Reliability And Fault Tolerance Research Of Distributed Ocean Storage System

Reliability is the key problem of the storage system, the high availability of data requires the storage system multiple implementation method.We research the reliability of the storage based on analysis of the general ocean storage structure in details; anglicize the structure and discuss the reliability design solution about the NAS and SAN. Then we focus on 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. We focus on the EVENODD, XCODE, HOVER, RS code in details and emphasize the RS code principle, optimize the RS code and apply it in the RAID by softer method, it increase the access speed efficiently, we also notice the bottle neck problem when it transferred the large block data. Then we have the new idea for multiple disk fault tolerance, we present a method to settle the multiple disk failure problem by change the step length. We discuss the tornado code application possibility in the RAID, while it has applied in the net transferation for the reliable restore. We analyze the benefit by discussing the tornado code principal; we show the outstanding real time and veracity with the Tornado code, we also indicate the problem toapply the Tornado code in the RAID, it can ensure the disk failure restoration and data reconstruct, so it brings a new path for the reliability of the storage system.