I've just "discovered" ODE and I'm pondering the merits of using BPEL for orchestrating calls.
My current understanding is that BPEL's goal is to provide a means of exporting a "packaged" web-service that resolves to a series of web service calls based on some control logic.
Now, I understand that this is useful if the "package" contains decision-making logic with a few alternative flows that consult business rules and such. However, for simple use cases, it seems an overkill.
For example, in my environment we have a very simple 2-step process that is being carried out (i.e. two web services are called in a specific order). There's no complex rules to apply, no alternative flows to consider, nothing. It's just "step1: class WS-x, upon success step 2: call WS-y, otherwise return immediately".
Now, converting this process to a BPEL package and having the application call that, seems to me like throwing away good performance just to be able to use SOA buzz-words in a design (i.e. wear an "I can BPEL" badge on one's chest).
Is ODE really inappropriate for such use-cases or am I just being to technology-phobic?
I adhere to KISS.
So if you do not already is using BPEL or have prior experience using it then I would think its overkill for your use case.
Any decent engineer and architect can understand the use-case you have involving only 2 webservices.