Improving the Performance of Software-Defined Network Load Balancer Using Open Flow Based Multi-Controller Topology

Abstract

Due to the emergence of internet of things and 5G networks there's an enormous pressure on the underlying communication networks in terms of demand, performance requirements and dynamic management. To manage incoming traffic, a load balancing technology is assigned to server clusters. In traditional networks routing protocols forward traffic in keeping with the shortest path to reduce cost. This might result in abnormal distribution of traffic causing overloading of communication links. Recently Software Defined Networks (SDNs) has become increasingly popular and potential candidate to beat traditional networks limitations. Software Defined Networks (SDNs), described by an ideal separation of the control and data planes, is being approved as a distinct paradigm for complex network management. In this research work, performance analysis is performed on random, round robin, weighted round robin and least load balancing algorithms in terms of response time /sec transaction rate (trans/sec) and throughput (MB/sec). Moreover a new Open Flow Model based Multi-Controller Topology is proposed and, the proposed topology is able to reduce the response time (sec) by an average of 30.12%, increase the transaction rate (trans/sec) by an average of 39.44% and also increase the throughput (KB/sec) by an average of 10.56% when compared with a single controller topology using random load balancing algorithms in SDN POX controller.