In his Outside the Box blog, Todd Biske writes about the (lack of) Uptake of Complex Event Processing. He points out several reasons for the lack of uptake — the primary reason being a shift in mindset that has not taken place in most businesses.
Todd goes on to suggest that organizations should start a CEP system just by publishing the events and collecting statistics about events. His suggestion is a very safe and risk-averse approach. It also a suggestion that I disagree with.
Don’t get me wrong, I agree that collecting statistics about your events is important. I don’t agree that just publishing events and then collecting statistics about those events is enough.
Why not publish the events and then throw a little fuel on the fire by letting power users create their own Enterprise Mashups using the events? Why not start creating analytical applications that integrate events, services, from your SOA, and data from your data warehouse? Why not get started using the events now?
Your users will get more value faster if you release services and events to them sooner.
For the record, I agree with you. The key is getting the information known, which was the purpose of the statistical collection and reporting. If you want to take it a step further and start leveraging it elsewhere, go for it. The potential consumers first need to be aware that events exist at all, however.
Comment by biske — February 23, 2007 @ 3:44 pm