A Contemporary Framework for P2P Connected Systems through Trust Model for Self Organizing Nodes

The open nature of the peer-to-peer systems exposes them to malicious actions. Here in this paper we introdusing distributed algorithms that can enable peer to get the trustworthiness of other peers based on interactions of those peers in the past with other peers. The peers in the network are able of creating a trust network on their own proximity by the help of the local information it already have with it and without trying to learn any global trust information. The contexts of the trust, recommendation and service contexts are defined for measuring the trustworthiness in providing services to the other peers and giving recommendations about the other peers. Interactions done with other peers and recommendation about other peers are evaluated based upon the recentness of the information, importance and the satisfaction of the peer parameters. As the satisfaction of the peer is also considered as the most vital one because the recommendation will be always based upon the satisfaction of the peer that is a peer will recommend any other peer only if it is satisfied by the interaction with that peer. In addition to this the recommender’s trust worthiness about a recommendation given about a peer will be considered important while evaluating the recommendations. While moving with the simulation experiments on file sharing application it showed that this proposed trust model can reduce attacks on nearly 16 different types of behaviour models. In these experiments we see that good peers were able to form trust relationships in their own proximity and are able of isolating the untrusted peers. Keywords— Trust Model, Peer-to-peer Systems, trust management