Striped Disk Data Rebalancing Across RAID Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data striping systems lose I/O load balancing when new RAID ranks are added to a set across which data striping has already been performed, requiring inefficient redistribution of data.
Innovation Solution
A method and system for rebalancing striped disk data by distributing extents evenly across both pre-existing and newly added RAID arrays with minimal data movement, using a storage control unit to organize extents into groups that span volumes and redistribute them equally among all arrays, minimizing the number of copy operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If new RAID ranks are added to an existing striped disk system, then storage capacity and performance are improved, but I/O load balancing is lost and data redistribution becomes inefficient
Solution Approach 1:
The patent segments the data distribution problem by treating extents as discrete units that can be independently redistributed. By dividing the data into individual extent units and redistributing them systematically across the expanded RAID array, the system maintains load balancing without requiring complete data remapping, thus resolving the contradiction between capacity expansion and I/O efficiency
Solution Approach 2:
The patent implements a dynamic data redistribution mechanism that automatically adapts when RAID ranks are added or removed. The system continuously monitors the RAID array configuration and dynamically rebalances data distribution across extents, ensuring optimal I/O load balancing at any given configuration state while allowing flexible capacity changes
2Productivity
If data is redistributed across all storage units after adding new RAID arrays, then I/O load balancing is restored, but the number of data copy operations increases
Solution Approach 1:
The patent applies local quality by selectively redistributing only the necessary portions of data rather than performing a complete system-wide redistribution. By identifying specific extents that need rebalancing and moving only those local units, the system restores I/O load balancing while minimizing the time and resources required for data movement
Solution Approach 2:
The patent implements partial action by performing only the minimum necessary data redistribution to achieve load balancing. Rather than redistributing all data across the array, the system calculates and executes only the essential extent movements needed to equalize the load, thereby reducing redistribution time while maintaining balanced I/O performance
3Manufacturing precision
If extents are moved multiple times during redistribution, then data is evenly distributed across all arrays, but the complexity and time of the redistribution process increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating the optimal destination for each extent before initiating the redistribution process. The system determines the target RAID array for each extent in advance based on the desired uniform distribution pattern, ensuring that each extent is moved directly to its final position in a single operation, thereby achieving perfect distribution uniformity while minimizing movement次数 and process complexity
Data Source
AI summary
Provided are a method, system, and article of manufacture, where a plurality of extents are stored in a first set of storage units coupled to a controller. A determination is made that a second set of storage units has been coupled to the controller. The plurality of extents are distributed among all storage units included in the first set of storage units and the second set of storage units.