Non-volatile Memory Failure Prediction via Fermi-Dirac Extrapolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In non-volatile memory systems, high error rates can occur during data programming, storage, and reading, which are often not correctable by Error Correction Code (ECC) circuits, requiring significant time and resources for statistical analysis using large sample populations to predict low-probability events.
Innovation Solution
The implementation of an analytic function, such as the Fermi-Dirac distribution, allows for the extrapolation of probabilities from a small sample size to model a wide range of events, including those with low probabilities, enabling rapid and cost-effective analysis and memory management decisions within the non-volatile memory system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If statistical analysis using large sample populations is performed to predict low-probability events, then prediction accuracy is improved, but time consumption and resource usage increase significantly
Solution Approach 1:
The patent transforms the statistical analysis approach by changing the parameter from using large sample populations to using a small sample population combined with an analytic function (Fermi-Dirac distribution). This parameter change enables accurate prediction of low-probability events without requiring extensive data collection and analysis time, thus resolving the contradiction between prediction accuracy and time consumption.
2Measurement precision
If statistical analysis using large sample populations is performed to predict low-probability events, then prediction accuracy is improved, but computational resources and cost increase significantly
Solution Approach 1:
The patent applies parameter changes by replacing the computationally intensive large-sample statistical analysis with a small-sample approach enhanced by the Fermi-Dirac analytic function. This reduces the computational resources and cost required while maintaining prediction accuracy for low-probability events, thereby resolving the contradiction between measurement precision and device complexity.
3Productivity
If small sample size is used for probability estimation, then analysis speed and cost-effectiveness are improved, but prediction accuracy for low-probability events deteriorates
Solution Approach 1:
The patent substitutes the mechanical process of collecting and analyzing large amounts of statistical data with a mathematical model (Fermi-Dirac distribution) that can directly compute probabilities. This substitution allows the system to achieve both high productivity (fast analysis) and high measurement precision (accurate predictions) by using a small sample size, thus resolving the contradiction between productivity and measurement precision.
4Reliability
If ECC circuits are used to detect and correct errors, then data reliability is improved, but system complexity and resource consumption increase
Solution Approach 1:
The patent applies preliminary action by using the Fermi-Dirac distribution model to predict the probability of uncorrectable errors before they occur. This allows the system to take preventive measures (such as wear leveling, garbage collection, and bad block identification) proactively, improving data reliability while avoiding the need for overly complex error correction systems, thus resolving the contradiction between reliability and device complexity.
Data Source
AI summary
A non-volatile storage apparatus includes a set of non-volatile memory cells and one or more control circuits in communication with the set of non-volatile memory cells, the one or more control circuits are configured to collect failure bit counts (FBCs) for data read from the set of non-volatile memory cells, obtain one or more metrics of a cumulative distribution of the FBCs, calculate an indicator from the one or more metrics of the cumulative distribution of the FBCs and a target FBC, obtain a probability for the target FBC from the indicator, and manage at least one of: garbage collection, wear leveling, and read threshold voltage adjustment of the set of non-volatile memory cells according to the probability for the target FBC.