NFV-Enabled Network Slicing

Tachun Lin, Zhili Zhou

Abstract: Next generation (5G) networks are expected to support multiple types of services and various customer segments, which can be realized through the network slicing and network function virtualization (NFV) techniques. Through these techniques, a single physical network can be partitioned into multiple virtual networks providing cost-effective, flexible, and on-demand networking services - the requirements of 5G networks. In this paper, we study the NFV-enabled network slicing problem with limited physical resources. We first present the mathematical formulation to generate individual NFV-enabled network slices as the candidates, followed by the network packing framework addressing users' requirement of multiple network slices. We propose both the mathematical formulation and heuristic algorithm based on column generation as the solution approaches. Our evaluation results show the saving on resource consumptions through network slicing.