A Novel Three Layer Filtering (3L-F) Framework for Prevention of DDoS Attack in Cloud Environment

Data security is an integral requirement of any modern information system as attackers are gaining chances due to the prompt improvement in digital technology. However, in the current decade, the use of cloud computing is rising steeply, and so is network traffic. As the cloud computing model is based on the distributed computing, cloud servers are widely distributed and cloud users can access the service from anywhere and at any time. This makes the cloud servers, a target for the adversaries. The most common attack in a cloud environment is the DDoS attack that causes bulky and abnormal traffic to the cloud server. The cloud server is incapable to manage such unusual traffic and stops momentarily by making the server down with excessive traffic. DDoS attacks can be avoided by diligent traffic control prior to the DDoS attack. This paper proposes a novel three-layer filtering mechanism to prevent various forms of DDoS attacks. The first layer of the proposed DDoS attack prevention mechanism uses two-level authentication processes. Second layer filtering verifies whether the user accesses the resources within the pre-defined limits and the third layer filtering sieves out the spoofed packets. The proposed model has been analyzed for evaluating the performance in terms of CPU overhead and load, the throughput of the victim, the reduction in connection delay. The result analysis shows that the proposed model has improved performance with a higher detection rate of 0.92 and a lower dropout rate of 0.10.