CEP in BPM: But Will It Blend? (aka BPMI ThinkTanks 2007 report)

Reading Time: 2 minutes

Although the BPMI ThinkTank events are mostly of interest to the BPM, as opposed to CEP, community, it is probably worth reporting some of the CEP-related results from the presentations and Round Table sessions made at the European MiniThinkTank in June and main ThinkTank in July this year [*1].

BPM is of course traditionally related to explicitly-defined orchestrations of processes and subprocesses; to some extent this is the direct opposite of EDA approaches that include CEP (ie top-down process decomposition vs bottom-up event propagation). Does CEP and real-time event processing fit in the BPM world? In other words, will it blend?

The “Business Rules and Process” Round Table observations from both events indicated the need for BPM, business rules and events to be modelled together, not in silo’d in different application areas / environments. The business does not want to bring up 2 screens to view a process and its embedded rules. This sounds obvious, but is difficult to achieve. Below is the “backgrounder” for this particular round table (click for PDF)…

Other interesting observations from the sessions included:

1. A major Systems Integrator mentioned to the effect that “standards are not in the interest of the major SIs” [*2]. Of course, one of the major themes of BPMI Think Tank is “standards in processes”.

2. The new OMG BPDM standard (developed and supported by at least 2 vendors in OMG, although with an undefined relationship to the existing and far more popular WfMC XPDL standard) includes “behavioral happenings” (aka events). BPDM is designed to try and provide a model for all types of process, not just procedural steps: it will be interesting to see if it can also accomodate a CEP model, and whether it resides mainly in the MDA business/computation-independent or platform-independent levels [*3].

3. There was very little mention of EDA and events from any of the BPM experts present. The roundtable format meant that there were plenty of interesting discussions going on, but you only got to hear some of them – so perhaps event-processing came up in some other sessions. There was, though, some mention of the BPM-SOA boundaries.

For a fuller discourse on the proceedings, check out Sandy Kemsley’s blog or Derek Miers’ BPMFocus blog & wiki.

[*1] Disclosure: TIBCO were on the Programme Committee and chaired the “Process and Rules” round tables.
[*2] One might consider the lack of interest in standards by the major SIs as just a cynical observation by those of us involved in newer software technologies and their related standards – SIs profits are from selling “bums on seats”, not customer results – so it was very surprising to hear this as a statement of fact!
[*3] A rough ‘n ready test for a BPM standard’s positioning in MDA is “what type of rules does it reference?”. If it references the SBVR standard, then it is business-level (ie MDA CIM). If it references PRR then it is an IT-level (ie MDA PIM) standard. Both are equally important of course.