A P2p-based Distributed File Sharing System Design And Implementation

With the rapid increase of the Internet nodes in computing power, disk space and network bandwidth, Internet applications for file resources sharing has been developing rapidly. Richer file resources, higher reliability, more efficient data transmission, better load balancing and lower cost have become the new goals to pursue for file sharing users. P2P system, by sharing some resources of nodes, such as processors, disk space, network bandwidth, etc, allows users to directly use other nodes'shared resources without the need to obtain form particular servers. Peer nodes in P2P network are not only owners of resources, but also resources requesters. P2P technology has been used widely in file sharing, distributed computing, collaborative work, real-time communication, search engines, etc.In this thesis, a distributed file sharing system based on P2P is presented, by analyzing and summarizing the development of file sharing system, and P2P technology. System nodes compose an overlay network through P2P technology, and nodes take advantage of these resources, to build a large-scale, low-cost, high scalability, high reliability, high-performance distributed file sharing system. The system uses the centralized topology, in order to efficiently implement the system load balancing and file upload, download, query, and system management. At the same time, the fully distributed unstructured topology is a backup architecture to use. If the index server fails, the system will automatically convert to the fully distributed unstructured topology. This greatly enhanced the reliability of the system.In order to improve the operating efficiency of the index server, the system uses the memory-mapped file technology to improve the efficiency of read and write of files, and use the socket I/O completion port model to improve the sockets I/O efficiency. Through load balancing, the system can take full advantage of disk space, computing and bandwidth resources of each node in the P2P overlay network, thereby enhancing the system scalability and disk space utilization. By using dynamic copy, the number and location of files can be changed according to system requirements. Many of the specific implementation strategy in this system can achieve the freedom to configure. If a file is deleted from a node, users can select the appropriate file-out algorithm according to the actual circumstances and needs. For file download and backup, users can select one or more of the node selection algorithm the system provides, and change the priority between them. The system's capabilities of log and filtering can enhance the system security and allow the system to be in a better management and control.