Trust-based Decentralized Service Discovery in Structured Peer-to-Peer Networks

Abstract

In a Peer-to-Peer(P2P) network, efficient cooperation among peers (nodes) is an important issue since a peer may not believe another peer for various service discovery applications. Chord protocol faces stale routing entries (finger table entries) due to the presence of malicious peers in the network. We propose a trust-based routing algorithm for efficient service discovery for Chord network. The proposed algorithm improves the data accessibility in the presence of malicious peers by computing a trust value for each peer. The trust value is calculated using a Bayesian network based trust model. In this algorithm, the finger table entries including the trust attribute are retrieved using an improved querying algorithm while identifying a trusted peer. The trust is computed based on entities such as age, number of uploads and downloads of a peer. The proposed approach is flexible(one or many entities can be chosen to compute trust) and can be customized based on application requirements. Performance evaluation has been performed by considering performance metrics, such as, responses of queries, number of uploads and downloads and trust value of a peer in the P2P network.