Throughput Maximization of Delay-Sensitive Request Admissions via Virtualized Network Function Placements and Migrations

Abstract

Network Function Virtualization (NFV) has attracted significant attentions from both industry and academia as an important paradigm change in network service provisioning. Most existing studies on NFV dealt with admissions of user requests through deploying Virtualized Network Function (VNF) instances for individual user requests, without considering sharing VNF instances among multiple user requests to provide better network services and improve network throughput. In this paper, we study the network throughput maximization problem by adopting two different VNF instance scalings: (i) horizontal scaling by migrating existing VNF instances from their current locations to new locations; and (ii) vertical scaling by instantiating more VNF instances if needed. Specifically, we first propose a unified framework that jointly considers both vertical and horizontal scalings to maximize the network throughput, by admitting as many requests as possible while meeting their resource demands and end-to-end transmission delay requirements. We then devise an efficient heuristic algorithm for the problem. We finally conduct experiments to evaluate the performance of the proposed algorithm. Experimental results demonstrate that the proposed algorithm outperforms a baseline algorithm.