|M.Sc Student||Gaidar Dmitry|
|Subject||TEM Model Implementation|
|Department||Department of Industrial Engineering and Management||Supervisor||Dr. Opher Etzion|
|Full Thesis text|
In situation where event-driven architecture is required for detection of desired patterns arises a demand of simple event processing methods to identify meaningful events (such as opportunities or threats) and respond to them. As an example of such system we can consider the detection of lost luggage at an airport, during transfer from check-in to baggage claim. Another example is the detection of suspicious money transfers in the banking system.
The Event Model (TEM), defined by Opher Etzion and colleagues, allows the implementation of the desired functionality. While many event processing engines exists, there is an inherent complexity in using them that require skilled developers. The TEM is a "computational independent model" aimed to enable semi-technical people to create applications by removing technical terms and decisions.
The purpose of my project was to create a proof of concept for this theoretical model.
In order to let a potential user apply the model, I created a simplified tool that allows defining event processing rules and using the model without preliminary studies. During the implementation of the model the following challenges were addressed:
1) Inferring missing information
2) Designing quick access database
3) Fast search for a desired combination of events
During the model implementation we predefined a set of tests that gave us a confirmation of TEM model's correctness.