QoS improvement with an optimum controller selection for software-defined networks.

Area of Science:

• Computer Science
• Network Engineering

Background:

• Software-Defined Networking (SDN) decouples data and control planes, centralizing network management in controllers.
• Controller features vary, making optimal selection a complex multi-criteria decision-making (MCDM) problem.

Purpose of the Study:

• To propose and validate a two-step approach for selecting optimal SDN controllers.
• To evaluate the performance of selected controllers based on qualitative features and quantitative analysis.

Main Methods:

• Utilized the Analytical Network Process (ANP) for feature-based controller ranking.
• Conducted experimental analysis on real-world Internet and Brite topologies for QoS comparison.
• Assessed controller performance under normal and high traffic load scenarios.

Main Results:

• The ANP-based selection approach outperformed existing methods in simulations.
• Optimized controller selection reduced topology discovery time and network delay.
• Achieved increased throughput and efficient CPU utilization with the proposed method.

Conclusions:

• The ANP is applicable and effective for SDN controller selection.
• The proposed method enhances network performance metrics like speed and resource utilization.
• This approach provides a systematic way to choose the best SDN controller for specific network needs.