS-TF standards effort.
Question 3: David says BPEL has issues with failure recovery, exception handling and multi-programming model support.
Did David get it right? Click the arrow to find out…Then click here to read the correct answer
So very sorry. Chris points that exception handling is covered by the specifications and that recovery isn’t something that belongs in the specification at all (This is a major differentiator among vendors, and something ActiveVOS does brilliantly — ed.). As for multi-programming model support, we aren’t quite sure what David means, but if he’s talking about multi-tasking, the spec has that covered, too.
Question 4: David says BPEL is not very good at adding a human as part of the process and as SOA moves forward, he’s finding that composites and workflows are more applicable than simple service binding and extending.
Did David get it right? Click the arrow to find out…Then click here to read the correct answer
Close…but no cigar. Chris points WS-HT and BPEL4People have been added to the specification for precisely this purpose. While David’s criticism was once correct, it’s not applicable today.
We hope that you’ve enjoyed our little episode of The BPEL Game Show. And sorry, David, but you didn’t win our prize. However, anytime you’d like to be brought up-to-date on why BPEL is at the heart of SOA development, we’re happy to update you so you can win the next time.