I was scanning the latest book for the business rules community – Smart Enough Systems by Neil Rayden and James Taylor – which provides an excellent introduction to its main course of BRE and BRMS technology, with a side-dish of analytics, all in the cause of (automated) business agility. James [*1] is a proponent (and likely the inventor) of the term EDM or Enterprise Decision Management, which basically is about using business rule management and services together with feedback loops through analytics and data warehouses. Its interesting to note the similarities between this book and Vivek Ranadivé (TIBCO CEO)’s book The Power Of Now – the former concentrating on rules and analytics -driven agility and the latter on infrastructure-driven agility via SOA and messaging through to BPM. They could almost be episodes from the same series!
James occasionally refers in his blog(s) to Event Driven Architecture and Complex Event Processing. In his latest blog entry, he mentions how decisions are critical to real-time event-driven systems (although possibly it might be more accurate to say that event-driven systems are critical to real-time decisions). This is of course why a rule-based CEP solution, handling rules for decisions as well as event processing, is proving so useful.
Back to James’ book: I was intrigued to see how a rules + analytics specialist referenced CEP and Event Stream Processing. CEP is mentioned in relation to Business Activity Monitoring (covered earlier in this blog here) in the “EDM and the IT Department” chapter, whereas ESP (under the moniker “Stream-Processing Engines”) is covered under the “Data and Analytics” chapter. The latter section is interesting because the authors clearly separate the idea of stream processing / event identification, and the subsequent analysis stage (via an “analysis engine”) using temporal rules, calculations, aggregations etc. This is describing a subset of CEP, so perhaps one can one infer that a CEP engine is effectively an event-processing + real-time analytics engine [*2]? Interesting food for thought.
Notes:
[*1] Disclosure: James is an ex-colleague as well as the editor for the OMG PRR standard (in which TIBCO is a participant/contributor).
[*2] There are many other functions to an analytics engine outside of the scope of real-time analytics: for example importing large data sets and extracting potential classification rules from them. However, there sure is a lot of overlap between analytics, BI, and CEP…
