Securing Dynamic Service Function Chain Orchestration in EC-IoT Using Federated Learning.

Area of Science:

• Computer Science
• Network Engineering
• Cybersecurity

Background:

• Advancements in IoT and edge computing necessitate dynamic service orchestration.
• Increasing edge devices and data volumes raise concerns for data security and efficient orchestration.
• Existing solutions struggle to balance security, privacy, and performance in dynamic EC-IoT environments.

Purpose of the Study:

• To propose a secure dynamic Service Function Chain (SFC) orchestration algorithm for Edge Computing-Internet of Things (EC-IoT) scenarios.
• To enhance communication efficiency and protect data privacy using a federated learning framework.
• To reduce end-to-end latency and improve convergence speed in dynamic orchestration.

Main Methods:

• Developed a dynamic SFC orchestration security algorithm based on federated learning.
• Integrated block coordinated descent and quadratic penalty algorithms for communication efficiency and privacy.
• Employed a deep reinforcement learning algorithm for adaptive orchestration and latency reduction.

Main Results:

• The proposed algorithm demonstrates superior convergence and latency performance compared to existing methods.
• Achieved approximately 33% reduction in overall latency.
• Improved overall convergence speed by about 9% while ensuring privacy protection.

Conclusions:

• The algorithm effectively secures data privacy for edge computing nodes.
• It meets the dynamic SFC orchestration requirements in EC-IoT settings.
• The approach offers a robust solution for balancing security, privacy, and performance in complex network environments.