Optical Network Anomaly Estimation via Performance Baseline Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for detecting anomalies in optical networks, such as WDM-NW, can only identify failures after a communication interruption occurs, making it impossible to estimate anomalies in advance.
Innovation Solution
An estimation apparatus and method that acquire performance information from nodes in an optical network and compare it to baseline data from normal conditions to estimate anomaly occurrence and identify affected nodes, using machine learning to model and predict anomalies before they cause significant disruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If failure detection is performed using existing techniques (Ping, connection tests), then failure location can be identified after communication interruption, but anomaly cannot be estimated in advance
Solution Approach 1:
The system performs preliminary actions by collecting performance information during normal operation and learning the normal state characteristics before anomalies occur. This enables the system to compare current performance against the learned normal state and estimate anomalies in advance, rather than waiting for communication interruptions to occur.
Solution Approach 2:
The system segments the detection approach by separating anomaly estimation from failure detection. It uses performance information (such as optical signal characteristics) to estimate anomalies before they cause communication interruptions, while existing failure detection methods remain for confirming actual failures. This segmentation allows proactive anomaly estimation without relying on reactive failure detection.
2Reliability
If performance information is continuously monitored and compared to normal state, then anomaly can be estimated in advance, but system complexity increases
Solution Approach 1:
The system applies self-service by using the optical network's own performance information (such as optical signal characteristics already present in the network) to monitor and estimate anomalies. Rather than introducing complex external monitoring equipment, the system leverages existing performance data from the network elements themselves, reducing additional hardware complexity while enabling anomaly estimation.
Data Source
AI summary
An estimation apparatus includes: an acquiring unit that acquires performance information corresponding to the performance of light from nodes configuring an optical network; and an estimating unit that compares the performance information to be an estimation target acquired by the acquiring unit with performance information at normal time at which no anomaly occurs, and performs at least one of estimation of anomaly occurrence and identification of an anomaly occurring node.