Storage Device Grouping for Data Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems, when the number of corrupted storage devices exceeds the threshold, data loss becomes a high risk due to unrecoverable corruption across multiple storage servers, as all data blocks are randomly stored across devices, leading to a lack of redundancy and integrity.
Innovation Solution
The method involves grouping storage devices into non-overlapping groups based on shared slot numbers or server configurations, ensuring that each data stripe is distributed across different devices within a target storage device grouping, and establishing relationships between monitoring points and storage device groupings to manage data storage and recovery efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data blocks are randomly stored across all storage devices, then storage utilization is high, but data reliability deteriorates when multiple storage devices are corrupted
Solution Approach 1:
The patent divides storage devices into multiple non-overlapping groups, where each group contains storage devices from different storage servers. Data stripes are then distributed to different groups rather than randomly placed across all devices. This segmentation ensures that corruption in one group does not affect other groups, maintaining data reliability even when multiple devices fail.
Solution Approach 2:
The patent introduces a new dimension of organization by creating group-level abstraction over individual storage devices. Instead of managing data placement at the device level only, the system adds a grouping dimension where data stripes are allocated to groups, and within each group, blocks are distributed across devices. This dimensional change enables controlled redundancy while maintaining high storage utilization.
2Reliability
If data is distributed across all storage devices, then storage capacity utilization is maximized, but data recovery becomes difficult when corruption exceeds threshold
Solution Approach 1:
By segmenting storage devices into non-overlapping groups and distributing data stripes across different groups, the patent ensures that each group contains a complete set of data blocks necessary for recovery. This segmentation allows recovery operations to be confined to specific groups rather than requiring access to all storage devices, improving recovery capability while maintaining efficient capacity utilization.
Solution Approach 2:
The patent applies local quality by ensuring that each storage device group has the specific property of containing all necessary data blocks for independent recovery operations. This local completeness allows any single group to serve as a recovery source, enhancing overall system reliability without sacrificing storage capacity utilization.
3Productivity
If storage devices are grouped by slot numbers, then data recovery efficiency is improved, but storage device flexibility decreases
Solution Approach 1:
The patent segments storage devices into groups based on slot numbers, creating organized units that facilitate efficient recovery operations. Within each group, devices with the same slot number are collected, ensuring that data blocks can be quickly located and recovered. This segmentation improves recovery efficiency while the non-overlapping group structure maintains flexibility by allowing devices to be reassigned between groups as needed.
Solution Approach 2:
The patent introduces dynamic flexibility by allowing the grouping structure to adapt to different scenarios. Groups are defined by slot numbers but are not fixed assignments; devices can be dynamically allocated to different groups based on availability and requirements. This dynamic approach maintains adaptability while preserving the efficiency benefits of organized grouping for recovery operations.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The present disclosure relates to the technical field of storage, and disclosed thereby are a data storage method and apparatus. The method comprises: receiving a storage request carrying target data, wherein the target data comprises at least one data stripe, and each data stripe comprises a plurality of data blocks; determining, among a plurality of preset storage device groupings, a target storage device grouping used to store the target data; storing each data block of the data stripes for each data stripe in different storage devices respectively among the target storage device grouping. Thus, the risk of data loss may be reduced by employing the present disclosure.