Multi-Access Edge Computing: Striking a balance between QoE and Energy Efficiency

The Multi-Access Edge Computing (MEC) came to address the issue of the current centralized cloud paradigm, with tight interactions between the computing and networking that is reaching its limits. The motivation is principally to enhance user Quality of Experience (QoE), increase manageability, reliability and performances. MEC is based on softwarization techniques as SDN and NFV, the aim consists of building federated and distributed clouds based on set of mini-clouds situated at the network access, to enable latency sensitive application as requested for 5G network. In this paper, we want to know how the user demands allocation is impacting the QoE and the provider energy efficiency? To answer to this question, we define a generic MEC service placement optimization model that aims to minimize the user maximum latency (QoE) and the number of active MEC servers (energy efficiency). By extensive simulations on realistic ISP topology, we draw conclusions on the tread-off between users' QoE and energy efficiency goals. Our results show that the demands placement strategy can improve by ten QoE and by four the energy efficiency. We believe that this is the first work that try to address these impacts for MEC service placement, and we believe that these results can help MEC enablers and designers to take the right choices when energy efficiency and users' QoE performance goals need to be taken into account.