Network Upgrade Performance Change Detection via Rule Learner
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network upgrades often result in unintended performance changes across multiple devices, making it challenging for network operators to predict and manage these changes effectively, as traditional testing methods may not account for all possible outcomes in a larger operational environment.
Innovation Solution
A method and system that identify common attributes across different behavior changes by correlating network triggers to performance changes using a rule learner, which extracts rules from data to determine the commonality of performance changes across multiple network devices, employing clustering and machine learning algorithms to analyze trigger/change point pairs and generate logical rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional testing methods are used for network upgrades, then the upgrade process is simple and quick, but the ability to predict and manage all possible performance outcomes is insufficient
Solution Approach 1:
The patent introduces an intermediary analysis system that acts as a mediator between network upgrades and performance outcomes. This system collects performance data, identifies change points, correlates them with triggers, and extracts common attributes to predict upgrade outcomes, thereby bridging the gap between simple upgrades and reliable performance prediction.
Solution Approach 2:
The system performs preliminary actions by collecting and analyzing performance data before actual network upgrades are implemented. It identifies potential performance changes and their triggers in advance, allowing operators to predict outcomes and make informed decisions before committing to upgrade actions.
2Measurement precision
If network operators manually monitor performance changes across multiple devices, then detailed observation is possible, but the complexity and time required increase significantly
Solution Approach 1:
The system enables self-service by automatically collecting performance data, identifying change points, correlating triggers, and extracting common attributes without requiring manual intervention. The automated rule learner and analysis engine perform monitoring tasks independently, delivering precise results without consuming operator time.
Solution Approach 2:
The patent replaces manual mechanical monitoring with an automated computational system. Instead of operators manually observing and recording performance changes, the system uses algorithms to automatically detect change points, correlate triggers, and extract patterns, substituting human effort with automated processing.
3Loss of information
If comprehensive data collection across all network devices is performed, then complete performance visibility is achieved, but the data processing complexity and computational resources required increase
Solution Approach 1:
The system applies extraction by selectively identifying and extracting only the most relevant information from comprehensive performance data. It extracts change points, correlates specific triggers with performance changes, and identifies common attributes across devices, filtering out unnecessary data while maintaining complete performance visibility.
Solution Approach 2:
The patent segments the comprehensive data collection process into distinct manageable components: data collection, change point identification, trigger correlation, and common attribute extraction. This segmentation reduces processing complexity by breaking down the monolithic data analysis task into sequential, specialized processing stages.
Data Source
AI summary
A system and method identify a set of rules for determining a commonality of attributes across different behavior changes for a network. The system performs the method by receiving a set of data correlating network triggers to performance changes of one or more network devices. The set of data further includes an indication of a sign of the performance change for each of the network devices based on the triggers. The method further includes extracting a set of rules relating to a set of relationships between the triggers and the performance changes. The rules identify a commonality of the performance changes for multiple network devices based on the triggers.