Energy-Saving and Resource-Efficient Algorithm for Virtual Network Function Placement With Network Scaling

Abstract

As an significant paradigm change in network service provisioning, network function virtualization (NFV) can accommodate sophisticated and heterogeneous network services by placing miscellaneous virtual network functions (VNFs) on the specific software-based middleboxes. However, with every-growing energy and construction cost, network operators still face several challenges in terms of adaptation to fluctuant network load and the reduction of energy consumption. Specifically, it is still an open problem to resource-efficiently place VNFs on substrate network while saving network's energy consumption. In this article, the problem of VNF placement is investigated for the optimization of energy consumption and resource utilization. We first present the formulations of power and resource utilization, and then we formulate the VNF placement problem as a binary integer programming (BIP) optimization problem. A novel VNF placement algorithm, named as VPANS, is designed to achieve the efficient VNF placement solution based on matching theory and the dynamical adjustment of the substrate network's scale while it is redundant or inadequate for real-time load. The performance evaluation shows that compared with other three contrast algorithms, our proposed VPANS algorithm not only achieve a maximum power gain of 34.67 percent, but also achieve a better resource utilization and revenue.