Stochastic Strategic Routing Reduces Attack Effects

In this paper we consider the problem of routing traffic between k source-destination pairs. Using game theoretic modeling we provide randomized strategies to minimize the threat of attacks on links by an adversary. The adversary is assumed to have a choice of c edges for attack. We propose iterative methods to find the Nash Equilibrium of the zero-sum game. The proposed schemes have been implemented using existing network models (GEANT in Europe and the AT&T network in US) and show marked reduction in the gain of the attacker. As the gain of the attacker is related to the congestion on the edges, our schemes also reduce congestion.