No abstract

In this chapter we consider the effect of “selfishness” on distributed MAC protocols in wireless local area network (WLAN). The inherently contentionbased medium access in distributed systems is modelled as a non-cooperative game: “access game.” Both quality of service (QoS) and battery power (BP) are incorporated in modelling the game. It is shown that the Nash equilibrium (NE) for incomplete information games is usually inefficient compared to the NE of complete information games. We propose some simple mechanisms to approximate the incomplete information scenario as complete information. For complete information games, we investigate whether fairness can be achieved by selfish users. Different cases are considered and it is shown that the NE does not result in fairness. We next compute the constrained NE (CNE) for the access game. Finally, we analyze the stability of the distributed system.

No abstract

Fair sharing of bandwidth in distributed systems is considered. The inherently contention-based medium access in distributed systems is modelled as a non-cooperative game: “Access Game”. It is shown that the Nash Equilibrium (NE) for incomplete information games is usually inefficient compared to the NE of complete information games. Consequently, simple mechanisms to approximate an incomplete information scenario as a complete information scenario are proposed. For complete information games, it is shown that the NE usually does not result in fairness. Therefore, we compute the Constrained Nash Equilibrium (CNE) of the “Access Game” with the fairness criteria as the constraints. It is shown that the CNE is unique.

Abstract-Fair sharing of bandwidth remains an unresolved issue for distributed systems. In this paper, the users of a distributed LAN are modeled as selfish users with independence to choose their individual strategies. With these selfish users, the contention-based distributed medium access scenario is modeled as a complete-information, noncooperative game, designated the "Access Game." A novel MAC strategy based on p-persistent CSMA is presented to achieve fairness in the "Access Game." It is proven that there are an infinite number of Nash Equilibria for the "Access Game," but they do not result in fairness. Therefore, it may be beneficial for the selfish users to adhere to a set of constraints that result in fairness in a noncooperative fashion. This leads to the formulation of a constrained "Access Game" with fairness represented as a set of algebraic constraints. It is proven that the solution of the constrained game, the Constrained Nash Equilibrium, is unique. Further, it is shown that, in addition to achieving fairness, this solution also optimizes the throughput. Finally, these results are extended to a more realistic incomplete-information scenario by approximating the incomplete-information scenario as a complete-information scenario through information gathering and dissemination.