The Research Of Regeneration Based On Ocean Network Storage System

Serveral aspects of disaster recovery and fault tolerance for ocean network storage are presented in this thesis.In storage system data lost and fault is ineluctable, however the method that data lost is reduced and avoided is the key problem.This paper research the reliability of the storage based on anglicizing the general ocean storage structure in details; anglicizes the structure and discusses the reliability design solution about the NAS and SAN. Then we discuss the esurience method in the access and control layer of storage system, we anglicize the reliability mathematics model of different RAID level and research the fault tolerance coding techniques for multiple disk failure and restore.It will increase the reliability of storage system when the coding techniques are applied in the disk fault tolerance, especially in the RAID and dRAID. At present this paper research the Reed-Solomon, HOVER code in details.The paper present a novel coding scheme on ocean storage fault tolerance base on HOVER code and new ideal for multiple disk fault tolerance by changing the step length due to introduce into new concept Step, and finally has resoved the problem of high fault tolerance that HoVer code never satisfy in realization. This paper is a sharp contrast between code of this paper and Reed-Solomon code, and apply it in the fault tolerance by softer method, it increase the access speed efficiently, we also discover the bottle problem when it transfer the large block data. Finally the paper present simulation and test on allusion to fault tolerance.We discuss the code application advantage in the ocean storage and the process of code and encode, while it has applied in the net transfer for the reliable restore. This paper presents the code Real-time and veracity, and then ensures quickly failover and disaster recovery, consequently provides a novel method on reliability in ocean storage.