Flash Memory Data Structure for Universal File System Support
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional flash file systems for non-volatile memory devices require knowledge of the erasable unit size, limiting flexibility and increasing manufacturing time and costs due to the need for multiple data images and complex software maintenance when dealing with different memory types and sizes.
Innovation Solution
A data structure that allows for the combination of data units into a single erasable unit with a data unit header for storing status information, enabling the use of a single file system image across various erasable unit sizes, including the ability to select a single data unit header for status updates during operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional flash file systems are used that require knowledge of erasable unit size, then the file system can manage write and erase operations, but the manufacturing flexibility is limited and multiple data images are needed for different memory types
Solution Approach 1:
The patent creates a universal data structure that can be used across different memory types and erasable unit sizes. By defining a standardized data structure with specific fields and organization, the same file system image can be used for various memory configurations, eliminating the need for multiple specialized data images and reducing software maintenance complexity.
Solution Approach 2:
The patent changes the approach from hardcoding erasable unit size knowledge into the file system to using a parameterized data structure where the erasable unit size is determined by the actual memory device characteristics. This allows the file system to adapt to different parameter values (erasable unit sizes) without requiring fundamental changes to the file system structure.
2Reliability
If multiple data images are created for different erasable unit sizes, then each memory type can be properly supported, but the manufacturing time and costs increase
Solution Approach 1:
By creating a universal data structure that works across different memory types and erasable unit sizes, the patent enables a single data image to serve multiple memory configurations. This eliminates the need to create and maintain multiple specialized data images, thereby reducing manufacturing time and increasing throughput while still ensuring proper support for each memory type.
Solution Approach 2:
The patent prepares a standardized data structure in advance that is designed to be adaptable to different memory configurations. This preliminary preparation of a universal structure allows for faster manufacturing since the same pre-prepared data image can be used across different memory types without requiring on-the-fly conversion or customization.
3Productivity
If erasable units are large in size, then fewer erase operations are needed, but the writing speed for small data items deteriorates
Solution Approach 1:
The patent segments the erasable unit into multiple data units, where each data unit can be independently managed. This segmentation allows the file system to write small data items to individual data units without requiring erase operations on the entire large erasable unit, thereby maintaining fast write speeds for small data while still benefiting from the larger erasable unit structure for erase efficiency.
Solution Approach 2:
The patent applies different management strategies to different parts of the erasable unit. Individual data units within the erasable unit can be managed with fine-grained control for small writes, while the erasable unit as a whole provides the efficiency benefits for erase operations. This local differentiation of quality allows optimization for both small writes and erase efficiency.
4Productivity
If data is written to unwritten portions first before erasing, then write operations are optimized, but the file system depends strongly on internal memory organization
Solution Approach 1:
The patent creates a universal data structure that abstracts the file system from specific memory hardware organizations. By defining a standardized structure with clear field definitions and organization, the same write optimization strategy can be applied across different memory types and erasable unit sizes without requiring file system changes, thereby maintaining both write speed optimization and hardware independence.
Data Source
AI summary
A method is provided for storing data on memory devices comprising a plurality of erasable units, wherein the size of said erasable units is an integer multiple of a first integer value, comprising providing a data structure comprising a plurality of data units each including a data unit header, wherein the size of said data units is equal to said first integer value, a plurality of data items and corresponding data item headers within each data unit, associating at least one data unit to each erasable unit, storing said data in said data items and storing data item status information in the corresponding data item headers, and storing data unit status information in said data unit headers.