IP-Optical Link Discovery via Neural Network Port Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current networks require manual provisioning of inter domain links between IP and optical layers, which is error-prone, time-consuming, and cumbersome, and existing methods for automatic discovery lack accuracy and scalability, especially in large networks.
Innovation Solution
A method using a convolutional neural network (CNN) for automatic discovery of IP-optical links by grouping network nodes, filtering ports, producing class IDs, matching IP ports to optical ports, and verifying links, allowing for parallel processing and accurate identification of IP-optical links without the need for hardware changes or LLDP support.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual provisioning of inter domain links is used, then configuration accuracy can be maintained, but time consumption and operational complexity increase significantly
Solution Approach 1:
The system performs automatic link discovery through self-learning neural networks that autonomously identify IP-optical mappings without human intervention. The controller automatically provisions inter domain links by learning traffic patterns and correlating IP addresses with optical ports, eliminating manual configuration while maintaining high accuracy through iterative learning and verification mechanisms.
Solution Approach 2:
The patent replaces manual mechanical provisioning operations with an automated neural network-based discovery system. The machine learning model substitutes human operators by analyzing traffic flows, identifying patterns, and automatically establishing link mappings between IP and optical layers, thereby reducing time consumption while maintaining configuration accuracy through intelligent algorithms.
2Loss of time
If existing automatic discovery methods are used, then time consumption is reduced, but accuracy and scalability deteriorate in large networks
Solution Approach 1:
The system segments the network discovery process into multiple hierarchical levels: first grouping nodes into discovery groups, then identifying links within each group, and finally verifying mappings. This segmented approach improves both accuracy and scalability by breaking down the complex discovery task into manageable stages that can be processed efficiently even in large networks.
Solution Approach 2:
The patent introduces multi-layer filtering across different dimensions of network data (traffic patterns, port characteristics, protocol information) to enhance discovery accuracy. By analyzing links from multiple dimensional perspectives simultaneously, the system achieves high accuracy in identifying IP-optical mappings while maintaining scalability through efficient parallel processing of multi-dimensional features.
3Measurement precision
If comprehensive link verification is performed, then topology accuracy is improved, but processing complexity and time increase
Solution Approach 1:
The system performs preliminary filtering and grouping of network nodes and ports before conducting detailed link verification. By pre-organizing data into discovery groups and filtering out obviously non-matching candidates, the system reduces the complexity of subsequent verification tasks while maintaining comprehensive topology accuracy through targeted verification of promising candidates.
Solution Approach 2:
The patent implements a multi-stage verification process where not all possible links are verified with equal depth. Instead, the system performs partial verification on high-probability candidates identified through initial filtering, and excessive verification only where necessary to resolve ambiguities. This selective verification approach maintains topology accuracy while minimizing overall processing complexity.
Data Source
AI summary
A method of identifying IP-optical links in a network having a plurality of nodes, including: grouping network nodes into discovery groups; for each group filtering ports of the nodes in the discovery group; for each group producing class IDs for each filtered port using a machine learning model; for each group matching IP ports to optical ports from the filtered ports using the class IDs for each port to identify IP-optical links; and verifying identified IP-optical links.