File Management via Discontiguous Fileset Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional file management systems are inefficient when generating incremental data snapshots for large filesets, such as a root filesystem, as they require scanning all files, leading to wasteful processing time due to many unchanged files between snapshots.
Innovation Solution
Classifying files into discontiguous filesets based on factors like frequency of change or purpose, allowing for differential processing and snapshot frequencies, rather than relying solely on hierarchical location, to optimize data management operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional file management scans all files in a hierarchical directory structure to generate incremental data snapshots, then complete file coverage is achieved, but processing time increases linearly with the number of files
Solution Approach 1:
The patent segments the fileset into multiple categories based on change frequency (frequently changed, occasionally changed, rarely changed files). This segmentation allows the system to apply different processing strategies to different segments, scanning only frequently changed files at full depth while using sampling or skipping techniques for less dynamic segments, thereby reducing overall processing time while maintaining reliable coverage of all files.
2Measurement precision
If the scan phase obtains metadata for every file to detect changes, then accurate change detection is achieved, but the time taken becomes directly proportional to the total number of files
Solution Approach 1:
The patent applies local quality by using different metadata scanning strategies for different file categories. Frequently changed files receive full metadata scanning with high precision change detection, while rarely changed files use sampling or lightweight metadata checks. This localized approach maintains high measurement precision for critical files while improving overall productivity through reduced scanning of stable files.
Solution Approach 2:
The patent implements partial action by performing complete metadata scanning only on frequently changed files rather than all files. For occasionally and rarely changed files, the system uses partial scanning or sampling techniques, accepting that some changes in these categories may be detected with lower precision but gaining significant productivity improvements overall.
3Reliability
If incremental snapshots are generated for all files in a large fileset, then comprehensive data protection is provided, but resources are wasted on files that remain unchanged between snapshots
Solution Approach 1:
The patent introduces dynamics by adjusting the snapshot generation strategy based on file change characteristics. Frequently changed files receive full incremental snapshot processing to ensure comprehensive data protection, while rarely changed files use reduced processing or skipping mechanisms. This dynamic approach maintains reliable data protection coverage for critical files while reducing computational resource waste on stable files.
4Stability of the object's composition
If traditional hierarchical file grouping is used, then location-based organization is maintained, but files with similar change patterns scattered across directories are processed together inefficiently
Solution Approach 1:
The patent adds another dimension to file organization by creating a categorical layer based on change frequency that operates alongside the traditional hierarchical directory structure. Files are grouped into categories (frequently changed, occasionally changed, rarely changed) regardless of their directory location, allowing the system to process files with similar change patterns together even when scattered across different directories, thereby improving processing efficiency without disrupting directory structure integrity.
Data Source
AI summary
According to various embodiments, with respect to a target set of files being managed (e.g., protected by data snapshots), each file in the target set of files is classified into one of two or more filesets (discontiguous filesets), where each of these filesets comprises one or more files that are related to each other by one or more factors, such as frequency of file change or purpose of existence (e.g., used by a software application). Once classified, files within the target set of files can be uniquely processed by a data management operation (e.g., incremental data snapshot process) according to their association to a discontiguous fileset.