Log Extraction Graphs for Variable Identifier Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing log extraction methods struggle to associate identifiers appropriately in information processing systems with varying names and values, especially when constructed by combining open software and existing applications, leading to difficulties in extracting logs with high correlation.
Innovation Solution
A log extraction system that generates a same type request related identifier list and calculates correlation degrees among identifier values, setting identifiers as nodes and correlation degrees as link metrics to create a related identifier graph, facilitating appropriate association of identifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If identifiers are associated based on fixed names and values, then log extraction is simple and fast, but identifier association becomes inappropriate when log names and values vary across different systems
Solution Approach 1:
The patent changes the parameters used for identifier association from fixed names and values to correlation degrees calculated from actual log data patterns. Instead of relying on predetermined identifier definitions, the system dynamically determines identifier relationships by analyzing how identifiers co-occur and correlate across different log types, allowing accurate association even when identifier names and values vary.
Solution Approach 2:
The patent replaces the mechanical system of fixed-rule-based identifier matching with a data-driven correlation analysis system. Rather than using static association rules, the system calculates correlation degrees between identifiers based on actual log patterns, substituting rigid mechanical matching with flexible statistical analysis.
2Loss of information
If all log patterns are analyzed to ensure comprehensive coverage, then log analysis completeness is improved, but analysis time and computational resources become excessive
Solution Approach 1:
The patent performs preliminary action by pre-calculating and storing correlation degrees between identifiers in a related identifier graph before actual log analysis. This preprocessing step creates a ready-to-use reference structure that enables rapid log extraction without requiring comprehensive analysis of all log patterns during the actual analysis phase.
Solution Approach 2:
The patent segments the log analysis process into two phases: (1) pre-computation of identifier correlation relationships using the related identifier graph, and (2) efficient log extraction using the pre-computed relationships. This segmentation allows comprehensive identifier relationship analysis to be performed once, while subsequent log analyses benefit from the pre-computed structure without requiring full re-analysis.
3Ease of manufacture
If identifier association uses fixed rules, then the system is simple to implement, but it cannot adapt to varying identifier names and values in different software combinations
Solution Approach 1:
The patent introduces dynamics into the identifier association system by using correlation degrees that are calculated from actual log data rather than fixed rules. The related identifier graph dynamically adapts to different identifier names and values by computing correlations based on observed patterns, allowing the system to automatically adjust to varying identifier schemes across different software combinations without requiring manual reconfiguration.
Data Source
AI summary
In a log extraction method, logs relating to requests of the same type are grasped and a same type request related identifier list for managing a correspondence relation between identifier keys for identifying identifiers included in the logs and identifier values representing values that the identifiers take in the logs is generated. In the log extraction method, correlation degrees of the identifier values among the requests of different types are calculated based on the identifier values of the same identifiers associated with the requests of the different types in the same type request related identifier list. In the log extraction method, the identifier keys are set as nodes, a relation between the identifier keys and the identifiers is set as a link, and the correlation degrees are set as link metrics and a related identifier graph obtained by grouping the identifier keys based on the link metrics is generated.