A method and system for realizing memory data reliability based on raid5

Abstract

The invention discloses a method and system for implementing reliability of memory data on the basis of an RAID5. The method comprises the steps of: partitioning part of memory data serving as a cache from each distributed storage server, and then constructing the RAID5; placing data accessed to usually into a memory database; besides storing basic data, storing additional metadata information and strip information for identifying the RAID5. When all storage servers normally operate, a client reads data from a memory; when a single-point fault is generated, the client can obtain data of the same strip by the stored metadata information, recover the data at the memory level and provide the data to a user without accessing a hard disk of which an access speed is lower than that of the memory by multiple orders of magnitude, so that reliability of a storage system is improved; and meanwhile, by constructing the RAID5, the data storage redundancy is reduced when reliability is achieved, so that cost is saved.

Description

technical field [0001] The invention relates to the field of computers, in particular to a method and system for realizing memory data reliability of a distributed storage system based on RAID5. Background technique [0002] With the deep popularization of the Internet and the vigorous development of the mobile Internet, more and more web applications have emerged, which have high requirements for performance parameters such as access delay and throughput, such as social networks, information retrieval and e-commerce. Moreover, with the popularization of smart phones, the number of web application users is also increasing day by day, and millions or even hundreds of millions of users visit at every turn, which poses a severe test to the system architecture and performance. [0003] Traditional relational databases have three problems in the face of massive data: first, high performance, and the need for high concurrent read and write of the database. A huge number of users ...

AI Technical Summary

Problems solved by technology

[0003] Traditional relational databases have three problems in the face of massive data: first, high performance, and the need for high concurrent read and write of the database

The backup system is a process of doubling the storage cost, and the performance of persisting to disk will be significantly affected

Images