Flash Mapping Format Using Biased CRC for UNC Marking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional non-volatile memory mapping information formats require a reserved uncorrectable flag bit, limiting the number of bits available for physical address identification and thereby reducing the storage capacity of data storage devices.
Innovation Solution
A mapping information format without any uncorrectable flag bits is introduced, allowing for a sufficient number of bits for physical address identification, and a cyclic redundancy check engine is used to encode data patterns with a biased encoding seed to generate biased cyclic redundancy check codes, enabling uncorrectable marking and error correction without returning incorrect data to the host.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If an uncorrectable flag bit is included in the mapping information format, then uncorrectable data can be marked, but the number of bits available for physical address identification is reduced
Solution Approach 1:
The patent extracts the uncorrectable data marking function from the mapping information format by introducing a separate uncorrectable marking unit that operates independently from the mapping information. This separation allows the mapping information to use all available bits for physical address identification while the uncorrectable marking unit handles error marking through a different mechanism using cyclic redundancy check codes.
Solution Approach 2:
The patent introduces a cyclic redundancy check engine as an intermediary component that generates biased CRC codes to mark uncorrectable data. This intermediary mechanism enables uncorrectable data marking without requiring additional flag bits in the mapping information format, as the marking is achieved through the CRC code generation process rather than direct flag setting.
2Quantity of substance
If mapping information format uses all bits for physical address identification, then storage capacity is increased, but uncorrectable data marking becomes difficult
Solution Approach 1:
The patent segments the data management functionality into separate components: mapping information management for address identification and an independent uncorrectable marking unit for error marking. This segmentation allows the mapping information format to be optimized for storage capacity with all bits dedicated to physical address identification, while the separate marking unit handles reliability functions through biased CRC code generation.
Solution Approach 2:
The patent changes the parameter used for uncorrectable data marking from flag bits in the mapping information to biased cyclic redundancy check codes generated by the CRC engine. This parameter change allows the system to maintain full bit utilization in mapping information for address identification while still providing robust uncorrectable data marking capability through the alternative CRC-based mechanism.
3Measurement precision
If uncorrectable marking unit is smaller than mapping information management unit, then marking precision is improved, but system complexity increases
Solution Approach 1:
The patent segments the error management system into a fine-grained uncorrectable marking unit that operates at a smaller granularity than the mapping information management unit. This segmentation enables precise marking of individual uncorrectable data units while the larger mapping information unit continues to handle broader address management, achieving high marking precision without requiring complete redesign of the entire system.
Data Source
AI summary
Mapping information management for data storage. A mapping information format without any uncorrectable flag bits (UNC bits) is shown. A controller provides a cyclic redundancy check (CRC) engine. In response to an uncorrectable marking command issued by a host, the controller operates the cyclic redundancy check engine to encode a data pattern with a biased encoding seed to generate biased cyclic redundancy check code. The controller programs the data pattern and the biased cyclic redundancy check code to the non-volatile memory. The data pattern, therefore, will not pass CRC. The uncorrectable marking command works.