Controller Selection in Software Defined Networks using Best-Worst Multi-Criteria Decision-making

Controllers are the key component of Software-defined Network (SDN) architecture. Given the diversity of open SDN controllers, the following question arises for the network administrators: How can we choose the appropriate SDN controller? Different characteristics of the controllers have greatly increased the complexity of the right decision. Multi-Criteria Decision-making Methods (MCDMs) is a family of robust mathematical tools to address complex problems regarding multiple objectives. In this paper, we study the most important features of SDN controllers. To this end, we compare the well-known SDN controllers including NOX, POX, Beacon, Floodlight, Ryu, ODL, and ONOS. Leveraging a novel MCDM technique called the Best–Worst Multi-criteria (BWM), we find the most appropriate SDN controller. We solve an optimization problem and evaluate its performance in terms of significant criteria such as throughput and latency. Initial evaluation revealed that the ONOS and ODL have the highest throughput, while the lowest throughputs belong to the NOX, POX, and Ryu. However, final evaluation concerning all criteria confirmed the robustness of the ONOS and the ODL compared to other controllers.