Zero-Day Vigilance: Proposing a Machine Learning Framework for DDoS Detection in Cloud Computing

The seemingly exponential growth of cloud computing has introduced menacing and crucial threats that are specifically DDoS attacks. These types of attacks are hard to counter because they exploit as yet unknown vulnerabilities thereby making it difficult to counter the zero-day DDoS attacks. Such threats are rarely detected by traditional anomaly-based systems because they do not rely on anomalies, but rather use real-time detection and response methodologies with the help of sophisticated frames. This paper provides a machine learning-based integrated framework to fight against zero-day DDoS attacks in the cloud context, whose goal is to enhance the detection accuracy and reaction speed including anomaly detection, deep learning, and feature engineering as well as collecting reliable data from cloud honeypots. The purpose of this framework is to become proactive against the increasing number of zero-day DDoS attack threats.