A hybrid layered architecture for detection and analysis of network based Zero-day attack

A Zero-day (0-day) susceptibility is an undisclosed computer software or application vulnerability that could be exploited to affect hardware, applications, data, or networks negatively. The main objectives of a Zero-day attack are for hackers or attackers to be able steal sensitive information, legal documents, enterprises data, and other information. We have analyzed the lifecycle of Zero-day vulnerabilities and different detection methodologies. In this paper, we propose a novel hybrid layered architecture framework for Zero-day attack detection and analysis in real-time, which is based on statistics, signatures, and behavior techniques.  To enhance our architecture, we used an SVM approach in order to provide unsupervised learning and minimize false alarm detection capabilities.