Machine Learning to Enhance Security in Cloud Computing: A Systematic Literature Review

The usage and popularity of Cloud Computing is continuously growing. Cloud Computing can help users and businesses improve their productivity without worrying about maintaining the software. Moreover, Cloud Computing offers users scalability to meet their needs. Both users and businesses are investing in Cloud Computing. However, despite the advantages that Cloud Computing has, it has a weakness. Cloud Computing can be vulnerable to attack. The rise of the Cloud has led to several security issues which security is one of the important aspects of Cloud Computing to protect. Machine Learning comes as the solution to help prevent attacks in Cloud Computing. The objective of the paper is to review the security of Cloud Computing and analyze Machine Learning techniques that might have the potential to be used. This paper has reviewed 30 relevant studies, and the results are categorized into three main research areas which are Cloud security areas, the potential of ML, and ML techniques used. This study has defined Cloud security areas, with DDoS being the most mentioned. To help protect security, the chosen ML model will learn and be trained from historical anomaly data to prevent attacks. This study found 14 Machine Learning techniques discussed. It has been discovered that a Machine Learning model can be trained using historical anomaly data, which is one of the methods for helping in attack prevention. Additionally, it was discovered that Random Forest, KNN, and SVM are the most often used techniques for anomaly identification.