Distributed Trust Management Model for Peer To Peer Systems

In a file sharing peer-to peer system trust is essential to attain enhanced cooperation between peers. Reputation is utilized in reputation-based peer-to-peer systems to form trust between peers. Commonly in these systems, extremely reputable peers will be chosen to upload requested files and thus diminishing malicious uploads in the system considerably. Conversely, malicious peers should be encouraged to fund confidently by uploading authentic files rather than malicious ones. In this work, d istributed algorithms is presented, which allow a peer to reason nearly trust worthiness of other peers depends on past interactions and recommendations using distributed trust management model. In this, peers build their own trust network in their v icin ity with the use of local information available trust-reputation, trust-service, and recommendation metrics which are well-defined to measure trustworthiness in providing services and providing recommendations with the use of a trust management system. In addition to that, Credib ility Behaviour is presented as the second dimension of the trust management framework. The Malicious Detector Algorithm (MDA) is utilized to discover liar peers. The new concept of suspicious transactions is presented to identify these liar peers. Performance evaluations show that the proposed scheme is able to detect and isolate malicious peers from the system, henceforth, giving higher peer satisfaction, improved network resource utilization and increasing peers' satisfaction.