Bad Column Replacement in Non-Volatile Storage via Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for replacing bad columns in non-volatile storage systems result in discontinuous data access, increased software and hardware costs, and performance degradation due to the need for large mapping tables and complex operations.
Innovation Solution
A method that allows continuous data access by bypassing bad columns and shifting data from adjacent good columns, reducing the need for extensive mapping tables and eliminating address switching, thereby simplifying operations and reducing costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional bad column replacement method is used with mapping tables, then bad columns can be replaced by good spare columns, but data access becomes discontinuous and performance degrades
Solution Approach 1:
The patent extracts and removes bad columns from the data area entirely, replacing them with good spare columns. By completely removing the problematic bad columns rather than mapping around them, the system achieves continuous data access without the performance degradation caused by address switching and mapping table lookups.
Solution Approach 2:
The patent performs preliminary identification and replacement of bad columns before normal data access operations begin. During initialization, bad columns are detected and replaced with good spare columns from the spare area, so that during normal operation, only continuous access to good columns is required without any runtime address switching.
2Reliability
If conventional bad column replacement method is used with mapping tables, then bad columns can be replaced, but the system requires larger mapping tables and complex operations
Solution Approach 1:
The patent eliminates the need for complex mapping tables by completely removing bad columns from the addressable space and using simple sequential addressing on the remaining good columns. This extraction approach replaces the need for large 2^n mapping tables with straightforward continuous memory access.
Solution Approach 2:
The patent copies data from good spare columns to replace bad columns in the data area. During initialization, data from spare columns is copied to fill the positions of bad columns, creating a simplified continuous address space that requires no complex mapping during operation.
3Reliability
If address switching is performed for bad column replacement, then bad columns can be accessed through good columns, but operations become complex and hardware cost increases
Solution Approach 1:
The patent removes bad columns from the addressable space entirely during initialization, replacing them with good spare columns. This eliminates the need for runtime address switching and complex access operations, as the system simply accesses continuous sequences of good columns without any address translation or switching required.
Data Source
AI summary
A method for replacing the address of some bad bytes (bad columns) of the data area and the spare area to the good address of bytes (good columns) in non-volatile storage system is disclosed. The steps of the method are: waiting for a command from a host; judging if there is still some data to be processed; if no, go back to the previous step; if yes, go to next step; judging if a bad column is used; if no, process data access and go back to the step of judging if there is still some data to be processed; and if yes, process data accessing as original operation and increase the address by one.