BPM Futures

BPM with IBM Websphere

I attended an IBM ‘Business Agility’ workshop at Sydney’s Shangri-La Hotel yesterday – the first IBM event to feature BPM that I’ve managed to get to since the Lombardi purchase. It was a Websphere event, which meant that it included Lombardi and excluded FileNet, so I was a little concerned that the BPM section might be dominated by talk of process orchestration and middleware layers, rather than end-to-end processes.

I needn’t have worried. The Websphere team has embraced IBM Lombardi (as we must now know Teamworks) with great enthusiasm, and started a day of real (yes, live) demonstrations with several that showed off Lombardi to good effect. Point and click SLA setup; process stats (such as wait or execution times) displayed through a mouse-over in the unified process model-define-simulate view; colourful monitoring views populated with whatever defined field you required – just click that checkbox on the field definition dialogue; and so on.

There were also Websphere Dynamic Process Edition (Process Server, as was) demos. The emphasis there was on architecture, integration and transactional integrity. The latter featured a high-wire demo, with 100 updates to two databases on separate servers, interrupted by the speaker who pulled out the connecting cable to the second (Oracle, as it happened) with a flourish. 56 updates had been processed successfully and, to the relief of all, the other 44 were in a ‘failed’ queue, from which they were dispatched – to a successful completion – by a single click on the ‘resume’ button once the cable was re-connected. We were told that the product was unique amongst BPMS’s in fully supporting two-phase commits, with resume, restart and ‘compensate’ options for system administrators.

All of which provided – to this viewer – a pretty clear, if unspoken, message. For the human side of BPM (the typical financial services back office, perhaps), Lombardi is IBM’s answer, packed with business-friendly features. Alternatively, if the business depends on multiple integration points that require sophisticated sequencing, error handling and recovery options – bullet proof delivery, in other words, WDPE does the job (telco provisioning comes to mind). And for the business that needs both, well, integration between the two is currently available through web services, with work under way to convert Lombardi to IBM’s Service Component Architecture, the basis of the Websphere product range.

One other demonstrated feature of WDPE that I liked, by the way, is the easy way in which routing rules can be changed without re-deploying (or even opening for editing) the process itself. This seems like an obvious feature, but by no means all BPMS’s share it. Isolating the change eliminates the need for system and regression testing and even (depending upon the process design and one’s perception of risk) UAT. Now there’s something that offers Business Agility.


Trackbacks & Pingbacks


  1. David, thanks for the truly interesting post on IBM’s direction. What this however truly shows that IBM WebSphere by itself does not delivery any of the features that Lombardi offers, which I consider in terms of Human BPM one of the best products on the market. I will not elaborate on how and why our ISIS Papyrus Platform surpasses Lombardi, especially in terms of ADAPTIVE capability — meaning learning feedback into the template. Neither WebSphere nor Lombardi offers integrated content management and generation as does Papyrus. You can integrate Filenet for the capture part. That means there is already substantial integration work necessary between WebSphere, Filenet and Lombardi to come to a reasonably functional set of the business actor. Despite the cool Lombardi features … STILL BIG PROJECTS.

    Two things are apparent from your observations. WebSphere and Filenet process capabilities are outdated, clunky and horribly expensive to implement and certainly NOT AGILE. Thats why IBM bought Lombardi, obviously. The other one is IBM’s other move towards their ACM (not Adaptive) but Adavanced Case Management ‘vision’. When challenged on their IOD presentations in Rome, where they showed cool user case creation, they had to admit that most of it was hardcoded demoes. Their partners will do the integration work, IBM said. Well, here is another piece of the puzzle that a business has to integrate.

    BTW, if IBM is so amazed (as apparently you are) that processes that do not get a positive feedback go into an error state and have to be restarted, then what in the world has IBM been doing up till now? In Papyrus you don’t even have to restart the process if a network connection fails, as the task that expects the data feed simply remains in ‘Waiting for service’ state until it happens or times out and gets routed to a supervisor. All other non-dependent tasks can be executed in the meantime.

    Anyway, thanks again for a great view behind the scenes of IBM’s BPM world. Too bad for them that standalone TLA’s are on the way out. Clearly, Lombardi saw that coming and that’s why they sold to IBM. Clever move …

    | Reply Posted 7 years, 2 months ago
  2. * the wall says:

    Its like you read my mind! You seem to know
    so much about this, like you wrote the book in it or something.
    I think that you can do with a few pics to drive the message home a little bit, but instead of
    that, this is fantastic blog. An excellent read.
    I’ll definitely be back.

    | Reply Posted 3 years, 2 months ago

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: