A Comprehensive Review of Vulnerabilities and Ai-Enabled Defense Against DDOS Attacks for Securing Cloud Services

Abstract

In this paper, it proposes a software defined end edge cloud architecture coupled with IP FPGA acceleration to integrate the two lines of checks, as well as a manner to use deep learning based IP FPGA to secure IoT networks. In this work, it uses the Caffe framework to implement multiple CNN architectures such as LeNet, AlexNet, and ResNet50, reaching detection accuracy rates of up to 97.8%. The performance improvement of such an engine on a Xilinx KU115 FPGA is shown by implementing it: 1000 packets per second, with low power consumption. It validates the results through a real use case of how this hybrid approach can be used for real time threat detection in large scale IoT deployments.