Energy-Aware Optimal Service Function Chain Embedding in Emulated Multi-Access Edge Computing for Internet of Things (to appear)

Abstract

Rising data processing demands from ever-increasing IoT devices have led to increased use of edge compute deployments, where energy consumption and associated carbon emissions have become critical challenges. By switching from dedicated hardware functions to chains of virtualised network functions (service function chains), energy consumption can be reduced when processing IoT network data. However, optimally embedding these functions to achieve this is $\mathcal{NP}$-hard, with existing solutions optimising only a subset of the problem space.

In this work, we explore a Genetic Algorithm-based approach that optimises all three sub-problems scalably to minimise energy consumption and carbon emission in SFCs. We test our proposed solution across two Multi-Access Edge Computing scenarios and show that our solution efficiently converges on an optimal solution in terms of the number of embedded SFCs and energy consumption.