Application of Smell Detection Agent Based Algorithm for Optimal Path Identification by SDN Controllers

Software Defined Networking separates the control plane and data plane with which the switches and routers become simply packet forwarding devices. The decision related to the path to be taken by the packet from the source to the destination is taken at the control plane. Thus the SDN controller has to identify the optimal path for the packets. Many of the SDN controllers use Dijkstra's algorithm for computing the shortest path and subsequently update the data plane devices. Many path computation algorithms including bio-inspired algorithms are published and are in use today in computer networks. In this paper, a novel bio inspired algorithm namely Smell Detection Agent based path computation algorithm is applied and studied for its performance in comparison with Dijkstra's algorithm, Extended Dijkstra's algorithm and the most commonly used bio inspired algorithm based on Ant Colony Optimisation. The Smell Detection Agent based algorithm inspired from the dog's smell detection capability for tracing and reaching a destination is found to be very useful and providing better results compared to the other algorithms.