A New Method of Detecting Network Traffic Anomalies
Jun Cai, Wai Xi Liu · Applied Mechanics and Materials · 2013
As Internet communications and applications become more and more complex, accurately describing network traffic information and rapidly monitoring network traffic anomalies have become increasingly challenging tasks. In this paper, we present a framework and method for monitoring network traffic through measuring the dynamic changes of host communities. An unweighted and undirected host interaction network (HIN) is established through extracting the social-behavioral characteristics of network traffic. Based on social-behavior similarity in HIN, host community is defined, and then five features are proposed to capture host community changes. Finally, this method is evaluated through two real-world network traffic, and the experimental results show that the method presented in this paper can effectively capture the dynamic changes of host communities to monitor network traffic.