Sensing anomalies with an optimal filter applied to the traffic matrix of an IP telecommunications network

In this paper, we present the application of the well-known Kalman Optimal Filter to the Traffic Matrix (TM) of an IP Telecommunications network in order to sense and localize anomalies. At the first step, the application did not reveal significant anomalies with high dynamics activities through the network. As a second step, we simulated real increases of data flows at certain links Origin/Destination (OD) of the TM. The results obtained have shown clear and neat anomalies goodly sensed and localized by the optimum Kalman filter.