Front End LUN Slice Relocation for Failure Domain Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Front end LUNs face increased failure probability due to the use of multiple backend devices for storage, as the failure of any device can lead to LUN failure, and existing systems struggle to optimize performance and failure domains effectively.
Innovation Solution
The method involves identifying misassigned slices across different data tiers and relocating them to optimize the failure domain by moving slices from one front end LUN to another based on performance benchmarks and data tier assessments, thereby reducing the failure domain while maintaining performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple backend devices are used to construct front end LUNs for improved performance, then storage performance is improved through parallel access, but the failure probability increases since any single device failure can cause LUN failure
Solution Approach 1:
The patent segments the LUN into multiple slices distributed across different backend devices and data tiers. By dividing the LUN into slices that can be independently allocated to different tiers (hot tier for frequently accessed data, cold tier for less frequently accessed data), the system maintains performance benefits while enabling selective redundancy. This segmentation allows the system to optimize the failure domain by determining which slices require replication based on their access patterns and importance.
Solution Approach 2:
The patent applies local quality by treating different slices of the LUN differently based on their access characteristics. Frequently accessed slices are placed in the hot tier with higher reliability, while less frequently accessed slices are placed in the cold tier. This localized differentiation allows the system to optimize reliability for critical data while maintaining cost-effectiveness for less critical data, resolving the contradiction between performance and reliability.
2Quantity of substance
If data is distributed across multiple data tiers with different performance characteristics, then storage cost is optimized, but the complexity of managing and adjusting data tier assignments increases
Solution Approach 1:
The patent implements feedback mechanisms to automatically adjust data tier assignments based on real-time access patterns and performance metrics. The system continuously monitors slice access characteristics and dynamically relocates data between tiers as needed. This automated feedback-driven adjustment reduces manual management complexity while optimizing the balance between storage cost and performance.
Solution Approach 2:
The patent introduces dynamic data tier assignment that adapts to changing access patterns and workload conditions. Rather than static tier placement, the system dynamically adjusts which slices reside in hot or cold tiers based on current access characteristics. This dynamic approach simplifies management by automating adjustments while optimizing storage resource allocation.
Data Source
AI summary
A method, computer program product, and computing system for identifying at least one misassigned slice that is associated with a first data tier and is located in a first front end LUN. The at least one misassigned slice is being accessed contrary to the first data tier. The failure domain of the first front end LUN is determined. At least one replacement slice that is associated with a second data tier and is located in a second front end LUN is identified based, at least in part, upon the failure domain of the first front end LUN. The at least one misassigned slice is moved to the second front end LUN. The at least one replacement slice is moved to the first front end LUN.