ICMPv6-based DDoS Flooding-Attack Detection Using Machine and Deep Learning Techniques

Abstract

IPv6 was created to resolve the problem of adopting IPv4 addresses by providing many address spaces. Currently, security is becoming an increasingly important concern in exploiting networks and reaping the benefits of IPv6. ICMPv6 is a key protocol in IPv6 implementation that is utilized for neighbor and router discovery. However, attackers can use this protocol to deny network services using ICMPv6 DDoS flooding attacks, which reduce network performance. DDoS is a difficult challenge on the internet since it is one of the most common attacks impacting a network, causing enormous economic harm to people as well as companies. This paper provides an intelligent ICMPv6-based DDoS flooding-attack detection system based on an artificial neural network to address this issue. This study additionally investigates and examines the suggested framework's detection accuracy. Using real datasets, we illustrate the efficiency of our methodology. To validate our system, we chose different machine learning algorithms and compared their outcomes. The findings show that the proposed framework can identify ICMPv6 DDoS flood assaults with detection accuracy rates of 99.98% for the first dataset and 85.91% for the second dataset.