Optimized deep learning‐based attack detection framework for secure virtualized infrastructures in cloud

Abstract

AbstractVirtualization is the foundation of cloud computing process and allows more effective utilization of physical computer hardware. Virtualization technologies are used in various tasks for the improvement of operating systems. Virtualization is meant for enabling the entire representation of cloud computing systems on normal speed. Still, the scale of virtualized environments and its complexity is found to be difficult in this process. This paper introduces a security analytics method in virtualized infrastructure for detecting the attacks of cloud computing. As the work relies on big data issues based on the features of network behavior, the detection phase is processed under two major phases: (1) feature extraction and (2) classification. In the feature extraction phase, proposed holoentropy features are extracted along with exponential moving average features. These extracted features are then subjected to the classification process, where the optimized DBN is used to detect the presence of an attack in network. To make the detection more accurate, the weights are optimally tuned by a new hybrid elephant monarch algorithm that helps in DBN learning. At last, the performance of the proposed work is computed over other traditional models in terms of certain measures.