File System Transaction Management Using Temporary Write Area
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The conventional shadow page technique for file management systems faces issues with large page tables, fragmentation, high commitment costs, and degraded performance, especially in large-scale databases, which affect disk access and data transfer rates, and prevents simultaneous backup and update processing.
Innovation Solution
A file system with a transaction management function that uses a temporary write area for updating files, a pair of map tables to record updated pages, and a real update mechanism to write back data to the original page position, allowing for efficient data management and reduced map size, thereby minimizing fragmentation and improving performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the shadow page technique is used to guarantee atomicity, then reliability is improved, but device complexity increases due to large page tables and multiple bit maps
Solution Approach 1:
The patent extracts the atomicity guarantee mechanism from the complex shadow page technique by using only a single bit map instead of multiple bit maps and large page tables. The bit map records which pages have been updated but not yet committed, providing sufficient information for atomicity without the overhead of comprehensive page table management.
Solution Approach 2:
The patent applies local quality by making the bit map size local to the actual updated pages rather than covering the entire database. The bit map only needs to track pages that are currently being modified, reducing the overall complexity while maintaining reliability for the active transaction set.
2Stability of the object's composition
If the shadow page technique is used to prevent fragmentation, then stability of composition is improved, but productivity decreases due to high commitment costs
Solution Approach 1:
The patent applies partial action by performing commitment operations only for the specific pages that have been updated and marked in the bit map, rather than committing the entire database or large portions of it. This selective approach reduces commitment overhead and improves productivity while maintaining data integrity.
Solution Approach 2:
The patent changes the parameter of commitment scope from global (entire database) to local (only updated pages). By modifying this parameter, the system achieves faster commitment times and improved productivity without sacrificing the stability of data composition.
3Productivity
If logging is used to improve productivity, then speed of recovery is improved, but loss of time increases due to roll forward recovery requirements
Solution Approach 1:
The patent applies preliminary action by pre-marking updated pages in the bit map before actual commit operations. This allows the system to quickly identify and recover only the necessary pages after a failure, rather than performing time-consuming roll forward recovery of the entire database, thus reducing loss of time.
Solution Approach 2:
The patent extracts the essential recovery information into a compact bit map structure, separating the critical recovery data from the full logging mechanism. This extraction enables faster recovery by focusing only on the minimal information needed to restore updated pages, improving both productivity and reducing recovery time.
4Reliability
If a large database is managed with shadow page technique, then reliability is improved, but device complexity increases due to enormous page table requirements
Solution Approach 1:
The patent extracts the essential tracking function from large-scale page table management by using a compact bit map that only monitors updated pages. This extraction reduces memory requirements and device complexity while maintaining data integrity for large databases, as the bit map scales with the number of updated pages rather than the total database size.
Solution Approach 2:
The patent changes the scaling parameter from total database size to number of updated pages. This parameter transformation allows the system to manage large databases with limited memory resources, as the bit map size depends on active transaction scope rather than the enormous size of the complete database.
Data Source
AI summary
The storage area of a storage unit includes a file area and temporary write area. A pair of map tables are allocated in the storage area. An update processing unit executes update of a page in a file stored in a file area by writing updated data to an unused page acquired from a temporary write area. A commit module alternately writes a list of effective pages in the temporary write area to the pair of map tables whenever each of transactions is committed. A checkpoint processing unit writes back the updated data of each effective page in the temporary write area to the corresponding original page position in the file area.