Cost-Efficient Virtual Network Function Placement and Traffic Steering

Yun Liu,Jianing Pei,Peilin Hong,Defang Li
DOI: https://doi.org/10.1109/icc.2019.8762060
2019-01-01
Abstract:Benefiting from Network Function Virtualization (NFV), Service Function Chain (SFC) that is composed of a set of ordered Virtual Network Functions (VNFs) has become a popular network service pattern. One of the most important issues for Internet Service Providers (ISPs) is to determine the optimal placement of VNFs and the traffic steering of SFC to optimize the total network cost while guaranteeing resource constraints. In this paper, we study the cost-efficient VNF Placement and Traffic Steering (VNFP-TS) problem. First, we formulate the problem as a Binary Integer Programming (BIP) model aiming to minimize the node running cost, VNF placement cost and communication cost jointly. Then a heuristic Dynamic Programming based Cost Optimization Algorithm (DP-COA) is proposed to divide and conquer the problem. Finally, we evaluate the proposed algorithm by numerical simulations and confirm that DP-COA can achieve high network performance in terms of acceptance rate of SFC Requests (SFCRs) and throughput, and efficiently reduce the total network cost comparing with algorithms in existing literatures.
What problem does this paper attempt to address?