Memory Management Method for Read Disturb Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The frequent read operations in non-volatile memory modules lead to read disturb issues, causing data errors and inefficiencies due to rapid accumulation of read counts, which results in decreased performance as valid data is repeatedly moved between physical blocks.
Innovation Solution
A memory management method that identifies physical units with high read counts exceeding a threshold, divides valid data into multiple copies, and distributes these copies across recycling units during a data merging process to reduce excessive data transfers and mitigate read disturb risks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If valid data is moved to a new physical block when read count is high, then read disturb risk is reduced, but data transfer frequency increases and overall efficiency decreases
Solution Approach 1:
The patent segments valid data from a single physical block into multiple copies distributed across multiple recycling physical blocks. When data is frequently read, instead of moving all data to one new block, the system creates multiple distributed copies, reducing the read count accumulation in any single block and minimizing the frequency of data movement operations.
Solution Approach 2:
The patent transitions from a single-dimension data storage approach (one physical block per logical unit) to a multi-dimensional distribution approach where data copies are spread across multiple recycling physical blocks. This dimensional expansion allows the system to handle high read counts more efficiently by distributing access patterns across multiple blocks.
2Ease of operation
If the same data is frequently read, then data accessibility is improved, but read count accumulates quickly causing repeated data movements
Solution Approach 1:
The patent creates multiple copies of valid data from risk physical units and distributes them to recycling physical blocks. This copying mechanism allows the system to maintain high data accessibility while reducing read count accumulation in any single block, as read operations can be distributed across multiple copies rather than concentrating on one physical block.
3Speed
If read count of physical block increases, then data access speed is maintained, but hardware loss and data errors increase
Solution Approach 1:
The patent implements preliminary identification of risk physical units by monitoring read counts and comparing them against thresholds. Before read disturb causes data errors, the system proactively identifies physical units with high read counts and redistributes their data to recycling blocks, preventing data integrity issues before they occur.
Solution Approach 2:
The patent employs a feedback mechanism where the memory control circuit unit continuously monitors read counts of physical units, compares them against predefined thresholds, and triggers data redistribution when thresholds are exceeded. This feedback loop maintains data integrity by responding to read count conditions in real-time.
Data Source
AI summary
A data management method, a memory storage device and a memory control circuit unit. The method includes: executing one or more read commands, and recording a physical unit having a first variation of a read count greater than a read disturb threshold as a risk physical unit; and when a data merging process is performed, dividing valid data stored in the risk physical unit into a plurality of copies and copying the copies into a plurality of recycling units.