Detecting DNS-DDOS Attacks in the Cloud Computing Services using Machine Learning

Abstract

Cloud computing provides a suitable platform for sharing resources, services, and knowledge. It gives businesses a flexible design that gradually provides an economic model for computing. Because of the widespread use of cloud services by a wide range of organizations, a slew of threats and problems must be addressed. Because cloud storage servers are web-based services, consumer protection, data leakage, and authentication appear to be major challenges in cloud computing environments.There are some security issues like poor IP protection, authorized access, malicious insiders, cyber-attack for analysis There are two ways. When it comes to data-oriented experiments, there are two types: real-time experiments and data-oriented experiments. learning is one of the best technologies to analyze cloud data. This paper implements many security-oriented techniques and conducted a comparative study of different machine learning algorithms, that will predict the classification accuracy. For this analysis, we have considered the CICDDOS-2019 attack dataset and it has different types of attacks. The proposed Intrusion Detection System (IDS) is classified into three parts: pre-processing, feature selection, and classification. The pre-processing applied different techniques: transformation, normalization, and stratified sample dataset. Some features are identified by feature selection techniques and created a new dataset for applying classification algorithms. Finally, we experimented with the comparative study using three algorithms: SVM, NAIVE BAYES, and RANDOM FOREST. The performance metrics prediction accuracy and computing time have been considered to identify effective IDS. Results are presented, conclusions are drawn.