节点文献
高容错分布式共享存储机制的研究与系统实现
Research and Realization on High Fault-Tolerance Distributed Shared Storage Mechanism and System Implementation
【作者】 郭琳燕;
【导师】 傅汝林;
【作者基本信息】 成都理工大学 , 信号与信息处理, 2004, 硕士
【摘要】 随着各种网络服务的广泛应用,数据可靠性越来越重要,因此,发展数据容错和恢复技术是极其重要的。在目前的存储技术中,RAID是广泛采用的保证数据可靠性的技术,RAID的实现是建立在分块、镜像和XOR校验的基础上的,但RAID是建立在服务器级的一种存储子系统,且它的XOR校验是指校验分块是所有分块数据进行异或(XOR)操作所得出的校验值。其缺点是容错性低,只允许单磁盘失败,此外,RAID无法远距离容错。SAN虽然实现了网络存储,但其容错性依赖RAID的容错性。 本文提出一种基于网络的高容错分布式共享存储机制—FTDSS(High Fault-Tolerance Distributed Shared Storage Mechanism),FTDSS利用网络中各节点主机的磁盘空间,构建一个大的共享存储空间,该空间对所有的网络授权用户开放,用户可以透明地访问全局空间,可以在全局空间读写文件,从而实现文件的分布式共享存储,FTDSS将文件分片,然后将数据分片和它们之间的校验分片存储在网络中的各节点主机,只要一定数目的节点在线,即使其它的机器不在线或崩溃,也可以从全局空间完整读出数据文件,从而获得高的容错性。FTDSS通过网络分布式存储,真正实现了数据的远程容错。同时,FTDSS将文件分片并行读写,能获得比单机高的读写性能。 通过系统软件,可以实现文件的分布式存放,对局域网用户备份数据而言,起到了小型RAID和部分文件服务器的功能且不需要额外的硬件支持。实验表明,系统能获得高容错性和较高的读写性能。
【Abstract】 With the wide application of all kinds of service of networks, security of datas has become more critical to the survival of an enterprise. So,It is very important to find a strategy for recovering their data should data lose. But within current technology of storage, RAID is a popular mechanism to offer good fault-tolerance.,the realization of RAID is builded based on flagments ,mirror iamge and XOR checkout. But RAID is a kind of subsystem of storage builded on class of server. And its XOR checkout is the checkout value that all flagments XOR. Its demerit is that fault-tolerance is low, RAID can not work well when one more disks fail, fault-tolerance of SAN is depended on fault-tolerance of RAID, though it has realized storage of networks.In this paper, we present an efficient network-based high Fault-Tolerance Distributed Shared Storage mechanism called FTDSS. FTDSS make use of disk space of node in Network to build a large public shared storage space. Users can read/write their file from/to the public storage space from node of network. Physically File is stored in each node in form of data fragment or XOR verify fragment. Because of redundant XOR fragments, file is available even when two more nodes fail. FTDSS realizes distant redundant storage. At the same time, FTDSS make file to wirte and read by fragments. FTDSS can offer more performance than stand_alone.By system software, files can be stored distributed. For user to store datas in LAN, system do a function just as little RAID and some file servers. And, it is not need more hardwares. At last, this paper uses experiment to prove that FTDSS can offer high fault-tolerance and advanced performance.
- 【网络出版投稿人】 成都理工大学 【网络出版年期】2004年 03期
- 【分类号】TP311.52
- 【下载频次】173