New metrics for reputation management in P2P networks

  • Authors:
  • Debora Donato;Mario Paniccia;Maddalena Selis;Carlos Castillo;Giovanni Cortese;Stefano Leonardi

  • Affiliations:
  • Yahoo! Research Barcelona, Spain;University of Rome, "La Sapienza", Rome, Italy;University of Rome, "La Sapienza", Rome, Italy;Yahoo! Research Barcelona, Spain;University of Rome, Tor Vergata - Rome, Italy;University of Rome, "La Sapienza", Rome, Italy

  • Venue:
  • AIRWeb '07 Proceedings of the 3rd international workshop on Adversarial information retrieval on the web
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this work we study the effectiveness of mechanisms for decentralized reputation management in P2P networks. We depart from Eigen Trust, an algorithm designed for reputation management in file sharing applications over p2p networks. EigenTrust has been proved very effective against three different natural attacks from malicious coalitions while it performs poorly on particular attack organized by two different kinds of malicious peers. We propose various metrics of reputation based on ideas recently introduced for detecting and demoting Web spam. We combine these metrics with the original EigenTrust approach. Our mechanisms are more effective than EigenTrust alone for detecting malicious peers and reducing the number of inauthentic downloads not only for all the cases previously addressed but also for more sophisticated attacks.