Last session of the day (also blogged on paper) was Charles Brett on Why Events Matter To The Business and what this means for application development professionals. I heard Charles talk on a similar subject at the IBM IMPACT event –Live from IMPACT – Business Event Processing.
While many more business and IT people are aware of and/or using event processing or Complex Event Processing (CEP) solutions than you would expect, there is still a fair amount of confusion around terms. To try and clarify this, Charles defined a number of different types of event processing:
- Business Event Processing
Assigning responses to events to specific individuals
- Business Activity Monitoring
Using dashboards and similar to inform people as to the status of their business
- Systems and Operations Event Processing
Classic systems and IT event monitoring and processing
- Rules-based processing
- Complex Event Processing
Complex due to volume of events, complexity of correlation or some combination
- General Purpose Event Processing
Application development tools with some ability to build event processing solutions
In general, he said, events are everywhere, there are different styles of event processing and events are integral to integration, architecture and extended processes. Charles shows Events, Process, Services and Information as different aspects of the same problem and thinks they should all be considered, regardless of where you start your thinking.
He made a number of specific points:
- Event processing can often be added without disrupting existing systems, making it interesting
- Some kinds of event processing can be largely configured by the business without IT (though IT should be supportive and engaged)
- Business people often have an intuitive understanding of events where IT people see only IT events
- Events can be internal or external, IT or non-IT creating four different groups
- Knowing which events matter to you can really help with selecting the right approach
He had a number of nice examples. Two of my favorites where remote monitoring of multiple ICUs each with multiple pieces of equipment – particularly good as most ICU staff turn off the alarms on machines because they go off too much (they are hooked up to sick people, after all) and so more sophisticated monitoring that includes data about the patient seems like it would work way better – and monitoring a multi-stage service job so that if the first few stages take too long the system can assume delay in completion and start rescheduling.
The role of decisions in linking events, processes, services and information is one on which I will write some more soon.