Establishing Trust In Cloud Using Machine Learning

Abstract

In the modern era of digitization, cloud computing is playing an important key role in it. Tons of user data is on the cloud and to manage it with all the odds is a mesmerizing task. All the major community is an on cloud having all their confidential data resides on a cloud and to provide user data integrity. Security is the major concern task for any service provider. In the last few decades research has been motivated to provide security on a cloud with reference to the categorization and classification of security concerns. Growing security risk in the last few years in various components of the cloud is a major concern. Research studies have been motivated to handle risk, threat, and vulnerability imposed within the environment of the cloud. In the cloud, trust is a major concern as a security point of view. In cloud computing, user’s data stored on the remote server which may be operating by others and can be accessed through the internet connection. The facilities provided by the cloud are too attractive for customers but it has distributed and non-transparent nature due to some obstacles using in cloud computing service because users lose their control over the data, and they are sure about whether cloud provider trust or not. So, customers confused with cloud providers regard the trust issue. This paper mainly focuses on establishing trust in the cloud using machine learning methods. There are high chances that data may be lost or compromised. So how people can trust that their data is secure or not on the cloud. Hence, trust is becoming a serious issue. In this research study, the main aim is to minimize the security risk, threat and vulnerability as a trusted perspective in a cloud environment. For this, we have study machine learning methods. In machine learning, there are mainly three methods: supervised learning, unsupervised learning and reinforcement learning. Supervised learning means the system already knows what is the output. If the system gets the desired output then closed the process. Unsupervised learning means the system doesn’t know what is the output, and the last one reinforcement learning, it is based on reward and feedback oriented. We also discuss trust parameters like availability, confidentiality, accuracy, and integrity, resource management, non-repudiation, risk management, protecting communication, hardware security, reliability, and secure architecture.