SSD Bad Block Conversion via Dynamic Memory Tiering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing number of bad blocks in NAND flash memory due to higher density and capacity, leading to decreased efficiency, increased wear-leveling, and shorter lifespan in solid state drives (SSDs), is not effectively addressed by current solutions like overprovisioning, which results in reduced performance and efficiency.
Innovation Solution
A system that monitors the condition of blocks in SSDs and formats them to store a fewer number of bits per cell when thresholds are met, downgrading from QLC to MLC to SLC and using an SLC cache to utilize previously wasted space, thereby improving performance and longevity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If higher density and capacity memory elements (e.g., QLC) are used to increase storage capacity, then the cost and capacity increase, but the data retention decreases and the number of bad blocks increases
Solution Approach 1:
The patent segments the storage system into multiple memory types (QLC, MLC, SLC) with different reliability characteristics. Bad blocks are dynamically reclassified and relocated to appropriate memory segments based on their condition, allowing high-capacity QLC to be used while maintaining system reliability through segmentation.
Solution Approach 2:
The patent changes the operational parameters of memory blocks by dynamically adjusting their classification (good, bad, reserved) and relocating them between different memory types. This parameter change allows the system to adapt to degradation and maintain reliability while using high-density memory.
2Reliability
If overprovisioning is used to handle bad blocks, then reliability is maintained, but performance and efficiency decrease due to frequent garbage collection and reduced IOPS
Solution Approach 1:
The patent implements dynamic bad block management where blocks are continuously monitored and reclassified based on their condition. This dynamic approach allows the system to adapt to block degradation in real-time, maintaining reliability without the performance penalty of static overprovisioning and frequent garbage collection.
Solution Approach 2:
The system employs feedback mechanisms by monitoring block conditions and using this information to dynamically adjust block classification and relocation strategies. This feedback loop enables the system to maintain reliability while optimizing performance by only relocating blocks when necessary.
3Reliability
If blocks are downgraded from QLC to MLC to SLC formats, then reliability and data retention are improved, but the storage capacity per block decreases
Solution Approach 1:
The patent adds a temporal dimension to storage by implementing a tiered memory hierarchy (QLC, MLC, SLC) with different retention characteristics. Data can be moved between these dimensions based on access patterns and block condition, allowing the system to maintain high capacity while improving reliability when needed.
Solution Approach 2:
The patent implements a nested memory structure where SLC blocks are embedded within MLC blocks, and MLC blocks are embedded within QLC blocks. This nesting allows efficient utilization of storage capacity across different memory types while maintaining the ability to downgrade blocks for improved reliability.
Data Source
AI summary
One embodiment facilitates data placement. During operation, the system monitors a condition of a plurality of blocks of a non-volatile memory. The system determines that a condition of a first block falls below a first predetermined threshold, wherein the first block has a first capacity. The system formats the first block to obtain a second block which has a second capacity, wherein the second capacity is less than the first capacity.