Chapter 9. Detecting event patterns

 

Art is the imposing of a pattern on experience, and our aesthetic enjoyment is recognition of the pattern.

Alfred North Whitehead

It’s no exaggeration to say that event pattern detection is the jewel in the crown of event processing. Pattern detection lets us go beyond individual events to look for specific collections of events and the relationship between them; a pattern that is detected can have a meaning that goes beyond the individual events that it comprises. Consider a personalized healthcare system where a patient is hooked up to multiple monitors. The individual events reported by the monitors might not in themselves be significant, but by considering the events from all the monitors you might spot a problem that you would miss by looking at each monitor separately. In this example you would be looking for particular combinations of events occurring in particular orders, particular trends emerging from the events, or situations where an expected event did not occur. These are all examples of what we call an event pattern.

In this chapter we shall discuss patterns and the pattern detection process, in particular:

  • The definition of an event pattern, and the Pattern detect event processing agent
  • Categories of event patterns
  • Specific patterns of various types
  • The use of pattern policies
  • Pattern detection examples from the Fast Flower Delivery application

9.1. Introduction to event patterns

9.2. Basic patterns

9.3. Dimensional patterns

9.4. Pattern policies

9.5. Patterns reference table

9.6. The Fast Flower Delivery patterns

9.7. Pattern detection in practice

9.8. Summary