3D Nonvolatile Memory Redundancy Calculation via Location Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Three-dimensional nonvolatile memory systems face challenges in recovering data that becomes uncorrectable due to failure modes related to the physical structure, where Error Correction Codes (ECC) may not be sufficient to correct large numbers of bad bits, leading to uncorrectable data (UECC) issues.
Innovation Solution
The method involves calculating redundancy data using an Exclusive OR (XOR) operation on selected portions of data stored in specific locations within the memory array, ensuring that no two portions are assigned to connected or neighboring word lines or blocks, allowing any portion to be reproduced from the redundancy data and other portions, even if ECC fails.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If Error Correction Codes (ECC) are used to correct data errors, then data reliability is improved, but ECC may not be sufficient to correct large numbers of bad bits resulting in uncorrectable data (UECC) issues
Solution Approach 1:
The patent performs preliminary actions by selecting specific physical locations for data portions before storage, ensuring they are not subject to the same failure modes. This preventive location selection, combined with preliminary XOR redundancy calculation, prepares the system to recover from failures before they occur, addressing the limitation of ECC in handling large numbers of bad bits.
Solution Approach 2:
The patent introduces an intermediary mechanism (XOR redundancy data) that mediates between the stored data portions and the recovery process. This redundancy data serves as a mediator that enables reconstruction of lost or corrupted data portions, complementing ECC and providing an additional layer of protection against uncorrectable errors.
2Reliability
If redundancy data is calculated using XOR operation on selected data portions, then data recovery capability is improved, but device complexity increases
Solution Approach 1:
The patent segments data into multiple portions and stores them in specific physical locations within the memory array. This segmentation allows the system to apply XOR operations on selected portions to generate redundancy data, enabling recovery without requiring complex global error correction mechanisms across the entire data set.
Solution Approach 2:
The patent changes the parameter of redundancy generation from complex ECC algorithms to simple XOR operations. This parameter change simplifies the computational complexity while maintaining effective data recovery capability, as XOR operations are computationally much less intensive than traditional ECC methods.
3Reliability
If data portions are selected from non-neighboring locations to reduce simultaneous failure risk, then data reliability is improved, but manufacturing precision requirements increase
Solution Approach 1:
The patent applies local quality by assigning different data portions to specific physical locations with distinct failure mode characteristics. By ensuring that selected portions are not subject to the same failure modes (e.g., not on the same word line or subject to the same block select circuit), the system creates local diversity that improves reliability without requiring extreme manufacturing precision.
Data Source
AI summary
Portions of data stored in a three dimensional memory array are selected based on their locations for calculation of redundancy data. Locations are selected so that no two portions in a set of portions for a given calculation are likely to become uncorrectable at the same time. Selected portions may be separated by at least one word line and separated by at least one string in a block.