RAID Data Shard Decommission via Geometry Shrinkage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems, restoring data in RAID geometry can be time-consuming and inefficient, as rebuilding entire segments is often required when a data shard or portion of it is corrupted, which consumes time and computational power.
Innovation Solution
A method is implemented where a storage system controller detects anomalies in a RAID group, determines if a data shard should not be corrected within the group, and removes it, recalculating error recovery codes from the remaining data shards without the affected shard, thereby improving RAID efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional RAID rebuild process is used when a data shard is corrupted, then data integrity is restored, but the process is time-consuming and consumes excessive computational power
Solution Approach 1:
The patent segments the RAID group into multiple data shards and independently identifies corrupted shards through anomaly detection. Instead of rebuilding the entire RAID group, only the specific corrupted shard is targeted for removal or repair, significantly reducing rebuild time while maintaining data integrity through selective rather than comprehensive restoration.
Solution Approach 2:
The patent applies partial action by performing anomaly detection and removal operations only on the specific corrupted data shard rather than the entire RAID group. This partial approach to data restoration reduces computational overhead and time consumption while still achieving the goal of restoring data integrity to the affected portion.
2Reliability
If traditional RAID rebuild process is used when a data shard is corrupted, then data integrity is restored, but computational power consumption increases
Solution Approach 1:
By segmenting the RAID group into individual data shards and applying anomaly detection to identify only the corrupted shard, the patent reduces the scope of computational operations. The error recovery code recalculation is performed only for the affected shard rather than the entire RAID group, significantly reducing computational power consumption while maintaining data integrity.
Solution Approach 2:
The patent performs partial computational action by limiting anomaly detection, shard removal, and error recovery code recalculation to only the corrupted data shard. This partial approach avoids the excessive computational power consumption of traditional full RAID group rebuilds while still achieving data integrity restoration for the affected portion.
3Reliability
If entire RAID group is rebuilt when a data shard is corrupted, then data integrity is ensured, but system productivity decreases
Solution Approach 1:
The patent segments the RAID group into independent data shards and applies targeted anomaly detection to identify only the corrupted shard. This segmentation enables selective removal and error recovery code recalculation for only the affected shard, maintaining data integrity while preserving overall system productivity by avoiding unnecessary operations on healthy data.
Solution Approach 2:
The patent applies partial action by performing anomaly detection, shard removal, and error recovery operations only on the corrupted data shard rather than the entire RAID group. This partial approach maintains data integrity for the affected shard while preserving system productivity by minimizing the impact on overall RAID operations and avoiding unnecessary computational overhead.
Data Source
AI summary
In one implementation, a method includes detecting, by a storage system controller associated with a plurality of storage devices, a first anomaly corresponding to a first data shard of a redundant array of independent disks (RAID) group and determining, by the storage system controller, that the first anomaly corresponding to the first data shard is not to be corrected within the RAID group. The method further includes removing the first data shard from the RAID group and recalculating, by the storage system controller, one or more error recovery codes corresponding to the RAID group from a plurality of remaining data shards of the RAID group, without the first data shard.