Abstract
Peer-to-peer (P2P) networks are gaining attention for their significant features such as anonymity, autonomy, decentralization, and self-policing. However, the anonymous environment of P2P makes it vulnerable to potential malicious attacks and threats. Reputation Management Systems (RMS) help to establish and evaluate trustworthiness among peers based on their previous transactions and feedbacks from other peers. There are two primary methods for RMS: peer-based reputation systems and file-based reputation systems. AuthenticPeer is one of the hybrid techniques that utilize both approaches by using EigenTrust, and file-based reputation algorithm together. We propose an improved version of AuthenticPeer which is named AuthenticPeer++. Our goal is to increase trust by preventing untrustworthy files from spreading and reducing the offensive and deceptive behavior of collective malicious peers. We have compared our proposed method with other reputation algorithms (RA) considering four different threat models. Simulation results indicated that the AuthenticPeer++ maintained a higher performance in minimizing the impact of Purely Collective Malicious, Disguised Malicious and Feedback Malicious threats.