Machine Learning Algorithms for Cloud Computing Security: A Systematic Review

Abstract: Cloud computing, also known as CC, refers to the on-demand availability of network resources, most notably data storage and processing power, which does not require any additional or direct administration on the part of the users. CC has recently surfaced as a collection of public and private data centres that provide customers with a unified online platform regardless of location. The term "edge computing" refers to a new computing paradigm that moves computation and information storage closer to the end users of a network to speed up reaction times and increase available transmission capacity. Mobile processing Cloud (MCC) is an approach to mobile application delivery that uses distributed processing. The rapid acceptance of computing models is slowed down by the fact that cloud computing (CC) and edge computing (edge computing) have security issues. These security issues include a vulnerability for clients and association acknowledgment. The study of computer techniques that can learn and become more efficient on their own through exposure to new data is known as machine learning (ML). In this review article, we present an analysis of CC security threats, issues, and solutions that utilized one or more ML algorithms. These solutions were implemented to combat these threats. We look at the various machine learning techniques, such as supervised, unsupervised, semi-supervised, and reinforcement learning, that are utilized to solve the problems associated with cloud security. After that, we evaluate the effectiveness of each method by contrasting its characteristics, as well as its positive and negative aspects. In addition to this, we outline potential future research directions to protect CC models.

Keywords: cloud security; security threats; cybersecurity; machine learning; network-based attacks; storage-based attacks.