Software Defined Networking (SDN) Intrusion Detection Using Machine Learning

Research Supervisor
Journal Title
Journal Title
Journal ISSN
Volume Title
Information security and data analysis systems for Big Data have recently taken on new significance because of the massive amount of data and its steady growth. The volume, velocity, and variety of Big Data necessitate the development of novel methods for dealing with it. An intrusion detection system (IDS) is a piece of hardware or software that looks at data and looks for any attack on a system or network. SDN anomaly detection method is a software program that uses machine algorithms to identify manufacturing anomalies. Anomaly detection is used to solve a variety of problems, including selective logging, privacy protection, reputation-based protection, multiple threat protection, and dynamic threat response. Distributed networks that create huge volumes of data on a daily basis include mobile phones, wearable gadgets, and self-driving automobiles. Anomaly detection services are critical for the device's security and privacy. Machine learning is a subset of artificial intelligence that allows software programs to anticipate outcomes more accurately without having to build them explicitly. Machine learning with network anomaly detection systems has gotten a lot of attention because of its high categorization accuracy. As a consequence of technological advancements, cyberattacks are expected to rise considerably in frequency by 2021. This research looks at a water manufacturing anomaly detection system that employs machine learning to improve its efficiency and accuracy. SDN dataset has been used in this research. Analyzing the combinations of the most popular feature selection techniques and classifiers, such as K-Nearest Neighbors (KNN) Classification, Decision Tree (DT) Classification, and Random Forest (RF) Classification, has a good union of feature selection techniques and classifiers. The decision tree was found to be 98 percent accurate, while KNearest Neighbors (KNN) was 99 percent accurate, and Random Forest was 95 percent accurate. The experiment demonstrated that the machine learning model is effective for big data, has high performance, and requires less time to train.
Department Name
Electrical and Computer Engineering
North South University
Printed Thesis