Controlling the Placement of Data in a Storage System

Abstract

A method, computer readable storage medium and computer system for controlling the allocation of data to one of a plurality of storage units of a storage system, the method comprising: accessing a source storage unit comprising the data; gathering file system level (FS-level) metadata from the source storage unit; analyzing the gathered FS-level metadata for determining if the data should be moved to one of the other storage units, said other storage unit acting as a destination storage unit; and in case the data should be moved, displaying an indication of the destination storage unit and / or automatically moving the data to the determined destination storage unit.

Description

CROSS-REFERENCE TO RELATED APPLICATION(S)[0001]This application is a continuation of and claims the benefit of priority to U.S. patent application Ser. No. 13 / 612,196, filed on Sep. 12, 2012 and entitled “Controlling a Placement of Data in a Storage System” which claims priority to EP 11193300.8 filed on Dec. 13, 2011, the contents of which are hereby incorporated by reference.FIELD OF THE INVENTION[0002]The invention relates to the field of storage management, and more particularly to a method for controlling the allocation of data in one of a plurality of storage units of a storage system.BACKGROUND[0003]Data management systems are used to automatically control the storage of data by automating the placement of data according to various criteria. Data placement in such systems comprises a decision on which one of a plurality of storage units the data should be stored for minimizing the costs without causing the quality of data provisioning services based on said data to decline si...

AI Technical Summary

Benefits of technology

The patent text discusses a storage management system that uses meta data collected at the file-system level to decide which data should be moved to another storage unit, in order to optimize data storage. This decision is based on higher-level abstraction, rather than specific details of the storage blocks, which allows for easier abstraction from different types of storage media. The data is moved by re-mapping physical storage block addresses and logical storage block addresses, and the movement is transparent to the user. The technical effect of the patent is to improve the efficiency of data storage and management by optimizing data placement and utilizing different types of storage media based on their performance and cost.

Problems solved by technology

One drawback of today's HSM implementations is that they often rely on a single type of file system as they require support in the file system for hiding from the user the actual physical location of the stored data.

Not every piece of data are, however, organized as an individual file system node, and different storage devices may comprise file systems of different types (ext2, ext3, FAT32, NTFS or the like) which may be incompatible with each other.

One major deficiency in prior art data management systems which are based on tiered storage is that for deciding whether to migrate a volume's data from a storage volume in one storage pool to another volume of another storage pool (or, analogously, for deciding if one or more physical storage volumes of another pool should be assigned to a particular logical volume) in order to optimize disk usage with regards to price / performance, said systems are often not able to correctly determine if the data of a particular storage unit is of relevance for the user and / or operator of the data management system.

The information used as input for said decision is often insufficient for determining the real relevance and future usage frequency of a particular piece of data.

Thus, information gathered by prior art systems is often unsuited for determining the optimum destination storage unit for storing data in a given use case scenario.

For example, data causing very high performance load (such as video stream data) may in fact be very unimportant for a given industrial control workflow and allocating such kind of data to highly efficient, expensive storage devices may actually be counterproductive.

Images