LARS

Abstract

Mobile ad hoc networks (MANETs) have nodes that are dynamically and arbitrarily located. The nodes are often battery powered and may behave selfishly to preserve power. A promising mechanism to deal with node selfishness is to use reputation methods. In this paper, we discuss some general issues with current reputation methods that are used to mitigate selfish node behavior and propose a simple and efficient Locally Aware Reputation System (LARS) for which the reputation of the nodes is derived by using direct observation. LARS is flexible and can deal with both extreme selfishness and selective selfishness.