Bloom Filter based Replication Management in Peer-to-Peer Storage Network
Peer-to-peer (P2P) storage system has emerged as an important model to implement distributed storage for its scalability and autonomous feature. Redundancy is an effective way to guarantee durability and availability in wide-area distributed storage systems. Replication strategies in P2P storage networks are always proposed as part of system accomplishment or routing mechanism. Former researches are always bounded up with specific network scene and ignore how to make it more flexible and easy to adjust to various P2P storage network conditions. This paper works over a more selfadaptive replication management to adapt to different P2P storage networks. It utilizes Bloom filter algorithm to help system make replication decision. We only give the basic strategy in both balanced and heterogeneous instances, and the detailed methods can be laid down according to specific situation. In our simulation, we designed the detailed replication management mechanism based on this idea. The results demonstrate its effectiveness in both scenes.
replication peer-to-peer storage bloom fliter
Fengjun Wang Meina Song Junde Song
ICT&SSME CENTER Beijing University of Posts and Telecommunications Beijing, China
国际会议
桂林
英文
429-432
2010-11-17（万方平台首次上网日期，不代表论文的发表时间）