SLAMHHA: A supervised learning approach to mitigate host location hijacking attack on SDN controllers

Abstract

Current era of networking world witnesses an increase in the number of devices that have intelligent component embedded in them, which leads to an unmanageable state. This has lead to a steady shift towards using Open Software Defined Networks to reduce operational expenditure. The whole of the intelligence lies in the central controller which obviously is the single point of failure. The attackers find it easy to bring down the whole network by saturating the control plane of the SDN which eventually leads to denial-of-service (DoS) to the data plane, through host location hijacking attack. The various strategies proposed to defend the SDN controller from host location hijack attack includes authentication of host which incurs overhead. In this paper we propose SLAMHHA, a Supervised Learning Approach to Mitigate Host location Hijacking Attack. SLAMHHA is implemented in the SDN controller which monitors the legitimacy of the hosts and identifies the clandestine users impersonating the hosts in the data plane. The SLAMHHA algorithm can be implemented in either of the two SDN controller setup (i.e.) in both centralized and decentralized controller setup, been used to set the flow rules and monitor the underlying network. MININET has been used to test the efficiency of the SLAMHHA algorithm. SLAMHHA algorithm was implemented in the POX controller. Numerical results show that SLAMHHA incurs less overhead in terms of CPU and memory consumption when compared to the authentication method. This algorithm blocks the attack within 3 seconds when 100 hosts are impersonated to perform DoS attack. Thus our proposed SLAMHHA algorithm mitigates the host location hijacking attack with less overhead.