Clustering NGN user behavior for anomaly detection

In the vision of both researchers and standardization committees, networks and services will evolve in the direction of increasing pervasiveness, convergence, and quality of service management capability. Consequently, users will gain an increasing dependency on the presence and availability of network connectivity and the huge plethora of provided services. Yet fostering the development of our society, such dependency on a relatively young technology poses serious threats, especially from the trustworthiness, security and privacy point of view. In this paper, we will describe and critically evaluate user behavior clustering aimed at monitoring and assuring the security of NGN-based applications. Different models of user behavior, developed within both ISP and academic research projects will be described, and several techniques for manipulating and exploiting such model for the anomaly detection purpose will be described and evaluated.