Thin Provisioning Storage Reclamation via Non-Zero Data Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage systems with thin provisioning face inefficiencies in utilizing physical storage space, as they do not effectively reclaim space allocated to data portions containing only zeros or optimally manage non-zero data units, leading to suboptimal resource utilization.
Innovation Solution
A storage system with a control layer that analyzes data portions to identify zero and non-zero units, reallocates non-zero units to designated areas, and returns allocated physical storage space to the pool when criteria are met, while ensuring data integrity and availability upon read requests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If thin provisioning is implemented to dynamically allocate storage space, then storage utilization efficiency is improved, but physical storage space allocated to data portions with mixed zero and non-zero data cannot be fully reclaimed
Solution Approach 1:
The patent segments a data portion into multiple data units and identifies non-zero data units within it. By separating the identification and handling of non-zero data units from the rest of the data portion, the system can selectively reclaim space for zero data units while preserving necessary non-zero data, thus resolving the contradiction between storage utilization efficiency and reclaimable space.
Solution Approach 2:
The patent extracts non-zero data units from data portions and relocates them to a designated area. This extraction allows the system to reclaim the physical storage space originally allocated to those data portions (now containing only zero data units) while maintaining data integrity by preserving the extracted non-zero data units in their new location.
2Productivity
If zero page reclamation is applied to data portions containing only zeros, then physical storage space is freed, but data portions with mixed zero and non-zero data cannot be optimized
Solution Approach 1:
The patent divides data portions into data units and identifies which units contain non-zero data. This segmentation enables the system to apply zero page reclamation selectively to portions containing only zero data units, while handling mixed data portions by extracting and preserving their non-zero data units, thus extending the applicability of reclamation to previously unoptimizable mixed data portions.
Solution Approach 2:
The patent applies different handling strategies to different parts of a data portion based on its content. Zero data units are reclaimed, non-zero data units are extracted and preserved, and the treatment varies by local data characteristics. This local quality approach enables optimization of mixed data portions while maintaining data integrity.
3Reliability
If traditional fat provisioning allocates predefined storage spaces to users, then data availability is ensured, but substantial parts of allocated storage space remain unused
Solution Approach 1:
The patent implements dynamic storage allocation where the physical storage space allocated to users is not fixed but adjusts based on actual data needs. The system continuously monitors data portions, identifies reclaimable space from zero data units, and reallocates it dynamically, ensuring both data availability and optimal storage utilization without the waste inherent in static fat provisioning.
Solution Approach 2:
The storage system automatically monitors its own allocated space, identifies data portions containing only zero data units or extractable non-zero data units, and performs reclamation and reallocation without external intervention. This self-service capability ensures data availability while continuously optimizing storage utilization to eliminate unused allocated space.
Data Source
AI summary
A method and system is disclosed for storage optimization of thin provisioning. Non-zero data units within data portions are re-allocated to specifically designated one or more areas in the physical storage space and the physical address of the non-zero data unit within the designated area is associated with the logical address of the non-zero data unit. In case the data portion is allocated to the physical storage, the physical storage space initially allocated to the data portion is returned to the pool of available physical storage space, thereby freeing the storage space for storing other data.


