Machine vs Non-Machine Learning Approaches to Cloud Security Solutions: A Survey Gopal Krishna Shyam #1 , Srilatha Doddi *2 # School of Computing and Information Technology, REVA University, Bangalore 560064, IN 1 gopalkrishnashyam@reva.edu.in * School of Computing and Information Technology, REVA University, Bangalore 560064, IN * Department of CSE, Sreenidhi Institute of Science & Technology, Hyderabad 501301, IN 2 doddisrilatha@gmail.com Abstract—Cloud computing is a trending paradigm that provides both physical and logical computational resources as services over the Internet. The basic advantages of the cloud are, a reduction of IT organization’s infrastructure cost, flexibility to access and use the services. Regardless of its advantages, it has raised several security concerns such as data availability, data privacy, data location, authentication, authorization, access control, network security, web security, and virtual machine security etc. which may potentially hamper its growth. In recent years, the expansion of several types of dynamic threats such as data breaches, account hijacking, insecure interfaces, advanced persistent threats, shared technology vulnerabilities, and distributed denial of service attacks target the cloud to disrupt cloud services and can compromise security. To tackle several security issues, solutions can be provided through a set of control based technologies such as next generation firewalls, cryptography techniques, intrusion detection systems, software defined networks, machine learning techniques etc. In this paper, we focus on comparative analysis of several cloud security issues through machine learning and non-machine learning approaches. Some open challenges for further research have also been suggested. Keyword-Attacks, Cloud computing, Cloud security, Intrusion detection system, Machine learning. I. INTRODUCTION Cloud computing is a trending model that provides computational resources such as compute, storage, network, operating system, application development and deployment environment, application software, media etc. as services over the Internet. Cloud Computing is gaining popularity due to its basic advantage of IT organizations’ infrastructure cost savings [1]. Cloud computing has become a social phenomenon used by most people every day as they are migrating the applications and data from a local machine to a remote machine. The basic idea behind cloud computing is layered and flexible architecture. The reference model of cloud computing is presented in the Fig. 1. The National Institute of Standards and Technology (NIST) defines five essential characteristics, four deployment models, and three service models of cloud computing. The five essential characteristics of cloud are broad network access, resource pooling, on-demand self service, rapid elasticity and measured service and four deployment models of cloud include public cloud, private cloud, hybrid cloud and community cloud. The infrastructure-as-a-service (IaaS), platform-as-a-service (PaaS), and software-as-a-service (SaaS) are the three service models of cloud. In IaaS layer, infrastructural services like compute (i.e. virtual machines), storage, and networks can be delivered to the end users. The PaaS layer provides databases, software libraries, application development and run-time environments for developing, testing, delivering and managing the application software’. In SaaS, application software’ are delivered to the end users without installing these on local machines. In a public cloud, the resources are openly accessible to several organizations. Whereas, in a private cloud, the resources are used exclusively by a single organization. A hybrid cloud is a mixture of private and public cloud that provides data and applications to be shared between several organizations. Whereas, in a community cloud, the cloud infrastructure supports a specific community that has shared mission, policies and service requirements. ISSN (Print) : 2319-8613 ISSN (Online) : 0975-4024 Gopal Krishna Shyam et al. / International Journal of Engineering and Technology (IJET) DOI: 10.21817/ijet/2019/v11i2/191102063 Vol 11 No 2 Apr-May 2019 376