Design of restricted boltzmann machine based secure cognitive protocol for software defined networks

Abstract

Software Defined Networks SDN have replaced the traditional network architecture by separating the control from forwarding planes It provides a global view of network topology and the network control is entirely softwarized for network management The softwarized network control is vulnerable to security attacks according to the functionality of network architecture Distributed Denial of Service DDoS attack is the most common security attack generated by an attacker in order to deny the availability of network resources In SDN architecture control plane acts as a brain of the network and hence it becomes an attractive target for an attacker DDoS attack occurs in various forms according to the functionality of SDN planes According to the Internet Security Report Q2/2016 the impact of DDoS attacks is higher in infrastructure layer The origin of DDoS attacks get varied in SDN based on the layers Attackers insert malicious applications to make controllers learn false information to forward the unknown flows that affect the network performance These kinds of attacks are considered to be a serious threat and it is highly important to secure the network in order to protect the resources such as Central Processing Unit CPU and network bandwidth It is highly essential to detect the attack traffic flows in a dynamic network environment This thesis attempts to design a secure cognitive protocol to detect and defend flooding based DDoS kind of attacks in SDN The main objective of this research work is to detect and mitigate these attacks with the help of an unsupervised Machine Learning ML algorithm and to deploy it in a routing protocol that automatically safeguards network resources The proposed protocol automatically defends against security attacks using trained network metrics and context-aware metrics to identify the pattern of the incoming network traffic flows newline