۱- Introduction

۲- System model

۳- Problem formulation

۴- Algorithm design

۵- Performance evaluation

۶- Related work

۷- Conclusion

References

Abstract

The ever-increasing computation tasks and communication traffic have imposed a heavy burden on cloud data centers and also resulted in a significantly high energy consumption. To ease such burden, edge computing is proposed to explore the distributed resources of edge devices (e.g., base stations) to provision the cloud services for latency-sensitive applications at the network edge. Owing to the geo-distribution of edge devices, edge computing is also an ideal energy efficient platform to leverage the distributed green energy for energy efficient computing. Thus, it is natural to integrate Energy Internet (EI) technology into edge computing for customizable energy scheduling. In such EI supported edge computing, both the green energy generation rates and the data processing demands vary in different time and space. To pursue high energy efficiency, it is desirable to maximize the utilization of green energy so as to reduce the brown energy consumption. This requires careful task allocation and energy scheduling to match the energy provision and demand. In this paper, we investigate the energy cost minimization problem with joint consideration of VM migration, task allocation and green energy scheduling and prove its NP-hardness. To tackle the computation complexity, a heuristic algorithm approximating the optimal solution is proposed. Through extensive simulations, we show that the proposed algorithm can efficiently reduce brown energy consumption and perform much close to the optimal solution.

Introduction

Nowadays, cloud computing provides various services to global users with high resource utilization, strong computing ability, and high service reliability. According to Cisco’s Global Cloud Index Report (2015–۲۰۲۰), ۹۲% of global computational workloads is processed in cloud. However, with the rapid increasing of user tasks, the bulk data transmission and processing impose a heavy burden on the communication bandwidth and computation resource of cloud, bringing unbearable service delay to end users [1– ۳]. Moreover, the significantly high energy consumption of cloud also becomes another critical issue with growing concerns [4,5]. To handle this challenge, the concept of edge computing, which extends cloud computing to the network edge, is proposed. Edge computing utilizes the distributed resources in routers, gateways, base stations and even mobile devices to offer ‘‘cloud’’ services to the users in proximity [6,7], as shown in Fig. 1. By directly processing user tasks in local edge devices, the transmission delay and the traffic congestion on the Internet can be effectively reduced, thereby improving the user experience [8]. Furthermore, edge computing platform is formed by geo-distributed edge devices, which can harvest green energy from the environment. Such paradigm naturally provides an ideal platform to utilize the local green energy [9]. Actually, many studies [10–۱۳] have already mentioned the possibility and advantages of utilizing green energy in edge cloud to reduce the brown energy consumption.