Cloud Security using Machine Learning

Abstract

Cloud computing is quickly gaining in popularity and utilisation. Several businesses are investing in this subject, either for their own benefit or as a service to others. The creation of numerous security risks for both industry and consumers is one of the consequences of Cloud development. Machine Learning is one of the methods for securing the cloud (ML). On the Cloud, machine learning techniques have been employed in a variety of methods to avoid or attack detection and security flaws. We conduct a SLR (Systematic Literature Review) of Machine Learning and Cloud Security methodology and strategies in this study. The results of the SLR were divided into three key study categories after we reviewed 63 relevant studies: (i) the many forms of Cloud security threats, (ii) the machine learning approaches employed, and (iii) the performance results. Eleven cloud security areas have been identified. Furthermore, with 16 percent and 14 percent usage, The most common Cloud security threats are distributed denial-of-service (DDoS) and data privacy topics. On the other hand, we discovered 30 ML approaches, Some of them were hybrids, while others were standalones. In both hybrid and standalone models, SVM is the most extensively used machine learning method. In addition, In order to illustrate the efficacy of their suggested model, 60 percent of the papers compared it to other models. There were also 13 other evaluation indicators specified.The rate of true positives is the highest widely used statistic, whereas training time is the least widely utilised. Finally, The most popular are KDD and KDD CUP'99widely utilised datasets research that are relevant, out of a total of 20.