Machine Learning Algorithms for Prediction of the Quality of Transmission in Optical Networks

Area of Science:

• Telecommunications Engineering
• Network Optimization
• Machine Learning Applications

Background:

• Rising traffic in Dense Wavelength Division Multiplexing (DWDM) networks necessitates cost-effective solutions for increased transmission speeds.
• Efficient network reconfiguration is crucial for managing DWDM traffic demands.
• Accurate quality of transmission assessment is a key challenge in DWDM network reconfiguration.

Purpose of the Study:

• To propose a Machine Learning (ML) based method for assessing the quality of transmission in DWDM networks.
• To develop an ML approach utilizing only data accessible through the DWDM network control plane.
• To evaluate the performance of various ML classifiers for transmission quality assessment in real DWDM network scenarios.

Main Methods:

• Development of a database using information exclusively from the DWDM network control plane.
• Implementation and testing of several ML classifier types.
• Performance comparison of ML classifiers on two distinct real DWDM network topologies.

Main Results:

• Promising results achieved in assessing transmission quality using ML methods.
• Demonstrated feasibility of using control plane data for ML-based quality assessment.
• Comparative analysis identified effective ML classifiers for DWDM network applications.

Conclusions:

• The proposed ML-based method offers a viable solution for quality of transmission assessment in DWDM networks.
• The approach simplifies and potentially accelerates network reconfiguration procedures.
• Further research is warranted to explore the full potential of ML in optimizing DWDM network operations.