Two Weighting Methods for Software-Defined Network Controllers Ranking as a Multicriteria Decision-Making Problem

Abstract

SDN is new paradigm to make the network manageable for network administrators. The controlling functions are managed by controller which is the brain of SDN design and has a wide-view of network topology. SDN controllers have different features, so choosing a suitable controller for the network is a crucial challenge for administrators. Although there are many controller proposals for different kinds of networks in the literature, but there is a limited comprehensive quantitative analysis for them. In this work, 14 of widely-used decentralized SDN controllers are evaluated and compared in terms of 8 criteria that each controller should possess. The comparison is done using 3 MCDM methods: MARCOS, TOPSIS, and MAIRCA for a comprehensive comparative study. Moreover, two weighting methods: Entropy and MEREC are applied to assign the importance degree for the criteria. This study presents a novel approach to evaluate SDN controllers using a combination of three MCDM methods and two weight determination methods, resulting in six different ranking solutions. The most notable finding is that all six solutions identified the same two SDN controllers, ONOS and ODL, as the best performers. These results demonstrate the effectiveness of using MCDM methods for comparing and evaluating SDN controllers.

DOI: 10.61416/ceai.v26i2.8924