An Efficient Load Balancing Mechanism in Software Defined Networks

Abstract

The growth of the network, increase in users, as well as surfacing of new technologies, such as cloud computing and big data has posed difficulties in managing traditional networks. To address this problem, Software Defined Network has been proposed which tends to govern the network in a better way. Traffic management technologies include traffic load balancing, QOS based scheduling. To fit in the prerequisites of quality of service, load balancing has to be considered as a crucial concept because it assists in controlling and regulating data traffic among numerous resources amplifying the network responsiveness, steadfastness and capacity. Network today needs to manage heavy traffic and serve clients in huge numbers at once. Using multiple servers with a load balancer is one of the solutions to this problem such that the clients can direct the request to load balancer which can further forward these requests to distinct servers after being subjected to load balancing strategy. Load balancing in a conventional network is based on the native information of the network and is nonprogrammable. But, SDN controllers have a global vision of the network and are also programmable which helps them design optimized load balancing mechanisms. In this paper, we have implemented Round Robin load balancing strategy and Random load balancing strategy using an OpenFlow switch. Mininet simulation results using round robin algorithm have led to higher throughput and decreased packet loss when compared to default random scheduling algorithm.