本文已被：浏览 1690次 下载 3917次
Received:June 09, 2011 Revised:July 18, 2011
Received:June 09, 2011 Revised:July 18, 2011
中文摘要: 研究了从数据密集大部分拜占庭容错分布式存储协议使用复制技术,但是当存储的数据块很大时,复制技术要求大量的存储空间并占用网络带宽,效率低下。采用纠删码技术则将数据块编码为长度相同的分片,然后把各个分片分别存储在对应的存储节点上,这样可以节省存储空间和网络带宽。拜占庭容错纠删码分布式存储需要额外的开销、附加的服务器甚至要版本化存储,才能保证数据的一致性。通过对通常的情况进行优化,采用新颖的机制并引入同态指纹检验码,使得需要最少的服务器、最小的额外计算量和最小的通信回合数目,达到了低开销,并且可以保证时间戳不跳
Abstract:Most byzantine fault-tolerant protocol distributed storage protocol use replication,but if the block is very huge,replication requires network width and many spaces, so its efficiency is low. Employing erasure coding makes the data block coded into fragments of the same length, then store the according fragment on corresponding node. This can save spaces and network width. Byzantine fault-tolerant erasure coded distributed storage usually requires much overhead, extra servers and versioned storage for the consistency of the data fragments. The paper optimizes for the common cases, employs novel mechanism and introduces homomorphic fingerprints. The protocol reaches the real overhead, because it requires minimal servers, least extra computation and minimal communication rounds. Furthermore, it can make timestamp no-skipping
keywords: web information extraction byzantine fault-tolerant erasure code replication homomorphic fingerprints
文章编号： 中图分类号： 文献标志码：
基金项目:安徽省教育厅自然科学基金(2005KJ004ZD)
| Author Name | Affiliation |
| CAI Luan-Jia | School of Electronic and Information Engineering, Tongji University, Shanghai 201804, China |
| Author Name | Affiliation |
| CAI Luan-Jia | School of Electronic and Information Engineering, Tongji University, Shanghai 201804, China |
引用文本：
蔡鸾佳.拜占庭容错纠删码分布式存储协议.计算机系统应用,2012,21(2):98-103
CAI Luan-Jia.Byzantine Fault-Tolerant Erasure Eoded Distributed Storage Protocol.COMPUTER SYSTEMS APPLICATIONS,2012,21(2):98-103
蔡鸾佳.拜占庭容错纠删码分布式存储协议.计算机系统应用,2012,21(2):98-103
CAI Luan-Jia.Byzantine Fault-Tolerant Erasure Eoded Distributed Storage Protocol.COMPUTER SYSTEMS APPLICATIONS,2012,21(2):98-103
