More about HKUST
Providing Incentive and Service Differentiation in P2P Networks
Speaker: Prof. John C.S. Lui Department of Computer Science & Engineering The Chinese University of Hong Kong Title : "Providing Incentive and Service Differentiation in P2P Networks" Date: Monday, 8 November 2004 Time: 4:00 - 5:00pm Venue: Lecture Theatre F (Leung Yat Sing Lecture Theatre, near lift nos. 25/26) HKUST Abstract: Traditional Peer-to-Peer (P2P) networks do not provide service differentiation and incentive for users. Consequently, users can easily access information without contributing any information or service to a P2P community. This leads to the "free-riding" problem and consequently, most of the information requests are directed toward a small number of P2P nodes which are willing to share information or to provide service, hence, the "tragedy of the commons" occurs. The aim of this work is to provide service differentiation and incentive features to P2P networks so as to encourage all nodes to share information or services. We first introduce a distributed resource distribution mechanism which has a linear time complexity and it guarantees the "Pareto-optimal" resource allocation. Secondly, we model the whole resource request and distribution process as a competition game between all competing nodes. We show that this game has a Nash equilibrium. To realize this game, we propose a protocol such that all competing nodes can interact with the information providing node such that Nash equilibra can be reached efficiently and dynamically. Lastly, we show that our protocol is adaptive to different nodes arrival and departure events, as well as to different forms of network congestion. ****************** Biography: John received his Ph.D. in Computer Science from UCLA. After his graduation, he joined the IBM Almaden Research Laboratory/San Jose Laboratory. He later joined the Department of Computer Science and Engineering at the CUHK. He has been a visiting professor in computer science departments at UCLA, Columbia University, University of Maryland at College Park, Purdue University, University of Massachusetts at Amherst. His current research interests are in theoretic/applied topics in data networks, distributed multimedia systems, network security, mathematical optimization and performance evaluation theory. John received the CUHK Vice-Chancellor's Exemplary Teaching Award in 2001. He is an associate editor in the Performance Evaluation Journal and an elected member in the IFIP WG 7.3. John is the TPC co-chair of ACM Sigmetrics 2005. His personal interests include films and general reading.