Abstract:

To correlate system events and to get the most fundamental cause of the problem, I adopted codebook and pool combined approach to event correlation analysis. By constantly shrinking the number of causes in cause pool, the real cause of the problem of system can be identified. This method can be applied to applications of distributed systems event correlation analysis. Test the model capability in connection with the case of the event and the reasons scale expanding which are recorded in codebook. Test the model function in connection with the indicators of recall rate, error rate, re-reported rate etc. The results of the test show that the model has the capability of anti-interference for the loss and reverse of the event, so it can analyze concurrent multi-system events independently without influencing each other.

Key words: computer application, event analysis, codebook based event correlation, cause pool