IoT DDoS Attack Detection System Using Machine Learning Classification Techniques

• Published in: Proceedings (IEEE Student Conference on Research and Development. Online) pp. 234 - 239
• Main Authors: Sa'Idi, Aiman Saftwan Bin, Jamil, Ameerah Muhsinah Binti
• Format: Conference Proceeding
• Language: English
• Published: IEEE 13.12.2023
• Subjects: Bayes methods; Biological system modeling; Classification algorithms; Decision Tree (DT); Decision trees; Denial-of-service attack; Distributed Denial of Service (DDOS); Internet of Things; Internet of Things (IoT); Machine Learning (ML); Naïve Bayes (NB); Radio frequency; Random Forest (RF)
• ISSN: 2643-2447
• DOI: 10.1109/SCOReD60679.2023.10563925

The Internet of Things (IoT) has experienced substantial expansion, facilitating the connection of a multitude of devices to the Internet and enhancing the level of automation in several aspects of our daily routines. Nevertheless, the expansion has also brought to light several weaknesses, most notably the occurrence of Distributed Denial of Service (DDoS) attacks. IoT devices, which frequently lack comprehensive security measures, are vulnerable to such attacks. The objective of this work is to explore the feasibility of employing machine learning (ML) classification techniques to identify DDoS attacks on IoT networks. This study focuses on three commonly employed machine learning classification techniques: (i) Decision Tree (DT), (ii) Random Forest (RF), and (iii) Naïve Bayes (NB). The models will undergo training using a dataset specifically designed for detecting and classifying DDoS attacks. The TON IoT (UNSWIoT20) dataset sourced from the School of Engineering at the University of New South Wales Canberra is included in the study as the selected dataset. Among the several datasets accessible from the source, the IoT Network dataset with a specific emphasis on DDoS attacks was deliberately selected. To determine the efficacy of the suggested solution, a comprehensive assessment of the dataset is conducted. The testing results are then compared and analyzed to determine the model with the utmost level of precision in detecting DDoS attacks. This paper is a valuable contribution to the subject of IoT security by offering possible efficient and resilient approaches to address DDoS attacks targeting IoT devices and networks. The detection system, which utilizes machine learning, provides a proactive approach to protection, thereby guaranteeing the security and dependability of Internet of Things (IoT) services. Additionally, it facilitates the sustained growth of the IoT ecosystem.