Traffic Management in IoT Backbone Networks Using GNN and MAB with SDN Orchestration.

Area of Science:

• Computer Science
• Network Engineering
• Artificial Intelligence

Background:

• Traditional traffic management in software-defined IoT networks (SDN-IoTs) struggles with dynamic traffic patterns.
• Queuing theory and static policies are insufficient for ensuring Quality of Service (QoS).

Purpose of the Study:

• To develop a dynamic traffic management approach for SDN-IoTs.
• To leverage Graph Neural Networks (GNNs) and multi-arm bandit algorithms for policy optimization.

Main Methods:

• Utilized GNNs for learning and predicting real-time network traffic patterns.
• Employed multi-arm bandit algorithms to dynamically optimize traffic management policies based on GNN predictions.
• Evaluated the approach on diverse datasets including KDD Cup 1999, CAIDA, and NSL-KDD.

Main Results:

• The proposed approach demonstrated superior performance compared to state-of-the-art methods.
• Achieved significant improvements in throughput, reduction in packet loss, and decreased network delay.
• Effectively identified anomalous traffic patterns within the network.

Conclusions:

• The GNN and multi-arm bandit approach offers a robust solution for dynamic traffic management in SDN-IoTs.
• Enables efficient network resource allocation and ensures end-user Quality of Service.
• Provides a promising direction for enhancing network stability and security.