Communication Network Failure Detection Using Alarm Log Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication network failure detection systems lack objectivity in evaluating the degree of abnormality and fail to comprehensively detect failures, as they rely on manual analysis and cannot distinguish between significant and less significant factors, leading to missed detections of critical failures, especially when multiple factors contribute to alarms.
Innovation Solution
A system that calculates the occurrence intensity of more significant factors causing alarms, using independent component analysis to extract time and space variation parameters, and compares these intensities against a probability distribution to objectively determine the degree of abnormality and detect network failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual analysis of alarm logs is performed by maintenance staff, then the system can detect abnormalities, but the detection process requires a lot of labor and time
Solution Approach 1:
The patent replaces manual mechanical analysis with automated information processing systems. The management server automatically collects alarm logs, determines significance levels, and performs statistical analysis, substituting human labor with computational processes that operate faster and without fatigue.
Solution Approach 2:
The system enables self-service by allowing the management server to autonomously perform failure detection without requiring continuous human intervention. The automated determination of significant factors and statistical evaluation allows the system to monitor itself and alert operators only when actual failures occur.
2Measurement precision
If multiple alarms are aggregated to identify fundamental factors, then the detection accuracy improves, but crucial alarms may be hidden by chain reaction alarms
Solution Approach 1:
The patent segments alarms into different significance levels (first significance level for crucial alarms, second significance level for chain reaction alarms). This segmentation allows the system to prioritize and analyze alarms hierarchically, ensuring that crucial alarms are not obscured by less important ones while still utilizing aggregated information for comprehensive failure detection.
3Loss of information
If individual alarm analysis is performed, then specific alarm details are captured, but the overall network failure state cannot be comprehensively evaluated
Solution Approach 1:
The patent merges individual alarm analyses with aggregate statistical evaluation. The management server both preserves details of individual significant alarms and performs statistical analysis on the collection of alarms to determine overall network failure states. This combining approach ensures that neither individual alarm information nor comprehensive network state evaluation is lost.
4Reliability
If experienced staff perform log analysis, then accurate failure detection is possible, but the system requires high expertise and cannot be easily scaled
Solution Approach 1:
The patent replaces the need for human expertise with automated algorithms. The management server uses predetermined statistical methods and significance determination rules to perform failure detection, eliminating the need for maintenance staff to have specialized knowledge while maintaining high detection accuracy. This substitution makes the system scalable without requiring additional expert personnel.
Data Source
AI summary
The system for detecting a failure on a communication network according to an objective basis by analyzing an alarm log output by a management server of the communication network and evaluating a degree of abnormality of the communication network tracing back to a more significant factor causing the alarm, calculates an occurrence intensity of the more significant factor 500 causing an alarm based on the recording contents of the alarm, and detects a failure derived from the more significant factor of the communication network based on the calculated intensity of occurrence of said more significant factor 500.