Cross-Environment Event Correlation via ML Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multi-domain computing environments, events or changes in one domain often go uncorrelated with their impacts on other domains, leading to prolonged debugging and difficulty in understanding risks, as existing systems lack effective cross-environment correlation mechanisms.
Innovation Solution
A computer-implemented method and system for cross-environment event correlation using machine learning to determine correlated events across domains, constructing correlation graphs, and providing logical reasoning descriptions to explain issue causes, facilitating diagnosis and corrective actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If events in different domains are reviewed independently, then domain-specific analysis is simplified, but cross-domain issue detection becomes difficult and debugging time increases
Solution Approach 1:
The patent combines independent domain event reviews with cross-domain correlation analysis by integrating machine learning models that process events across multiple domains simultaneously. The system merges domain-specific event data into a unified correlation framework that identifies cross-domain relationships without losing the benefits of specialized domain analysis.
Solution Approach 2:
The patent introduces a machine learning-based event correlation system as an intermediary layer between independent domain event reviews and cross-domain issue detection. This intermediary automatically correlates events across domains using trained models, reducing the time required to detect cross-domain issues while preserving domain-specific analysis capabilities.
2Measurement precision
If machine learning techniques are used to determine correlated events across domains, then discovery of hidden correlations improves, but system complexity and computational resources increase
Solution Approach 1:
The patent segments the machine learning system into specialized components including association rule learning algorithms, clustering algorithms, and supervised learning techniques. Each component handles specific aspects of event correlation, making the overall complex system manageable through modular design while maintaining high correlation detection accuracy.
Solution Approach 2:
The patent applies preliminary action by pre-training machine learning models using historical event data and synthetic data before deployment. This preliminary training phase establishes the correlation patterns and relationships that the system will use for real-time event analysis, reducing the complexity of real-time processing while maintaining high detection accuracy.
3Measurement precision
If unsupervised learning techniques are used to discover correlations, then detection of previously unknown event relationships improves, but training time and computational resources increase
Solution Approach 1:
The patent applies partial action by using unsupervised learning techniques selectively for discovering new correlation patterns rather than applying them to all event analysis tasks. The system uses unsupervised learning specifically for exploring unknown relationships while relying on other methods for routine correlation detection, balancing discovery capability with training efficiency.
Solution Approach 2:
The patent uses synthetic data as a copy of real event data to train machine learning models. This synthetic data replication allows the system to practice correlation discovery without the computational cost and time requirements of processing actual large-scale event data during training, reducing training time while maintaining correlation discovery capabilities.
4Loss of information
If correlation graphs are generated to trace issues and group events, then issue tracking and understanding improve, but data processing complexity and time increase
Solution Approach 1:
The patent applies preliminary action by pre-processing and normalizing event data before generating correlation graphs. The system prepares event data in advance by standardizing formats, extracting key attributes, and organizing data structures that facilitate efficient graph generation. This preliminary preparation reduces the time required for actual correlation graph creation while maintaining comprehensive issue context information.
Data Source
AI summary
A computer-implemented method of cross-environment event correlation includes determining one or more correlated events about an issue across a plurality of domains. A knowledge data is extracted from the issue determined from the one or more correlated events is performed. A correlation graph is generated from the extracted knowledge to trace the issue and group the correlated events into one or more event groups to represent their relationship with the issue. A logical reasoning description is constructed based on the generated correlation graph for a domain-space exploration related to how the issue in one domain affects another domain of the plurality of domains. The one or more event groups of correlated events is provided with an explanation about a cause of the issue based on the logical reasoning description.