Optimization of Network Performance in Complex Environments with Software Defined Networks

Abstract

Software-defined networks (SDN) have emerged as a promising approach to address the limitations of conventional networks. Its architecture can be implemented using either a single controller or multiple controllers. Although a single controller is inadequate for managing networks, deploying multiple controllers introduces the challenge of controller placement (CPP) in a network environment. To address these issues, this study presents a Software Defined Networks-Fault-Tolerant Method (SDN-FTM) where, in the event of a network failure, the SDN controller automatically reroutes traffic through an alternate, pre-configured network path, thereby maintaining uninterrupted service. The proposed SDN-FTM was tested and evaluated in real-time using Mininet simulation tools on a real-life small scale network data from tracking unit department in Walter Sisulu University (WSU), with a focus on performance measures such as latency and throughput. From the result obtained, the proposed method produced throughput and latency on Ryu with 2.15m/s and 18408m/s respectively. Furthermore, the findings indicate that Ryu controllers generally outperform OpenFlow controllers in terms of throughput, while OpenFlow controllers exhibit lower latency. The proposed method demonstrates significant improvements in network management by providing a robust solution for maintaining high network availability and performance in the presence of faults

Keywords

• software-defined networking, controller, latency, throughput