Cloud disk as well as realization method and realization device thereof

Abstract

The invention discloses a realization method for a cloud disk. The realization method comprises the following steps: receiving a file request message, and applying a lock service corresponding to the file request message according to the file request message; generating an SCSI commend corresponding to the file request message according to the lock service; receiving file information corresponding to the file request message, which is looked up according to the SCSI commend. The invention further discloses a cloud disk, a server, a cloud storage server, a cloud mapping server and a distributed lock server.

Description

technical field [0001] The invention relates to cloud computing technology, in particular to a cloud disk and its implementation method and device. Background technique [0002] Currently, the more popular cloud disks include Netease Cloud Disk, Microsoft Cloud Disk, and Amazon Cloud Disk; and large-capacity email spaces, such as QQ Space, Fetion Space, etc. In the prior art, cloud disks are mainly realized through the following three storage technologies: First, the storage technology of Storage Area Network (SAN, Storage Area Network) is realized based on the optical fiber network with circular threaded (FC) interface, referred to as FC-SAN storage technology; second, SAN storage technology based on Internet Protocol (IP, Internet Protocol), referred to as IP-SAN storage technology; third, distributed file system. [0003] However, the above three storage technologies have the following problems: [0004] (1) The input and output (I / O) performance of the distributed file...

AI Technical Summary

Problems solved by technology

[0004] (1) The input and output (I / O) performance of the distributed file system is poor, because the distributed file system is generally implemented by a two-layer file system, and the two-layer file system includes a distributed file subsystem and a local file subsystem , therefore, there are more layer structures and non-I / O access operations on the data I / O path, resulting in poor I / O performance of the distributed file system;

[0005] (2) Although the FC-SAN storage technology has the characteristics of high scalability, etc., the FC-SAN storage technology is limited to the FC-SAN storage network and cannot effectively utilize the cheap and widespread IP network; moreover, the In the FC-SAN storage technology, the client must use a host bus adapter whose interface is FC, that is, FC-HBA and driver software, etc. Therefore, the FC-SAN storage technology cannot be seamlessly connected with the IP network;

[0006] (3) In the IP-SAN storage technology, there is a data consistency problem when multiple servers perform parallel I / O, and an additional cluster technology or distributed file system technology is needed to solve the data consistency problem; moreover, the Multiple devices in IP-SAN storage technology cannot be effectively cloudified, and cloud storage with unlimited space cannot be realized;

[0007] (4) The deployment and maintenance costs of the existing three storage technologies are high, especially the FC-SAN storage system, which requires professional high-tech personnel to maintain and manage

Images