Towards an efficient VNF placement in network function virtualization

Abstract

Network Function Virtualization (NFV) decouples network function (also called middlebox function) software from specified appliances onto general shared servers. Thus, it has been being regarded as a promising technology to overcome high Capital Expenditures (CAPEX) and Operational Expenditures (OPEX) on the middlebox deployment and maintenance. In NFV, the network function deployed on servers with virtual machine is named as the virtualized network function (VNF). One critical issue is VNF placement for provisioning service function chains (SFC), which deals with the resource allocation to VNFs and routing path between them. The problem is inherently NP-hard. Current VNF placement algorithms do not scale with respect to the network size, leading to these algorithms not applicable in large-scaled scenarios where VNFs have to be placed in a timely way. Therefore, this paper aims to solve the problem of VNF placement in a scalable way. We attempt to narrow the target searching space of VNF placement by introducing a smaller accessible scope where the locations of VNFs are confined. The accessible scope constraint is generic for different conventional VNF placement algorithms, which can be used in conjunction with existing algorithms to improve time efficiency. Two algorithms to be evaluated are chosen to run with the accessible scope constraint under medium and large scales of scenarios. Results show that the algorithms with the constraint of accessible scope have significant time efficiency improvements especially in large-scale scenario and the solution quality is at least comparable.