A DDoS Attack Detection on Cloud Framework Using Improved Features Based Machine Learning Approach
Ravi Bhargav, Vishal Kumar Jain, Manish Verma · 2022 Second International Conference on Advances in Electrical, Computing, Communication and Sustainable Technologies (ICAECT) · 2022
A Denial of Service (DDoS) attack consumes the network bandwidth and computing resources of a targeted system, preventing the target system from being DDoS attacked by unauthorized users. This research work proposed a modified feature selection-based neural network for efficient DDoS attack detection. The proposed method is divided into three stages: feature selection, training, and testing. For the implementation of the proposed work using MATLAB 2020 software, MATLAB is a well-known academic as well as an industrial research software package for this work. In R2020 MATLAB, the proposed method is designed and simulated. There are different types of DDoS attacks present on the cloud internet. This research work uses the Canadian Institute of Cyber Security (CICIDS2017) data set to perform the proposed methodology. This data set was selected for classification because it consists of 80 parameters. Other than the benign traffic, as per the tools used, the flow records are labeled as 'Slowloris', 'Slowhttptest', 'Hulk', and 'Begian'. The proposed method shows good results in terms of accuracy, precision, selectivity, sensitivity, and confusion matrix (C.M.). The presented method shows an accuracy of 99% and the other parameters are discussed in the simulation and result section.