Title: Estimation of trustworthiness for P2P systems in a collusive attack

Authors: Fumiaki Sato

Addresses: Department of Information Science, Toho University, 2–2–1 Miyama, Funabashi, 274–8510, Japan

Abstract: In Peer-to-Peer (P2P) systems, it is an important problem to evaluate the trustworthiness of the peer for safe communications. In many existing methods, the trustworthiness is computed based on the reputation feedback of the peer. Especially, the method using the maximum likelihood estimation can compute a stochastic estimation value from comparatively little feedback information. Moreover, even if a lie is included in the feedback, the trustworthiness is calculated with an acceptable accuracy. However, this method does not have enough strength in a collusive attack. This paper proposes a new estimation method in consideration of a collusive rate of the peer. Moreover, it proposes an algorithm which calculates trustworthiness by proposing the method of presuming a collusive rate from the similarity of the feedback of each peer. This method was evaluated by the simulation. As a result, trustworthiness was presumed at a higher accuracy when there was a collusive attack.

Keywords: trustworthiness; reputation systems; P2P systems; maximum likelihood estimation; collusive attack; peer-to-peer networks; simulation.

DOI: 10.1504/IJWGS.2008.018488

International Journal of Web and Grid Services, 2008 Vol.4 No.1, pp.24 - 34

Published online: 25 May 2008 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article