A Game Model and Fault Recovery Algorithm for SDN Multi-Domain

Area of Science:

• Computer Science
• Network Engineering
• Distributed Systems

Background:

• Software-Defined Networking (SDN) enables flexible Wireless Sensor Network (WSN) management by decoupling control logic.
• SDN controller failures pose a significant challenge to network stability and performance.
• Existing fault recovery mechanisms often struggle with efficiency and scalability in multi-domain environments.

Purpose of the Study:

• To propose a novel game-theoretic approach for autonomous fault recovery in SDN controllers.
• To develop an algorithm that ensures timely and cost-effective recovery from controller failures.
• To enhance the resilience and reliability of WSNs managed by SDN.

Main Methods:

• A game-theoretic model for multi-domain SDN controller fault recovery is developed.
• A game-enhanced autonomous fault recovery algorithm is proposed, incorporating controller capacity and device transition relationships.
• Linear programming and a multi-population particle swarm optimization algorithm are used for optimal switch mapping and migration.
• The algorithm iteratively optimizes controller-switch mapping to minimize migration costs and load.

Main Results:

• The proposed algorithm achieves fast fault recovery with low migration costs for SDN controllers.
• It effectively balances switch migration costs and controller load pressure.
• Experimental results show a significant reduction in fault recovery time compared to existing methods.
• Reduced propagation delay in SDN controllers is observed.

Conclusions:

• The game-enhanced autonomous fault recovery algorithm offers a robust solution for SDN controller failures in WSNs.
• The approach enhances network resilience by optimizing recovery processes.
• This method provides a significant improvement in fault recovery speed and resource management efficiency.