Storage clustering system and method for providing access to clustered storage

Abstract

The invention discloses a storage cluster system and a method for providing access to the cluster storage. The system includes multiple storage front ends and multiple cluster modules. At least one of the clustering modules receives an access command from a client. When an access instruction indicates to write a data item, one of the clustering modules calls at least one operation module to calculate at least one derived value of the data item, and at least one of the clustering modules writes the derived value according to the index through one of the storage front ends or at least part of the data item and update the metadata accordingly. When an access command indicates to read a data item, one of the clustering modules checks the metadata, so that one of the clustering modules reads the data item through the aforementioned storage front end. When the storage front end returns a derived value instead of a data item, the read clustering module checks the index based on the derived value to synthesize the data item to the client.

Description

technical field [0001] The present invention relates to storage clustering, and more particularly to performance considerations and data de-duplication in clustered storage. Background technique [0002] Traditional storage architectures generally can only scale up vertically but cannot scale out horizontally. In other words, the number and specifications of hosts in the architecture remain the same, and the only way to install or replace more hard disks is to install or replace more hard disks when more storage space is needed. Therefore, vertical expansion cannot be expanded without limit, and it is not beneficial to performance. It is very time-consuming to migrate data from the earlier smaller hard drive to the newly purchased larger hard drive during vertical expansion, not to mention that the capacity of the hard drive is not directly proportional to the asking price. [0003] Clustering storage and managing nodes as units can partially solve the above problems. Howe...

AI Technical Summary

Problems solved by technology

However, in the case of Small Computer System Interface (SCSI) storage clusters, the clustering and granting of access rights occurs in the logical volume management (LVM) behind the SCSI target. , the client itself needs to have the ability to identify targets, and each target can only control eight to sixteen SCSI devices, and if distributed lock management (DLM) is also used for all targets, the performance will be reduced. Unbelievably low

Images