Performance evaluation of Software-Defined Network (SDN) controllers using Dijkstra’s algorithm

Abstract

The advancement of technology, specifically the development of information technology (IT) has had a great influence on everyone life. Due to the rapid development of technology, the traditional network architectures are no longer adequate for the nowadays communication, education, businesses, carriers, and end-users, etc. This indeed is a serious issue, as every organization, enterprise, and educational institute, etc. wants to keep itself on top and to facilitate their customers as much as possible by providing high quality services with minimum delay, etc. In order to tackle this problem, a new emerging network architecture known as software defined networking (SDN) may be employed, as it is more interactive, more flexible, controllable, scalable, and programmable. A network system has two main planes known as control and data planes. The SDN architecture/design separates the control and data planes, legalizes network information and state, and keeps the network infrastructure out of the applications. In contrast, a network may reach a point where the computer or network resources restrict the data flow governed by the bandwidth. In SDN architecture, the controller is the most significant and central component, and large SDN networks might have numerous controllers or controller domains that share network administration. Due to the obvious importance of controllers, different studies have been conducted to compare, test, and assess their performance. This study examines the implementation of Dijkstra’s algorithm using two of the most important SDN open-source controllers (POX and RYU), which permits packet acquisition and transmission between end devices via the network’s shortest and/or lowest load pathways. Performance of the two utilized controllers is measured in terms of various quality of service (QoS) metrics such as throughput, packet delivery ratio, jitter, and packet loss on a specific network architecture under a variety of workloads. Further, we developed a customized network topology using the Mininet which is an emulator tool for SDN. In addition, we also measured the performance of the utilized controllers in terms of various QoS metrics, using the Mininet tool. The experimental results show that the proposed system attained promising results in terms of all QoS metrics.