A framework for detecting selfish misbehavior in wireless mesh community networks

Wireless Mesh Networks (WMNs) have recently emerged as a flexible and low-cost extension of wired infrastructure networks. They consist of mesh routers and clients, where mesh routers are almost static and form the backbone of the WMN. The complete absence of an infrastructure and the flexibility provided by the wireless mesh technology has fostered the development of new network paradigms like Wireless Mesh Community Networks. Such networks are usually composed of heterogeneous mesh routers managed by different users (a subset of participants to the community), that collaborate to extend the network coverage. However, in such environment some participants can exhibit selfish behaviors, by dropping selectively the packets sent by other mesh routers, in order to prioritize their own traffic and increase their network utilization. In this paper we propose a complete scheme to detect selfish behavior of the mesh routers that participate to the community network. Each node evaluates the trustworthiness of the other mesh routers by combining the direct observations on the relaying behavior of neighbor nodes with the trust information provided by other mesh routers. The proposed scheme has been integrated in the AODV routing protocol, and tested in several network scenarios. The numerical results show that our scheme provides a high detection accuracy, even when a high percentage of network nodes provide false trust values (bad-mouthing attack).