Performance Evaluation Of Open Network Operating System And Floodlight Controllers In Software Defined Network Based Internet Of Things

Abstract

The increase in the number of smart connected things has affected the performance of IoT applications, the problem statement of this research is that organizations requirements are becoming heterogeneous and stringent in terms of bandwidth, response time, throughput. Software Defined Networking (SDN) came as a solution that enables the network to adapt its behavior dynamically according to the traffic type which will increase the performance. In this research, the performances of the most used SDN controllers (ONOS, Floodlight) are analyzed and compared. Controllers analyzed based on some metrics such as throughput, delay, bandwidth using different topologies (single, linear, tree and SDN based IoT) using IoT traffic flow transmitted through MQTT protocol. ONOS shows throughput of 223.985, 340.3, 69.24 and 222.6 MB and floodlight shows throughput of 207.46 ,138.26, 255.33 and 228.5 MB in single, linear, tree and SDN-based IoT topologies respectively. ONOS shows bandwidth of 1.876, 2.85, 587.55 and 1.77 Gbit/sec and floodlight show bandwidth of 1.741, 1.4, 2.157 and 1.86 Gbit/sec in single, linear, tree and SDN-based IoT topologies respectively. ONOS show delay of 3.442, 7.15, 8.3and 0.062 msec and floodlight show delay of 2.85, 4.9, 5.28 and 0.039 msec in single, linear, tree and SDN-based IoT topologies respectively.