Tim Bass
12-18-2008 03:43 AM
In August of this year, we issued the challenge,* The Bot Hunter: An Event Processing Challenge (Bot or Not). Not surprisingly, none of the “self-described CEP vendors” with grand claims of CEP greatness, responded.* Real-time detection of threats and opportunities, the core premise of CEP, is very different than algo trading, order routing and process orchestration. Solving real problems is quite different than marketing.

One would think that vendors would take advantage of a real CEP challenge to “show their stuff” versus the releasing marketing fluffy awards. * Folks believe in solutions tested by independent experts, not self-proclaimed proclaimations of greatness.* Every major company has a web site, so it is quite useful to create a CEP solution that can accurately detect automated bots in real-time (with low false positives and false negatives) and then block the bots with high confidence.

One problem is that it is non-trivial to write a set of rules that examine all the real-time web log entries and determine, with high confidence, which transactions are from bots and which are from humans.* This useful application is a bit “complex” and hence, a great CEP application.** Unfortunately,* CEP vendors, claiming “quick, low coding application development” can’t silence their critics!

Readers might recall from The Attack of the Spiders from the Clouds my musings on how cloud computing services can easily be used as a platform for massive denial-of-service attacks.* This is a real-time threat, ripe for a CEP solutions, but we cannot find a single CEP vendor that can meet the challenge.



Source...