DDoS Attack Detection using Artificial Intelligence

Abstract

Distributed Denial of Service (DDoS) attacks have become one of the most critical threats to the availability of modern network infrastructures and online services. These attacks aim to overwhelm system resources with massive volumes of malicious traffic, causing service disruption and preventing legitimate users from accessing network resources. Traditional detection approaches based on signature matching or static thresholds often struggle to detect modern and evolving DDoS attacks due to their limited adaptability. This work presents the design and implementation of a real-time Machine Learning-based DDoS detection system. The proposed architecture consists of four main components: (1) a packet capture and traffic monitoring module using the Scapy library, (2) a statistical network flow feature extraction engine, (3) a Random Forest classifier for distinguishing between be￾nign and malicious traffic, and (4) a hybrid detection mechanism combining Baseline Moni￾toring, Entropy Analysis, and Emergency Logic to improve detection reliability. An interac￾tive Streamlit dashboard was also developed to provide real-time visualization, monitoring, and intelligent alert generation. The proposed system was evaluated using the CIC-DDoS2019 dataset together with con trolled attack simulations in an isolated environment. Experimental results achieved a classi fication accuracy of 97.99% with an average inference latency of approximately 5 millisec onds per analysis window. These results demonstrate that DDoSentinel provides an efficient, lightweight, and scalable solution suitable for real-time DDoS detection and integration with modern Intrusion Detection and Prevention Systems (IDS/IPS).