News Archives

RSS Feed

Peer-to-peer Trust and Reputation Systems

May 3, 2007

  • Date: Thursday, May 3rd, 2007 
  • Time: 11 am— 12:15 pm 
  • Place: ECE 118

Baruch Awerbuch 
John Hopkins University

Abstract: The purpose of P2P reputation systems is to enable peers to leverage experience of other peers, whose taste and integrity are questionable. For example, electronic commerce engines like eBay depend heavily on reputation systems, which allow a peer to check the reputation system before any transaction to avoid dealing with un-reputable peers. Other systems like Amazon and Netflix provide recommendation on movies and books based on experience of other peers with widely varying taste.

We show that a “peer group” of honest peers with same or similar taste can effectively cooperate by essentially filtering out the inputs of irrelevant or malicious peers, even if the above peer group is in the minority. That is, the net effect is as if the members of the peer group would know in advance who are the other members of that group.

This holds for any behavior pattern, and in particular for completely adversarial behavior of the peers outside the peer group. The talk will review some briefly some of the past work in this area.

Baruch Awerbuch is a professor of Computer Science at the Johns Hopkins University. He received his Ph.D. in 1984 from the Technion, Haifa. From 1984 till 1994 he was at MIT as post-doc, research associate, and faculty at the Mathematics department. His research interests are in the areas of algorithms, wireless networks, learning, security, and peer-to-peer systems with the emphasis on algorithmic approaches to real systems and networks.