AI-Based DDoS Protection System for Cloud (Netlify) Architecture and Tool

Abstract

Distributed Denial-of-Service (DDoS) attacks are one of the major cybersecurity threats faced by cloud- hosted websites and online services. These attacks flood servers with a massive volume of fake traffic, causing service slowdown, disruption, or complete unavailability for genuine users. This project presents an AI-Based DDoS Protection System for Cloud: Architecture and Tool that leverages machine learning and real-time traffic monitoring to detect and mitigate malicious requests automatically. The system is developed using the CICIDS 2018 public intrusion detection dataset, which contains realistic network traffic scenarios including benign and attack records. Feature engineering transforms raw traffic data into meaningful inputs such as Flow Duration and Total Forward Packets. A Random Forest Classifier is trained and selected as the final model based on its strong accuracy, fast prediction speed, and resistance to overfitting. The model is integrated with a threshold-based rate limiting mechanism to create a hybrid detection system. When suspicious activity is identified, the system automatically flags malicious traffic, blocks attacker IP addresses, and ensures continued service availability with minimum downtime. A Flask-based backend API manages traffic analysis and security actions, while an interactive web dashboard provides live request logs, attack status, blocked IP management, confidence scores, and monitoring statistics. A simulated attack module is also developed to generate mixed benign and malicious traffic for testing system performance. This solution offers a cost-effective and scalable cloud security framework that helps organizations proactively protect their digital infrastructure from DDoS attacks