Resource Optimization and Delay-aware Virtual Network Function Placement for Mapping SFC Requests in NFV-enabled Networks

Abstract

Network Function Virtualization (NFV) enables cloud service providers (CSPs) to flexibly place their network functions on common off-the-shelf servers in the form of virtual network functions (VNF), without incurring costly capital and operating expenses (CAPEX/OPEX). In the NFV-enabled network, service function chains (SFCs) are responsible for accomplishing users' service requests by steering traffic through a set of VNFs in a specified order. Therefore, it is an important but intractable issue for CSPs to devise an optimal VNF placement scheme to enhance network performance and profits. In this paper, we focus on the VNF placement problem for mapping SFC requests (SFCRs) in NFV-enabled networks, considering the delay requirement of SFCRs. To improve resource utilization, we consider the basic resource overheads and sharability of VNF instances. Then we formulate the problem as an integer linear programming (ILP) model, with the purpose of total resource consumption minimization. Afterward, the novel SFCR mapping algorithm (SMA) and VNF request adjustment algorithm (VAA) are proposed to map SFCRs and optimize the placement of VNF requests. Simulation results show that our approach is near-optimal in terms of node resource consumption. Besides, it provides higher performance in terms of node resource consumption, average VNF utilization and the number of activated servers compared with the benchmarks.