Mapped RAID Extent Sub-Groups for Parallel Rebuild
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional RAID data storage systems face limitations in adding new disks and suffer from increased rebuild times due to write bandwidth bottlenecks, especially when concurrent failures occur during the rebuilding process, leading to potential data loss and performance degradation during sequential write operations.
Innovation Solution
The implementation of mapped RAID technology, which divides physical data storage drives into contiguous regions called drive extents, allocates these extents into RAID extents organized in sub-groups across multiple physical drives, allowing for concurrent rebuilding and improved performance by distributing write operations across multiple drives, thereby reducing rebuild time and enhancing fault tolerance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional RAID systems are used with fixed stripe width, then data redundancy and fault tolerance are provided, but the system cannot flexibly add new disks and rebuild times increase due to write bandwidth bottlenecks
Solution Approach 1:
The patent divides the RAID storage space into multiple extent sub-groups, where each sub-group contains a subset of RAID extents. This segmentation allows independent allocation and rebuilding of subsets of data, enabling parallel rebuild operations across multiple sub-groups. When a disk fails, only the affected sub-groups need to be rebuilt rather than the entire RAID array, significantly reducing rebuild time and allowing flexible disk addition without system-wide reconfiguration.
2Productivity
If data is striped across multiple drives in traditional RAID, then write throughput is improved, but write bandwidth bottlenecks occur during rebuild operations
Solution Approach 1:
By segmenting the RAID array into multiple extent sub-groups, the patent enables parallel write operations to different sub-groups during normal operation, maintaining high write throughput. During rebuild operations, multiple sub-groups can be rebuilt simultaneously across different physical drives, distributing the write bandwidth demand and eliminating the bottleneck that occurs in traditional RAID where all rebuild writes must pass through a single drive.
Solution Approach 2:
The patent introduces a new dimension of organization by creating extent sub-groups that are independent of the traditional stripe width constraint. This allows the system to operate in multiple dimensions simultaneously: traditional striping within each extent for data distribution, and extent sub-grouping for parallel rebuild operations. This multi-dimensional approach enables the system to achieve both high write throughput during normal operation and efficient parallel rebuilds when failures occur.
3Reliability
If concurrent disk failures occur during rebuild in traditional RAID, then data loss risk increases, but system complexity must be maintained
Solution Approach 1:
The patent segments the RAID array into multiple independent extent sub-groups, where each sub-group maintains its own data and parity information. This segmentation provides inherent fault isolation: if concurrent disk failures occur during a rebuild operation, only the affected sub-groups are impacted while other sub-groups remain operational. This reduces the overall data loss risk compared to traditional RAID where a failure during rebuild can compromise the entire array, while maintaining similar system complexity through the use of existing RAID metadata structures.
Data Source
AI summary
Technology is provided within a mapped RAID system that creates multiple RAID extent sub-groups in a RAID mapping table. Each of the RAID extent sub-groups contains a unique set of consecutively located RAID extents within the RAID mapping table, and a unique corresponding set of drive extents is allocated to the RAID extents contained in each one of the RAID extent sub-groups. Drive extents are allocated to each RAID extent sub-group such that no drive extent is allocated to any RAID extent contained in the RAID extent sub-group that is located on the same hard disk drive as any other drive extent allocated to any other RAID extent contained in that RAID extent sub-group. Data striping may be performed based on the RAID extent sub-groups, such that host data is divided into data elements that are striped across the RAID extents contained in one or more individual RAID extent sub-groups.