Models
and Games for Data Dissemination in Peer-to-Peer Systems
George
Kesidis
Pennsylvania
State University
Date:
Friday, January 5, 2pm
Location: BA 1200
Abstract:
We begin with an overview of incentive systems for cooperation in
peer-to-peer file distribution including those that are: based on
monetary exchange, based on cumulative reputations, and rule-based.
Incentives work for both the process of query resolution (via
structured, unstructured or hybrid systems) and for the exchange of
data itself. For data-exchange, we will formulate a game involving
provably incentive-compatible cumulative reputation systems and
uplink-bandwidth control. An epidemiological (coagulation) model for
BitTorrent swarms will also be described. Finally, I will conclude with
a brief description of epidemiological modeling of the Slammer worm and
discuss the quarantined DETER testbed based on Emulab.
Bio:
George
Kesidis received his M.S. and Ph.D. in EECS from U.C. Berkeley in 1990
and 1992 respectively. He was a professor in the E&CE
Dept of the University of Waterloo, Canada, from 1992 to 2000. Since
April 2000, he has taught in both the CS&E and EE Depts of the
Pennsylvania State University. His research experience spans
several areas of computer/communication networking including security,
incentive engineering, efficient simulation, and traffic
engineering. Currently, he a senior member of the IEEE and
the TPC co-chair of IEEE INFOCOM 2007.