Enhancing Network Security in IoT Applications through DDoS Attack Detection Using ML

Abstract

A significant issue that affects contemporary network infrastructures is the Distributed Denial of Service (DDoS) attack, which presents serious dangers to organizations, people, and even governments. By flooding a target server, network, or website with deceptive traffic, this kind of cyberattack seeks to prevent it from providing services to legitimate users. For those in charge of maintaining network security, the prevalence and sophistication of these attacks have both grown significantly. DDoS attacks have the potential to lead to significant financial losses and service interruptions. Anomaly-based systems, traffic filtering, and Machine Learning (ML) algorithms are employed to spot them and lessen the effects of their influence. To successfully defend against DDoS attacks, it's imperative to have a proactive attitude and keep up with new security risks. In this study, the CICDDoS 2019 dataset was used to train and evaluate different ML algorithms, including Stochastic Gradient Boosting (SGB), Decision Tree (DT), K Nearest Neighbour (K-NN), Naive Bayes (NB), Support Vector Machine (SVM), and Logistic Regression (LR). The results showed that all the ML algorithms effectively detected DDoS attacks with high accuracy, precision, and recall. However, the SVM algorithm outperformed the other techniques, achieving the highest accuracy =0.99%.