Dynamic SSD Over-Provisioning for Service-Based Capacity Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The fixed over-provisioning (OP) size in solid state disks (SSDs) results in capacity waste, as it is determined by hardware and manufacturer policy, and cannot be optimally utilized by the storage system.
Innovation Solution
A method to dynamically allocate over-provisioning capacity based on the configuration requirements of the service, allowing the SSD to partition its capacity into user and system available spaces and OP, adjusting these based on events or state information to optimize space utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed over-provisioning (OP) size is set by hardware and manufacturer policy, then the SSD can maintain stable internal logic operations, but the storage system experiences capacity waste due to inability to optimally utilize the OP space
Solution Approach 1:
The patent implements dynamic OP allocation where the over-provisioning capacity is no longer fixed but can be adjusted based on service requirements. The storage system can dynamically allocate OP space to different services, allowing the OP capacity to adapt to changing workload demands while maintaining the stability benefits of reserved space for SSD internal operations.
Solution Approach 2:
The patent changes the parameter of OP capacity from a static fixed value to a dynamic adjustable parameter. By allowing the OP size to be modified based on service configuration requirements, the system optimizes the balance between maintaining reliable internal operations and maximizing storage capacity utilization for user services.
2Reliability
If a large fixed OP capacity is reserved to ensure sufficient space for wear leveling and garbage collection, then SSD endurance and write performance are improved, but user available capacity is reduced
Solution Approach 1:
The patent enables dynamic adjustment of OP capacity allocation based on actual service requirements. Instead of reserving a large fixed OP capacity that reduces user space, the system can dynamically allocate the necessary OP space for wear leveling and garbage collection while maximizing user available capacity. The OP size adapts to the actual needs of different services rather than being overly provisioned.
Solution Approach 2:
The patent modifies the OP capacity parameter from a large fixed value to an optimized variable value. By adjusting the OP parameter based on service configuration and actual workload requirements, the system maintains sufficient space for SSD internal operations (wear leveling, garbage collection) while minimizing the impact on user available capacity.
3Productivity
If the OP capacity is set to meet high-performance service requirements, then write performance and flash endurance are enhanced, but the OP space is excessive for low-performance services, causing capacity waste
Solution Approach 1:
The patent implements service-specific dynamic OP allocation where each service can be configured with its own OP capacity requirements. High-performance services receive larger OP allocations to maximize write performance and endurance, while low-performance services receive smaller OP allocations to minimize capacity waste. The system dynamically adjusts OP distribution based on the performance needs of different services rather than applying a uniform fixed allocation.
Solution Approach 2:
The patent applies different OP capacity qualities to different services based on their specific requirements. Instead of a uniform OP allocation, the system provides locally optimized OP space for each service - high-performance services receive more OP space for wear leveling and garbage collection, while low-performance services receive less. This local quality differentiation eliminates capacity waste for services that don't require high OP allocations.
Data Source
AI summary
Example hard disk management methods, hard disk control methods, and related devices are provided. One example method includes obtaining an available physical capacity of a hard disk, and allocating a hard disk over-provisioning (OP) capacity to the hard disk based on a configuration requirement of a service configured by the hard disk. A setting instruction is generated, where the setting instruction instructs the hard disk to divide the available physical capacity into a hard disk available capacity and the hard disk OP capacity. The setting instruction is then sent to the hard disk.