In the original proposal the supported runtime platforms were:
- Oracle BPEL Process Manager runtime
- Eclipse Test and Performance Tools Platform (TPTP) 4.0 Choreography
engine (based on BPEL)
Is this list still up to date?
Thomas Schulze wrote:
> In the original proposal the supported runtime platforms were:
> - Oracle BPEL Process Manager runtime
> - Eclipse Test and Performance Tools Platform (TPTP) 4.0 Choreography
> engine (based on BPEL)
> Is this list still up to date?
The design and implementation of our "runtime framework" will be
inclusive enough that someone could theoretically write a set of
bindings for any runtime that they like. Having said that, we will
select one runtime to "support" in this way as part of the project.
"Support" means we will write the actual bindings for this runtime. From
- This runtime will be open source. This means the Oracle runtime is out
(as is IBM's runtime or any other proprietary engine).
- TPTP is certainly a possibility.
- We had been following the progress of the Apache Agila project, as
this would be a possibility as well.
- There are other open source runtimes available as well.
Selection of which reference runtime we will target in the open source
project is an item in the Milestone 3 (May 15th) plan at the moment. We
hope to have more information at that time about the completeness and
status of each of the open source runtimes - with that information, and
a better idea of what our runtime framework will look like, we will be
able to make a more informed decision.
I'd encourage any discussion (here or on the mailing list) about the
various open source runtimes as possible candidates!
is a decision in May not to late? This would imply the then selected runtime
supports all BPEL issues which the project has implemented. This can
narrow the possible runtimes - in the badest case to 0.
Thomas Schulze wrote:
> Hi James,
> is a decision in May not to late? This would imply the then selected runtime
> supports all BPEL issues which the project has implemented. This can
> narrow the possible runtimes - in the badest case to 0.
It's a tough call. Regardless of the available runtime support, we aim
to have full WSBPEL 2.0 compliance in the editor, model and validator;
from that point of view, choosing a reference runtime too early might
mean that this runtime won't pan out according to our timeline. On the
other hand, waiting longer limits our runway to add support for this
runtime. We could certainly make the decision earlier if there appears
to be a clear winner among the available runtimes.