Dispersion analysis of network traffic for intrusion detection in Smart Grids

Abstract

We consider the systems of detection and prevention of intrusions in modern telecommunication systems and networks. Methods of monitoring events, consisting of analysis of network activity of individual services and information services, are analyzed. It is proposed to use the mathematical apparatus of analysis of variance for processing results of modeling telecommunication systems and studying the statistical properties of network traffic in determining the significance of discrepancies or coincidence of characteristics. The proposed approach is to use the Fisher statistical criterion based on an estimate of the ratio of sample variances. This allows you to test the hypothesis about the homogeneity of statistical properties of network traffic with respect to the variance index (variance) with a given level of significance. The obtained results of experimental studies are recommended to be used to improve mechanisms for monitoring the network activity of individual services and information services, including for detecting and preventing intrusions in telecommunications systems and networks of promising Smart Grids.