Thursday, February 20th, 2003
Room 104, Gates Computer Science Building

Incentives for Cooperation in the Internet

Kevin Lai

About the talk:

Many Internet systems rely on cooperation between two or more self-interested parties. In such systems, each party has an incentive to exploit the benefits of the system without contributing. For example, in peer-to-peer file sharing, people can maximize their utility by downloading from others while not allowing uploads. Existing solutions are not effective in the large, dynamic, zero-cost identity systems that are common in the Internet. We present collusion-resistant techniques that encourage cooperation in such systems and demonstrate their effectiveness using Prisoner's Dilemma simulations.

This is joint work with Michal Feldman, John Chuang, and Ion Stoica.

About the speaker:

Kevin Lai is currently a post-doctoral scholar in both the School of Information Management and Systems (SIMS) and EECS Department at U.C. Berkeley. He received his Ph.D. in Computer Science from Stanford University and his AB from U.C. Berkeley. Do not ask him whether Cal or Stanford is better. He has worked on operating systems benchmarking, mobility, wireless ad hoc networking, and network bandwidth measurement.