Logo image
Fair VNF Provisioning in NFV Clusters via Node Labeling
Conference paper

Fair VNF Provisioning in NFV Clusters via Node Labeling

Tzu-Wen Chang, Tung-Wei Kuo and Ming-Jer Tsai
2020 IEEE Global Communications Conference, GLOBECOM 2020 - Proceedings, Vol.2020-January, 9347991
12/2020

Abstract

Dominant Resource Fairness Max-Min Fairness Multi-Resource Allocation Network Function Virtualization Media Technology Modeling and Simulation Instrumentation Artificial Intelligence Computer Networks and Communications Hardware and Architecture Software Safety Risk Reliability and Quality
We study fair multi-resource allocation in Network Function Virtualization (NFV) clusters, where the relative amounts of (multiple) resources allocated for a virtual network function (VNF) can be flexibly adjusted. In NFV clusters, the fairness across users can benefit from the flexibility of the multi-resource allocation for VNFs, but we also have to address a research challenge: What relative amounts of resources should be allocated to a VNF? Although many studies address fair multi-resource allocation in the literature, they all assume that the relative amounts of resources allocated for a VNF are pre-determined and fixed, which would lead to the poor fairness across users. In this paper, we make the first attempt to propose an algorithm to allocate resources to users under the circumstance of flexible multi-resource allocation for VNFs. Our algorithm is shown to achieve max-min fairness and satisfy two beneficial properties of fair multi-resource allocation: Pareto efficiency and envy-freeness. Simulations also show our algorithm can allocate resources in a fair and efficient way in NFV clusters.

Metrics

1 Record Views

Details

Logo image