In Delay Tolerant Networks (DTNs), routing protocols work in an environment where there is no guarantee of end to end path between source and destination. Mobility of nodes is an important factor that helps to deliver packets in DTN. Intermediate nodes receive, store and forward packets to other intermediate nodes or destination. The packets are stored in entirety at intermediate nodes even if intermediate nodes may not have path to destination. Since mobile nodes have limited buffer space and are connected intermittently, therefore, finding a node that is a good forwarder is important for delivery of messages. Probabilistic routing protocols suggest predictability measure on information stored to know which node in contact is potential forwarder in terms to deliver message to destined node.
However, any malicious node in network may inject other nodes with bogus messages in order to degrade network performance. This paper presents updated metric called reputation of node which captures predictability as well as enables node in network to decide whether to accept messages from node in contact or not. When node transfers genuine messages it gains reputation as well as predictability with respect to destined node. Simulation results show that algorithm prevents flow of bogus messages in the network and thereby increases message delivery and reduces overhead in presence of malicious nodes.