Improved Recommendation Filtering Component Resilient to Trust Distortion Attacks in a MANET

Abstract

The resilience of a mobile ad hoc network (MANET) to malicious packet droppers is directly related to the design of a trust management framework (TMF) which assigns a trust metric value to each node through behavioral monitoring. As the network topology in a MANET changes dynamically, a node may be uncertain about the trust metric associated with some other node with which it never had interacted. Under such circumstances, recommendations from other trusted peers may help in the computation of trust which is known as indirect trust component. Even though recommendations facilitate the trust computation of a TMF, malicious nodes may launch trust distortion attacks which intend to misguide the nodes about the trustworthiness of other nodes. Dishonest recommenders form a category of malicious nodes which launch the trust distortion attacks. To cope up with such attackers, a TMF should encompass a recommendations filtering component. The current work serves to improvise upon the recommendations filtering scheme used in an uncertainty reasoning based TMF proposed in our earlier work.