Expert System for Distributed Computing Log Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed computing environments, identifying and resolving application execution problems is cumbersome and time-consuming due to the need for manual analysis of log files across multiple resources, limiting visibility and requiring involvement from multiple operations groups, which delays solution identification.
Innovation Solution
An expert system that uses pattern recognition and self-learning techniques to analyze log files, plot events on a graph, identify patterns associated with problem conditions, and provide corrective actions, allowing for real-time detection and automated identification of solutions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual analysis of log files is performed across multiple resources, then comprehensive problem detection is achieved, but time consumption and operational complexity increase significantly
Solution Approach 1:
The patent segments the distributed computing environment into individual topological resources, each maintaining its own log files. The expert system divides the analysis task by extracting events from each resource's log files separately, then integrating them through graph plotting. This segmentation allows parallel processing of log data from multiple resources, reducing overall analysis time while maintaining comprehensive problem detection capability.
Solution Approach 2:
The patent introduces an expert system as an intermediary between raw log files and problem analysis. This intermediary automatically extracts events from log files, plots them on graphs with standardized axes (time of occurrence and topological resource), and identifies patterns. This intermediary layer eliminates manual analysis requirements, significantly reducing time consumption while preserving detection comprehensiveness.
2Measurement precision
If manual analysis of log files is performed across multiple resources, then detailed problem understanding is achieved, but operational complexity and coordination requirements increase
Solution Approach 1:
The expert system performs self-service by automatically extracting events from log files, plotting them on graphs, and identifying patterns without human intervention. The system autonomously correlates events across different topological resources, determines problem conditions, and identifies solutions by comparing pre-problem and post-problem configurations. This automation eliminates the need for manual coordination among operations groups, reducing operational complexity while maintaining deep problem understanding.
Solution Approach 2:
The patent creates a universal expert system that handles multiple functions: event extraction from various log file formats, graph plotting with standardized axes, pattern recognition across different resource types, and solution identification. This multi-functional system replaces multiple specialized manual analysis processes, reducing operational complexity while preserving comprehensive problem understanding across the distributed environment.
3Reliability
If visibility into application execution problems is enhanced across distributed resources, then solution accuracy is improved, but system complexity and data processing requirements increase
Solution Approach 1:
The patent transforms log data into a graphical dimension with two axes: time of occurrence (x-axis) and topological resource (y-axis). Each event becomes a plot point representing a combination of error type and log type. This dimensional transformation enables visual pattern recognition across time and resources, improving solution accuracy by revealing temporal and spatial relationships that would be difficult to detect in raw log data, while the automated graphing process manages the complexity of processing distributed data.
Solution Approach 2:
The expert system implements feedback by continuously monitoring log files for new events, updating graphs in real-time, and re-evaluating patterns as new data arrives. The system compares current problem conditions against historical patterns and previously identified solutions, refining its accuracy through iterative learning. This feedback mechanism improves solution reliability by incorporating ongoing observations while managing system complexity through automated iterative processing rather than manual re-analysis.
Data Source
AI summary
An expert system extracts events associated with executing an application from log files generated by various topological resources in a distributed computing environment. The events are plotted as plot points on a time series graph. Patterns are identified in the plot points that are associated with application problems, along with the computing environment configurations both before the problem and after the problem was resolved. The difference in the configurations represents a corrective action for the application problem, and the expert system links the corrective action to the pattern. When a pattern repeats in conjunction with another application problem, the corrective action is identified as a possible solution to the new problem. A confidence level associated with the pattern/corrective action may be increased when a user accepts the corrective action and may be decreased when a user rejects the corrective action.