A Review of Edge Computing Nodes based on the Internet of Things

Yunqi Dong, Jiujun Bai and Xuebo Chen

School of Electronic and Information Engineering, University of Science and Technology Liaoning,

Qianshan Road, Anshan 114051, Liaoning, China

Keywords: Edge Computing Nodes, Trust, Virtualization Technology, Isolation Technology.

Abstract: Due to the limitations of resources on the IoT device side, it is necessary to provide users with services not

only by means of a long-distance cloud computing center node, but also by some edge computing nodes. If

all the data on the device side is transmitted to the cloud center node, it will be returned to the device side

after unified processing. This transmission method will bring great pressure to the network link and data

center, and it will also cause the cloud center node to overload and refuse service. In order to speed up the

data processing and reduce delay, we briefly summarize the edge computing node model in this paper. Firstly,

considering the properties of privacy, security, trust and resource scheduling, etc., the edge computing is

analyzed. Then, based on these properties, the definition, architecture, and collaboration with cloud-edge-net

of edge computing are discussed. We also introduce the current key technologies used in edge computing,

such as network, virtualization, isolation, deep learning, and access control technologies. Finally, we give a

prospect of the possible application of edge computing in the future.

1 INTRODUCTION

Edge computing is an open platform based on edge

nodes, which integrates network, computing, storage,

application and other core capabilities of distributed

computing. The United States is the birthplace of the

concept of edge computing. In 2003, AKAMAI began

collaborating with IBM on "Edge computing",

providing edge-based services on Web Sphere (IBM &

AKAMAI, 2003). In 2011, Cisco was the first one to

proposed the concept of fog Computing, which extends

the concept of Cloud Computing (BONOMI, 2011).

Compared with Cloud Computing, it is closer to

the place where data is generated. Data, data-related

processing and applications are concentrated in

devices at the edge of the network. From fog

computing to edge computing, mapping along the

edge cloud network hierarchy. It makes possible for a

variety of computing tasks to achieve different levels

of intelligence at different costs and energy budgets

(Martin and Diaz, 2018), as shown Figure 1. In May

2016, a team of professors at Weisong Shi state

university in the United States came up with a formal

definition of edge. Edge computing is a new network

computing mode that performs edge computing.

https://orcid.org/0000-0001-6799-7667

Figure 1: A simplified architecture of edge-to-cloud

network with cloud computing.

The downlink data of edge computing represents

cloud service and the uplink represents the

interconnection of everything (Shi and Cao, 2016).

Edge computing refers to the computing and network

resources between the data source and any path to the

cloud computing center.

In November 2018, the Edge Cross Alliance was

established. The alliance was used a two-pronged

approach to address the critical issues involved in

accurately modeling data at the edge. The alliance

Dong, Y., Bai, J. and Chen, X.

A Review of Edge Computing Nodes based on the Internet of Things.

DOI: 10.5220/0009407003130320

In Proceedings of the 5th International Conference on Internet of Things, Big Data and Security (IoTBDS 2020), pages 313-320

ISBN: 978-989-758-426-8

313

also aimed to promote standardization of these data

models to further pursue more precise and simplified

the edge processing methods. They discussed the

edge computing nodes based on the Internet of things

(Boubendir and Guillemin, 2018).

This paper introduces the development history of

edge computing node in section 1. The section 2, 3

introduce the reasons for the generation of edge

computing nodes, and then give an overview from

definition, architecture, and Cloud-Edge-Network

Coordination. And then the key techniques to

promote the development of edge computing are

discussed in the 4

section, the 5

section is the the

direction of application, the 6

is the summary and

outlook.

2 EDGE COMPUTING NODE

GENERATION

In daily life, when users use various applications

on the network, the application will obtain all the

user's data, including the private data. Most of the

data is uploaded to the cloud center for storage. In

the process of data transmission and storage, due to

the openness of the application environment of the

Internet of things, diversity of services, energy

nodes limited and other characteristics, the risk of

user privacy data leakage increases, so the security

problem between nodes remains to be solved.

When the two edge clouds interact, they need to be

accessed through the edge trust node on the edge

cloud. Under the condition of reaching a

consensus, resources are allocated among the edge

clouds. The framework for this interaction can be

seen in Figure 2.

2.1 Privacy, Security and Trust

Saurabh proposed a framework where each sensor

node maintains reputation metrics which both

represent past behavior of other nodes and are used as

an inherent aspect in predicting their future behavior.

He employed a Bayesian formulation and combine

with the beta reputation system, for the algorithm

steps of reputation representation, updates,

integration and trust evolution (Saurabh and

Ganeriwal, 2008). Integrated comprehensive trust

theory, rough set theory, analytic hierarchy process

(AHP) and the combination of weight and grey

correlation analysis, in order to improve user

satisfaction and the success rate of node interaction

(Wang and Wen, 2019).

Figure 2: Trust and resource collaboration between edge

nodes.

Resource efficiency and dependability are the

most basic requirements of the trust model. The

definition of a self-adapted weighted method for trust

aggregation at CH level surpasses the limitations of

the subjective definition of weights in traditional

weighted method. The fuzzy degree of nearness is

used to evaluate the reliability of the recommended

trust values from the third-party nodes (Shao and

Zhou, 2015). Based on service preference to identify

user community effectively, a dynamic cloud service

trust evaluation model based on service-level

agreement (SLA) and privacy awareness is proposed,

which improves the satisfaction of service requesters

and avoids malicious interference (Wang and Wen,

2018). Shahzad provides a study of the issues related

to the data sharing through these smart devices over

service providing cloud. It proposes communication

architecture by introducing an intermediate layer of

data sharing control consisting of privacy agents. It

used to define custom privacy policies for different

personal attributes in different business models

(Shahzad and Iqbal, 2017).

2.2 Resource Allocation

2.2.1 Global Delay

In order to achieve the minimize global delay, a flow

scheduling policy based on link bandwidth

proportional fair allocation was designed by Feng.

The virtual forwarding space (VFS) of flood lighting

is proposed to realize the fair allocation of link

bandwidth and the minimum delay flow scheduling in

the data plane (Feng and Bi, 2014). Considering

dynamic changes in user requirements and limited

available resources in Fog devices. Naha propose

resource allocation and provisioning algorithms by

using resource ranking and provision of resources in

a hybrid and hierarchical fashion. Use the Cloud Sim

toolkit to simulate a real fog environment. The

performance of the proposed algorithms is better

IoTBDS 2020 - 5th International Conference on Internet of Things, Big Data and Security

314

compared with existing algorithms in terms of overall

data processing time and network delay (Naha and

Garg, 2019). Zenith is a novel model for allocating

computing resources in an edge computing platform

which allows the service providers to establish

resource sharing contracts with edge infrastructure

providers (Xu and Palanisamy, 2017). The latency-

aware scheduling and resource provisioning

algorithm that enables tasks to complete and meet

their latency requirements. A resource allocation

model of edge cloud based on auction mechanism is

proposed (Xu, 2016), and the time-delay requirement

of application is effectively guaranteed by the use of

delay-aware resource scheduling algorithm.

2.2.2 Quality-of-Service

Song presents an effective approach to periodically

distributing incoming tasks in the edge computing

network so that the number of tasks can be processed

in the edge computing network is increased. In order to

satisfied all the accommodated tasks quality-of-service

(QoS) specifications required in the IoT applications

(Song and Yau, 2017). To solve the problem of

resource scheduling in edge applications, under the

principle of not violating the service quality

requirements, a resource scheduling method based on

genetic algorithm is proposed (Skarlat and Nardelli,

2017).

In the context of smart cities, the concept of mobile

edge computing (MEC) is taken as a key factor to

improve the Quality of Service (QoS). Not only does it

enhance the autonomous creation of MEC services,

allowing data to be accessed any-where-anytime, with

the best Quality of Experience (QoE), but it also

reduced latency (Taleb and Dutta, 2017).

2.2.3 Bandwidth Allocation, Offloading

Distributed gradient-push algorithm can solve the

bandwidth allocation problems over WLAN with

directed typologies. Compared with the existing

distributed algorithms, we use weighted running

average bandwidth to replace the current state variables

meanwhile, the convergence performance is improved

and the robustness against noise is enhanced (Shi and

Zhou, 2019). Ascigil proposes a set of practical,

uncoordinated service placement policies. Through a

large number of simulations using synthetic and real

trace data, it has been demonstrated that the

uncoordinated strategy performs better than the

optimal placement solution, satisfying the maximum

amount of user requests (Ascigil and Phan, 2017).

For the power delay trade off of multi-user moving

edge computing system, an online algorithm for

determining local execution and calculating unloading

strategy is presented based on Lyapunov optimization

algorithm (Mao and Zhang, 2016). Lyapunov

optimization-based dynamic computation offloading

algorithm named low-complexity online algorithm,

which jointly decides the offloading decision, the

CPU-cycle frequencies for mobile execution and the

transmit power for computation offloading. A unique

advantage of this algorithm is that the decisions depend

only on the current system state. The implementation

of the algorithm only requires to solve a deterministic

problem in each time slot, for which the optimal

solution can be obtained either in closed form or by

bisection search (Mao and Zhang, 2016).

3 DEFINITION OF EDGE

COMPUTING

3.1 Definition

Edge computing is a new computing model. It can be

seen from the definition of edge computing that edge

computing is not to replace cloud computing but to

supplement cloud computing, providing a better

computing platform for mobile computing and

Internet of things. "Edge" in edge computing refers to

the computing and storage resources on the edge of

the network. The edge of the network here is opposite

to the data center, which is closer to the user both in

terms of geographical distance and network distance.

Professor Satyanarayanan describes edge

computing as: "edge computing is a new computing

model that deploys computing and storage resources

(e.g., cloudlets, micro data centers, or fog nodes) on

the edge of the network closer to the mobile device or

sensor"(Satyanarayanan, 2017). Wayne state

university in the United States WeiSong Shi et al (Shi

and Cao, 2016). The calculation is defined as: "edge

of computing refers to the network edge perform

calculations of a new type of computing model, in

computing the edge on the edge of the downstream

data cloud services, uplink said that everything

connected, and the edge of the edge of computing

refers to from the data source to cloud computing

center path between any of the computing and

network resources."

Edge computing is the technology of using these

resources to provide services to users on the edge of

the network so that applications can process data near

the data source. ISO think edge computing is a main

business processing and data storage in the network

edge node of the distributed computing form (Arndt,

A Review of Edge Computing Nodes based on the Internet of Things

315

2017). However, the Edge Computing Consortium

(ECC) is defined margin calculation is close to the

network edge side of data source, integrating core

ability of network computing storage application

development platform, etc (ECC, 2017).

3.2 Architecture

Resources and information are gathered to the center

by cloud computing, realizing the centralized

computing with complete control over user

information and behavior. Edge computing is a

distributed operation that transfers applications, data

and services from network center nodes to network

logical edge nodes for processing, you can see the

difference between Figure 3 and 4. Large services

originally handled entirely by cloud center nodes are

decomposed by edge computing, cut into smaller and

more manageable parts and distributed to edge. Based

on the previous work, λ-CoAP architecture, a new

edge computing architecture is proposed. Edge

computing deployments that cover the entire field of

vision, from IOT devices, edge smart gateways to the

cloud infrastructure (Martin and Diaz, 2018).

3.3 Cloud-Edge-Network Coordination

There is force synergy between edge computing and

cloud computing at the center. Force network refers

to the realization of cloud network collaboration,

cloud edge collaboration, and even edge collaboration

between edge computing and cloud computing and

networks, in order to realize the optimization of

resource utilization. On the basis of studying the

demand of edge computing force distribution and

dispatching, a deep fusion of cloud-network-edge

computing force network is proposed. This scheme

can effectively deal with the multi-stage deployment

of computing storage network and even algorithm

resources as well as the flexible scheduling among

nodes (Lei and Liu, 2019).

4 KEY TECHNOLOGY

The rapid development of edge computing is due to

technological advances. Promote the development of

core technologies of edge computing including

network, isolation technology, virtualization

technology, deep learning technology, access control

technology, algorithm execution framework security

and privacy, data processing platform, etc.

Figure 3: Cloud computing.

Figure 4: Edge computing.

4.1 Network

Edge computing pushes the computation closer to the

data source, and even deplores the entire computation

to nodes on the transport path from the data source to

the cloud computing center.

Edge devices generate a large amount of data,

while edge servers provide a large number of

services. Therefore, according to the edge servers and

network conditions, how to dynamically schedule

these data to the appropriate computing service

providers, load balancing will be one of the core

issues in edge computing. The maximal uniform

distribution of the load across closer and simpler

nodes can help managing and providing the big data

and large workloads more easily. Verma proposed an

efficient load balancing algorithm for a Fog-Cloud

based architecture. The ultimate goal is to balance

load through Fog networks and make internet less

Cloud dependent by having data available closer to

the user end (Verma and Yadav, 2016).

Edcs are used for reduce latency and network

congestion by processing data streams and user

requests in real time. The new load balancing

technology implements authentication of the edcs and

finds less heavily loaded edcs for task assignments.

Load balancing techniques enable workload

redistribution between edcs to improve resource

utilization and job response time (Puthal and Obaidat,

2018).

IoTBDS 2020 - 5th International Conference on Internet of Things, Big Data and Security

316

4.2 Virtualization Technology

Virtualization technology includes memory

virtualization, storage virtualization, hardware

virtualization, software virtualization and other

technologies. It is a process of turning a physical unit

into multiple logical units. It is applicable to multiple

applications to achieve efficient utilization of resources

and efficient management of various resources.

Virtualization technology provides high

availability for critical applications and simplifies

application deployment and migration. It performed

well in terms of cost, time and energy savings (Ding

and Ghansah, 2015). The virtualization technology

can achieve the rational utilization of hardware

resources through virtual software and assigns these

resources into corresponding application software

(Kong 2014). Container is a new virtualization

replaces the traditional virtual machine technology.

Changes have been made in the architecture of

Infrastructure as a Service (Iaas) and Platform as a

Service (Paas) platforms (Wu, 2017).

4.3 Isolation Technology

Isolation technology is a research technology that

supports the robust development of edge computing.

Edge equipment needs effective isolation technology

to ensure the reliability and quality of service.

1) The isolation of computing resources, that is,

the applications shall not interfere with each other;

2) Data isolation, that is, different applications

should have different access rights.

The core idea of Docker is to use the extended

LXC (Linux Container) scheme to achieve a

lightweight virtualization solution. By using kernel

namespace, Docker realizes isolation of container to

ensure the operating environment isolation among

each virtual machine service. The isolation

mechanism of Docker can reduce the memory

overhead and ensure the virtual instance density. (Liu

and Li, 2015).

Containers are constantly gaining ground in the

virtualization landscape as a lightweight and efficient

alternative to hypervisor-based Virtual Machines,

with Docker being the most successful representative.

This strategy allows developers to easily pack

applications into Docker image layers and distribute

them via public registries. Ioannis proposed and

demonstrate a mechanism for secure Docker image

manipulation throughout its life cycle: The creation,

storage and usage of a Docker image is backed by a

data-at-rest mechanism. (Giannakopoulos and

Papazafeiropoulos，2017).

4.4 Deep Learning Technology

Deep learning (DL) is a promising approach for

extracting accurate information from raw sensor data

from IoT devices deployed in complex environments.

Because of its multilayer structure, deep learning is

also appropriate for the edge computing environment

(Li, Ota and Dong, 2018).

DL as the representative technique of artificial

intelligence, it can be integrated into edge computing

frameworks to build intelligent edge for dynamic,

adaptive edge maintenance and management.

Promoting the fusion of edge intelligence and

intelligent edge, Edge Deep Learning (Han and Wang,

2019). Figure 5 is a frame about combine deep learning

with edge computing. Deep learning technology

require substantial computation resources to run

quickly. Edge computing is a viable way to meet the

high computation and low-latency requirements of

deep learning on edge devices and also provides

additional benefits in terms of privacy, bandwidth

efficiency, and scalability (Chen and Ran, 2019).

4.5 Access Control Technology

Access control technology is based on satisfying

users' demands for maximum enjoyment of resource

sharing, in order to achieve the management of user

access rights, prevent information from being

tampered with or abused by unauthorized users. It is

a reliable tool to ensure system security and protect

user' privacy. In edge computing, access control

becomes more difficult, mainly because in multi-user

access environments, access control functions need to

be provided by edge computing service providers.

Secondly, access control supports remote provision of

user basic information and policy information, and

also supports regular update of access control

information. Finally, access control for highly

distributed and dynamically heterogeneous data is an

important challenge in itself.

Figure 5: Combine deep learning with edge computing.

A Review of Edge Computing Nodes based on the Internet of Things

317

5 DIRECTION OF APPLICATION

5.1 Content Delivery Network

The combination of content distribution network and

edge computing is the future development trend. With

the evolution of CDN technology, the initial content

cache needs to be upgraded to smart edge computing.

The combination of CDN and mobile edge computing

device can fully reuse its nodes and provide edge

computing services. CDN USES global load

technology to direct user access to the nearest

working cache server.

Because of CDN distribution of nodes in the

network edge, service providers can take advantage

of the existing CDN further down to near the user's

network edge node, make each CDN nodes near the

user into edge mobile computing devices, to achieve

operational and support for various applications, in

order to meet the high frequency, the fragmentary

IOT equipment demand of calculation and

transmission. Amazon s AWS Lambda shows how

custom processing can enhance standard content

delivery networks (CDNs) without incurring

additional latency (Klas, 2017). A CDN with the help

of distributed denial of service (DDoS) mitigation can

enhance and maintain the website security from

DDoS attacks (Alrowaily and Lu, 2018).

5.2 5G

As the most important innovation scenario in the 5G

era, edge computing can provide customers with

multiple business guarantees such as low delay and

large bandwidth. According to IDC's data times 2025

report, 50% of data will be analyzed, processed and

stored on the edge of the network by 2025 (Tran and

Hajisami, 2017). At the same time, edge computing is

also considered as an important combination point of

5G, industrial Internet and Internet of things, which

can promote relevant industries to bring about rapid

development. However, with in-depth research and

practice, the concept of edge computing is not limited

to 5G, it has been extended to PON, wifi, 4G, etc.,

that is, multi-access edge computing.

5.3 Industry 4.0

Edge computing can play an important role in the

virtualization and business management of industrial

intelligent manufacturing systems. Firstly, through

the real-time connection and perception of edge

devices, an independent and reconfigurable digital

device model is established to enable the virtual

modeling and relational retrieval of production

resources. Secondly, adaptive allocation of network

resources is realized through SDN technology to

provide effective information transfer means for

reconfigurable devices.

6 CONCLUSIONS

The so-called Internet of everything extends on the X-

axis of time. The largest network is the Internet of

things. The huge scale of the Internet of things

requires fast computing power. If cloud computing is

the coordinator, edge computing is the executor.

Cloud computing is responsible for long-term big

data analysis. Edge computing is the calculation,

processing, optimization and storage of data close to

the edge of the Internet of things. Edge computing

and cloud computing work together and complement

each other. Edge computing is close to the device,

which also contributes to cloud data collection and

supports big data analysis of cloud applications. In

addition, trust is the foundation of secure interaction

between edge computing nodes. Edge compute nodes

control trust relationships and sensitive information

flows in a secure manner. Integrate distributed

computing into edge computing in a centralized cloud

computing architecture to reduce reliance on the

cloud center. Through the core technology of edge

calculation, the calculation efficiency is greatly

improved. Edge computing is still a work in progress.

In the future, the technical standards and

standardization of edge computing can be improved,

and the development of edge computing technology

can also be combined with the new generation of

information technology, so that the research on edge

computing will be more in-depth.

ACKNOWLEDGEMENTS

The research reported herein was supported by the

NSFC of China under Grants Nos. 71571091,

71771112 and 71371092, and by University of

Science and Technology Liaoning Talent Project

Grants No.：601011507-03.

REFERENCES

IBM & Akamai., 2003. Develop edge computing

application [EB].

IoTBDS 2020 - 5th International Conference on Internet of Things, Big Data and Security

318

Bonomi, F., 2011. Connected vehicles, the internet of

things, and fog computing. The Eighth ACM

International Workshop on Vehicular Inter-Networking

(VANET). ACM.

Martin, Fernandez, C., Diaz, Rodriguez, M., & Rubio

Munoz, B., 2018. An Edge Computing Architecture in

the Internet of Things. 2018 IEEE 21st International

Symposium on Real-Time Distributed Computing

(ISORC). IEEE.

ETSI ISG MEC., 2016. Multi-access edge computing

(MEC) framework and reference architecture.

Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L., 2016. Edge

Computing: Vision and Challenges. IEEE Internet of

Things Journal. IEEE.

Boubendir, A., Guillemin, F., Le, Toquin, C., 2018.

Federation of cross-domain edge resources: a brokering

architecture for network slicing. 2018 4th IEEE

Conference on Network Softwarization and Workshops

(NetSoft). IEEE.

Saurabh, Ganeriwal, Laura., 2008. Reputation-based

framework for high integrity sensor networks. ACM

Transactions on Sensor Networks. ACM.

Wang, Y., Wen, J., Wang, X., Tao, B., & Zhou, W., 2019.

A Cloud Service Trust Evaluation Model Based on

Combining Weights and Gray Correlation Analysis.

Security and Communication Networks.

Shao, N., Zhou, Z., Sun, Z., 2015. A Lightweight and

Dependable Trust Model for Clustered Wireless Sensor

Networks. International Conference on Cloud

Computing and Security. SPRINGER.

Wang, Y., Wen, J., Zhou, W., & Luo, F., 2018. A Novel

Dynamic Cloud Service Trust Evaluation Model in

Cloud Computing. 2018 17th IEEE International

Conferences. IEEE.

Shahzad, S., Iqbal, M., Ahmad, N., 2017. Privacy Agents

for IoT Cloud Communication. In Proceedings of the

2nd International Conference on Internet of Things, Big

Data and Security. IEEE.

Feng, T., Bi, J., Wang, K., 2014. Allocation and Scheduling

of Network Resource for Multiple Control Applications

in SDN. 2014 IEEE/IFIP Network Operations and

Management Symposium. IEEE.

Naha, R, K., Garg, S., Chan, A., & Battula, S, K., 2019.

Deadline-based dynamic resource allocation and

provisioning algorithms in Fog-Cloud environment.

Future Generation Computer Systems. ELSEVIER.

Xu, J., Palanisamy, B., Ludwig, H., & Wang, Q., 2017.

Zenith: Utility-Aware Resource Allocation for Edge

Computing. 2017 IEEE International Conference on

Edge Computing. IEEE.

Xu, J., 2016. A Cloud Computing Resource Allocation

Model Based on Combinatorial Double Auction. 2016

3rd International Conference on Information Science

and Control Engineering (ICISCE). IEEE.

Song, Y., Yau, S, S., Yu, R., 2017. An Approach to QoS-

based Task Distribution in Edge Computing Networks

for IoT Applications. 2017 IEEE International

Conference on Edge Computing (EDGE). IEEE.

Skarlat, O., Nardelli, M., Schulte, S., 2017. Optimized loT

service placement in the fog. Service Oriented

Computing and Applications. SPRINGER.

Taleb, T., Dutta, S., Ksentini, A., 2017. Mobile Edge

Computing Potential in Making Cities Smarter. IEEE

Communications Magazine. IEEE.

Zheng, Qing., Shi., Chuan, Zhou., 2019. An Improved

Distributed Gradient-Push Algorithm for Bandwidth

Resource Allocation over Wireless Local Area

Network. Journal of Optimization Theory and

Applications. SPRINGER.

Ascigil, O., Phan, T, K., Tasiopoulos, A, G., 2017. On

uncoordinated service placement in edge-clouds. 2017

IEEE International Conference on Cloud Computing

Technology and Science (CloudCom). IEEE.

Mao, Y., Zhang, J., Song, S, H., 2016. Power-Delay

Tradeoff in Multi-User Mobile-Edge Computing

Systems. Oxford University Press.

Mao, Y., Zhang, J., Letaief, K, B., 2016. Dynamic

Computation Offloading for Mobile-Edge Computing

with Energy Harvesting Devices. IEEE Journal on

Selected Areas in Communications. IEEE.

Satyanarayanan, M., 2017. The Emergence of Edge

Computing. Computer.

Arndt, T., 2017. International Organization for

Standardization. ISO.

Edge Computing Consortium (ECC)., 2017. Alliance

Internet Industrial (AII). Edge computing reference

architecture2.0.

Martin Fernandez, C., Diaz Rodriguez, M., & Rubio

Munoz, B. 2018. An Edge Computing Architecture in

the Internet of Things. 2018 IEEE 21st International

Symposium on Real-Time Distributed Computing

(ISORC). IEEE.

Lei, b., Liu, Z, Y., Wang, X, L., 2019. A new edge

computing scheme based on cloud, network and edge

fusion: Calculate power network. Telecommunications

Science.

Verma, S., Yadav, A, K., Motwani, D., 2016. An efficient

data replication and load balancing technique for

computing environment. International Conference on

Computing for Sustainable Global Development. IEEE.

Puthal., Deepak., Obaidat., Mohammad, S., Nanda,

Priyadarsi., Secure and Sustainable Load Balancing of

Edge Data Centers in Fog Computing. IEEE

Communications Magazine. IEEE.

Ding, W, M., Ghansah, B., & Wu, Y, Y., 2015. Research

on the Virtualization Technology in Cloud Computing

Environment. International Journal of Engineering

Research in Africa. ELSEVIER.

Kong, L, L., 2014. Study of Cloud Computing and

Virtualization Technology. Applied Mechanics and

Materials.

Zhixue, W. U., 2017. Advances on virtualization

technology of cloud computing. Journal of Computer

Applications. ELSEVIER.

Si-Yao, L., Qiang, L. I., Bin, L. I., 2015. Research on

isolation of container based on docker technology.

Computer Engineering & Software.

A Review of Edge Computing Nodes based on the Internet of Things

319

Giannakopoulos, I., Papazafeiropoulos, K., Doka, K.,

Koziris, N., 2017. Isolation in Docker through Layer

Encryption. 2017 IEEE 37th International Conference

on Distributed Computing Systems (ICDCS). IEEE.

Li, H., Ota, K., Dong, M., 2018. Learning IoT in Edge:

Deep Learning for the Internet of Things with Edge

Computing. IEEE Network. IEEE.

Han, Y., Wang, X., Leung, V, C, M., 2019. Convergence of

Edge Computing and Deep Learning: A

Comprehensive Survey.

Chen, J., & Ran, X., 2019. Deep Learning with Edge

Computing: A Review. Proceedings of the IEEE. IEEE.

Klas, G., 2017. Edge Computing and the Role of Cellular

Networks. Computer.

Cloudflflare. What is a cdn? How does a cdn work?

[Online]. Available: https://www.cloudflare.com/

learning/cdn/what-is-a-cdn/

Alrowaily, M., Lu, Z., 2018. Secure Edge Computing in IoT

Systems: Review and Case Studies. 2018 IEEE/ACM

Symposium on Edge Computing (SEC). ACM.

Tran, Tuyen. X., Hajisami, Abolfazl, Pandey, Parul, &

Pompili, Dario., 2017. Collaborative mobile edge

computing in 5G networks: new paradigms, scenarios,

and challenges. IEEE Communications Magazine.

IEEE.

IoTBDS 2020 - 5th International Conference on Internet of Things, Big Data and Security

320