Optimizing Cloud Network Security with Anomaly Detection for DDoS Attacks

The rapid rise of cloud computing has revolutionized infrastructure management for organizations, offering scalability, flexibility, and cost-effectiveness. These advantages come with significant security challenges, especially the increasing threat of Distributed Denial of Service(DDoS) attacks. Traditional security methods like firewalls and signature-based Intrusion Detection Systems(IDS) are insufficient against sophisticated attacks in dynamic cloud environments. These systems rely on static rules and predefined signatures, making them ineffective against new and evolving threats. This thesis proposes an adaptive, scalable, and intelligent solution: a cloud-based anomaly detection system powered by machine learning. The system is designed to detect network anomalies, especially DDoS attacks, in real-time by using both supervised and unsupervised machine learning models. It continuously learns from network traffic, adapting to new attack vectors, and improving detection accuracy over time minimizing false positives and negatives. Cloud environments, which operate on shared, distributed resources, pose additional challenges due to unpredictable traffic surges. Distinguishing between legitimate traffic spikes and malicious activity is difficult, requiring the system to process high volumes of data in real time. Moreover, the security solution must scale dynamically with the cloud infrastructure without compromising performance or increasing latency. A continuous learning scenario to ensure the system can adjust detection thresholds, providing robust protection in dynamic cloud settings. Testing on real-world and synthetic datasets showed that the system outperforms traditional methods, with high accuracy. It scales efficiently, handling large traffic volumes with low latency, making it ideal for cloud service providers and enterprises. This research offers a novel, cost-effective solution to the evolving threat of DDoS attacks in modern cloud environments.