Edge Computing in Industrial Internet of Things

A Report Submitted

in Partial Fulfilment for the degree of

Masters of Technology
in
Computer Science and Engineering
by
SURENDRA KUMAR

Pursued in

Department of Computer Science and Engineering

Graphic Era Deemed to be University

Graphic Era Deemed to be University, Dehradun

MAY, 2023
 Certificate
This is to certify that the seminar project Edge Computing in Industrial Internet
of Things will be submitted by Surendra Kumar under the supervision of Dr.
Vikas Tripathi in the academic year 2022-23.

Sign of Teacher Sign of External

 Abstract
A significant study area developed from the Internet of Things (IoT) is the
Industrial Internet of Things (IIoT). The IIoT establishes data collecting,
exchange, and analysis systems, optimises processes and services, and connects
all sorts of industrial equipment across the network to save costs and increase
production. Edge computing (EC) can help IIoT drastically reduce decision
latency, conserve bandwidth, and, to a certain extent, safeguard privacy. This
report describes the advancements in IIoT EC research. First, the ideas of EC and
IIoT are explored, and then the development of EC research is thoroughly
covered. Next, the projected future architecture for EC in the IIoT is proposed.
Then, the opportunities and challenges of EC in the IIoT are covered, including
edge communication based on 5G, load balancing and data offloading, edge
intelligence, and data sharing security. Last but not least, some common IIoT EC
application scenarios, including prognostics and health management (PHM),
smart grids, factory coordination, intelligent connected vehicles (ICV), and smart
logistics are provided.
 Acknowledgment

I take this opportunity to express my profound gratitude and deep regard to the
respected faculties for their exemplary guidance, monitoring, and constant
encouragement throughout the course of this report.

I am obliged to sir Doctor Vikas Tripathi, for the valuable information provided
by him and am grateful for his cooperation during this report making. Lastly, the
most sincere and important acknowledgment and gratitude are due to my
Computer Science Department and my Friends having their suggestions and
moral boosting support and encouragement at every stage of this endeavour. In
the end, I convey my thanks to all those who extended their cooperation directly
or indirectly in the completion of my research.
 Table Of Contents
Abstract……………………………………………………………………………………..

Acknowledgement………………………………………………………………………….

Table of Contents…………………………………………………………………………..

Chapter 1. Introduction……………………………………………………………………

1.1 IIoT and Related Concepts………………………………………………………

1.2 Deploying Edge Computing in IIoT…………………………………………….

1.3 Motivation and Paper Organization……………………………………………..

Chapter 2. Related Advances and Surveys of Edge Computing………………………..

2.1 Fog Computing………………………………………………………………….

2.2 MEC…………………………………………………………………………….

2.3 Edge Networking……………………………………………………………….

2.4 Edge Security…………………………………………………………………..

2.5 Edge Computing in IIoT……………………………………………………….

Chapter 3. Reference Architecture of Edge Computing in IIoT……………………….

3.1 Existing Reference Architectures………………………………………………

3.2 Proposed Reference Architecture of EC in IIoT……………………………….

3.2.1 Device Layer………………………………………………………….

3.2.2 Edge Layer…………………………………………………………....

3.2.3 Cloud Application Layer……………………………………………..

Chapter 4. Challenges and Opportunities……………………………………………….

4.1 5G Based Edge Communication……………………………………………….

4.2 Data Offloading and Load Balancing………………………………………….

 4.2.1 Data Offloading………………………………………………………

4.2.2 Load Balancing………………………………………………………

4.3 Edge Intelligence……………………………………………………………….

4.4 Data Sharing Security………………………………………………………….

Chapter 5. Application Scenarios……………………………………………………….

5.1 PHM…………………………………………………………………………...

5.2 Smart Grid……………………………………………………………………..

5.3 Manufacturing…………………………………………………………………

5.4 ICV……………………………………………………………………………

Conclusion……………………………………………………………………………….

References……………………………………………………………………………….
 CHAPTER – 1

INTRODUCTION

The Internet of Things (IoT) is a dynamic global network architecture with self-configuring
capabilities built on open, standardised communication protocols that enable all things to
connect with one another and enable collaborative problem solving and information sharing
[1]. To realise dynamic information exchange, the Internet of Things (IoT) connects various
objects with independent addresses and identities. The Industrial Internet of Things (IIoT) is a
brand-new study field that was created as a result of the application of IoT to the industrial
sector. The IIoT is a brand-new service-oriented industrial ecosystem that makes use of the
network interconnection of industrial resources, data interoperability, and system
interoperability to enable flexible resource allocation, on-demand process execution, rational
process optimisation, and quick environment adaptation [2]. In order to perform deeper data
analysis and mining to increase productivity and optimise operations, IIoT abstracts industrial
processes into data types, transforms devices into data terminals, gathers underlying basic data
in all directions, and combines the powerful data storage and computing functions of cloud
computing. Its implementation will bring about significant changes in how businesses produce,
operate, and are managed, establishing the groundwork for more logical supply chain resource
allocation and increases in the effectiveness of their output and services.

1.1 IIoT and Related Concepts

Despite the IIoT being descended from the IoT, the IIoT and IoT both have an emphasis on
principles and actual application situations.. The widely adopted Internet of Things (IoT) is
mostly consumer-focused and seeks to enhance people's quality of life. Smart homes, health
monitoring, indoor localisation, and other IoT application scenarios are the most common [3].
The IIoT is focused on improving industrial production efficiency and is production-oriented.
Smart logistics, remote maintenance, and intelligent factories are examples of typical IIoT
application scenarios [4]. The system frameworks of various IoT application scenarios
typically need to be constructed from scratch, and the sensor deployment scale has low
precision requirements [2]. However, because the system frameworks of IIoT application
scenarios are created based on conventional industrial infrastructure, the deployment size of
sensors is quite big and has high precision requirements. In contrast to the IIoT, where most
devices are fixed in place, generate large amounts of perceived data, and have low delay
tolerance, IoT devices often have considerable mobility, generate medium volumes of data,
and have good delay tolerance.

Fig.1. CPS, IoT, Industrial Internet, IIoT and Industry 4.0 in venn diagram.

The Industrial Internet of Things (IIoT) concept is closely related to certain well-known ideas
including cyber-physical systems (CPS), IoT, the Industrial Internet, and Industry 4.0. In order
to achieve highly synergistic and autonomous informationization capabilities, real-time and
flexible feedback, and a positive cycle between the physical and information worlds, Helen
Gill's CPS, which was first proposed in 2006, emphasises the deep integration of various
information technologies, including sensing technology, embedded technology, and software
& hardware technology. IoT, a subclass of CPS, focuses primarily on the Internet-based
interactions between items based on distinctive identifications. The IoT concept is supported
by the Internet's globalisation, openness, interoperability, and socialisation. The Industrial
Internet is a concept put forth by the Industrial Internet Consortium (IIC), which was
established by the top five US corporations, including GE, AT&T, IBM, Intel, and Cisco. It
primarily focuses on the development, application, and standardisation of cutting-edge
networks, the improvement of data flow, and the digital transformation of the entire industrial
sector. Industry 4.0, a sub-concept of IIoT that was introduced in Germany, is an information
CPS that is mostly used in the sphere of smart manufacturing and is globally focused. In
conclusion, CPS is the most comprehensive of the aforementioned notions since it offers a
roadmap for the interaction between the physical and informational worlds based on the
connectivity of things. IoT emphasises the connections between things through physical
addresses, whether they are business- or consumer-oriented, whereas the Industrial Internet
shows probable future patterns of sectors based on new technology. The Industrial Internet and
IoT are subcategories of CPS in this regard [5], [6]. In the meantime, IIoT refers to the point at
which IoT and the Industrial Internet converge. Industry 4.0 is also a part of the IIoT, which is
a manufacturing-related field. Figure 1 depicts the connections between the Industrial Internet,
IIoT, IoT, CPS, Industry 4.0, and the internet of things.

1.2 Deploying Edge Computing in IIoT

A sizable number of heterogeneous node devices make up the IIoT system, which is connected
by a variety of wired and wireless heterogeneous networks. Sensor networks, wireless Wi-Fi
networks, mobile communication 3G/4G/LTE/5G networks, and specialised industrial buses
are some examples of heterogeneous networks [12]. An edge network is made up of a large
number of dispersed heterogeneous industrial devices that gather industrial data in real time
and send it to a cloud server for processing and control. Traditional cloud data centre networks
fail to meet the real-time, security, and reliability needs of IIoT for huge data transmission and
processing as the IIoT expands and increases the scale of such networks. For instance, in a
smart factory, numerous production equipments gather a lot of sensory data; occasionally, the
data volume reaches the level of GB per second [13]. The amount of bandwidth resources used,
the cost of operation, and the overall processing latency all go higher if all the data are
transferred to the cloud platform for processing. Due of the fast-moving cars in the intelligent
connected vehicles scenario, the task is vulnerable to delays. The effective time phase of
information may always be missed if the cloud platform receives the data gathered by cars for
processing and decision-making [14]. Data transmission to the cloud platform for processing
and then back will have more catastrophic repercussions in some time-sensitive jobs, such as
emergency equipment shutdown in a smart factory and emergency braking in a smart connected
vehicle. EC, a research solution for the above issues, involves offloading some data from IIoT
cloud data centres to edge networks. In various IIoT applications, such as remote equipment
monitoring [15], [17], predictive maintenance [18], and quality control [19], EC is gradually
being used.

Using the computing, storage, and network resources dispersed on the pathways between data
sources and the cloud computing centre, EC is a new computing model that analyses and
processes a piece of data [20]. In order to execute local preparation of source data, make some
quick judgements, and upload computing results or pre-processed data to the cloud computing
centre, EC employs edge devices with adequate computing capability. Following are the
primary benefits of EC for the IIoT:

 Improve System Performance: The most significant contribution made by EC to the

IIoT is the achievement of ms level data processing, in addition to data collection and
transmission to the cloud platform. It is effective to decrease the system's overall
latency, lower the need for communication bandwidth, and raise the system's overall
performance.
 Protect Data Security and Privacy: Customers can get a whole system of centralised
data security protection solutions from cloud platform service providers. However, if
centrally stored data leaks, there would be catastrophic repercussions. EC in IIoT
enables businesses to implement the best security measures close at hand, lowering the
likelihood of data leakage during transmission and the amount of data retained in the
cloud platform to reduce security and privacy threats.
 Reduce Operational Costs: Data migration, enough bandwidth, and low latency
requirements come at a high operating expense when data is transmitted directly to the
cloud platform. EC in the IIoT may lower the amount of data uploaded, which lowers
the amount of data migrated, bandwidth usage, and latency, which lowers operating
expenses.

1.3 Motivation and Paper Organization

This study focuses on EC in the IIoT and sifts through several research accomplishments
related to EC in the IIoT, analysing the architecture, advancements, challenges, and
applications of EC in the IIoT. This is being done to aid in delineating the connection between
EC and the IIoT and to encourage their continued growth and integration. Following is the
format for the remaining portion of this report:

The studies and concepts linked to EC are compiled in Section II. Section III makes a
suggestion for the IIoT EC reference architecture. In Section IV, possibilities and problems
related to the growth of EC in the IIoT are next covered. A few EC applications for the IIoT
are also described in Section V. Section VI provides a conclusion to the report.
 CHAPTER – 2

Related Advances and Surveys of Edge Computing

Due to the fact that EC in the IIoT is still in its infancy while attracting a lot of interest, there
are just a few surveys that address its current state. As a result, in this section we first summarise
previous EC surveys including a variety of research approaches (for example, fog computing
(FC), mobile edge computing (MEC), edge networking, and edge security), and then we present
an overview and a comparison of the surveys connected to EC in IIoT. This will show how EC
is evolving, make it easier to grasp how EC has changed in the IIoT, and prove that our work
is innovative.

2.1 Fog Computing

Edge computing, which Cisco proposed in 2012 [21], is a notion that fog computing (FC) is
comparable to. By creating a fog layer, which consists of certain servers distributed between
the cloud platform and edge devices, Vaquero and Rodero-Merino characterised FC as a novel
computing paradigm that migrates workloads from the cloud computing centre to light servers
close to devices [22]. The main goals of EC and FC are to move data processing close to the
data source, minimise network transmission burden, and reduce decision-making delay. FC, on
the other hand, is concerned with the infrastructure between edge devices and the cloud
platform, whereas EC is also concerned with the massive number of edge devices in addition
to the infrastructure that FC was concerned with. Consequently, we think that FC is a part of
EC. Table I, which is based on [23], [24], compares FC with EC.

The study on FC has been summarised in different surveys. For instance, Naha et al. [26]
summarised the ideas, definitions, research trends, and architectures of FC; they also spoke
about the variations and constraints of the current research paths in resource allocation and
scheduling; they also put out their personal thoughts on these constraints. A thorough study of
FC was conducted by Mouradian et al. [27], who also reviewed the state of technology, looked
at fog system architecture, algorithms, research challenges, and future research directions. They
also speculated on the crucial role that fog may play in future emerging technologies like the
tactile internet. The overview and underlying principles of the FC architecture were outlined
by Mukherjee et al. [25], along with a summary of the major studies on resource and service
allocation networks, outstanding problems, and research trends.

Table I

Comparison of Fog Computing and Edge Computing

Features Fog Computing Edge Computing

Network Access Mostly LAN [25] LAN/WLAN [23]

Mobility Support Low [25], [26] High [23]

Latency Low [25]-[27] Low or very low, depending

on layers [24]
Scalability of Servers Distributed and low [25]-[27] Distributed and high [24]

Computing Infrastructure between cloud All computing resources

Power Location and terminals [25]-[27] except cloud [23], [24], [28]-
[30]
Distribution Distributed [25]-[27] Widely distributed [23], [24],
[28]
Reliability High [26], [27] Low [29], [30]
Deployment Cost Very low [26] Low [30]

Maintenance Experts of enterprise [25] Experts, engineers,

employees of enterprise [30]
Standardization Yes [27] No [29], [30]

2.2 MEC

The European Telecommunications Standards Institute (ETSI) put out the MEC idea. MEC's
central concept is the efficient use of computation and storage resources by moving some
requests and tasks from the mobile core network to the mobile access network. In order to
improve the user experience, introduce new changes to the operation modes of the telecom
operator, and establish a new industrial chain and network ecosystem, this concept deeply
integrates the traditional telecom cellular network with the Internet service. Its goals are to
reduce end-to-end delays in the delivery of mobile services and explore the inherent capabilities
of the wireless network. Multi-access edge computing, as defined by the ETSI in 2017, is the
extension of the MEC idea from cellular telecom networks to other access networks including
Wi-Fi and fixed access. As a result, multi-access edge computing (MEC) has been
enthusiastically adopted by researchers and is now commonly referred to by that term.

There are several surveys of MEC research accessible. Taleb et al. [28] discussed the current
standardisation effort and challenges of open research, analysed the MEC reference
architecture, and gave an overview of research on MEC and the major technologies for
implementing MEC. The use of MEC technology to develop IoT applications and their
synergies was thoroughly covered by Porambage et al. [29], who also covered the technical
elements of making MEC possible in the IoT and gave some insight into alternative integration
solutions. Moura and Hutchison [27] examined the primary difficulties of MEC services for
wireless resources based on classic and evolutionary games, and they explored some particular
elements influencing the effectiveness of game theory and how to manage the previously
mentioned components in order to establish a performance-cost balance in practical edge
network scenarios, as well as projected future trends and research areas in the use of game
theory in future MEC services. By illustrating automobile use scenarios connected to MEC,
Giust et al. [12] gave an in-depth explanation of the MEC technologies described and
investigated by the ETSI Industry Specification Group.

2.3 Edge Networking

EC networks are evolving in a multi-technology shift direction, and they are increasingly
integrating with innovative network protocols like software defined networks (SDN), network
function virtualization (NFV), and the fifth generation (5G) cellular network technology. SDN
makes network control and configuration easier, introduces many new functions to the network
by introducing virtualized components, and achieves reflecting control by liberating the
forwarding device from the routing management task and performing this independently on the
logically centralised control plane. Virtualized networks are encouraged to develop by the
concept of isolating the plane from the data-forwarding plane. A typical network's software
and hardware are separated by NFV, which also virtualizes network functions and makes it
possible to share network hardware resources. This allows for the quick deployment of network
functions and the flexible distribution of service capabilities as needed. Emerging
commercially accessible technologies for communication including massive MIMO, beam
formation, and micro-base-stations are brought together to create 5G [13]. Faster delivery rates,
minimised latencies, more enormous base stations, and improved quality of service (QoS) will
all be features of the 5G era.

Computing using fog will be a part of the integration of NFV and 5G, according to Lingen et
al. [34]. They also proposed an open, converging framework that uses MANO that offers
integrated governance for IoT services across the entire spectrum from the cloud to the edge.
According to Baktir et al. [23], SDN may be used as an accelerator to simplify EC and
maximise its capabilities. They mentioned the features of SDN, mixed it with the technical
shortcomings of EC executions, offered a clear SDN-EC interacting conjunction model,
demonstrated that SDN-related strategies can operate successfully in the framework for EC,
and discussed upcoming paths of SDN development to guarantee the successful fusion of EC
and SDN. To address the issue that the conventional NFV is challenging to implement at the
edge of an infrastructure, Cziva and Pezaros [15] created a container-based NFV structure.
Through three scenarios, they showed the framework's efficiency in lowering delays and
preserving the main network's consumption, thereby validating the potential for virtualization
at the edge of the network.

For the upcoming advancement of bright interconnected automobiles Huang et al. [16]
recommended a 5G-supported software-defined vehicle network (5G-SDVN). They
additionally utilised EC on mobile devices to improve the network management of 5G-SDVN,
creating a configurable, adaptable, and manageable 5G-SDVN network framework. They
explored unresolved issues with 5G-SDVN and used a case study of automotive cloud
computing to show how it may enhance usage of resources and facilitate handling networks.

2.4 Edge Security

By permitting certain data to be utilised and examined at the data origin, computing at the edge
not only reduces overall data communication but also eliminates the risk of data theft and
privacy breaches by preventing data transfer and communication. EC hence has the inherent
property of limiting data leakage and preserving data privacy. Despite this, computing at the
edge still has significant safety and confidentiality problems, including the possibility of attacks
on edge nodes, edge servers, or edge networks and the leakage of personal information as a
result of access privileges of the edge operating system or the irrational layout of the secured
conveyance procedure. Related studies are rather popular since several researchers have
explored the security issues in EC over an extended period of time, making some progress.
Zhang et al. [17] provided an overview of the fundamental definition and designs of EC,
examined the requirements, difficulties, and methods for data security and privacy, summarised
cryptographic-based techniques for resolving confidentiality and security of data problems,
looked into the edge relationship framework of the most cutting-edge data confidentiality as
well as safety solutions, and offered multiple open-access studies instructions for data
confidentiality in the area of EC.

Ni et al. [18] proposed the fundamental design and a number of offering edge-assisted IoT
applications, looked at the confidentiality, security, as well as performance difficulties related
to mobile computing at the edge in handling data, covered possibilities for enhancing the
computing effectiveness of data security in IoT with the aid of EC in encrypted data collection,
preserve data splitting, and reliable computational outsourcing, and offered a number of
intriguing areas for study. The architecture of data safety and confidentiality in FC was
examined by Guan et al. [19]. They highlighted the reasons why the data security strategy that
works in cloud computing doesn't work in FC and suggested a few data confidentiality and
safety design problems that FC has brought about.

2.5 Edge Computing in IIoT

In the IIoT, EC is currently in the rising stage, and several pertinent studies have been
conducted. Yet, due to the IIoT's early stages of development, only a small number of polls
include EC. Some studies on EC in the IIoT only took a tiny portion of the study areas into
account. In one case, Seitz et al. [23] published two research reports that demonstrate the
practicality of the fog architecture for IIoT applications. Georgakopoulos et al. [22] introduced
Internet of Things and edge cloud computing agenda for production. The primary EC reference
design concepts for Industry 4.0 that have been suggested by various organisations were
outlined by Sitton-Candanedo et al. [20] and analysed. FC was put up as a structural method to
execute the IIoT by Steiner and Poledna [24], who also covered the supporting technologies.
These studies in certain respects highlight the issues regarding EC in IIoT. They are not long
enough to provide a thorough and organised introduction, though. Consequently, these studies
are not included in the IIoT EC summary table.

Although required themes were somewhat addressed in several studies on EC in the IIoT, a
thoroughness of the debate was still lacking. For instance, Aazam et al. [40] outlined the IIoT
and industry 4.0 design, covered several use cases and new research problems, and addressed
how fog offered computing assistance in an IIoT context. It yet failed to explore the pertinent
cutting-edge technology, and the conversation of difficulties and application situations is also
very simplistic. Basir et al. [21] concentrated on FC and explored important problems as well
as the applications of FC in many areas. They also analysed the history of the industrial boom
and significant developments as the foundation of the industrial shift. Still, the study is not
sufficiently thorough since it seems more directly connected to IIoT than EC, and it only
concentrated on interaction and connectivity when discussing novel innovations and
difficulties. Table II provides an overview and analysis of several EC studies.
 CHAPTER – 3

Reference Architecture of Edge Computing in IIoT

A reference architecture is a thorough representation of advanced concept for a particular

technology type that is applied to direct the creation of software frameworks that are
appropriate for various application situations. EC integration with IIoT is now a trend that will
undoubtedly continue in the future. This part proposes an EC reference architecture for the IIoT
that uses the reference architectures already in use.

3.1 Existing Reference Architectures

To establish the quantity and capabilities of layers, architectures are typically offered based on
the technical, enterprise, and operational requirements of a specific situation. Since EC is still
a young technology, there is no one standard reference architecture for it. FC, an aspect of EC,
has several reference designs that have been suggested, many of which are based on the
standardised efforts of OpenFog. These reference architectural perspectives for FC are,
however, beginning to converge. In order to study the potential architecture of EC in the IoT,
one starting point may be the FC infrastructure. The majority of experts think that the reference
architecture for FC can be separated into three layers [25]: the equipment layer, the fog layer,
and the cloud computing layer [27]. The three levels have different names, but they all provide
essentially the same purposes. Typically, the first layer is made up of physical devices with
observing capabilities, which is in charge of gathering the necessary data. The next layer, which
processes some of the data, is made up of numerous gadgets that send data to the cloud channel.
The third layer, which is made up of cloud computing resources, is in charge of analysing large
amounts of data and making decisions.

Integrating EC into IIoT also necessitates having a fundamental grasp of the IIoT framework.
The hardware layer, transmission layer, and software layer are the three layers that make up a
common reference architecture for IIoT [9]. As an alternative, an infrastructure layer can be
appended in between communication and application levels to create a four-tier architecture
[25] from a three-tier design. Reference designs, like the standardised architecture for Industry
4.0, sometimes demand for modification and greater difficulty for some particular IIoT
businesses. Although the reference designs for FC and the IIoT are quite developed and
aligned, there has been limited study on the standard framework for EC in the IIoT since this
is an emerging field of study. A few studies have attempted to discuss the IIoT EC reference
model. As an example, Sitton-Candanedo et al. [30] described the EC reference model for
industry 4.0, while Aazam et al. [20] offered the fog-based IIoT reference model, which
includes equipment layer, fog layer, and cloud layer. These designs, however, simply described
a few levels in broad strokes and were unable to thoroughly explain all of them in relation to
IIoT features. Additionally, certain apparent trends should be taken into account with the
gradual integration of EC into IIoT as a way to render the reference design of edge computing
in IIoT more reasonable and applicable, such as the fact that traditional FC models sometimes
ignore the computational capacity of equipment in the fog layer, which can vary by numerous
orders of magnitude.

Table II

A Summary of Relatively Surveys of Edge Computing

Aspect Ref. Main Contribution Relevance to EC in IIoT

[16] An overview of architectures, Discussed the relationship of FC
open issues and future research and IoT and how to use it, but did
directions for FC. not focus on EC in IIoT.
[15] A comprehensive survey on Discussed combination of FC
Fog architectures, research with IoT, but no explicit focus on
Computing directions and prospects of FC. EC in IIoT.
(FC) [17] A survey of architecture, No focus on EC in IIoT.
applications, research aspects
and trends in FC.
[18] A survey on key enabling No explicit focus on EC in IIoT.
technologies, reference
architecture, and open
challenges for MEC.
MEC [19] A holistic overview on the Discussed the vital role MEC
exploitation of MEC will play in enabling future IIoT
technology for the realization of applications briefly.
IoT applications.
 A discussion of MEC-relevant
automotive use cases and the Discussed a typical application
[22] technologies specified and scenario of EC in IIoT.
investigated by the ETSI.

[24] An elaboration of architecture No explicit focus on EC in IIoT.

based on MANO models, and
the application of models for
the combination of NFV, 5G,
Edge and FC.
Networking [23] A comparative discussion of Discussed IoT and wireless
approaches and technologies, sensor networks (WSN) that can
capabilities, possible modes be learnt to deploy IIoT
and future directions for the networks, bot no focus on EC in
SDN. IIoT.
[25] A discussion of the Discussed the role of IIoT
opportunities of virtualization gateways played at edge
at the network edge, a networking.
container-based NFV platform
that runs lightweight container
VNF’s.
[28] An overview of architecture, Discussed some typical security
applications, opportunities and solutions for edge empowered
Edge Security interesting directions on edge IoT that can be used in EC in
empowered IoT data analysis. IIoT.
[29] A discussion of the design No explicit focus on EC in IIoT.
issues for data security and
privacy in FC.
[30] An overview of the architecture Discussed how to deploy FC in
of IIoT and industry 4.0, FC IIoT.
EC in IIoT approach to providing local
computing support, the core
elements and building blocks of
 IIoT, and some of the emerging
research challenges associated
with IIoT.
[31] A review of the application Discussed the prospectus and
areas of IIoT, key technologies technical challenges of
under industry transformation, deploying FC in IIoT.
FC’s solutions to key
challenges.

3.2 Proposed Reference Architecture of EC in IIoT

IIoT EC focuses on integrating EC into various IIoT situations to lessen traffic on the network
and latency in making decisions. As a result, the IIoT EC reference model has to be enhanced
and developed from the current EC reference models. Additionally, it must take into account
all of the IIoT and EC features. The proposed reference model is made up of three layers that
were taken from the current edge computing reference architectures: the device layer, the edge
layer, and the cloud application layer. The proposed reference model here takes into account
the features of IIoT and EC in a comprehensive manner, including what roles each layer ought
to perform and how layers interact with one another, opposite to the present EC models.

3.2.1 Device Layer

All types of sensors, portable devices, metres, intelligent machines, intelligent vehicles,
robotics, and other equipment or devices are included in this Layer. These gadgets use a variety
of wired or wireless networks to gather large amounts of variable data from different kinds of
sensors, pass on to the Edge Layer, and then wait for the Edge Layer's command instructions,
recognising the connectivity of data flow and control flow among the Device Layer and the
Edge Layer.

3.2.2 Edge Layer

The foundational element of the IIoT EC reference model is the edge layer. In addition to
offering immediate services like edge security and privacy protection, edge data analysis,
intelligent computing, process optimisation, and real-time control, this Layer is primarily in
charge of getting, analysing, and transmitting data flow from the Device Layer. The Edge Layer
can be separated into three sub-layers based on the data processing ability of various devices
in the Edge Layer: The Near-Edge Layer, Mid-Edge Layer, and Far-Edge Layer. This is
because the processing capacity of devices in the Edge Layer can occasionally vary by multiple
orders of magnitude.

3.2.3 Cloud Application Layer

The primary responsibility for extracting useful information from large data and realising
efficient allocation of assets within an organisation, an area, or sometimes the entire country
falls on this layer. In order to assist upper layer functions like product or process design,
enterprise comprehensive management, sales and after-sales service, and feedback model and
micro service, the cloud application layer receives data from the Edge Layer via the public
network. In order to accomplish more diverse and in-depth data value mining, this layer may
also exchange data via cloud collaboration across sets of different qualities, such as managers,
cooperative companies, developers, and clients. Such event’s decision times are often
expressed in terms of days.
 CHAPTER – 4

Challenges and Opportunities

Analysing the state of technological development reveals certain gaps in how well different
technologies address particular issues. The growth of novel innovations has created both
significant potential and obstacles for the advancement of EC in the IIoT. We examine some
of the potential and problems for EC in the IIoT in this part, covering edge connection using
5G, data offloading and load balancing, edge intelligence, and data sharing security.

4.1 5G Based Edge Communication

IIoT networks currently have two configuration options: wired mode and wireless mode [17].
In an IIoT network, the wired mode is more reliable, but the wireless mode is still in its early
stages of development due to its low transmission rate and unstable operation [18], [19]. High
transmission rate and minimal latency are the primary criteria and aims of IIoT EC, which
frequently schedules and exchanges data between edge devices and edge servers at each tier.
The creation and implementation of 5G, the next phase of cellular mobile communication
technology, will be crucial for network organisation and information exchange in the edge-
based IIoT system [23]. In terms of the rate of transfer and minimal delay necessary to minimise
certain financial losses and life-threatening emergencies, like sudden interrupting of industrial
machinery or autopilot vehicles, an IIoT system powered by EC can precisely meet these needs
thanks to the three key features of 5G technology, including an ultra-high rate, super-large
connection, and ultra-low delay [16]. Additionally, 5G may remove a lot of gadgets from wires,
increasing physical deployment flexibility and requiring less maintenance. Despite the bright
industrial future, there are still plenty of possible issues with the integration of 5G into an EC-
based IIoT. EC-based 5G and IIoT integration may accelerate the invention of common IIoT
application scenarios including VANET, digital twin, remote control, and maintenance. Long-
term integration continues to face several of the probable technological difficulties listed above.

4.2 Data Offloading and Load Balancing

A distributed system with a range of data sources and distributed computing power makes up
an IIoT system based on EC [18]. Task dividing and data offloading must be taken into account
when an individual device or edge server hits its computing power limit. The need for
appropriate load balancing is greater when an item's performance is poor, as in the MEC
scenario [19]– [21]. To accomplish the total load balancing of the system, jobs and data are
divided and routed to various devices or servers, and outcomes from all of them are combined.

4.2.1 Data Offloading

While there are many devices in a typical IIoT system, gadgets are often not taken into account;
instead, data offloading between a few big servers is the only thing that has to be taken into
account. The complexity of this challenge would, however, grow exponentially in an IIoT
system based on EC, where the data offloading of every gadget with data handling features
should be taken into account. Additionally, data offloading will often take place to equalise a
system's total load due to the restricted computational and storage capabilities of the edge
devices. As a result, the edge network will experience high throughput, significant bandwidth
resource utilisation, and eventually, it's possible that task delays may become more noticeable
overall.

4.2.2 Load Balancing

Time delay and energy expenditure are typically taken into account in tandem when
determining the data offloading strategy. However, the computed plan can cause some devices
to be overloaded, creating a new issue with load balancing [24]. The goal of load balancing is
to achieve equitable load distribution across edge nodes and reliable communication links,
resulting in an equilibrium between the use of computing resources and network resources. In
order to adapt the characteristics of the IIoT edge system, it is important to modify the current
load balancing techniques due to the size and frequency of scheduling being considerably
greater.

A load balancing service may establish the load balancing tactics based on a machine learning
model for each layer in the IIoT based on EC, which sings the task data and bringing data from
the devices and edge servers of the far-edge, mid-edge, and near-edge layers. In order to reduce
the difficulty of scheduling and routing, load balancing scheduling may also be carried out
using SDN from the edge network's worldwide viewpoint.
4.3 Edge Intelligence

Edge gadgets can only carry out modest computing activities in existing IIoT systems based
on EC. Edge intelligence (EI) is implemented on the edge of the IIoT to make edge devices
and edge servers smarter so they can carry out more sophisticated activities with a greater data
handling efficiency and reduced latency. This has developed into an IIoT EC concept.
However, it is possible to train an AI model to generate predictions and make judgements with
high precision, but this requires a significant quantity of training and validation data. Due to
the constrained computation and storage capabilities of edge devices, training and using the AI
model might be challenging. To find a solution to the tension between the AI model's intricate
nature and the edge device’s scarce resources. There are two fundamental strategies: boosting
edge device processing capability and streamlining or dividing the AI model installed on edge
gadgets.

4.4 Data Sharing Security

The enormous volume of real-time data through numerous devices, locations, and Facilities is
one of the benefits of IIoT. Industrial output effectiveness will be greatly increased through
data mining and multi-dimensional corporate decision-making [21]. The IIoT system
continuously produces data islands because the typical IIoT system is governed by vertical,
closed programmes that are only concerned with ensuring the appropriate operation of a single
unit or location. IIoT gains greater versatility and splits the data by using EC. The topic of data
security sharing is becoming somewhat more complicated, though. A tendency of EC-based
IIoT systems will be to effectively knock down data islands and safely share identical real-time
data with any kind and number of specialised programmes. There are two issues with edge data
sharing: the efficiency of edge devices is constrained, and effective security techniques are
sometimes challenging to execute instantly on edge devices; and the unavoidable growth of
data ports may result in more catastrophic effects, such as infiltration and loss. The adoption
of blockchain in IIoT EC presents novel possibilities and problems for the safe exchange of
data [22]. Additionally, blockchain presents commercial opportunities in a variety of emerging
contexts, such as the potential to bill for future IIoT services. Blockchain thereby holds great
potential for programmes that ensure the integrity of edge data sharing. To ensure safe edge
data exchange, there hasn't been much study done yet on the specifics of methodically
integrating blockchain with EC in IIoT. Thus, there are still several issues that require
investigation and solutions.
 CHAPTER-5

Application Scenarios

This section presents multiple application scenarios including Prognostic and Health
Management (PHM), Smart Grids, Manufacturing, Intelligent Connected Vehicles (ICV) to
demonstrate how EC in IIoT is implemented in real-world applications.

5.1 PHM

PHM is a cutting-edge solution that is highly appreciated by businesses all over the world [25].
It uses the operational data of the system elements gathered by different sensors, monitors and
assesses the system's general health, and predicts and takes necessary action before system
breakdown. The PHM system should typically be able to identify performance issues, faults,
isolate and anticipate faults, perform advanced diagnosis, manage health, monitor component
life, etc. In order to forecast the timing and location of a breakdown in advance, PHM connects
the self-sustaining assistance network with the joint distributed information system (JDIS).
This leads to lower maintenance costs, increased operational dependability, and condition-
based repairs.

PHM is used to keep an eye on element deterioration, ageing, corrosion, and breakdown while
the device is in use. This helps to avoid unanticipated downtime, which poses a major risk to
people's safety and property. A crucial IIoT application case, the PHM system makes use of
several sensors to track the state of elements in real-time [26]. The introduction of EC to PHM
and executing the initial data processing adjacent to sensors will substantially decrease the
quantity of data submitted to the cloud and decrease the decision delays of emergency events.
This is due to the enormous volume of sensors and data, as well as the unexpected
consequences that might arise from a delay of the submitting and final choice in the cloud.

The safety monitoring of railway tracks is one of the primary EC applications in IIoT in PHM
at the moment, which are rather broad. The benefits of EC are used for real-time feature
extraction [27] and anomaly detection of railway tracks, tracking the real-time effectiveness of
railway and train operations, and forecasting possible failures to avoid unexpected downtime
and assist optimisation choices. Drones can also be a source of data for the detection of railway
tracks [28].
5.2 Smart Grid

One prominent IIoT application case is the smart grid. Its goal is to facilitate node monitoring
and information exchanges for the transfer of electrical energy from the power plant to the
customers [29]. A smart power grid has an advantage over a typical power grid in that it
integrates generation, transmission, distribution, safety precautions, and other linkages utilising
cutting-edge information technology. In order to increase the overall effectiveness of the power
grid, power grid firms and consumers can access real-time information about the system's
condition and electricity use. The smart grid now has a significant number of smart metres and
different kinds of sensing equipment deployed. The immediate data volume is enormous, the
data kinds are diverse, and the overall framework is intricate. To address the issues mentioned
above, an edge server can be installed close to smart metres and other sensing equipment. In
order to implement regional oversight of equipment and energy efficiency optimisation, the
obtained data is analysed on the spot, and some decisions are taken. This increases management
effectiveness and satisfies real-time demands. Each edge server gathers and sends the data
necessary for maintaining machinery and optimising building structures to the cloud centre,
which centrally processes, analyses, and trains the gathered data.

The edge computing-based smart grid system may efficiently route the intelligent assets of the
entire power system nearer to end-users [25], effectively notice the grid structure [30],
distribute computing, storage, and control services to the edge network, and accomplish high-
demand acts like smart low-voltage area management, user energy administration, and
distribution of outside forces damage risk tracking [31].

5.3 Manufacturing

There are several types of industrial equipment, a significant amount of real-time data, intricate
communication network topologies, and high-performance needs for the correct and timely
transfer of data [32]. Synchronisation of software management tools with manufacturing tools
on an industrial site is an important task. EC deployment in an industrial manufacturing site,
particularly when coupled with NFV technology and real-time network communication
technology, can create an excellent network link from the cloud platform to the industrial field
EC platform, accomplish flexible and detached application deployment capabilities, and
provide smart, real-time, secure, and quality-guaranteed industrial network and EC.
Real-time processing of photos or videos, which may be used for product defect identification
and categorization [29], worker motion correction, or equipment component assembly error
checking [33], is a typical use for production site coordination. Acquisition, preprocessing,
image segmentation, feature extraction, and matching recognition are all steps in the processing
of images and videos [34], and these steps are typically combined with an artificial intelligence
method. The image detection model is trained beforehand using industrial image or video
datasets. It is frequently important to employ an incremental learning method to continuously
adjust the model to increase the recognition accuracy in order to attain a higher performance
[35]. In order to ensure recognition accuracy and shorten recognition delays, EC technology
may retain the combination identification process on the EC infrastructure while the training
procedure is kept on the cloud platform.

Manufacturing facilities use a variety of network connection techniques, like commercial

Ethernet or Fieldbus, each of which has many procedures, making it challenging to link them.
To fix the connection issue between the diverse networks of an industrial network, the
EC architecture may transform disparate protocols into a single protocol. At the same time, an
EC platform for a production location offers administration and data layouts and makes use of
a portable network and application virtualization governance to remotely control, improve, and
care for a huge number of gadgets and programmes for remote setup and tracking. Additionally,
it combines chip-level protected launching and encrypted key verification with data cleaning
and desensitisation to assure reliability of data and the confidentiality of critical data, creating
an encrypted atmosphere for industrial networks.

5.4 ICV

ICV will play a significant role in future IIoT scenarios when the 5G era arrives [26]. ICV's
primary solution will be the Edge-Cloud interaction. When it comes to handling rather difficult
tasks like traffic predictions in a specific region, cloud computing is like the super brain of
automobiles [24], [27]. EC is also comparable to the nerve endings in cars, performing some
"subconscious" functions such autonomous emergency braking or gathering driving data from
nearby vehicles.

One of the main areas of EC research used in ICV is automatic driving. Regional automatic
driving is rather straightforward. In order to achieve the autonomous operation of the vehicle
in the authorised area, the course and speed of a vehicle are often planned in advance using
surroundings information of the whole running zone as input to the system. EC technology is
utilised to evaluate the picture data from the camera with the current road information in the
onboard system and make a judgement, realising a rapid response, in the case of a sudden entry
of people or cars. When compared to conventional autonomous driving, responsive
autonomous driving is more complicated and variable. A variety of situations and aspects need
to be taken into account, including autonomous parking, lane-change support, intersection
passage, and path planning. Under these circumstances, delays in transferring data to the cloud
cannot tolerate delays in nearby vehicle recognition, traffic light recognition, and intrusion of
emergency obstructions. Therefore, in the scenario described above, EC will act as the
processing hub. Prioritising events is still necessary in order to decide which events are handled
on the edge system and which are handled on the main-board system.

ICV relies on the amenities and entertainment provided on board. The edge server is mounted
to the side of the road and allows for real-time communication with the vehicle while it is
moving at high speed. This is comparable to a mobile phone's service mode, although a mobile
phone and an automobile have distinct travelling distances and speeds. As a result, the ICV
application situation is somewhat comparable with the MEC study [34]. In order to fulfil the
requirements of this situation, the regular communications between cars, edge servers, and the
cloud infrastructure frequently need effective data navigation, caching, and offloading
solutions.
 CONCLUSION

IIoT links gadgets as a growing industrial solution with the help of cutting-edge communication
technologies. This allows the system to monitor, gather, exchange, and analyse data, make
high-value decisions in a previously unheard-of manner, and increase industry efficiency and
effectiveness. Utilising EC technology in the IIoT may analyse a portion of massive amounts
of real-time sensing data near to the data source on the edge of the network, resolving the main
issues of constrained transmission capacity and prolonged latency when making decisions from
cloud platforms.

In this work, the development and integration process of IIoT and EC are explained, and the
reference model of EC in IIoT is provided through a thorough examination of EC in IIoT. In
the IIoT, there are two layers—the cloud layer and the edge layer—and three sublayers inside
the edge layer. A thorough elaboration for the advanced EC methods used in IIoT from the
viewpoints of routing, job scheduling, data storage and analytics, security, and standardisation
have been created. From the viewpoints of load balancing, edge AI, safe data sharing, and 5G-
based edge communication, the difficulties of EC in the IIoT are analysed and debated. EC in
conjunction with blockchain, machine learning, SDN, and 5G will undoubtedly become
popular. In an effort to aid in the marketing and implementation of EC in IIoT, we also
addressed a number of common application cases. In contrast to other IIoT or EC survey
papers, the primary objective of this paper is to emphasise the combined application of IIoT
and EC, attempting to clarify the significance of the future use of EC in IIoT, in order to garner
greater interest for the issue from other scholars and facilitate quicker and easier industry
development.
 References

[1] A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, and M. Ayyash, “Internet of Things: A

survey on enabling technologies, protocols, and applications,” IEEE Commun. Surveys Tuts., vol. 17,
no. 4, pp. 2347–2376, 4th Quart., 2015.

[2] E. Sisinni, A. Saifullah, S. Han, U. Jennehag, and M. Gidlund, “Industrial Internet of Things:
Challenges, opportunities, and directions,” IEEE Trans. Ind. Informat., vol. 14, no. 11, pp. 4724–4734,
Nov. 2018.

[3] P. P. Gaikwad, J. P. Gabhane, and S. S. Golait, “A survey based on smart homes system using
Internet-of-Things,” in Proc. Int. Conf. Comput. Power Energy Inf. Commun. (ICCPEIC), Chennai, India,
2015, pp. 0330–0335.

[4] L. Xu, W. He, and S. Li, “Internet of Things in industries: A survey,” IEEE Trans. Ind. Informat., vol.
10, no. 4, pp. 2233–2243, Nov. 2014.

[5] D. S. Nunes, P. Zhang, and J. S. Silva, “A survey on human-in-the-loop applications towards an

Internet of all,” IEEE Commun. Surveys Tuts., vol. 17, no. 2, pp. 944–965, 2nd Quart., 2015.

[6] I. Stojmenovic, “Machine-to-machine communications with in-network data aggregation,

processing, and actuation for large-scale cyberphysical systems,” IEEE Internet Things J., vol. 1, no. 2,
pp. 122–128, Apr. 2014.

[7] M. Ammar, G. Russello, and B. Crispo, “Internet of Things: A survey on the security of IoT
frameworks,” J. Inf. Security Appl., vol. 38, pp. 8–27, Feb. 2018.

[8] A. C. Panchal, V. M. Khadse, and P. N. Mahalle, “Security issues in IIoT: A comprehensive survey of
attacks on IIoT and its countermeasures,” in Proc. IEEE Global Conf. Wireless Comput. Netw. (GCWCN),
Lonavala, India, 2018, pp. 124–130.

[9] H. Xu, W. Yu, D. Griffith, and N. Golmie, “A survey on Industrial Internet of Things: A cyber-physical
systems perspective,” IEEE Access, vol. 6, pp. 78238–78259, 2018.

[10] Y. Lu, “Industry 4.0: A survey on technologies, applications and open research issues,” J. Ind. Inf.
Integr., vol. 6, pp. 1–10, Jun. 2017.

[11] J. Li, F. R. Yu, G. Deng, C. Luo, Z. Ming, and Q. Yan, “Industrial Internet: A survey on the enabling
technologies, applications, and challenges,” IEEE Commun. Surveys Tuts., vol. 19, no. 3, pp. 1504–
1526, 3rd Quart., 2017.
[12] T. Qiu, B. Li, X. Zhou, H. Song, I. Lee, and J. Lloret, “A novel shortcut addition algorithm with particle
swarm for multisink Internet of Things,” IEEE Trans. Ind. Informat., vol. 16, no. 5, pp. 3566–3577, May
2020.

[13] P. K. Illa and N. Padhi, “Practical guide to smart factory transition using IoT, big data and edge
analytics,” IEEE Access, vol. 6, pp. 55162–55170, 2018.

[14] A. Thakur and R. Malekian, “Fog computing for detecting vehicular congestion, an Internet of
Vehicles based approach: A review,” IEEE Intell. Transp. Syst. Mag., vol. 11, no. 2, pp. 8–16, Mar. 2019.
[15] H. Wang, Q. Wang, Y. Li, G. Chen, and Y. Tang, “Application of fog architecture based on multi-
agent mechanism in CPPS,” in Proc. 2nd IEEE Conf. Energy Internet Energy Syst. Integr. (EI2), Beijing,
China, 2018, pp. 1–6.

[16] N. Yoshikane et al., “First demonstration of geographically unconstrained control of an industrial

robot by jointly employing SDNbased optical transport networks and edge compute,” in Proc. 21st
OptoElectron. Commun. Conf. (OECC) Held Jointly Int. Conf. Photon. Switching (PS), Niigata, Japan,
2016, pp. 1–3.

[17] I. A. Tsokalo, H. Wu, G. T. Nguyen, H. Salah, and F. H. P. Fitzek, “Mobile edge cloud for robot control
services in industry automation,” in Proc. 16th IEEE Annu. Consum. Commun. Netw. Conf. (CCNC), Las
Vegas, NV, USA, 2019, pp. 1–2.

[18] T. M. Jose, “A novel sensor based approach to predictive maintenance of machines by leveraging
heterogeneous computing,” in Proc. IEEE SENSORS, New Delhi, India, 2018, pp. 1–4.

[19] L. Li, K. Ota, and M. Dong, “Deep learning for smart industry: Efficient manufacture inspection
system with fog computing,” IEEE Trans. Ind. Informat., vol. 14, no. 10, pp. 4665–4673, Oct. 2018.

[20] W. Shi, J. Cao, Q. Zhang, Y. Li, and L. Xu, “Edge computing: Vision and challenges,” IEEE Internet
Things J., vol. 3, no. 5, pp. 637–646, Oct. 2016.

[21] B. Flavio, M. Rodolfo, Z. Jiang, and A. Sateesh, “Fog computing and its role in the Internet of
Things,” in Proc. 1st Edition MCC Workshop Mobile Cloud Comput., 2012, pp. 13–16.

[22] L. M. Vaquero and L. Rodero-Merino, “Finding your way in the definition fog: Towards a
comprehensive of fog computing,” ACM SIGCOMM Comput. Commun. Rev., vol. 44, no. 5, pp. 27–32,
2014.

[23] A. C. Baktir, A. Ozgovde, and C. Ersoy, “How can edge computing benefit from software-defined
networking: A survey, use cases, and future directions,” IEEE Commun. Surveys Tuts., vol. 19, no. 4, pp.
2359–2391, 4th Quart., 2017.
[24] P. Mach and Z. Becvar, “Mobile edge computing: A survey on architecture and computation
offloading,” IEEE Commun. Surveys Tuts., vol. 19, no. 3, pp. 1628–1656, 3rd Quart., 2017.

[25] M. Mukherjee, L. Shu, and D. Wang, “Survey of fog computing: Fundamental, network
applications, and research challenges,” IEEE Commun. Surveys Tuts., vol. 20, no. 3, pp. 1826–1857, 3rd
Quart., 2018.

[26] R. K. Naha et al., “Fog computing: Survey of trends, architectures, requirements, and research
directions,” IEEE Access, vol. 6, pp. 47980–48009, 2018.

[27] C. Mouradian, D. Naboulsi, S. Yangui, R. H. Glitho, M. J. Morrow, and P. A. Polakos, “A

comprehensive survey on fog computing: State-ofthe-art and research challenges,” IEEE Commun.
Surveys Tuts., vol. 20, no. 1, pp. 416–464, 1st Quart., 2018.

[28] T. Taleb, K. Samdanis, B. Mada, H. Flinck, S. Dutta, and D. Sabella, “On multi-access edge
computing: A survey of the emerging 5G network edge cloud architecture and orchestration,” IEEE
Commun. Surveys Tuts., vol. 19, no. 3, pp. 1657–1681, 3rd Quart., 2017.

[29] P. Porambage, J. Okwuibe, M. Liyanage, M. Ylianttila, and T. Taleb, “Survey on multi-access edge
computing for Internet of Things realization,” IEEE Commun. Surveys Tuts., vol. 20, no. 4, pp. 2961–
2991, 4th Quart., 2018.

[30] I. Sitton-Candanedo, R. S. Alonso, S. Rodriguez-Gonzalez, J. A. Garcia Coria, and F. De La Prieta,

“Edge computing architectures in industry 4.0: A general survey and comparison,” in Proc. 14th Int.
Conf. Soft Comput. Models Ind. Environ. Appl., vol. 950, 2020, pp. 121–131.

[31] J. Moura and D. Hutchison, “Game theory for multi-access edge computing: Survey, use cases,
and future trends,” IEEE Commun. Surveys Tuts., vol. 21, no. 1, pp. 260–288, 1st Quart., 2019.

[32] F. Giust et al., “Multi-access edge computing: The driver behind the wheel of 5G-connected cars,”
IEEE Commun. Stand. Mag., vol. 2, no. 3, pp. 66–73, Sep. 2018.

[33] M. Agiwal, A. Roy, and N. Saxena, “Next generation 5G wireless networks: A comprehensive
survey,” IEEE Commun. Surveys Tuts., vol. 18, no. 3, pp. 1617–1655, 3rd Quart., 2016.

[34] F. Lingen et al., “The unavoidable convergence of NFV, 5G, and fog: A model-driven approach to
bridge cloud and edge,” IEEE Commun. Mag., vol. 55, no. 8, pp. 28–35, Aug. 2017.

[35] R. Cziva and D. P. Pezaros, “Container network functions: Bringing NFV to the network edge,” IEEE
Commun. Mag., vol. 55, no. 6, pp. 24–31, Jun. 2017