Memory Controller Wear-Leveling via Random Data Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Nonvolatile memory systems, particularly Multi-Level Cell (MLC) flash memories, experience threshold voltage shifts due to frequent programming and erasing, leading to data errors and reduced memory accuracy over time.
Innovation Solution
A nonvolatile memory system that logically combines data with a random pattern to generate encoded data, which is then programmed, using a memory controller that generates the random pattern based on a reference value such as an erase count value, to distribute program stresses evenly across memory cells, thereby maintaining uniform threshold voltage levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data is frequently programmed and erased in nonvolatile memory, then memory capacity and functionality are improved, but threshold voltage levels shift causing data errors
Solution Approach 1:
The patent applies preliminary action by encoding data with a random pattern before programming it into the nonvolatile memory. This pre-processing step modifies the data pattern to reduce stress concentration on specific memory cells, thereby preventing threshold voltage shifts before they occur. The random pattern is generated based on wear-leveling information that predicts future stress distribution, allowing proactive protection against data errors.
Solution Approach 2:
The patent changes the parameter of data representation by transforming original data into encoded data through logical combination with a random pattern. This parameter transformation alters the stress distribution characteristics during programming operations, distributing wear more uniformly across memory cells and maintaining threshold voltage stability despite frequent write/erase cycles.
2Stability of the object's composition
If wear-leveling information is used to encode data, then threshold voltage stability is improved, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary element - a random pattern generator - that mediates between the wear-leveling information and the data encoding process. This intermediary component simplifies the overall system by providing a straightforward mechanism to generate random patterns based on wear-leveling metrics, avoiding the need for complex adaptive encoding algorithms while still achieving threshold voltage stability.
Data Source
AI summary
A nonvolatile memory system is operated by providing data to be written to a nonvolatile memory, logically combining the data to be written to the nonvolatile memory with a random pattern to generate encoded data; and programming the encoded data in the nonvolatile memory.