Cloud Network Topology Optimization Using Graph Convolutional Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network topology optimization methods are inefficient in identifying optimal network topologies that balance performance, availability, and scalability, particularly in complex cloud environments.
Innovation Solution
A graph convolutional network (GCN) model is trained using network topology datasets to identify an optimum network topology by evaluating performance, availability, and scalability, utilizing a neural network architecture that includes input, convolutional, pooling, and fully connected layers to determine the best network configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional network topology optimization methods are used, then the optimization process is simple, but the efficiency in identifying optimal network topologies is low
Solution Approach 1:
The patent replaces traditional mechanical or manual network topology optimization methods with a graph convolutional network (GCN) model, which is a type of neural network. This substitution enables automated, data-driven optimization that efficiently identifies optimal topologies by learning from training datasets, thereby resolving the contradiction between simplicity and efficiency.
Solution Approach 2:
The patent introduces a graph convolutional network model as an intermediary between the input network topology and the optimization outcome. The GCN model processes topology data through multiple layers (convolutional, pooling, fully connected) to generate optimized topology configurations, serving as a sophisticated mediator that achieves high efficiency while managing complexity through structured processing.
2Measurement precision
If a graph convolutional network model is used, then the accuracy in identifying optimal topologies is improved, but the computational complexity increases
Solution Approach 1:
The patent segments the neural network architecture into distinct functional modules: graph convolutional layers for feature extraction, pooling layers for dimensionality reduction, and fully connected layers for final prediction. This segmentation allows each component to specialize in specific tasks, improving overall accuracy while making the complex system more manageable and interpretable.
Solution Approach 2:
The patent transforms the network topology problem into a graph-structured data format that the GCN model can process effectively. By representing topologies as graphs with nodes and edges, the model can capture complex relationships and dependencies that traditional flat data structures cannot, thereby improving measurement precision in evaluating optimality functions.
3Adaptability or versatility
If multiple optimality parameters are considered, then the comprehensiveness of optimization is improved, but the difficulty of evaluation increases
Solution Approach 1:
The patent merges multiple optimality parameters (performance, availability, scalability) into a single unified optimality function that the GCN model evaluates comprehensively. The model processes all parameters simultaneously through its graph convolutional operations, producing a unified assessment of each candidate topology's overall quality, thereby resolving the contradiction between comprehensiveness and evaluation difficulty.
Solution Approach 2:
The graph convolutional network model serves multiple functions: it evaluates performance metrics, assesses availability, measures scalability, and generates optimized topology configurations all within a single unified framework. This multi-functionality allows the system to comprehensively consider multiple optimality parameters while simplifying the evaluation process through integrated processing.
Data Source
AI summary
A method, computer program product, and computer system for network topology optimization. A graph convolutional network (GCN) model is trained using training network topology datasets as input. Each training network topology dataset includes: (i) a specified network topology and an associated optimal network topology and an optimality function value for the optimal network topology, (ii) relative weights of optimality parameters including performance, availability, and scalability, and (iii) an identification of an optimality function of the optimality parameters weighted by the relative weights. The specified network topology includes components, relationships between the components, and metadata pertaining to the components. Each output node in an output layer of the GCN model includes an optimality function value for a different candidate network topology. One of the output nodes identifies an optimum network topology relative to the specified network topology as being the candidate network topology having a highest optimality function value.