A chaotic grey wolf controller allocator for Software Defined Mobile Network (SDMN) for 5th generation of cloud-based cellular systems (5G)

摘要

There has been a significant increase in mobile data traffic in recent years and by the advent of the loT (Internet of Things), it will continue to rise. Since the current wireless systems cannot handle this amount of traffic, the next generation of the mobile standard is planning to overcome the obstacles by using new technologies such as Cloud Computing, Software Defined Networking (SDN), and Network Function Virtualization (NFV). Cloud Computing will provide sufficient resources, required for the mobile networks, to handle the ever-increasing amount of traffic. Additionally, using SDN as well as NFV in cloud datacenters can lead to better manageability in the networks and easier development of network applications in the future. Since cloud datacenters as well as mobile networks are inherently large-scale networks, a single SDN controller (traditional SDN) cannot handle decision-making processes of these networks and control plane must be distributed between several software SDN controllers. Therefore, allocating SDN controllers efficiently in these networks becomes important, which can lead to lesser energy consumption, lesser CAPEX, and lesser OPEX. In addition, the highly variable pattern of these networks makes it necessary to allocate controllers dynamically. However, by the growth of the networks, dynamic allocation of controllers can turn into an NP-Hard problem and using metaheuristic algorithms instead of deterministic ones to solve this problem can show better results as well as better computation time. In this study, we proposed a metaheuristic-based framework for dynamic controller allocation for the 5th generation of mobile technology (5G). We simulated our framework in MATLAB, compared our framework with static allocation technique, and compared our algorithm with another metaheuristic algorithm PSO. Simulation results show that our algorithm ameliorates computation time and the computed solutions for different Quality of Services (QoS) are feasible, acceptable, and accurate. (C) 2017 Elsevier B.V. All rights reserved.