Pattern Detection via ML Correlation of Configuration Metrics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Pattern and anti-pattern detection in software engineering and enterprise systems is a difficult process, as many instances may go unidentified, hindering system design, configuration, and the identification of potential issues before they arise.
Innovation Solution
A system and method for identifying, utilizing, and sharing patterns and anti-patterns through the evaluation of performance metric values over time, associating changes in configuration settings with performance metrics, and providing visualization tools for IT administrators to monitor system health and risk, with mechanisms to implement identified patterns and anti-patterns to improve system performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pattern and anti-pattern detection is performed manually or through traditional methods, then detection accuracy may be improved, but detection efficiency and coverage are significantly reduced
Solution Approach 1:
The patent replaces manual pattern detection (mechanical human analysis) with automated machine learning models and algorithms. The system uses trained ML models to automatically analyze configuration data, performance metrics, and system logs to identify patterns and anti-patterns, thereby maintaining detection accuracy while dramatically improving efficiency and coverage.
Solution Approach 2:
The patent introduces an intermediary layer of machine learning models and analysis algorithms that mediate between raw system data and pattern identification. This intermediary layer processes configuration data, performance metrics, and logs through trained models to detect patterns, enabling automated high-efficiency detection without sacrificing the precision that comes from sophisticated analysis.
2Adaptability or versatility
If comprehensive pattern detection is implemented across the entire system, then detection coverage is improved, but system complexity and computational resources increase
Solution Approach 1:
The patent segments the pattern detection system into modular components: data collection modules that gather configuration and performance data, machine learning model modules that analyze specific pattern types, and reporting modules that present findings. This segmentation enables comprehensive system-wide detection coverage while managing complexity through modular, independently deployable components.
Solution Approach 2:
The patent implements preliminary action by pre-training machine learning models with historical configuration data and known patterns before deployment. These pre-trained models are then applied to new systems, enabling comprehensive pattern detection coverage without requiring complex real-time analysis of every possible pattern type, thus reducing operational complexity.
3Loss of time
If frequent performance metric collection is performed to identify patterns, then pattern identification timeliness is improved, but system overhead and resource consumption increase
Solution Approach 1:
The patent implements periodic action by collecting performance metrics at optimized intervals rather than continuously. The system determines appropriate sampling frequencies based on the specific pattern types being detected and system characteristics, achieving timely pattern identification while minimizing the overhead and resource consumption associated with excessive data collection.
Solution Approach 2:
The patent applies partial action by collecting only the specific performance metrics and configuration data relevant to the patterns being detected, rather than gathering all possible system data. This selective data collection approach maintains timely pattern identification while reducing the overhead and computational resources required for processing unnecessary information.
Data Source
AI summary
Performance of a computer system is measured based, at least in part, on a performance metric. In response to determining that the computer system is experiencing a performance issue based on measuring the performance, the performance metric is matched with an anti-pattern to identify a performance issue, wherein the anti-pattern defines an incorrect solution to a defined problem occurring in the computer system. Also, a pattern repository is queried to identify a pattern that defines a correct solution to the defined problem based, at least in part, on the match between the performance metric and the anti-pattern. In response to identifying the pattern, implementing the pattern in the computer system to improve the performance.