Optimizing the service capacity of SDN-based cellular networks with service chaining and NFV

In this paper, we investigate the service capacity of SDN-based cellular networks with service chaining and network function virtualization. Due to service chaining requirements, the service capacity is different from the well-known maximum-flow-minimum-cut capacity. To maximize the service capacity, we take a novel approach of traffic engineering. In particular, we formulate and solve a linear optimization problem to obtain optimal service paths for flows and an optimal bandwidth allocation for virtual machines. We also include numerical results that justify the usage of the proposed approach.