Predictive Network Service Assurance for Critical Alarm Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network service assurance systems face challenges in efficiently identifying critical issues due to the complexity and cost associated with determining legitimate problems in network performance, leading to high development and maintenance costs.
Innovation Solution
A predictive model is trained using network monitoring data to automate the identification of critical issues and prioritize network monitoring results, utilizing artificial intelligence and machine learning to quickly and automatically determine the severity of alarms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional network service assurance systems use multiple algorithms, rules, and processes to identify network problems, then measurement precision is improved, but device complexity increases and development costs increase
Solution Approach 1:
The patent extracts and isolates the core problem identification function into a dedicated predictive analytics engine that uses machine learning models. This separates the complex analytical functions from the traditional monitoring infrastructure, allowing precise problem detection while simplifying the overall system architecture by consolidating multiple algorithms into unified predictive models.
Solution Approach 2:
The patent replaces traditional mechanical rule-based systems with intelligent predictive analytics using machine learning. Instead of relying on numerous predefined algorithms and rules, the system uses trained models that automatically learn patterns from historical data, reducing system complexity while maintaining or improving problem identification accuracy.
2Reliability
If traditional network service assurance systems implement multiple algorithms and processes, then problem detection capability is improved, but development cost increases and maintenance cost increases
Solution Approach 1:
The patent implements self-service through automated predictive analytics that continuously learn from network data without requiring manual rule updates. The machine learning models automatically adapt to changing network conditions, eliminating the need for continuous manual development and maintenance of multiple algorithms, thereby reducing both development and maintenance costs while maintaining reliable problem detection.
Solution Approach 2:
The patent changes the fundamental parameter of problem detection from rule-based thresholds to predictive probability scores generated by machine learning models. This parameter transformation allows the system to maintain high reliability in problem detection while reducing complexity, as the models process multiple factors simultaneously and produce prioritized alerts without requiring manual configuration of numerous separate algorithms.
3Measurement precision
If manual analysis of network monitoring data is performed, then measurement precision is maintained, but productivity decreases and time consumption increases
Solution Approach 1:
The patent introduces an intermediary predictive analytics layer between raw network monitoring data and human operators. This intermediary automatically processes and prioritizes alerts using machine learning models, maintaining measurement precision by preserving critical information while dramatically improving productivity by filtering and ranking issues before they reach human analysts, reducing time consumption for issue identification.
Solution Approach 2:
The patent performs preliminary analysis of network data using predictive models before human intervention is needed. The system pre-processes monitoring data, identifies potential issues, and prioritizes them based on predicted impact and likelihood, so that when human operators review the data, the most critical issues are already highlighted, maintaining accuracy while significantly improving productivity and reducing time to identify critical problems.
Data Source
AI summary
Aspects of the subject disclosure may include, for example, identifying a network service assurance objective and obtaining first network monitoring data indicative of a first operational status of a group of network devices. The first network monitoring data are prioritized according to the network operations objective to obtain first prioritized results and a predictive model is trained based on the first network monitoring data and the first prioritized results. The second monitoring data indicative of a second operational status of the group of network devices are evaluated according to the trained predictive model to obtain second prioritized results, wherein the network monitoring data is prioritized according to the second prioritized results. Other embodiments are disclosed.