A recent CEP RFP requested marketshare data for the “CEP market“, worldwide and in specific territories. Presumably the contracts departments for large enterprises are hoping that software development tool purchases are tracked by some industry body (Nielsens maybe) just like Cornflake packets and Persil washing powder. Unfortunately, this is not the case - and of course the definition of “customer” is itself open to interpretation (someone who uses the software? someone who paid money for the software? a licensee?). Some companies also embed “CEP software” in larger categories of software (e.g. messaging or BAM); others sell only stacks that might include CEP licenses, without necessarily any use of said licenses.

Nonetheless, it is interesting to try and find some data points here. Software companies generally make “customer numbers” claims via press releases and in analyst reports, as well as SEC etc reports, and these can be collated and cross-referenced to try and come to a “big picture”. Doing so finds some interesting details such as:

• the most recent available data (for 2008-2009) gives TIBCO BusinessEvents the leading market share figure of 28% (by number of customers)
• this market share number has pretty much remained constant (except for one anomoly year, when it was 10 points higher!) for the previous 4 years
• other established CEP vendor shares are 21%, 18% and 14%  - a pretty even spread overall, if the data is accurate
• the cutoff date for available data (early 2009) excludes any major effects from the large companies getting involved in CEP (IBM, Oracle, Microsoft) which, along with CEP market growth and the addition of many new startups, will likely make future numbers difficult to interpret
• vendors 2, 3 and 4 (in order of market share) report being almost exclusively (>80%) in financial / capital markets - so probably ~50% of business in CEP (to-date) is in the financial area
• overall market growth, year on year, was up to 60% initially, but dropped to just over 30% in the latest numbers. Something to do with the financial market meltdown, perhaps?

It seems that publishing the CEP Market survey is a “preceding event” for some change in the market - no sooner do we update the 2010 market view then another vendor adjustment takes place. Last month mid-tier RDBMS vendor Sybase got acquired by ERP vendor SAP (completing a chain of new corporate identities for Coral8 employees who became Aleri in Mar09, Sybase in Feb10, and now SAP in May10)…

We also made a few other updates, such as adding in the IBM R&D Amit product - the one that IBM management decided to bypass when it acquired Aptsoft to be its “version of Business Events” in 2008. I hadn’t realised that Amit was embedded in a few other IBM product offerings… so is worthy of inclusion in the chart. I adjusted a few other start times based on evidence of “first customer project” too - but clearly that data is still more subjective than it could be.

So - what will change next?

So more consolidation announcements mean its time to update the CEP Market graphic. And an observation that the 10th anniversary of CEP technology is nigh, per some vendors’ start dates. Probably we should really celebrate the first decade of CEP on the anniversary of David Luckham’s book in 2 years.

I had a comment on the “start dates” for solutions in the last iteration… however my excuse is that neither I, nor Google, have good access to this (occasionally sensitive) information. I have tried to set the “start point” for a “CEP offering” as the first customer announcement / sale, except for the solutions which no longer exist (which are simply there for as a historical record). Of course, updates, suggestions and ommisions gratefuly received.

One thing I will try and expand for next time is the *type* of CEP technology used or offered. For example, which of these offer cache, or graphical dashboards, or other techniques? It might also be useful to have a rough idea (where available) of the number of users / market share, which is data usually given to the analyst companies.

I thought it would be a good time to update the CEP Market diagram. This covers all the main event processing application development tools (as far as I know). Compared to, say, the EPTS membership list there will be a mismatch - some EPTS members have not announced products yet, are in conventional (as in, not continuous / complex) event processing, or are focused on research.

I also thought it might be a good idea to bend the original rule for inclusion to add some of the fruits of the main players in CEP such as Dr Luckham’s Stanford spin-off ePatterns, as mentioned in his “Short History of CEP”.

It’s quite likely we’ll see more CEP merger and acquisition activity in 2010. Meanwhile it is the BPMS market that is currently in the news with a spate of takeovers (i.e. IBM with Lombardi, and Progress with Savvion). BPM is of course related to (and indeed a type of) event processing, albeit with a focus on standard operating procedures and processes  (exemplified by the “human-oriented simple event processing” that is workflow); but then, BPM is also considered a much larger market than CEP - at least in the first decade of the the 21st centory.

TIBCO’s contributions to event processing innovations in 2010 start appearing in merely a few weeks - it should be a good year!

Notes:

Start dates for tools and tool classification are not guaranteed:

• Start date is based on “available data” for “commercial delivery” (hence excluding lab-only “products”)
  
• Font size is meant to give a very approximate indication of number of (CEP tool) customers and thence indication of market importance. Yes, VERY approximate!

Yet Another Analyst report means that Forrester has joined Gartner, IDC, and Bloor (and possibly a few others) with coverage of, and comparisons between, CEP vendors.

The Forrester CEP Wave for Q3 2009 was done by Mike Gualtieri and John Rymer who, whilst perhaps being new to the field of CEP, have previously covered the BRE and BRMS market [*1] for Forrester. Although many people realise there is both an affinity and synergy between event processing and business rules, one suspects the authors are keeping such thoughts to themselves in describing CEP as a “hot new enterprise middleware category” (per Mike’s blog). Presumably they use this description because CEP systems can be event sources as well as sinks - but on the other hand so can many other types of system, so overall “middleware” probably isn’t the best description one could use.

Interestingly this report does not use or refer to Forrester’s previous classification of event processing systems.

The good news for CEP users is that there are clearly no “weak” offerings in the CEP market!