Dynamic SSD Tier Management for I/O Performance and Wear Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage apparatuses using SSDs do not effectively utilize the variable performance and recovery properties of SSDs, leading to performance degradation and inefficient data migration between tiers, as they do not consider differences in SSD types and performance recovery.
Innovation Solution
A storage apparatus and tier control method that classifies storage areas into multiple tiers based on SSD performance and properties, dynamically changing tier assignments as SSD performance varies, utilizing a controller to manage SSDs of different types and performance levels to optimize I/O performance and extend SSD life.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If wear leveling processing is adopted to reduce the life for the number of rewrites, then storage reliability is improved, but metadata size increases and search overhead increases
Solution Approach 1:
The patent extracts and removes invalid metadata entries through reclamation processing. When wear leveling creates invalid mappings, the system identifies and removes these invalid entries from the metadata structure, thereby reducing metadata size and search overhead while maintaining the reliability benefits of wear leveling.
Solution Approach 2:
The system discards invalid metadata entries and recovers storage space by reclamation processing. Invalid mappings are identified and removed, allowing the metadata structure to be optimized and reduced in size, which directly addresses the contradiction between reliability and metadata complexity.
2Productivity
If reclamation processing is executed to release old source data, then storage efficiency is improved, but database search efficiency in metadata decreases due to invalid entries
Solution Approach 1:
The system performs preliminary reclamation processing to remove invalid metadata entries before they affect search operations. By proactively identifying and removing invalid mappings, the system maintains database search efficiency while achieving storage efficiency through space recovery.
Solution Approach 2:
The patent converts the harmful effect of invalid metadata entries into a beneficial process by implementing reclamation processing. The presence of invalid entries triggers a systematic removal process that actually improves overall system efficiency by cleaning up the metadata structure and reducing search overhead.
3Speed
If SSDs are used to process inputs and outputs at high speed, then I/O performance is improved, but performance varies depending on SSD type and property
Solution Approach 1:
The patent implements dynamic tier management that adapts to varying SSD performance characteristics. The system dynamically adjusts data placement and migration strategies based on real-time performance monitoring, allowing optimal utilization of different SSD types and properties while maintaining consistent overall performance.
Solution Approach 2:
The system changes operational parameters such as tier assignments and data migration policies based on SSD performance characteristics. By monitoring and adjusting parameters like access patterns and performance thresholds, the system optimizes I/O performance across diverse SSD types while managing the variability in their properties.
4Productivity
If data are migrated between tiers based on access frequency, then I/O performance is improved, but performance degradation occurs as storage areas are used
Solution Approach 1:
The system implements feedback-based tier management that continuously monitors storage area performance and usage patterns. Based on this feedback, the system dynamically adjusts data migration strategies to balance I/O performance with storage area lifespan, preventing premature degradation while maintaining high performance.
Solution Approach 2:
The patent employs periodic reevaluation of tier assignments and data migration policies. By periodically analyzing usage patterns and performance metrics, the system optimizes the balance between maintaining high I/O performance and extending storage area lifespan, adapting to changing conditions over time.
Data Source
AI summary
A storage apparatus and tier control method capable of executing hierarchical control based on variable performance of SSDs are suggested. With a storage apparatus equipped with a plurality of drives composed of flash memories, the storage apparatus includes a controller for hierarchically controlling storage areas provided by the plurality of drives by classifying the storage areas into a plurality of tiers with different performances and managing them, wherein the controller: decides a first tier, to which each of the drives should belong, based on performance of the drive; decides to change the tier, to which the drive should belong, from the first tier to a second tier different from the first tier based on the drive's performance, which varies depending on property of the drive; and changes the tier, to which the drive determined to change its tier should belong, from the first tier to the second tier.