Graph Embedding Virtual Network Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual network mapping methods face challenges in dynamically changing network environments, particularly in embedding virtual networks onto substrate networks regardless of network size and scenario, leading to inefficiencies in mapping and resource allocation.
Innovation Solution
A graph embedding-based method using a network encoder and decoder, employing graph convolution networks (GCN) and multi-layer perceptrons (MLP), which generates embedding values for nodes and determines optimal mappings by comparing difference thresholds, allowing for dynamic and efficient virtual network mapping across varying network sizes and scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If virtual network mapping is performed using traditional embedding methods (CNN or MLP), then the mapping can be achieved, but the structure of reinforcement learning must be changed in environments with dynamic network size
Solution Approach 1:
The patent applies graph embedding technology to create a universal mapping framework that handles different network sizes and scenarios without requiring changes to the reinforcement learning structure. The graph embedding approach embeds both virtual and substrate networks into a common vector space, enabling the same mapping algorithm to work across diverse network configurations, thus achieving multi-functionality and adaptability.
Solution Approach 2:
The patent changes the fundamental parameters of the mapping approach by transitioning from traditional CNN/MLP embedding to graph embedding. This parameter change involves representing networks as graphs with nodes and edges, using graph neural networks to capture topological relationships, and embedding nodes into a continuous vector space. This fundamental parameter change enables the system to handle dynamic network sizes without structural modifications.
2Measurement precision
If graph embedding is used to map virtual network to substrate network, then mapping accuracy is improved, but computation complexity increases due to embedding calculations for each node
Solution Approach 1:
The patent segments the virtual network mapping problem into individual node-level embedding tasks. Instead of computing a single global mapping, the system generates embedding vectors for each node independently using graph neural networks. This segmentation allows for parallel computation and enables the use of pre-computed embeddings, reducing overall computation complexity while maintaining high mapping accuracy at the node level.
3Measurement precision
If embedding values are generated for every node, then mapping precision is improved, but mapping time increases
Solution Approach 1:
The patent performs preliminary action by pre-computing embedding vectors for substrate network nodes and storing them in advance. When a virtual network mapping request arrives, the system retrieves these pre-computed embeddings and performs only the necessary matching calculations, rather than computing embeddings from scratch. This preliminary action significantly reduces mapping time while maintaining high precision through the use of pre-generated embedding values.
Data Source
AI summary
The present invention provides a graph embedding-based virtual network mapping method including inputting a virtual network and a substrate network corresponding to a virtual network mapping scenario, in an optimized mapping mode, generating an embedding value for every substrate node by applying a graph convolution network (GCN), by a network encoder, to embed the virtual network, determining whether a difference between an embedding value for every substrate node and an embedding value for every previous substrate node which is previously embedded exceeds a set threshold, by a network decoder, and mapping an allocation node of the virtual network to a mapping node of a previous substrate network which is previously mapped according to an embedding value for every previous substrate node when the difference value does not exceed the threshold value, by the network decoder.