A Hybrid Optimization Algorithm Based on Ant Colony and Particle Swarm Algorithm to Address IP Traceback Problem

Abstract

The Internet today is highly vulnerable to security threats. The rate of cybercrime has increased proportionately with its usage. Out of numerous possible attacks, the most precarious is Denial of Service (DoS) attack. In DoS the attacker uses the vulnerabilities of compromised hosts in a network and create an attack network called Botnet. The identity of the bots is disguised by using fake source addresses in Internet Protocol (IP) header known as address spoofing. Further, the stateless nature of IP does not allow verification of source address thus making the attack easier. The best way to handle DoS attacks is to reach the source of the attack and block it. IP traceback is a proactive and effective approach to detect the origin of the DoS attack. Once attack origin is detected attack can be blocked, routine network traffic can be restored, chances of future attacks can be prevented and most importantly the responsible attacker can be brought in front of the law. The technique of backtracking for finding an anonymous attacker on a vast network is a complex combinatorial optimization problem, which falls under NP-hard category. In this paper, we have proposed a hybrid approach by integrating Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO), to find the efficient solution of IP traceback problem. The main focus of our work is to increase the convergence rate and further reduce the computational complexity of ACO algorithm by combining the distance-based search technique used by ACO with particle velocity based search used by PSO algorithm. The performance of proposed algorithm is evaluated by simulating it on network simulator 2 and the results show that the method can successfully and efficiently detect the DoS attack path with reduced convergence time and computational complexity.