A Study On The Impact Of Network Topologies On The Performance Of Openflow Controller

ABSTRACT

Software-Defined Network (SDN) is a new and booming network paradigm that promises to improve network performance. This is achieved by the separation of the control plane and the data plane. The traditional network faced a lot of challenges particularly on how to improve network performance, manual configurations of the network devices, and many more others. SDN overcome all these challenges by introducing the SDN controller that controls the entire network. This research work studied the impact of SDN network topologies on the performance of the OpenFlow controller by employing the Floodlight controller which is java based and supported by various network engineers from Big Switch Network. The floodlight controller is among of the OpenFlow controller and is not only the controller but also the correction of applications that offer tremendous performance. This study also employed three different network topologies, 2-D toroid topology, Generalized Star topology, and linear topology. Mininet simulator and emulator were used to provide an environment for the simulation that emulated switches, hosts and the controller. Two network parameters were used for this study, throughput, and delay. The results showed that Floodlight controller performed better on a generalized star topology in terms of throughput over the rest of all network topologies that were employed in this study, this is because the generalized star topology offers multiple paths for a network packet to transverse over and reach the destination. In terms of delay, linear and generalized topology showed good results, but 2-d toroid topology has a higher delay than the two network topologies when the Floodlight controller was employed.