Global Write Stamp Controller for Flash Memory Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Emerging storage media like flash memory and phase change memory face reliability issues due to limited write cycles and data disturbance, leading to complexity and high calculation overheads when using separate wear leveling and disturb algorithms.
Innovation Solution
A method involving a global write stamp counter to identify and move 'to-be-moved' data blocks, which includes cold and disturbed blocks, by reconfiguring their mapping relationships, thereby reducing calculation overheads and improving storage medium reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If separate wear leveling and disturb algorithms are used to process reliability problems, then the storage medium reliability is improved, but the device complexity and calculation overheads increase
Solution Approach 1:
The patent combines separate wear leveling and disturb algorithms into a unified processing framework. The controller uses a single set of data structures (physical address mapping tables and write count information) to simultaneously track and manage both wear leveling and disturbance issues, eliminating the need for separate independent processing systems.
Solution Approach 2:
The physical address mapping table and write count information serve multiple functions: they track write counts for wear leveling, identify disturbed blocks, and enable unified management of both reliability issues. This multi-functional approach reduces the overall system complexity while maintaining comprehensive reliability protection.
2Reliability
If separate wear leveling and disturb algorithms are used to process reliability problems, then the storage medium reliability is improved, but the calculation overheads increase
Solution Approach 1:
The patent merges wear leveling and disturb processing into a single algorithmic framework that operates on unified data structures. By combining these operations, the controller performs fewer separate calculations and reduces redundant processing, thereby lowering calculation overheads and energy consumption while maintaining reliability improvements.
3Reliability
If frequent inspection and movement of data blocks is performed, then the storage medium reliability is improved, but the productivity and write speed decrease
Solution Approach 1:
The controller performs preliminary identification of cold and disturbed blocks using physical address mapping tables and write count tracking, but delays the actual data movement operation. By preparing the identification in advance and executing movement only when necessary, the system maintains reliability protection while minimizing interruptions to write operations and preserving write speed.
Solution Approach 2:
The system implements periodic inspection of data blocks based on write count thresholds and disturbance detection, rather than continuous monitoring. This periodic approach ensures reliability through regular checks while avoiding excessive intervention that would slow down write operations, thus balancing reliability improvement with productivity maintenance.
Data Source
AI summary
Embodiments of this application disclose example methods for processing data blocks and controllers. One example method includes configuring, by a controller, a counter of a global write stamp where the global write stamp is used to represent a total quantity of times that the controller performs writing on a storage medium. When a write operation is performed on a physical data block in the storage medium, a global write stamp can be used at a current moment as an instant write stamp of the physical data block, and the instant write stamp can be stored. Physical data blocks can be inspected periodically. When the physical data block is inspected, a determination as to whether the physical data block is a to-be-moved data block can be made based on the stored instant write stamp.