Hypervisor Based Intrusion Detection Using Enhanced Radial Basis Neural Network on Cloud Environment

Cloud computing has grown to be a significant component of the IT sector in the current digital economy. Despite its advantages such as reduced capital costs and operational costs, there are some constraints that limit its use. Previous research has proposed various intrusion detection systems and authentication systems, but they fail to detect intruders. Hypervisor-based intrusion detection systems have recently been developed that monitor both virtual machines (VM) and the host simultaneously. Furthermore, it has become popular because it provides a comprehensive view of the cloud environment and allows for more accurate threat detection. Considering these, we propose novel hypervisor-based intrusion detection in this work. Initially, we collect intrusion data and pre-process it. The best features are then chosen using an adaptive artificial flora algorithm (AAF), which lowers processing time and complexity. Then, the selected features are passed to the ERBNN classifier to classify the data as normal or intrusive data. For validation of the proposed method, KDDCup-99 and NSL-KDD datasets are used. The performance of the suggested technique is validated based on different metrics and the proposed system is implemented using Python.