TROD: Throughput-Optimal Dynamic Data Traffic Management in Software-Defined Networks

In this paper, we study the problem of dynamic data traffic management in software-defined networks (SDNs) in the presence of Internet-of-things (IoT) devices, where IoT devices act as the data traffic sources. In the existing literature, researchers focused on efficient utilization of the ternary content-addressable memory (TCAM) in data traffic management. However, in the presence of IoT-devices, the unbalanced data traffic in SDN deteriorates the overall network performance. Hence, there is a need to design a throughput-optimal dynamic data traffic management scheme for SDN in the presence of IoT-devices, while minimizing the network delay. In this work, we propose an extensive game theory-based scheme, named TROD, for dynamic data traffic management. Additionally, we prove that the dynamic data traffic management, while minimizing the delay and maximizing the throughput of SDN in the presence of IoT-devices, is an NP-complete problem. Hence, we use an evolutionary game theoretic approach to obtain a sub-optimal problem. Thereafter, using linear programming, we obtain the optimal time division matrix, which ensures optimal throughput and delay in SDN. Through simulation, we observe that using TROD, proper distribution of data traffic among the available switches is ensured, and the volumetric overhead per switch reduces by 23.4-29.7%.