Revealing the Impact of Dynamic Traffic on NFV Networks Planned Under Static Assumptions

Abstract

5G/6G networks offer high-capacity, low-latency and high-speed communications to new services associated with vertical industries like smart cities, automotive and energy sectors. To achieve this objective, new networking and computing paradigms such as Network Function Virtualization (NFV), Multi-access Edge Computing (MEC) and Software Defined Networking (SDN) are required to bring flexibility and adaptability to networks, while optical networks are envisioned to build the backhaul of 5G/6G networks thanks to their high capacity. Current studies on NFV-5G network planning with MEC resources follow either the online trend or the offline trend. In the online trend, resource allocation to service connection requests is determined upon request arrival based on the available resources. Conversely, in the offline trend, network planning is conducted assuming that the expected traffic (or the full set of service requests) is known in advance. However, to the best of our knowledge, previous works have not assessed the performance of these static offline network plannings when operating under dynamic traffic. In this paper, we evaluate the performance of a previously proposed static planning algorithm for NFV-5G networks under dynamic traffic conditions and analyze the impact of dynamic traffic on the request blocking ratio. Moreover, we also examine the performance of the static planning for each individual service type in a dynamic network scenario with three different available services. The simulation results show that overdimensioning is necessary when planning NFV-5G networks statically, and that the overdimensioning should consider different criteria based on the network expected performance objectives.