Kaoru Chujo

Peer-to-peer traffic in metro networks: analysis, modeling, and policies

We report studies on peer-to-peer (P2P) traffic analysis, modeling, and evaluation of policies for the management P2P traffic in metropolitan networks. We conducted crawler-based measurement and analysis of P2P traffic using Gnutella. Based on the empirical parameters obtained through the measurement, known modeling techniques of P2P peers and contents, and GeoPoint, a geomapping technology, we built a P2P traffic simulator using J-Sim, a Java-based network simulation tool. The simulator is scenario-driven, allowing its user to customize network settings and service environment to examine the behavior of P2P traffic. With the P2P simulator, we were able to recreate a spatio-temporal traffic pattern characteristic to P2P services observed in August 2002. We then evaluated three different P2P traffic management policies, showing that traffic localization using a peer selection policy at super peers is possible, containing P2P traffic to the local metropolitan network as much as 40%.