SDN Routing Control via Deep Neural Network Q-Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional network routing algorithms in software-defined networks (SDNs) lack adaptability and fail to effectively manage real-time traffic distribution, leading to network congestion and poor load balancing due to their reliance on fixed topologies, which requires manual reconfiguration when network structures change.
Innovation Solution
A method using a deep neural network model combined with a Q-learning algorithm to generate routing control actions in SDNs, based on current network state parameters such as link utilization and flow completion times, allowing for dynamic and adaptive control adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional network routing algorithms are used, then routing control is simple, but adaptability to network topology changes is poor and load balancing is difficult to implement
Solution Approach 1:
The patent replaces traditional mechanical routing algorithms with a deep neural network model that processes network state parameters to generate routing control actions. The neural network model takes input features including link utilization, queue lengths, and flow characteristics, processes them through multiple layers, and outputs optimized routing decisions, thereby achieving high adaptability while managing complexity through automated intelligence
Solution Approach 2:
The patent changes the routing control approach by using real-time network state parameters (link utilization, queue lengths, flow completion times) as input to the neural network model. This dynamic parameter-based approach enables the system to adapt to changing network conditions automatically, transforming static routing algorithms into dynamic, data-driven control mechanisms
2Productivity
If fixed topology routing algorithms are used, then implementation is simple, but real-time traffic distribution is not considered leading to network congestion
Solution Approach 1:
The patent implements feedback mechanisms by continuously monitoring network state parameters (link utilization, queue lengths, flow characteristics) and using this real-time information as input to the neural network model. The model processes this feedback data to generate routing control actions that optimize traffic distribution in real-time, preventing congestion through data-driven decision making
Solution Approach 2:
The system performs self-service by automatically adjusting routing decisions based on its own monitored network state. The neural network model processes internal network metrics and generates appropriate routing control actions without external intervention, enabling the system to self-optimize traffic distribution and respond to congestion conditions autonomously
3Loss of time
If manual routing configuration is used, then control precision is high, but response time to topology changes is slow
Solution Approach 1:
The patent applies preliminary action by having the neural network model continuously process network state parameters and prepare routing control actions in advance. The system monitors network conditions continuously and pre-computes optimal routing decisions based on predicted traffic patterns, enabling rapid response when topology changes occur without requiring manual reconfiguration
Solution Approach 2:
The patent substitutes manual routing configuration with an automated neural network system that continuously learns from network state data and generates routing control actions. This replacement eliminates the slow manual response process while maintaining high control precision through intelligent decision-making algorithms that process real-time network metrics
Data Source
AI summary
Embodiments of this application provide a method for generating a routing control action in a software-defined network and a related device, to provide optimum control actions for the SDN. The method includes: obtaining a current network state parameter of the SDN; determining a Q function of the SDN based on the current network state parameter of the SDN and a deep neural network model, where the deep neural network model is determined based on a current topology structure of the SDN; and determining a routing control action for the SDN based on the Q function and a link state parameter of each link in the SDN. In the technical solution, the deep neural network model is combined with a Q-learning algorithm of reinforcement learning, and optimum control actions can be determined.