Feature processing method for block-level data in hierarchical storage system

Abstract

The invention provides a feature processing method for block-level data in a hierarchical storage system. The method includes: setting a hierarchical storage system architecture comprising a data feature collector, a storage resource manager, a strategy configuration manger and a migration manager; allowing the storage resource manager to integrate storage resources to form a storage resource pool capable of hierarchical storage, and mapping the storage resources into a logical volume to use in an upper-level file system; allowing the data feature collector to be responsible for monitoring upper-level transmitted I / O requests, counting reference frequency features for a certain data block, analyzing by a data feature analysis method, and informing the migration manager of analysis results to allow for subsequent hierarchical management. Compared to the prior art, the feature processing method for block-level data in the hierarchical storage system allows hot spot data to be discovered and judged during the process of hierarchical data storage and management and allows the hot spot data to be judged more accurately.

Description

technical field [0001] The invention relates to the field of intelligent data management, in particular to a block-level data feature processing method in a hierarchical storage system. Background technique [0002] Data hierarchical storage technology is mainly based on the data characteristics of the monitored object to reasonably match storage resources with hot data at the storage virtual layer to achieve efficient utilization of storage resources. Among the existing hierarchical storage data feature analysis methods, one is to use various data feature attributes included in the file object, such as file size, type, etc. to obtain and analyze data features, and classify and manage data with different features. However, the hierarchical storage based on the file level uses the file as the basic unit of data feature statistics. If the file is relatively large and only local information is frequently accessed, then there is a problem of inaccurate positioning of hot data, w...

AI Technical Summary

Problems solved by technology

However, the acquisition and analysis of data features is difficult, especially for the determination of block-level hotspot data.

If the data migration decision is made solely based on the comparison between the number of references to the block and the threshold value as the criterion for judging whether the data is hot or cold, it is likely to cause data jitter problems, that is, the data will be repeatedly upgraded / downgraded in a short period of time operation, this method will lead to the consumption of system resources, which is not conducive to improving system performance

Images