A Performance Analysis Model for BitTorrent-Like Peer-to-Peer Systems Based on the Notion of Emergence

As a typical and practical peer-to-peer system, BitTorrent gets amount of attention in theory and practice. The work that has been taken great efforts is to provide an effective model for analyzing BitTorrent. Based on the notion of emergence in system science, this paper investigates BitTorrent in a broad view which is from microscopic to macroscopic, and gives a micro-macro analytical model for them. To obtain this model, we start from the states of microscopic BitTorrent, i. e. nodes' states, and get the parameters that represent a node,called microscopic parameters. Based on these parameters and the activities of a node,the mi-croscopic behaviors of BitTorrent are characterized. At the same time, several macroscopic parameters are chosen to describe system-atic states of BitTorrent. Based on the microscopic behaviors, by exploring the relationship between microscopic and macroscopic pa-rameters, the macroscopic equations characterizing BitTorrent's evolvement are established. And based these equations,macro prop-erties of BitTorrent can be obtained. Experiment results indicate that this model can effectively catch the behaviors of BitTorrent.