ReviewEnergy aware edge computing: A survey☆
Introduction
The advances in sensors, and wireless communication technologies promote the wide deployment of mobile devices and smart things in industrial and personal scenarios. With higher computing capability and higher storage capacity, these mobile devices and smart things can provide various kinds of services [1], such as data aggregation, real-time local data analytics, content caching and transmission relay. The devices deployed at the network edge are usually referred to as edge devices. Edge devices with more functionalities and more powerful computing capabilities are referred to as edge servers since they are closer to network edge and farther away from the cloud data centers. Edge computing can fully exploit the computing capability of edge devices and edge servers [2], [3]. Applications can be executed at the network edge closer to data sources. Processing the data at the network edge provides shorter response time and less pressure on the network bandwidth. As such, edge computing improves the user experience for time-sensitive application significantly.
In particular, mobile edge computing (MEC) is becoming a promising paradigm to provide more responsive computing services at the network edge [4], [5]. However, edge devices are usually resource-constrained with limited computing capability and power supply. In some cases edge devices are prohibited from executing applications that require a high-power supply and a large amount of data processing [6]. To this end, migrating computing tasks to nearby edge servers or cloud data centers can improve the application performance for large volume data analytics or resource-hungry computing intensive processing such as deep learning model training and inference [7]. In edge computing environment, computing tasks can be offloaded from the cloud data center to edge servers or edge devices for low latency and data privacy preservation.
In edge computing environment, both devices and servers are usually heterogeneous in terms of hardware capabilities, architectural and programming interoperability, operating system, and services stacks. Many edge devices are resource constrained in computing capability, storage capacity and network connectivity. In many edge computing scenarios increasing the energy consumption could have a negative impact on the power-constrained IoT device or edge cloud side with limited power sources. Since billions of edge devices are deployed in edge computing environment, their total energy consumptions are immense and as important as those of cloud data centers. For example, for battery powered devices or power constrained edge nodes, energy aware edge computing can extend their lifetime, provide quality of service guarantee, or increase system throughput under specific power budget. Different from energy aware computing in server systems and cloud data centers, energy awareness in edge computing involves all operations conducted along data’s whole life cycle, including data generation, transmission, aggregation, storage, processing, and etc. Therefore, energy aware computing is urged for all aspects of edge computing, including architecture, operating system, middleware, service provisioning, and computing offloading.
In edge computing, computation offloading is frequently invoked for latency minimization and Quality of Services guarantee. Specifically, in order to tradeoff among system overheads, energy consumption, and system performance, tasks may be offloaded to edge devices from the cloud data centers. However, computing offloading may lead to jitter in quality of service along with energy shifting and redistribution among different edge nodes. Nowadays, energy efficiency has become one of the most important concerns for both cloud servers and mobile devices. Though energy efficiency in cloud data centers has been thoroughly investigated, energy efficiency in edge computing is largely left uninvestigated due to the complicated interactions between edge devices, edge servers, and cloud data centers.
In this paper, we conduct a thorough survey on the energy aware edge computing. From the hardware layer to the application layer, we provide a bottom-up systematic review of the existing work on energy efficiency in edge computing. A systematic view of the survey is presented in Fig. 1.
This remainder of this paper is organized as follows. Section 2 reviews energy-aware hardware design in edge computing. We review the work on energy-aware edge architectures in Section 3 and energy-aware edge operating systems in Section 4. We discuss the energy-aware edge middleware in Section 5 and energy-aware applications and services provisioning in Section 6. Energy-aware computing offloading is surveyed in Section 7 and we conclude our paper in Section 8.
Section snippets
Energy-aware edge hardware design
Edge devices are widely deployed in various scenarios, such as smart cities, vehicles, industrial workplace, smart home, and information and communication infrastructures. However, these edge devices are resource-constrained with limited power supply, or they are equipped with lower powerful computing and storage resources for computing-intensive tasks. Hence, energy efficiency should be the primary design goal for edge devices, such as processors, sensors, edge servers, switches, and routers.
Energy-aware edge computing architecture
Energy aware architectural design is vital for edge computing although hardware level energy reduction capabilities are available in current edge devices. Architecture design may harness and integrate different level of energy aware approaches and provide functions and capabilities for energy awareness programming interface design and implementation. In this section, we survey the research work of energy-aware edge computing architecture, including memory system, networking, compiler and
Energy-aware edge OS
A typical operating system (OS) manages computer hardware, software resources, and provides common services for computer programs. With operating system, developers can build applications without thinking too much about the device underneath and in most cases, applications take advantage of services and APIs available in the OS without having to deal directly with hardware resources and even network connectivity. With the advance in computer technology, smaller, lighter, and faster computing
Energy-aware edge middleware
Middleware provides abstraction of underlying resources to upper service applications. In order to provide energy efficient resource transparency and interpretation, edge middleware must have elaborate system design on networking, storage, and computing. For example, it is usually assumed that in edge computing scenario, IoT services are depending on always-connected networking. However, the always-connected approach can sometimes generate boatloads of data and not all of it is necessary at the
Energy aware edge services & applications
In this section, we focus on the topic of energy-aware edge devices and applications. The related work of energy awareness in edge devices and applications is listed in Fig. 7. According to Fig. 7, we provide Table 7 for a reference to the existing work.
Energy aware computing offloading
Computing offloading from the edge to the cloud is a promising technique that saves battery while improving the computing capabilities of edge devices [214]. However, it is not always energy efficient when applications are offloaded to the cloud. Since the energy efficiencies of mobile devices, edge servers, and cloud servers are different, the energy consumption of each type of device/server varies and depends on the type of device/server where the tasks are executed. An inefficient offloading
Conclusions and future work
Edge computing is an emerging paradigm to meet the ever-increasing computing and communication demands from billions of edge devices. Edge computing is promising for various application scenarios due to its less network bandwidth usage and thus less data center side processing pressure, as well as enhanced service responsiveness and data privacy protection. Edge Computing is becoming feasible to handle the data deluge occurring in the network edge to gain insights that assist in real-time
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (227)
A survey of computation offloading strategies for performance improvement of applications running on mobile devices
J. Netw. Comput. Appl.
(2015)- et al.
From cuda to opencl: Towards a performance-portable solution for multi-platform gpu programming
Parallel Comput.
(2012) - et al.
A review of thin film solar cell technologies and challenges
Renew. Sustain. Energy Rev.
(2017) - et al.
Faultdiagnosis and quantitative analysis of micro-short circuits for lithium-ion batteries in battery packs
J. Power Sources
(2018) - et al.
Real-time diagnosis of micro-short circuit for li-ion batteries utilizing low-pass filters
Energy
(2019) - et al.
A comparative cost analysis offault-tolerance mechanisms for availability on the cloud
Sustain. Comput.: Inform. Syst.
(2018) - et al.
A distributed multi-level model with dynamic replacement for the storage of smart edge computing
J. Syst. Archit.
(2018) - et al.
The promise of edge computing
Computer
(2016) - et al.
Edge computing: Vision and challenges
IEEE Internet Things J.
(2016) - et al.
Performance guaranteed computation offloading for mobile-edge cloud computing
IEEE Wirel. Commun. Lett.
(2017)
Mobile-Edge Computing Introductory Technical White PaperWhite paper
Mobile edge computing: A survey on architecture and computation offloading
IEEE Commun. Surv. Tutor.
Efficient multi-user computation offloading for mobile-edge cloud computing
IEEE/ACM Trans. Netw.
GPU computing
GPGPU: general-purpose computation on graphics hardware
Eurora: a european architecture toward exascale
Workload partitioning strategy for improved parallelism on FPGA-CPU heterogeneous chips
Xeon+ FPGA platform for the data center
A reconfigurable fabric for accelerating large-scale datacenter services
ACM SIGARCH Comput. Archit. News
The chimera: an off-the-shelf CPU/GPGPU/FPGA hybrid computing platform
Int. J. Reconfig. Comput.
FPGA-GPU architecture for kernel SVM pedestrian detection
FPGA-GPU-CPU heterogenous architecture for real-time cardiac physiological optical mapping
Single-chip heterogeneous computing: Does the future include custom logic, FPGAs, and GPGPUs?
Accelerating recurrent neural networks in analytics servers: Comparison of FPGA, CPU, GPU, and ASIC
Accelerating binarized neural networks: Comparison of FPGA, CPU, GPU, and ASIC
The architectural implications of autonomous driving: Constraints and acceleration
Scalable synthesis of silicon-nanolayer-embedded graphite for high-energy lithium-ion batteries
Nat. Energy
Rechargeable nickel–3d zinc batteries: An energy dense, safer alternative to lithium-ion
Science
Thin-film solar cells: an overview
Prog. Photovolt., Res. Appl.
Thin-film cuinse2/cds heterojunction solar cells
Appl. Phys. Lett.
New results on the development of a thin-film p-cdte-n-cds heterojunction solar cell
Amorphous silicon solar cell
Appl. Phys. Lett.
Solar cells using discharge-produced amorphous silicon
J. Electron. Mater.
High photocurrent polycrystalline thin-film cds/cuinse2 solar cellar
Appl. Phys. Lett.
Complete microcrystalline solar cell crystalline or amorphous cell behavior?
Appl. Phys. Lett.
Growth and characterization of thin-film compound semiconductor photovoltaic heterojunctions
J. Vac. Sci. Technol.
An energy efficient task scheduling algorithm in DVFS-enabled cloud environment
J. Grid Comput.
Data center energy consumption modeling: A survey
IEEE Commun. Surv. Tutor.
Design and implementation of a critical speed-based DVFS mechanism for the android operating system
Low power hevc software decoder for mobile devices
J. Real-Time Image Process.
Happe: Human andapplication-driven frequency scaling for processor power efficiency
IEEE Trans. Mob. Comput.
User-satisfaction-aware power management in mobile devices based on perceptual computing
IEEE Trans. Fuzzy Syst.
Memory power management via dynamic voltage/frequency scaling
Memscale: active low-power modes for main memory
Energy-performance trade-offs on energy-constrained devices with multi-component DVFS
Optic: Optimizing collaborative cpu–gpu computing on mobile devices with thermal constraints
IEEE Trans. Comput.-Aided Des. Integr. Circuits Syst.
Cited by (0)
- ☆
This work is supported by Natural Science Foundation of China (No. 61972118, No. 61972358, and No. 61572163), and Key Research and Development Program of Zhejiang Province, China (No. 2018C01098 and No. 2019C01059).