Studying misbehavior in CSMA/CA Wireless LANs

Abstract

Security is a fundamental aspect for achieving high availability and reliability in wireless networks. Wireless medium access control (MAC) protocols use distributed contention resolution mechanisms for sharing the wireless medium. Wireless nodes decide access to their channel independently and accessing the channel by a node has an influence on those of other nodes. In this environment, selfish nodes that do not obey the operation of the MAC protocol may obtain an unfair share of the channel bandwidth at the expense of well-behaved nodes. Game theory is a useful and powerful tool to research this kind of systems. We study the effect of misbehavior in CSMA/CA wireless LANs and in particular in Idle Sense access method which has been designed to optimize the performance and fairness of the network using game theory. We show that Nash equilibrium point of our game is not an optimal point of the game. Then we find the Pareto-optimal point of our game using static game model from Nash bargaining framework. We design a prevention model to force nodes converge to the Pareto-optimal point and make the Pareto-optimal point as the Nash equilibrium point of our game. Finally, we devise a detection system to detect deviating cheaters from optimal point.