Distributed Denial of Service (DDoS) attacks present a significant challenge to the security and availability of cloud-based services. Exploiting inherent cloud features such as resource pooling, rapid scalability, and multi-tenancy, attackers launch high-rate and low-rate DDoS attacks that disrupt operations, escalate costs, and Erode customer trust. This paper reviews the vulnerabilities of cloud environments and existing detection and mitigation strategies, highlighting gaps in real-time adaptive frameworks. A novel layered architecture is proposed to enhance DDoS resilience. It integrates real-time traffic monitoring with sandboxing, IP and session validation, and Captcha challenges, and machine learning-based pattern analysis to differentiate between legitimate and malicious traffic. The system dynamically quarantines threats while ensuring uninterrupted service for genuine users. The methodology addresses cloud-specific vulnerabilities, offering a scalable, efficient, and cost-effective solution to evolving DDoS threats.