Interesting times here at TIBCO and beyond. In fact, it seems like according to the zeitgeist, it’s one of those tumultuous times where everyone is constantly reminded of pending change.

There’s also change of some sort in the political environment coming here in the U.S. I’m not going to take any political stands in this blog, but if you’re familiar with Greg Reemler aka “Greg the Architect” and the folks at Techrotech, they’re also contemplating change too. This one was too good for me not to share.

Lots of reasons for us all to re-evaluate how we cope with change… and CEP interestingly enough can help companies do just that. Recently, we concluded our event processing Online Summit and had the author, blogger and enterprise decision management proponent James Taylor as one of the keynote presenters. We had over 900 attendees listen and ask questions — many of which were along this line — What is the relationship and difference between CEP and BI (real-time and operational BI)?

The most noticeable difference is the event driven, real-time nature of CEP versus the query driven approach of traditional BI tools. CEP, at least in terms of TIBCO’s approach, is declarative which means the presence (or absence) of an event can determine how the business rules fire and ultimately how the system responds to the detected situation. In other words, BI makes you ask questions in order to get an answer and CEP gives you an answer based on real detected events.

CEP could utilize BI or analytics as a source of events or source of rules. You might be able to use BI to query a database to determine rules, but you would be better off applying the rules in real time in a CEP system to take advantage of CEP’s real time nature. This means you can detect fraud when it happens — not after it’s too late.

It’s a fundamental difference that event processing is based on data in motion versus the static “data at rest” in a database. CEP systems can usually detect more event sources and types than traditional BI which relies on a database. That’s the old way of doing things. CEP is also much better at finding out root cause and why something happened: users can drill down from a complex event to find what source events led to that situation (or complex event).

So that’s my story and I’ll stick with it. To those of you dealing with some sort of change, breathe in and hang in there. To those of you in the US, happy election – and make sure to get out and vote, even if it’s for Paris!

Finishing up BRForum this week were 3 sessions:

• A panel of worthies discussed questions on “Rules on the Web” in front of a large audience, chaired by John Hall, which turned into a panel on Rule Standards (all the panellists were involved in standards like W3C RIF, RuleML, and SBVR). So presumably “the web” is pretty irrelevant for most users.
•  Another panel of worthies then discussed “Emerging Trends and Decisioning” in front of a much smaller audience, chaired by Kristen Seer. This had a bit more of interest from a CEP perspective:
  • Jim Sinur commented that events, rules, processes and agility were coming together, pretty much mirroring Paul Haley’s view the day before.
  • Ron Ross “confessed”, if I heard correctly, that the biggest business value of business rules were the decisions they drive.
  • Jim Sinur commented on the move from a market-driven to a resource-driven economy.
  • James Taylor commented that predictive analytics on pre-bust data is not going to be any worse than making no decision analysis at all.
  • Jim Sinur added that there was a need to simulate rules against complex event (driven) scenarios.
• The last panel of worthies discussed Rule Standards, in front of an even smaller audience. Must have been the end of the day! But we had a good discussion on the merits and roles of PRR, RIF, SBVR, BPMN etc.

So was this a worthwhile conference? Yes, as usual there were a number of great sessions. Will it happen next year? I guess that depends on how the rule market adapts to the big guys being involved. And I apologize to the TIBCO customers present for not giving any updates on iProcess Decisions…

One of the nice things about Business Rules Forum is that the Rule Markup Conference (RuleML08) is co-located with it. So some of us change rooms today to discuss progress on rule standards, with TIBCO presenting OMG PRR, Ilog presenting W3C RIF, and IBM presenting SBVR’s date-time vocabulary. Is this relevant to Complex Event Processing? Well, later David Luckham presents on CEP aspects. And one of the slides shows the huge overlap between rules and, say, the TIBCO BusinessEvents CEP solution.

Apart from the PRR metamodel we also added some background slides on:

• the relationship between business rules, decision models and rules,

  

• the RuleML classification of executable rule types, extended with state transition rules, and

  

• a graphic of the various standards versus other rule representations.

  

Some of these topics were a cause of confusion for the BRForum folks earlier in this week, and these graphics are subject to change and feedback…

Paul Haley, no doubt somewhat bemused that his old company (still using his name) is now under the Oracle banner, presented his proposal for the development of an upper ontology joining together the concepts of events (and event processing), process, state and rules.

Not surprisingly, given that TIBCO is here presenting event processing in a business rules conference, and is a vendor of a leading BPM product, this is somewhat relevant to us. Points covered included:

• Few rules are used to drive processes to make them dynamic; normally BPM vendors concentrate on single-execution point decision services.
  That is a fair point, but at TIBCO we would point to iProcess Conductor, and even BusinessEvents CEP being used to control iProcess workflows.
• Conventional rule systems languages don’t handle temporal aspects very well.
  Again, true. Paul’s example was the statement “underwriting precedes approval“, which seems more to me to define a process diagram constraint rather than something a BRMS would handle. 
• Process people know about events, they just don’t dwell on them.
  I think this quip went wa-a-a-y over the heads of the audience…
• Decisions ignore process state.
  More likely, the process state is implicit in the decision service. Which certainly implies a potential for issues with re-use and maintenance.
• The semantics of process, state, rule and event are not joined, but need to be.
  That is Paul’s thesis.
  • He commented further on the lack of semantic progress (bravely, considering the SBVR experts in the room) with the example that Siebel represents money types as an amount, currency, and time. The latter is needed as conversions between currencies are time-dependent.
  • I had always thought semantic-meeting-of-the-minds was the role of OMG BPDM, proposed as the “semantic BPMN 2.0″. But BPDM’s capability to model state and rules (and probably continuous events) is still unproven, AFAIK.

Paul tag’s this area as a whole as “Semantic Corporate Performance Management”.

Just to show how close the BRE/BRMS and CEP fields are to converging was ably demonstrated by Kevin Forbes of Healthways in his presentation on the massive data requirements and event loads being handled by a conventional BRMS, albeit with much of the “heavy lifting” being done by a Coherance cache, with traceability being handled by Amberpoint. Points of note were:

• “Network and data access are the limiting factors” in decision automation.
• Distributed cache and data grid approaches are “key for decision service performance” and “multi event correlation”.
• They are looking at / planning to do real-time analytics in terms of updating rules from analysis of data on the datagrid. Which makes perfect sense.

In some respects these guys have constructed their own custom equivalent (albeit with a web services level for invoking the rules, on a Microsoft platform) of a CEP tool like TIBCO BusinessEvents. Needless to say I thought it was a great presentation…