An Adaptive Multi-layered Approach for DoS Detection and Mitigation

Abstract

A Denial of Service (DoS) attack imposes a heavy load on a system rendering it unavailable to the benign traffic. One of the most popular approaches to carry out the attack is to send a multitude of requests to the targeted site or network, causing the host or network to become unable to reply to the benign traffic or to respond slowly. The complexity and frequency of these attacks have been increasing in recent years. Hence, there is a need to design an efficient system that would detect any suspicious activity in the network and dispatch a timely and appropriate response to counter the same. In this paper, different design models and implementations of contemporary intrusion detection systems have been reviewed and analyzed for shortcomings. A multi-level design for an Intrusion Detection and Prevention System (IDPS) that aims to efficiently detect the DoS attack with minimal response time and high accuracy has been proposed. A UDP flood is simulated inside a virtual network environment to emulate the attack and the results demonstrate the successful detection and mitigation of the DoS attack.