Memory Programming Using Read-Error Feedback to Cut RBER
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Non-volatile memory devices face issues with read errors that lead to increased latency, reduced performance, and decreased lifetime due to heroic recovery mechanisms and marking of blocks as bad, affecting effective overprovisioning and write amplification.
Innovation Solution
Adjusting program verify levels and voltages based on previous read errors to reduce uncorrectable bit error rates, thereby minimizing the need for heroic recovery and conserving storage capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If heroic recovery mechanism is applied to recover uncorrectable read errors, then data recovery capability is improved, but read latency increases significantly and read performance decreases
Solution Approach 1:
The system performs preliminary actions by adjusting program verify levels and voltages based on prior read errors before new write operations occur. This proactive adjustment prevents future read errors from occurring in the first place, eliminating the need for heroic recovery mechanisms and their associated latency penalties.
Solution Approach 2:
The system implements feedback by using information from prior read errors to adjust subsequent program verify levels. The controller analyzes read error patterns and modifies programming parameters accordingly, creating a closed-loop system that continuously improves reliability without requiring costly recovery operations.
2Reliability
If heroic recovery mechanism is applied to recover read errors, then data recovery capability is improved, but quality of service decreases
Solution Approach 1:
By performing preliminary adjustments to program verify levels based on prior errors, the system prevents read errors before they occur, maintaining high read performance without the need for heroic recovery operations that degrade quality of service.
Solution Approach 2:
The feedback mechanism uses read error information to optimize future programming operations, ensuring data is written with sufficient margin to prevent read errors, thereby maintaining high read performance and quality of service.
3Ease of manufacture
If blocks are marked as bad due to uncorrectable read errors, then data recovery is simplified, but effective overprovisioning decreases and write amplification increases
Solution Approach 1:
The system performs preliminary programming adjustments to prevent read errors before they occur, eliminating the need to mark blocks as bad and preserving effective overprovisioning, which in turn maintains high write performance and reduces write amplification.
Solution Approach 2:
By using feedback from read error analysis to adjust program verify levels, the system prevents read errors that would otherwise lead to block marking, thereby preserving storage capacity and maintaining optimal write performance.
4Reliability
If program verify levels are adjusted based on prior read errors, then uncorrectable bit error rate is reduced, but device complexity increases
Solution Approach 1:
The system changes programming parameters (program verify levels and voltages) based on analyzed read error patterns. This parameter adjustment approach is straightforward and implementable, improving reliability without requiring complex architectural changes to the storage device.
Data Source
AI summary
In some implementations, a storage device may perform a read operation using a first read voltage. The storage device may identify a read error associated with the read operation. The storage device may identify a difference between the first read voltage and a second read voltage, the second read voltage associated with a reduced raw bit error rate (RBER). The storage device may perform a write operation using a program verify voltage that is based on the difference.