Research On Fault-Tolerant Technology Of Structured Peer-to-Peer System

Peer-to-Peer (P2P) has become one of the most important architectures for Internet applications for its inherent scalability, fault tolerance, and high performance. The research of DHT systems is one of the hottest issues in P2P, more and more applications on DHT come into use, and it plays an important role in Internet. Nevertheless, every peer is meant to join and leave the system arbitrarily, which makes failure more common than other systems and makes the system run slowly, even break down. To handle the churn, a number of approaches have been proposed to date.In this paper, we summarized the content and relative problem of P2P fault-tolerant firstly, and then introduced a series of techniques for structured P2P fault-tolerant under churn. We analyzed five important aspects on structured P2P fault-tolerant research: system dynamic, failure detection, overlay partition, data redundancy and new topology for fault-tolerant, and outlined the shortcomings and problems of existing method.Secondly, this paper discussed the node reliability evaluation problem in P2P file-sharing system from the quality of service perspective. Based on artificial neural network method, it designed a node reliability evaluation model in P2P system, and presented a node reliability measurement framework in P2P file-sharing system, which had some significance in reliability evaluation of P2P file-sharing system node, as well as the super node selection in topologies.On this basis, this paper designed a redundancy strategy based on cluster in structured P2P system-RSFE. This strategy included RSFE routing algorithm, node join and depart algorithm on FissionE, and achieved the purpose of fault-tolerant by using of back up node information with redundant hardware. The experiment showed that this strategy can improve the fault-tolerant capability effectively in the P2P network.In the end, the future trend of the P2P fault-tolerant technology is discussed.