Moving to Green Edges: A Cooperative MEC Framework to Reduce Energy Demand of Clouds

Abstract

Multi-Access Edge Computing (MEC), extending cloud computing capabilities to network edges, is recognized as one of key pillars to improve service responsiveness and data privacy, and reduce backhaul traffic. However, one greatly potential advantage of MEC, i.e., more energy efficiency with respect to serving mobile users in contrast to cloud computing, has not been well explored. This advantage makes MEC be an effective way to slow the rapidly increasing energy demand of cloud computing while meeting growing demand for computing resources. To excavate the advantage, we first use a toy model to illustrate that the more the mobile users are served by MEC, the more the energy consumption of cloud computing can be saved; the less the energy is consumed to serve a given set of mobile users' demands, the more energy efficiency the MEC service is. As such, how to make full use of MEC resources in an energy-efficient way to handle users' demands becomes the fundamental problem to determine the effects on reducing the growth of the energy consumption of clouds. To address the fundamental problem, we take one key use case of MEC, MEC caching, as an example and make the following contributions. We formulate the energy efficient joint content placement and scheduling problem, considering practical constraints on wireless and backhaul transmissions. Since the formulated problem is proved to be a quadratic assignment problem, we design the efficient joint caching, and wireless and backhaul scheduling (JCWBS) algorithm. Finally, utilizing a real-world dataset, we testify the proposed algorithm.