CEP: The Tech That Dare Not Speak Its Name

IBM has decided to go after the complex event processing (CEP) market with InfoSphere Streams a bit sooner than expected (by about a year, it seems.) The Product Formerly Known As System S (apologies to Prince – but at least it doesn’t have a weird symbol instead of a name) has been released into the wild because it’s ready enough. Or because Sam Palmisano wanted to talk about it at a recent analyst event. Or (Curt Monash theorizes here) because Microsoft announced that their CEP offering will be in SQL Server Real Soon Now. IBM tells me it’s serious about the – wait for it – real time analytic processing (RTAP) market. In standard IBM fashion these days, this spans a number of related acquisitions: the Aptsoft acquisition for WebSphere, associated with  Business Event Management (BEM), which IBM wanted to distinguish from Business Process Management (BPM). IBM subsequently enhanced its BPM capability by acquiring iLog, a leader in Business Rules Management Systems (BRMS). And the SolidDB in memory database adds a nice capability for temporary persistence, if such a thing makes any sense to you (frighteningly, it does to me.)

In any case, the messaging seems a bit hasty – the first IBM link above points to a Business Event Processing (BEP) page on IBM’s site, which has only a placeholder so far (May 18) for InfoSphere Streams. It leads with WebSphere Event Processing (WEP) – “Correlate without Coding”, perhaps a subtle reference to Coral8, which is the engine in WebSphere’s RFID product. Nomenclature galore. Are you confused yet?

Convoluted messages aside, entry into this market by IBM and Microsoft signals that it may well be time to take it all seriously again. The bloom seemed to be a bit off the rose for the market(s), as interesting as it (they) seemed a while ago when Progress Apama, Coral8 (now part of Aleri), Streambase, SQLStream and others were chasing a few exciting-sounding use cases in the financial services markets. Everyone agreed that there were lots of applications with a need to filter, reduce, enrich, and otherwise process high-volume data flows with very low latency requirements. But things seemed to settle down into a few very specific use cases that Monash describes in several posts on his sites. So, has anything really changed?

Yes. At the very least, IBM’s plans to enter markets change things. It ramps up a formidable engine; the team has presented at conferences like WebSphere’s Impact, and Information on Demand is coming in Europe. It’s presenting at the Security Industry’s FMA on Wall Street in late June with the CIO from TD Securities, an early reference customer. And IBM has been doing internal sales webcasts and calls with individual territory management teams. It’s established an EMEA competency center, loaded the product into the Dublin cloud in Ireland, and hosted scientists from the Tokyo lab.

In another post, I’ll get into the technology and some of the interesting opportunities IBM sees – this is worthy of some serious discussion. But for now, at least, it would be nice to decide, branding aside, what we’re going to call all this stuff. “You never step into the same stream twice” may be wise, but it ain’t product marketing.