Mathematical Modeling of Incentive Policies in P2P Networks

By

Prof. John Lui
Chairman / Professor, Department of Computer Science & Engineering, CUHK
 
 
 

Date: July 7, 2009 (Tuesday)

Time: 3:00p.m. - 5:00 p.m.

Venue: Rm. 121, Ho Sin Hang Engineering Building, CUHK

Abstract :

In order to stimulate cooperation among nodes in P2P systems, some form of incentive mechanism is necessary so as to encourage service contribution. Hence, designing and evaluating the stability, robustness and performance of incentive policies is extremely critical. In this presentation, we propose a general mathematical framework to evaluate the stability and evolution of a family of shared history based incentive policies. To illustrate the utility of the framework, we present several incentive policies and show why some incentive policy can lead to a total system collapse while the other is stable and operates at the optimal point. One can use this mathematical framework to design and analyze various incentive policies and verify whether they match the design objectives of the underlying P2P systems.