An optimized approach for detection and mitigation of DDoS attack cloud using an ensembled deep learning approach

Abstract

As cloud computing gains in popularity, safety becomes an increasingly important consideration. One of the most challenging issues in cloud computing is the detection of Distributed Denial-of-Service (DDoS) attacks (Gupta, B. B., et al., 2009). One of the most crucial aspects of cloud architecture is the ability to provide self-service whenever it is needed. Applications built on the cloud computing model are available on demand and at low cost. As cloud computing grows in popularity, so too is the amount of cyberattacks aimed against it. One such attack is a Distributed Denial of Service attack, which is designed to overload the cloud's hardware/software, resources, and services, making them difficult to use for everyone. The difficulty of this assault stems from the fact that it can overwhelm the victim's ability to communicate or compute in a short amount of time with little to no notice. It's getting harder to spot and stop these assaults as they get more sophisticated and more numerous. Several Machine Learning methods, including Logistic Regression, K-Nearest Neighbors, Support Vector Machine, Decision Tree, Naive Bayes, Multi-layer Perceptron, XGBoost, and SGD have been implemented for accurate DDoS flooding attack detection. When compared to current methods, the suggested strategy of utilizing deep learning with Quadratic discriminant appears to result in higher accuracy. There is also a thorough comparison and evaluation of the abovementioned algorithms with respect to the accuracy measures used.