Adaptive Memory Scans for NAND Wear and QoS Balance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory sub-systems perform memory scans and data integrity checks based on static, end-of-life specifications, leading to over-scanning and over-folding, which degrade performance and quality of service, especially at the beginning of a memory device's life.
Innovation Solution
Adaptive adjustment of scan frequency and data state metric thresholds based on the number of program erase cycles and performance metrics, using lookup tables to optimize scan parameters throughout the memory device's lifetime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If static end-of-life scan parameters are used throughout the memory device lifetime, then data integrity is ensured at end-of-life, but performance and quality of service degrade at the beginning of life due to over-scanning and over-folding
Solution Approach 1:
The patent applies dynamics by transitioning from static scan parameters to dynamic, adaptive scan parameters that change based on the memory device's age and wear level. The scan management component continuously adjusts scan frequency and data state metric thresholds based on program erase cycle counts and performance metrics, allowing the system to optimize between data integrity and performance at different stages of the device lifecycle.
Solution Approach 2:
The patent implements parameter changes by modifying scan frequency and data state metric thresholds based on the number of program erase cycles and measured performance metrics. Lookup tables store different parameter sets that are selected based on the device's wear level, enabling the system to use stricter parameters when needed and more relaxed parameters when performance is prioritized.
2Reliability
If frequent scans are performed to ensure data integrity, then reliability improves, but performance and quality of service deteriorate due to excessive scanning operations
Solution Approach 1:
The patent applies partial or excessive action by performing scans at varying frequencies rather than consistently at maximum frequency. When the memory device is new and reliable, scans are performed less frequently (partial action). As the device ages and wear increases, scan frequency increases (excessive action) to ensure data integrity, thus optimizing the balance between reliability and performance overhead.
Solution Approach 2:
The patent implements feedback by continuously monitoring performance metrics and program erase cycle counts, then using this information to adjust scan frequency and data state metric thresholds. The scan management component receives feedback about device wear and performance, and dynamically modifies scanning behavior accordingly, reducing unnecessary scans when the device is healthy and increasing scans when wear is detected.
Data Source
AI summary
One or more media scan parameters associated with a memory device are maintained. A number of program erase cycles associated with the memory device is identified. Responsive to determining that the number of program erase cycles satisfies a criterion, one or more adjusted media scan parameters are generated by adjusting the one or more media scan parameters. A media scan of the memory device is performed according to the one or more adjusted media scan parameters.


