Abstract

In today's digitally connected society, the cloud is vital in communications. There is a lot of valuable information online, and the services are available anywhere at any time. On the other hand, cloud services attract intruders to exploit online information, bringing considerable changes to the attack landscape in cyberspace. Distributed Denial of Service (DDoS) attackers cause the site to be overflowing with errant traffic, causing the network's data layer to be compromised or completely disabled. DDoS attacks are increasing in severity and frequency. Harm makes many governments and major financial institutions with an online presence the target of attacks. The DDoS attacks to pertinacity are raising increasingly based on extortion and data theft. This paper proposes a feature selection using Denoising AutoEncoder (DAE) and a Convolutional Neural Networks (CNN) classification model for DDoS attack detection in network traffic data; the NSL-KDD dataset is used, we compared related works that used the NSL-KDD dataset. The suggested proposed evaluation shows a detection accuracy of (97.7%). This result is higher accuracy than previous works to identify hackers.