The Research On Sharing File Consistency Maintenance In Unstructured P2P System

Growth in the volume of data promotes that computer storage technology continues to develop. Comparing with traditional distributed systems, P2P technology has unparalleled advantages and broad application prospects, which make it becoming the most popular technology in 21st century. As an important component of the P2P system, data replication is a well-known and widely accepted technique to reduce data response time and network bandwidth affect the performance of the system. However, because current network may be dynamic with nodes frequently joins and leaves and objects updated highly for adapting to operation developments, maintaining consistency of replications is more challenging in these environments. First, if there aren't effective consistency maintenance algorithms, the P2P applications only apply for sharing static and low-frequent updated objects; Secondly, new operation development needs which ensured the dynamic content updated. Meanwhile, it is absolutely necessary for advancing quality of service and development P2P systems.Present researches have these mainly problems: the balance between updated overhead and successful updated ratio; sometimes emphasize expansion and highly performance, so ignore the impact of performance from different replication methods; P2P system of functions low-efficiently, because their structures don't fit that of the underlying network.In this paper, firstly, it is analyzed the characteristics of P2P systems, the origin of replica, and study algorithms based on replication and replica location, present a new consistency maintain algorithm.This paper focuses the balance between updated overhead and successful update ratio, presents a region push then pull algorithm in highly dynamic networks. This algorithm is mainly made up of two parts: conceive the replication maintenance model and present a new region update mechanism based on push/pull. It based on the percentage of offline nodes and the transform ratio between replicated nodes and un-replicated nodes to decide the nodes which are responsible of easing the overload of file node, maintaining update consistency and avoid network congestion. Besides, in the improved push/pull method, the update messages broadcasted by the path within replication maintenance model, and the TTR could change when the different replications have the different demand for consistent standard, so can reduce the unnecessary messages by weakening the aggression of pull. At last, it is simulated the algorithms mentioned above to test the properties, performance and validity.