Tim Bass
07-15-2008 02:29 AM
In an earlier post I explained why situation modelling, and preferable an object-oriented situation model, is one of the key attributes of CEP. Unfortunately, I have yet to find a situation model for complex events, so I offer a few simple baseline concepts here.* Your comments and improvements are much appreciated.
1. A situation model of a complex event is an abstract representation of a described or experienced situation that we wish to detect in real-time.
2. Situation models are composed of four primary objects:
a. A spatial-temporal reference framework (spatial locations, time frames, window size)
b. Entities objects (people, objects, system)
c. Properties of entities objects (velocity, amount, size, price, direction)
d. Object relational information (spatial, temporal, causal, dependence, proximity, network, taxonomy, classification)
3. Situation models of complex events may have*three levels of model representation:
a. Situation model (event-specific)
b. Episodic model (coherence sequences of events)
c. Comprehensive model (a comprehensive collection of episodes)
Hence, in a nutshell, it is imperative that we have a situation model for representing complex events if we are going to move CEP forward.*** The simple model in this post may or may not be the right one to develop, but at least we have something to talk about.* Ideally, the model should be object-oriented, althought it does not have to be.
When we have a workable model for situations in the context of event processing, we will have a working model for complex events.** Then, with a working model of complex events, we can build a working model for complex event processing.*
References:
The New Theory for Situation Models
Source...
