A survey on cloud security threats using Deep learning algorithms

Abstract

Enterprises have transformed to the usage of Cloud computing technology as it provides facilities for storage, infrastructure and software's on demand. Huge amount of confidential data is stored in cloud. The popularity of cloud computing has increased rapidly due to this the main concern is security for data. These data are vulnerable to different types of attacks. One of the way to handle security is using deep learning(DL) algorithms. DL algorithms are used in various ways to detect and prevent the threat. This survey is a systematical analyze of 25 relevant research articles on various cloud computing threats, which has incorporated deep learning algorithms to detect them. The results of the study are categorized into three main research areas: i) various cloud security threats, ii) deep learning algorithms used to overcome the threat, iii) performance measure and dataset used. 5 threats are defined and the most vulnerable cloud security threat is denial of service and distributed denial of service with 43% of detection. 9 deep learning algorithms are used in standalone and hybrid mode for detection of threat. Moreover 9 evaluation metrics are enumerated. In the last from the 6 dataset which is found in the relevant studies NSL-KDD is the most used.