Toward a Mapping Cost Minimization Model under Delay and Resource Constraints in NFV

摘要

Network function virtualization (NFV) has improved the way Telecommunication Service Providers (TSP) use to design their network by providing the required flexibility of the network and permitting to facilitate the reach of ever-increasing demand for high quality of services. However, as NFV is an emerging technology, many challenges still need to be addressed. This paper focuses on one of the most important issues from those challenges, namely resource allocation. A two-plan architecture is proposed in order to well explain the service resource allocation process in NFV. So as to show the impact that mapping process has on operators' CAPEX, OPEX and service performances, we also model the mapping cost minimization problem under delay and resource constraints as an integer linear programming problem. We afterward propose the Cost-Aware Service Execution Algorithm (CASEXA) to present the service execution process scenario. Simulation results show the effectiveness and the stability of our approach on mapping costs and delay when the number of service requests being simultaneously mapped increases, compared to the random mapping approach.