Storage Controller OVS Analysis for NAND Read Voltage Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage controllers for non-volatile memory devices lack efficient methods to determine and adapt operations based on the distribution type of memory regions, leading to potential reliability issues and increased input/output load.
Innovation Solution
The storage controller incorporates an on-chip valley search (OVS) management module and a type determiner to analyze OVS count data generated during normal read operations, determining the distribution type of memory regions using machine learning, and adjusting read voltages and operations accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the storage controller uses traditional read operations without distribution type analysis, then the input/output load increases and latency increases, but the device complexity remains low
Solution Approach 1:
The storage controller performs on-chip valley search (OVS) during normal read operations to pre-capture threshold voltage distribution information before actual data reading. This preliminary action allows the controller to determine memory region distribution types in advance, enabling optimized read operations that reduce I/O load and latency without requiring complex external analysis systems
Solution Approach 2:
The storage controller integrates the OVS management module and type determiner within itself, allowing the controller to autonomously analyze its own memory region distribution types using internally generated OVS count data. This self-service capability eliminates the need for external profiling systems while improving read operation efficiency through distribution-type-aware voltage optimization
2Reliability
If the storage controller implements distribution type determination using machine learning, then the reliability improves through optimized read voltages, but the device complexity increases
Solution Approach 1:
The type determiner acts as an intermediary component that bridges the OVS management module and the read operation controller. It analyzes OVS count data to determine memory region distribution types and translates these determinations into optimized read voltage selections, thereby improving reliability through a structured intermediate analysis layer without requiring complex machine learning implementations
Solution Approach 2:
The storage controller dynamically adjusts read voltage parameters based on determined distribution types. By changing voltage levels according to the specific distribution characteristics of each memory region (e.g., normal, retention, read-disturb types), the system improves read reliability while maintaining a relatively simple controller structure through parameter optimization rather than architectural complexity
3Measurement precision
If the storage controller performs on-chip valley search during normal read operations, then the measurement precision of threshold voltage distribution improves, but the use of energy increases
Solution Approach 1:
The on-chip valley search operation is merged with normal read operations, allowing the controller to capture threshold voltage distribution information using the same read infrastructure already in place. This combining approach achieves precise distribution measurement without requiring separate dedicated measurement operations, thereby limiting additional energy consumption while maintaining high measurement precision
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is a method of operating a storage controller which communicates with a non-volatile memory device. The method includes outputting a first command including a request for on-chip valley search (OVS) count data of a memory region of the non-volatile memory device to the non-volatile memory device, wherein the OVS count data includes a first count value and a second count value of a first read voltage and a third count value and a fourth count value of a second read voltage, receiving the OVS count data from the non-volatile memory device, determining a distribution type of the memory region to be a predicted distribution type, from among a plurality of distribution types, based on the OVS count data, and determining a subsequent operation, based on the predicted distribution type.