Nonvolatile Memory Defect Avoidance via Distributed Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In nonvolatile storage media, such as NAND flash memory, defective blocks cannot be efficiently managed due to large defect information tables that cannot be fully expanded into volatile storage, leading to overhead during data writing and reading operations, especially in storage media with higher defect rates.
Innovation Solution
A memory system with multiple nonvolatile memories and a controller that includes both data storage areas and defect information storage areas, where defect information from one nonvolatile memory is stored in another, allowing parallel access and overlapping read times for data and defect information, thereby enabling effective defect avoidance without performance degradation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the defect information table is entirely expanded into a volatile storage medium at start-up, then defect avoidance control can be performed efficiently, but the system cannot handle storage media with higher defect rates that require more granular defect information
Solution Approach 1:
The defect information table is divided into multiple segments stored in different nonvolatile memory devices. Each nonvolatile memory device stores a portion of the defect information table, allowing the system to handle large defect information tables that cannot be entirely expanded into volatile storage. This segmentation enables the system to manage storage media with higher defect rates by distributing the defect information across multiple devices.
Solution Approach 2:
The patent transitions from a single-dimension approach (one defect information table in one location) to a multi-dimensional distributed storage approach. Defect information is distributed across multiple nonvolatile memory devices, and the system selects the optimal device for data storage based on defect information from other devices. This dimensional change allows efficient handling of large-scale defect information without requiring complete expansion into volatile storage.
2Adaptability or versatility
If the defect information table is stored in nonvolatile storage medium and referred to during data operations, then the system can handle high defect rate media, but reading defect information creates overhead and degrades performance
Solution Approach 1:
The system performs preliminary actions by expanding the required portion of the defect information table into the volatile storage medium of the selected nonvolatile memory device before data operations begin. This preliminary expansion ensures that defect information is readily available in fast volatile storage during data read/write operations, eliminating the overhead of repeatedly accessing defect information from nonvolatile storage during normal operations.
Solution Approach 2:
The volatile storage medium acts as an intermediary between the nonvolatile memory devices and the data processing operations. Instead of directly accessing defect information from nonvolatile storage during data operations, the system uses volatile storage as a buffer to hold the required defect information, thereby mediating the access pattern and reducing the performance overhead associated with nonvolatile storage access.
3Measurement precision
If more granular defect avoidance control in units of bits is implemented, then defect avoidance precision is improved, but the defect information table size increases and cannot be expanded into volatile storage
Solution Approach 1:
The system applies local quality by expanding only the specific segments of the defect information table that are relevant to the selected nonvolatile memory device into volatile storage. Instead of expanding the entire defect information table, the system selectively loads only the portions needed for current operations, thereby maintaining high defect avoidance precision while managing the memory footprint efficiently.
Solution Approach 2:
The system performs partial action by loading only the necessary portion of the defect information table into volatile storage rather than the entire table. This partial expansion approach provides sufficient defect avoidance capability for the current operational context without requiring the excessive resource allocation that would be needed to load complete granular defect information for all possible scenarios.
Data Source
AI summary
According to an embodiment, a memory system includes multiple nonvolatile memories to/from each of which data can be written/read independently of one another; and a controller configured to control writing of data to and reading of data from the nonvolatile memories. Each of the nonvolatile memories includes a data storage including a normal data storage area for storing the data and a redundant data storage area for writing the data avoiding defect positions in the normal data storage area; and a defect information storage configured to store defect information indicating information on a defect of the data storage included in another nonvolatile memory different from the present nonvolatile memory.