The College of Computer Science and Information Technology discussed the master’s thesis titled: “Improving Performance Analysis in Software-Defined Networks and the Internet of Things.”

The College of Computer Science and Information Technology discussed the master’s thesis titled: “Improving Performance Analysis in Software-Defined Networks and the Internet of Things.”

The thesis, presented by student Rawand Nouri Salih, was supervised by Professor Dr. Issa Ibrahim Issa.

The study aimed to develop an integrated framework to improve the performance of SDN-IoT networks. This was achieved by employing Metaheuristic Algorithms (ACO, PSO, GA, and DE) in conjunction with machine learning classifiers (Random Forest, KNN, and XGBoost) to enhance feature selection and reduce dimensionality while maintaining accuracy. The research also sought to strengthen the capabilities of Intrusion Detection Systems (IDS) in IoT environments.

Using the IoT-SDN IDS dataset, the study applied optimization algorithms (ACO, PSO, GA, DE) and demonstrated the superior performance of hybrid models in improving attack detection. The combination of the ACO algorithm with Random Forest showed the best results, while the combination of GA and XGBoost offered a good balance between feature reduction and high accuracy.

The researcher recommended adopting hybrid frameworks that combine optimization algorithms with classifiers to enhance the accuracy of intrusion detection systems and reduce temporal complexity. The recommendation also focused on selecting the most important features to ensure higher speed and efficiency in SDN-IoT environments and to test these models in practical applications across various fields.