Storage System Deduplication with Access-Aware Data Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deduplication technologies using HDDs as storage media face performance degradation due to fragmentation, leading to deteriorated I/O performance, as they randomly read data from multiple addresses, which is inefficient compared to sequential read access.
Innovation Solution
A storage system that verifies data duplication for each storage area within a logical volume, maintains unshared and shared data separately, and releases storage areas based on access frequency to optimize sequential read performance by maintaining data order and reducing unnecessary deduplication of frequently accessed data blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If deduplication is executed in a storage apparatus using HDDs, then data amount is reduced and storage cost is lowered, but I/O performance deteriorates due to random read access from multiple addresses
Solution Approach 1:
The patent segments the storage space into primary storage areas and secondary storage areas. Unshared data is stored in primary storage areas, while shared (duplicated) data is stored in secondary storage areas. This segmentation allows the system to maintain sequential access patterns for frequently accessed unshared data while still achieving deduplication benefits for shared data, thereby resolving the contradiction between data reduction and I/O performance.
2Quantity of substance
If deduplication is executed, then storage capacity is optimized, but fragmentation occurs leading to random read access patterns
Solution Approach 1:
The patent applies local quality by treating different types of data differently in terms of storage location. Unshared data that requires sequential access is placed in primary storage areas, while shared data that can tolerate random access is placed in secondary storage areas. This localized differentiation of storage quality resolves the fragmentation issue while maintaining storage capacity optimization.
3Quantity of substance
If HDDs are used as storage media, then large capacity and low cost are achieved, but random access performance is inherently lower than sequential access
Solution Approach 1:
The patent performs preliminary action by pre-organizing data into unshared and shared categories before storage. By identifying which data will be shared and which will not, the system can pre-place data in appropriate storage areas, ensuring that sequential access patterns are maintained for performance-critical data while still achieving deduplication for shared data. This preliminary organization prevents performance degradation.
Data Source
AI summary
It is provided a storage system comprising at least one controller and a storage device. The at least one controller verifies, for each predetermined storage area within a logical volume provided to the host computer, whether data of the each predetermined storage area is duplicated to another storage area. The storage device holds unshared data associated only with the storage area and shared data associated with the storage area and the another storage area in the case where the data of the each predetermined storage area is identical to the data of the another storage area. The at least one controller reads the unshared data in the case where a request to read the data is received under a state in which the unshared data and the shared data are held, and releases an area in which the unshared data is stored at predetermined timing.