Using Probabilistic Availability Measures for Predicting Targeted Attacks on Network Nodes

Abstract

This paper deals with targeted attacks on the nodes of a communication network. We present an optimization approach that may be useful for the network operator when deploying the so-called control nodes (called controllers) for resistance to attacks. A key element of our investigations is selecting an appropriate list of attacks that should be included in the optimization of controller placement. For this purpose, we propose innovative probabilistic network availability measures that could be used in planning the most dangerous attacks based on the attacker’s knowledge of the network. The operator can anticipate the set of such attacks and then incorporate them into optimizing the controller placement. In the paper, we discuss the proposed measures and present optimization problems appropriate for the deployment of controllers and attack planning. The numerical results illustrating our considerations are also included.