A New Game Model of Nodes' Strategies in P2P Network

Abstract

The heterogeneity of P2P nodes' strategies has serious affect on the service performance of P2P network. Many incentive mechanisms based on game models of nodes' strategies are proposed to limit node's malicious actions. However, most of them don't take the diversity of nodes' types and the complexity of nodes' strategies into account, as well as the asymmetry of nodes' information. So the game models wouldn't imitate actual behavior and the incentive mechanisms based on them are not effective. This paper presents a new game model of nodes' strategies in P2P network, referring to the perspective of human society. The model classifies nodes' strategies to several sets according to their types, and analyzes the transfer conditions of different strategies. Finally, this paper simulates the evolution of the game model and the evolutionary results show similarity with real P2P network environment.

A distributed storage model for selfish and rational nodes

“…Those researches simply abstract nodes' action such as corporation or non-corporation, cheat or non-cheat, etc. Wang haoyun [7]divides nodes into two types: 1) good nodes with corporation and non-corporation strategy, 2) bad nodes with attack, collusion and free-riding strategy. Both types have the strategy of betray when one node changes its type to another, and strategy of white wash when the node reenters system using a new identity.…”