Enhancing Detection and Prediction of DDoS Attacks Through Regression Modeling
Kayalvizhi Subrmanian, Gunasekar Thangarasu, Zhao Yanyan, K Nattar Kannan · 2024
Detecting Distributed Denial of Service (DDoS) attacks poses a significant challenge in the realm of machine learning particularly in the context of cloud computing due to the computational complexity involved. A Denial of Service (DoS) attack is an intentional effort by attackers typically originating from a single source to render an application inaccessible to its intended users. These attacks impair network bandwidth and consume system resources thereby preventing genuine users from accessing the application. DDoS attacks on the other hand leverage numerous sources orchestrated by the attacker. The network, transport, presentation, and application layers of the seven-layer OSI model are often targeted in such attacks. This study aims to address the problem of identifying DDoS attacks within a cloud environment by utilizing the widely adopted CSE-CIC-IDS2018 dataset. Multiple logistic regression analysis is employed to predict DDoS and bot attacks. The analysis focuses on a traffic subset file collected on a Friday afternoon. By employing machine learning techniques on real-world datasets, we seek to advance the field of DDoS attack detection in cloud environments.