VNF Placement Strategy for Availability and Reliability of Network Services in NFV

Network Functions Virtualization (NFV) is a new networking platform for fast and economic provisioning of Network Services (NSs) to the end users. NS in NFV is provided through a chain of Virtual Network Functions (VNFs) that are hosted on the underlying physical computing machines. Actually, placement of the VNFs on the physical machines has a big impact on the quality of the provisioned NS. In this paper, we study the impact of the VNF placement on both the availability and reliability of NS in NFV platform. In addition, we propose an optimized VNF placement strategy from the availability and reliability perspectives of the NS. We map the placement problem to Integer Linear Programming (ILP) optimization model. We build a prototype to evaluate the proposed strategy, and compare it with other existing placement strategies named: BinPack, Random, and Sparse. The results of the experiments show the ability of the proposed strategy to enhance the availability and reliability levels of NS compared to the other strategies.