A Game Theoretic Approach to Modelling Jamming Attacks in Delay Tolerant Networks

Abstract

Cyberspace plays a prominent role in our social, economic and civic welfare and cyber security issues are of paramount importance today. Growing reliance of the intertwined military and civilian applications on wireless computer networks makes these networks highly vulnerable to attacks of which jamming attacks are a vital and exigent problem. In this paper, we study defence against jamming attacks as game in a delay tolerant network, with two adversarial players: the jammer playing against the transmitter. The transmitters seek to choose an optimal time to schedule his transmission securely, so as to maximize the probability of successful delivery before his session expires, while these transmissions are subject to inference from the jammer, who attempts to minimize this probability . We design strategies for the transmitters that offset transmission period based inference of network traffic by the jammer. We model these interactions and decisions as a game and use simulation as a tool to evaluate the games. Probability distribution functions over finite set of strategies are proposed to compute the expected payoff of both the players. Simulation results are used to evaluate the expected payoff along with the resulting equilibrium in cases where players are biased and unbiased. These results are used to strategically decide on the optimal time for both the players, and evaluate the efficiency of the strategies used by the transmitters against jammer attacks.