Identification of Flooding Attack (Zero Day Attack) at Application Layer Using Mathematical Model and Detection Using Correlations

Authors: Hamsini Pulugurtha, V.S. Lakshmi Jagadmaba Paluri

Abstract:

Distributed denial of service attack (DDoS) is one altogether the top-rated cyber threats presently. It runs down the victim server resources like a system of measurement and buffer size by obstructing the server to supply resources to legitimate shoppers. Throughout this text, we tend to tend to propose a mathematical model of DDoS attack; we discuss its relevancy to the choices like inter-arrival time or rate of arrival of the assault customers accessing the server. We tend to tend to further analyze the attack model in context to the exhausting system of measurement and buffer size of the victim server. The projected technique uses an associate in nursing unattended learning technique, self-organizing map, to make the clusters of identical choices. Lastly, the abstract applies mathematical correlation and so the standard likelihood distribution on the clusters and analyses their behaviors to look at a DDoS attack. These systems not exclusively interconnect very little devices exchanging personal data, but to boot essential infrastructures news standing of nuclear facilities. Although this interconnection brings many edges and blessings, it to boot creates new vulnerabilities and threats which might be conversant in mount attacks. In such sophisticated interconnected systems, the power to look at attacks as early as accomplishable is of paramount importance.

Keywords: application attack, bandwidth, buffer correlation, DDoS distribution flooding intrusion layer, normal prevention probability size