The trust management framework for peer-to-peer networks

Abstract

<p>Popularity of peer-to-peer (P2P) networks exposed a number of security vulnerabilities. Among those is a problem of finding reliable communication partners. In this thesis, we present an integrated trust framework for peer-to-peer networks that quantifies the trustworthiness of a peer via reputation-based trust mechanism and anomaly detection techniques. As opposed to other known techniques in P2P networks, our trust management schema is fully decentralized and does not rely on the co-operation of peers. Furthermore, the reputation computation is based on traffic coming from other peers. We also describe an anomaly detection procedure that analyses peer activity on the network and flags potentially malicious behavior by detecting deviation from peer profile. We present integration of our anomaly detection to trust management scheme and study the performance of reputation-based approach using implementation and performance of trust framework through simulation.</p>