A machine learning application for reducing the security risks in hybrid cloud networks

Abstract

Cloud computing facilitates the enormous support of the public, business and emerging applications such as healthcare and nature disasters. Based on their services and characteristics, it can be classified as private cloud and public cloud. In this scenario, we need these two kinds of cloud services for an organization to provide the better service to the society. For that purpose, introduced a new cloud service called hybrid cloud service which is the combination of both private and public cloud. Security to the cloud environment is a challenging issue today especially for the hybrid cloud due to the combination of both. Many security mechanisms available in the literature but these are not achieved the sufficient level of security. For this purpose, we propose a new machine learning application for providing security to the hybrid cloud networks while storing the data and retrieving or accessing the data from the cloud. This proposed algorithm combines an existing Enhanced C4.5, newly proposed deduplication processing algorithm and the newly proposed dynamic access control mechanism. Moreover, we introduce a new deduplication processing algorithm for secure storage and retrieval without duplication or redundancy. In addition, a newly proposed dynamic access control mechanism called Dynamic Spatial Role Based Access Control Algorithm is also used in the proposed security framework. The proposed security framework has been implemented and also evaluated that the security level of the hybrid cloud while storing, retrieving and accessing the data from the cloud database.