A Trust Model Based on the Multinomial Subjective Logic for P2P Network

Junfeng TIAN, Chao Li, Xuemin HE, Rui TIAN
2009 International Journal of Communications, Network and System Sciences  
In order to deal with the problems in P2P systems such as unreliability of the Service, security risk and attacks caused by malicious peers, a novel trust model MSL-TM based on the Multinomial Subjective Logic is proposed. The model uses multinomial ratings and Dirichlet distribution to compute the expectation of the subjective opinion and accordingly draws the peer's reputation value and risk value, and finally gets the trust value. The decay of time, rating credibility and the risk value are
more » ... ntroduced to reflect the recent behaviors of the peers and make the system more sensitive to malicious acts. Finally, the effectiveness and feasibility of the model is illustrated by the simulation experiment designed with peer-sim. Nowadays the research of P2P trust is mainly focused on building reliable trust management model. Trust Management (TM) is first proposed by Blaze M. in 1996 [5], and then it became a research focus of network security. The PeerTrust [6,7] model proposed by L. Xiong combines both local and global reputation with confidence coefficient, and considers several factors influencing credibility quantification, the model can cope with virtual ratings well. However, the PeerTrust model does not offer measurements for factors of trust and methods for defining confidence coefficient. The P2Poriented and reputation-based trust management model proposed in reference [8] introduces risk factor, and proposes to quantize risk with information entropy. This model is superior to some existing trust models in terms of both security and other aspects. Jøsang makes research on trust management based on Subjective Logic [9], and proposes Evidence Space and Opinion Space that are used to describe and measure trust relationship. Also, he offers a series of Subjective
doi:10.4236/ijcns.2009.26060 fatcat:pcgvi3pm4jfrbmkbnklhpviei4