Simulating peer-to-peer networks

Abstract

Peer-to-Peer (P2P) systems are emerging as a new form of distributed computing with a strong emphasis on self-organization, decentralization, and autonomy of the participating nodes. The characteristics of self-organization, autonomy, and decentralization allow for highly adaptive, robust, and scalable networks, making P2P an increasingly interesting way to design distributed systems. Since the deployment of P2P systems involves significant resources, e.g., hundreds of hosts and users, it is often not possible to run realistic tests prior to the rollout of the system. Consequently, simulation is the only realistic approach for testing or predicting the behavior of large P2P networks. However, the majority of the existing simulators tend to provide limited flexibility in simulating the details of the users, application, protocol, and physical network. In this research, the impact of user behavior, protocol, and physical network characteristic on the overall P2P system are being observed. The aim is to investigate the importance of simulating P2P systems in such detail.