I have been seeing an increasing amount on interest in the marriage between Business Transaction Management (BTM) and Complex Event Processing (CEP). On July 29th Dr Dobbs Journal published an article called Complex Event Processing: IT Liberator or Over-Engineering Hell? This article was about the synergy of BTM and CEP (although I felt it was rather biased towards one company). Also, last week Jean Pierre Garbani at Forrester published this blog in which he discussed the evolution towards BTM and CEP working together.
Business Transaction Management is a rapidly growing area of Application Performance Management (APM). BTM enables users to look into the transaction flows within their business and ensure everything is running as expected. BTM enables problems in transaction flows to be discovered – such as a bottleneck in an important business process. The really appealing aspect of BTM is it can do this without the need to change the applications in the business; BTM can “discover the transaction flows” by tapping non-intrusively into the flows going through application servers, middleware buses, business process management systems and other systems within the environment. Over time, BTM can build up a picture of the environment’s business flows, look inside the transactions and flag up immediately problems that can really hurt the business. Thus BTM works really well in legacy environments – not just modern SOA environments. And of course it appeals to business executives and operations users – not just IT users.
Complex Event Processing is the ability to correlate events flowing through a business - to identify patterns in real-time. These patterns might indicate opportunities and/or threats to the business that have just happened, are in the process of happening or are likely to happen right now. Events are occurrences in the business, such as stock market quotes in trading, call data records being generated in communications or packages changing location in logistics. An example of a real-time opportunity is a trading “statistical arbitrage” opportunity – to sell one instrument and buy another at a micro profit; another is the ability to upsell something to a customer who has just purchased an item on their credit card – based on their spending and buying patterns, their location and context. Threats to be detected include risk exceeding a certain key level in a bank or gaming fraud occurring in a casino. This kind of business level visibility and immediate response also appeals to business users as well as IT.
Listening to the descriptions of BTM and CEP, does it sound like there is a little overlap? Well there is some. What BTM is really good at is non-intrusively discovering process endpoints and the events they exchange – and then tracking these events. What CEP is really good at is correlating complex real-time business events in real-time, including arbitrary user-defined patterns, which can evolve over time as the business evolves. So it makes perfect sense to put these capabilities of BTM and CEP together. For BTM this strengthens the real-time correlation and pattern detection capabilities. For CEP this enables discovery of services without the need to do expensive and time-consuming instrumentation of the environment.
At Progress we have two leading products in the BTM and CEP categories: Actional and Apama. We believe that BTM and CEP capabilities are converging for certain business use cases, so as part of our Responsive Process Management (RPM) suite we now provide seamless integration between these capabilities. Of course RPM does much more than that. More on that later!