BPEL will not survive, long live BPMN

Note: the article depicts my personal view on role of BPEL in evolution of SOA and BPM. 

BPEL is for Business Process Execution Language, one of approved OASIS standards almost 10 years ago. It was promising standard for executable business processes in process automation at that time. Now future of BPEL is under question.

BPEL is applicable for 2 areas: services composition (ESB) and process orchestration (BPM). For ESB, it is usualy easier to use either 3rd gen languages like Java or C#, or, use ESB capabilities which usually are not BPEL based. For instance, TIBCO ESB – BusinessWorks use proprietary modeling notation for ESB/SOA, and BPEL support is optional plugin. Oracle ESB supports BPEL, though it also may be implemented w/o BPEL. Read the rest of this entry »

IBM Impact 2012 – Final 2 days

This year I visit one of most promised enterprise IT conferences in Las Vegas – IBM Impact. There are my impressions and take-aways from world-wide IT industry hardware, software and IT consulting monster – IBM.

This is wrap up of final 2 days.

Business Process Optimization. Is always possible if you know your goals. The ways include task flow optimization (reduce round trips, cycles, collect more information from one task), task time optimization, end-to-end scenario optimization, optimization for throughput. The human services and automated services interfaces have huge difference and should be designed in different ways. For example, human service requires a lot of denormalized information to be passed, while automated activity need only valuable information for algorithm-based decisions. Big difference in error handling (for human service errors are passed to user, for automated should be handled via policy, with retries and recovering). Optimistic (human) vs. pessimistic locking (automated).  Response time – for human interfaces is critical for UX, for automated activities response may be longer, but throughput is more important.

Optimization could take place w/o automation, though automation enable new horizons such as better visibility onto task progress, no chance for task to be lost, process intelligence and activity monitoring. Without optimization measurement are very difficult if not impossible. Finally, 20% of optimization should be applied in green-field (before-production) and 80% using production process analytics (“brown field” optimization).

ECM and BPM integration. So far it works one-way: BPM interfaces may display list of documents or document internals as well as manipulate those documents via automated activities. Great news that common interfaces to ECM are standardized via CMIS standard, approved at May 2011 and currently supported by major vendors (Microsoft, Alfresco, IBM, Oracle). No process triggering from ECM (though it’s possible via just web service call).

BPM in Healthcare. Sample case study from Ottawa Hospital proved that healthcare processes could be dramatically optimized with  use of BPM and mobile iPad interfaces. But this require not just one project, but several-years mobility strategy. iPad is now one of the way to better engage doctor and patient via sharing info on the screen, but initially it got a lot of resistance from doctors who did not want to carry such bulk devices in their pockets (now pockets are bigger).

Oracle vs IBM – series of talks, on how IBM outperforms Oracle dramatically (Oracle claims same but vice versa). I personally believe that WLS vs. WAS is question of company expertise in any of these and available other products from one of companies. Regarding Exalogic and PureSystem – IBM PureSystem is just released, and Oracle will add same features soon, I believe. For BPM portfolio IBM outperforms Oracle, especially with new IBPM 8.5. But in DB space Oracle definitely rocks over DB2.

Mobile security – again confirmed that this is strong trend, much stronger than development of mobile apps (which is comodized already).

And generally, the event was great chance to network with tons of experts and IBM customers.