DWDM Smart Channel Launch Energy Optimization Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current networking management technologies in DWDM networks lack the capacity to learn and analyze the complex optical characteristics of the entire network, leading to inefficiencies in launching channels and optimizing launch energy, power, and efficiency.
Innovation Solution
A Software Defined Domain Controller (SDDC) utilizing machine learning algorithms to monitor and predict optimal launch energy, power, and efficiency by training on network topology, parameters, and characteristics, recommending smart channel configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional networking management technologies are used in DWDM networks, then the network can maintain basic channel launch operations, but the system lacks the capacity to learn and analyze complex optical characteristics, leading to inefficiencies in launching channels and optimizing launch energy, power, and efficiency
Solution Approach 1:
The patent replaces traditional mechanical/network management systems with a machine learning-based system. The ML model processes complex optical characteristics and network topology data to predict optimal launch energy, power, and efficiency parameters, substituting manual or rule-based network management with intelligent automated decision-making.
Solution Approach 2:
The system changes the approach to network management by introducing new parameters such as spectral efficiency, launch energy, and power optimization. The ML model analyzes these parameters along with network topology and optical characteristics to determine optimal channel launch configurations, transforming how network performance is measured and optimized.
2Loss of energy
If machine learning algorithms are introduced to optimize launch energy and power, then spectral efficiency is maximized and operational expenditures are reduced, but the system complexity increases due to training requirements and computational resources
Solution Approach 1:
The system performs preliminary training of the machine learning model using historical network data, topological information, and optical characteristics before actual channel launch operations. This pre-training phase allows the model to learn patterns and relationships, enabling it to make accurate predictions during operational use without requiring complex real-time computations.
Solution Approach 2:
The patent creates a virtual copy of the network system in the form of an ML model that replicates network behavior and characteristics. This digital twin allows the system to simulate and analyze network performance without physically modifying the actual network infrastructure, reducing operational complexity while improving optimization capabilities.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The SDDC optimizes channel launch energy, power, and efficiency by predicting and recommending optimal configurations, maximizing spectral efficiency and reducing operational expenditures.
Implementation Method 1
training a machine learning (ML) model using the set of the parameters and the characteristics of the channel in the fiber optic network; and predicting a target launch energy, power and efficiency using the ML model for the channel in the fiber optic network
Implementation Method 2
To compensate for optical signal loss, an optical amplifier must be used using a common method of stimulated emission of photons
Implementation Method 3
The most common type of optical amplifiers is Erbium-Doped Fiber Amplifiers (EDFAs)
Data Source
AI summary
Aspects of the subject disclosure may include, for example, a device, including: a processing system including a processor; and a memory that stores executable instructions that, when executed by the processing system, facilitate performance of operations of: determining a network topology in a fiber optic network, wherein the network topology comprises a plurality of network elements joined by fiber optic links; selecting parameter values of a set of parameters for a channel between a first network element and a second network element in the plurality of network elements; applying the parameter values to create parameterized dense wavelength division multiplexing (DWDM) signals between the first network element and the second network element; responsive to the applying the parameter values, determining characteristics of the channel in the fiber optic network; repeating the selecting and applying of the parameter values to determine the characteristics of the channel using different selected parameter values; training a machine learning (ML) model using the set of the parameters and the characteristics of the channel in the fiber optic network; and predicting a target launch energy, power and efficiency using the ML model for the channel in the fiber optic network. Other embodiments are disclosed.