Stripe Management via Device Correlation for Storage Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing RAID-based storage systems face challenges in evenly distributing extents across storage devices, leading to uneven workloads and potential access bottlenecks, which can result in longer reconstruction times and increased data loss risks during disk failures.
Innovation Solution
A method for managing stripes in a storage system by determining device correlations between storage devices, selecting a storage device with a low device correlation for adding an extent to the stripe, and using stripe correlations to distribute subsequent extents, ensuring balanced workloads and improved access performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If extents are distributed across storage devices without considering device correlation, then the stripe creation process is simple, but the workload distribution becomes uneven leading to access bottlenecks
Solution Approach 1:
The system pre-calculates and stores device correlation values before stripe creation. When a stripe needs to be created, the management module directly queries pre-computed correlation data to select storage devices, avoiding real-time complex calculations and achieving both balanced workload distribution and efficient stripe creation
Solution Approach 2:
The patent introduces a device correlation metric as an intermediary parameter that mediates between simple distribution and balanced distribution. This correlation value serves as a selection criterion that guides the stripe creation process to achieve balanced workload without requiring complex real-time analysis
2Loss of time
If storage devices are selected randomly for extent distribution, then the device selection process is fast, but reconstruction time increases when disk failures occur
Solution Approach 1:
Device correlation values are pre-calculated and stored in the system before any stripe creation or disk failure occurs. When selection is needed, the system simply queries this pre-computed data, making both stripe creation and failure reconstruction efficient without real-time computation overhead
Solution Approach 2:
The system uses pre-computed device correlation values as feedback to guide storage device selection. This feedback mechanism ensures that devices with lower correlations (indicating better distribution characteristics) are selected, optimizing both normal operation and failure reconstruction scenarios
3Reliability
If extents are concentrated on fewer storage devices, then the stripe creation process is simpler, but the risk of data loss increases during disk failures
Solution Approach 1:
The system pre-computes device correlation values that reflect the distribution characteristics of storage devices. These pre-calculated values guide the selection process to automatically distribute extents across devices with lower correlations, achieving both simplified management and improved data safety
Solution Approach 2:
The patent changes the selection parameter from simple random or sequential selection to correlation-based selection. By using device correlation values as the selection criterion, the system transforms the distribution strategy to achieve better extent dispersion and reduced data loss risk while maintaining manageable complexity
Data Source
AI summary
Techniques manage stripes in a storage system. In accordance with such techniques, in the storage system including a plurality of storage devices, a device correlation of a storage device in the plurality of storage devices is determined according to a request for creating a stripe in the storage system. The device correlation indicates distribution of a set of stripes created in the storage system between the storage device and other storage devices than the storage device. One storage device is selected from the plurality of storage devices based on the determined device correlation. One extent in the selected storage device is added to the stripe. As such, all extents in the stripe can be distributed as evenly as possible in the plurality of storage devices in the storage system, thus ensuring load balancing of the storage system and more effective management of the stripes in the storage system.