| As the Internet's development and the capability of PC enhancement, the technology of P2P network become more and more important in people life. Various applications based on P2P technology are springing out. People do not satisfy with the P2P system which can only provide some basic capability such as searching and downloading resources. A system can support dynamic business have becoming new goals and fields of research.Replication management is an important part to the P2P system. Data replication is a well-known and widely accepted technique to reduce data response time and network bandwidth consuming. However, because the node may join and leave freely and frequently, the P2P network presents a kind of significant dynamic characteristic, maintaining consistency of replications is more challenging in these environments.In this thesis, firstly, we analyze the characteristics of P2P systems and introduce several classic model and replications creation strategies, furthermore, present a new consistency maintaining algorithm. We choose the nodes to download files whose heat level of searching or performance is greater than a threshold. Otherwise, in order to reduce the creation of replication and the burden of system, the nodes only store the location information of nodes which have those files. What deserves to be mentioned, it does not decrease the probability of success of searching. In succession, a host node is adapted to maintain replications' consistency. Host node stores some location information of all replication nodes, and replication node stores the location of host node. When a replication has been changed, replication node send information to the host node, then host node will notice other replication nodes. It makes sure that updating messages will be transferred through it without flooding blindly. At last, we simulate the algorithms mentioned above under laboratory environment, and prove the validity of the algorithm. |
PDF Full Text Request