Optimization-in-the-Loop for Energy-Efficient 5G

TL;DR

This paper presents an optimization framework integrated into SDN and NFV systems to improve energy efficiency in 5G networks by jointly managing node activation, VNF placement, and traffic routing based on real-time data.

Contribution

It introduces a novel optimization model for 5G network management that considers both networking and computational features, implemented within SDN/NFV environments.

Findings

Significant power savings demonstrated in real-world tests

Effective joint optimization of node switching, VNF placement, and routing

Quantitative analysis of contribution impacts on power consumption

Abstract

We consider the problem of energy-efficient network management in 5G systems, where backhaul and fronthaul nodes have both networking and computational capabilities. We devise an optimization model accounting for the main features of 5G backhaul and fronthaul, and jointly solve the problems of (i) node switch on/off, (ii) VNF placement, and (iii) traffic routing. We implement an optimization module within an application on top of an SDN controller and NFV orchestrator, thus enabling swift, high-quality decisions based on current network conditions. Finally, we validate and test our scheme with real-world power consumption, network topology and traffic demand, assessing its performance as well as the relative importance of the main contributions to the total power consumption of the system.