Flash Memory Storage Using Bloom Filter Digest for Space Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage systems for flash memories are inefficient in utilizing fast reading/writing characteristics and occupy large storage space due to complex maintenance operations and inefficient data management.
Innovation Solution
A data storage method using a flash storage structure with a data recording area, a log area, and a log digest area, where data is sequentially written and log information is generated with a physical address and ID, and Bloom filter data is constructed to reduce storage space and simplify maintenance, improving query efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If JFFS3 uses a B+ tree index structure with logs to delay updates, then searching efficiency is improved, but storage space is occupied and data maintenance becomes complex
Solution Approach 1:
The patent extracts the essential function of the B+ tree index by using a simplified log structure with sequential writes. Instead of maintaining a complex B+ tree, it uses a linear log area where only necessary information (physical address, ID, timestamp) is stored, reducing maintenance complexity while preserving search capability
Solution Approach 2:
The patent creates a simplified copy of the indexing mechanism. Rather than copying the entire B+ tree structure, it copies only the essential log entries needed for searching, storing minimal information in a sequential format that is easier to maintain while still enabling efficient data retrieval
2Productivity
If JFFS3 delays B+ tree updates by using logs, then writing operations are combined, but storage space increases and writing efficiency does not improve
Solution Approach 1:
The patent extracts only the essential information needed for tracking changes (physical address, ID, timestamp) and stores it in a compact sequential log format. This eliminates the need to store complete B+ tree structures or redundant data, reducing storage space while still enabling effective change tracking and writing consolidation
3Adaptability or versatility
If flash memories are packaged as block devices with intermediate mapping, then compatibility is achieved, but working efficiency decreases and fast reading/writing characteristics are not utilized
Solution Approach 1:
The patent segments the flash storage into distinct functional areas: data recording area, log area, and log digest area. This segmentation allows the system to maintain compatibility with block device interfaces while implementing optimized flash-specific operations in each area, enabling fast reading/writing through direct access to the data recording area without unnecessary intermediate mapping operations
Data Source
AI summary
A data storage processing method, a data searching method, and devices thereof are provided. The data storage processing method includes: sequentially writing data to a data recording area in a flash; generating log information according to a physical address of the data in the data recording area and an identifier (ID) of the data, and sequentially writing the log information to a log area in the flash; and constructing a Bloom filter data for the log information in the log area, and sequentially writing the Bloom filter data to a log digest area in the flash. A flash storage structure including the data recording area, the log area, and the log digest area is adopted, thereby reducing the occupied storage space of the flash. In addition, since all the areas adopt a sequential storage mode, the data maintenance is quite simple.