An enhanced intrusion detection framework for securing network layer of cloud computing

Virtualization is a key enabler of Cloud computing. Due to the numerous vulnerabilities in current implementations of virtualization, security is the major concern of Cloud computing. In this paper, we propose an enhanced security framework to detect intrusions at the virtual network layer of Cloud. It combines signature and anomaly based techniques to detect possible attacks. It uses different classifiers viz; naive bayes, decision tree, random forest, extra trees and linear discriminant analysis for an efficient and effective detection of intrusions. To detect distributed attacks at each cluster and at whole Cloud, it collects intrusion evidences from each region of Cloud and applies Dempster-Shafer theory (DST) for final decision making. We analyze the proposed security framework in terms of Cloud IDS requirements through offline simulation using different intrusion datasets.