Improving the Load Balancing in Software Defined Networks

Abstract

Computer networks design and management has become more innovative throughout the past few years with the aid of Software Defined Networks (SDN). Although this technology had a sudden outbreak, it is actually part of a long history in attempt to make computer networks programmable. SDN for short, are ready to de-complicate computer network design and management through presenting a clean and open interface between devices which are run on networks and the programs controlling these devices. The best describe of SDN is a network connection method in which the control plane is separated from the hardware, unlike traditional network equipment in which control and data planes co-exist in the same device. The control plane was abstracted to a software application, called the Controller, which is considered as network operating system. Subsequently, control plane has been centralized. Furthermore, the Controller has a comprehensive view of full topology of the network that it runs; thus, it is responsible for traffic routing and load balancing of entire network.

SDN have shown great advantages in many respects in contrast to traditional networks. However, the distribution of traffic routing in SDNs negatively affects efficiency and offers many challenges. For example, unequal load distribution gravely affects network performance, so many load balancing techniques have been used to improve SDN efficiency. this paper is a presentation of a mechanism for load balancing in traffic routing inside networks. This mechanism is an application of light and cost-efficient traffic routing. The Mininet emulator was used for execution of the network and connecting it to the Controller HPE VAN. The algorithm was applied using python, experiments were done using commands of iperf tool.