Self-Adaptive Model Of Two-Level For Dynamic P2P Network

All existing P2P model can be classified to three categories by network logic topology:Centralized model, Decentralized unstructured model and Decentralized structured model.Centralized model depends on a single server, Decentralized unstructured model is ineffectivefor its flooding search mechanism, and Decentralized structured model based on DHT(Distribute Hash Table) enjoys great advantage of simplicity and extensibility. However suchrigid structured model is not appropriate for dynamic network as it brings large number ofmessages when peer churns.The present thesis puts forward a hierarchical P2P model: self-adaptive P2P model oftwo-level. Super peers compose a self-adaptive DHT network on the top layer, and lowerordinary peers take super peer in the same group as their center server to form a cluster. Dataare stored in ordinary peers and super peers hold indices of data and answer for request fromordinary peers. When super peer joins or leaves, it does not maintain consistency of entirenetwork's logic topology but only update routing table of its predecessor and successor,besides, super peers update their routing table cache when they transmit messages. Theordinary peers keep backup of super peer for system restore when super peer fails. System canoptimize network structure by select peers with high capacity to be super peer and restrictamount of peers in each group.The experimental results indicate that if k is average amount of peers in each group, thensearch in self-adaptive P2P model of two-level only brings 1/k hops compare to Chord ofsame net size, when peers join or leave it produce messages at constants level. It also haseffective mechanism for fault tolerance. The conclusion indicates that the model can be moreapplicable in ad hoc P2P network with high churn rate.