Network Slicing

IEEE COMMUNICATIONS SURVEYS & TUTORIALS, VOL. 23, NO.
2, SECOND QUARTER 2021 957
Survey on Network Slicing for Internet of Things

Realization in 5G Networks
Shalitha Wijethilaka , Student Member, IEEE, and Madhusanka Liyanage , Senior Member, IEEE,
Abstract—Internet of Things (IoT) is an emerging technology

that makes people’s lives smart by conquering a plethora of
diverse application and service areas. In near future, the fifth-
generation (5G) wireless networks provide the connectivity for
this IoT ecosystem. It has been carefully designed to facilitate
the exponential growth in the IoT field. Network slicing is one of
the key technologies in the 5G architecture that has the ability
to divide the physical network into multiple logical networks
(i.e., slices) with different network characteristics. Therefore,
network slicing is also a key enabler of realisation of IoT in
5G. Network slicing can satisfy the various networking demands
by heterogeneous IoT applications via dedicated slices. In this
survey, we present a comprehensive analysis of the exploita- Fig. 1. Evolution of the Internet.
tion of network slicing in IoT realisation. We discuss network
slicing utilisation in different IoT application scenarios, along
with the technical challenges that can be solved via network
slicing. Furthermore, integration challenges and open research application scenarios demand various performance require-
problems related to the network slicing in the IoT realisation ments such as low latency, ultra-reliability, high security and
are also discussed in this paper. Finally, we discuss the role of high data rates.
other emerging technologies and concepts, such as blockchain Most of these connected devices will use the wireless
and Artificial Intelligence/Machine Learning (AI/ML) in network infrastructure to communicate. However, the existing wireless
slicing and IoT integration.
infrastructure is not able to handle the rapid growth of IoT
Index Terms—Network slicing, IoT, 5G, SDN, NFV, network connections along with typical mobile connections. Moreover,
architecture, latency, reliability. such networks are incapable of satisfying the heterogeneous
QoS requirements for different IoT application scenarios. As
a solution, the 5G architecture is designed with the aid of
various new technologies to handle these requirements.
I. I NTRODUCTION Software Defined Networking (SDN), Network Function
NTERNET has evolved over the last four decades, from
I simple peer-to-peer networks to an advance IoT ecosystem
(Figure 1). With the ubiquitous utilisation of IoTs, everything
Virtualization (NFV), Multi-access Edge Computing (MEC)
and network slicing are some of the predominant technologies
in 5G architecture [4]–[7]. SDN offers the ability to control
around us is becoming smart. It is possible to connect people the network traffic routing centrally and intelligently, using
and things at any time from any place with network access, to software applications [8]. Centralised control, network pro-
receive information of the thing, to operate the thing or even grammability and abstraction are the key benefits of SDN.
both, to make the life of the mankind easier [1]. The number NFV is a concept which is used to package network func-
of connected devices proliferates exponentially. According to tions such routing, load balancing and firewalls as software
Machina research, it is expected that the number of connected applications, so that they can run on commodity and general
devices will be 27 billion by 2024 [2]. Diversity of these hardware devices [9]. Tight coupling between network func-
IoT applications will also expand from simple smart home tions and specific hardware units was a huge bottleneck to
solutions to mission-critical health care systems [3]. These the evolution of such network functions. This can be elimi-
nated by using NFV. Providing cloud computing capabilities
Manuscript received September 25, 2020; revised December 16, 2020 to the edge network is the main purpose of MEC technol-
and January 26, 2021; accepted March 11, 2021. Date of publication
March 22, 2021; date of current version May 21, 2021. (Corresponding ogy [10]. It will minimise the network congestion on backhaul
author: Shalitha Wijethilaka.) and improve the resource optimisation, user experience and the
Shalitha Wijethilaka is with the School of Computer Science, overall performance of the network [11].
University College Dublin, Dublin 4, D04 V1W8, Ireland (e-mail:
mahadurage.wijethilaka@ucdconnect.ie). Dividing the physical network into separate logical networks
Madhusanka Liyanage is with the School of Computer Science, University known as slices is called network slicing. Each slice can be
College Dublin, Dublin 4, D04 V1W8, Ireland, and also with the Centre for configured to offer specific network capabilities and network
Wireless Communications, University of Oulu, 90014 Oulu, Finland (e-mail:
madhusanka.liyanage@oulu.fi). characteristics [12]. Thus, the End-to-End (E2E) network slic-
Digital Object Identifier 10.1109/COMST.2021.3067807 ing can help to deploy various 5G based services [13]–[18].
1553-877X
c 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: University of Exeter. Downloaded on May 28,2021 at 02:53:50 UTC from IEEE Xplore. Restrictions apply.
958 IEEE COMMUNICATIONS SURVEYS & TUTORIALS, VOL. 23, NO. 2, SECOND QUARTER 2021
TABLE I
S UMMARY OF ACRONYMS slice for each service scenario is needed, to allocate required
resources to fulfill the such requirements.
Similarly, different IoT applications demand different
networking requirements. For instance, more than 99.99% reli-
ability level and less than 1 ms E2E latency are required for
mission-critical communication IoT use cases. 10−9 packet
loss rate and E2E latency range of 250 µs to 10 ms are
required factory automation applications [20]. Such diverse
requirements of different IoT applications can be fulfilled by
allocating a dedicated E2E network slice for each application
(Figure 2).
The Network Slice as a Service (NSlaaS) concept allows
operators to create customised network slices for their cus-
tomers as a service. This generates a new business model for
Mobile Network Operators (MNOs) [31], [32]. Moreover, an
E2E network slicing framework that has the ability to hori-
zontally slice computation and communication resources was
proposed in [33] for supporting vertical industry applications.
Network slicing improves the efficiency of resource utili-
sation by dynamically adjusting network resources between
the slices. It is possible to implement autonomous systems
for such resource allocations and dynamic adjustment of
resources [34]. This method can be used to improve the
scalability of future IoT applications.
High security and privacy for sensitive IoT applications,
such as healthcare, can be achieved through the slice iso-
lation. For instance, a secure service-oriented authentication
framework has been proposed by Ni et al. for the realisa-
tion of IoT services in 5G networks through fog computing
and network slicing [35]. Since IoT devices are generally
resource-constraint, they are highly vulnerable to attacks. The
rapid growth in the number of IoT devices and pervasiveness
are also attracting many attackers to the IoT ecosystem [36].
An adversary can easily dominate IoT devices and can lead
them to generate Distributed Denial-of-Service (DDoS) attacks
to the network. The severity of these kinds of attacks can be
minimised through isolating IoT applications using network
slicing. Also, dynamic allocation of idle resources to the vic-
timised network slice is possible to keep the service without
any degradation during an attack.
B. Paper Motivation
IoT expands its roots over each and every field by permut-
ing unintelligible dumb items into smart things over the last
few years. A large number of vertical industries are engaged
in developing smart solutions at a very high speed and the
community rapidly adopts these into their lifestyle. Multiple
SDN and NFV can be identified as key enabling technologies surveys related to IoT have been conducted during the past few
for network slicing realisation in 5G networks [19]. years over several subareas. Standardisation, security, archi-
tecture, privacy, trust, visions, challenges and future research
A. Role of Network Slicing for IoT directions are some key areas that are exploited in existing
In a nutshell, 5G network architecture is designed to sup- surveys and researches.
port three main fundamental service classes: enhanced Mobile Network slicing is a neoteric technology and numerous
BroadBand (eMBB), massive Machine Type Communications network slicing-related areas have to be investigated. The
(mMTC) and Ultra-Reliable Low-Latency Communications standardisation of network slicing is still ongoing by telecom-
(URLLC) [12]. Each service class has a diverse set of network munication organisations. Though it is possible to find surveys
requirements. Therefore, the creation of an E2E network and researches associated with network slicing within the past
WIJETHILAKA AND LIYANAGE: SURVEY ON NETWORK SLICING FOR IoT REALIZATION IN 5G NETWORKS 959
Fig. 2. E2E Network Slicing.
TABLE II
S UMMARY OF I MPORTANT S URVEYS ON N ETWORK S LICING
few years, most of them tend to be conceptual rather than In [26], Foukas et al. reviewed state-of-the-art network slicing
technical. Table II consists of a summary of existing surveys and they have presented a framework and evaluated it under
related to network slicing. Slicing architecture, technologies the maturity of current proposals. Resource allocation algo-
that support network slicing realisation, standards, security and rithms are analysed in [27]. Issues in allocation, isolation,
use cases are the areas that are covered by most surveys. guaranteeing the inter-operability of the resources between
In [21], they have carried out a comprehensive review slices discussed in [28]. Enabling smart services in future
of network slicing with enabling technologies, standardisa- networks with network slicing and different parameters in
tion efforts, industrial projects that accelerate network-slicing network slicing are discussed in [29]. Future research areas of
usage and future research directions. The latest status of network slicing with open research problems are highlighted
3GPP standardisation, solutions to reduce the complexity in [19], [21], [22], [25], [26], [28], [29].
introduced by network slicing and future research areas dis- As with the references that we could find, there is
cussed in [22]. Network slicing architecture and some par- no single survey that specifically analyses the contri-
ticular technologies that can be used with network slicing bution of network slicing to the IoT realisation. Since
are discussed in [23]. Reference [24] discussed a signifi- network slicing is the prominent technology that rein-
cant use case in future IoT: smart grid, E2E network slicing forces IoT realisation in 5G networks, it is pertinent
with concepts, technologies, solutions and use cases are con- to analyse the associativity between network slicing and
sidered in [19]. The network slicing concept with aspects IoT, in terms of technical aspects, applications and
that support network slicing realisation is discussed in [25]. challenges.
Fig. 3. Paper Organization.
C. Our Contribution 4) A concise summary of the network slicing-related

In our survey, we broadly investigate how network slic- projects that have a impact on IoT.
ing will overcome the challenges of IoT technologies and 5) A discussion on the impact of the emerging technologies
their related applications in future networks. Since there are in network slicing and IoT.
enough surveys related to network slicing integration technolo- 6) A pool of future research directions in applying network
gies, such as SDN and NFV, they will not be covered in this slicing on several IoT applications and technical aspects.
survey. A comprehensive overview of the state-of-the-art tech-
nical aspects that can be used to IoT realisation via network D. Paper Organization
slicing will be provided through the survey. The impact of the The paper consists of nine sections and the organisation
emerging technologies and concepts, such as blockchain and is as follows. Section II provides the background knowl-
AI/ML, in network slicing and different IoT applications will edge of network slicing in order to understand the concept.
be discussed in here. Our contribution to the IoT realisation via Section III focuses on improving the technical aspects related
network slicing through this paper is enumerated as follows. to IoT via network slicing. Scalability, dynamicity, security,
1) Technical aspects that can be improved via network privacy, QoS, E2E orchestration and resource allocation and
slicing in the IoT realisation. prioritisation are discussed in here. Network slicing utilisation,
2) Detailed discussion on the utilisation of network slic- in several IoT application areas, is described in Section IV.
ing for the realisation of divergent IoT use cases with Section V is allocated to discuss the technical challenges of
different network requirements. network slicing that can rise with the advancement of IoT.
3) A comprehensive analysis of the technical challenges of Network slicing-related projects are described in Section VI
network slicing that will rise due to the advancement of and Section VII is dedicated to discuss the lessons learned
the IoT. in each application and technical aspect with future research
Fig. 4. Overall Network Slicing Architecture.
directions respectively. The impact of emerging technolo- • Infrastructure layer is responsible for providing
gies in the IoT realisation via network slicing is described physical or virtual resources such as storage, computing
in Section VIII. Finally, Section IX concludes the paper. resource and connectivity.
Graphical representation of the paper organisation is shown in • Network slice instance layer which runs over the infras-
Figure 3. We included the definitions of the frequently used tructure layer, consists of NSIs which form E2E logical
acronyms in Table I. network slices.
• Service instance layer which runs over all other layers,
represents end user and business services. These services
II. N ETWORK S LICING will be provided via service instances by the network
A. What Is Network Slicing? operator or by a third party.
Future mobile networks will have heterogeneous service A functional network slice basically consists of two sub-
requirements due to the wide set of new networking services. slices: the RAN subslice, specific to Next-Generation Radio
Thus, the ‘one size fits all’ networking concept is not Access Network (NG RAN) and the core subslice, specific to
applicable for 5G and beyond. Next Generation Mobile core network. An overview of the network slicing architecture
Network(NGMN) Alliance first introduced the network slicing is shown in Figure 4 and it mainly consists of four seg-
concept in 2015 to address the above issue. Third Generation ments: RAN subslice, slice-paring functions, core subslice
Partnership Project (3GPP) considers network slicing as a key and the NSM. Fully data flow of some particular applica-
feature in 5G networks [11]. The concept of dividing the phys- tions throughout the network is shown in the Figure 4. A
ical network into multiple logical networks (network slices) fully functional network slice is able to route and control a
so that each logical network can be specialised to provide particular packet over the network without influencing other
specific network capabilities and characteristics for a particu- slices.
lar use case can be identified as network slicing [18]. 3GPP Multiple types of User Equipment (UE) can be con-
defines network slicing as a “technology that enables the nected through air interface or fixed-line interface. Thus,
operator to create networks, customised to provide optimised RAN is sliced according to different tenant requirements
solutions for different market scenarios which demand diverse of each application. Radio resource management, slice spe-
requirements (e.g., in terms of functionality, performance and cific admission control policies, the configuration rules for
isolation)” [37]. Control Plane (CP) and User Plane (UP) functions and
According to [19], network slicing is built upon seven main UE awareness on the RAN configuration for the different
principles: isolation, elasticity, automation, programmability, services are the key design aspects for RAN slicing [39].
customisation, E2E and hierarchical abstraction. Separation of RAN resources can be achieved through
frequency, time, code, hardware equipment, software and other
dimensions [12].
B. Network Slicing Architecture RAN subslice and core subslice will be connected through
As stated by NGMN, network slicing architecture consists slice-paring functions. Paring among RAN/core slices can be
of three layers: infrastructure layer, network slice instance 1:1 or 1:N, e.g., a RAN slice could be connected with multiple
layer and service instance layer [38]. core slices [33].
Fig. 5. Network Slicing Life Cycle.
Core network-related functionalities related to a particular edge computing and computation offloading, the gener-
application will be provided via a core slice. Core network ated traffic flow can be made non-uniform. This scenario
resources can be sliced over servers, virtual machines or con- can be identified as horizontal network slicing. It removes
tainers or hardware elements. Logical separation between CP the need of high-resource requirements like computation,
and UP functions and the corresponding implemented VNFs communication and storage in a device itself. Complex
should be considered while designing CN slices. All the opera- tasks that required high computation power can be handed
tional and management tasks related to NSIs will be facilitated over to the next higher layer. As an example, AR and VR
via the NSM. devices need a very high computation capability, which
NSM plays a key role in network slicing architecture and cannot be provided through existing hardware that can be
provides ample services. Composition and deployment of new realised through edge computing technologies.
Network Slice Templates (NSTs) is a significant service pro- Vertical and horizontal network slicing is possible with
vided by NSM. NST is the blueprint of the NSIs and it is static and dynamic network slicing.
fabricated according to the particular requirements of the ten- • Static network slicing: In here, network slices will be
ants using the network capabilities in each technical domain. pre-instantiated and devices need to select the slice which
NSI-related activities such as status monitoring, scale in/out it is going to have the connection.
and destruction are some other functionalities provided by the • Dynamic network slicing: Operators are able to dynam-
NSM. ically design, deploy, customise and optimise the slices
A network slice consists of network functions and those according to the service requirements or conditions in the
network functions can be categorised as slice-specific network network [41]. It encourages the emergence of the novel
functions and slice-independent network functions. Since there concept known as Network Slice as a Service(NSaaS).
is no consensus on selecting slice-independent network func-
tions, this categorisation is considered as an unresolved design
issue in network slicing realisation in 5G networks [40]. Due
to the relative ease of implementing core NFs as software D. Network Slice Life Cycle
applications, core slicing is a less complex task than RAN NSI life cycle can be divided into four phases: preparation,
slicing. commissioning, operation and decommissioning (Figure 5).
• Preparation phase Life cycle starts from the preparation
C. Different Types of Network Slices phase and NSI does not exist in this phase. Tasks in this
Network slices can be categorised using several models: phase are as follows.
vertical and horizontal, static and dynamic and RAN and core. – NST creation and verification.
RAN and core slicing have been discussed earlier. – Preparation of necessary network environment which
• Vertical network slicing: This can be defined as parti- is used to facilitate the NSI life cycle.
tioning the network according to the use cases, whilst – Capacity planning of the network slice.
simplifying the traditional QoS problem. Here, each – Onboarding NSTs.
network domain will be sliced and then those slices will – Evaluation of the network slice requirements.
be paired using the slice paring functions to create the – Preparation of any other requirements in the network.
complete slice. With the advent of 5G, several new appli- • Commissioning phase: In this phase, NSI will be created
cations, such as smart wearables, Augmented Reality from the NST. Tasks in here:
(AR)/ Virtual Reality (VR) and UHD video, became – Allocating the required network resources.
increasingly popular among the community and for each – Performing the necessary configurations to facilitate
application, a vertical network slice can be allocated. Few the slice requirements.
vertical network slices can be identified in 4G LTE either: Network slice will be ready for the operation phase after
NB-IoT slice and MTC slice [33]. this phase.
• Horizontal network slicing: A large amount of traffic • Operation phase: This phase consists of multiple sub-
is generated at the very edge of the network. Through tasks related to network slice instance:
– Activation: Activating the NSI is done in here devices operate in sleep and wake cycles. This causes network
through the processes, such as diverting traffic to the resources to be idle or over-utilised from time to time. Network
slice and provisioning databases. slicing facilitates the dynamic allocation of idle network
– Supervision: NSI will be continuously supervised. resources of a particular network slice to another slice that
– Key Performance Indicator (KPI) monitoring : NSI demands high resource requirements. This process improves
will be continuously monitored. the scalability of the IoT network and finally, it leads to
– Modification: Upgrading NSI, reconfiguration, enhance the network performance, increase the efficiency of
changes in NSI topology, association and dis- resource utilisation and reduce the infrastructure cost [38].
association of network functions with NSI, NSI 3) Existing Slicing Based Solutions: According to the
scaling and altering NSI capacity will be handled proposed existing solutions, network slicing is identified as
under here. a solution to improve the scalability of the 5G based IoT
– Deactivation: Taking the NSI out of the active duty networks. In [45], they concluded that the network slic-
is done in here. ing has the potential to address diverse requirements of 5G
• Decommissioning phase: IoT networks that finally lead to improve the scalability of
– Reclamation of dedicated resources. future IoT applications. Their solution is more conceptual
– Reclamation of configurations from the and implementations have not been provided. The presented
shared/dependent resources. dynamic network slicing framework to improve the scala-
After this phase, the NSI does not exist anymore. bility of fog networks in [46] needs to extend to 5G core
network. Furthermore, AI/ML-based prediction algorithms can
III. T HE ROLE OF N ETWORK S LICING IN I OT be developed to enhance resource allocation that supports
R EALIZATION improving scalability of the IoT networks.
This section focuses on the technical aspects that can
be improved for the IoT realisation via network slicing. B. Improving Dynamicity
Improving scalability, dynamicity, security, privacy, QoS, and Dynamicity describes the continuously changing nature of
E2E orchestration, through network slicing, are discussed here. something. Introduction of IoT and IoT application will change
the fixed nature of the traditional telecommunication networks.
A. Improving Scalability Dynamicity should be a key requirement for 5G and beyond
Scalability, often a sign of stability and competitiveness, network to facilitate heterogeneous IoT ecosystem and also to
is an attribute that is used to describe the ability of a pro- optimise resource utilisation.
cess, software, organisation, or network, to grow and manage 1) Description of the Limitation: IoT requires to accom-
increased demand [42]. A proper scalable network should be plish dynamicity of its network facilities in two main
able to handle the influx of traffic while utilising a limited scenarios: dynamic network resource requirements of IoT
number of resources. Network slicing has been identified by applications, and rapid temporary network deployments for
3GPP as a technology to improve the flexibility and scalability emergency IoT applications. The quantity of resource is
in networking systems including IoT networks [43]. Software- approximately remains constant for a long period in most
based network functions in network slices allow deploying telecommunication networks. Such limited and constant
network functions according to the network traffic to achieve resources have to be optimally utilised among heteroge-
scalability. neous networking applications including IoT applications. The
1) Description of the Limitation: The number of con- amount of resource utilization dynamically varies according to
nected IoT devices will be exponentially increasing in mod- the traffic demand which leads to resources been under-utilised
ern telecommunication networks with heterogeneous network or over-utilised time to time. By nature, IoT applications
requirements. According to the statistics, the number of IoT are extremely dynamic and their traffic patters are fluctuat-
devices in 2020 which is 31 billion devices, will increase to ing constantly. Not only allocation of dedicated portions of
75 billion IoT devices by 2025, and 127 new IoT connections the physical network for different IoT applications, but also
will be established at every second [44]. This growth of IoT dynamically change of the resource quantity of those portions
devices will be under myriads of new IoT based applications according to traffic flow, are a vital requirement to operate an
such as smart healthcare, smart grid, autonomous vehicles, efficient IoT services. Moreover, IoT uses in applications such
industrial automation and AR/VR technologies. Frequent con- as emergency situations (floods, earthquakes) and military sce-
nections and disconnections of nodes of various applications narios. Hence, rapid deployment of temporary IoT networks
with the network are possible. It might pave the way to cause is required to facilitate communication requirements of such
performance degradation, as well as the inability to fulfill the IoT applications. The solution should be fast and effortlessly
network requirements of a diverse set of applications. To han- deployable, and cost-effective.
dle these recurring changes in the IoT traffic flow, the network 2) Benefits of Using Network Slicing: Network slicing can
should be scalable. support to eliminate the static nature of the networks to
2) Benefits of Using Network Slicing: The amount of make them dynamic. Dynamic slice allocation for hetero-
the IoT traffic flow through the network is not always the geneous IoT applications and dynamic resource allocation
same. Due to the energy saving communication links, IoT between slices are two ways of improving dynamicity of the
TABLE III
K EY I OT S ECURITY ATTACKS
networks via network slicing. Learning theory schemes such ransomware [59]–[62]. Network-level security solutions are
as deep learning and reinforcement learning, support to predict vital requisites to mitigate these attacks. Security requirements
user demands and change the resource allocation of slices are diverse in various IoT applications. For instance, security
dynamically according to the predictions [29]. requirements in smart healthcare applications are completely
Moreover, network slicing can be used for the realization of different from that of a smart grid scenario. Therefore, dif-
IoT applications in critical situations. Network slicing allows ferent IoT security mechanisms are needed to fulfill these
rapid deployment of dedicated network slices with relatively diverse security requirements [63] of each IoT application in
effortless configurations over the public network. Thus, the a telecommunication network.
slicing based approach is cost-effective and faster than deploy- 2) Benefits of Using Network Slicing: Network slicing
ing whole new networks for each IoT application. In this can be used to implement various security solutions for the
way, network slicing provides the required dynamicity for IoT plethora of IoT applications. Slice isolation facilitates reduc-
applications. ing the impact of the security attacks, as well as protecting
3) Existing Slicing Based Solutions: Improving dynamic- sensitive information collected via IoT devices. Dynamic
ity in the 5G network through network slicing is a favoured deployment of security NFs in the slices allows to tackle the
area among researchers. Two areas of enabling dynamicity in attacks in run-time, without affecting other IoT applications.
network slicing, i.e., dynamic resource allocation [47], [48] The ability to create quarantine slices to isolate devices that
and dynamic slice creations for different applications [49], have suspicious behaviours facilitates to operate those devices
that finally improve dynamicity of IoT networks have been under tighter restrictions, until taking necessary actions. In this
addressed in existing solutions. Moreover, efficient resource way, network slicing can be beneficial in improving security
allocation algorithms needs to be formulated to improve of IoT applications. Table III summarises the possible attacks
dynamicity. in the IoT applications, along with the possible solutions that
can be provided via network slicing to mitigate those.
C. Improving Security
3) Existing Slicing Based Solutions: According to the exist-
Security is a fundamental requirement in IoT systems. ing research work, network slicing is identified as a way of
Billions of resource-constrained devices in heterogeneous IoT improving security of 5G IoT applications. In [50] and [64],
applications are now connected into people’s day-to-day lives. they have proposed methods to mitigate IoT based DDoS
Because they share sensitive information or are involved in attacks through slice isolation and edge computing respec-
safety-critical operations [57], attackers tend to move towards tively. While secure keying scheme for network slicing was
IoT applications. presented in [56], service-oriented authentication framework
1) Description of the Limitation: Most of the IoT devices was presented in [35]. However, new scientific investigations
are resource constraints in design. Thus, it is difficult to imple- have to be conducted in mitigating other types of security
ment security solutions on IoT devices. Device vendors tend to attacks by using network slicing.
use common credentials for IoT devices and most of device
consumers do not care to change these default credentials.
This increases the vulnerability of the IoT devices to dic- D. Improving Privacy
tionary attacks, while being used by consumers [58]. This Along with the expansion of IoT systems worldwide, the
will help adversaries gain control of the IoT devices. Such necessity of schemes for protecting the privacy of exchanged
potential reasons may result in a series of different types of IoT data has increased. Preserving data privacy plays an
attacks in IoT, such as Distributed Denial of Services (DDoS), ever-increasing role in the IoT applications, which collect
Man In The Middle (MITM), Zero-day, IoT botnets and sensitive measurements [65].
TABLE IV
K EY I OT P RIVACY I SSUES
TABLE V
Q O S R EQUIREMENTS OF D IFFERENT I OT A PPLICATIONS
1) Description of the Limitation: The worldwide spreading 3) Existing Slicing Based Solutions: Network slicing can
of IoT devices results in the need for transmitting collected be used to ensure the privacy of IoT applications, but scien-
data to remote data centres that requires to be further pro- tific investigations that are directly addressing this area are
cessed. This has to follow several data privacy standards quite limited. Moreover, some of the existing work can be
and regulations, which may be international or specialised further extended. An IoT user case-specific privacy-preserving
for a particular geographical region, such as General Data communication scheme presented in [67] can be extended to
Protection Regulation (GDPR) [66]. Implementing this diverse general applications. In [35], Ni et al. proposed an efficient and
set of standards and regulations, corresponding to the geo- secure service-oriented authentication framework supporting
graphical specifications is a vital requirement. Privacy require- network slicing 5G-enabled IoT services.
ments are drastically difference in different IoT applications.
For instance, the privacy requirements in a smart healthcare
application are crucial than the requirements in an environmen- E. Improving Quality of Service (QoS)
tal monitoring application. Facilitating such diverse privacy QoS can be defined as a form of traffic control mechanism
requirements of IoT applications over traditional telecommu- that guarantees the ability to run high-priority applications and
nication networks is a burdensome task. traffic under limited network capacity. The measurements of
2) Benefits of Using Network Slicing: Allocation of sepa- concern in QoS are bandwidth (throughput), delay (latency),
rate slices with proper slice isolation mechanisms for each IoT jitter (variance in latency) and error rate. Different combina-
application, along with required privacy-preserving network tions of these properties are essential in increasing the quality
functions, helps to satisfy the diverse privacy requirements of IoT applications.
of IoT applications. Robust authentication mechanisms and 1) Description of the Limitation: The IoT has prolifer-
strong slice isolation techniques deny the access to a partic- ated within massive application areas that need diverse QoS
ular slice from other slices to protect the confidentiality of requirements for their optimal behavior. As an example, QoS
the IoT application data. The ability to changing the VNF requirements in autonomous vehicle applications, such as
structure of the slices dynamically enhances the difficulty of very low latency and ultra-reliability, are completely different
breaking privacy preserving mechanisms. In this way, network from the requirements in environmental monitoring applica-
slicing can improve the privacy of IoT applications. Table [IV] tions. Facilitating all these heterogeneous QoS requirements
describes the privacy issues in different IoT scenarios, along of different IoT applications through the same infrastruc-
with possible solutions that can be provided through network ture entails a revolutionising re-engineering of the network
slicing for ensuring privacy. architecture [68]. Even in congestion situations, QoS control
allows us to fulfill sufficient service quality requirements of G. Better Resource Allocation and Prioritization
the IoT applications with high priority classification. Table V IoT networks uses several types of network resources such
illustrates the requirement of different QoS for different IoT as computing, storage and networking resources. It is required
applications [69]. the allocate these resource efficiently to optimized the resource
2) Benefits of Using Network Slicing: Network slicing utilization.
ensures facilitating QoS requirements of different IoT appli- 1) Description of the Limitation: IoT networks are con-
cations, through allocating dedicated slices for each use case. sists billions of connected devices which support different use
Dynamic resource allocation between slices allows us to cases with heterogeneous network requirements. The alloca-
accomplish QoS requirements in congestion situations [70]. tion of physical network resources for such vast amount of IoT
3) Existing Slicing Based Solutions: Based on the existing devices is a complex task. Specially, the dynamic nature of IoT
related works, QoS requirements of different IoT applications services will leads to constant fluctuation in resource utiliza-
can be facilitated through network slicing. In [68], Yousaf et al. tion. Moreover, the amount of available network resources is
presented an architecture to provide QoS requirements of the always limited in SP networks. Thus, it is essential to man-
slices. In [71], Höyhtyä et al. used network slicing as an age these limited resources efficiently. However, most of the
enabling technology in providing QoS requirements in critical time, SPs struggle to utilise the available resources efficiently.
scenarios over the public networks. Automated slice resource Moreover, some of the mission critical and delay sensitive IoT
allocation frameworks can be utilized to maintain the QoS applications such as autonomous vehicles, robotics and AR
level of highly dynamic IoT applications. applications might need prioritization of allocated resources
over other IoT applications [76].
2) Benefits of Using Network Slicing: Network slicing, that
F. Providing End-to-End Orchestration implies the allocation of resources [28], can be considered as
Whenever a service enabled in a network, the life time the optimal technology to satisfy diverse network requirements
of that service has to be managed for proper operation. The of 5G IoT applications. The primary idea of network slic-
service End-to-End orchestrator is responsible for provision, ing is analogous to the concept of Infrastructure as a Service
management and optimization of resources for that particular (IaaS) in cloud computing, that shares computing, storage and
network service. networking resources among tenants [27]. Moreover, the chal-
lenge of varying nature of amount of required resources for a
1) Description of the Limitation: Rapid expansion of IoT
particular IoT use case can be overcome by dynamic adjust-
within a plethora of applications, increases the number of ten-
ment of allocated resources for each slice via automated slice
ants and different stakeholders who involve in managing the
manager functions [77]. Moreover, network programmability
ecosystem. As an example, in a smart transportation applica-
is a basic element of network slicing that allows us to alter
tion, road authorities, manufacturing companies, maintenance
the network resource utilisation dynamically among IoT appli-
teams, need to cooperate for the well-being of the applica-
cations. Network slicing allows us to prioritize traffic in two
tion. Cooperation these multiple parties is a challenging task.
techniques: user/traffic prioritization via simultaneous manage-
With the augmentation of IoT services, the amount of con-
ment of the priority among different slices, and prioritization
trol channel data related to IoT flown through the network
users belong to same slice [47].
increases. Hence, a proper mechanism is required to mitigate
3) Existing Slicing Based Solutions: Network slicing
the challenge of managing the amount of IoT control data.
is recognised as a superlative technology for allocat-
2) Benefits of Using Network Slicing: Network slicing pro-
ing required network resources to miscellaneous applica-
vides a solution for this matter through reducing the network
tions in current related works. Radio resource allocation
complexity, by dividing the network into small manageable
frameworks [78], [79] and a hierarchical resource allocation
parts. It supports a wide range of customers’ and operators’
framework [80] for network slicing have been presented for
requirements by allowing them to execute required config-
this. Mathematical models for resource allocation between
uration changes at run time to their slices [72]. Increased
slices were developed in [27]. Since most of the proposed
operational complexity due to a large number of created cus-
resource allocation algorithms are specific to single domain,
tomised networks can be resolved by an end-to-end network
E2E resource allocation algorithms can be identified as a
slice orchestrator. The IoT world can be envisioned as the most
possible future research direction in this aspect [27].
suitable exploitation for a self-managed and isolated slice of
network resources [73].
3) Existing Slicing Based Solutions: E2E orchestration of IV. N ETWORK S LICING BASED I OT A PPLICATIONS
divergent applications is identified as a critical requirement in IoT has been used in many application domains. These IoT
modern networks and have been proposed several architectures applications can be categorised into four key areas accord-
facilitating this requirement. Network slicing management and ing to their requirements (Figure 6). This section focuses
orchestration architectures [72], [74] and such an architec- on discussing key use cases/applications of IoT and how
ture with federated slicing [75] were proposed to serve this network slicing can be used to overcome the challenges in
requirement. Slice orchestration over multi-domains is dis- these applications. Table VI describes a concise summary of
cussed as a novel scope in E2E orchestration and it can be possible slicing solutions for each IoT application scenarios
further investigated. with required technical aspects.
Fig. 6. IoT application scenarios.
A. Smart Transportation
With the mass evolution of the transportation system,
IoT has tightly coupled with several areas in trans-
portation, along with vehicle-to-vehicle communication,
vehicle-to-infrastructure communication, autonomous or
semi-autonomous driving and in-car infotainment systems.
Ultrareliability and very low latency are critical com-
munication requirements in Vehicle-to-Everything (V2X)
applications [111]. Four different types of communication
modes of V2X are identified by 3GPP: vehicle-to-vehicle Fig. 7. Network Slicing & Smart Transportation.
(V2V), vehicle-to-infrastructure (V2I), vehicle-to-pedestrian
(V2P) and vehicle-to-network (V2N).
1) The Role of Network Slicing in Smart Transportation: the resource requirements of the V2X use cases. Network
V2X covers multiple use cases, such as V2V, V2P, V2I and slicing can provide guaranteed level of network resources
V2N. Each use case has heterogeneous service and connec- for V2X use cases via isolating V2X resources from other
tivity requirements that cannot be facilitated through a single application specific slices. Dynamic deployment of network
network infrastructure. Network slicing is the optimal solution functions according to the time of the day (peak and off-
to fulfill these requirements in a cost-effective manner. New peak) or type of area (rural or urban) increases the efficiency
players such as road authorities, vehicle manufacturers and of the network resource utilisation. Moreover, strong security
municipalities that provide multiple services will participate in mechanisms can be provided to the communication of dif-
the V2X scenario, rather than the traditional network provider. ferent V2X applications through dedicated security network
It is difficult to support these multiple tenants via infras- slicing.
tructure owned by different operators. A proper set of slice Figure 7 shows a brief overview of the network slicing
templates can be developed to consummate the requirements of utilisation in smart transportation use case. Four different
multiple tenants. To manage the high density of moving vehi- slices (autonomous vehicles, teleoperated vehicles, remote
cles, deployment of network functions in the network should diagnostic and management and vehicle infotainment systems
be changed dynamically. Also, variations of resource utilisa- with diverse network requirements) are depicted here. A
tion in network resources by other use cases should not violate slice for autonomous vehicles can be used to facilitate
TABLE VI
ROLE OF S LICING FOR I OT AND I TS P ERTINENT D EPLOYMENT C HALLENGES
communication requirements of self-driving cars. Another vehicle simultaneously: slice for autonomous driving and other
slice for teleoperated vehicles can be utilised to manage vehi- safety-critical services, slice supporting teleoperated driv-
cles remotely in environments that are either dangerous or ing, slice for vehicular infotainment and slice for vehicle
unfavourable for humans. Communication between car manu- remote diagnostic and management. A reference network-
facturers or diagnostic centres and vehicles can be facilitated slicing architecture for V2X services is proposed in [82], based
via a slice dedicated for this specific purpose. To provide info- on a three-layer model [112]: infrastructure layer, service layer,
tainment services such as Web browsing, HD video streaming business layer and Management and Orchestration (MANO).
and social media access for passengers, a separate slice can In [83], Khan et al. have analysed the performance of network
be allocated. slicing in a vehicular network, using a multi-lane highway sce-
2) Related Works: Facilitating communication require- nario with two logical slices (the infotainment slice and the
ments of smart transportation applications through network autonomous driving slice). They showed that their network
slicing have been comprehensively discussed in existing solu- slicing approach outperforms the direct Road Side Unit (RSU)
tions. method, while attaining high reliability and throughput. Air-
In [81], Campolo et al. stated that heterogeneous require- Ground Integrated VEhicular Network (AGIVEN) architecture
ments of V2X services cannot be mapped into reference is proposed in [113], to provide high capacity with seamless
slices in 5G: eMBB, mMTC, URLCC or into a single coverage. The proposed architecture is divided into multiple
V2X slice. Besides that, they propose a set of slices for slices to support a specific application while guaranteeing QoS
identified V2X use cases that may be consumed by a single requirements.
2) Related Works: Several slicing-based implementations

can be found in Industrial automation domain. In [84]
and [117], Wu et al. demonstrated a practical scenario in con-
ditional monitoring to achieve self-organisation and flexibility,
with optimal utilisation of network resources in IIoT networks
via network slicing. One of the challenges in IIoT networks
that they addressed in their demonstration was the require-
ment of a quickly accessible and self-organising network.
Another challenge they addressed was the necessity for a flexi-
ble network to achieve diverse QoS requirements from various
Fig. 8. Network Slicing & Industrial automation.
services, as well as on-demand optimisation of the network
efficiency. Results from their demonstration show the useful-
ness of network slicing in IIoT systems. In [85], Kalør, et al.
However, improving security and privacy of V2X network discussed how network slicing can be used to overcome the
slices and efficient resource allocation algorithms for ensuring challenges in traditional telecommunication systems, by intro-
network requirements can be further investigated in transporta- ducing programmability and flexibility to IIoT networks. They
tion applications. have used Industry 4.0 use case to show the network slic-
ing usage to achieve a diverse set of requirements. In [86],
Theodorou et al. presented a cross-domain network slicing
B. Industrial Automation/ Industrial IoT (IIoT) solution developed for EU-funded research project VirtuWind,
Fourth industrial revolution (Industry 4.0) [114], the novel for an industrial wind park scenario. A two-level hierarchi-
enhancement in the industrial automation systems, intro- cal structure had been proposed to management and control
duces modern communication and computation technolo- of the network resources. The network slicing process also
gies such as cloud computing and IoT to industrial man- depends on this two-level hierarchical structure. In [118],
ufacturing systems [115]. As a result, a large number of Baddeley et al. proposed initial efforts to create dedicated SDN
IoT devices, machines and applications with heterogeneous control slices in IIoT networks. They have demonstrated the
network requirements will connect with each other through task using IETF 6TiSCH tracks.
the network for the realisation of industry 4.0 [116]. Utilising Moreover, federated slicing and hierarchical slicing can be
connected IoT based machines rather than human labor with considered as novel research directions related to network
remote monitoring and supervision in a production line will slicing in industrial networks.
increase the efficiency of the system while reducing the
cost. Different types of IoT based sensors can be distributed
throughout the factory to monitor the environmental conditions C. Smart Healthcare
and status of the machines for reducing unplanned downtime. Similar to the other application areas, IoT is heavily used in
1) The Role of Network Slicing in Industrial Automation: the healthcare system under various use cases [119]. Recent
Existing traditional networks are not capable of fulfilling the advances in IoT are identified as a potential solution to alle-
diverse communication requirements, like high reliability, low viate the pressures on the healthcare systems, such as medical
latency and high data rates, of Industrial IoT (IIoT) applica- staff shortage and high healthcare costs, while maintaining
tions. Network slicing is a viable option for supporting the quality care to patients [120]. Remote surgery is one of
diverse set of network requirements using the same physical the mission-critical IoT healthcare applications that enables
network infrastructure cost-effectively. Changing the NFs, in patients to get the service of a set of consultants who are
terms of location and structuring of the slices, achieves vari- in different places around the world. Remote monitoring of
ous combinations of the properties, such as security, mobility, elders and patients who needs continuous monitoring, is an IoT
latency and bandwidth. It is possible to allocate a dedicated application that requires communication services with ultra-
slice with guaranteed network resources for a particularly reliability and low latency [121], [122]. As with the prodigious
large-scale factory line to accomplish their communication development in the robotic field, employment of humanoid
requirements. robots near elderly people will support them to live indepen-
Possible use of network slices in a factory environment dently at home. Wearable IoT devices is becoming popular in
are shown in the Figure 8. Factory environment management, multiple remote monitoring scenarios, such as glucose, ECG,
remote operation of the machines and remote diagnostic and blood pressure and heart rate, for the purpose of avoiding
maintenance are the probable scenarios that require separate preventable deaths [123].
slices. Thousands of sensors and actuators in the factory can 1) The Role of Network Slicing in Smart Healthcare:
be connected together via a dedicated slice, for managing the Heterogeneous communication service requirements in the
factory environment effectively. Workers and machines in the diverse set of IoT use cases in the modern healthcare system
factory can be connected together through a separate slice for cannot be accomplished through the traditional telecom-
remote operation. Machine manufacturers or diagnose centres munication network. Higher data capacity and extremely
and the machines can be coupled securely through a dedicated fast response time, received through novel 5G technolo-
slice to optimize the machines maintenance process. gies, support faster and more accurate results needed for
use case for highlighting the necessity of managing network

slices through defining the remote healthcare slice for various
requirements of dynamic healthcare environments. In [124],
Pries et al. demonstrated allocating a network slice for a smart
health use case, that connects a smart wearable device with
a specific traffic pattern to the cloud efficiently. Innovative
eHealth system, powered by 5G network slicing, was proposed
in [89]. In there, they have studied the health data collec-
Fig. 9. Network Slicing & Smart Healthcare. tion and analysis from various IoT medical devices via 5G
networks. Dynamic provisioning of 5G slices according to the
medical devices with diverse requirements was considered in
patients’ remote monitoring and examination in the health- their proposed architecture.
care domain [87]. Since the lion’s share of the healthcare
applications directly deal with the patients’ lives, communi-
cation services used in the healthcare applications such as D. Smart Home and City
remote surgeries should have supreme reliability. While main- The smart home concept is extended up to smart offices,
taining the promising network resource allocation for these smart buildings and finally, to smart cities. Making people’s
critical healthcare applications, disturbances that will gener- lives smarter through automating day-to-day activities, cost
ate through the variation of resource utilisation from other reduction and energy preservation are major advantages of IoT
applications should be minimised. Network slicing is the pro- utilisation in smart home and city applications. Smart home
pitious technology that allows achieving this requirement via applications are ranging from a simple temperature sensing
slice isolation. The ability to allocate resources dynamically system that automates the functionality of the air conditioner,
to healthcare slices alleviates the resource consumption surges to an advanced image processing system that identifies intrud-
generated through the Internet of Medical Things (IoMT) ers, to increase the protection. Despite the advantages of IoT
devices. Security and privacy are primary concerns in the from a smart home perspective, the rate of adoption of IoT
healthcare industry, since it deals with sensitive data of the devices to their homes by users depends on their desire to
people. Additional security and privacy can be provided by buy those devices. Security and convenience have been iden-
deploying more security and privacy related network functions tified as the key factors that influence their decision [125].
in healthcare slices. Slice isolation removes the visibility of Rapid population growth and urbanisation around cities tend to
healthcare traffic flow to other application slices. utilise most of the resources around the earth in a very limited
Figure 9 explains how different network slices can be used space. This dramatic expansion of cities should be addressed
in the smart health applications. For this sector, separate slices sustainability while increasing the quality of life. The IoT has
for remote surgeries, remote health monitoring and smart hos- been made use of to tackle these issues with the smart city
pitals are identified as required. Applications involving remote concept [91].
surgeries that demand network requirements, such as very 1) The Role of Network Slicing in Smart Home and City:
high throughput, ultra-reliability and very low latency, can be Heterogeneity of IoT devices in homes is varied from very
included into a separate slice. Remote health monitoring appli- primitive devices with very limited energy, to very power-
cations and wearable devices can be connected to a dedicated ful devices that need a continuous power supply and from
slice to provide a secure communication. Medical appliances insecure devices, to devices with very high-security mecha-
in a hospital can be networked to a centralised location via a nisms [92]. Since security devices like door locks, surveillance
dedicated slice. cameras, lights, sirens, smoke detectors and garage door
2) Related Works: In terms of providing required secu- openers deal directly with the security of homes, intruders
rity, latency and reliability requirements of healthcare-related who take control of those devices are able to create neg-
IoT applications, network slicing is recognised as a dominant ative effects to the homeowners. Network slicing is a way
solution in existing related works. to eliminate such negative effects to some extent through
In [87], Mavrogiorgou et al. proposed a platform to manage slice isolation and implementing various security functions
healthcare data efficiently with the use of the latest tech- in the smart home slice. Smart city applications such as
niques in network slicing, data acquisition and data operability. waste-management systems, automated-road-lamp systems,
Through the developed platform, they were able to identify connected-traffic-control systems and smart-parking systems
unknown devices and collect their data to assign that into a will be connected together to facilitate centralised manage-
particular slice to inter-operate with other IoMT devices. ment. Since a majority of these IoT devices are very simple
In [88], Celdrán et al. proposed an architecture to manage and use batteries, energy-efficient communication services is
the life cycle of network slicing, while addressing resource a vital requirement with very high security, as it deals directly
orchestration problems, in terms of what, when and how. A with the daily living styles of the people who live in the city.
policy-based system with two types of policies, intra-slice poli- Network slicing is the prominent technology to fulfill the com-
cies and inter-slice policies, is defined to control the network munication requirements of such devices. This can be done via
slices, as well as to change the resource allocation to the slices allocating a separate slice with lightweight network functions
dynamically. They have thoroughly examined the remote care and isolating this traffic stream from other traffic.
Fig. 11. Network Slicing & AR/VR/Gaming.
Fig. 10. Network Slicing & Smart City & Home. contribution to facilitate indispensable services like remote
surgeries, tourism industry, remote meetings, retail, market-
ing and smart education, as well as entertainment activities.
Some of the possible network slices with their network In [96], Alam et al. proposed a monitoring and safety system
capabilities in a smart home and city environment is depicted using AR/VR technologies and various IoT devices. Remote
in Figure 10. Though a large number of slices is realisable monitoring and supervision are possible with the proposed
in a smart home and city environment, five key scenarios solution. The traditional gaming industry has faced a mas-
among them are shown in the figure: eMBB services, home sive transformation along with AR/VR technologies with IoT
management, waste-management system, traffic-management devices. It is viable to connect multiple players around the
system and energy-management system. Since eMBB devices world to the gaming environment and enhance the gaming
are distributed over smart homes and smart cities, they can experience, rather than playing in a isolated environment. A
be connected to a single dedicated slice. Smart devices in a rapid growth of the VR-based gaming market was achieved
home can be connected to a distinct slice to assure secure over the past few years, due to the development of innova-
communication. tive accessories and wearables. The majority of startups who
2) Related Works: In [92], Dzogovic et al. proposed a enter the electronic games industry tend to engage with VR
network slicing-based smart home system, using three differ- technology to develop their games [127].
ent end-to-end slices. Slice dedicated to the home security 1) The Role of Network Slicing in AR, VR, and Gaming:
system, eMBB slice to devices that demand high data rates More latency causes experiencing disorientation and dizziness
and massive IoT network slice to provide low-data rates (cyber-sickness) to users while consuming AR/VR devices.
for low-power devices are proposed in the paper. In [93], Hence, the Motion-To-Photon (MTP) latency is required to
Chaabnia and Meddeb presented a model for smart home, keep less than 20ms. Accordingly, considerable range for
using network slicing. In their scheme, smart home appli- network-side latency must be 5 to 9ms [128]–[130]. Existing
cations are sliced into four different classes according to traditional networks are not capable to facilitate this require-
their usage, bandwidth requirements and traffic type. In [126], ment. In the cases of remote surgeries and virtual meetings,
Boussard et al. presented an end-to-end research solution, security and privacy are critical requirements, since they are
called Future Spaces, based on SDN and NFV, to dynami- directly dealing with the sensitive information of mankind. AR
cally control people’s digital assets. They have used network and VR-related applications need very high data rates. For a
slices to provide secure access to devices. low resolution, 360 degree AR/VR video requires bandwidth
around 25Mbits/s, and according to the quality of the video,
the required bandwidth ramps up significantly [128], [129].
E. AR, VR and Gaming These heterogeneous network requirements can be facilitated
Augmented Reality (AR), which brings digital elements to through network slicing. The standard eMBB service scenario
live view, and Virtual Reality (VR), which completely replaces in the novel 5G architecture is designed to facilitate these
the live view with a digital view, are the novel technologies kinds of applications. Customised network slices can be used
that have the ability to join the physical world and digi- to provide specific network characteristics of the AR/VR appli-
tal world. AR, one of the disruptive technologies developed cations that cannot be facilitated through the standard eMBB
under the expansion of IoT, is used to improve the interaction slice.
between human and computer in a more entertaining manner Figure 11 shows the allocation of different network slices
within smart environments [94]. Mobile phones can be used in the AR/VR/gaming applications. Since AR/VR entered a
as AR visors at the early stage, to evaluate the effectiveness plethora of application scenarios, it is feasible to categorise
of AR in IoT. After this, the transition can be done to multiple those scenarios within three slices, according to the network
alternative options, such as head-mounted displays like Google requirements as depicted in the figure: slice for AR/VR in
Glass, Magic Leap Lightwear and Microsoft Hololons, and entertainment that includes real-time games, slice for AR/VR
other wearables that allow users hands-free interaction with in healthcare and slice for AR/VR in general applications such
IoT services and objects [95]. VR with IoT provides a huge as education, business, and tourism.
2) Related Works: In [131], Esteves et al. proposed a Proof-

of-Concept (PoC) to use the user’s location to the placement
of a network slice. They considered an interactive-gaming
event that has multiple players who access the gaming appli-
cation simultaneously, in implementing their PoC. A dedicated
network slice is used to facilitate the QoS/QoE requirements
of the gaming service provider to users. In [132], SK Telecom
Fig. 12. Network Slicing & Military applications.
and Ericsson reported that their ability to create different
virtual network slices optimised for AR services.
Since AR/VR is relatively a recent application area, network how to implement critical communication services over a com-
slicing utilisation is not much covered in existing scien- mercial 5G mobile network. They also talked about creating
tific investigations. However, AR/VR realisation via network rapidly deployable networks for emergency and tactical oper-
slicing has been proved only by some companies through ations, using techniques such as network slicing and licensed
demonstrations [132]. shared access.
Figure 12 explains how network slicing facilitates network
requirements in the military IoT. Two important cases for allo-
F. Military Applications cating slices have been identified and they are shown in the
figure: slice for a battlefield and slice for general defense com-
IoT integration with technologies, such as wireless sensor
munication. Fast temporary deployments of slices dedicated
networks, embedded systems, M2M communications, cloud
to a battlefield can facilitate communication requirements of
computing and mobile applications, has the potential to
drones, smart weapons and soldiers. A portion of the public
serve mission-critical military applications. IoT deployment
network with high security mechanisms can be permanently
in military applications has primarily focused on battlefield
allocated to general-defense communications.
applications. Millions of sensors deployed on multiple appli-
2) Related Works: In [133], Grønsund et al. discussed the
cations are used to serve the systems of Command, Control,
implementation of military communication services over the
Communications, Computers, Intelligence, Surveillance and
5G network via network slicing. They proposed two network
Reconnaissance (C4ISR). The situational awareness gained
slices for military mobile users: military slice for military
through those sensors helps senior commanders to steer the
services which requires more restricted access and security
battalion and helps warfighters to combat strategically [97].
and commercial slice for traditional services. The methodol-
In [98], Wrona discuss some important use cases for IoT
ogy of providing the basic service requirements of military
deployment in military applications, such as smart equipment,
communications such as isolation, security, high availability,
situational awareness, logistics and medical care. Military IoT
QoS, and performance, via network slicing has been discussed.
(MIoT) supports to realise the concept of “anytime, anyplace
Moreover, they described the basic implementation of the
connectivity for anything, ubiquitous network with ubiquitous
military slice based on the 5G URLCC slice.
computing” in the military domain. It increases the efficiency
of the utilisation of military resources via implementing and
managing military diversification affairs more accurately and G. Smart Energy
dynamically [99]. Use of IoT-aided robotics in military appli- The smart grid can be identified as an electrical system
cations are discussed in [100] within the following activities: that consists of operations such as electricity generation,
detection of hazardous chemicals and biological weapons, transmission, distribution, control and consumption. It uses
autonomous vehicles in battlefields, support on civil opera- two-way flows of electricity and information, along with
tions in war fields, deactivation of nuclear weapons and access computational and communication technologies to achieve an
control of people in restricted areas. automated and distributed energy-delivery network, with prop-
1) The Role of Network Slicing in Military Applications: erties such as clean, safe, secure, efficient, sustainable, reliable
End-to-end secured communication is a crucial requirement and resilient [101]. The overall efficiency of the integrated
in the battlefields for scenarios like communication between power grid can be improved by using the IoT to form an
soldiers and the control centre and transmission of the interactive real-time network connection between the users
information gathered by the IoT devices. Using an exist- and power equipment [102]. All kinds of components such
ing public network infrastructure with large coverage is a as transformers, breakers, meters and capacitors in the grid
cost-effective solution. Without specifying a separate slice in can be made intelligible through IoT. It facilitates the advan-
the physical network with appropriate security functions and tages such as enhancing customer engagement, optimisation of
strong slice isolation mechanisms, security requirements can’t renewable energy and reducing maintenance costs. In [103],
be accomplished. In the applications such as IoT-aided robotics Sarwat et al. divide the smart grid into three main logical enti-
and Unmanned Aerial Vehicles (UAVs), very low latency, ties: Functional Entity (FE), Operational Entity (OE) and IoT
ultra-reliability, and high bandwidth are critical network Entity (IE). The IE entity is responsible for making the smart
requirements. Traditional telecommunication networks are not grid intelligent by constantly sensing accurate measurements
capable to facilitate these requirements along with the required from FE and reporting them to OE. A shift into more renew-
security in military use cases. In [71], Höyhtyä et al. discuss able and distributed energy generation to reduce greenhouse
Return On Investment (ROI) of network slicing in smart grids

have been analysed from the operator’s perspective. In [137],
Dorsch et al. discussed the economic advantages that can be
achieved through SDN and network slicing when transmit-
ting critical measurements and control commands that required
ultra-reliability and low latency in smart grids. They have
proposed a techno-economic evaluation approach that con-
sists of smart-grid-traffic modelling, network dimensioning,
operator modelling and cost modelling. In [136], Kurtz et al.
proposed an SDN and NFV based network slicing solution
Fig. 13. Network Slicing & Smart Energy.
for critical communications in 5G shared infrastructure. They
have evaluated the solution using a testing setup for two real-
world scenarios - smart grid and Intelligent Transportation
gas emissions, generates instability in the power grid. IoT System (ITS) - allocating separate slices for each scenario
helps to remove the instability in the grid through continuous under three cases: performance study, scalability analysis
real-time monitoring. and critical-infrastructure communication. They have achieved
1) The Role of Network Slicing in Smart Energy: Typically, higher scalability and low end-to-end delay via their proposed
a smart grid consists of a massive number of connected solution.
components, distributed across a very large geographical
area [103]. These smart components have diverse network
requirements [45] across multiple application areas, such as H. UAVs and Drones
online monitoring for the power transmission line, smart The popularity of Unmanned Aerial Vehicles (UAVs), also
homes and smart vehicles [134]. Some of the complications in known as drones, increased over the past few years in a wide
the existing network infrastructure include providing a good range of areas related to civilian, military, commercial and
network connectivity, handling the massive amount of data governmental sectors. Drones have the ability to facilitate
generated through those components, and facilitating diverse multiple IoT services from great heights, forming the airborne
network requirements. Allocation of separate slices for smart IoT through sensors, cameras and GPS modules fixed into
grid-based applications in the 5G network can mitigate these drones and emerging telecommunication technologies such
complications cost-effectively. The smart grid applications are as 5G [104]. Surveillance scenarios (e.g., traffic, power line,
more attractive to cyber-attacks [135], since a successful attack agriculture and environment) are one of the key aspects in
on a grid can cause a bad impact to the whole country, through IoT. UAVs can be used to facilitate the requirements of these
creating disturbances in the daily routine of the people, as well applications [138], [139]. In [140], Motlagh et al. presented
as damaging the electric assets. Communication through pub- an integrative IoT platform operational in the sky, using
lic networks increases the vulnerability for attacks and proper UAVs equipped with diverse IoT devices. They have imple-
security mechanisms should be implemented to mitigate these mented UAV-based crowd surveillance applications through
vulnerabilities. Network slicing is a prominent way to imple- the proposed platform. Using drones as aerial base stations
ment security functions for grid systems. Slice isolation is a is a novel approach in wireless networks to enhance coverage,
potential solution for segregating the grid-based traffic over capacity, energy efficiency and reliability. Attributes of UAVs
the public network. such as mobility, adaptive altitude and flexibility support some
Different application areas in the smart energy service sec- potential applications in wireless systems [105]. This helps to
tor, along with allocating separate slices for each application, improve the network coverage in remote areas to facilitate
are shown in Figure 13. Four different feasible slices have remote IoT applications, such as wildlife monitoring and IoT
been identified in the smart-grid use case: firstly, for UAV- devices that have small transmit power. In addition to sup-
based-grid inspection, secondly, for distribution automation, porting IoT applications, UAVs are considered part of the IoT,
thirdly, for smart metering and lastly, for load control. Slices since their functionality depends on a combination of sensors,
have been allocated according to the network requirements in antennas and embedded software [106].
each use case. 1) The Role of Network Slicing and UAVs: There are two
2) Related Works: Allocation of separate slices for iden- main kinds of data that need to be communicated with UAVs,
tified smart grid applications [24], [136] and increasing the one being control data that needs to handle the UAV. The
revenue of MNOs through network slicing [137] have been other is payload data collected from cameras and sensors
covered in current related works. that needs to be sent to another location for further process-
In [24], Zhang et al. thoroughly examined network slicing. These data types have diverse network requirements. It
ing in smart grid applications. They have identified four use needs to facilitate these requirements cost-effectively. High
cases in smart grids, along with different network require- reliability and low latency are key QoS requirements in con-
ments: Advanced Metering Infrastructure (AMI), Distribution trol data. In payload data, a massive amount of data may
Automation, UAV-based grid Inspection and Millisecond-level need to be transmitted. Along with existing technologies, it
precise load control and they proposed separate slices for is not possible to separate UAV-related traffic from other traf-
each use case. The Total Cost of Ownership (TCO) and fic, as well as control traffic from payload traffic. Network
Fig. 15. Network Slicing & Massive IoT.
Fig. 14. Network Slicing & Drones.
slicing in 5G networks is a noteworthy technology to facil- devices. Agriculture and environment-monitoring applications
itate this requirement via slice isolation, while facilitating are an example for such context, where scale becomes the crit-
heterogeneous network requirements in UAV-related traffic. ical factor, rather than speed [142]. Traditional farms that used
Figure 14 shows how network slicing can be used in UAV to be small units distributed over small land areas have now
service areas with diverse network requirements. Three possi- evolved into massive farms that have thousands of animals
ble use cases of drones that required separate slices are shown over large geographic areas, armed with heterogeneous sen-
in the figure: UAVs in emergency situations, UAVs as floating sors and actuators. These highly accurate embedded sensors
base stations and UAVs in commercial activities like delivery enable precision agriculture by measuring the environmental
services and photography. Each drone use case can be fur- context inside farms [143], [144]. Improving productivity and
ther sliced to communicate control signals and payload data maximising yields and profitability, and reducing the envi-
respectively. ronmental footprint, are resulted by the optimal usage of
2) Related Works: Accomplishing communication require- fertilizers, pesticides and efficient irrigation mechanisms [108].
ments of UAVs and drones, in terms of controlling and other The smart farming concept that includes real-time data gather-
operations such as video streaming, through dedicated network ing and processing, along with automating necessary actions
slices, is discussed in current related works. Further investi- over farming procedures, is realised by precision agriculture.
gations should be conducted in the realisation of other UAV In [109], Ruan et al. divided IoT techniques in agriculture into
use cases, such as aerial base stations and military operations, four categories: controlled equipment farming, livestock breed-
through network slicing. ing, aquaculture and aquaponics and open-field planting. They
In [107], Garcia et al. discussed the performance impact of discussed challenges in exploitation of IoT into agriculture.
using network slicing in aerial vehicle communications. They IoT enables environmental monitoring with an assemblage
have allocated separate network slices to control data and pay- of connected sensors to monitor environmental parameters
load data, referred to as control slice and payload slice. The such as temperature, humidity, wind speed, gases and pres-
experiment has been done under three trials. In the first trial, sure. It provides advantages such as maintaining healthy
they have measured Ipref values in control slice and payload growth of crops and ensuring safe working environments in
slice in Down Link (DL). The second trial had the same prin- industries [110].
ciple with Up Link (UP) and DL. In the final trial, they have 1) The Role of Network Slicing and Massive IoT:
checked whether the delay of the control slice is affected by International standard organisations identified the massive IoT
the payload use of the network. They have shown that required as one of the main service areas in the novel 5G architecture,
throughput of the control slice can be maintained while allo- which is referred as mMTTC. Network slicing is a promis-
cating remaining capacity to payload slice during the flight, ing technology to facilitate network requirements in massive
via first and second trials. Using the results in the third trial, IoT applications that have billions of low power devices with
they have shown that the Round Trip Time (RTT) of control a small amount of data. Narrow Band (NB) IoT is a kind
slice is not influenced by the varied payload throughput. They of network slicing technology in Long Term Evolution (LTE)
have concluded that slicing performs effectively in facilitating networks, which runs over the same physical network infras-
network resources for control communication in aerial vehi- tructure. It is designed to facilitate low-cost devices, high
cle communications. Reference [141] demonstrated how 5G, coverage, long device battery life and massive capacity [145].
along with network slicing, can be used in disaster and emer- Power consumption of the IoT devices can be optimised via
gency situations using drones. The demo is executed under two allocating a separate E2E network slice with optimised NFs.
scenarios (delivering supplies and video streaming) in given Normally, in smart farming and environmental monitoring
situations and a dedicated slice is used to provide network applications, IoT devices will be distributed in rural areas that
connectivity to the drone with required QoS requirements. do not have a proper network coverage. These devices connect
They have shown how networks can be deployed in disaster to the network intermittently to transmit collected information,
situations using network slicing. in order to optimise the battery life. Just like NB-IoT, allocat-
ing a separate slice facilitates network requirements of these
I. Massive IoT (Farming, Agriculture and Environment) devices over common infrastructure with cost optimisation.
The concept of Massive IoT lies on transmitting and con- Two application areas, farming and environmental monitor-
suming a small amount of data from a massive number of IoT ing, in massive IoT are shown in Figure 15 and it describes
how network slicing can be used to facilitate network require- correspondingly to mitigate management and orchestration dif-
ments of those applications. Furthermore, dedicated slices with ficulties evolved through the massive number of slices due to
different NFs can be created to other massive IoT applications. the rapid expansion of IoT.
2) Related Works: In [146], Popovski et al. discussed 2) Existing Solutions: In [170], Kukliński and
heterogeneous non-orthogonal sharing (H-NOMA) of RAN Tomaszewski proposed a scalable approach to mitigate
resources in uplink communications of a set of eMBB, mMTC difficulties (i.e., massive increment in the number of network
and URLLC devices, which connects to a common base sta- slices and involvement of third parties to have slices for their
tion in RAN slicing. They have analysed slicing of resources needs) in network slicing management and orchestration.
between eMBB and URLCC and between eMBB and mMTC, They have introduced a concept called DASMO (Distributed
with illustrations of tradeoffs between H-OMA and H-NOMA. and Autonomic Slice Management and Orchestration), which
In both cases, they have shown that non-orthogonal solu- reduces the management delays and management related
tions must be guided by reliability diversity to achieve more traffic, along with enabling the formation of distributed and
efficiency. automated network slicing management solutions. However,
the proposed solution is only conceptual and not implemented.
J. Other Applications In [169], Afolabi et al. proposed a novel E2E Network Slicing
In addition to the discussed IoT application areas, there are Orchestration System (NSOS) and a Dynamic Auto-Scaling
several other applications that can lay hold of network slic- Algorithm (DASA) for it. Their auto-scaling algorithm scales
ing for their realisation. Typically, these applications have less the resources of their proposed multi-domain orchestration
impact from network slicing rather than discussed applications, system by handling the large number of slice creation requests
but have specific advantages. to maintain system-wide stability.
Smart retail is an application that IoT has dominated in Furthermore, efficient algorithms for resource sharing can
many segments, including supply-chain management, smart be investigated to improve the scalability of network slices.
vending machines and digital signage. Smart retail is iden-
tified as the second-largest MEC use case [147]. Technologies B. Recursion
such as AR/VR that require network slicing for its realisa- The ability to create larger functional blocks by aggregating
tion have the ability to influence the customer’s decision, multiple numbers of smaller functional blocks can be defined
finally impacting smart retail [148]. The smart wearable is as recursion [171]. Applying recursion property in network
another application domain that is comprised with several dis- slicing means the creation of new network slices using existing
cussed applications, such as smart healthcare, smart city and slices [172]. In other terms, network slicing recursion can be
AR/VR/gaming. Networking slicing with MEC is recognised defined as methods for network slice segmentation, allowing
as a solution to overcome the critical concerns in wearable a slicing hierarchy with parent-child relationships [173].
communications, such as short battery life and limited com- 1) Requirements: IoT applications can span across multiple
puting capability [149], [150]. Smart supply chain is listed in tenets and network domains. Therefore, creation a completely
the top ten IoT applications in [151]. Tracking goods while new slice is a complex task when comparing with using exist-
they are on the roads and exchanging inventory information, ing slices. Due the dynamicity of IoT use cases, the creation of
are some possible use cases in the smart supply chain that can new slices is a frequent operation which has to accommodate
be facilitated through mMTC slice over 5G networks. in many networks. Most of the time, there is no be significant
Table VII summarises the potential network slices and differences between the use cases and hence the characteristics
the technical requirements of each proposed slice in each of the requested slices. Therefore, it is beneficial for most of
discussed IoT application. the IoT applications to create new slices by inheriting prop-
erties from parent slices. This will optimise the slice creation
V. T ECHNICAL C HALLENGES R ELATED TO N ETWORK operation in IoT context.
S LICING IN I OT R EALIZATION 2) Existing Solutions: In [174], Héno et al. presented
This section is allocated to describe the technical challenges a recursive network slicing model. The proposed model is
related to the network slicing, due to the IoT realisation. The based on a parallel analysis between IT virtualisation and
state of the art of each challenge in network slicing will be network virtualisation, using virtualisation theory introduced
discussed in detail here. by Popek and Goldberg. Isolation, recursiveness and indepen-
dence are considered in their model. They showed how to
A. Scalability enable new Business-to-Business-to-Consumer (B2B2C) busi-
ness models through network slicing by re-renting rented
1) Requirements: Along with the exponentially rising num-
resources achieved by recursive building of network slices
ber of different IoT applications, the number of slices in the
layered and nested within each other.
network will also increase. Furthermore, service area of a
network slice that offer critical services with ultra-low latency
requirements will also shrink due to high number of IoT C. Adaptive Service Function Chaining
devices. This might finally increase the required number of The network softwarization concepts, especially NFV
slices, thus increasing the slice orchestration requests [169]. proposed to used virtualized network functions or services than
Thus, the network slice management function should scale hardware based network services such as firewalls, Network
TABLE VII
S UMMARY OF E XPECTED P ERFORMANCE IN THE I OT A PPLICATIONS W ITH P OSSIBLE S LICES
Address Translation (NAT) and intrusion protection. These policy and operational requirements. Linkage of these network
vitalized networks services have to execute in certain order functions to form service can be defined as Service Function
to satisfy a particular business need and facilitate the service Chaining (SFC) [175]. Improved operational efficiency and
automated handling of traffic flows with advance features (i.e., In [180], Khettab et al. proposed and evaluated a novel
more security, lower latency or overall high QoS for connected Security as a Service (SECaaS) architecture that leveraged
services) can be achieved through SFC [176]. A network slice SDN and NFV to secure a network slice. This ensured
can be also identified as a collection of connected VNFs. an optimal resource allocation that efficiently manages the
According the nature of supported applications, it is neces- slice-security strategy. Their proposed architecture consists of
sary to dynamically change the structure of the slices. This four main parts: cloud network, each managed by a Virtual
can be realized via by implementing adaptive service function Infrastructure Manager (VIM), VNF managers to monitor a set
chaining. of VNFs in a slice, an NFV orchestrator responsible for life
1) Requirements: IoT applications are highly dynamic cycle managing of VNFs and an SDN controller to monitor
and the network requirements, such as security, latency and and control traffic flows. Along with deploying and manag-
QoS are always changing. Using dynamic service chaining security VNFs (including Intrusion Detection/Prevention
ing, network operators will be able to create, scale, modify System [IDS/IPS] and Deep Packet Inspection [DPI]), the
and remove network functions of a network slice according capabilities in their proposed architecture include monitor-
to the such changing demand, networks and cloud context ing their performance and predictive auto-scaling based on
information [177]. Recurring changes of network requirements pre-defined policies, to ensure the elasticity.
of IoT applications entails continuous updates of SFCs of In [65], Cunha et al. presented a detailed discussion on
network slices to reduce the network cost. Hence, adaptive security challenges of network slicing at the packet core.
SFC in network slicing is a challenge that has to address due Applying AI techniques in achieving security in network slices
to the evolution of IoT applications. is discussed in [181]. They have discussed various security
2) Existing Solutions: In [178], Li et al. introduced a challenges with possible AI mechanisms, in order to mitigate
novel concept of an elastic service function chain. They those attacks in network slicing. Reference [182] discusses
addressed network slicing with such chaining constraints as implementation of a quarantine slice to isolate devices that
a Software Defined Topology (SDT) (performs slice framing) were subjected to a security attack to perform security-related
problem. They formulated the SDT problem (when VF shar- operations for those devices and to minimise the effect of the
ing across services, node overlapping may appear between attack on the other systems.
logical topologies) with elastic SFC as a combinatorial optimi-
sation problem. Then they devised a fast multi-stage heuristic
algorithm to tackle it. E. Privacy
1) Requirements: Network slicing has its own set of pri-
D. Security vacy issues and integration of IoT will further increase the
1) Requirements: Several security vulnerabilities can be impact of it. Information leakages between slices expose the
identified within the network slicing ecosystem. The prolif- sensitive data to third parties, endangering the privacy. Strong
eration of IoT and network slicing exploitation of different slice isolation techniques with secure inter-slice communica-
IoT applications, intensify the effect of these vulnerabilities. tion via secure channels are a critical requirement in using
Such issues can be categorized in to three different areas[179]. network slicing for communications in order to preserve the
• Life cycle security aspects for security-related in different privacy of the users. Also, slice isolation mechanisms should
phases of the network slice life cycle be implemented in customer-end devices to assure the privacy
• Intra-slice security for security aspects in slice itself and of the data from the customer end intruders. Some of the end
• Inter-slice security for security aspects between slices IoT devices can access more than one slice simultaneously
Proper slice isolation is a critical security requirement in may also be a cause for information leakages between slices.
network slicing. Since end-device can access more than one In [65], Cunha et al. discussed the probable security and
slice at once, if there are no strong isolation mechanisms, privacy breaches in network slicing descriptively. Network
attackers can launch attacks to other slices relatively easily. functions that may be shared between slices are allowing
Due to the use of lightweight and less complex security mech- unauthorised inter-slice communication, challenging the confi-
anisms, IoT devices become attractive for many attackers to dentiality and the integrity of the data communicated through
perform such attacks. slices. IoT tenet might ask to use such third party network
Involvement of third parties increases due to the IoT expan- functions in their slices which might jeopardize the whole
sion. Specially, new set of attackers such as cyber criminals slicing system. Successful impersonation of the NSM allows
and terrorist will also target IoT application when they are used subverting slice isolation, resulting in unauthorised access to
in Critical National Infrastructures (CNIs). In such scenar- all the slices that could be resulted in a breach of system
ios, slice-related configuration and management operations are confidentiality and integrity. The ability to change the manage-
given to third parties via Application Programming Interfaces ment and configuration via APIs for third party tenants allows
(APIs). These APIs are an entry point for intruders to enter them to impersonate the host platform and run as an autho-
and perform unauthorised activities. rised platform. By exposing the network and private services
2) Existing Solutions: Exploitation of IoT in a plethora to an attacker, confidentiality and the integrity of the system
of applications increases the security challenges of network can be endangered. The mentioned privacy issues can become
slicing and several mechanisms have been proposed to mitigate more challenging along with the rapid exploitation of IoT with
these challenges. network slicing.
2) Existing Solutions: Based on the existing solutions, slice In [186], Meneses et al. proposed the SliMANO framework
isolation [183] and securing inter-slice communication [184], for slice management and orchestration with an implemented
[185] have been identified as solutions to address privacy chal- and experimentally evaluated prototype. The proposed frame-
lenges related to network slicing. Furthermore, investigations work consists of three parts: SliMANO core for coordinating
can be conducted in preserving privacy in communications SliMANO functionalities, SliMANO-plugin framework for
when sharing network resources between slices. facilitating the continuous deployment of new plugins and
In [184], Liu et al. proposed two heterogeneous signcryp- respective agents and SliMANO-agent framework for perform-
tion schemes to accomplish mutual communications between ing the actions requested by plugins to external network entities.
the Public Key Infrastructure (PKI) and the CertificateLess In [188], Yousaf et al. shed the light on the concept of MANO-as-
public key Cryptography (CLC) environment, in order a-service (MANOaaS) to customise and distribute the MANO
to ensure secure communication between network slices. instances between different network tenants for improving the
PKI-CLC Heterogeneous Signcryption (PCHS) and CLC- control of network slices. In [187], Abbas et al. presented
PKI Heterogeneous Signcryption (CPHS) are the proposed different aspects of slice management for 5G networks. They
schemes in there. They presented how to secure communica- proposed a framework that exploited Mobile-Central Office
tion between a mobile Internet slice and a vehicle Internet slice Re-Architected as Datacentre (M-CORD) for introducing a
using proposed schemes. In [183], Kotulski et al. presented a slicing mechanism for transport network. It enables slices to
detailed discussion on slice isolation, along with types of iso- be available every time a UE requests them.
lation, isolation parameters and dynamic slice isolation. These
address inter-slice communication to accomplish security per- G. Other Technical Challenges
spectives. In [185], Suárez et al. presented a mathematical Apart from the technical challenges above discussed, a few
model based on the concept of network slice chains, to manage other technical challenges can be found due to the rise of the
the inter-slice communication securely when there are differ- IoT realisation. The impact of these challenges is far away
ent security requirements and attributes between slices. Their from the above-discussed challenges, by reason of lack of large
proposed model is extensible for application in any service scale network slicing implementations.
and it complies with any access control model. Limited nature of the RAN resources such as frequency
F. E2E Management and Orchestration and time than network resources in the core network such as
servers and databases, causes difficulties in RAN slicing. It
1) Requirements: The proliferation of the IoT applications affects negatively for the E2E implementation and the rapid
increases the need for dedicated slices in the network for their deployment of the application specific slices. In mobility man-
optimal behavior, finally causing the rise of the number of agement and handling roaming scenarios, the same serving
slices that have to be managed. Resource allocation between slice needs to exist in different operators [189]. IoT devices
slices, changing configurations, facilitating QoS requirements should be able to cross these boundaries without experienc-
and managing the security and privacy aspects for this massive ing any disturbances in communications. Secure and efficient
number of slices, are complex tasks that need to be han- slice information (slice configuration information and user-
dled efficiently. An abundance of third-party tenants due to specific information) transmission between MNOs is required
IoT collaboration within diverse fields will involve manag- in accomplishing this challenge.
ing and configuring the slices. Managing the involvement of
third parties is a challenging task. Temporary usage of slices VI. P ROJECTS
in emergency situations such as earthquakes, floods and fires
make the need for rapid deployment of slices. In addition to This section is allocated to discuss some important projects
the conventional IoT security attacks that are potential to harm that explicitly work with IoT and network slicing technolo-
telecommunication networks, IoT intensifies the severity of gies. Most of the projects are based on the Europe region
the security threats in the network slicing ecosystem. More and the recent Horizon 2020 (H2020) funding scheme has
robust network slicing management facilities are required to fueled for most ongoing projects. Discussed projects along
overcome these security challenges. Utilization of a centralised with their technical aspects and the focused research areas,
approach for slice orchestration and management will increase are summarized in the Table IX.
the scalability and reliability issues, along with the escalated
network slices. According to these facts, IoT systems origi- A. SliceNet
nate several challenges to the network slice management and SliceNet [191] (1st June 2017 - 31st May 2020) is a project
orchestration. funded by the EU commission under the H2020 program. It
2) Existing Solutions: According to the existing solutions, aims to develop a management/control framework to support
slice management and orchestration can become a complex 5G vertical services built as ‘slices’ of network resources. This
activity, along with the rapid increment of the number of project drives in achieving three main objectives based on
slices. A couple of MANO frameworks [186], [187] and the the softwarization of network elements, one being removing
MANOaaS concept [188] have been identified as solutions that the limitations of current network infrastructure. Another is
address this challenge. Future researches can be directed on addressing the associated challenges in managing, controlling
developing more efficient slice management and orchestration and orchestrating the new services running in 5G infrastructures
frameworks. and thirdly, maximising the potential of 5G infrastructures and
S UMMARY OF R ELATED W ORKS

TABLE VIII
(Continued.) S UMMARY OF R ELATED W ORKS

TABLE VIII
TABLE IX
C ONTRIBUTION OF G LOBAL L EVEL P ROJECTS ON N ETWORK S LICING AND I OT
their services. The SliceNet managed domain consists of five Centre Technologic de Telecomunicacions de Catalunya. In
layers, from bottom to top respectively: physical infrastructure, this project, they will create an advanced research and training
virtual infrastructure, service, control and sliced service. The structure for multi-GHz spectrum communications, MEC-
project is running across three main use cases: 5G smart grid empowered service provisioning and end-to-end network slic-
self-healing use case, 5G e-Health-connected ambulance use ing. The goal of this project is training Early Stage Researchers
case and smart-city use case. (ESRs) to form highly-trained academic researchers and indus-
try professionals in the mentioned areas. This aims to over-
B. AutoAir come the technical gaps towards the journey to 5G with the
AutoAir [192] project,led by Airspan, is a unique, acceler- guidance from industry professionals.
ated program for 5G technology. This is based on small cells
that operate on a ‘Neutral Host’ basis that uses network slic- E. MATILDA
ing for allowing multiple public and private mobile operators Matilda [195] (1st June 2017 – 30th November 2019) was
to use the same infrastructure. This project uses the mobile funded by H2020-EU.2.1.1, aiming to develop a holistic,
infrastructure installed at Millibrook. The launch event was innovative 5G framework over sliced programmable infras-
held on 12th February 2019 with a demonstration of providing tructure. This is for design, development and orchestration
gigabit access to vehicles. phases of the 5G-ready applications and the network services.
One of the main objectives of this project is facilitating ver-
C. 5G!Pagoda
tical industries by enabling the development, deployment and
5G!Pagoda [193] (July 2016 - December 2019) is funded orchestration of network-aware applications via dynamically
by European Commission’s H2020 program. The principal created application-aware network slices in the 5G ecosystem.
goal of the project is to explore relevant standards and align It incorporated technological and business requirements that
views on 5G network infrastructure. It uses dynamic creation rose from various parties, such as industry, service providers,
and management of network slices for various mobile services application users and the research community. The project
using federated Japanese and European 5G testbeds. They consisted of four IoT use cases: 5G emergency infrastructure
have presented a demo on IoT services via 5G network slic- and services orchestration with SLA enforcement, smart-city
ing technology at the IoT week in Aarhus in 2019. The top intelligent-lighting system, remote control and monitoring of
network-slicing-related objective in this project is the devel- automobile electrical systems and an Industry 4.0 smart factory.
opment of scalable 5G architecture with a scalable network
slice management and orchestration framework. The IoT-based F. 5G!Drones
objective of this project is to develop emerging 5G applica- H2020 research and innovation program funded for
tions based on IoT use cases and human communication, that 5G!Drones [196] (June 2019 - May 2022) project and it
require high scalability and customisation of diverse end-user is aiming to trial several UAV use cases distributed over
requirements. eMBB, URLCC and mMTC service areas in 5G architec-
ture. It is also validating 5G KPIs for such challenging
D. SEMANTIC use cases. The objectives of this project are validating 5G
end-to-end Slicing and data-drivEn autoMAtion of Next- KPIs and evaluating and validating the performance of dif-
generation cellular neTworks with mobIle edge Clouds (1st ferent UAV applications. This project covers four use cases:
January 2020 – 31st December 2023) SEMANTIC [194] UAV traffic management, public safety/ saving lives, sit-
is a project funded by H2020-EU.1.3.1 and coordinated by uation awareness and connectivity during crowded events.
Network slicing is featured as the key component in this J. 5G-MONARCH

project, to run UAV services across the network using the 5G MObile Network ARCHitecture (5G-
same 5G infrastructure. It demonstrates that each UAV appli- MONARCH) [200] (1st July 2017 - 30th June 2019) is
cation runs independently without affecting the performance a project that was aiming to build a flexible, adaptable
of other applications. The results obtained through this and programmable architecture for 5G. Network slicing
project will be used by the UAV association to make further utilization is the specific technical goal of this project. The
recommendations. project consists of two phases. In phase 1, they designed
such an architecture at a conceptual level with three enabling
G. MonB5G innovations. In phase 2, the addressed architecture was
MonB5G [197] (November 2019 - November 2022) was brought into practice. The three enabling innovations in this
funded by the European Union’s Horizon 2020 research project are inter-slice control and cross-domain management,
and innovation program. It is aiming a novel zero-touch- experiment-driven optimisation and cloud-enabled protocol
management and orchestration framework for network slicing, stack. The developed architecture deployed in two testbeds:
to increase the scalability massively for 5G LTE and beyond. the smart sea port related to vertical industry use case and
This will leverage distribution of operations, together with the tourist city related to media and entertainment use case.
AI-based mechanisms. Security and energy efficiency are Resilience, security and resource elasticity are the functional
considered key features in their conceptual, hierarchical, fault- innovations of the instantiated slices in use cases respectively.
tolerant, data-driven network management system, in order
to orchestrate a plethora of parallel network slices specific K. 5GCity
for diverse services. This project consists of two PoCs: zero-
5GCity [201] (June 2017 – June 2019) is funded by the
touch network and service management with end-to-end SLAs
H2020 research and innovation program. The vision is design-
and AI-assisted policy-driven security monitoring and enforce-
ing, developing, deploying and demonstrating a distributed
ment. The architecture of this project is about splitting the
cloud and radio platform for 5G neutral hosts. This project
centralised management system into multiple management
was focused on bringing benefits of smart city infrastruc-
subsystems that contain one or more distributed management
ture based on resource sharing and end-to-end virtualisation
elements. It has the purpose of distributing intelligence and
to 5G neutral hosts. The project targeted three different cities
design making across various components.
(Barcelona, Bristol and Lucca) under three use cases: media,
neutral host and unauthorised-waste-dumping prevention. This
H. 5GMOBIX
was to enable the creation of dynamic E2E slices with vir-
5GMobix [198] (November 2018 - October 2021) is funded tualised edge and network resources to lease to third-party
via H2020 program and this project is examining the impli- operators. Under the media use case, they will consider three
cations of 5G and its role in future autonomous driving. They different scenarios: mobile real-time transmission, UHD video
will develop and test the functionalities of automated vehicles distribution and real-time video acquisition and production
using 5G technology in various environments, such as city and in edge. For dumping use case, they would use surveillance
urban. This will be under several conditions like vehicular cameras to monitor and detect unauthorised dumping.
traffic and network coverage and service demand, consider-
ing social and business aspects. Network slicing will be used
L. Inspire-5Gplus
to facilitate Corporative, Connected and Automated Mobility
(CCAM) trial activities in this project. INtelligent Security and PervasIve tRust for 5G and Beyond
(INSPIRE-5Gplus) [202] (November 2019 - November 2022)
I. PRIMO-5G which is a Horizon 2020 project, aims to address the cyber-
security risks of 5G and beyond networks, and vertical appli-
The goal of the Primo-5G [199] (July 2018 - June 2021)
cations, ranging from connected cars to industry 4.0, through
project , funded by H2020 program, is to demonstrate an
an innovative security management concept. The project will
end-to-end 5G system providing immersive video services for
improve security in several dimensions including, vision,
moving objects. The project consists of three objectives, one
use cases, integration and management, architecture, models,
being demonstrating a 5G system to provide video services to
and assets. Ensuring the expected Security SLAs (SSLAs)
moving objects. The second is developing technologies related
and the regulation requirements via a fully automated E2E
to mmWave access and 5G core networks. Finally, the third
smart network and security management framework, is another
objective is AI-assisted communications to support objective
objective of this project.
1, 5G standardisation and spectrum regulation activities. The
main envisaged scenario of this project is the firefighting sce-
nario that is a part of the Public Protection and Disaster Relief M. Hexa-X
(PPDR). Drones with cameras will be used by the crew to Hexa-X [203] (January 2021 - June 2023) is the European
get real-time situation awareness of the fire scene prior to Commission’s 6G flagship initiative project which aims to
when they reach to the location, through AR/VR technologies. develop and define the architecture, technologies and vision
Network slicing will be used to provide network services for for next generation 6G wireless networks. The research areas
drones and other moving objects in this project. of Hexa-X project is focusing on six key topics, i.e., connected
Fig. 16. Future research directions.
intelligence, networks of networks, sustainability, global ser- of them are conceptual. Researches with actual slicing imple-
vice coverage, extreme experience, and trustworthiness. mentations can be considered as a valuable movement in the
realisation of smart transportation. Since security and privacy
VII. L ESSONS L EARNED AND F UTURE D IRECTIONS are extremely important aspects of autonomous vehicles, the
implementation of those aspects using network slicing can
This section focuses on the lessons learned and the future
be identified as a future research direction. Traffic-prediction
research directions with respect to network slicing and IoT
algorithms using AI/ML and reinforcement techniques need to
integration. Network slicing in different IoT applications, and
be formulated to increase the efficiency of resource allocation
technical aspects and challenges related to IoT realisation via
between slices, to tackle the traffic surges.
network slicing will be discussed in here. Figure 16 shows a
summary of the discussed future research directions. B. Industrial Automation
1) Lessons Learned: Network slicing tackles the gener-
A. Smart Transportation ated complexity to the industrial networks. This includes
1) Lessons Learned: We explored how IoT can be engaged managing and controlling the network, according to the intro-
in different smart transportation applications with diverse duced new range of requirements by Industry 4.0. Facilitating
network requirements. Autonomous vehicles play a key role the network requirements through allocating resources effi-
in smart transportation. Several pieces of research have been ciently via network slicing has been examined in the existing
conducted to discuss the realisation of autonomous vehicles, researches. Hierarchical slicing can be used in large scale fac-
including fulfilling network requirements. We identified that tories to reduce the complexity in MNO’s network and taking
network slicing can be used to facilitate diverse network the management of the factory network to the factory author-
requirements of different applications in smart transporta- ity. Utilising network slicing to implement secure industry
tion. Moreover, MEC helps to fulfill latency requirements in networks has to be investigated further.
smart transportation applications. Slice isolation is lucrative 2) Future Directions: Using AR/VR technologies in indus-
in implementing security and privacy requirements of trans- trial automation [204], [205] for operating machines remotely
portation communications. Dynamic allocation of network is a significant requirement to enhance the performance of a
resources with enhanced scalability might be required in slices factory line and reduce the cost. Another requirement is remote
allocated to transportation due to sudden traffic surges induced diagnosing faults and performing necessary maintenance activ-
from irregular patterns of using vehicles. ities in machines. Network slicing is a salient technology
2) Future Directions: Though some researches can be that enables this feature and more researches can be con-
found regarding network slicing and smart transportation, most ducted in the future in this area. In [206], Backman et al.
proposed a novel concept of blockchain slice leasing ledger to Network slicing is recognised as the way of providing the
autonomously and dynamically acquire the slices needed for network requirements, such as very low latency and high
increasing the efficiency in future factories. Future researches security required to AR/VR applications over the common
can be directed in using blockchain technology to the manage- infrastructure. Modern games enforced with smart wearables
ment of 5G network slices that enable several new business and AR/VR tools that connect people around the world need
models for smart factories, as well as IT infrastructure. network slicing to attain its network requirements, such as low
latency and high throughput over the public network cost-
C. Smart Healthcare efficiently. Edge computing [210] is a supported technology
1) Lessons Learned: A myriad of use cases of the for the realisation of the AR/VR.
IoMT [207] intensifies the need of network slicing to facilitate 2) Future Directions: To the best of our knowledge at the
heterogeneous network requirements of each use case. Since time that we are doing the survey, no researches can be found
IoMT deals with the sensitive information of people, secu- in network slicing and AR/VR applications. Future researches
rity and privacy of the exchanged healthcare data should be can be directed in the realisation of AR/VR applications using
preserved during the communication [208]. Technical aspects network slicing including satisfying AR/VR network require-
in network slicing, such as slice isolation and adaptive SFC to ments, such as security and high throughput. Implementing the
change the security-related VNFs dynamically, are spotted as same serving slice within multiple MNOs spread over different
preserving ways to healthcare data. Network slicing and the geographical regions is required in modern games.
5G network support the realisation of remote surgeries [209]
that need ultra-reliable network requirements with very low F. Military Applications
latency and guaranteed bandwidth. 1) Lessons Learned: Military applications need extremely
2) Future Directions: Remote patients’ monitoring frame- secure communication with very high privacy. Network slicing
work collects valuable information from wearables, heart is recognised as the way of enabling secure and privacy-
monitors and infusion pumps to provide better treatment to preserved communication in military applications over the
patients in remote. This is highlighted as a future direction public network. Scalability, dynamicity, end-to-end orches-
of remote healthcare [88]. Healthcare-data transmission and tration, recursion, security, adaptive SFC and privacy are
data-management operations can be identified as very chal- all critical requirements in military networks, that can be
lenging research topics in healthcare [89]. So, the attention facilitated through network slicing.
of future researches can be paved the way to healthcare-data 2) Future Directions: As far as we are aware, there are no
transmission via network slicing. AI/ML techniques can be scientific investigations in this domain. Thus, this area is a
used to clean the collected massive healthcare data in MEC novel domain that researchers can direct their researches on
servers prior transmitting to the cloud for analysis. Facilitating the network slicing utilisation in military applications. Since
network requirements in applying AR/VR technologies in security and privacy are supreme aspects in the defense sector,
remote surgeries and telemedicine via network slicing is an modern methods for enhancing these aspects are needed. Rapid
enthralling research direction. deployment of new slices for covering battlefields and multi-
domain slicing, need to be investigated to facilitate military
D. Smart City and Home applications.
1) Lessons Learned: Network slicing enhances the effi-
ciency of network resource utilisation, to enable the realisation G. Smart Grid
of a smart city concept that has billions of devices connected 1) Lessons Learned: Network slicing is the solution for the
together under various use cases. Recursion is discerned as realisation of the smart grid that needs several isolated slices
a salient aspect that supports the fast deploying of network for services such as smart metering and UAV-based power-
slices for homogeneous smart city use cases. Among the smart line inspection to facilitate diverse network requirements.
home diversified-network requirements, security, performance, Slices need to be distributed over a very large geographical
cost and management shortcomings are addressed via network area [211]. MEC can be considered as a viable solution for
slicing elaborately. delay-sensitive applications in smart grids. Higher scalability
2) Future Directions: More elaborate pilots with more in the network requested by the massive number of connected
devices in multiple home applications are required, despite devices, such as smart meters to the grid, can be facilitated
primitive implementations of smart-home solutions. This is via network slicing.
to recognise the viability of network slicing in smart homes. 2) Future Directions: It is possible to find high-level inves-
Recursive composition of network slices can be researched in tigations related to smart grid applications in network slicing.
order to support smart city use cases. However, scientific investigations specified for a particular
application related to the grid are difficult to find and it will
E. AR/VR/Gaming open a new research direction. Diverse energy sources and
1) Lessons Learned: AR/VR is becoming a realisation consumers can be connected together to build a more effective
through the 5G architecture. It extends its roots into multiple smart grid. More researches can be conducted in combining
service areas including mission-critical scenarios, like health- diverse sectors into the grid. Methods to improve the slice
care, and entertaining applications, like gaming and movies. coverage are required in smart grid applications.
H. UAVs and Drones is recognised as two main areas in achieving dynamicity via
1) Lessons Learned: In addition to the very fast communi- network slicing. We have identified that the number of active
cation requirements for controlling the UAV, high-bandwidth connections and the amount of traffic flowed vary rapidly
requirements exist for applications like transferring the video in the IoT applications. Dynamic resource allocation using
recordings collected by the UAV. Network slicing provides network slicing is a possible solution to increase the effi-
these diverse requirements on top of the same physical cient utilisation of network resources. Rapid deployments of
infrastructure. Slice isolation for providing required resources new slices for situations such as disaster management and
uninterruptedly is a critical requirement in UAVs, to assure the battlefields, intensify the need of dynamicity of the network.
best functionality. 2) Future Directions: AI/ML-based resource-requirement
2) Future Directions: Contribution of Multiple Input prediction algorithms are a vital requirement in enhanc-
Multiple Output (MIMO) systems in UAV communications ing the flexibility of dynamic resource allocation between
can be considered as future research directions. Improving the slices and it will be a sophisticated research direction. For
network performance in order to increase the flying time can efficient dynamic resource allocation, resource-allocation algo-
also be considered. Application-specific optimisation of deploy- rithms should be designed with a very low computational
ment, mobility and operations of drones [212] are an important complexity [27]. Due to the heterogeneity of Radio Access
research topic that can be tackled using network slicing. Technologies (RATs), slicing the RAN is a complex task and
it aggravates dynamic RAN resource allocation.
I. Massive IoT
L. Recursion
1) Lessons Learned: A massive number of connected
devices with low power and low data rates increases the 1) Lessons Learned: Slice creation is a frequent activity
requirement of dedicated network slices with simple network due to various slice requirements with slight changes among
functions over the common network infrastructure, due to sim- them. Recursion supports creating slices using the existing
ple power-preserving communication requirements. Massive slices and it reduces the complexity of creating new slices
amounts of data collected through the sensors in this applica- while fastening the deployment of new ones. We recognised
tion can be processed in the edge to reduce the burden in the that the recursion supports the rapid expansion of the IoT in
core network. novel application areas.
2) Future Directions: Simple authentication mechanisms 2) Future Directions: Since recursion is a novel concept,
for these kinds of devices can be implemented as VNFs there is a lot to investigate under this aspect. To the best of
so that they can be deployed in network slices. They are our knowledge at the time we are conducting the research,
a vital requirement and further investigations can be done we couldn’t find a dedicated scientific investigation regarding
in this area. Scientific investigations related to RAN-slicing implementing recursion in network slicing. Required algo-
techniques for massive IoT scenarios are also an interesting rithms, entities, protocols and interfaces needs to be researched
research direction. Methods to improve the coverage of the in implementing recursion.
slices are required when facilitating environmental monitoring.
M. Adaptive SFC
J. Scalability 1) Lessons Learned: Network slice that is allocated for a
1) Lessons Learned: With the exploitation of the IoT for particular use case can be identified as an SFC itself. It runs
a myriad of applications, an enormous quantity of devices on the principles of SFC [213]. The chain of the NFs will be
with diverse network requirements exist for connecting. The needed to change according to the different situations, such as
number of actively connected devices is varying from time to deploying a security function when there is a security attack
time. This escalates the need for a scalable network. Through and removing the security function after the attack. The adapt-
slicing the network and changing the network-resource allo- ability of the SFC in the network slice supports minimising
cation dynamically (with respect to the requirements like the the cost related to the network in terms of the resources and
amount of traffic through the slice), network slicing increases time, as well as increasing the productivity of the use case.
the scalability of the network. 2) Future Directions: As stated in [29], there are two main
2) Future Directions: Extending the proposed dynamic challenges in SFC: selecting the optimal node and the routing
network-slicing framework for fog computing systems into scheme for the selected nodes. AI/ML-based algorithms can be
the 5G core network is stated as a future research direction developed in selecting optimal nodes and the optimum routing
in [46]. Dynamic resource allocation is an important aspect of scheme for the SFC.
the operation of improving scalability. The use of prediction
algorithms based on the AI/ML techniques to predict the traffic N. Security
pattern is an enthralling direction that can be used to enhance 1) Lessons Learned: IoT exploitation in critical applica-
the scalability of the network. tions like smart grids and military applications and sensitive
applications like smart health, increases the security require-
K. Dynamicity ments. To accomplish the different security levels in the IoT
1) Lessons Learned: Dynamic resource allocation between applications, network slicing can be considered as one of the
slices and the ability to deploy network slices dynamically optimum technologies. Slice isolation in network slicing is
identified as the basic way of providing security between dif- Q. Resource Management
ferent applications. Deployment of different security functions 1) Lessons Learned: IoT increases the resource require-
in the slices is another way to achieve security requirements. In ments in the telecommunication networks that can’t be
addition to the traditional security challenges, a series of new satisfied cost-efficiently using traditional methods. We iden-
security challenges have been introduced due to network slic- tified that the network slicing is the modern solution for
ing. Security in network slicing can be divided into three main addressing this problem. Assigning a dedicated amount of
parts: life cycle security, inter-slice security and intra-slice resources for each slice, along with the ability to dynami-
security [179]. cally reassign or overbook the allocated resources, increases
2) Future Directions: Due to the lack of large-scale imple- the resource utilisation efficiency and cost-efficiency of the
mentations of network slicing and the ongoing 5G security system [214]. Assuring the SLAs of the connected third-party
specifications being prone to change, the analysis of network tenants due to heterogeneous applications, including the IoT,
slicing security is still in an incipient phase [179]. Slice iso- becomes a reality due to network slicing [215].
lation is a critical requirement in achieving security. Several 2) Future Directions: Radio resource management with
possible research directions can be identified, such as level of guaranteeing the diverse QoS requirements is a challenging
isolation and inter-slice communication with slice isolation. As task [28]. Investigating efficient methods for RAN resource
discussed in [179], solutions for life cycle security, inter-slice management for RAN slicing is a possible future research
security and intra-slice security can be investigated. Due to the direction. In [215], Khodapanah et al. highlighted the require-
third-party involvement in management scenarios of network ment of an intelligent slice management function for fulfilling
slicing, a series of unknown security attacks will be generated the SLAs of different network slices in the network. AI/ML-
and those can be examined in future researches. related algorithms for predicting resource requirements is a
novel approach that can be used in resource management in
O. Privacy
network slicing.
1) Lessons Learned: Since IoT has entered to the appli-
cations that deal with sensitive information like healthcare VIII. E MERGING AND F UTURE R ESEARCH D IRECTIONS
applications and military applications, privacy became a FOR N ETWORK S LICING AND I OT I NTEGRATION
momentous requirement. Allocating separate slices for differ-
ent applications with strong slice isolation mechanisms allows In this section, we present some of the novel technolo-
privacy-preservation of the data communicated in different gies and their impact on network slicing and IoT realisation.
slices. The ability to change the privacy-related network func- Blockchain, AI/ML, multi-domain slicing and hierarchical
tions of the network slice dynamically facilitates to corporate slicing will be covered in here.
more privacy-preserving facilities into the IoT data streams.
2) Future Directions: As slice isolation is a critical A. Network Slicing and Blockchain
requirement in preserving privacy, implementing an E2E Blockchain is a novel technology that began to attract the
isolated network slice needs to be investigated. Novel privacy- attention as the basis of cryptocurrencies such as Bitcoin. It
protecting algorithms that can be implemented as VNFs and has the ability to vastly improve the existing technology appli-
deployed in a slice can be addressed. Effectuating concepts cations, as well as to enable the realisation of the new applica-
like Privacy by design and Sensitive Data Protection (SDP) in tions that were never previously practical to be deployed [216].
network slicing to preserve privacy can be studied. Blockchain can be identified as a time-stamped series of
data that are immutable. This means once data is added to
P. End-to-End Orchestration the blockchain, they cannot be deleted or modified. Data is
1) Lessons Learned: Network slicing is an end-to-end stored in the blocks that are connected to each other and
technology that enables the deployment of slices, from the cus- secured, using cryptographic principles. Different kinds of
tomer end to the application backend, through the network. It existing blockchain-based activities can be categorised into
has the ability to change the configurations in slices according three areas: Blockchain 1.0 for currency-related operations,
to the SLAs between the tenants and the network providers. Blockchain 2.0 for contracts such as stocks, bonds and loans
This smooths the joining of new applications into the 5G mar- and Blockchain 3.0 for applications such as health, science and
ket, including diverse IoT applications. We recognised that the literacy [217], [218]. Blockchain has been rapidly utilised to
diverse management requirements during the end-to-end traf- register, authenticate and validate assets and transactions. It has
fic flow of various IoT applications can be accomplished using also been utilised to record data, manage the identification and
the end-to-end orchestration property of network slicing. govern interactions among multiple parties [219]. As well as
2) Future Directions: Though there are proposed E2E in the other application areas, blockchains can be utilised in
network slicing management and orchestration frameworks, the telecommunication applications such as fraud detection,
a sophisticated E2E orchestration and management plane Identity-as-a-Service and data management, 5G enablement
is required. It must have adaptive solutions for man- and the IoT connectivity [220]–[222]. In [219], Chaer et al.
aging resources holistically and efficiently, through mak- identified the main five areas that can be used blockchain
ing the decisions according to the current state of in 5G networks: 5G infrastructure and crowdsourcing, 5G-
the slice, as well as the predictions of future user infrastructure sharing, international roaming, network slicing
demands [21]. and management and authentication of mMTC and uRLLC.
Fig. 17. Role of Blockchain for Network slicing in IoT.
Figure 17 explains how blockchain can be used to achieve serve a plethora of vertical services including the IoT [227].
benefits in network slicing and in different IoT applications. Network slicing needs integration and more intelligent capabil-
Network slice broker is a popular research area of using ities based on ML and big data applications, in order to achieve
blockchain in network slicing and several researches can be functionalities such as self-configuration, self-optimisation and
found in this domain [206], [223]–[225]. In [223], Nour et al. fault management [228]. ML techniques (i.e., supervised learn-
proposed a network slice broker design based on blockchain ing, unsupervised learning and reinforcement learning) have
technology. It supports the slice provider that is a new role been extensively used in improving network security, includ-
included in 5G, to select resources from different resource ing authentication, access control, malware detection and
providers and create E2E slices. They envisioned a subslice- anti-jamming offloading [229]. A comprehensive survey on
deployment-brokering mechanism as a series of small con- resource management in cellular and IoT networks using ML
tracts. In [224], Zanzi et al. proposed a novel network-slicing- techniques, was presented in [230]. Figure 18 shows how
brokering solution that leverages blockchain technology called AI/ML can be used in network slicing and in different IoT
NSBchain. It allows the allocation of network resources from application scenarios.
Infrastructure Providers (InPs) to the Intermediate broker (IB) In [231], Thantharate et al. implemented a deep learning
via smart contracts. It also allows IB to assign and redistribute neural network, to develop a DeepSlice model to manage
their resources to tenants in a secure, automated and scal- network-load efficiency and network availability. It covers
able manner. In [225], Valtanen et al. presented a blockchain three main goals: appropriate slice selection for a device, cor-
network-slice-brokering use case value analysis. They used an rect slice prediction and allocating required resources based on
industrial automation scenario as a use case in their paper that the traffic prediction and adaption of slice assignments, in case
acquired the slices needed autonomously and dynamically. of network failures. In [232], Mei et al. proposed an intelligent
network-slicing architecture for V2X services that leverages
B. Network Slicing and Machine Learning/Artificial the recent advancements of ML technologies. A novel deep
Intelligence reinforcement algorithm is proposed to automate the deploy-
Artificial Intelligence (AI) and Machine Learning (ML) ment of network slices based on the collected historical data
are revolutionary technologies that have become ubiqui- of vehicular networks. A deep reinforcement algorithm was
tous in every field, making machines intelligent in order proposed in [233] for resource mapping in 5G network slic-
to make decisions themselves without any human interven- ing. The proposed RLCO algorithm was able to solve the
tion. Telecommunication operators tend to use ML techniques problems of poor efficiency of existing algorithms in virtual
in multifarious areas such as network automation, customer network mapping, low resource utilisation and poor coordi-
experience, business process automation, infrastructure main- nation between node mapping and link mapping. In [234],
tenance and new digital services [226]. ML can be recognised De Bast et al. proposed a fast-learning Deep Reinforcement
as a necessary part of any 5G network, due to its higher com- Learning (DRL) model that has the ability to optimise the
plexity than previous generation networks and the ability to slice configuration of unplanned Wi-Fi networks dynamically
Fig. 18. Role of AI/ML for network slicing in IoT.
Fig. 19. Multi-domain network slicing.
without expert knowledge. The proposed approach was able high-quality connectivity infrastructure in specific locations
to optimise various Wi-Fi parameters per slice dynamically. such as schools and transport hubs, the context-driven and
location-specific needs for wireless connectivity in different
C. Multi-Domain and Multi-Operator Slicing facilities, different business requirements for deployment of
Network slices that are allocated to different vertical appli- 5G networks, and the growing interest for local 5G networks
cations, may be spread over large geographical areas or to serve restricted set of customers such as in a factory envi-
encompassing areas where coverage can’t be provided by a ronment, are some facts that burgeon the concept of local
single operator, may need to combine resources from differ- 5G operators concept [235]. Therefore, the involvement of
ent operators to provide the coverage [75]. Provisioning of multiple administrative domains in creating and operating the
a hierarchical resource allocation architecture for network

slicing. In there, they use Global Radio Resource Manager
(GRRM) to allocate resources for a slice and Local Radio
Resource Manager (LRRM) that is specific for the slice, to
further divide the allocated amount of resources for the slice
among the UEs. To the best of our knowledge, there are no
significant scientific investigations that were conducted in hier-
archical slicing and hence, this is a rich area for conducting
future researches.
IX. C ONCLUSION
Network slicing is becoming a reality in future telecommu-
Fig. 20. Hierarchical slicing Architecture.
nication networks, due to the support of several technologies,
such as SDN, NFV and cloud computing. It has been iden-
network slices for providing a better service is a significant tified as an inevitable technology in the realisation of the
requirement in novel 5G networks. IoT solutions that are proliferating in several applications
Figure 19 shows the concept of multi-domain network with heterogeneous network requirements. We comprehen-
slicing that depicted a network slice spanned over three sively discussed how network slicing can be used in different
administrative domains. As shown in the figure, a minimum IoT applications, along with the existing researches and future
of three orchestrators are needed to facilitate multi-domain research directions for each application. Technical aspects that
network slicing. One is domain-specific-slice orchestrator for can be improved in the IoT through network slicing were
the management and orchestration of the Network Slice Subnet explored. Although network slicing imparts a lot of advan-
Instances (NSSIs) in each domain. Another is domain-specific- tages in the IoT realisation, several technical challenges in
NFVI orchestrator for managing network resources. Thirdly, network slicing will mature, due to the evolution of the IoT.
multi-domain E2E-slice orchestrator for E2E management We identified those challenges during the survey. Since the
of the NS. The significant challenges in multi-domain slic- lack of large-scale implementations of slicing and the spec-
ing includes identifying the necessary administrative domains ifications related to network slicing that are still prone to
with required resources to deploy the NSSIs, stitching them changes, network slicing is a rich area that has several possi-
to create the federated NSI, and the run-time coordination ble future research directions. In all essence, network slicing
of the management operations across different administra- and IoT are two complementary technologies that, if well har-
tive domains [75]. The open-research challenges in the area nessed, have the potential of enabling the smart world since
of multi-domain network slicing are service-management 5G networks and beyond.
interfaces and service profiling, resource sharing and isolation
and service-based network management [75]. R EFERENCES
[1] P. F. P. Guillemin. (2009). The Industrial Internet of Things Volume G1:
Reference Architecture. [Online]. Available: http://www.internet-of-
D. Hierarchical Slicing things-research.eu/pdf/IoT Cluster Strategic Research Agenda 2009.pdf
Hierarchical slicing is an exotic area in network slicing that [2] M. Hatton, “The global M2M market in 2013,” London, U.K., Machina
Research, White Paper, Jan. 2013.
will generate new business models. Ordinarily, MNO creates [3] Y. Zhou, F. R. Yu, J. Chen, and Y. Kuo, “Cyber-physical-social systems:
network slices according to the requirements received from the A state-of-the-art survey, challenges and opportunities,” IEEE Commun.
third party tenants. Though it satisfies the basic communication Surveys Tuts., vol. 22, no. 1, pp. 389–425, 1st Quart., 2020.
[4] B. Blanco et al., “Technology pillars in the architecture of future 5G
requirements of the tenant, he may need to further slice the mobile networks: NFV, MEC and SDN,” Comput. Stand. Interfaces,
received network slice into more subslices that are specified vol. 54, pp. 216–228, Nov. 2017.
for separate service areas of the particular tenant. As an exam- [5] V. W. Wong, Key Technologies for 5G Wireless Systems. Cambridge,
U.K.: Cambridge Univ. Press, 2017.
ple, a giant-sized smart factory can get a dedicated slice from a [6] W. H. Chin, Z. Fan, and R. Haines, “Emerging technologies
large MNO. But the factory needs to further slice the received and research challenges for 5G wireless networks,” IEEE Wireless
network slice into more slices that are specialised to different Commun., vol. 21, no. 2, pp. 106–112, Apr. 2014.
[7] M. Liyanage, A. Gurtov, and M. Ylianttila, Software Defined Mobile
applications in the factory, such as operating machines, mon- Networks (SDMN): Beyond LTE Network Architecture. Chichester,
itoring the factory environment and communication between U.K.: Wiley, 2015.
employees. This can be recognised as the concept of hierar- [8] M. Jarschel, T. Zinner, T. Hoßfeld, P. Tran-Gia, and W. Kellerer,
“Interfaces, attributes, and use cases: A compass for SDN,” IEEE
chical slicing. Reference [236] highlights a set of key features Commun. Mag., vol. 52, no. 6, pp. 210–217, Jun. 2014.
in hierarchical network slicing concepts, such as multi-tenant [9] Y. Li and M. Chen, “Software-defined network function virtualization:
Virtual Service Networks (VSNs) with embedded slices, hier- A survey,” IEEE Access, vol. 3, pp. 2542–2553, 2015.
[10] P. Ranaweera, A. D. Jurcut, and M. Liyanage, “Realizing multi-access
archical SDN control and NFV integration and distributed slice edge computing feasibility: Security perspective,” in Proc. IEEE Conf.
selection. Stand. Commun. Netw. (CSCN), 2019, pp. 1–7.
Figure 20 shows the concept of hierarchical slicing. [11] P. Porambage, J. Okwuibe, M. Liyanage, M. Ylianttila, and T. Taleb,
“Survey on multi-access edge computing for Internet of Things real-
Allocated slice is sliced again into three slices that are speci- ization,” IEEE Commun. Surveys Tuts., vol. 20, no. 4, pp. 2961–2991,
fied for the services of the tenant. In [80], Sun et al. proposed 4th Quart., 2018.
[12] P. C. Chih and I. Lin, End to End Network Slicing, Wireless World [37] S. Antipolis, “Study on architecture for next generation system, release
Res. Forum, Zurich, Switzerland, 2017. 14,” 3GPP, Sophia Antipolis, France, Rep. 23.799, Dec. 2016.
[13] Nokia, Huawei, and Ericsson, “5G network slicing for vertical indus- [38] M. A. Habibi, B. Han, and H. D. Schotten, “Network slicing in 5G
tries,” Global Mobile Suppliers Assoc., London, U.K., Rep., 2017. mobile communication architecture, profit modeling, and challenges,”
[Online]. Available: https://www.huawei.com/minisite/5g/img/5g- 2017. [Online]. Available: arXiv:1707.00852.
network-slicing-for-vertical-industries-en.pdf [39] P. Porambage and M. Liyanage, “Security in network slicing,” Wiley 5G
[14] “Service requirements for the 5G system: Stage I,” 3GPP, Sophia Ref: The Essential 5G Reference Online. Hoboken, NJ, USA: Wiley,
Antipolis, France, Rep. 22.261, Aug. 2016. 2019, pp. 1–12.
[15] “Feasibility study on new services and markets technology enablers [40] T. Yoo, “Network slicing architecture for 5G network,” in Proc.
network operation,” 3GPP, Sophia Antipolis, France, Rep. 22.864, Int. Conf. Inf. Commun. Technol. Convergence (ICTC), 2016,
Jan. 2016. pp. 1010–1014.
[16] “System architecture for the 5G,” 3GPP, Sophia Antipolis, France, Rep. [41] S. Staff. (Dec. 2017). What is Dynamic Network Slicing? What is
23.501, Dec. 2016. Dynamic Network Slicing? [Online]. Available: https://www.sdxcentral.
[17] “Study on architecture for next generation syste,” 3GPP, Sophia com/5g/definitions/dynamic-network-slicing/
Antipolis, France, Rep. 23.799, Dec. 2015. [42] Techopedia. (Jan. 2017). Scalability. [Online]. Available: https://www.
[18] “Description of network slicing concept,” NGMN Alliance, techopedia.com/definition/9269/scalability
Frankfurt, Germany, Rep., Jan. 2016. [Online]. Available: [43] 3GPP. (2016). Network Sharing; Architecture and Functional
https://www.ngmn.org/wp-content/uploads/Publications/2016/161010_ Descriptio. [Online]. Available: https://portal.3gpp.org/
NGMN_Network_Slicing_framework_v1.0.8.pdf desktopmodules/Specifications/SpecificationDetails.aspx?
[19] I. Afolabi, T. Taleb, K. Samdanis, A. Ksentini, and H. Flinck, “Network specificationId=830
slicing and softwarization: A survey on principles, enabling technolo- [44] (Jan. 2020). The IoT Rundown for 2020: Stats, Risks, and Solutions.
gies, and solutions,” IEEE Commun. Surveys Tuts., vol. 20, no. 3, [Online]. Available: https://securitytoday.com/Articles/2020/01/13/The-
pp. 2429–2453, 3rd Quart., 2018. IoT-Rundown-for-2020.aspx?Page=2
[20] P. Schulz et al., “Latency critical IoT applications in 5G: Perspective on [45] P. Rost et al., “Network slicing to enable scalability and flexibility in
the design of radio interface and network architecture,” IEEE Commun. 5G mobile networks,” IEEE Commun. Mag., vol. 55, no. 5, pp. 72–79,
Mag., vol. 55, no. 2, pp. 70–78, Feb. 2017. 2017.
[21] A. A. Barakabitze, A. Ahmad, R. Mijumbi, and A. Hines, “5G [46] Y. Xiao and M. Krunz, “Dynamic network slicing for scalable fog com-
network slicing using SDN and NFV: A survey of taxonomy, archi- puting systems with energy harvesting,” IEEE J. Sel. Areas Commun.,
tectures and future challenges,” Comput. Netw., vol. 167, Feb. 2020, vol. 36, no. 12, pp. 2640–2654, Dec. 2018.
Art. no. 106984. [47] M. Jiang, M. Condoluci, and T. Mahmoodi, “Network slicing man-
[22] A. Kaloxylos, “A survey and an analysis of network slicing in 5G agement & prioritization in 5G mobile systems,” in Proc. Eur. 22nd
networks,” IEEE Commun. Stand. Mag., vol. 2, no. 1, pp. 60–65, Wireless Conf., 2016, pp. 1–6.
Mar. 2018. [48] Y. L. Lee, J. Loo, T. C. Chuah, and L.-C. Wang, “Dynamic network
[23] Q. Chen and C. Liu, “A survey of network slicing in 5G,” in Proc. Int. slicing for multitenant heterogeneous cloud radio access networks,”
Conf. Comput. Sci. Mech. Autom. (CSMA), 2017, pp. 27–35. IEEE Trans. Wireless Commun., vol. 17, no. 4, pp. 2146–2161,
[24] L. Zhang et al., “A survey on 5G network slicing enabling the smart Apr. 2018.
grid,” in Proc. IEEE 25th Int. Conf. Parallel Distrib. Syst. (ICPADS), [49] A. Mayoral, R. Vilalta, R. Casellas, R. Martinez, and R. Munoz, “Multi-
2019, pp. 911–916. tenant 5G network slicing architecture with dynamic deployment of
[25] J. Ordonez-Lucena, P. Ameigeiras, D. Lopez, J. J. Ramos-Munoz, virtualized tenant management and orchestration (MANO) instances,”
J. Lorca, and J. Folgueira, “Network slicing for 5G with SDN/NFV: in Proc. 42nd Eur. Conf. Opt. Commun., 2016, pp. 1–3.
Concepts, architectures, and challenges,” IEEE Commun. Mag., vol. 55, [50] D. Sattar and A. Matrawy, “Towards secure slicing: Using slice iso-
no. 5, pp. 80–87, May 2017. lation to mitigate DDoS attacks on 5G core network slices,” in Proc.
[26] X. Foukas, G. Patounas, A. Elmokashfi, and M. K. Marina, “Network IEEE Conf. Commun. Netw. Security (CNS), 2019, pp. 82–90.
slicing in 5G: Survey and challenges,” IEEE Commun. Mag., vol. 55, [51] P. Schneider, C. Mannweiler, and S. Kerboeuf, “Providing strong 5G
no. 5, pp. 94–100, May 2017. mobile network slice isolation for highly sensitive third-party services,”
[27] R. Su et al., “Resource allocation for network slicing in 5G telecom- in Proc. IEEE Wireless Commun. Netw. Conf. (WCNC), 2018, pp. 1–6.
munication networks: A survey of principles and models,” IEEE Netw., [52] D. Sattar and A. Matrawy, “Optimal slice allocation in 5G core
vol. 33, no. 6, pp. 172–179, Nov./Dec. 2019. networks,” IEEE Netw. Lett., vol. 1, no. 2, pp. 48–51, Jun. 2019.
[28] M. Richart, J. Baliosian, J. Serrat, and J.-L. Gorricho, “Resource slicing [53] A. Mathew, “Network slicing in 5G and the security concerns,” in Proc.
in virtual wireless networks: A survey,” IEEE Trans. Netw. Service 4th Int. Conf. Comput. Methodol. Commun. (ICCMC), 2020, pp. 75–78.
Manag., vol. 13, no. 3, pp. 462–476, Sep. 2016. [54] T. Haugen, A. A. Hassan, and P. F. Menezes, “Device quarantine in a
[29] L. U. Khan, I. Yaqoob, N. H. Tran, Z. Han, and C. S. Hong, “Network wireless network,” U.S. Patent 9 717 006, Jul. 25, 2017.
slicing: Recent Advances, taxonomy, requirements, and open research [55] V. K. Choyi, A. Abdel-Hamid, Y. Shah, S. Ferdi, and A. Brusilovsky,
challenges,” IEEE Access, vol. 8, pp. 36009–36028, 2020. “Network slice selection, assignment and routing within 5G networks,”
[30] W. Rafique, L. Qi, I. Yaqoob, M. Imran, R. U. Rasool, and W. Dou, in Proc. IEEE Conf. Stand. Commun. Netw. (CSCN), 2016, pp. 1–7.
“Complementing IoT services through software defined networking and [56] P. Porambage, Y. Miche, A. Kalliola, M. Liyanage, and M. Ylianttila,
edge computing: A comprehensive survey,” IEEE Commun. Surveys “Secure keying scheme for network slicing in 5G architecture,” in Proc.
Tuts., vol. 22, no. 3, pp. 1761–1804, 3rd Quart., 2020. IEEE Conf. Stand. Commun. Netw. (CSCN), 2019, pp. 1–6.
[31] X. Zhou, R. Li, T. Chen, and H. Zhang, “Network slicing as a service: [57] M. Nawir, A. Amir, N. Yaakob, and O. B. Lynn, “Internet of Things
Enabling enterprises’ own software-defined cellular networks,” IEEE (IoT): Taxonomy of security attacks,” in Proc. 3rd Int. Conf. Electron.
Commun. Mag., vol. 54, no. 7, pp. 146–153, Jul. 2016. Design (ICED), 2016, pp. 321–326.
[32] V. Sciancalepore, F. Cirillo, and X. Costa-Perez, “Slice as a ser- [58] D. Stiawan, M. Idris, R. F. Malik, S. Nurmaini,
vice (SlaaS) optimal IoT slice resources orchestration,” in Proc. IEEE N. Alsharif, and R. Budiarto, “Investigating brute force
Global Commun. Conf., 2017, pp. 1–7. attack patterns in IoT network,” J. Electr. Comput. Eng.,
[33] Q. Li, G. Wu, A. Papathanassiou, and U. Mukherjee, “An end-to-end vol. 2019, Jan. 2019, Art. no. 4568368. [Online]. Available:
network slicing framework for 5G wireless communication systems,” https://www.hindawi.com/journals/jece/2019/4568368/
2016. [Online]. Available: arXiv:1608.00572. [59] N. JOSHI. (May 2019). 8 Types of Security Threats to IoT. [Online].
[34] V. P. Kafle, Y. Fukushima, P. Martinez-Julia, T. Miyazawa, and Available: https://www.allerin.com/blog/8-types-of-security-threats-to-
H. Harai, “Adaptive virtual network slices for diverse IoT services,” iot
IEEE Commun. Stand. Mag., vol. 2, no. 4, pp. 33–41, Dec. 2018. [60] R. Khan, P. Kumar, D. N. K. Jayakody, and M. Liyanage, “A survey
[35] J. Ni, X. Lin, and X. S. Shen, “Efficient and secure service-oriented on security and privacy of 5G technologies: Potential solutions, recent
authentication supporting network slicing for 5G-enabled IoT,” IEEE advancements and future directions,” IEEE Commun. Surveys Tuts.,
J. Sel. Areas Commun., vol. 36, no. 3, pp. 644–657, Mar. 2018. vol. 22, no. 1, pp. 196–248, 1st Quart., 2020.
[36] C. Kolias, G. Kambourakis, A. Stavrou, and J. Voas, “DDoS in the [61] I. Butun, P. Österberg, and H. Song, “Security of the Internet of Things:
IoT: Mirai and other botnets,” Computer, vol. 50, no. 7, pp. 80–84, Vulnerabilities, attacks, and countermeasures,” IEEE Commun. Surveys
Jul. 2017. Tuts., vol. 22, no. 1, pp. 616–644, 1st Quart., 2020.
[62] S. A. Hamad, Q. Z. Sheng, W. E. Zhang, and S. Nepal, “Realizing an [85] A. E. Kalør, R. Guillaume, J. J. Nielsen, A. Mueller, and P. Popovski,
Internet of Secure Things: A survey on issues and enabling technolo- “Network slicing in industry 4.0 applications: Abstraction methods
gies,” IEEE Commun. Surveys Tuts., vol. 22, no. 2, pp. 1372–1391, and end-to-end analysis,” IEEE Trans. Ind. Informat., vol. 14, no. 12,
2nd Quart., 2020. pp. 5419–5427, Dec. 2018.
[63] 5G Security Recommendations Package# 2: Network Slicing, NGMN [86] V. Theodorou, K. V. Katsaros, A. Roos, E. Sakic, and V. Kulkarni,
Alliance, Frankfurt, Germany, 2016. “Cross-domain network slicing for industrial applications,” in Proc.
[64] K. Bhardwaj, J. C. Miranda, and A. Gavrilovska, “Towards IoT-DDOS Eur. Conf. Netw. Commun. (EuCNC), 2018, pp. 209–213.
prevention using edge computing,” in Proc. {USENIX} Workshop [87] A. Mavrogiorgou, A. Kiourtis, M. Touloupou, E. Kapassa, and
Hot Topics Edge Comput. (HotEdge), 2018. [Online]. Available: D. Kyriazis, “Internet of Medical Things (IoMT): Acquiring and trans-
https://www.usenix.org/system/files/conference/hotedge18/hotedge18- forming data into HL7 FHIR through 5G network slicing,” Emerg. Sci.
papers-bhardwaj.pdf J., vol. 3, no. 2, pp. 64–77, 2019.
[65] V. A. Cunha et al., “Network slicing security: Challenges and direc- [88] A. H. Celdrán, M. G. Pérez, F. J. G. Clemente, F. Ippoliti, and
tions,” Internet Technol. Lett., vol. 2, no. 5, p. e125, 2019. G. M. Pérez, “Dynamic network slicing management of multimedia
[66] General Data Protection Regulation. Accessed: Aug. 4, 2020. [Online]. scenarios for future remote healthcare,” Multimedia Tools Appl.,
Available: https://gdpr-info.eu/ vol. 78, no. 17, pp. 24707–24737, 2019.
[67] Y. Zhang, J. Li, D. Zheng, P. Li, and Y. Tian, “Privacy-preserving com- [89] E. Kapassa et al., “An innovative eHealth system powered by 5G
munication and power injection over vehicle networks and 5G smart network slicing,” in Proc. 6th Int. Conf. Internet Things Syst. Manag.
grid slice,” J. Netw. Comput. Appl., vol. 122, pp. 50–60, Nov. 2018. Security (IOTSMS), 2019, pp. 7–12.
[68] F. Z. Yousaf et al. “Network slicing with flexible mobility and [90] F. K. Santoso and N. C. Vun, “Securing IoT for smart home system,”
QoS/QoE support for 5G Networks,” in Proc. IEEE Int. Conf. Commun. in Proc. Int. Symp. Consum. Electron. (ISCE), 2015, pp. 1–2.
Workshops (ICC Workshops), 2017, pp. 1195–1201. [91] E. Theodoridis, G. Mylonas, and I. Chatzigiannakis, “Developing an
[69] (Jul. 2020). QoS Class Identifier. [Online]. Available: https://en. IoT smart city framework,” in Proc. IISA, 2013, pp. 1–6.
wikipedia.org/wiki/QoS_Class_Identifier [92] B. Dzogovic, B. Santos, J. Noll, V. T. Do, B. Feng, and T. van Do,
[70] D. N. Skoutas, N. Nomikos, D. Vouyioukas, C. Skianis, and “Enabling smart home with 5G network slicing,” in Proc. IEEE 4th
A. Antonopoulos, “Hybrid resource sharing for QoS preservation in Int. Conf. Comput. Commun. Syst. (ICCCS), 2019, pp. 543–548.
virtual wireless networks,” Cloud and Fog Computing in 5G Mobile [93] S. Chaabnia and A. Meddeb, “Slicing aware QoS/QoE in software
Networks: Emerging Advances and Applications, vol. 70. Stevenage, defined smart home network,” in Proc. IEEE/IFIP Netw. Oper. Manag.
U.K.: IET, 2017, p. 303. Symp., 2018, pp. 1–5.
[71] M. Höyhtyä et al., “Critical communications over mobile operators’ [94] B. Pokric et al., “Augmented reality enabled IoT services for environ-
networks: 5G use cases enabled by licensed spectrum sharing, network mental monitoring utilising serious gaming concept” J. Wireless Mobile
slicing and QoS control,” IEEE Access, vol. 6, pp. 73572–73582, 2018. Netw. Ubiquitous Comput. Dependable Appl., vol. 6, no. 1, pp. 37–55,
[72] A. Devlic, A. Hamidian, D. Liang, M. Eriksson, A. Consoli, and 2015.
J. Lundstedt, “NESMO: Network slicing management and orchestra- [95] G. White, C. Cabrera, A. Palade, and S. Clarke, “Augmented reality in
tion framework,” in Proc. IEEE Int. Conf. Commun. Workshops (ICC IoT,” in Proc. Int. Conf. Service-Oriented Comput., 2018, pp. 149–160.
Workshops), 2017, pp. 1202–1208.
[96] M. F. Alam, S. Katsikas, O. Beltramello, and S. Hadjiefthymiades,
[73] C. Pereira, A. Pinto, D. Ferreira, and A. Aguiar, “Experimental char-
“Augmented and virtual reality based monitoring and safety system: A
acterization of mobile IoT application latency,” IEEE Internet Things
prototype IoT platform,” J. Netw. Comput. Appl., vol. 89, pp. 109–119,
J., vol. 4, no. 4, pp. 1082–1094, Aug. 2017.
Jul. 2017.
[74] S. Redana, A. Kaloxylos, A. Galis, P. Rost, and V. Jungnickel,
[97] D. E. Zheng and W. A. Carter, Leveraging the Internet of Things for
“View on 5G architecture,” Germany, 5G-PPP, White Paper, 2016.
a More Efficient and Effective Military. Lanham, MD, USA: Rowman
[Online]. Available: https://5g-ppp.eu/wp-content/uploads/2014/02/5G-
& Littlefield, 2015.
PPP-View-on-5G-Architecture-For-public-consultation.pdf
[75] T. Taleb, I. Afolabi, K. Samdanis, and F. Z. Yousaf, “On multi-domain [98] K. Wrona, “Securing the Internet of Things a military perspective,”
network slicing orchestration architecture and federated resource con- in Proc. IEEE 2nd World Forum Internet Things (WF-IoT), 2015,
trol,” IEEE Netw., vol. 33, no. 5, pp. 242–252, Sep./Oct. 2019. pp. 502–507.
[76] Y. Siriwardhana, P. Porambage, M. Liyanage, and M. Ylinattila, [99] L. Yushi, J. Fei, and Y. Hui, “Study on application modes of military
“A survey on mobile augmented reality with 5G mobile edge Internet of Things (MIOT),” in Proc. IEEE Int. Conf. Comput. Sci.
computing: Architectures, applications and technical aspects,” Autom. Eng. (CSAE), vol. 3, 2012, pp. 630–634.
IEEE Commun. Surveys Tuts., early access, Feb. 25, 2021, [100] L. A. Grieco et al., “IoT-aided robotics applications: Technological
doi: 10.1109/COMST.2021.3061981. implications, target domains and open issues,” Comput. Commun.,
[77] H. Wang, Y. Wu, G. Min, J. Xu, and P. Tang, “Data-driven dynamic vol. 54, pp. 32–47, Dec. 2014.
resource scheduling for network slicing: A deep reinforcement learning [101] X. Fang, S. Misra, G. Xue, and D. Yang, “Smart grid—The new and
approach,” Inf. Sci., vol. 498, pp. 106–116, Sep. 2019. improved power grid: A survey,” IEEE Commun. Surveys Tuts., vol. 14,
[78] M. Leconte, G. S. Paschos, P. Mertikopoulos, and U. C. Kozat, “A no. 4, pp. 944–980, 4th Quart., 2012.
resource allocation framework for network slicing,” in Proc. IEEE [102] M. Yun and B. Yuxin, “Research on the architecture and key technology
IEEE Conf. Comput. Commun., 2018, pp. 2177–2185. of Internet of Things (IoT) applied on smart grid,” in Proc. Int. Conf.
[79] H. Zhang, N. Liu, X. Chu, K. Long, A.-H. Aghvami, and Adv. Energy Eng., 2010, pp. 69–72.
V. C. M. Leung, “Network slicing based 5G and future mobile [103] A. I. Sarwat, A. Sundararajan, and I. Parvez, “Trends and future direc-
networks: Mobility, resource management, and challenges,” IEEE tions of research for smart grid IoT sensor networks,” in Proc. Int.
Commun. Mag., vol. 55, no. 8, pp. 138–145, Aug. 2017. Symp. Sens. Netw. Syst. Security, 2017, pp. 45–61.
[80] Y. Sun, M. Peng, S. Mao, and S. Yan, “Hierarchical radio resource [104] T. Lagkas, V. Argyriou, S. Bibi, and P. Sarigiannidis, “UAV IoT frame-
allocation for network slicing in fog radio access networks,” IEEE work views and challenges: Towards protecting drones as “Things”,”
Trans. Veh. Technol., vol. 68, no. 4, pp. 3866–3881, Apr. 2019. Sensors, vol. 18, no. 11, p. 4015, 2018.
[81] C. Campolo, A. Molinaro, A. Iera, and F. Menichella, “5G network [105] M. Mozaffari, W. Saad, M. Bennis, Y.-H. Nam, and M. Debbah,
slicing for vehicle-to-everything services,” IEEE Wireless Commun., “A tutorial on UAVs for wireless networks: Applications, challenges,
vol. 24, no. 6, pp. 38–45, Dec. 2017. and open problems,” IEEE Commun. Surveys Tuts., vol. 21, no. 3,
[82] C. Campolo, A. Molinaro, A. Iera, R. R. Fontes, and C. E. Rothenberg, pp. 2334–2360, 3rd Qurt., 2019.
“Towards 5G network slicing for the V2X ecosystem,” in Proc. [106] C. Luo, J. Nightingale, E. Asemota, and C. Grecos, “A UAV-cloud
4th IEEE Conf. Netw. Softwarization Workshops (NetSoft), 2018, system for disaster sensing applications,” in Proc. IEEE 81st Veh.
pp. 400–405. Technol. Conf. (VTC Spring), 2015, pp. 1–5.
[83] H. Khan, P. Luoto, M. Bennis, and M. Latva-aho, “On the application [107] A. E. Garcia et al., “Performance evaluation of network slicing for
of network slicing for 5G-V2X,” in Proc. 24th Eur. Wireless Conf., aerial vehicle communications,” in Proc. IEEE Int. Conf. Commun.
2018, pp. 1–6. Workshops (ICC Workshops), 2019, pp. 1–6.
[84] H. Wu, I. A. Tsokalo, D. Kuss, H. Salah, L. Pingel, and F. H. Fitzek, [108] A. Kamilaris, F. Gao, F. X. Prenafeta-Boldu, and M. I. Ali, “Agri-IoT:
“Demonstration of network slicing for flexible conditional monitor- A semantic framework for Internet of Things-enabled smart farming
ing in industrial IoT networks,” in Proc. 16th IEEE Annu. Consum. applications,” in Proc. IEEE 3rd World Forum Internet Things (WF-
Commun. Netw. Conf. (CCNC), 2019, pp. 1–2. IoT), 2016, pp. 442–447.
[109] J. Ruan et al., “A life cycle framework of green IoT-based agriculture [131] J. J. A. Esteves, A. Boubendir, F. Guillemin, and P. Sens, “Optimized
and its finance, operation, and management issues,” IEEE Commun. network slicing proof-of-concept with interactive gaming use case,” in
Mag., vol. 57, no. 3, pp. 90–96, Mar. 2019. Proc. 23rd Conf. Innovat. Clouds Internet Netw. Workshops (ICIN),
[110] S. R. Shinde, A. Karode, and D. S. Suralkar, “Review on-IoT based 2020, pp. 150–152.
environment monitoring system,” Int. J. Electron. Commun. Eng. [132] Augmented Reality for Enterprise Alliance. (Oct. 2015). SK Telecom
Technol., vol. 8, no. 2, pp. 103–108, 2017. and Ericsson Demonstrate 5G Network Slices for Augmented
[111] O. Zakaria, J. Britt, and H. Forood, “Internet of Things (IoT) auto- Reality. [Online]. Available: https://thearea.org/ar-news/sk-telecom-
motive device, system, and method,” U.S. Patent 9 717 012, Jul. 25, and-ericsson-demonstrate-5g-network-slices-for-augmented-reality/
2017. [133] P. Grønsund et al., “5G service and slice implementation for a mil-
[112] T. Soenen, R. Banerjee, W. Tavernier, D. Colle, and M. Pickavet, itary use case,” in Proc. IEEE Int. Conf. Commun. Workshops (ICC
“Demystifying network slicing: From theory to practice,” in Workshops), 2020, pp. 1–6.
Proc. IFIP/IEEE Symp. Integr. Netw. Service Manag. (IM), 2017, [134] J. Liu, X. Li, X. Chen, Y. Zhen, and L. Zeng, “Applications of Internet
pp. 1115–1120. of Things on smart grid in China,” in Proc. 13th Int. Conf. Adv.
[113] S. Zhang, W. Quan, J. Li, W. Shi, P. Yang, and X. Shen, “Air-ground Commun. Technol. (ICACT), 2011, pp. 13–17.
integrated vehicular network slicing with content pushing and caching,” [135] C. Bekara, “Security issues and challenges for the IoT-based smart
IEEE J. Sel. Areas Commun., vol. 36, no. 9, pp. 2114–2127, Sep. 2018. grid,” in Proc. FNC/MobiSPC, 2014, pp. 532–537.
[114] Industry 4.0: The Fourth Industrial Revolution—Guide to Industrie [136] F. Kurtz, C. Bektas, N. Dorsch, and C. Wietfeld, “Network slicing for
4.0. Accessed: Jul. 14, 2020. [Online]. Available: https://www:i-scoop: critical communications in shared 5G infrastructures—An empirical
eu/industry-4–0/#:∼:text=Industry%204:0%20is%20the%20digital; evaluation,” in Proc. 4th IEEE Conf. Netw. Softwarization Workshops
of%20the%20industrial%20value%20chain (NetSoft), 2018, pp. 393–399.
[115] T. Qiu, J. Chi, X. Zhou, Z. Ning, M. Atiquzzaman, and D. O. Wu, [137] N. Dorsch, F. Kurtz, and C. Wietfeld, “On the economic benefits
“Edge computing in industrial Internet of Things: Architecture, of software-defined networking and network slicing for smart grid
advances and challenges,” IEEE Commun. Surveys Tuts., vol. 22, no. 4, communications,” NETNOMICS Econ. Res. Electron. Netw., vol. 19,
pp. 2462–2488, 4th Quart., 2020. nos. 1–2, pp. 1–30, 2018.
[116] A. Gurtov, M. Liyanage, and D. Korzun, “Secure communication and [138] S. Petkovic, D. Petkovic, and A. Petkovic, “IoT devices vs. drones
data processing challenges in the industrial Internet,” Balt. J. Modern for data collection in agriculture,” in DAAAM International Scientific
Comput., vol. 4, no. 4, pp. 1058–1073, 2016. Book, vol. 16. Vienna, Austria: DAAAM Int., 2017, pp. 63–80.
[117] H. Wu, G. T. Nguyen, A. K. Chorppath, and F. Fitzek, “Network [139] A. Rajakaruna, A. Manzoor, P. Porambage, M. Liyanage, M. Ylianttila,
slicing for conditional monitoring in the industrial Internet of and A. Gurtov, “Enabling end-to-end secure connectivity for low-power
Things,” Proc. IEEE Softw., Jan. 2018, pp. 1–2. [Online]. Available: IoT devices with UAVs,” in Proc. IEEE Wireless Commun. Netw. Conf.
https://sdn.ieee.org/newsletter/january-2018/network-slicing-for- Workshop (WCNCW), 2019, pp. 1–6.
conditional-monitoring-in-the-industrial-internet-of-things [140] N. H. Motlagh, M. Bagaa, and T. Taleb, “UAV-based IoT platform:
[118] M. Baddeley, R. Nejabati, G. Oikonomou, S. Gormus, A crowd surveillance use case,” IEEE Commun. Mag., vol. 55, no. 2,
M. Sooriyabandara, and D. Simeonidou, “Isolating SDN control pp. 128–134, Feb. 2017.
traffic with layer-2 slicing in 6TiSCH industrial IoT networks,” in [141] S. Wray. (May 2018). Autonomous Drone Services Tested Over
Proc. IEEE Conf. Netw. Funct. Virtualization Softw. Defined Netw. Intercontinental 5G. [Online]. Available: https://5g.co.uk/news/drone-
(NFV-SDN), 2017, pp. 247–251. services-over-intercontinental-5g/4374/
[142] M. Contento. (Mar. 2020). Massive IoT and 5G: What’s Next for
[119] Y. A. Qadri, A. Nauman, Y. B. Zikria, A. V. Vasilakos, and S. W.
Large-Scale Cellular IoT. [Online]. Available: https://www.telit.com/
Kim, “The future of healthcare Internet of Things: A survey of
blog/massive-iot-5g-whats-next/
emerging technologies,” IEEE Commun. Surveys Tuts., vol. 22, no. 2,
[143] P. Chivenge and S. Sharma, “Precision agriculture in food production:
pp. 1121–1167, 2nd Quart., 2020.
Nutrient management,” in Proc. Int. Workshop ICTS Precision Agr.,
[120] L. Catarinucci et al., “An IoT-aware architecture for smart health-
2019, p. 12.
care systems,” IEEE Internet Things J., vol. 2, no. 6, pp. 515–526,
[144] J. P. S. Sundaram, W. Du, and Z. Zhao, “A survey on LoRA networking:
Dec. 2015.
Research problems, current solutions, and open issues,” IEEE Commun.
[121] P. S. Ranaweera, M. Liyanage, and A. D. Jurcut, “Novel MEC based Surveys Tuts., vol. 22, no. 1, pp. 371–388, 1st Quart., 2020.
approaches for smart hospitals to combat COVID-19 pandemic,” IEEE [145] N. Mangalvedhe, R. Ratasuk, and A. Ghosh, “NB-IoT deployment
Consum. Electron. Mag., vol. 10, no. 2, pp. 80–91, Mar. 2021. study for low power wide area cellular IoT,” in Proc. IEEE 27th
[122] Y. Siriwardhana, G. Gür, M. Ylianttila, and M. Liyanage, “The role of Annu. Int. Symp. Pers. Indoor Mobile Radio Commun. (PIMRC), 2016,
5G for digital healthcare against COVID-19 pandemic: Opportunities pp. 1–6.
and challenges,” ICT Exp., to be published. [146] P. Popovski, K. F. Trillingsgaard, O. Simeone, and G. Durisi,
[123] M. S. Hossain and G. Muhammad, “Cloud-assisted Industrial Internet “5G wireless network slicing for eMBB, URLLC, and mMTC:
of Things (IIoT)—Enabled framework for health monitoring,” Comput. A communication-theoretic view,” IEEE Access, vol. 6,
Netw., vol. 101, pp. 192–202, Jun. 2016. pp. 55765–55779, 2018.
[124] R. Pries, H.-J. Morper, N. Galambosi, and M. Jarschel, “Network as a [147] The Business Case for MEC in Retail: A TCO Analysis and
service—A demo on 5G network slicing,” in Proc. 28th Int. Teletraffic Its Implications in the 5G Era. Accessed: Jul. 21, 2020. [Online].
Congr. (ITC), vol. 1, 2016, pp. 209–211. Available: https://builders.intel.com/docs/networkbuilders/the-business-
[125] Y. Chang, X. Dong, and W. Sun, “Influence of characteristics of the case-for-mec-in-retail-a-tco-analysis-and-its-implications-in-the-5g-
Internet of Things on consumer purchase intention,” Social Behav. Pers. era.pdf
Int. J., vol. 42, no. 2, pp. 321–330, 2014. [148] S. G. Dacko, “Enabling smart retail settings via mobile augmented real-
[126] M. Boussard et al., “Future spaces: Reinventing the home network for ity shopping apps,” in Technol. Forecasting Social Change, vol. 124,
better security and automation in the IoT era,” Sensors, vol. 18, no. 9, pp. 243–256, Nov. 2017.
p. 2986, 2018. [149] H. Sun, Z. Zhang, R. Q. Hu, and Y. Qian, “Challenges and enabling
[127] Srikanth. (Jan. 2020). How IoT is Empowering Virtual Reality. technologies in 5G wearable communications,” 2017. [Online].
[Online]. Available: https://www.techiexpert.com/how-iot-is- Available: arXiv:1708.05410v1.
empowering-virtual-reality/ [150] P. Ranaweera, A. D. Jurcut, and M. Liyanage, “Survey on multi-access
[128] “Empowering consumer-focused immersive VR and AR expe- edge computing security and privacy,” IEEE Commun. Surveys Tuts.,
riences with mobile broadband,” Huawei, Shenzhen, China, early access, Feb. 26, 2021, doi: 10.1109/COMST.2021.3062546.
Rep., 2016. [Online]. Available: https://www-file.huawei.com/-/media [151] (2018). The Top 10 IoT Segments in 2018–Based on 1,600 Real
/corporate/pdf/mbb/huawei_whitepaper_vr_ar_final.pdf?la=sw-ke IoT Projects. [Online]. Available: https://iot-analytics.com/top-10-iot-
[129] “VR and AR pushing connectivity limits,” Qualcomm, segments-2018-real-iot-projects/
San Diego, CA, USA, Rep., 2018. [Online]. Available: [152] How Network Latency Affects the Future of Autonomous Vehicles.
https://www.qualcomm.com/media/documents/files/vr-and-ar-pushing- Accessed: Jul. 13, 2020. [Online]. Available: https://www.orcad.com/
connectivity-limits.pdf jp/node/6591
[130] G. F. Networks. (Apr. 2019). Cloud AR/VR Whitepaper. [Online]. [153] Autonomous Cars Will Generate More Than 300 TB of Data Per Year.
Available: https://www.gsma.com/futurenetworks/wiki/cloud-ar-vr- Accessed: Jul. 18, 2020. [Online]. Available: https://www.tuxera.com/
whitepaper/ blog/autonomous-cars-300-tb-of-data-per-year/
[154] (May 2017). Putting Sensors to Work in the Factory Environment: Data [174] M. L. Héno, A. Boubendir, and N. Simoni, “Telco network slicing
to Information to Wisdom. [Online]. Available: https://itpeernetwork. models enabling recursive multi-tenancy,” in Proc. 10th Int. Conf. Netw.
intel.com/putting-sensors-to-work-in-the-factory-environment/#gs. Future (NoF), 2019, pp. 40–47.
9lkvhx [175] P. Q. Carlos Pignataro. (Apr. 2017). Service Function Chaining.
[155] (2018). New Ethernet Applications—Industrial Networking [Online]. Available: https://www.cisco.com/c/dam/m/en_us/network-
Requirements. [Online]. Available: http://www.ieee802.org/3/ad_ intelligence/service-provider/digital-transformation/knowledge-
hoc/ngrates/public/18_03/woods_nea_01_0318.pdf network-webinars/pdfs/0426-techad-ckn.pdf
[156] 5G in Healthcare: Boosting Remote Brain Surgeries, Connected [176] S. Raynovich. (Jan. 2016). What Is Network Service Chaining?
Health, or Medical VR. Accessed: Jul. 25, 2020. [Online]. Available: Definition. [Online]. Available: https://www.sdxcentral.com/
https://medicalfuturist.com/5g-in-healthcare-boosting-telehealth-vr- networking/virtualization/definitions/what-is-network-service-chaining/
connected-health/ [177] S. Zhang, “An overview of network slicing for 5G,” IEEE Wireless
[157] Q. Zhang, J. Liu, and G. Zhao, “Towards 5G enabled tactile robotic Commun., vol. 26, no. 3, pp. 111–117, Jun. 2019.
telesurgery,” 2018. [Online]. Available: arXiv:1803.03586. [178] X. Li, J. Rao, H. Zhang, and A. Callard, “Network slicing with elas-
[158] What Is the Recommended Bandwidth for Different Types of Health tic SFC,” in Proc. IEEE 86th Veh. Technol. Conf. (VTC-Fall), 2017,
Care Providers? Accessed: Jul. 26, 2020. [Online]. Available: pp. 1–5.
https://www.healthit.gov/faq/what-recommended-bandwidth-different- [179] R. F. Olimid and G. Nencioni, “5G network slicing: A security
types-health-care-providers overview,” IEEE Access, vol. 8, pp. 99999–100009, 2020.
[159] Why a Smart Hospital Needs to Look After the Security of [180] Y. Khettab, M. Bagaa, D. L. C. Dutra, T. Taleb, and N. Toumi, “Virtual
Its Network. Accessed: Jul. 26, 2020. [Online]. Available: security as a service for 5G verticals,” in Proc. IEEE Wireless Commun.
https://internationalsecurityjournal.com/why-a-smart-hospital-needs-to- Netw. Conf. (WCNC), 2018, pp. 1–6.
look-after-the-security-of-its-network/ [181] L. Suárez, D. Espes, P. Le Parc, F. Cuppens, P. Bertin,
[160] How Much Internet Speed Does Your Smart Home Need? and C.-T. Phan, “Enhancing network slice security via Artificial
Accessed: Jun. 8, 2020. [Online]. Available: Available: https://www: Intelligence: Challenges and solutions,” in Proc. Conf. C&ESAR,
smarthomeblog:net/bandwidthsmart-home/#:∼:text=Speed%3A% Rennes, France, Nov. 2018. [Online]. Available: https://hal.archives-
20As%20a%20general%20recommendation;you%20can%20adjust% ouvertes.fr/hal-01959251
20from%20there [182] (Jul. 2019). The Evolution of Security in 5G. [Online]. Available:
[161] P. Fedchenkov, A. Zaslavsky, A. Medvedev, T. Anagnostopoulos, https://www.5gamericas.org/wp-content/uploads/2019/08/5G-Security-
I. Sosunova, and O. Sadov, “Supporting data communications in IoT- White-Paper_8.15.pdf
enabled waste management,” in Internet of Things, Smart Spaces, and [183] Z. Kotulski et al., “Towards constructive approach to end-to-end slice
Next Generation Networks and Systems. Cham, Switzerland: Springer, isolation in 5G networks,” EURASIP J. Inf. Security, vol. 2018, no. 1,
2017, pp. 163–174. p. 2, 2018.
[184] J. Liu, L. Zhang, R. Sun, X. Du, and M. Guizani, “Mutual heteroge-
[162] E. Mohyeldin. Minimum Technical Performance Requirements for
neous signcryption schemes for 5G network slicings,” IEEE Access,
IMT-2020 Radio Interface(s). Accessed: Jun. 20, 2020. [Online].
vol. 6, pp. 7854–7863, 2018.
Available: https://www:itu:int/en/ITU-R/study-groups/rsg5/rwp5d/imt-
2020/Documents/S01–1Requirements%20for%20IMT-2020Rev:pdf [185] L. Suárez, D. Espes, F. Cuppens, C.-T. Phan, P. Bertin, and P. Le Parc,
“Managing secure inter-slice communication in 5G network slice
[163] Bandwidth Requirements for Virtual Reality (VR) and Augmented
chains,” in Proc. IFIP Annu. Conf. Data Appl. Security Privacy, 2020,
Reality (AR). Accessed: Aug. 8, 2020. [Online]. Available: https://www.
pp. 24–41.
mushroomnetworks.com/infographics/bandwidth-requirements-for-
[186] F. Meneses, M. Fernandes, D. Corujo, and R. L. Aguiar, “SliMANO:
virtual-reality-vr-and-augmented-reality-ar-infographic/
An expandable framework for the management and orchestration of
[164] H. Mendis, P. E. Heegaard, and K. Kralevska, “5G network slicing end-to-end network slices,” in Proc. IEEE 8th Int. Conf. Cloud Netw.
as an enabler for smart distribution grid operations,” in Proc. CIRED, (CloudNet), 2019, pp. 1–6.
2019, pp. 1–5.
[187] M. T. Abbas, T. A. Khan, A. Mahmood, J. J. D. Rivera, and
[165] 5G Network Slicing Enabling the Smart Grid. Accessed: May 25, 2020. W.-C. Song, “Introducing network slice management inside M-CORD-
[Online]. Available: http://www-file.huawei.com/-/media/ based-5G framework,” in Proc. IEEE/IFIP Netw. Oper. Manag. Symp.,
CORPORATE/PDF/News/5g-network-slicing-enabling-the-smart- 2018, pp. 1–2.
grid.pdf [188] F. Z. Yousaf, V. Sciancalepore, M. Liebsch, and X. Costa-Perez,
[166] Growing Use of UAVs Strains Bandwidth. Accessed: May 28, 2020. “MANOaaS: A multi-tenant NFV MANO for 5G network slices,” IEEE
[Online]. Available: https://spacenews.com/growing-use-uavs-strains- Commun. Mag., vol. 57, no. 5, pp. 103–109, May 2019.
bandwidth/ [189] R. Wen, G. Feng, J. Zhou, and S. Qin, “Mobility management for
[167] Enterprise Application of Autonomous UAVs: QoS & QoE From network slicing based 5G networks,” in Proc. IEEE 18th Int. Conf.
Modern Telecom Networks. Accessed: Jun. 1, 2020. [Online]. Commun. Technol. (ICCT), 2018, pp. 291–296.
Available: https://www.itu.int/en/ITU-T/Workshops-and-Seminars/qos/ [190] M. Richart, J. Baliosian, J. Serrati, J.-L. Gorricho, R. Agüero, and
201809/Documents/R_Roy_Presentation.pdf N. Agoulmine, “Resource allocation for network slicing in WiFi access
[168] What Value Can High-Speed Broadband Create for Farming points,” in Proc. 13th Int. Conf. Netw. Service Manag. (CNSM), 2017,
Communities. Accessed: Jun. 2, 2020. [Online]. Available: https://www. pp. 1–4.
smartfarmnet.com/farming-with-high-speed-broadband.html [191] SliceNet. Accessed: Aug. 15, 2020. [Online]. Available: https://slicenet.
[169] I. Afolabi, J. Prados, M. Bagaa, T. Taleb, and P. Ameigeiras, “Dynamic eu/
resource provisioning of a scalable E2E network slicing orchestration [192] AutoAir. Accessed: Aug. 15, 2020. [Online]. Available: https://www.
system,” IEEE Trans. Mobile Comput., vol. 19, no. 11, pp. 2594–2608, airspan.com/autoair/
Nov. 2019. [193] 5G!Pagoda. Accessed: Aug. 17, 2020. [Online]. Available: https://5g-
[170] S. Kukliński and L. Tomaszewski, “DASMO: A scalable approach to pagoda.aalto.fi/
network slices management and orchestration,” in Proc. IEEE/IFIP [194] SEMANTIC. Accessed: Aug. 16, 2020. [Online]. Available:
Netw. Oper. Manag. Symp., 2018, pp. 1–6. https://cordis.europa.eu/project/id/861165
[171] L. Geng, S. Bryant, K. Makhijani, A. Galis, X. de Foy, and S. Kuklinsk. [195] MATILDA. Accessed: Aug. 20, 2020. [Online]. Available: https://www.
(2017). Network Slicing Architectur. [Online]. Available: https://tools. matilda-5g.eu/index.php
ietf.org/id/draft-geng-netslices-architecture-01.html#rfc.section.3.3.1 [196] 5G!Drones. Accessed: Aug. 19, 2020. [Online]. Available:
[172] C. E. Curtis Collicutt. (2018). 5G Network Slicing and OpenStac. https://5gdrones.eu/
[Online]. Available: https://object-storage-ca-ymq-1.vexxhost.net/ [197] MonB5G. Accessed: Aug. 19, 2020. [Online]. Available: https://www.
swift/v1/6e4619c416ff4bd19e1c087f27a43eea/www-assets-prod/ monb5g.eu/
presentation-media/5G-Network-Slicing-and-OpenStack-Presentation- [198] 5GMOBIX. Accessed: Aug. 21, 2020. [Online]. Available: https://www.
Copy-C03.pdf 5g-mobix.com/
[173] A. Galis. (Nov. 2018). Network Slicing—A Holistic Architectural [199] PRIMO-5. Accessed: Aug. 21, 2020. [Online]. Available: https://primo-
Approach, Orchestration and Management with Applicability in Mobile 5g.eu/
and Fixed Networks and Cloud. [Online]. Available: http://cnsm-conf. [200] 5G-MONARCH. Accessed: Aug. 22, 2020. [Online]. Available:
org/2018/files/CNSM18_SlicingTutorial_AlexGalis_5–10-2018.pdf https://5g-monarch.eu/
[201] 5GCity. Accessed: Aug. 23, 2020. [Online]. Available: https://www. [226] B. Ciszewski. (Oct. 2018). Machine Learning in Telecommunications:
5gcity.eu/ The Sector Bound to Lead the AI Revolution. [Online]. Available:
[202] INSPIRE-5Gplus. Accessed: Aug. 23, 2020. [Online]. Available: https://www.netguru.com/blog/ai-in-telecommunications
https://www.inspire-5gplus.eu/ [227] E. Schmit. (Dec. 2019). The Powerful Combination of Machine
[203] Hexa-X. Accessed: Jan. 25, 2021. [Online]. Available: https://hexa-x.eu/ Learning and 5G Networks. [Online]. Available: https://shape.att.com/
[204] J. Martin and J. Bohuslava, “Augmented reality as an instrument for blog/combination-of-machine-learning-and-5g
teaching industrial automation,” in Proc. Cybern. Informat. (K&I). [228] L.-V. Le, B.-S. P. Lin, L.-P. Tung, and D. Sinh, “SDN/NFV, machine
2018, pp. 1–5. learning, and big data driven network slicing for 5G,” in Proc. IEEE
[205] I. Malỳ, D. Sedláček, and P. Leitao, “Augmented reality experiments 5G World Forum (5GWF), 2018, pp. 20–25.
with industrial robot in industry 4.0 environment,” in Proc. IEEE 14th [229] L. Xiao, X. Wan, X. Lu, Y. Zhang, and D. Wu, “IoT security techniques
Int. Conf. Ind. Informat. (INDIN), 2016, pp. 176–181. based on machine learning: How do IoT devices use AI to enhance
[206] J. Backman, S. Yrjölä, K. Valtanen, and O. Mämmelä, “Blockchain security?” IEEE Signal Process. Mag., vol. 35, no. 5, pp. 41–49,
network slice broker in 5G: Slice leasing in factory of the future use Sep. 2018.
case,” in Proc. Internet Things Bus. Models Users Netw., 2017, pp. 1–8. [230] F. Hussain, S. A. Hassan, R. Hussain, and E. Hossain, “Machine learn-
[207] Why the IoMT Will Start to Transform Healthcare. ing for resource management in cellular and IoT networks: Potentials,
Accessed: Aug. 27, 2020. [Online]. Available: https://www.forbes. current solutions, and open challenges,” IEEE Commun. Surveys Tuts.,
com/sites/bernardmarr/2018/01/25/why-the-internet-of-medical-things- vol. 22, no. 2, pp. 1251–1275, 2nd Quart., 2020.
iomt-will-start-to-transform-healthcare-in-2018 [231] A. Thantharate, R. Paropkari, V. Walunj, and C. Beard, “DeepSlice:
[208] Using Network Slicing to Enhance Security in e-Health Use Cases. A deep learning approach towards an efficient and reliable network
Accessed: Aug. 28, 2020. [Online]. Available: https://protego-project. slicing in 5G networks,” in Proc. IEEE 10th Annu. Ubiquitous Comput.
eu/2019/05/using-network-slicing-to-enhance-security-in-e-health-use- Electron. Mobile Commun. Conf. (UEMCON), 2019, pp. 762–767.
cases/ [232] J. Mei, X. Wang, and K. Zheng, “Intelligent network slicing for
[209] How 5G Mobile Networks are Opening the Door to Remote V2X services toward 5G,” IEEE Netw., vol. 33, no. 6, pp. 196–204,
Surgery. Accessed: Aug. 28, 2020. [Online]. Available: https://www. Nov./Dec. 2019.
medicaldevice-network.com/features/5g-remote-surgery/ [233] L. Zhao and L. Li, “Reinforcement learning for resource mapping in
[210] How 5G and Edge Computing Can Enhance Virtual Reality. 5G network slicing,” in Proc. 5th Int. Conf. Comput. Commun. Syst.
Accessed: Aug. 30, 2020. [Online]. Available: https://www.ericsson. (ICCCS), 2020, pp. 869–873.
com/en/blog/2020/4/how-5g-and-edge-computing-can-enhance-virtual- [234] S. De Bast, R. Torrea-Duran, A. Chiumento, S. Pollin, and H. Gacanin,
reality “Deep reinforcement learning for dynamic network slicing in IEEE
[211] 5G Network Slicing Enabling the Smart Grid. Accessed: Aug. 30, 2020. 802.11 networks,” in Proc. IEEE Conf. Comput. Commun. Workshops
[Online]. Available: http://www-file.huawei.com/-/media/ (INFOCOM WKSHPS), 2019, pp. 264–269.
CORPORATE/PDF/News/5g-network-slicing-enabling-the-smart- [235] P. Ahokangas et al., “Business models for local 5G micro opera-
grid.pdf tors,” IEEE Trans. Cogn. Commun. Netw., vol. 5, no. 3, pp. 730–740,
[212] S. Sekander, H. Tabassum, and E. Hossain, “Multi-tier drone architec- Sep. 2019.
ture for 5G/B5G cellular networks: Challenges, trends, and prospects,” [236] (Jul. 2018). 5G and Network Slicing: A New Era in Networking.
IEEE Commun. Mag., vol. 56, no. 3, pp. 96–103, Mar. 2018. [Online]. Available: https://ieee-wf-5g.org/wp-content/uploads/sites/
[213] SFC/Network slicing. Accessed: Sep. 1, 2020. [Online]. Available: 102/2018/11/POLYCHRONOPOLISSlides.pdf
https://newslab.iith.ac.in/research/sfc.html
[214] C. Marquez, M. Gramaglia, M. Fiore, A. Banchs, and X. Costa-Pérez,
“Resource sharing efficiency in network slicing,” IEEE Trans. Netw. Shalitha Wijethilaka (Student Member, IEEE)
Service Manag., vol. 16, no. 3, pp. 909–923, Sep. 2019. received the bachelor’s degree (First Class Hons.)
[215] B. Khodapanah, A. Awada, I. Viering, J. Francis, M. Simsek, and in electronics and telecommunication engineering
G. P. Fettweis, “Radio resource management in context of network slic- from the University of Moratuwa, Sri Lanka, in
ing: What is missing in existing mechanisms?” in Proc. IEEE Wireless 2017. He is currently pursuing the Ph.D. degree with
Commun. Netw. Conf. (WCNC), 2019, pp. 1–7. the School of Computer Science, University College
[216] S. Underwood, “Blockchain beyond bitcoin,” Commun. ACM, vol. 59, Dublin, Ireland. His main research interests include
no. 11, pp. 15–17, 2016. network slicing, network security, and IoT.
[217] M. Swan, Blockchain: Blueprint for a New Economy, Sebastopol, CA,
USA: O’Reilly Media, Inc., 2015.
[218] T. Hewa, M. Ylianttila, and M. Liyanage, “Survey on blockchain based
smart contracts: Applications, opportunities and challenges,” J. Netw.
Comput. Appl., vol. 117, Mar. 2021, Art. no. 102857.
[219] A. Chaer, K. Salah, C. Lima, P. P. Ray, and T. Sheltami, “Blockchain for Madhusanka Liyanage (Senior Member, IEEE)
5G: Opportunities and challenges,” in Proc. IEEE Globecom Workshops received the B.Sc. degree (First Class Hons.) in elec-
(GC Wkshps), 2019, pp. 1–6. tronics and telecommunication engineering from the
[220] How Blockchain Can Impact the Telecommunications Industry and Its University of Moratuwa, Moratuwa, Sri Lanka, in
Relevance to the C-Suite. Accessed: Sep. 5, 2020. [Online]. Available: 2009, the M.Eng. degree from the Asian Institute
https://www2.deloitte.com/content/dam/Deloitte/za/Documents/ of Technology, Bangkok, Thailand, in 2011, the
technology-media-telecommunications/za_TMT_Blockchain_TelCo. M.Sc. degree from the University of Nice Sophia
pdf Antipolis, Nice, France, in 2011, and the Doctor
[221] A. Braeken, M. Liyanage, S. S. Kanhere, and S. Dixit, “Blockchain of Technology degree in communication engineer-
and cyberphysical systems,” Computer, vol. 53, no. 09, pp. 31–35, ing from the University of Oulu, Oulu, Finland, in
Sep. 2020. 2016. From 2011 to 2012, he worked a Research
[222] N. Weerasinghe, T. Hewa, M. Dissanayake, M. Ylianttila, and Scientist with the I3S Laboratory and Inria, Shopia Antipolis, France. He is
M. Liyanage, “Blockchain-based roaming and offload service platform currently an Assistant Professor/Ad Astra Fellow with the School of Computer
for local 5G operators,” in Proc. IEEE Consum. Commun. Netw. Conf. Science, University College Dublin, Ireland. He is also acting as an Adjunct
(CCNC), 2020, pp. 1–6. Processor with the Center for Wireless Communications, University of Oulu.
[223] B. Nour, A. Ksentini, N. Herbaut, P. A. Frangoudis, and H. Moungla, From 2015 to 2018, he has been a Visiting Research Fellow with the CSIRO,
“A blockchain-based network slice broker for 5G services,” IEEE Netw. Australia, the Infolabs21, Lancaster University, U.K., Computer Science and
Lett., vol. 1, no. 3, pp. 99–102, Sep. 2019. Engineering, The University of New South Wales, Australia, School of IT,
[224] L. Zanzi, A. Albanese, V. Sciancalepore, and X. Costa-Pérez, University of Sydney, Australia, LIP6, Sorbonne University, France, and
“NSBchain: A secure blockchain framework for network slicing bro- Computer Science and Engineering, The University of Oxford, U.K. His
kerage,” 2020. [Online]. Available: arXiv:2003.07748. research interests are 5G/6G, SDN, IoT, Blockchain, MEC, mobile and virtual
[225] K. Valtanen, J. Backman, and S. Yrjölä, “Creating value through network security. He was a recipient of prestigious Marie Skłodowska-Curie
blockchain powered resource configurations: Analysis of 5G network Actions Individual Fellowship from 2018 to 2020. In 2020, he has received the
slice brokering case,” in Proc. IEEE Wireless Commun. Netw. Conf. 2020 IEEE ComSoc Outstanding Young Researcher Award by IEEE ComSoc
Workshops (WCNCW), 2018, pp. 185–190. EMEA. More info: www.madhusanka.com.

Network Slicing

Uploaded by

Copyright:

Available Formats

Network Slicing

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Network Slicing

Uploaded by

Copyright:

Available Formats

IEEE COMMUNICATIONS SURVEYS & TUTORIALS, VOL. 23, NO.

2, SECOND QUARTER 2021 957

Survey on Network Slicing for Internet of Things

Abstract—Internet of Things (IoT) is an emerging technology

Fig. 2. E2E Network Slicing.

Fig. 3. Paper Organization.

C. Our Contribution 4) A concise summary of the network slicing-related

Fig. 4. Overall Network Slicing Architecture.

Fig. 5. Network Slicing Life Cycle.

Fig. 6. IoT application scenarios.

2) Related Works: Several slicing-based implementations

use case for highlighting the necessity of managing network

Fig. 11. Network Slicing & AR/VR/Gaming.

2) Related Works: In [131], Esteves et al. proposed a Proof-

Return On Investment (ROI) of network slicing in smart grids

Fig. 15. Network Slicing & Massive IoT.

Fig. 14. Network Slicing & Drones.

S UMMARY OF R ELATED W ORKS

(Continued.) S UMMARY OF R ELATED W ORKS

Network slicing is featured as the key component in this J. 5G-MONARCH

Fig. 16. Future research directions.

Fig. 17. Role of Blockchain for Network slicing in IoT.

Fig. 18. Role of AI/ML for network slicing in IoT.

Fig. 19. Multi-domain network slicing.

a hierarchical resource allocation architecture for network

You might also like