۱- Introduction

۲- System model

۳- Proposed power consumption model

۴- Proposed Power Efficient Clustering Scheme (PECS)

۵- Numerical results

۶- Conclusion

References

Abstract

Mobile edge computing (MEC), which is an evolution of cloud computing, is acknowledged as a promising technology for meeting low latency and bandwidth efficiency required in fifth generation (5G) era. Accordingly, the enlargement of distributed MEC installments will be realized and their power consumption might be a significant problem in terms of operating costs for service providers. Thus, this paper proposes a theoretical framework for MEC server clustering to minimize power consumption of the MEC environment. To do this, considering power consumption behavior of MEC servers using CPUs with dynamic voltage frequency scaling, we propose a power-efficient clustering scheme (PECS) that minimizes power consumption of MEC servers by obtaining the optimal number of clusters through convex optimization. Numerical results reveal the proposed PECS reduces power consumption of the MEC environment by 12.32% relative to an existing scheme while sustaining average delay of inflows processed in MEC servers at the acceptable level without turning off MEC servers.

Introduction

Mobile edge computing (MEC), which sustains multiple distributed edge servers by moving computing points from centralized points to near end-user points, has emerged as a promising technology for pervasive fifth generation (5G) services such as internet of things (IoT), device-to-device (D2D), and low-latency video streaming (e.g., virtual reality/ augmented reality video streaming) etc., to meet low latency and bandwidth efficiency [1–۴]. Subsequently, the enlargement of distributed MEC installments will be realized and their power consumption might be a significant issue in terms of operating costs for service providers. Thus, with much attention being directed towards energy efficient data center management, power consumption is also expected to be an important criterion in MEC management along with delay. According to iGR [5], the number of U.S. MEC installations is expected to reach 563,000 divided into 2,000 locations by 2026. They generate more than 2.66 TWh annually in the U.S. alone and the corresponding cost amounts to more than 217 million dollars when the average power consumption of each MEC installation is 300W and the power utilization effectiveness (PUE) is 1.8 [6]. Moreover, the PUE of MEC is getting worse since MEC is restricted to being located physically near the end-user point, whereas a traditional data center can reduce power consumption by more than 20% by locating the data center in a cool climate region with free cooling [7]. To date, researches on MEC has mostly focused on reducing delay [8, 9]. MEC is implemented as a virtualized platform instead of a custom hardware appliance, leveraging recent advances in network function virtualization (NFV) [1, 10, 11].