Memory Die Failure Detection Using Consecutive Blockstripe Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting die failure in memory sub-systems are inadequate as they fail to accurately track block failures, leading to ineffective handling of program failures, especially in high-capacity memory devices, resulting in potential dead loops and drive hangs.
Innovation Solution
Implement a die failure detection method using channel CPUs to monitor program failures across multiple blockstripes, incrementing a program failure count for each die experiencing consecutive failures, and initiating a retirement process when the count reaches a threshold.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional methods are used to detect die failure, then the system structure remains simple, but the reliability of die failure detection deteriorates due to inaccurate block failure tracking
Solution Approach 1:
The patent segments the memory device into multiple blockstripes, each associated with a specific die. By tracking program failures at the blockstripe level rather than individual blocks, the system achieves accurate die failure detection while reducing tracking complexity. Each blockstripe serves as a representative unit that, when failed, indicates die-level failure without requiring monitoring of every individual block within the die.
2Measurement precision
If all blocks are monitored for failure, then measurement precision improves, but the loss of time increases due to extensive monitoring requirements
Solution Approach 1:
Instead of monitoring every individual block within a die, the patent implements partial monitoring by selecting and monitoring only one representative block per blockstripe. This partial action approach maintains sufficient measurement precision for die failure detection while dramatically reducing the time and computational resources required compared to exhaustive block-level monitoring.
3Speed
If program failure count threshold is set low, then the speed of die failure detection improves, but the loss of substance increases due to retirement of good dies
Solution Approach 1:
The patent implements a preliminary verification step before retiring a die by checking whether the program failure is isolated to a single blockstripe or persists across multiple consecutive blockstripes. This preliminary action of verifying consecutive failures across multiple blockstripes prevents premature retirement of good dies while maintaining fast detection speed through the use of a threshold-based counting mechanism.
Data Source
AI summary
A plurality of commands comprising data to be written to a plurality of blockstripes, each blockstripe comprising at least one block from each of a plurality of memory devices, can be received from a high-speed processing device. Based on determining a first memory device associated with a first blockstripe experiences program failure in response to execution of a first command, whether a second memory device associated with a second blockstripe experiences program failure in response to execution of a second command, can be determined. Based on detecting consecutive program failure on the memory device, whether a program failure count associated with the memory device satisfies a program failure threshold can be determined. Based on determining the program failure count associated with the memory device satisfies the program failure threshold, retirement of the memory device can be initiated.