Data Folding in 3D NAND Memory for Defective String Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In three-dimensional NAND flash memory, data management is inefficient due to the need to copy and fold data from Single Level Cell (SLC) to Multi Level Cell (MLC) blocks, which can be affected by defective NAND strings, leading to data scrambling and interleaving challenges that impact storage density and write performance.
Innovation Solution
A method is implemented to copy data in fold-sets from SLC blocks to MLC blocks, where each fold-set is stored exclusively on individual separately-selectable sets of NAND strings, with data scrambling using a repetitive pattern to avoid pattern-related issues, and a system that interleaves data copying and uses a bad string record to manage defective strings, ensuring data is not stored on faulty sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is copied from SLC to MLC blocks in three-dimensional NAND flash memory, then storage density is improved, but write performance deteriorates due to the copying process
Solution Approach 1:
The patent applies preliminary action by pre-processing data into fold-sets with specific structures (interleaved patterns, scrambling) before copying to MLC blocks. This preparation work is done in advance to minimize the complexity and time of the actual copying operation, thereby improving write performance while maintaining high storage density.
Solution Approach 2:
The patent segments data into fold-sets that are copied to specific MLC blocks in a structured manner. By dividing data into manageable fold-sets and organizing them across multiple MLC blocks, the system achieves high storage density while the segmented structure enables parallel processing that maintains write performance.
2Quantity of substance
If data is folded into MLC blocks with multiple word lines, then storage density is improved, but data management complexity increases
Solution Approach 1:
The patent segments data into fold-sets that map to specific MLC blocks and word lines. This segmentation creates a structured organization where each fold-set is associated with particular physical locations, making data management more systematic and less complex despite the multi-word-line structure of MLC blocks.
Solution Approach 2:
The patent introduces fold-sets as an intermediary structure between logical data and physical storage locations in MLC blocks. This intermediary layer simplifies data management by providing a structured mapping mechanism that handles the complexity of multi-word-line organization without exposing it to higher-level data management functions.
3Ease of manufacture
If defective NAND strings are present in MLC blocks, then manufacturing yield is improved, but reliability deteriorates due to potential data corruption
Solution Approach 1:
The patent extracts or isolates defective NAND strings from the data storage process. By identifying and excluding bad strings from the fold-set copying process, the system maintains manufacturing yield (by not requiring complete string perfection) while ensuring data integrity (by preventing data from being stored on defective strings).
Solution Approach 2:
The patent converts the presence of defective strings (a harmful factor) into a benefit by using them as opportunities to demonstrate robust data management. The system acknowledges defective strings exist and builds mechanisms to handle them, ultimately improving reliability through proven fault tolerance while maintaining manufacturing yield.
4Reliability
If data scrambling is applied during folding, then data security and pattern avoidance are improved, but processing time increases
Solution Approach 1:
The patent applies scrambling as a preliminary action during the data folding process itself, rather than as a separate post-processing step. By integrating scrambling into the fold-set creation and copying operation, the system achieves data security without significant additional processing time, as the scrambling is performed concurrently with the structured data organization.
Data Source
AI summary
Data that is initially stored in Single Level Cell (SLC) blocks is subsequently copied (folded) to a Multi Level Cell (MLC) block where the data is stored in MLC format, the data copied in a minimum unit of a fold-set, the MLC block including a plurality of separately-selectable sets of NAND strings, data of an individual fold-set copied exclusively to two or more word lines of an individual separately-selectable set of NAND strings in the MLC block.