A GRU-based traffic situation prediction method in multi-domain software defined network.

Area of Science:

• Computer Science
• Network Engineering
• Artificial Intelligence

Background:

• Software-Defined Networking (SDN) enables multi-domain architectures for large-scale networks.
• Managing complex dynamics in multi-domain SDN presents significant network management challenges.
• Accurate traffic prediction is crucial for maintaining stability in these advanced networks.

Purpose of the Study:

• To propose an effective traffic situation prediction method for multi-domain SDN environments.
• To enhance the prediction accuracy of the Gated Recurrent Unit (GRU) network.
• To improve overall network stability and management through precise traffic forecasting.

Main Methods:

• Analysis of factors influencing data and control traffic to create time-series data.
• Application of the Salp Swarm Algorithm (SSA) for automatic hyperparameter optimization of GRU networks.
• Implementation of the hyperparameter-optimized GRU for traffic situation prediction.

Main Results:

• The proposed method demonstrates superior traffic situation prediction accuracy in multi-domain SDN.
• Optimized GRU network significantly enhances prediction performance compared to standard GRU.
• The approach effectively addresses the complexities of traffic dynamics in multi-domain SDN.

Conclusions:

• The Salp Swarm Algorithm-optimized GRU network provides a robust solution for traffic prediction in multi-domain SDN.
• This method offers a significant improvement over traditional machine learning algorithms for network traffic forecasting.
• Accurate traffic prediction is vital for ensuring the stability and efficient management of complex multi-domain networks.