Consolidation of Services in Mobile Edge Clouds using a Learning-based Framework

Mobile edge computing offers numerous advantages over centralized computing platforms as it can deliver ultra-low-latency services for mobile users and consume less energy. The management of services on these systems is challenging due to the mobility of the users. Previous edge service place-ment/migration solutions have focused on minimizing the latency or maximizing the throughput of the services without regard to their energy consumption. The consolidation of services onto physical machines can be used to reduce energy consumption but this can have negative effects on the latency of services. The exact latency requirements of the services, however, varies considerably and it is possible to strike a balance between providing a reasonable Quality of Service for mobile users and reducing the energy consumption of the mobile edge cloud. Thus, in this paper we propose a framework for optimizing the trade-off between these two factors which can be adjusted depending on the requirements of the service as well as reinforcement learning techniques to achieve this optimization goal. Preliminary results demonstrate that our learning based approach can achieve a balanced performance between approaches which focus solely on consolidation or latency performance.