Comprehensive Survey on 5G: Next Generation Mobile Communication
VerifiedAdded on 2022/10/18
|30
|26153
|370
Essay
AI Summary
This document presents a comprehensive survey on 5G, the next generation of mobile communication. It begins by highlighting the limitations of 4G networks and the increasing demands of mobile devices, voluminous data, and higher data rates, setting the stage for the necessity of 5G. The s...
Accepted Manuscript
A survey on 5G: The next generation of mobile communication
Nisha Panwar, Shantanu Sharma, Awadhesh Kumar Singh
PII: S1874-4907(15)00053-1
DOI: http://dx.doi.org/10.1016/j.phycom.2015.10.006
Reference: PHYCOM 302
To appear in: Physical Communication
Received date: 30 June 2015
Revised date: 11 October 2015
Accepted date: 30 October 2015
Please cite this article as: N. Panwar, S. Sharma, A.K. Singh, A survey on 5G: The next
generation of mobile communication, Physical Communication (2015),
http://dx.doi.org/10.1016/j.phycom.2015.10.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a
service to our customers we are providing this early version of the manuscript. The manuscript
will undergo copyediting, typesetting, and review of the resulting proof before it is published in
its final form. Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
A survey on 5G: The next generation of mobile communication
Nisha Panwar, Shantanu Sharma, Awadhesh Kumar Singh
PII: S1874-4907(15)00053-1
DOI: http://dx.doi.org/10.1016/j.phycom.2015.10.006
Reference: PHYCOM 302
To appear in: Physical Communication
Received date: 30 June 2015
Revised date: 11 October 2015
Accepted date: 30 October 2015
Please cite this article as: N. Panwar, S. Sharma, A.K. Singh, A survey on 5G: The next
generation of mobile communication, Physical Communication (2015),
http://dx.doi.org/10.1016/j.phycom.2015.10.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a
service to our customers we are providing this early version of the manuscript. The manuscript
will undergo copyediting, typesetting, and review of the resulting proof before it is published in
its final form. Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
Secure Best Marks with AI Grader
Need help grading? Try our AI Grader for instant feedback on your assignments.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
A Survey on 5G: The Next Generation of Mobile Communication1
Nisha Panwar1, Shantanu Sharma1, and Awadhesh Kumar Singh2
2
1Department of Computer Science, Ben-Gurion University of the Negev, Israel.3
{panwar, sharmas}@cs.bgu.ac.il.4
2Department of Computer Engineering, National Institute of Technology, Kurukshetra, India.5
aksinreck@nitkkr.ac.in6
Abstract7
A rapidly increasing number of mobile devices, voluminous data, and higher data rate are pushing to rethink8
the current generation of the cellular mobile communication. The next or fifth generation (5G) cellular networks9
are expected to meet these requirements. The 5G networks are broadly characterized by three unique features:10
ubiquitous connectivity, very low latency, and very high-speed data transfer. The 5G networks will provide novel11
architectures and technologies beyond state-of-the-art architectures and technologies. In this paper, we will find12
an answer to the question: “what will be done by 5G and how?” We investigate and discuss serious limitations13
of the fourth generation (4G) cellular networks and corresponding new features of 5G networks. We identify14
challenges in 5G networks, new technologies for 5G networks, and a comparative discussion of the proposed15
architectures that can be categorized on the basis of energy-efficiency, network hierarchy, and network types.16
Interestingly, implementation issues, e.g., interference, QoS, handoff, security-privacy, channel access, and load17
balancing, hugely effect the realization of 5G networks. Furthermore, our discussion highlights the feasibility18
of these models through an evaluation of existing real-experiments and testbeds.19
Keywords: Cloud radio access networks; cognitive radio networks; D2D communication; dense deployment; multi-tier20
heterogeneous network; privacy; security; tactile Internet.21
1 Introduction22
The evolution of the cellular network generations is primarily influenced by a continuous growth in wireless user23
devices, data usage, and the need for a better quality of experience (QoE). It is expected that more than 50 billion24
connected devices will utilize the cellular network services by the end of the year 2020 [1], and it will result in a25
tremendous increase in data traffic, as compared to the year 2014 [2]. However, state-of-the-art solutions are not26
sufficient for the challenges mentioned above. In short, the increase of 3D (‘D’evice, ‘D’ata, and ‘D’ata transfer27
rate) encourages the development of 5G networks.28
Specifically, the fifth generation (5G) of the cellular networks will highlight and address three broad29
views, as: (i) user-centric (by providing 24 ×7 device connectivity, uninterrupted communication services,30
and a smooth consumer experience), (ii) service-provider-centric (by providing a connected intelligent31
transportation systems, road-side service units, sensors, and mission critical monitoring/tracking services),32
and (iii) network-operator-centric (by providing an energy-efficient, scalable, low-cost, uniformly-monitored,33
programmable, and secure communication infrastructure). Therefore, 5G networks are perceived to materialize34
the three main features as below:35
• Ubiquitous connectivity: In the future, many types of devices will connect ubiquitously and provide an36
uninterrupted user experience. In fact, the user-centric view will be realized by ubiquitous connectivity.37
• Zero latency: 5G networks will support life-critical systems and real-time applications and services with zero38
delay tolerance. Hence, it is envisioned that 5G networks will realize zero latency, i.e, very low latency of the39
order of 1 millisecond [3, 47]. In fact, the service-provider-centric view will be realized by the zero latency.40
• High-speed Gigabit connection: The zero latency property could be achieved using a high-speed connection for41
fast data transmission and reception, which will be of the order of Gigabits per second to users and machines [3].42
1
Manuscript
Click here to view linked References
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
A Survey on 5G: The Next Generation of Mobile Communication1
Nisha Panwar1, Shantanu Sharma1, and Awadhesh Kumar Singh2
2
1Department of Computer Science, Ben-Gurion University of the Negev, Israel.3
{panwar, sharmas}@cs.bgu.ac.il.4
2Department of Computer Engineering, National Institute of Technology, Kurukshetra, India.5
aksinreck@nitkkr.ac.in6
Abstract7
A rapidly increasing number of mobile devices, voluminous data, and higher data rate are pushing to rethink8
the current generation of the cellular mobile communication. The next or fifth generation (5G) cellular networks9
are expected to meet these requirements. The 5G networks are broadly characterized by three unique features:10
ubiquitous connectivity, very low latency, and very high-speed data transfer. The 5G networks will provide novel11
architectures and technologies beyond state-of-the-art architectures and technologies. In this paper, we will find12
an answer to the question: “what will be done by 5G and how?” We investigate and discuss serious limitations13
of the fourth generation (4G) cellular networks and corresponding new features of 5G networks. We identify14
challenges in 5G networks, new technologies for 5G networks, and a comparative discussion of the proposed15
architectures that can be categorized on the basis of energy-efficiency, network hierarchy, and network types.16
Interestingly, implementation issues, e.g., interference, QoS, handoff, security-privacy, channel access, and load17
balancing, hugely effect the realization of 5G networks. Furthermore, our discussion highlights the feasibility18
of these models through an evaluation of existing real-experiments and testbeds.19
Keywords: Cloud radio access networks; cognitive radio networks; D2D communication; dense deployment; multi-tier20
heterogeneous network; privacy; security; tactile Internet.21
1 Introduction22
The evolution of the cellular network generations is primarily influenced by a continuous growth in wireless user23
devices, data usage, and the need for a better quality of experience (QoE). It is expected that more than 50 billion24
connected devices will utilize the cellular network services by the end of the year 2020 [1], and it will result in a25
tremendous increase in data traffic, as compared to the year 2014 [2]. However, state-of-the-art solutions are not26
sufficient for the challenges mentioned above. In short, the increase of 3D (‘D’evice, ‘D’ata, and ‘D’ata transfer27
rate) encourages the development of 5G networks.28
Specifically, the fifth generation (5G) of the cellular networks will highlight and address three broad29
views, as: (i) user-centric (by providing 24 ×7 device connectivity, uninterrupted communication services,30
and a smooth consumer experience), (ii) service-provider-centric (by providing a connected intelligent31
transportation systems, road-side service units, sensors, and mission critical monitoring/tracking services),32
and (iii) network-operator-centric (by providing an energy-efficient, scalable, low-cost, uniformly-monitored,33
programmable, and secure communication infrastructure). Therefore, 5G networks are perceived to materialize34
the three main features as below:35
• Ubiquitous connectivity: In the future, many types of devices will connect ubiquitously and provide an36
uninterrupted user experience. In fact, the user-centric view will be realized by ubiquitous connectivity.37
• Zero latency: 5G networks will support life-critical systems and real-time applications and services with zero38
delay tolerance. Hence, it is envisioned that 5G networks will realize zero latency, i.e, very low latency of the39
order of 1 millisecond [3, 47]. In fact, the service-provider-centric view will be realized by the zero latency.40
• High-speed Gigabit connection: The zero latency property could be achieved using a high-speed connection for41
fast data transmission and reception, which will be of the order of Gigabits per second to users and machines [3].42
1
Manuscript
Click here to view linked References
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
A few more key features of 5G networks are enlisted and compared to the fourth generation (4G) of the cellular43
networks, as below [4, 5, 6]: (i) 10-100xnumber of connected devices, (ii) 1000x higher mobile data volume per44
area, (iii) 10-100xhigher data rate, (iv) 1 millisecond latency, (v) 99.99% availability, (vi) 100% coverage, (vii) x
1045
energy consumption as compared to the year 2010, (viii) real-time information processing and transmission, (ix) x
546
network management operation expenses, and (x) seamless integration of the current wireless technologies.47
5G
Networks
Increased
data rate
&
network
capacity
Densification, FDD,
CRN, mMIMO, D2D
communication, full
duplex radio
Multi-RAT, self-heal,
densification, CRN,
NFV, SDN, C-RAN,
RANaaS, CONCERT,
Low latency
Cache, fast
handoff, D2D
communication,
mobile small-
cells, self-heal Scalability
Environmental
friendly & less
money
QoSSecurity &
privacy
Interference &
handoff
management
NFV, SDN,
C-RAN,
RANaaS,
CONCERT
Delay-bound QoS,
Quality management
equipment, multi-links
with multi-flow and
multi-QoS
C-RAN, VLC,
mmWave,
mMIMO, small-
cells, D2D
communication,
user separation
Monitoring and
encryption-decryption
SIC, CRN,
advance receiver,
joint
detection/decodi
ng
Inter-tier, intra-
tier, and
multi-RAT
handoff,
Figure 1: Requirements and proposed solutions for the
development of 5G networks. The inner, middle, and
outermost layers present requirements, solutions, and
applications of 5G networks, respectively. Two colored
wedges highlight primary features of 5G networks.
Therefore, the revolutionary scope and the48
consequent advantages of the envisioned 5G49
networks demand new architectures, methodologies,50
and technologies (see Figure 1), e.g., energy-efficient51
heterogeneous frameworks, cloud-based52
communication (software-defined networks (SDN)53
and network function virtualization (NFV)),54
full duplex radio, self-interference cancellation55
(SIC), device-to-device (D2D) communications,56
machine-to-machine (M2M) communications, access57
protocols, cheap devices, cognitive networks (for58
accessing licensed, unlicensed, and shared frequency59
bands), dense-deployment, security-privacy60
protocols for communication and data transfer,61
backhaul connections, massive multiple-input and62
multiple-output (mMIMO), multi-radio access63
technology (RAT) architectures, and technologies64
for working on millimeter wave (mmWave) 30–30065
GHz. Interestingly, 5G networks will not be a mere66
enhancement of 4G networks in terms of additional67
capacity; they will encompass a system architecture68
visualization, conceptualization, and redesigning at69
every communication layer [51].70
Several industries, Alcatel-Lucent [7],71
DOCOMO [8], GSMA Intelligence [5], Huawei [9],72
Nokia Siemens Networks [3], Qualcomm [10], Samsung [11], Vodafone,1 the European Commission supported73
5G Infrastructure Public Private Partnership (5GPPP) [4], and Mobile and Wireless Communications Enablers for74
the Twenty-Twenty Information Society (METIS) [6], are brainstorming with the development of 5G networks.75
Currently, the industry standards are yet to be explored about the expected designs and architectures for 5G76
networks.77
Scope of the paper. In this paper, we will review the vision of the 5G networks, advantages, applications, proposed78
architectures, implementation issues, real demonstrations, and testbeds. The outline of the paper is provided in79
Figure 2. In Section 2, we will discuss the vision of 5G networks. Section 3 presents challenges in the development80
of 5G networks. Section 4 address the current proposed architectures for 5G networks, e.g., multi-tier, cognitive81
radio based, cloud-based, device proximity based, and energy-efficient architectures. Section 5 presents issues82
regarding interference, handoff, quality of services, load balancing, channel access, and security-privacy of the83
network. Sections 6, 7, and 8 present several methodologies and technologies involved in 5G networks, applications84
of 5G networks, and real demonstrations and testbeds of 5G networks, respectively.85
We would like to emphasize that there are some review works on 5G networks by Andrews et al. [20],86
Chávez-Santiago et al. [35], and Gavrilovska et al. [50], to the best of our knowledge. However, our perspective87
about 5G networks is different, as we deal with a variety of architectures and discuss several implementation affairs,88
technologies in 5G networks along with applications and real-testbed demonstrations. In addition, we intentionally89
avoid an mmWave oriented discussion in this paper, unlike the current work [20, 35, 50].90
We encourage our readers to see an overview about the generations of the cellular networks (see Table 1) and91
crucial limitations of the current cellular networks in the next section.92
1http://www.surrey.ac.uk/5gic/research
2
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
A few more key features of 5G networks are enlisted and compared to the fourth generation (4G) of the cellular43
networks, as below [4, 5, 6]: (i) 10-100xnumber of connected devices, (ii) 1000x higher mobile data volume per44
area, (iii) 10-100xhigher data rate, (iv) 1 millisecond latency, (v) 99.99% availability, (vi) 100% coverage, (vii) x
1045
energy consumption as compared to the year 2010, (viii) real-time information processing and transmission, (ix) x
546
network management operation expenses, and (x) seamless integration of the current wireless technologies.47
5G
Networks
Increased
data rate
&
network
capacity
Densification, FDD,
CRN, mMIMO, D2D
communication, full
duplex radio
Multi-RAT, self-heal,
densification, CRN,
NFV, SDN, C-RAN,
RANaaS, CONCERT,
Low latency
Cache, fast
handoff, D2D
communication,
mobile small-
cells, self-heal Scalability
Environmental
friendly & less
money
QoSSecurity &
privacy
Interference &
handoff
management
NFV, SDN,
C-RAN,
RANaaS,
CONCERT
Delay-bound QoS,
Quality management
equipment, multi-links
with multi-flow and
multi-QoS
C-RAN, VLC,
mmWave,
mMIMO, small-
cells, D2D
communication,
user separation
Monitoring and
encryption-decryption
SIC, CRN,
advance receiver,
joint
detection/decodi
ng
Inter-tier, intra-
tier, and
multi-RAT
handoff,
Figure 1: Requirements and proposed solutions for the
development of 5G networks. The inner, middle, and
outermost layers present requirements, solutions, and
applications of 5G networks, respectively. Two colored
wedges highlight primary features of 5G networks.
Therefore, the revolutionary scope and the48
consequent advantages of the envisioned 5G49
networks demand new architectures, methodologies,50
and technologies (see Figure 1), e.g., energy-efficient51
heterogeneous frameworks, cloud-based52
communication (software-defined networks (SDN)53
and network function virtualization (NFV)),54
full duplex radio, self-interference cancellation55
(SIC), device-to-device (D2D) communications,56
machine-to-machine (M2M) communications, access57
protocols, cheap devices, cognitive networks (for58
accessing licensed, unlicensed, and shared frequency59
bands), dense-deployment, security-privacy60
protocols for communication and data transfer,61
backhaul connections, massive multiple-input and62
multiple-output (mMIMO), multi-radio access63
technology (RAT) architectures, and technologies64
for working on millimeter wave (mmWave) 30–30065
GHz. Interestingly, 5G networks will not be a mere66
enhancement of 4G networks in terms of additional67
capacity; they will encompass a system architecture68
visualization, conceptualization, and redesigning at69
every communication layer [51].70
Several industries, Alcatel-Lucent [7],71
DOCOMO [8], GSMA Intelligence [5], Huawei [9],72
Nokia Siemens Networks [3], Qualcomm [10], Samsung [11], Vodafone,1 the European Commission supported73
5G Infrastructure Public Private Partnership (5GPPP) [4], and Mobile and Wireless Communications Enablers for74
the Twenty-Twenty Information Society (METIS) [6], are brainstorming with the development of 5G networks.75
Currently, the industry standards are yet to be explored about the expected designs and architectures for 5G76
networks.77
Scope of the paper. In this paper, we will review the vision of the 5G networks, advantages, applications, proposed78
architectures, implementation issues, real demonstrations, and testbeds. The outline of the paper is provided in79
Figure 2. In Section 2, we will discuss the vision of 5G networks. Section 3 presents challenges in the development80
of 5G networks. Section 4 address the current proposed architectures for 5G networks, e.g., multi-tier, cognitive81
radio based, cloud-based, device proximity based, and energy-efficient architectures. Section 5 presents issues82
regarding interference, handoff, quality of services, load balancing, channel access, and security-privacy of the83
network. Sections 6, 7, and 8 present several methodologies and technologies involved in 5G networks, applications84
of 5G networks, and real demonstrations and testbeds of 5G networks, respectively.85
We would like to emphasize that there are some review works on 5G networks by Andrews et al. [20],86
Chávez-Santiago et al. [35], and Gavrilovska et al. [50], to the best of our knowledge. However, our perspective87
about 5G networks is different, as we deal with a variety of architectures and discuss several implementation affairs,88
technologies in 5G networks along with applications and real-testbed demonstrations. In addition, we intentionally89
avoid an mmWave oriented discussion in this paper, unlike the current work [20, 35, 50].90
We encourage our readers to see an overview about the generations of the cellular networks (see Table 1) and91
crucial limitations of the current cellular networks in the next section.92
1http://www.surrey.ac.uk/5gic/research
2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Generations Year Features Limitations
1G 1980s Analog signals for voice only communications Very less security
2G 1990s Digital signals, voice communications, and text
messaging
Very less support for the Internet
3G 1998-99 Voice communications, wireless mobile and fixed
Internet access, video calls, and mobile television (TV)
Less support for high-speed
Internet
4G 2008-09 Higher data rate (hundreds of megabits per second) No support for 50 billion ubiquitous
connected devices
5G 2020 Mentioned in Section 1
Table 1: The generations of the cellular networks.
1. Introduction
An introduction
of the paper and
scope of the paper
2. Desideratum
of 5G Networks
Dramatic upsurge in
device scalability,
massive data
streaming and high
data rate, spectrum
utilization,
ubiquitous
connectivity, and
zero latency
1.1 Limitations of
the current cellular
networks
No support for
bursty data
traffic, inefficient
utilization of
processing
capabilities of a
base-station, co-
channel
interference, no
support for
heterogeneous
wireless
networks, and no
separation of
indoor and
outdoor users
4. Architectures of the
Future/5G Mobile
Cellular Networks
4.1 Two-tier Architectures
How small-cells are
deployed under macrocells?
A Survey on 5G: The Next Generation of Mobile Communication
3. Challenges in
the Development
of 5G Networks
Increase data rate
and network
capacity with low
power consumption,
scalability and
flexibility, handling
interference,
environmental
friendly, low latency
and high reliability,
price, high mobility,
self-healing
infrastructures, QoS,
and security and
privacy of the
network and UEs
5.Implementation
Issues in 5G
Networks
6. Methodologies
and Technologies
for 5G Networks
Remaining
methodologies and
technologies are
discussed, e.g., SIC,
DUD, NFV, SDN,
mmWave, M2M
communication,
mMIMO, VLC
7. Applications
of 5G Networks
Personal usages,
virtualized homes,
smart societies,
smart grids, the
tactile Internet,
automation, health-
care systems,
logistics and
tracking, and
industrial usages
8. Real
Demonstrations
of 5G Networks
How industries and
academia are
looking towards
5G? What kind of
real
implementations
and testbeds they
are doing?
5.1 Interference
Management
4.2 CRN-based
Architectures
How CRNs are deployed
under a macrocell?
4.3 D2D Communication
Architectures
How devices communicate
to their close devices
without involving a MBS?
4.4 Cloud-based
Architectures
How the cloud facilitate
communication in 5G
networks?
4.5 Energy-efficient
Architectures
How to save energy in 5G
networks?
5.2 Handoff
Management
5.3 QoS
Management
5.4 Load
balancing
5.5 Channel
Access Control
Management
5.6 Security and
Privacy
Management in
5G Networks
Figure 2: Schematic map of the paper.
1.1 Limitations of the Conventional Cellular Systems93
4G networks are not substantial enough to support massively connected devices with low latency and significant94
spectral efficiency, which will be crucial in the future. In this section, we discuss a few crucial aspects in which95
conventional cellular networks lag behind, thereby motivating the evolution of 5G networks.96
No support for bursty data traffic. There are several mobile applications that send heartbeat messages to their97
servers and occasionally ask for a very high data transfer speed for a very short duration. Such types of data98
transmission may consume more battery life of (mobile) user equipments (UEs) with increasing bursty data in the99
network, and hence, may crash the core network [123]. However, only one type of signaling/control mechanism is100
designed for all types of the traffic in the current networks, creating high overhead for bursty traffic [64, 25].101
Inefficient utilization of processing capabilities of a base-station. In the current cellular networks, the processing102
power of a base-station (BS) can only be used by its associated UEs, and they are designed to support peak time103
traffic. However, a BS’s processing power can be shared across a large geographical area when it is lightly loaded.104
For example: (i) during the day, BSs in business areas are over-subscribed, while BSs in residential areas are105
almost idle, and vice versa [115], and (ii) BSs in residential areas are overloaded in weekends or holidays while106
BSs in business areas are almost idle [92]. However, the almost idle BSs consume an identical amount of power as107
over-subscribed BSs, hence, the overall cost of the network increases.108
Co-channel interference. A typical cellular network uses two separate channels, one as a transmission path from a109
UE to a BS, called uplink (UL), and the reverse path, called downlink (DL). The allocation of two different channels110
for a UE is not an efficient utilization of the frequency band. However, if both the channels operate at an identical111
frequency, i.e., a full duplex wireless radio [27], then a high level of co-channel interference (the interference112
between the signals using an identical frequency) in UL and DL channels is a major issue in 4G networks [86]. It113
3
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Generations Year Features Limitations
1G 1980s Analog signals for voice only communications Very less security
2G 1990s Digital signals, voice communications, and text
messaging
Very less support for the Internet
3G 1998-99 Voice communications, wireless mobile and fixed
Internet access, video calls, and mobile television (TV)
Less support for high-speed
Internet
4G 2008-09 Higher data rate (hundreds of megabits per second) No support for 50 billion ubiquitous
connected devices
5G 2020 Mentioned in Section 1
Table 1: The generations of the cellular networks.
1. Introduction
An introduction
of the paper and
scope of the paper
2. Desideratum
of 5G Networks
Dramatic upsurge in
device scalability,
massive data
streaming and high
data rate, spectrum
utilization,
ubiquitous
connectivity, and
zero latency
1.1 Limitations of
the current cellular
networks
No support for
bursty data
traffic, inefficient
utilization of
processing
capabilities of a
base-station, co-
channel
interference, no
support for
heterogeneous
wireless
networks, and no
separation of
indoor and
outdoor users
4. Architectures of the
Future/5G Mobile
Cellular Networks
4.1 Two-tier Architectures
How small-cells are
deployed under macrocells?
A Survey on 5G: The Next Generation of Mobile Communication
3. Challenges in
the Development
of 5G Networks
Increase data rate
and network
capacity with low
power consumption,
scalability and
flexibility, handling
interference,
environmental
friendly, low latency
and high reliability,
price, high mobility,
self-healing
infrastructures, QoS,
and security and
privacy of the
network and UEs
5.Implementation
Issues in 5G
Networks
6. Methodologies
and Technologies
for 5G Networks
Remaining
methodologies and
technologies are
discussed, e.g., SIC,
DUD, NFV, SDN,
mmWave, M2M
communication,
mMIMO, VLC
7. Applications
of 5G Networks
Personal usages,
virtualized homes,
smart societies,
smart grids, the
tactile Internet,
automation, health-
care systems,
logistics and
tracking, and
industrial usages
8. Real
Demonstrations
of 5G Networks
How industries and
academia are
looking towards
5G? What kind of
real
implementations
and testbeds they
are doing?
5.1 Interference
Management
4.2 CRN-based
Architectures
How CRNs are deployed
under a macrocell?
4.3 D2D Communication
Architectures
How devices communicate
to their close devices
without involving a MBS?
4.4 Cloud-based
Architectures
How the cloud facilitate
communication in 5G
networks?
4.5 Energy-efficient
Architectures
How to save energy in 5G
networks?
5.2 Handoff
Management
5.3 QoS
Management
5.4 Load
balancing
5.5 Channel
Access Control
Management
5.6 Security and
Privacy
Management in
5G Networks
Figure 2: Schematic map of the paper.
1.1 Limitations of the Conventional Cellular Systems93
4G networks are not substantial enough to support massively connected devices with low latency and significant94
spectral efficiency, which will be crucial in the future. In this section, we discuss a few crucial aspects in which95
conventional cellular networks lag behind, thereby motivating the evolution of 5G networks.96
No support for bursty data traffic. There are several mobile applications that send heartbeat messages to their97
servers and occasionally ask for a very high data transfer speed for a very short duration. Such types of data98
transmission may consume more battery life of (mobile) user equipments (UEs) with increasing bursty data in the99
network, and hence, may crash the core network [123]. However, only one type of signaling/control mechanism is100
designed for all types of the traffic in the current networks, creating high overhead for bursty traffic [64, 25].101
Inefficient utilization of processing capabilities of a base-station. In the current cellular networks, the processing102
power of a base-station (BS) can only be used by its associated UEs, and they are designed to support peak time103
traffic. However, a BS’s processing power can be shared across a large geographical area when it is lightly loaded.104
For example: (i) during the day, BSs in business areas are over-subscribed, while BSs in residential areas are105
almost idle, and vice versa [115], and (ii) BSs in residential areas are overloaded in weekends or holidays while106
BSs in business areas are almost idle [92]. However, the almost idle BSs consume an identical amount of power as107
over-subscribed BSs, hence, the overall cost of the network increases.108
Co-channel interference. A typical cellular network uses two separate channels, one as a transmission path from a109
UE to a BS, called uplink (UL), and the reverse path, called downlink (DL). The allocation of two different channels110
for a UE is not an efficient utilization of the frequency band. However, if both the channels operate at an identical111
frequency, i.e., a full duplex wireless radio [27], then a high level of co-channel interference (the interference112
between the signals using an identical frequency) in UL and DL channels is a major issue in 4G networks [86]. It113
3
Secure Best Marks with AI Grader
Need help grading? Try our AI Grader for instant feedback on your assignments.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
also prevents the network densification, i.e., the deployment of many BSs in a geographical area.114
No support for heterogeneous wireless networks. Heterogeneous wireless networks (HetNets) are composed of115
wireless networks of diverse access technologies, e.g., the third generation (3G), 4G, wireless local area networks116
(WLAN), WiFi, and Bluetooth. HetNets are already standardized in 4G; however, the basic architecture was not117
intended to support them. Furthermore, the current cellular networks allow a UE to have a DL channel and a UL118
channel must be associated with a single BS that prevents the maximum utilization of HetNets. In HetNets, a UE119
can select a UL channel and a DL channel from two different BSs belonging to two different wireless networks for120
performance improvement [29, 42].121
No separation of indoor and outdoor users. The current cellular networks have a single BS installed near the122
center of the cell and interacts with all the UEs irrespective of the indoor or outdoor location of the UEs; while123
UEs stay indoors and outdoors for about 80% and 20% of the time, respectively. Furthermore, the communication124
between an indoor UE and an outside BS is not efficient in terms of data transfer rate, spectral efficiency, and125
energy-efficiency, due to the attenuation of signals passing through walls [107].126
Latency. When a UE receives an access to the best candidate BS, it takes several hundreds of milliseconds in the127
current cellular networks [121], and hence, they cannot support the zero latency property.128
2 Desideratum of 5G Networks129
A growing number of UEs and the corresponding surge in the bandwidth requirement for the huge amount of data130
transmission certainly necessitate the novel enhancement to the current technology. In this section, we highlight131
requirements of the future 5G networks.132
Dramatic upsurge in device scalability. A rapid growth of smart phones, gaming consoles, high-resolution TVs,133
cameras, home appliances, laptops, connected transportation systems, video surveillance systems, robots, sensors,134
and wearable devices (watches and glasses) is expected to continue exponentially in the near future. Therefore, 5G135
networks are perceived to support massively connected devices [107, 1, 15].136
Massive data streaming and high data rate. A vast growth in a number of wireless devices will of course137
result in a higher amount of data trading (e.g., videos, audio, Web browsing, social-media data, gaming, real-time138
signals, photos, bursty data, and multimedia) that will be 100-times more as compared to the year 2014 and would139
overburden the current network. Thus, it is mandatory to have matching data transfer capabilities in terms of new140
architectures, methods, technologies, and data distribution of indoor and outdoor users [61, 15, 60].141
Spectrum utilization. The two different channels (one for a UL and another for a DL) seem redundant from142
the point of view of the spectrum utilization [59]. In addition, the currently allocated spectrums have their143
significant portions under-utilized [12]. Hence, it is necessary to develop an access control method that can144
enhance the spectrum utilization. Furthermore, the spectrum utilization and efficiency have already been stretched145
to the maximum. It definitely requires spectrum broadening (above 3 GHz) along with novel spectrum utilization146
techniques [34].147
Ubiquitous connectivity. Ubiquitous connectivity requires UEs to support a variety of radios, RATs, and bands148
due to the global non-identical operating bands. In addition, the major market split between time division duplex149
(e.g., India and China) versus frequency division duplex (e.g., US and Europe) so that UEs are required to support150
different duplex options. Hence, 5G networks are envisioned for seamless connectivity of UEs over HetNets [13].151
Zero latency. The future mobile cellular networks are expected to assist numerous real-time applications, the tactile152
Internet [47, 46], and services with varying levels of quality of service (QoS) (in terms of bandwidth, latency, jitter,153
packet loss, and packet delay) and QoE (in terms of users’ and network-providers’ service satisfaction versus154
feedback). Hence, 5G networks are envisioned to realize real-time and delay-bound services with the optimal QoS155
and QoE experiences [15, 86].156
3 Challenges in the Development of 5G Networks157
The vision of 5G networks is not trivial to achieve. There are several challenges (some of the following challenges158
are shown in Figure 1 with their proposed solutions) to be handled in that context, as mentioned below:159
4
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
also prevents the network densification, i.e., the deployment of many BSs in a geographical area.114
No support for heterogeneous wireless networks. Heterogeneous wireless networks (HetNets) are composed of115
wireless networks of diverse access technologies, e.g., the third generation (3G), 4G, wireless local area networks116
(WLAN), WiFi, and Bluetooth. HetNets are already standardized in 4G; however, the basic architecture was not117
intended to support them. Furthermore, the current cellular networks allow a UE to have a DL channel and a UL118
channel must be associated with a single BS that prevents the maximum utilization of HetNets. In HetNets, a UE119
can select a UL channel and a DL channel from two different BSs belonging to two different wireless networks for120
performance improvement [29, 42].121
No separation of indoor and outdoor users. The current cellular networks have a single BS installed near the122
center of the cell and interacts with all the UEs irrespective of the indoor or outdoor location of the UEs; while123
UEs stay indoors and outdoors for about 80% and 20% of the time, respectively. Furthermore, the communication124
between an indoor UE and an outside BS is not efficient in terms of data transfer rate, spectral efficiency, and125
energy-efficiency, due to the attenuation of signals passing through walls [107].126
Latency. When a UE receives an access to the best candidate BS, it takes several hundreds of milliseconds in the127
current cellular networks [121], and hence, they cannot support the zero latency property.128
2 Desideratum of 5G Networks129
A growing number of UEs and the corresponding surge in the bandwidth requirement for the huge amount of data130
transmission certainly necessitate the novel enhancement to the current technology. In this section, we highlight131
requirements of the future 5G networks.132
Dramatic upsurge in device scalability. A rapid growth of smart phones, gaming consoles, high-resolution TVs,133
cameras, home appliances, laptops, connected transportation systems, video surveillance systems, robots, sensors,134
and wearable devices (watches and glasses) is expected to continue exponentially in the near future. Therefore, 5G135
networks are perceived to support massively connected devices [107, 1, 15].136
Massive data streaming and high data rate. A vast growth in a number of wireless devices will of course137
result in a higher amount of data trading (e.g., videos, audio, Web browsing, social-media data, gaming, real-time138
signals, photos, bursty data, and multimedia) that will be 100-times more as compared to the year 2014 and would139
overburden the current network. Thus, it is mandatory to have matching data transfer capabilities in terms of new140
architectures, methods, technologies, and data distribution of indoor and outdoor users [61, 15, 60].141
Spectrum utilization. The two different channels (one for a UL and another for a DL) seem redundant from142
the point of view of the spectrum utilization [59]. In addition, the currently allocated spectrums have their143
significant portions under-utilized [12]. Hence, it is necessary to develop an access control method that can144
enhance the spectrum utilization. Furthermore, the spectrum utilization and efficiency have already been stretched145
to the maximum. It definitely requires spectrum broadening (above 3 GHz) along with novel spectrum utilization146
techniques [34].147
Ubiquitous connectivity. Ubiquitous connectivity requires UEs to support a variety of radios, RATs, and bands148
due to the global non-identical operating bands. In addition, the major market split between time division duplex149
(e.g., India and China) versus frequency division duplex (e.g., US and Europe) so that UEs are required to support150
different duplex options. Hence, 5G networks are envisioned for seamless connectivity of UEs over HetNets [13].151
Zero latency. The future mobile cellular networks are expected to assist numerous real-time applications, the tactile152
Internet [47, 46], and services with varying levels of quality of service (QoS) (in terms of bandwidth, latency, jitter,153
packet loss, and packet delay) and QoE (in terms of users’ and network-providers’ service satisfaction versus154
feedback). Hence, 5G networks are envisioned to realize real-time and delay-bound services with the optimal QoS155
and QoE experiences [15, 86].156
3 Challenges in the Development of 5G Networks157
The vision of 5G networks is not trivial to achieve. There are several challenges (some of the following challenges158
are shown in Figure 1 with their proposed solutions) to be handled in that context, as mentioned below:159
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Data rate and network capacity expansion with energy optimization. The deployment of more BSs in a160
geographical area, use of the higher frequency bands, and link improvement might support the network capacity161
expansion, billions of UEs, high data rate, high volume of data, and efficient backhaul data transfer to the core162
network. However, the implementation of these solutions is a cumbersome task in terms of economy and energy163
intake. Hence, the network capacity is required to be significantly increased, keeping the energy consumption and164
cost under strict control.165
Proposed solutions: Network densification or small-cell deployment [15, 28, 107] (Section 4.1), cognitive166
radio networks (CRNs) [16] (Section 4.2), mMIMO [71, 81, 87] (Section 6), network offload using D2D167
communication [33, 104, 113] (Section 4.3), efficient backhaul networks [51, 88] (Section 4.1.1), energy-efficient168
architectures [62, 83] (Section 4.5), full duplex radios [27] (Section 6), NFV, and SDN based architectures [14, 78,169
97, 119] (Section 6).170
Scalability and flexibility. These are the most prominent features of the future mobile communication. The171
future cellular infrastructures and methodologies must be designed to work in HetNets. Moreover, a vast number172
of potential users might request simultaneously for a set of services. Therefore, 5G networks must be powerful173
enough to support a scalable user demand across the coverage area [78, 94].174
Proposed solutions: NFV- and SDN-based architectures [14, 78, 97, 119] (Section 6).175
Single channel for both UL and DL. A full duplex wireless radio [27] uses only a single channel for transmitting176
and receiving signals at identical time and frequency. Thus, a full duplex system achieves an identical performance177
as having different UL and DL channels, and hence, increases link capacity, saves the spectrum, and cost. However,178
the implementation of full duplex systems is not trivial, because now a radio has to use sophisticated protocols179
for the physical and the data link layers [122], and mechanisms to remove the effects of interference [59]. The180
advantages of a full duplex radio in 5G networks are given in [56, 59, 64].181
Handling interference. Handling interference among communicating devices is a well-known challenge in the182
wireless communication. Due to a growing number of UEs, technologies (e.g., HetNets, CRNs, full duplex, and183
D2D communication) and applications, the interference will also increase in 5G networks, and the state-of-the-art184
technique may not perform well in the future cellular networks [61]. In 5G networks, a UE may receive interference185
from multiple macrocell base-stations (MBSs), various UEs, and small-cell base-stations (SBSs). Hence, it is186
required to develop an efficient (in terms of avoiding network overload) and reliable (in terms of perfect interference187
detection and decoding) interference management technique for channel allocation, power control, cell association,188
and load balancing.189
Proposed solutions: Self-interference cancellation [64, 59], an advance receiver with interference joint190
detection/decoding, and network-side interference management [86]. We will discuss these solutions in Section 5.1.191
Environmentally friendly. The current radio access network (RAN) consumes 70%-80%of the total power [64,192
114]. The wireless technologies consume lots of energy that lead to huge CO 2 emission and inflate the cost. It is193
a serious threat to the environment [107]. Thus, it is required to develop energy-efficient communication systems,194
hardware, and technologies, thereby the ratio between the network throughput and energy consumption is equitable.195
Proposed solutions: Cloud-RAN (C-RAN) [114, 62], visual light communication (VLC) [114], mmWave [114],196
separation of indoor and outdoor users [114], joint investigation of spectral efficiency and energy-efficacy [64, 62],197
multi-tier architectures [62], D2D communication [33, 104, 113], mMIMO architectures [62], and full duplex198
radios [64]. Except the above mentioned solutions, we will discuss some special techniques/architectures in the199
context of energy-efficiency in 5G networks in Section 4.5.200
Low latency and high reliability. Low latency and high reliability are critical in several real-time applications,201
e.g., message transmission by robots monitoring patients, life safety systems, cloud-based gaming, nuclear reactors,202
sensors, drones, and connected transportation systems. However, it is very challenging to have very low latency and203
reliable delivery of data over a large scale network without increasing the network infrastructure cost, as it requires204
the development of techniques providing fast connections, quick handovers, and high data transfer rate.205
Proposed solutions: Caching methods [29, 112], VLC, mmWave, mMIMO (Section 6), fast handover206
techniques [40, 93, 102] (Section 5.2), and D2D communication (Section 4.3).207
Network performance optimization. The performance parameters, e.g., peak data rate, geographical area208
coverage, spectral efficiency, QoS, QoE, ease of connectivity, energy-efficiency, latency, reliability, fairness of209
users, and implementation complexity, are crucial for a cellular network [107]. Hence, a general framework for 5G210
5
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Data rate and network capacity expansion with energy optimization. The deployment of more BSs in a160
geographical area, use of the higher frequency bands, and link improvement might support the network capacity161
expansion, billions of UEs, high data rate, high volume of data, and efficient backhaul data transfer to the core162
network. However, the implementation of these solutions is a cumbersome task in terms of economy and energy163
intake. Hence, the network capacity is required to be significantly increased, keeping the energy consumption and164
cost under strict control.165
Proposed solutions: Network densification or small-cell deployment [15, 28, 107] (Section 4.1), cognitive166
radio networks (CRNs) [16] (Section 4.2), mMIMO [71, 81, 87] (Section 6), network offload using D2D167
communication [33, 104, 113] (Section 4.3), efficient backhaul networks [51, 88] (Section 4.1.1), energy-efficient168
architectures [62, 83] (Section 4.5), full duplex radios [27] (Section 6), NFV, and SDN based architectures [14, 78,169
97, 119] (Section 6).170
Scalability and flexibility. These are the most prominent features of the future mobile communication. The171
future cellular infrastructures and methodologies must be designed to work in HetNets. Moreover, a vast number172
of potential users might request simultaneously for a set of services. Therefore, 5G networks must be powerful173
enough to support a scalable user demand across the coverage area [78, 94].174
Proposed solutions: NFV- and SDN-based architectures [14, 78, 97, 119] (Section 6).175
Single channel for both UL and DL. A full duplex wireless radio [27] uses only a single channel for transmitting176
and receiving signals at identical time and frequency. Thus, a full duplex system achieves an identical performance177
as having different UL and DL channels, and hence, increases link capacity, saves the spectrum, and cost. However,178
the implementation of full duplex systems is not trivial, because now a radio has to use sophisticated protocols179
for the physical and the data link layers [122], and mechanisms to remove the effects of interference [59]. The180
advantages of a full duplex radio in 5G networks are given in [56, 59, 64].181
Handling interference. Handling interference among communicating devices is a well-known challenge in the182
wireless communication. Due to a growing number of UEs, technologies (e.g., HetNets, CRNs, full duplex, and183
D2D communication) and applications, the interference will also increase in 5G networks, and the state-of-the-art184
technique may not perform well in the future cellular networks [61]. In 5G networks, a UE may receive interference185
from multiple macrocell base-stations (MBSs), various UEs, and small-cell base-stations (SBSs). Hence, it is186
required to develop an efficient (in terms of avoiding network overload) and reliable (in terms of perfect interference187
detection and decoding) interference management technique for channel allocation, power control, cell association,188
and load balancing.189
Proposed solutions: Self-interference cancellation [64, 59], an advance receiver with interference joint190
detection/decoding, and network-side interference management [86]. We will discuss these solutions in Section 5.1.191
Environmentally friendly. The current radio access network (RAN) consumes 70%-80%of the total power [64,192
114]. The wireless technologies consume lots of energy that lead to huge CO 2 emission and inflate the cost. It is193
a serious threat to the environment [107]. Thus, it is required to develop energy-efficient communication systems,194
hardware, and technologies, thereby the ratio between the network throughput and energy consumption is equitable.195
Proposed solutions: Cloud-RAN (C-RAN) [114, 62], visual light communication (VLC) [114], mmWave [114],196
separation of indoor and outdoor users [114], joint investigation of spectral efficiency and energy-efficacy [64, 62],197
multi-tier architectures [62], D2D communication [33, 104, 113], mMIMO architectures [62], and full duplex198
radios [64]. Except the above mentioned solutions, we will discuss some special techniques/architectures in the199
context of energy-efficiency in 5G networks in Section 4.5.200
Low latency and high reliability. Low latency and high reliability are critical in several real-time applications,201
e.g., message transmission by robots monitoring patients, life safety systems, cloud-based gaming, nuclear reactors,202
sensors, drones, and connected transportation systems. However, it is very challenging to have very low latency and203
reliable delivery of data over a large scale network without increasing the network infrastructure cost, as it requires204
the development of techniques providing fast connections, quick handovers, and high data transfer rate.205
Proposed solutions: Caching methods [29, 112], VLC, mmWave, mMIMO (Section 6), fast handover206
techniques [40, 93, 102] (Section 5.2), and D2D communication (Section 4.3).207
Network performance optimization. The performance parameters, e.g., peak data rate, geographical area208
coverage, spectral efficiency, QoS, QoE, ease of connectivity, energy-efficiency, latency, reliability, fairness of209
users, and implementation complexity, are crucial for a cellular network [107]. Hence, a general framework for 5G210
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Methodologies/Technologies
Section
Increase data
rate
Increase
network
capacity
Massive
device support
Energy-efficient
Low latency
Economic
Security and
privacy
Interference
Mobility
support
Small-cells 4.1 X X X X P 2
Mobile small-cells 4.1 X X X X P X
CRN 4.2 X X
D2D 4.3 X P X P X X P X P
C-RANs 4.4 X X X X X X X
Full duplex radio 3,5.1 X X
Advance receiver 5.1 X X
SIC 5.1,6 X X P X X P
DUD 6 X P X
mmWave 6 X X X X X
mMIMO 6 X X X X X X
VLC 6 X X X X X
CCN-based caching 6 X
Table 2: Summary of methodologies and technologies for 5G networks.
networks should substantially optimize these parameters. However, there are some tradeoffs among all parameters,211
which further emphasize the need of a joint optimization algorithm.212
Economical impacts. A revolutionary change in the future mobile communication techniques would have drastic213
economical impacts in terms of deployment and motivation for user participation. It is critical to provide an entirely214
new infrastructure due to economical stretch. Therefore, the cost of deployment, maintenance, management, and215
operation of an infrastructure must be affordable from the perspective of governments, regulating authorities, and216
network operators. Also, the cost of using D2D communication should be feasible, so that devices involved in D2D217
communication should not charge more than using the services of a BS [15, 46]. Further, the projected revenue218
growth is much lower than the traffic growth [14]; hence, it is required to develop 5G networks in a manner that219
both network operators and users get honey in their hands.220
High mobility and handoff. The 5G wireless UEs are meant for retaining an active service connection while221
frequently moving from one cell to another or from one RAT (e.g., 3G, 4G, 5G, WiFi, Bluetooth, and WLAN) to222
another. The mobility adaptation for the wireless services should not back-off even at a very high speed as a UE223
inside a moving vehicle. Moreover, during a particular interval, many UEs move from one place to another; for224
example, moving to offices from residential areas in the morning. As a result, 5G networks are envisioned to use225
the spectrum in the best manner and to cope up with pace of the device movement.226
Proposed solutions: Inter-tier, intra-tier, and multi-RATs handoff mechanisms, and a mechanism for secure227
handoff [40, 93, 102, 52], which we will discuss in Section 5.2.228
Self-healing infrastructures. A self-healing infrastructure finds a failed macrocell or small-cell (i.e., a cell that is229
unable to work because of hardware failures, software failures, or misconfigurations) with the help of neighboring230
cells and provides a way for communication to the affected users by adjusting the transmission power and231
operating channels in the neighboring cells [41, 111]. The design of a self-healing network insists on the frequent232
communication among cells; hence, it brings in the following challenges, as: (i) develop an efficient algorithm233
that can detect and reconfigure a failed cell with insignificant communication and computational overheads in the234
minimal detection time, and (ii) reconfiguration of a failed cell should not lead to degradation of nearby cells’235
services.236
Proposed solutions: A small-cell network with self-healing property is suggested in [111], which we will237
discuss in Section 4.1.2.238
QoS. QoS guarantee in 5G networks has inherent difficulties, e.g., node mobility, multi-hop communication,239
resource allocation, and lack of central coordination. In addition, in 5G networks, a huge amount of bursty and240
2X P : Partial support
6
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Methodologies/Technologies
Section
Increase data
rate
Increase
network
capacity
Massive
device support
Energy-efficient
Low latency
Economic
Security and
privacy
Interference
Mobility
support
Small-cells 4.1 X X X X P 2
Mobile small-cells 4.1 X X X X P X
CRN 4.2 X X
D2D 4.3 X P X P X X P X P
C-RANs 4.4 X X X X X X X
Full duplex radio 3,5.1 X X
Advance receiver 5.1 X X
SIC 5.1,6 X X P X X P
DUD 6 X P X
mmWave 6 X X X X X
mMIMO 6 X X X X X X
VLC 6 X X X X X
CCN-based caching 6 X
Table 2: Summary of methodologies and technologies for 5G networks.
networks should substantially optimize these parameters. However, there are some tradeoffs among all parameters,211
which further emphasize the need of a joint optimization algorithm.212
Economical impacts. A revolutionary change in the future mobile communication techniques would have drastic213
economical impacts in terms of deployment and motivation for user participation. It is critical to provide an entirely214
new infrastructure due to economical stretch. Therefore, the cost of deployment, maintenance, management, and215
operation of an infrastructure must be affordable from the perspective of governments, regulating authorities, and216
network operators. Also, the cost of using D2D communication should be feasible, so that devices involved in D2D217
communication should not charge more than using the services of a BS [15, 46]. Further, the projected revenue218
growth is much lower than the traffic growth [14]; hence, it is required to develop 5G networks in a manner that219
both network operators and users get honey in their hands.220
High mobility and handoff. The 5G wireless UEs are meant for retaining an active service connection while221
frequently moving from one cell to another or from one RAT (e.g., 3G, 4G, 5G, WiFi, Bluetooth, and WLAN) to222
another. The mobility adaptation for the wireless services should not back-off even at a very high speed as a UE223
inside a moving vehicle. Moreover, during a particular interval, many UEs move from one place to another; for224
example, moving to offices from residential areas in the morning. As a result, 5G networks are envisioned to use225
the spectrum in the best manner and to cope up with pace of the device movement.226
Proposed solutions: Inter-tier, intra-tier, and multi-RATs handoff mechanisms, and a mechanism for secure227
handoff [40, 93, 102, 52], which we will discuss in Section 5.2.228
Self-healing infrastructures. A self-healing infrastructure finds a failed macrocell or small-cell (i.e., a cell that is229
unable to work because of hardware failures, software failures, or misconfigurations) with the help of neighboring230
cells and provides a way for communication to the affected users by adjusting the transmission power and231
operating channels in the neighboring cells [41, 111]. The design of a self-healing network insists on the frequent232
communication among cells; hence, it brings in the following challenges, as: (i) develop an efficient algorithm233
that can detect and reconfigure a failed cell with insignificant communication and computational overheads in the234
minimal detection time, and (ii) reconfiguration of a failed cell should not lead to degradation of nearby cells’235
services.236
Proposed solutions: A small-cell network with self-healing property is suggested in [111], which we will237
discuss in Section 4.1.2.238
QoS. QoS guarantee in 5G networks has inherent difficulties, e.g., node mobility, multi-hop communication,239
resource allocation, and lack of central coordination. In addition, in 5G networks, a huge amount of bursty and240
2X P : Partial support
6
Paraphrase This Document
Need a fresh take? Get an instant paraphrase of this document with our AI Paraphraser
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
multimedia data, multi-RATs, and low latency bound for different applications and services are major hurdles in241
achieving the desired QoS. Hence, it is challenging to design fast and efficient algorithms to maintain real-time242
QoS without overloading a BS [123, 120].243
Proposed solutions: Delay-bound QoS [62, 120], intelligent equipment [123], and multi-link with multi-flow244
and multi-QoS [69] have been suggested, which we will discuss in Section 5.3.245
Security and privacy of the network and UEs. The promising features of 5G networks bring in hard challenges246
in the design of security and privacy oriented 5G networks. For example, a huge number of new types of social247
(all-time connected) devices may originate several types of attacks like impersonation, denial-of-services (DoS),248
replay, eavesdropping, man-in-the-middle, and repudiation attacks. Also, the transfer of a huge volume of data in249
secure and high speed manners is critical while preventing malicious files to penetrate. In addition, the network250
densification needs to be secure and requires fast-secure handoff of UEs. We further highlight challenges in security251
and privacy of the network and UEs in Section 5.6.252
Proposed solutions: Physical layer security [118], monitoring [105, 48, 76], secret adaptive frequency253
hopping [76], encrypted- [79], and policy-based communications [67], which we will discuss in Section 5.6.254
All the above mentioned methodologies and technologies are comparatively studied in Table 2.255
4 Architectures for the Future/5G Mobile Cellular Networks256
In this section, we elaborate on the existing architectures for 5G networks, namely multi-tier, CRN-based, D2D257
communication based, and the cloud-based architectures. These proposed 5G architectures will be explained in the258
light of relevant advantages, disadvantages, and the challenges that are yet to be resolved.259
4.1 Two-tier Architectures260
A mobile small-cell
on a train
DR-OC
DC-OC
Relay device
Destination
Source
DR-DC
DC-DC
A SBS in
a home
CRN
The core
networkA mobile-small-
cell on a bus
A SBS on a factory
Car
communication
MBS
Figure 3: A multi-tier architecture for 5G networks with small-cells, mobile small-cells,
and D2D- and CRN-based communications.
Several two-tier261
architectures have262
been proposed for 5G263
networks, where a MBS264
stays in the top-tier265
and SBSs work under266
the supervision of the267
MBS in the lower tier.268
A macrocell covers269
all the small-cells of270
different types, e.g.,271
femtocell, picocell,272
and microcell (see273
Table 3), and both the274
tiers share an identical275
frequency band. The276
small-cell enhances the277
coverage and services278
of a macrocell, and the advantages of small-cells are mentioned at the end of this section. In addition, D2D279
communication and CRN-based communication enhance a 2-tier architecture to a multi-tier architecture; see280
Figure 3. Note that in this section, we confine ourselves on the deployment of small-cells under the cover of a281
macrocell; the discussion of CRN-based and D2D communications will be carried out in Sections 4.2 and 4.3,282
respectively.283
Cells Range Users
Femtocell 10-20 meters A few users
Picocell 200 meters 20 40
Microcell 2 kilometers > 100
Macrocell 30-35 kilometers Many
Table 3: Classification of the cells.
Wang et al. [107] suggested a way for separating indoor and284
outdoor users and using a mobile small-cell on a train or a bus.285
For separating indoor and outdoor users, a MBS holds large286
antenna arrays with some antenna elements distributed around the287
macrocell and connected to the MBS using optical fibers. A288
SBS and large antenna arrays are deployed in each building for289
7
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
multimedia data, multi-RATs, and low latency bound for different applications and services are major hurdles in241
achieving the desired QoS. Hence, it is challenging to design fast and efficient algorithms to maintain real-time242
QoS without overloading a BS [123, 120].243
Proposed solutions: Delay-bound QoS [62, 120], intelligent equipment [123], and multi-link with multi-flow244
and multi-QoS [69] have been suggested, which we will discuss in Section 5.3.245
Security and privacy of the network and UEs. The promising features of 5G networks bring in hard challenges246
in the design of security and privacy oriented 5G networks. For example, a huge number of new types of social247
(all-time connected) devices may originate several types of attacks like impersonation, denial-of-services (DoS),248
replay, eavesdropping, man-in-the-middle, and repudiation attacks. Also, the transfer of a huge volume of data in249
secure and high speed manners is critical while preventing malicious files to penetrate. In addition, the network250
densification needs to be secure and requires fast-secure handoff of UEs. We further highlight challenges in security251
and privacy of the network and UEs in Section 5.6.252
Proposed solutions: Physical layer security [118], monitoring [105, 48, 76], secret adaptive frequency253
hopping [76], encrypted- [79], and policy-based communications [67], which we will discuss in Section 5.6.254
All the above mentioned methodologies and technologies are comparatively studied in Table 2.255
4 Architectures for the Future/5G Mobile Cellular Networks256
In this section, we elaborate on the existing architectures for 5G networks, namely multi-tier, CRN-based, D2D257
communication based, and the cloud-based architectures. These proposed 5G architectures will be explained in the258
light of relevant advantages, disadvantages, and the challenges that are yet to be resolved.259
4.1 Two-tier Architectures260
A mobile small-cell
on a train
DR-OC
DC-OC
Relay device
Destination
Source
DR-DC
DC-DC
A SBS in
a home
CRN
The core
networkA mobile-small-
cell on a bus
A SBS on a factory
Car
communication
MBS
Figure 3: A multi-tier architecture for 5G networks with small-cells, mobile small-cells,
and D2D- and CRN-based communications.
Several two-tier261
architectures have262
been proposed for 5G263
networks, where a MBS264
stays in the top-tier265
and SBSs work under266
the supervision of the267
MBS in the lower tier.268
A macrocell covers269
all the small-cells of270
different types, e.g.,271
femtocell, picocell,272
and microcell (see273
Table 3), and both the274
tiers share an identical275
frequency band. The276
small-cell enhances the277
coverage and services278
of a macrocell, and the advantages of small-cells are mentioned at the end of this section. In addition, D2D279
communication and CRN-based communication enhance a 2-tier architecture to a multi-tier architecture; see280
Figure 3. Note that in this section, we confine ourselves on the deployment of small-cells under the cover of a281
macrocell; the discussion of CRN-based and D2D communications will be carried out in Sections 4.2 and 4.3,282
respectively.283
Cells Range Users
Femtocell 10-20 meters A few users
Picocell 200 meters 20 40
Microcell 2 kilometers > 100
Macrocell 30-35 kilometers Many
Table 3: Classification of the cells.
Wang et al. [107] suggested a way for separating indoor and284
outdoor users and using a mobile small-cell on a train or a bus.285
For separating indoor and outdoor users, a MBS holds large286
antenna arrays with some antenna elements distributed around the287
macrocell and connected to the MBS using optical fibers. A288
SBS and large antenna arrays are deployed in each building for289
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
communicating with the MBS. All UEs inside a building can have a connection to another UE either through the290
SBS or by using WiFi, mmWave, or VLC. Thus, the separation of users results in less load on a MBS.291
Wang et al. [107] also suggested to use a mobile small-cell that is located inside a vehicle to allow292
communication among internal UEs, while large antenna arrays are located outside the vehicle to communicate293
with a MBS. Thus, all the UEs inside a vehicle (or a building) appear to be a single unit with respect to the294
corresponding MBS, and clearly, the SBS appears as a MBS to all these UEs.295
In [28], a two-tier architecture is deployed as a process of network densification that is a combination of spatial296
densification (increasing the number of antennas per UE and MBS, and increasing the density of BSs) and spectral297
aggregation (using higher frequency bands > 3 Ghz). A tradeoff between the transmission power of a macrocell298
and the coverage area of small-cells is presented in [28]; i.e., on the one hand, if the transmission power of a299
macrocell is high, then many adjacent UEs to a small-cell may find themselves in the service area of the macrocell,300
and hence, it will decrease the coverage area of that small-cell. On the other hand, if the transmission power of301
a macrocell is low, then the coverage area of the small-cell will increase. Therefore, cell range expansion (i.e.,302
a biased handoff in the favor of small-cells) is carried out to serve more UEs by small-cells to which they are303
closer. Moreover, SBSs, deployed in offices or homes, can be used to serve outdoor users, e.g., pedestrians and304
low-mobility vehicles, in their neighborhoods, and the approach is called indoor-to-outdoor user service [28].305
Hossain et al. [61] presented a multi-tier architecture consisting of several types of small-cells, relays, and306
D2D communication for serving users with different QoS requirements in spectrum-efficient and energy-efficient307
manners. Interestingly, all of these architectures consider that UEs spontaneously discover a SBS. Zhang et308
al. [121] proposed a centralized system in which a MBS assists UEs to have connections to particular SBSs, thereby309
interference between UEs and SBSs is reduced. However, this approach overburdens the MBS.310
Advantage of the deployment of small-cells.311
• High data rate and efficient spectrum use: The small physical separation between a SBS and UEs (served by312
the same SBS) leads to a higher data rate and a better indoor coverage. Also, the spectrum efficiency increases313
due to fewer UEs in direct communication with a MBS [111].314
• Energy saving: The use of small-cells reduces the energy consumption of the network (by not involving MBSs)315
and of UEs (by allowing UEs to communicate at a shorter range with low signaling overhead) [51].316
• Money saving: It is more economical to install a SBS without any cumbersome planning as compared to a MBS,317
and also the operational-management cost is much lower than the cost associated with a MBS [41, 28].318
• The plug-and-play utility of small-cells boosts the on-demand network capacity [123].319
• Less congestion to a MBS: SBSs offload UEs from a MBS so that the MBS is lightly loaded and less congested,320
and hence, improve the system capacity [41].321
• Easy handoff : Mobile small-cells also follow the advantages of small-cells. Moreover, they provide an attractive322
solution to highly mobile UEs by reducing handoff time overheads, since a mobile small-cell is capable to do323
the handoff on behalf of all related UEs [28].324
Disadvantage of small-cells. Despite numerous prominent benefits as mentioned above, there are a few realistic325
issues such as implementation cost and operational reliability. The small-cells indeed impose an initial cost to the326
infrastructure, but less than the cost associated with a MBS. Moreover, a frequent authentication is mandatory due327
to frequent handoff operations. Further, an active or passive (on/off) state update of any small-cell would definitely328
result in frequent topological updates.329
Open issues in the deployment of 2-tier architectures using small-cells.330
• Interference management: The deployment of small-cells results in several types of interferences, as: inter-tier331
interference (i.e., interference from a MBS to a SBS, interference from a MBS to a SBS’s UEs, and interference332
from a SBS to a MBS’s UEs), and intra-tier interference (i.e., interference from a SBS to other SBSs’ UEs).333
Hence, it is also required to develop models and algorithms to handle these interferences [41, 38].334
• Backhaul data transfer: Though we have models to transfer data from a SBS to the core network, which we335
will discuss next in Section 4.1.1, an extremely dense-deployment of small-cells requires a huge amount of data336
transfer, and certainly, requires cost efficient architectures.337
4.1.1 Backhaul data transfer from small-cells338
Data transfer from a SBS to the core network is a challenging task, and in general, there may be three approaches339
to transfer (backhaul) data to the core network, as follows:340
8
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
communicating with the MBS. All UEs inside a building can have a connection to another UE either through the290
SBS or by using WiFi, mmWave, or VLC. Thus, the separation of users results in less load on a MBS.291
Wang et al. [107] also suggested to use a mobile small-cell that is located inside a vehicle to allow292
communication among internal UEs, while large antenna arrays are located outside the vehicle to communicate293
with a MBS. Thus, all the UEs inside a vehicle (or a building) appear to be a single unit with respect to the294
corresponding MBS, and clearly, the SBS appears as a MBS to all these UEs.295
In [28], a two-tier architecture is deployed as a process of network densification that is a combination of spatial296
densification (increasing the number of antennas per UE and MBS, and increasing the density of BSs) and spectral297
aggregation (using higher frequency bands > 3 Ghz). A tradeoff between the transmission power of a macrocell298
and the coverage area of small-cells is presented in [28]; i.e., on the one hand, if the transmission power of a299
macrocell is high, then many adjacent UEs to a small-cell may find themselves in the service area of the macrocell,300
and hence, it will decrease the coverage area of that small-cell. On the other hand, if the transmission power of301
a macrocell is low, then the coverage area of the small-cell will increase. Therefore, cell range expansion (i.e.,302
a biased handoff in the favor of small-cells) is carried out to serve more UEs by small-cells to which they are303
closer. Moreover, SBSs, deployed in offices or homes, can be used to serve outdoor users, e.g., pedestrians and304
low-mobility vehicles, in their neighborhoods, and the approach is called indoor-to-outdoor user service [28].305
Hossain et al. [61] presented a multi-tier architecture consisting of several types of small-cells, relays, and306
D2D communication for serving users with different QoS requirements in spectrum-efficient and energy-efficient307
manners. Interestingly, all of these architectures consider that UEs spontaneously discover a SBS. Zhang et308
al. [121] proposed a centralized system in which a MBS assists UEs to have connections to particular SBSs, thereby309
interference between UEs and SBSs is reduced. However, this approach overburdens the MBS.310
Advantage of the deployment of small-cells.311
• High data rate and efficient spectrum use: The small physical separation between a SBS and UEs (served by312
the same SBS) leads to a higher data rate and a better indoor coverage. Also, the spectrum efficiency increases313
due to fewer UEs in direct communication with a MBS [111].314
• Energy saving: The use of small-cells reduces the energy consumption of the network (by not involving MBSs)315
and of UEs (by allowing UEs to communicate at a shorter range with low signaling overhead) [51].316
• Money saving: It is more economical to install a SBS without any cumbersome planning as compared to a MBS,317
and also the operational-management cost is much lower than the cost associated with a MBS [41, 28].318
• The plug-and-play utility of small-cells boosts the on-demand network capacity [123].319
• Less congestion to a MBS: SBSs offload UEs from a MBS so that the MBS is lightly loaded and less congested,320
and hence, improve the system capacity [41].321
• Easy handoff : Mobile small-cells also follow the advantages of small-cells. Moreover, they provide an attractive322
solution to highly mobile UEs by reducing handoff time overheads, since a mobile small-cell is capable to do323
the handoff on behalf of all related UEs [28].324
Disadvantage of small-cells. Despite numerous prominent benefits as mentioned above, there are a few realistic325
issues such as implementation cost and operational reliability. The small-cells indeed impose an initial cost to the326
infrastructure, but less than the cost associated with a MBS. Moreover, a frequent authentication is mandatory due327
to frequent handoff operations. Further, an active or passive (on/off) state update of any small-cell would definitely328
result in frequent topological updates.329
Open issues in the deployment of 2-tier architectures using small-cells.330
• Interference management: The deployment of small-cells results in several types of interferences, as: inter-tier331
interference (i.e., interference from a MBS to a SBS, interference from a MBS to a SBS’s UEs, and interference332
from a SBS to a MBS’s UEs), and intra-tier interference (i.e., interference from a SBS to other SBSs’ UEs).333
Hence, it is also required to develop models and algorithms to handle these interferences [41, 38].334
• Backhaul data transfer: Though we have models to transfer data from a SBS to the core network, which we335
will discuss next in Section 4.1.1, an extremely dense-deployment of small-cells requires a huge amount of data336
transfer, and certainly, requires cost efficient architectures.337
4.1.1 Backhaul data transfer from small-cells338
Data transfer from a SBS to the core network is a challenging task, and in general, there may be three approaches339
to transfer (backhaul) data to the core network, as follows:340
8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
• Wired optical fiber: by establishing a wired optical fiber link from each SBS to a MBS; however, it is341
time-consuming and expensive.342
• Wireless point-to-multipoint (PTMP): by deploying a PTMP-BS at a MBS that communicates with SBSs and343
transfers data to the core network.344
• Wireless point-to-point (PTP): by using directional antennas in line-of-sight (LOS) environments; hence, it345
provides high capacity PTP links (same as with wired optical fibers), at a significantly lower cost.346
Ge et al. [51] presented two architectures based on the wireless PTMP approach. In the first (centralized)347
architecture, all SBSs send data using mmWave to a MBS that eventually aggregates the received data and forwards348
the same to the core network using fiber. In the second (distributed) architecture, all small-cells cooperatively349
forward data using mmWave to a specified SBS that transfers data to the core network using fiber without the350
explicit involvement of a MBS.351
Ni et al. [88] proposed an adaptive backhaul architecture based on the wireless PTP approach and frequency352
division duplex for UL and DL channels. A tree structure is used, where the root node is connected to the core353
network using fiber, the leaf nodes represent UEs, and other nodes represent SBSs. The data is transferred from the354
leaf nodes to the root node that transfers the same to the core network. The bandwidth is selected dynamically for355
backhaul links, as per the current bandwidth requirements, interference conditions, and the network topology. A356
similar approach is also presented in [28].357
4.1.2 Two-tier architectures with self-healing property358
An automatic detection and recovery of a failed cell is an important issue in densely deployed multi-tier359
architectures. Wang and Zhang [111] provided three approaches for designing a self-healing architecture such360
as below:361
1. Centralized approach: A dedicated server is responsible for detecting a failed cell by measuring and analyzing362
abnormal behavior of users, e.g., received signal strengths (RSSs) at users and handoff by several users at any363
time from a particular cell. The server collects global information and reconfigures the failed cell. However, the364
approach suffers with a high communication overhead and a high computational cost.365
2. Distributed approach: Each SBS detects failed small-cells in neighborhoods by measuring and analyzing users’366
handoff behavior and the neighboring small-cells’ signals. Consequently, on detecting a failed cell, a SBS might367
increase the transmission power in order to incorporate users of the failed cell. However, the approach might368
not work efficiently in case that users are scattered sparsely.369
3. Local cooperative or hybrid approach: This approach combines the benefits of both the previous approaches,370
and therefore, minimizes the drawback. Essentially, two steps are utilized, namely distributed trigger and371
cooperative detection. In the distributed trigger, each SBS collects information about users’ behavior.372
Subsequently, a trigger message is sent to a dedicated server in case the received information thrives below373
a certain threshold. Hence, it does not require communication among small-cells. In the cooperative detection,374
the dedicated server takes the final decision based on the information received from several small-cells, resulting375
in a high accuracy and lower latency.376
4.2 Cognitive Radio Network based Architectures377
A cognitive radio network (CRN) [16] is a collection of cognitive radio nodes (or processors), called secondary378
users (SUs) that exploit the existing spectrum opportunistically. The SUs have the LEIRA (learning, efficiency,379
intelligence, reliability, and adaptively) property for scanning and operating on multiple heterogeneous channels380
(or frequency bands) in the absence of the licensed user(s), termed as primary user(s) (PUs), of the respective381
bands [98]. Each PU has a fixed bandwidth, high transmit power, and high reliability; however, the SUs work on a382
broad range of bandwidth with low transmit power and low reliability.383
A CRN in 5G networks is used for designing multi-tier architectures, removing interference among cells, and384
minimizing energy consumption in the network [63, 41, 110, 60, 72, 83].385
4.2.1 CRN-based architectures for 5G networks386
A CRN creates a 2-tier architecture, similar to architectures discussed in Section 4.1; however, it is assumed that387
either a MBS or a SBS has cognitive properties for working on different channels.388
9
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
• Wired optical fiber: by establishing a wired optical fiber link from each SBS to a MBS; however, it is341
time-consuming and expensive.342
• Wireless point-to-multipoint (PTMP): by deploying a PTMP-BS at a MBS that communicates with SBSs and343
transfers data to the core network.344
• Wireless point-to-point (PTP): by using directional antennas in line-of-sight (LOS) environments; hence, it345
provides high capacity PTP links (same as with wired optical fibers), at a significantly lower cost.346
Ge et al. [51] presented two architectures based on the wireless PTMP approach. In the first (centralized)347
architecture, all SBSs send data using mmWave to a MBS that eventually aggregates the received data and forwards348
the same to the core network using fiber. In the second (distributed) architecture, all small-cells cooperatively349
forward data using mmWave to a specified SBS that transfers data to the core network using fiber without the350
explicit involvement of a MBS.351
Ni et al. [88] proposed an adaptive backhaul architecture based on the wireless PTP approach and frequency352
division duplex for UL and DL channels. A tree structure is used, where the root node is connected to the core353
network using fiber, the leaf nodes represent UEs, and other nodes represent SBSs. The data is transferred from the354
leaf nodes to the root node that transfers the same to the core network. The bandwidth is selected dynamically for355
backhaul links, as per the current bandwidth requirements, interference conditions, and the network topology. A356
similar approach is also presented in [28].357
4.1.2 Two-tier architectures with self-healing property358
An automatic detection and recovery of a failed cell is an important issue in densely deployed multi-tier359
architectures. Wang and Zhang [111] provided three approaches for designing a self-healing architecture such360
as below:361
1. Centralized approach: A dedicated server is responsible for detecting a failed cell by measuring and analyzing362
abnormal behavior of users, e.g., received signal strengths (RSSs) at users and handoff by several users at any363
time from a particular cell. The server collects global information and reconfigures the failed cell. However, the364
approach suffers with a high communication overhead and a high computational cost.365
2. Distributed approach: Each SBS detects failed small-cells in neighborhoods by measuring and analyzing users’366
handoff behavior and the neighboring small-cells’ signals. Consequently, on detecting a failed cell, a SBS might367
increase the transmission power in order to incorporate users of the failed cell. However, the approach might368
not work efficiently in case that users are scattered sparsely.369
3. Local cooperative or hybrid approach: This approach combines the benefits of both the previous approaches,370
and therefore, minimizes the drawback. Essentially, two steps are utilized, namely distributed trigger and371
cooperative detection. In the distributed trigger, each SBS collects information about users’ behavior.372
Subsequently, a trigger message is sent to a dedicated server in case the received information thrives below373
a certain threshold. Hence, it does not require communication among small-cells. In the cooperative detection,374
the dedicated server takes the final decision based on the information received from several small-cells, resulting375
in a high accuracy and lower latency.376
4.2 Cognitive Radio Network based Architectures377
A cognitive radio network (CRN) [16] is a collection of cognitive radio nodes (or processors), called secondary378
users (SUs) that exploit the existing spectrum opportunistically. The SUs have the LEIRA (learning, efficiency,379
intelligence, reliability, and adaptively) property for scanning and operating on multiple heterogeneous channels380
(or frequency bands) in the absence of the licensed user(s), termed as primary user(s) (PUs), of the respective381
bands [98]. Each PU has a fixed bandwidth, high transmit power, and high reliability; however, the SUs work on a382
broad range of bandwidth with low transmit power and low reliability.383
A CRN in 5G networks is used for designing multi-tier architectures, removing interference among cells, and384
minimizing energy consumption in the network [63, 41, 110, 60, 72, 83].385
4.2.1 CRN-based architectures for 5G networks386
A CRN creates a 2-tier architecture, similar to architectures discussed in Section 4.1; however, it is assumed that387
either a MBS or a SBS has cognitive properties for working on different channels.388
9
Secure Best Marks with AI Grader
Need help grading? Try our AI Grader for instant feedback on your assignments.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Hong et al. [60] presented two types of CRN-based architectures for 5G networks, as: (i) non-cooperative and389
(ii) cooperative CRNs. The non-cooperative CRN establishes a multi-RATs system, having two separate radio390
interfaces that operate at the licensed and temporary unoccupied channels by PUs, called cognitive channels.391
The SUs work only on cognitive channels and form a CRN, which overlays on the existing licensed cellular392
network. The two networks can be integrated in the upper layers while must be separated in the physical layer.393
This architecture can be used in different manners, as: (i) the cognitive and licensed channels are used by users394
near a MBS and users far away from the MBS, respectively, (ii) the cognitive and licensed channels are used for395
relaxed QoS and strict QoS, respectively. The cooperative CRN uses only a licensed channel, where SUs access the396
channel in an opportunistic fashion when the PU of the channel is absent. This architecture can be used in different397
manners, as: (i) a SBS communicates with a MBS using the licensed channel and provides service to its UEs via398
an opportunistic access to the licensed frequency band, (ii) a licensed channel is used to serve UEs by a SBS and399
the opportunistic access to the licensed channel is used to transfer backhaul data to the MBS.400
In short, the cooperative CRN [60] provides a real intuition of incorporating CRNs in 5G networks, where a401
SBS works as a SU, which scans activities of a macrocell and works on temporarily unoccupied frequency bands402
(known as spectrum holes [17]) by a PU to provide services to their UEs with minimally disrupting macrocell403
activities.404
A dynamic pricing model based on a game theoretic framework for cognitive small-cells is suggested in [65].405
Since in reality SBSs’ operators and MBSs’ operators may not be identical and small-cells’ UEs may achieve a406
higher data rate as compared to macrocells’ UEs, the pricing model for both UEs must be different.407
4.2.2 Interference Management using CRNs408
Huang et al. [63] provided an approach for avoiding inter-tier interference by integrating a cognitive technique409
at a SBS. The cognitive technique consists of three components, as: (i) a cognitive module, which senses the410
environment and collects information about spectrum holes, collision probability, QoS requirements, macrocell411
activities, and channel gains, (ii) a cognitive engine, which analyzes and stores the collected information for412
estimating available resources, and (iii) a self-configuration module, which uses the stored information for413
dynamically optimizing several parameters for efficient handoff, interference, and power management. Further,414
the channel allocation to a small-cell is done in a manner to avoid inter-tier and intra-tier interferences, based415
on Gale-Shapley [49] spectrum sharing scheme, which avoids collisions by not assigning an identical channel to416
neighboring small-cells.417
Wang et al. [110] suggested an approach for mitigating inter-tier interference based on spectrum sensing,418
spectrum sharing, and cognitive relay, where links between a MBS and its UEs are considered as PUs and links419
between a SBS and its UEs are considered as SUs. Cognitive techniques are used for detecting interference from420
a MBS to a SBS and vice versa, and a path loss estimation algorithm is provided for detecting interference from a421
small-cell’s UEs to a macrocell’s UEs. After detecting inter-tier interference, a small-cell shares spectrum with a422
macrocell using either overlay spectrum sharing scheme (i.e., SUs utilize unoccupied channels, and it is applicable423
when a MBS and a SBS’s UEs are very close or no interference is required by a macrocell’s UEs) or underlay424
spectrum sharing scheme (i.e., SUs and PUs transmit on an identical channel while restricting transmit power of425
SUs, and hence, resulting in a higher spectrum utilization).426
Note that a CRN can be used to support D2D communication and mitigate interferences caused by D2D427
communication, which we will see in Section 4.3.428
Advantages of CRNs in 5G networks.429
• Minimizing interference: By implementing a CRN at small-cells, cognitive small-cells can avoid interference430
very efficiently by not selecting identical channels as the channels of neighboring small-cells.431
• Increase network capacity: The spectrum holes can be exploited for supporting a higher data transfer rate and432
enhancing bandwidth utilization [83].433
Open issues. Usually, cellular networks are not energy-efficient as they consume energy in circuits, cooling434
systems, and also radiate in air. Hence, an energy-efficient deployment of a CRN in a cellular network is of435
utmost importance [60]. Further, there is a tradeoff between the spatial frequency reuse and the outage probability,436
which requires the selection of an efficient spectrum sensing algorithm [41].437
10
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Hong et al. [60] presented two types of CRN-based architectures for 5G networks, as: (i) non-cooperative and389
(ii) cooperative CRNs. The non-cooperative CRN establishes a multi-RATs system, having two separate radio390
interfaces that operate at the licensed and temporary unoccupied channels by PUs, called cognitive channels.391
The SUs work only on cognitive channels and form a CRN, which overlays on the existing licensed cellular392
network. The two networks can be integrated in the upper layers while must be separated in the physical layer.393
This architecture can be used in different manners, as: (i) the cognitive and licensed channels are used by users394
near a MBS and users far away from the MBS, respectively, (ii) the cognitive and licensed channels are used for395
relaxed QoS and strict QoS, respectively. The cooperative CRN uses only a licensed channel, where SUs access the396
channel in an opportunistic fashion when the PU of the channel is absent. This architecture can be used in different397
manners, as: (i) a SBS communicates with a MBS using the licensed channel and provides service to its UEs via398
an opportunistic access to the licensed frequency band, (ii) a licensed channel is used to serve UEs by a SBS and399
the opportunistic access to the licensed channel is used to transfer backhaul data to the MBS.400
In short, the cooperative CRN [60] provides a real intuition of incorporating CRNs in 5G networks, where a401
SBS works as a SU, which scans activities of a macrocell and works on temporarily unoccupied frequency bands402
(known as spectrum holes [17]) by a PU to provide services to their UEs with minimally disrupting macrocell403
activities.404
A dynamic pricing model based on a game theoretic framework for cognitive small-cells is suggested in [65].405
Since in reality SBSs’ operators and MBSs’ operators may not be identical and small-cells’ UEs may achieve a406
higher data rate as compared to macrocells’ UEs, the pricing model for both UEs must be different.407
4.2.2 Interference Management using CRNs408
Huang et al. [63] provided an approach for avoiding inter-tier interference by integrating a cognitive technique409
at a SBS. The cognitive technique consists of three components, as: (i) a cognitive module, which senses the410
environment and collects information about spectrum holes, collision probability, QoS requirements, macrocell411
activities, and channel gains, (ii) a cognitive engine, which analyzes and stores the collected information for412
estimating available resources, and (iii) a self-configuration module, which uses the stored information for413
dynamically optimizing several parameters for efficient handoff, interference, and power management. Further,414
the channel allocation to a small-cell is done in a manner to avoid inter-tier and intra-tier interferences, based415
on Gale-Shapley [49] spectrum sharing scheme, which avoids collisions by not assigning an identical channel to416
neighboring small-cells.417
Wang et al. [110] suggested an approach for mitigating inter-tier interference based on spectrum sensing,418
spectrum sharing, and cognitive relay, where links between a MBS and its UEs are considered as PUs and links419
between a SBS and its UEs are considered as SUs. Cognitive techniques are used for detecting interference from420
a MBS to a SBS and vice versa, and a path loss estimation algorithm is provided for detecting interference from a421
small-cell’s UEs to a macrocell’s UEs. After detecting inter-tier interference, a small-cell shares spectrum with a422
macrocell using either overlay spectrum sharing scheme (i.e., SUs utilize unoccupied channels, and it is applicable423
when a MBS and a SBS’s UEs are very close or no interference is required by a macrocell’s UEs) or underlay424
spectrum sharing scheme (i.e., SUs and PUs transmit on an identical channel while restricting transmit power of425
SUs, and hence, resulting in a higher spectrum utilization).426
Note that a CRN can be used to support D2D communication and mitigate interferences caused by D2D427
communication, which we will see in Section 4.3.428
Advantages of CRNs in 5G networks.429
• Minimizing interference: By implementing a CRN at small-cells, cognitive small-cells can avoid interference430
very efficiently by not selecting identical channels as the channels of neighboring small-cells.431
• Increase network capacity: The spectrum holes can be exploited for supporting a higher data transfer rate and432
enhancing bandwidth utilization [83].433
Open issues. Usually, cellular networks are not energy-efficient as they consume energy in circuits, cooling434
systems, and also radiate in air. Hence, an energy-efficient deployment of a CRN in a cellular network is of435
utmost importance [60]. Further, there is a tradeoff between the spatial frequency reuse and the outage probability,436
which requires the selection of an efficient spectrum sensing algorithm [41].437
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
4.3 Device-to-Device Communication Architectures438
Device-to-Device (D2D) communication allows close proximity UEs to communicate with each other on a licensed439
cellular bandwidth without involving a MBS or with a very controlled involvement of a MBS. The standards440
and frameworks for D2D communication are in an early stage of research. In this section, we will review D2D441
communication networks in short. For a detailed review of D2D communication, interested readers may refer442
to [24].443
Challenges in D2D Communication.444
• Interference management: UEs involved in D2D communication, say D-UEs, face (or create) interference from445
(or to) other UEs, or from (or to) a BS, based on the selection of a DL or UL channel, respectively. The following446
types of interferences are investigated in [113]:447
– When using a DL channel: (i) interference from BSs in the same cell, (ii) interference from other co-channel448
D-UEs in the same cell, and (iii) interference from BSs and co-channel D-UEs from other cells.449
– When using a UL channel: (i) interference from all co-channel C-UEs 3 in the same cell and other cells, and450
(ii) interference from all co-channel D-UEs in the same cell and other cells.451
Proposed solutions: A simple solution may exist by implementing CRNs in D2D communication, as: D-UEs452
are considered as SUs and C-UEs are considered as PUs that should not be interfered. Consequently, any453
mechanism of CRNs can be implemented in D2D communication for interference removal.454
• Resource allocation: When UEs involved in D2D communication, it is required to allocate a sufficient amount of455
resources, particularly bandwidth and channels. However, the allocation of optimum resources to D-UEs must456
be carried out in a fashion that C-UEs do not have interference from D-UEs, and D-UEs can also communicate457
and exchange data efficiently [113, 77].458
Proposed solutions: SARA [33], frame-by-frame and slot-by-slot channel allocation methods [74], and a459
social-aware channel allocation [77].460
• Delay-sensitive processing: Audio, video streaming, and online gaming, which are natural in close proximity461
UEs, require real-time and delay-sensitive processing. Hence, it is required to consider delay-sensitive and462
real-time processing in D2D communication [109].463
Proposed solutions: Solutions based on channel state information (CSI) and QoS are provided in [109].464
• Pricing: Sometimes a D-UE uses resources (e.g., battery and data storage) of other UEs for relaying information,465
where the other UE may charge for providing its resources. Hence, the design of a pricing model is needed,466
thereby a D-UE is not charged more money than that involved to communicate through a MBS.467
Proposed solutions: Some solutions based on game theory, auction theory, and bargaining are suggested468
in [104].469
D2D communication types. D2D communication can be done in the following four ways [104], as follows:470
1. Device relaying with operator controlled link establishment (DR-OC): A UE at the edge of a cell or in a poor471
coverage area can communicate with a MBS by relaying its information via other UEs, which are within the472
stronger coverage area and not at the edge; see Figure 3.473
2. Direct D2D communication with operator controlled link establishment (DC-OC): Source and destination UEs474
communicate directly with each other without involving a MBS, but they are assisted by the MBS for link475
establishment; see Figure 3.476
3. Device relaying with device controlled link establishment (DR-DC): Source and destination UEs communicate477
through a relay without involving a MBS, and they are also responsible for link establishment; see Figure 3.478
4. Direct D2D communication with device controlled link establishment (DC-DC): Source and destination UEs479
communicate directly with each other without involving a MBS, and they are also responsible for link480
establishment; see Figure 3.481
Note that DR-OC and DC-OC involve a MBS for resource allocation and call setup, and hence, prevent interference482
among devices to some extent.483
Two types of coding schemes (or communication types) are described in [113]: (i) two-way relay channel484
(TRC), where a source and a destination communicate through a relay, and (ii) multiple-access relay channel485
(MRC), where multiple sources communicate to a destination through a relay with direct links. Note that the486
workings of DR-OC and MRC, and DR-DC and TRC are identical. Two types of node discovery and D2D487
3A UE that is not involved in D2D communication and communicates to a MBS, we call it a cellular user equipment (C-UE) in this section.
11
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
4.3 Device-to-Device Communication Architectures438
Device-to-Device (D2D) communication allows close proximity UEs to communicate with each other on a licensed439
cellular bandwidth without involving a MBS or with a very controlled involvement of a MBS. The standards440
and frameworks for D2D communication are in an early stage of research. In this section, we will review D2D441
communication networks in short. For a detailed review of D2D communication, interested readers may refer442
to [24].443
Challenges in D2D Communication.444
• Interference management: UEs involved in D2D communication, say D-UEs, face (or create) interference from445
(or to) other UEs, or from (or to) a BS, based on the selection of a DL or UL channel, respectively. The following446
types of interferences are investigated in [113]:447
– When using a DL channel: (i) interference from BSs in the same cell, (ii) interference from other co-channel448
D-UEs in the same cell, and (iii) interference from BSs and co-channel D-UEs from other cells.449
– When using a UL channel: (i) interference from all co-channel C-UEs 3 in the same cell and other cells, and450
(ii) interference from all co-channel D-UEs in the same cell and other cells.451
Proposed solutions: A simple solution may exist by implementing CRNs in D2D communication, as: D-UEs452
are considered as SUs and C-UEs are considered as PUs that should not be interfered. Consequently, any453
mechanism of CRNs can be implemented in D2D communication for interference removal.454
• Resource allocation: When UEs involved in D2D communication, it is required to allocate a sufficient amount of455
resources, particularly bandwidth and channels. However, the allocation of optimum resources to D-UEs must456
be carried out in a fashion that C-UEs do not have interference from D-UEs, and D-UEs can also communicate457
and exchange data efficiently [113, 77].458
Proposed solutions: SARA [33], frame-by-frame and slot-by-slot channel allocation methods [74], and a459
social-aware channel allocation [77].460
• Delay-sensitive processing: Audio, video streaming, and online gaming, which are natural in close proximity461
UEs, require real-time and delay-sensitive processing. Hence, it is required to consider delay-sensitive and462
real-time processing in D2D communication [109].463
Proposed solutions: Solutions based on channel state information (CSI) and QoS are provided in [109].464
• Pricing: Sometimes a D-UE uses resources (e.g., battery and data storage) of other UEs for relaying information,465
where the other UE may charge for providing its resources. Hence, the design of a pricing model is needed,466
thereby a D-UE is not charged more money than that involved to communicate through a MBS.467
Proposed solutions: Some solutions based on game theory, auction theory, and bargaining are suggested468
in [104].469
D2D communication types. D2D communication can be done in the following four ways [104], as follows:470
1. Device relaying with operator controlled link establishment (DR-OC): A UE at the edge of a cell or in a poor471
coverage area can communicate with a MBS by relaying its information via other UEs, which are within the472
stronger coverage area and not at the edge; see Figure 3.473
2. Direct D2D communication with operator controlled link establishment (DC-OC): Source and destination UEs474
communicate directly with each other without involving a MBS, but they are assisted by the MBS for link475
establishment; see Figure 3.476
3. Device relaying with device controlled link establishment (DR-DC): Source and destination UEs communicate477
through a relay without involving a MBS, and they are also responsible for link establishment; see Figure 3.478
4. Direct D2D communication with device controlled link establishment (DC-DC): Source and destination UEs479
communicate directly with each other without involving a MBS, and they are also responsible for link480
establishment; see Figure 3.481
Note that DR-OC and DC-OC involve a MBS for resource allocation and call setup, and hence, prevent interference482
among devices to some extent.483
Two types of coding schemes (or communication types) are described in [113]: (i) two-way relay channel484
(TRC), where a source and a destination communicate through a relay, and (ii) multiple-access relay channel485
(MRC), where multiple sources communicate to a destination through a relay with direct links. Note that the486
workings of DR-OC and MRC, and DR-DC and TRC are identical. Two types of node discovery and D2D487
3A UE that is not involved in D2D communication and communicates to a MBS, we call it a cellular user equipment (C-UE) in this section.
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
communication methods are also studied in [77], namely network-controlled approach and ad hoc network approach488
that work in a similar manner as DC-OC and DC-DC, respectively.489
Resource allocation methods. Now, we will review an architecture and some methods for resource allocation in490
D2D communication.491
Social-Aware D2D Architecture: As D2D communication is very efficient for close proximity UEs, keeping492
this fact in view, Li et al. [77] suggested a social-aware D2D communication architecture based on the social493
networking. The architecture, see Figure 4, has four major components as:494
• Ties: They are similar to friend relations in a social media, and hence, may be used as a trust measurement495
between two UEs. Allocating more spectrum and energy resources to UEs with strong ties can increase the peer496
discovery ratio, avoid congestion, and improve spectral efficiency.497
• Community: It is similar to a group on Facebook and helps in allocating more resources to all the UEs in a498
community to decrease content duplication and increase the network throughput.499
• Centrality: It is similar to a node that has more communication links/friends in a social network. The concept of500
centrality in D2D communication reduces congestion and increases the network throughput by allocating more501
resources to a central node.502
• Bridges: They are similar to a connection between two communities. Hence, two devices forming a bridge can503
be allocated more resources as compared to other devices.504
Centrality
Tie
Community
Community
Figure 4: A D2D communication architecture based
on a social networking.
Channel Allocation Methods: Two cooperative channel505
allocation methods, frame-by-frame and slot-by-slot, are506
given in [74]. Consider three zones A, B , and C with some507
UEs such that A and B , B and C intersect, but A and C508
do not intersect, and B holds a UE that communicates with509
UEs of A and C. In the frame-by-frame channel allocation510
method, UEs of A and C do intra-zone communication at511
different frames, and the UEs of B also communicate at different frames. However, in the slot-by-slot channel512
allocation method, UEs of A and C do intra-zone communication at an identical time, and of course, UEs of B513
communicate at a different time. Both the methods improve the efficiency of frequency division multiplexing and514
increase the network throughput.515
Hoang et al. [58] provided an iterative algorithm for subcarrier 4 and power allocation such that the minimal516
individual link data rates and proportionate fairness among D2D links are obtained. A 2-phase service-aware517
resource allocation scheme, called SARA, is proposed [33]. In the first phase of SARA, resources are allocated518
on-demand to meet different service requirements of D-UEs, and in the second phase of SARA, the remaining519
resources are allocated to D-UEs such that the system throughput increases.520
Wang et al. [109] provided a delay-aware and dynamic power control mechanism that adapts the transmit power521
of D-UEs based on instantaneous values of CSI, and hence, finds the urgency of the data flow. The dynamic power522
control selects a power control policy so that the long-term average delay and the long-term average power cost of523
all the flows minimize.524
Advantages of D2D Communication. D2D communication results in link reliability among D-UEs, a higher data525
rate to D-UEs, instant communication, an easy way for peer-to-peer file sharing, local voice services, local video526
streaming, local online gaming, an improved spectral efficiency, decreased power consumption of D-UEs, and the527
traffic offload from a MBS.528
Open issues.529
• Security and privacy: In D2D communication, D-UEs may take helps from other UEs as relay nodes; hence,530
it is required to communicate and transfer data in secure and privacy-preserving manners. Consequently, the531
designing of energy-efficient and trust-making protocols is an open issue.532
• Network coding scheme: When D2D communication uses relay nodes, an efficient network coding scheme may533
be utilized for improving the throughput [113].534
• Multi-mode selection: In the current design of D2D communication, UEs can do either D2D communication or535
communication to a BS [77]; however, it is not efficient. Hence, there is a need to design a system that allows536
a UE to engage the two types communication modes (i.e., D2D communication and communication to a BS)537
simultaneously.538
4A subcarrier is a signal carried on a main radio signal, and hence, two signals are transmitted at an identical time.
12
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
communication methods are also studied in [77], namely network-controlled approach and ad hoc network approach488
that work in a similar manner as DC-OC and DC-DC, respectively.489
Resource allocation methods. Now, we will review an architecture and some methods for resource allocation in490
D2D communication.491
Social-Aware D2D Architecture: As D2D communication is very efficient for close proximity UEs, keeping492
this fact in view, Li et al. [77] suggested a social-aware D2D communication architecture based on the social493
networking. The architecture, see Figure 4, has four major components as:494
• Ties: They are similar to friend relations in a social media, and hence, may be used as a trust measurement495
between two UEs. Allocating more spectrum and energy resources to UEs with strong ties can increase the peer496
discovery ratio, avoid congestion, and improve spectral efficiency.497
• Community: It is similar to a group on Facebook and helps in allocating more resources to all the UEs in a498
community to decrease content duplication and increase the network throughput.499
• Centrality: It is similar to a node that has more communication links/friends in a social network. The concept of500
centrality in D2D communication reduces congestion and increases the network throughput by allocating more501
resources to a central node.502
• Bridges: They are similar to a connection between two communities. Hence, two devices forming a bridge can503
be allocated more resources as compared to other devices.504
Centrality
Tie
Community
Community
Figure 4: A D2D communication architecture based
on a social networking.
Channel Allocation Methods: Two cooperative channel505
allocation methods, frame-by-frame and slot-by-slot, are506
given in [74]. Consider three zones A, B , and C with some507
UEs such that A and B , B and C intersect, but A and C508
do not intersect, and B holds a UE that communicates with509
UEs of A and C. In the frame-by-frame channel allocation510
method, UEs of A and C do intra-zone communication at511
different frames, and the UEs of B also communicate at different frames. However, in the slot-by-slot channel512
allocation method, UEs of A and C do intra-zone communication at an identical time, and of course, UEs of B513
communicate at a different time. Both the methods improve the efficiency of frequency division multiplexing and514
increase the network throughput.515
Hoang et al. [58] provided an iterative algorithm for subcarrier 4 and power allocation such that the minimal516
individual link data rates and proportionate fairness among D2D links are obtained. A 2-phase service-aware517
resource allocation scheme, called SARA, is proposed [33]. In the first phase of SARA, resources are allocated518
on-demand to meet different service requirements of D-UEs, and in the second phase of SARA, the remaining519
resources are allocated to D-UEs such that the system throughput increases.520
Wang et al. [109] provided a delay-aware and dynamic power control mechanism that adapts the transmit power521
of D-UEs based on instantaneous values of CSI, and hence, finds the urgency of the data flow. The dynamic power522
control selects a power control policy so that the long-term average delay and the long-term average power cost of523
all the flows minimize.524
Advantages of D2D Communication. D2D communication results in link reliability among D-UEs, a higher data525
rate to D-UEs, instant communication, an easy way for peer-to-peer file sharing, local voice services, local video526
streaming, local online gaming, an improved spectral efficiency, decreased power consumption of D-UEs, and the527
traffic offload from a MBS.528
Open issues.529
• Security and privacy: In D2D communication, D-UEs may take helps from other UEs as relay nodes; hence,530
it is required to communicate and transfer data in secure and privacy-preserving manners. Consequently, the531
designing of energy-efficient and trust-making protocols is an open issue.532
• Network coding scheme: When D2D communication uses relay nodes, an efficient network coding scheme may533
be utilized for improving the throughput [113].534
• Multi-mode selection: In the current design of D2D communication, UEs can do either D2D communication or535
communication to a BS [77]; however, it is not efficient. Hence, there is a need to design a system that allows536
a UE to engage the two types communication modes (i.e., D2D communication and communication to a BS)537
simultaneously.538
4A subcarrier is a signal carried on a main radio signal, and hence, two signals are transmitted at an identical time.
12
Paraphrase This Document
Need a fresh take? Get an instant paraphrase of this document with our AI Paraphraser
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
4.4 Cloud-based Architectures539
Mobility management
RRH
Baseband unit
QoS management
Network management
Congestion control
Security and privacy
Handoff management
Channel access
Control
layer
Data layer
Very high speed connections
Figure 5: A basic cloud-based architecture for
5G networks.
Cloud computing [85] infrastructure provides on-demand, easy,540
and scalable access to a shared pool of configurable resources,541
without worrying about the management of resources. The542
inclusion of the cloud in the mobile cellular communication543
can provide its benefits to the communication system. In this544
section, we will see cloud-based architectures or cloud-based545
radio access networks (C-RANs) for 5G networks. A detailed546
review of C-RANs is given in [32].547
The main idea of a C-RAN. The first C-RAN is provided by548
China Mobile Research Institute [14]. The basic idea behind549
any C-RAN is to execute most of the functions of a MBS in550
the cloud, and hence, divide the functionality of a MBS into a551
control layer and a data layer; see Figure 5. The functions of552
the control and the data layers are executed in a cloud and in a553
MBS, respectively. Thus, a C-RAN provides a dynamic service554
allocation scheme for scaling the network without installing555
costly network devices.556
Specifically, a MBS has two main components, as: (i) a557
baseband unit (BBU, for implementing baseband processing using baseband processors), and (ii) a remote radio558
head (RRH, for performing radio functions). In most of the C-RANs, BBUs are placed in the cloud and RRHs stay559
in MBSs. Thus, a C-RAN provides an easily scalable and flexible architecture. We will see advantages of C-RANs560
at the end of this section.561
Challenges in the deployment of a C-RAN.562
• An efficient fronthaul data transfer technique: A flexible cloudification of the functions of a MBS comes at the563
cost of efficient fronthaul data transfer from RRHs to BBUs. The fast and efficient data transfer to the cloud has564
a proportionate impact on the performance of a C-RAN [97].565
• Real-time performance: Since C-RANs will be used instead of a MBS that provides all the services to users, it566
is required to transfer and process all the data in the cloud as fast as a MBS can do; otherwise, it is hard to find567
solutions to real-time problems using a C-RAN [97].568
• Reliability: The cloud provider does not ensure any guarantee of failure-free executions of their hardware and569
software. Thus, it is hard to simulate an error-free MBS using a C-RAN.570
• Security: The resources of the cloud are shared among several users and never be under the control of a single571
authority. Hence, a malicious user may easily access the control layer of a C-RAN, resulting in a more severe572
problem.573
• Manageability: It is clear that a non-secure C-RAN may be accessed by any cloud user, which poses an574
additional challenge in manageability of C-RANs. Further, the dynamic allocation of the cloud resources at575
a specific time interval is a critical issue; otherwise, a C-RAN may face additional latency [55].576
Now, we will see some C-RAN architectures in brief.577
2-layered C-RAN architectures. The authors [14] provided two C-RAN architectures based on the division of578
functionalities of a MBS, as: (i) full centralized C-RAN, where a BBU and all the other higher level functionalities579
of a MBS are located in the cloud while a RRH is only located in the MBS, and (ii) partially centralized C-RAN,580
where a RRH and some of the functionalities of a BBU are located in the MBS while all the remaining functions581
of the BBU and higher level functionalities of the MBS are located in the cloud. Thus, the authors [14] proposed582
the use of only two layers, namely a control layer and a data layer for implementing C-RANs, as follows:583
1. Data layer: It contains heterogeneous physical resources (e.g., radio interface equipment) and performs signal584
processing tasks (e.g., channel decoding, demultiplexing, and fast Fourier transformation).585
2. Control layer: It performs baseband processing and resource management (application delivery, QoS, real-time586
communication, seamless mobility, security, network management, regulation, and power control); see587
Figure 5.588
Rost et al. [97] introduced RAN-as-a-service (RANaaS) concept, having the control and the data layers.589
However, in RANaaS, a cloud provides flexible and on-demand RAN functionalities (such as network management,590
13
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
4.4 Cloud-based Architectures539
Mobility management
RRH
Baseband unit
QoS management
Network management
Congestion control
Security and privacy
Handoff management
Channel access
Control
layer
Data layer
Very high speed connections
Figure 5: A basic cloud-based architecture for
5G networks.
Cloud computing [85] infrastructure provides on-demand, easy,540
and scalable access to a shared pool of configurable resources,541
without worrying about the management of resources. The542
inclusion of the cloud in the mobile cellular communication543
can provide its benefits to the communication system. In this544
section, we will see cloud-based architectures or cloud-based545
radio access networks (C-RANs) for 5G networks. A detailed546
review of C-RANs is given in [32].547
The main idea of a C-RAN. The first C-RAN is provided by548
China Mobile Research Institute [14]. The basic idea behind549
any C-RAN is to execute most of the functions of a MBS in550
the cloud, and hence, divide the functionality of a MBS into a551
control layer and a data layer; see Figure 5. The functions of552
the control and the data layers are executed in a cloud and in a553
MBS, respectively. Thus, a C-RAN provides a dynamic service554
allocation scheme for scaling the network without installing555
costly network devices.556
Specifically, a MBS has two main components, as: (i) a557
baseband unit (BBU, for implementing baseband processing using baseband processors), and (ii) a remote radio558
head (RRH, for performing radio functions). In most of the C-RANs, BBUs are placed in the cloud and RRHs stay559
in MBSs. Thus, a C-RAN provides an easily scalable and flexible architecture. We will see advantages of C-RANs560
at the end of this section.561
Challenges in the deployment of a C-RAN.562
• An efficient fronthaul data transfer technique: A flexible cloudification of the functions of a MBS comes at the563
cost of efficient fronthaul data transfer from RRHs to BBUs. The fast and efficient data transfer to the cloud has564
a proportionate impact on the performance of a C-RAN [97].565
• Real-time performance: Since C-RANs will be used instead of a MBS that provides all the services to users, it566
is required to transfer and process all the data in the cloud as fast as a MBS can do; otherwise, it is hard to find567
solutions to real-time problems using a C-RAN [97].568
• Reliability: The cloud provider does not ensure any guarantee of failure-free executions of their hardware and569
software. Thus, it is hard to simulate an error-free MBS using a C-RAN.570
• Security: The resources of the cloud are shared among several users and never be under the control of a single571
authority. Hence, a malicious user may easily access the control layer of a C-RAN, resulting in a more severe572
problem.573
• Manageability: It is clear that a non-secure C-RAN may be accessed by any cloud user, which poses an574
additional challenge in manageability of C-RANs. Further, the dynamic allocation of the cloud resources at575
a specific time interval is a critical issue; otherwise, a C-RAN may face additional latency [55].576
Now, we will see some C-RAN architectures in brief.577
2-layered C-RAN architectures. The authors [14] provided two C-RAN architectures based on the division of578
functionalities of a MBS, as: (i) full centralized C-RAN, where a BBU and all the other higher level functionalities579
of a MBS are located in the cloud while a RRH is only located in the MBS, and (ii) partially centralized C-RAN,580
where a RRH and some of the functionalities of a BBU are located in the MBS while all the remaining functions581
of the BBU and higher level functionalities of the MBS are located in the cloud. Thus, the authors [14] proposed582
the use of only two layers, namely a control layer and a data layer for implementing C-RANs, as follows:583
1. Data layer: It contains heterogeneous physical resources (e.g., radio interface equipment) and performs signal584
processing tasks (e.g., channel decoding, demultiplexing, and fast Fourier transformation).585
2. Control layer: It performs baseband processing and resource management (application delivery, QoS, real-time586
communication, seamless mobility, security, network management, regulation, and power control); see587
Figure 5.588
Rost et al. [97] introduced RAN-as-a-service (RANaaS) concept, having the control and the data layers.589
However, in RANaaS, a cloud provides flexible and on-demand RAN functionalities (such as network management,590
13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
congestion control, radio resource management, medium access control, and physical layer management),591
according to the network requirements and characteristics, unlike [14]. Hence, there is no need to split592
functionalities in advance to the control and the data layers, as a result RANaaS provides more elasticity.593
Till now, it is clear that how a C-RAN will work. However, in order to achieve real-time performance, a RRH594
executing latency-critical applications may connect to a nearby small cloud while other RRHs that are not adhered595
to real-time applications may connect to a far larger cloud [15]. SoftAir [18] is also a two-layered C-RAN that596
performs mobility-management, resource-efficient network virtualization, and distributed and collaborative traffic597
balancing in the cloud.598
3-layered C-RAN architectures. The full-centralized C-RAN architecture [14] has some disadvantages, as:599
continuous exchange of raw baseband samples between the data and the control layers, and the control layer is600
usually far away from the data layer resulting in a processing delay.601
In order to remove these disadvantages, Liu et al. [78] proposed convergence of cloud and cellular systems602
(CONCERT). In this architecture, one more layer, called a software-defined service layer, is introduced at the top603
of the control layer. The functioning of the layers in CONCERT is as follows:604
1. Data layer: is identical to the full centralized C-RAN’s data layer [14], having RRHs with less powerful605
computational resources for application level computations.606
2. Control layer: works just as a logically centralized entity. The control layer coordinates with the data layer607
resources and presents them as virtual resources to the software-defined service layer. The control layer608
provides a few services as: radio interfacing management, wired networking management, and location-aware609
computing management to the data layer.610
3. Software-defined services layer: works as a virtual BS and provides services (e.g., application delivery, QoS,611
real-time communication, seamless mobility, security, network management, regulation, and power control) to612
the data layer.613
Wu et al. [115] enhanced C-RAN architecture [14] and RANaaS [97], by moving the whole RAN to a cloud.614
The proposed architecture also has three layers, where the data layer and the control layer are same to the respective615
layers of C-RAN [14, 78]. The third layer, called a service layer, executes in the cloud and provides some more616
functionalities than the software-defined services layer of [97], e.g., traffic management, the cell configuration,617
interference control, allocation of functional components to the physical elements, and video streaming services.618
The authors [119] proposed an all-software-defined network using three types of hierarchical network619
controllers, namely MBS controller, RAN controller, and network controller, where except the MBS controller620
all the others can be executed in the cloud, as follows:621
1. MBS controller: usually stays nearby UEs, and performs wireless resource management and packet creation.622
2. RAN controller: stays at the top of MBS controllers, and performs connectivity, RAT selection, handoff, QoS,623
policies, mobility management.624
3. Network controller: stays at the top of RAN controllers, ensures end-to-end QoS, and establishes625
application-aware routes.626
Advantages of C-RANs in 5G networks. C-RANs provide a variety of services as a software, power efficient,627
flexible, and scalable architecture for the future cellular communication. Here, we enlist some advantages of628
C-RANs, as follows:629
• An easy network management: C-RANs facilitate on-demand installation of virtual resources and execute630
cloud-based resources that dynamically manage interference, traffic, load balance, mobility, and do coordinated631
signal processing [30, 97].632
• Reduce cost: It is very costly and time-consuming to deploy and install a MBS to increase the network capacity.633
However, the deployment of C-RANs involves less cost, while it provides usual services like a MBS [115]. As634
a result, operators are required to only deploy, install, and operate RRHs in MBSs.635
• Save energy of UEs and a MBS: C-RANs offload data-intensive computations from a MBS and may store data636
of UEs and MBSs. Consequently, C-RANs allow UEs and MBSs to offload their energy-consuming tasks to a637
nearby cloud, which saves energy of UEs and MBSs [26].638
• Improved spectrum utilization: A C-RAN enables sharing of CSI, traffic data, and control information of mobile639
services among participating MBSs, and hence, results in increased cooperation among MBSs and reduced640
interference [115].641
Open issues. Transferring data from RRHs to BBUs, i.e., from the data layer to the control layer, is a crucial642
step based on the selection of the functions of a MBS that has to be sent to a cloud, resulting in the minimal data643
14
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
congestion control, radio resource management, medium access control, and physical layer management),591
according to the network requirements and characteristics, unlike [14]. Hence, there is no need to split592
functionalities in advance to the control and the data layers, as a result RANaaS provides more elasticity.593
Till now, it is clear that how a C-RAN will work. However, in order to achieve real-time performance, a RRH594
executing latency-critical applications may connect to a nearby small cloud while other RRHs that are not adhered595
to real-time applications may connect to a far larger cloud [15]. SoftAir [18] is also a two-layered C-RAN that596
performs mobility-management, resource-efficient network virtualization, and distributed and collaborative traffic597
balancing in the cloud.598
3-layered C-RAN architectures. The full-centralized C-RAN architecture [14] has some disadvantages, as:599
continuous exchange of raw baseband samples between the data and the control layers, and the control layer is600
usually far away from the data layer resulting in a processing delay.601
In order to remove these disadvantages, Liu et al. [78] proposed convergence of cloud and cellular systems602
(CONCERT). In this architecture, one more layer, called a software-defined service layer, is introduced at the top603
of the control layer. The functioning of the layers in CONCERT is as follows:604
1. Data layer: is identical to the full centralized C-RAN’s data layer [14], having RRHs with less powerful605
computational resources for application level computations.606
2. Control layer: works just as a logically centralized entity. The control layer coordinates with the data layer607
resources and presents them as virtual resources to the software-defined service layer. The control layer608
provides a few services as: radio interfacing management, wired networking management, and location-aware609
computing management to the data layer.610
3. Software-defined services layer: works as a virtual BS and provides services (e.g., application delivery, QoS,611
real-time communication, seamless mobility, security, network management, regulation, and power control) to612
the data layer.613
Wu et al. [115] enhanced C-RAN architecture [14] and RANaaS [97], by moving the whole RAN to a cloud.614
The proposed architecture also has three layers, where the data layer and the control layer are same to the respective615
layers of C-RAN [14, 78]. The third layer, called a service layer, executes in the cloud and provides some more616
functionalities than the software-defined services layer of [97], e.g., traffic management, the cell configuration,617
interference control, allocation of functional components to the physical elements, and video streaming services.618
The authors [119] proposed an all-software-defined network using three types of hierarchical network619
controllers, namely MBS controller, RAN controller, and network controller, where except the MBS controller620
all the others can be executed in the cloud, as follows:621
1. MBS controller: usually stays nearby UEs, and performs wireless resource management and packet creation.622
2. RAN controller: stays at the top of MBS controllers, and performs connectivity, RAT selection, handoff, QoS,623
policies, mobility management.624
3. Network controller: stays at the top of RAN controllers, ensures end-to-end QoS, and establishes625
application-aware routes.626
Advantages of C-RANs in 5G networks. C-RANs provide a variety of services as a software, power efficient,627
flexible, and scalable architecture for the future cellular communication. Here, we enlist some advantages of628
C-RANs, as follows:629
• An easy network management: C-RANs facilitate on-demand installation of virtual resources and execute630
cloud-based resources that dynamically manage interference, traffic, load balance, mobility, and do coordinated631
signal processing [30, 97].632
• Reduce cost: It is very costly and time-consuming to deploy and install a MBS to increase the network capacity.633
However, the deployment of C-RANs involves less cost, while it provides usual services like a MBS [115]. As634
a result, operators are required to only deploy, install, and operate RRHs in MBSs.635
• Save energy of UEs and a MBS: C-RANs offload data-intensive computations from a MBS and may store data636
of UEs and MBSs. Consequently, C-RANs allow UEs and MBSs to offload their energy-consuming tasks to a637
nearby cloud, which saves energy of UEs and MBSs [26].638
• Improved spectrum utilization: A C-RAN enables sharing of CSI, traffic data, and control information of mobile639
services among participating MBSs, and hence, results in increased cooperation among MBSs and reduced640
interference [115].641
Open issues. Transferring data from RRHs to BBUs, i.e., from the data layer to the control layer, is a crucial642
step based on the selection of the functions of a MBS that has to be sent to a cloud, resulting in the minimal data643
14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
movement in the network [97]. However, the selection of functions to be executed in a cloud and a MBS is a644
non-trivial affair [97]. The security and privacy issues involved in the cloud computing effect C-RANs, and hence,645
the development of a C-RAN has to deal with inherent challenges associated with the cloud and the wireless cellular646
communication simultaneously.647
4.5 Energy-Efficient Architectures for 5G Networks648
Energy-efficient infrastructures are a vital goal of 5G networks. Researchers have proposed a few ways of649
reducing energy in the infrastructure. Rowell et al. [64] considered a joint optimization of energy-efficiency and650
spectral-efficiency. A user-centric 5G network is suggested in [64] so that UEs are allowed to select UL and DL651
channels from different BSs depending on the load, channel conditions, services and application requirements. In652
a similar manner, decoupling of signaling and data is useful for energy saving; for example, a MBS may become a653
signaling BS while SBSs may serve all data requests. Thus, when there is no data traffic in a SBS, it can be turned654
off. A similar approach for decoupling of signaling and data is presented in [121]. However, in [121], a UE gets655
connected to a SBS according to instructions by a MBS, and hence, it results in less energy consumption at UEs’656
side due to less interference, faster small-cells’ discovery, and MBS-assisted handover.657
Hu and Qian [62] provided an energy-efficient C-RAN architecture in a manner that RRHs serve almost a same658
number of UEs. They also present an interference management approach so that the power consumption of SBSs659
and MBSs can be decreased. Like Rowell et al. [64], Hu and Qian [62] also suggested that the association of a UE660
cannot be done based on entirely a DL channel or a UL channel, and a UE must consider both the channels at the661
time of association with a BS. Lin et al. [80] suggested to include an energy harvesting device (to collect energy)662
and a spectrum harvesting controller (to collect spectrum) at SBSs.663
5 Implementation Issues in 5G Networks664
In this section, we will see issues regarding the interference, handoff, QoS, load balancing, and channel access665
management in the context of 5G networks.666
5.1 Interference Management in 5G Networks667
We have already seen challenges in interference management (Section 3). In this section, we will review some668
techniques/methods for interference management in 5G networks.669
Nam et al. [86] handled UE-side interference by using a new type of receiver equipment, called an advanced670
receiver, which detects, decodes, and removes interference from receiving signals. In addition, the network-side671
interference is managed by a joint scheduling, which selects each UE according to the resources needed (e.g.,672
time, frequency, transmission rate, and schemes of multiple cells) for its association with a BS. Hence, the joint673
scheduling, which can be implemented in a centralized or distributed manner, requires a coordination mechanism674
among the neighboring cells.675
Hossain et al. [61] proposed distributed cell access and power control (CAPC) schemes for handling676
interference in multi-tier architectures. CAPC involves: (i) prioritized power control (PPC), which assumes that677
UEs working under a SBS have a low-priority than UEs working under a MBS, and hence, low-priority UEs set their678
power so that the resulting interference must not exceed a predefined threshold; (ii) cell association (CA), which679
regards dynamic values of resources, traffic, distance to a MBS, and available channels at a MBS for selecting a680
MBS with the optimum values of the parameters; and (iii) resource-aware CA and PPC (RCA-PPC), which is a681
combination of the first two approaches and allows a UE to connect simultaneously with multiple BSs for a UL682
channel and a DL channel based on criteria of PPC and CA.683
Hong et al. [59] suggested to use self-interference cancellation (SIC) in small-cells’ networks. As we have684
seen that SBSs require methods to transfer backhaul data to a MBS (Section 4.1.1), the use of SIC can eliminate685
the need of such methods and result in self-backhauled small-cells (SSCs). SSCs use SIC for providing services686
and backhaul data transfer, and more importantly, they gain almost the same performance as having a small-cell687
connected with a wired optical fiber. It works as: in the DL channel, a SBS may receive from a MBS and688
simultaneously transmit to UEs. In the UL channel, a SBS may receive from UEs and simultaneously transmit689
15
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
movement in the network [97]. However, the selection of functions to be executed in a cloud and a MBS is a644
non-trivial affair [97]. The security and privacy issues involved in the cloud computing effect C-RANs, and hence,645
the development of a C-RAN has to deal with inherent challenges associated with the cloud and the wireless cellular646
communication simultaneously.647
4.5 Energy-Efficient Architectures for 5G Networks648
Energy-efficient infrastructures are a vital goal of 5G networks. Researchers have proposed a few ways of649
reducing energy in the infrastructure. Rowell et al. [64] considered a joint optimization of energy-efficiency and650
spectral-efficiency. A user-centric 5G network is suggested in [64] so that UEs are allowed to select UL and DL651
channels from different BSs depending on the load, channel conditions, services and application requirements. In652
a similar manner, decoupling of signaling and data is useful for energy saving; for example, a MBS may become a653
signaling BS while SBSs may serve all data requests. Thus, when there is no data traffic in a SBS, it can be turned654
off. A similar approach for decoupling of signaling and data is presented in [121]. However, in [121], a UE gets655
connected to a SBS according to instructions by a MBS, and hence, it results in less energy consumption at UEs’656
side due to less interference, faster small-cells’ discovery, and MBS-assisted handover.657
Hu and Qian [62] provided an energy-efficient C-RAN architecture in a manner that RRHs serve almost a same658
number of UEs. They also present an interference management approach so that the power consumption of SBSs659
and MBSs can be decreased. Like Rowell et al. [64], Hu and Qian [62] also suggested that the association of a UE660
cannot be done based on entirely a DL channel or a UL channel, and a UE must consider both the channels at the661
time of association with a BS. Lin et al. [80] suggested to include an energy harvesting device (to collect energy)662
and a spectrum harvesting controller (to collect spectrum) at SBSs.663
5 Implementation Issues in 5G Networks664
In this section, we will see issues regarding the interference, handoff, QoS, load balancing, and channel access665
management in the context of 5G networks.666
5.1 Interference Management in 5G Networks667
We have already seen challenges in interference management (Section 3). In this section, we will review some668
techniques/methods for interference management in 5G networks.669
Nam et al. [86] handled UE-side interference by using a new type of receiver equipment, called an advanced670
receiver, which detects, decodes, and removes interference from receiving signals. In addition, the network-side671
interference is managed by a joint scheduling, which selects each UE according to the resources needed (e.g.,672
time, frequency, transmission rate, and schemes of multiple cells) for its association with a BS. Hence, the joint673
scheduling, which can be implemented in a centralized or distributed manner, requires a coordination mechanism674
among the neighboring cells.675
Hossain et al. [61] proposed distributed cell access and power control (CAPC) schemes for handling676
interference in multi-tier architectures. CAPC involves: (i) prioritized power control (PPC), which assumes that677
UEs working under a SBS have a low-priority than UEs working under a MBS, and hence, low-priority UEs set their678
power so that the resulting interference must not exceed a predefined threshold; (ii) cell association (CA), which679
regards dynamic values of resources, traffic, distance to a MBS, and available channels at a MBS for selecting a680
MBS with the optimum values of the parameters; and (iii) resource-aware CA and PPC (RCA-PPC), which is a681
combination of the first two approaches and allows a UE to connect simultaneously with multiple BSs for a UL682
channel and a DL channel based on criteria of PPC and CA.683
Hong et al. [59] suggested to use self-interference cancellation (SIC) in small-cells’ networks. As we have684
seen that SBSs require methods to transfer backhaul data to a MBS (Section 4.1.1), the use of SIC can eliminate685
the need of such methods and result in self-backhauled small-cells (SSCs). SSCs use SIC for providing services686
and backhaul data transfer, and more importantly, they gain almost the same performance as having a small-cell687
connected with a wired optical fiber. It works as: in the DL channel, a SBS may receive from a MBS and688
simultaneously transmit to UEs. In the UL channel, a SBS may receive from UEs and simultaneously transmit689
15
Secure Best Marks with AI Grader
Need help grading? Try our AI Grader for instant feedback on your assignments.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
data to the MBS. Therefore, a small-cell can completely remove the need of a separate backhaul data transfer690
method, resulting in reduced cost.691
The authors [64] suggested that the measurement of inter-user interference channel, and then, the allocation692
of UL and DL channels by a MBS can mitigate inter-user UL-to-DL interference in a single-cell full duplex radio693
network. However, in the case of a multi-cell full duplex radio network, interference mitigation becomes more694
complex, because of the existence of interference in UL and DL channels between multi-cells’ UEs that work on695
identical frequency and time.696
Open issues. Interference cancellation in a full duplex radio network is still an open problem for multi-cell. The697
design of algorithms for inter-BS DL-to-UL-interference and inter-user UL-to-DL-interference cancellation in a698
full duplex radio network is still open and to be explored.699
5.2 Handoff Management in 5G Networks700
Handoff provides a way to UEs connected to a BS to move to another BS without disconnecting their sessions.701
Challenges in the handoff process in 5G networks. The handoff management in 5G networks has inherent702
challenges associated with the current cellular networks, e.g., minimum latency, improved routing, security, and less703
uncertainty of having no services. The network densification, very high mobility, the zero latency, and accessing704
multi-RATs make handoff management in 5G networks harder. Also, the current cellular networks do not provide705
efficient load balancing for a BS at the time of handoff. For example, movement of UEs from houses to offices in706
the morning creates a load imbalance at BSs of respective areas.707
Types of handoff in 5G networks. Three types of handoffs are presented in the context of 5G networks, as follows:708
1. Intra-macrocell handoff : refers to handoff between small-cells that are working under a single MBS [102].709
2. Inter-macrocell handoff : refers to handoff between macrocells. It may also lead to handoff between two710
small-cells that are working under different MBSs. Note that if the handover between small-cells of two711
different MBSs is not done properly, then the inter-macrocell handoff also fails [102].712
3. Multi-RATs handoff : refers to handoff of a UE from a RAT to other RAT.713
Song et al. [102] provided a handoff mechanism for highly mobile users, where a UE sends some parameters714
(e.g. QoS, signal-to-interference ratio (SIR), and time to handoff) in a measurement report to the current MBS. SIR715
is considered as a primary factor for finding a situation for an initiation of the handoff. The Gray system model716
predicts the (N + 1)th measurement report from the N th measurement report. The predicted value is used for the717
final decision for the handover process. Zhang et al. [121] proposed a handoff mechanism assisted by a MBS. The718
MBS collects several parameters from UEs, and if the MBS finds the values of the received parameters below a719
threshold, then it finds a new SBS or MBS for handoff and informs to the UEs.720
For handoff over different RATs, Orsino et al. [93] proposed a handoff procedure so that a UE can select the721
most suitable RAT without any performance loss. A UE collects received signal strength (i.e., RSRP) or quality (i.e.,722
RSRQ) from the current MBS, and then, it initiates handoff if RSRQ is below a threshold. The UE collects several723
parameters (e.g., transmitted power, the cell’s traffic load, and UE requested spectral efficiency) from adjacent BSs,724
and then, selects the most suitable BS.725
Duan et al. [40] provided an authenticated handoff procedure for C-RANs and multi-tier 5G networks. The726
control layer holds an authentication handover module (AHM) for monitoring and predicting the future location727
of UEs (based on the current location) and preparing relevant cells before UEs’ arrival in that. The AHM holds a728
master public-private key pair with each RRH, which are authenticated by the AHM in off-peak hours, and UEs729
are verified before accessing the network services by RRHs. UEs sends ID, the physical layer’s attributes, location,730
speed, and direction to the control layer in a secure manner for the handoff process. The proposed approach reduces731
the risk of impersonation and man-in-the-middle attacks. Giust et al. [52] provided three distributed mobility732
management protocols, the first is based on the existing Proxy Mobile IPv6 (PMIPv6), the second is based on733
SDN, and the third is based on routing protocols.734
Open issues. Handoff mechanisms are yet to be explored. It will be interesting to find solutions to extremely735
dense HetNets. Furthermore, the handoff process may also create interference to other UEs; hence, it is required736
to develop algorithms while considering different types of interferences in 5G networks and a tradeoff between the737
number of handoffs and the level of interference in the network [73].738
16
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
data to the MBS. Therefore, a small-cell can completely remove the need of a separate backhaul data transfer690
method, resulting in reduced cost.691
The authors [64] suggested that the measurement of inter-user interference channel, and then, the allocation692
of UL and DL channels by a MBS can mitigate inter-user UL-to-DL interference in a single-cell full duplex radio693
network. However, in the case of a multi-cell full duplex radio network, interference mitigation becomes more694
complex, because of the existence of interference in UL and DL channels between multi-cells’ UEs that work on695
identical frequency and time.696
Open issues. Interference cancellation in a full duplex radio network is still an open problem for multi-cell. The697
design of algorithms for inter-BS DL-to-UL-interference and inter-user UL-to-DL-interference cancellation in a698
full duplex radio network is still open and to be explored.699
5.2 Handoff Management in 5G Networks700
Handoff provides a way to UEs connected to a BS to move to another BS without disconnecting their sessions.701
Challenges in the handoff process in 5G networks. The handoff management in 5G networks has inherent702
challenges associated with the current cellular networks, e.g., minimum latency, improved routing, security, and less703
uncertainty of having no services. The network densification, very high mobility, the zero latency, and accessing704
multi-RATs make handoff management in 5G networks harder. Also, the current cellular networks do not provide705
efficient load balancing for a BS at the time of handoff. For example, movement of UEs from houses to offices in706
the morning creates a load imbalance at BSs of respective areas.707
Types of handoff in 5G networks. Three types of handoffs are presented in the context of 5G networks, as follows:708
1. Intra-macrocell handoff : refers to handoff between small-cells that are working under a single MBS [102].709
2. Inter-macrocell handoff : refers to handoff between macrocells. It may also lead to handoff between two710
small-cells that are working under different MBSs. Note that if the handover between small-cells of two711
different MBSs is not done properly, then the inter-macrocell handoff also fails [102].712
3. Multi-RATs handoff : refers to handoff of a UE from a RAT to other RAT.713
Song et al. [102] provided a handoff mechanism for highly mobile users, where a UE sends some parameters714
(e.g. QoS, signal-to-interference ratio (SIR), and time to handoff) in a measurement report to the current MBS. SIR715
is considered as a primary factor for finding a situation for an initiation of the handoff. The Gray system model716
predicts the (N + 1)th measurement report from the N th measurement report. The predicted value is used for the717
final decision for the handover process. Zhang et al. [121] proposed a handoff mechanism assisted by a MBS. The718
MBS collects several parameters from UEs, and if the MBS finds the values of the received parameters below a719
threshold, then it finds a new SBS or MBS for handoff and informs to the UEs.720
For handoff over different RATs, Orsino et al. [93] proposed a handoff procedure so that a UE can select the721
most suitable RAT without any performance loss. A UE collects received signal strength (i.e., RSRP) or quality (i.e.,722
RSRQ) from the current MBS, and then, it initiates handoff if RSRQ is below a threshold. The UE collects several723
parameters (e.g., transmitted power, the cell’s traffic load, and UE requested spectral efficiency) from adjacent BSs,724
and then, selects the most suitable BS.725
Duan et al. [40] provided an authenticated handoff procedure for C-RANs and multi-tier 5G networks. The726
control layer holds an authentication handover module (AHM) for monitoring and predicting the future location727
of UEs (based on the current location) and preparing relevant cells before UEs’ arrival in that. The AHM holds a728
master public-private key pair with each RRH, which are authenticated by the AHM in off-peak hours, and UEs729
are verified before accessing the network services by RRHs. UEs sends ID, the physical layer’s attributes, location,730
speed, and direction to the control layer in a secure manner for the handoff process. The proposed approach reduces731
the risk of impersonation and man-in-the-middle attacks. Giust et al. [52] provided three distributed mobility732
management protocols, the first is based on the existing Proxy Mobile IPv6 (PMIPv6), the second is based on733
SDN, and the third is based on routing protocols.734
Open issues. Handoff mechanisms are yet to be explored. It will be interesting to find solutions to extremely735
dense HetNets. Furthermore, the handoff process may also create interference to other UEs; hence, it is required736
to develop algorithms while considering different types of interferences in 5G networks and a tradeoff between the737
number of handoffs and the level of interference in the network [73].738
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
5.3 QoS Management in 5G Networks739
We have already seen challenges in QoS management in Section 3. In this section, we will review some740
techniques/methods for QoS management in 5G networks.741
Zhang et al. [120] provided a mechanism for different delay-bounded QoS for various types of services,742
applications, and users having diverse requirements, called heterogeneous statistical delay-bounded QoS743
provisioning (HSP). HSP maximizes the aggregate effective capacity of different types of 5G architectures,744
reviewed in Sections 4.1, 4.2, and 4.3. HSP [120] algorithm claims better performance over other approaches;745
however, it imposes new challenges in terms of the assignment of different resources for different links under the746
cover of delay-bounded QoS requirements.747
The authors [123] proposed the deployment of a quality management element (QME) in the cloud for748
monitoring inter-UEs and inter-layer (the control and the data layers in C-RANs) QoS. RRUs send wireless749
information (e.g., CSI, reference signal received quality, and resource block utilization) to the QME. Consequently,750
the QME executes service control algorithms (to manage QoS and some other activities like traffic offloading and751
customized scheduling), and then, sends scheduling strategies to the RRUs to achieve a desired level of QoS.752
Kim et al. [69] provided a routing algorithm for multi-hop D2D communication. The algorithm takes into753
account different QoS for each link so that it can achieve better performance than max-min routing algorithms. The754
algorithm increases flow until it provides the desired QoS or reaches the maximum capacity of the link. Since the755
algorithm considers individual links, there is a high probability that some of the channels will serve multiple-links756
with the desired QoS. Zhou et al. [124] provided a QoS-aware and energy-efficient resource allocation algorithm757
for DL channels, where UEs are allocated an identical power in one case and non-identical power in the second758
case. The algorithm maximizes energy-efficiency while minimizes transmit power. Hu and Qian [62] suggested759
that a UE must consider the source data rate, delay bound, and delay-violation probability before connecting to a760
MBS.761
Open issues. The 5G networks are supposed to satisfy the highest level of QoS. The tactile Internet [47] requires the762
best QoS, especially, latency of the order of 1 millisecond for senses such as touching, seeing, and hearing objects763
far away, as precise as human perception. However, the current proposed architectures do not support efficient764
tactile Internet services. In future, it would be a promising area as to encode senses, exchange data satisfying the765
zero latency, and enable the user to receive the sensation.766
5.4 Load balancing in 5G Networks767
Load balancing means allocation of resources to a cell such that all the users meet their demands. It is an important768
issue in the cellular wireless networks. In a 2-tier architecture, discussed in Section 4.1, user offloading to a769
small-cell is useless if there is no resource partitioning [99]. Singh and Andrews [99] provided an analytical and770
tractable framework for modeling and analyzing joint resource partitioning and offloading in a 2-tier architecture.771
Hossain et al. [61] provided a technique for cell association based on dynamic resources and traffic in a cell, as we772
have already discussed in Section 5.1. In a fast moving vehicle, e.g., a train, it is very hard to allocate resources773
without any service interruption at all. A distributed load balancing algorithm for fast moving vehicles is presented774
in [53]. Load balancing architectures for D2D communication are given in [77, 33, 58], which we discussed in775
Section 4.3. Interested readers may refer to [21] for further details of load balancing in 5G networks.776
5.5 Channel Access Control Management in 5G Network777
Channel access protocols allow several UEs to share a transmission channel without any collision while utilizing778
the maximum channel capacity.779
Challenges in channel access control management in 5G networks. Channel access control management in 5G780
networks faces inherent challenges associated with the current cellular networks, e.g., synchronization, fairness,781
adaptive rate control, resource reservation, real-time traffic support, scalability, throughput, and delay.In addition,782
providing the currently available best channel in 5G networks is prone to additional challenges, as: high mobility of783
UEs, working at the higher frequencies (> 3 GHz), different RATs, dense networks, high QoS, high link reliability,784
and the zero latency for applications and services.785
The authors [101] proposed a frame-based medium access control (FD-MAC) protocol for mmWave-based786
small-cells. FD-MAC consists of two phases, as: (i) scheduling phase, when a SBS collects the traffic demands787
17
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
5.3 QoS Management in 5G Networks739
We have already seen challenges in QoS management in Section 3. In this section, we will review some740
techniques/methods for QoS management in 5G networks.741
Zhang et al. [120] provided a mechanism for different delay-bounded QoS for various types of services,742
applications, and users having diverse requirements, called heterogeneous statistical delay-bounded QoS743
provisioning (HSP). HSP maximizes the aggregate effective capacity of different types of 5G architectures,744
reviewed in Sections 4.1, 4.2, and 4.3. HSP [120] algorithm claims better performance over other approaches;745
however, it imposes new challenges in terms of the assignment of different resources for different links under the746
cover of delay-bounded QoS requirements.747
The authors [123] proposed the deployment of a quality management element (QME) in the cloud for748
monitoring inter-UEs and inter-layer (the control and the data layers in C-RANs) QoS. RRUs send wireless749
information (e.g., CSI, reference signal received quality, and resource block utilization) to the QME. Consequently,750
the QME executes service control algorithms (to manage QoS and some other activities like traffic offloading and751
customized scheduling), and then, sends scheduling strategies to the RRUs to achieve a desired level of QoS.752
Kim et al. [69] provided a routing algorithm for multi-hop D2D communication. The algorithm takes into753
account different QoS for each link so that it can achieve better performance than max-min routing algorithms. The754
algorithm increases flow until it provides the desired QoS or reaches the maximum capacity of the link. Since the755
algorithm considers individual links, there is a high probability that some of the channels will serve multiple-links756
with the desired QoS. Zhou et al. [124] provided a QoS-aware and energy-efficient resource allocation algorithm757
for DL channels, where UEs are allocated an identical power in one case and non-identical power in the second758
case. The algorithm maximizes energy-efficiency while minimizes transmit power. Hu and Qian [62] suggested759
that a UE must consider the source data rate, delay bound, and delay-violation probability before connecting to a760
MBS.761
Open issues. The 5G networks are supposed to satisfy the highest level of QoS. The tactile Internet [47] requires the762
best QoS, especially, latency of the order of 1 millisecond for senses such as touching, seeing, and hearing objects763
far away, as precise as human perception. However, the current proposed architectures do not support efficient764
tactile Internet services. In future, it would be a promising area as to encode senses, exchange data satisfying the765
zero latency, and enable the user to receive the sensation.766
5.4 Load balancing in 5G Networks767
Load balancing means allocation of resources to a cell such that all the users meet their demands. It is an important768
issue in the cellular wireless networks. In a 2-tier architecture, discussed in Section 4.1, user offloading to a769
small-cell is useless if there is no resource partitioning [99]. Singh and Andrews [99] provided an analytical and770
tractable framework for modeling and analyzing joint resource partitioning and offloading in a 2-tier architecture.771
Hossain et al. [61] provided a technique for cell association based on dynamic resources and traffic in a cell, as we772
have already discussed in Section 5.1. In a fast moving vehicle, e.g., a train, it is very hard to allocate resources773
without any service interruption at all. A distributed load balancing algorithm for fast moving vehicles is presented774
in [53]. Load balancing architectures for D2D communication are given in [77, 33, 58], which we discussed in775
Section 4.3. Interested readers may refer to [21] for further details of load balancing in 5G networks.776
5.5 Channel Access Control Management in 5G Network777
Channel access protocols allow several UEs to share a transmission channel without any collision while utilizing778
the maximum channel capacity.779
Challenges in channel access control management in 5G networks. Channel access control management in 5G780
networks faces inherent challenges associated with the current cellular networks, e.g., synchronization, fairness,781
adaptive rate control, resource reservation, real-time traffic support, scalability, throughput, and delay.In addition,782
providing the currently available best channel in 5G networks is prone to additional challenges, as: high mobility of783
UEs, working at the higher frequencies (> 3 GHz), different RATs, dense networks, high QoS, high link reliability,784
and the zero latency for applications and services.785
The authors [101] proposed a frame-based medium access control (FD-MAC) protocol for mmWave-based786
small-cells. FD-MAC consists of two phases, as: (i) scheduling phase, when a SBS collects the traffic demands787
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
from the supported UEs and computes a schedule for data transmission, and (ii) transmission phase, when the788
UEs start concurrent transmissions following the schedule. A schedule, which is computed using a graph-edge789
coloring algorithm, consists of a sequence of topologies and a sequence of time intervals for indicating how long790
each topology should sustain.791
Liu et al. [80] provided MAC protocols for UEs of small-cells and macrocells. Two types of MAC protocols792
for SBS’s UEs are suggested, as: contention-based random channel access (CRCA), where UEs randomly access793
the channel and send messages if the channel is available, and reservation-based channel access (RCA), which uses794
time division multiple access. For macrocells’ UEs, they provided a MAC protocol for CRN-based 5G networks,795
where SUs sense a licensed channel until it is free or the residual energy of SUs exceeds a predetermined threshold,796
for saving their battery. Further, they evaluated a tradeoff between the network throughput and sensing overhead.797
The authors [41] suggested two CRN-based channel access techniques for cognitive SBSs. The first channel798
access scheme is termed as contention-resolution-based channel access (CCA), which is similar to CRCA [80], is799
based on carrier sense multiple access. The other channel access scheme is termed as uncoordinated aggressive800
channel access (UCA), which works identically as RCA [80], for aggressively using channels in a small-cell for801
increasing opportunistic spectrum access performance. Nikopour et al. [90] provided a multi-user sparse code802
multiple access (MU-SCMA) for increasing DL’s spectral efficiency. MU-SCMA does not require the complete803
CSI, and hence, provides high data rate and the robustness to mobility. Nikopour et al. [90] also provided an uplink804
contention based SCMA for massive connectivity, data transmission with low signaling overhead, low delay, and805
diverse traffic connectivity requirements.806
Open issues. The current channel access protocols do not regard QoS and latency challenges in 5G networks.807
Hence, there is a scope of designing algorithms for finding multiple reliable links with the desired QoS and the zero808
latency.809
5.6 Security and Privacy Management in 5G Networks810
In this section, we present security and privacy related challenges and a discussion of security and privacy protocols811
in the context of 5G networks.812
Challenges in security and privacy in 5G networks. Authentication is a vital issue in any network. Due to813
the zero latency guarantee of 5G networks, authentication of UEs and network devices is very challenging, since814
the current authentication mechanisms use an authentication server that takes hundreds of milliseconds delay in a815
preliminary authentication phase [40]. A fast and frequent handover of UEs over small-cells requires for a robust,816
efficient, and secure handoff process for transferring context information [40]. Security to multi-RATs selection817
is also challenging, since each RAT has its own challenges and certain methods to provide security; clearly, there818
is a need to provide overlapped security solutions across the various types of RATs. C-RANs also inherit all the819
challenges associated with the cloud computing and wireless networks. In addition, several other challenges (e.g.,820
authorization and access control of UEs, availability of the network, confidentiality of communication and data821
transfer, integrity of communication and data transmission, accounting and auditing of a task, low computation822
complexity, and communication cost) require sophisticated solutions to make a secure 5G network. The security823
and latency are correlated as a higher level of security and privacy results in increased latencies. Therefore, the824
communication satisfying the zero latency is cumbersome when combined with secure and privacy-preserving 5G825
networks.826
Monitoring is suggested for securing the network and detecting intruders [105, 48]. However, monitoring a827
large number of UEs (by a trusted authority) is not a trivial task; hence, we do not see monitoring as a preferred way828
to secure networks. Yang et al. [118] focused on the physical layer security, which is independent of computational829
complexity and easily handles any number of devices. The physical layer security protocol considers locations of830
UEs and provides the best way for UEs for securely selecting a MBS or a SBS without overloading the network.831
Tehrani et al. [104] provided a method for secure and private D2D communications, called close access, where832
D-UEs have a list of other trusted D-UEs devices, and all such UEs can communicate directly using an encryption833
scheme while the remaining UEs not in the list utilize a MBS-assisted communication. An encryption based video834
sharing scheme is also presented in [79]. Kantola et al. [67] proposed a policy-based scheme that can prevent DoS835
and spoofing attacks. A solution to secure handoff is given in [40] and discussed in Section 5.2.836
Open issues. The current security and privacy solutions to 5G networks are not impressive and, hopefully, unable837
to handle massive connections. We can clearly visualize a potential scope for developing latency-aware protocols838
18
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
from the supported UEs and computes a schedule for data transmission, and (ii) transmission phase, when the788
UEs start concurrent transmissions following the schedule. A schedule, which is computed using a graph-edge789
coloring algorithm, consists of a sequence of topologies and a sequence of time intervals for indicating how long790
each topology should sustain.791
Liu et al. [80] provided MAC protocols for UEs of small-cells and macrocells. Two types of MAC protocols792
for SBS’s UEs are suggested, as: contention-based random channel access (CRCA), where UEs randomly access793
the channel and send messages if the channel is available, and reservation-based channel access (RCA), which uses794
time division multiple access. For macrocells’ UEs, they provided a MAC protocol for CRN-based 5G networks,795
where SUs sense a licensed channel until it is free or the residual energy of SUs exceeds a predetermined threshold,796
for saving their battery. Further, they evaluated a tradeoff between the network throughput and sensing overhead.797
The authors [41] suggested two CRN-based channel access techniques for cognitive SBSs. The first channel798
access scheme is termed as contention-resolution-based channel access (CCA), which is similar to CRCA [80], is799
based on carrier sense multiple access. The other channel access scheme is termed as uncoordinated aggressive800
channel access (UCA), which works identically as RCA [80], for aggressively using channels in a small-cell for801
increasing opportunistic spectrum access performance. Nikopour et al. [90] provided a multi-user sparse code802
multiple access (MU-SCMA) for increasing DL’s spectral efficiency. MU-SCMA does not require the complete803
CSI, and hence, provides high data rate and the robustness to mobility. Nikopour et al. [90] also provided an uplink804
contention based SCMA for massive connectivity, data transmission with low signaling overhead, low delay, and805
diverse traffic connectivity requirements.806
Open issues. The current channel access protocols do not regard QoS and latency challenges in 5G networks.807
Hence, there is a scope of designing algorithms for finding multiple reliable links with the desired QoS and the zero808
latency.809
5.6 Security and Privacy Management in 5G Networks810
In this section, we present security and privacy related challenges and a discussion of security and privacy protocols811
in the context of 5G networks.812
Challenges in security and privacy in 5G networks. Authentication is a vital issue in any network. Due to813
the zero latency guarantee of 5G networks, authentication of UEs and network devices is very challenging, since814
the current authentication mechanisms use an authentication server that takes hundreds of milliseconds delay in a815
preliminary authentication phase [40]. A fast and frequent handover of UEs over small-cells requires for a robust,816
efficient, and secure handoff process for transferring context information [40]. Security to multi-RATs selection817
is also challenging, since each RAT has its own challenges and certain methods to provide security; clearly, there818
is a need to provide overlapped security solutions across the various types of RATs. C-RANs also inherit all the819
challenges associated with the cloud computing and wireless networks. In addition, several other challenges (e.g.,820
authorization and access control of UEs, availability of the network, confidentiality of communication and data821
transfer, integrity of communication and data transmission, accounting and auditing of a task, low computation822
complexity, and communication cost) require sophisticated solutions to make a secure 5G network. The security823
and latency are correlated as a higher level of security and privacy results in increased latencies. Therefore, the824
communication satisfying the zero latency is cumbersome when combined with secure and privacy-preserving 5G825
networks.826
Monitoring is suggested for securing the network and detecting intruders [105, 48]. However, monitoring a827
large number of UEs (by a trusted authority) is not a trivial task; hence, we do not see monitoring as a preferred way828
to secure networks. Yang et al. [118] focused on the physical layer security, which is independent of computational829
complexity and easily handles any number of devices. The physical layer security protocol considers locations of830
UEs and provides the best way for UEs for securely selecting a MBS or a SBS without overloading the network.831
Tehrani et al. [104] provided a method for secure and private D2D communications, called close access, where832
D-UEs have a list of other trusted D-UEs devices, and all such UEs can communicate directly using an encryption833
scheme while the remaining UEs not in the list utilize a MBS-assisted communication. An encryption based video834
sharing scheme is also presented in [79]. Kantola et al. [67] proposed a policy-based scheme that can prevent DoS835
and spoofing attacks. A solution to secure handoff is given in [40] and discussed in Section 5.2.836
Open issues. The current security and privacy solutions to 5G networks are not impressive and, hopefully, unable837
to handle massive connections. We can clearly visualize a potential scope for developing latency-aware protocols838
18
Paraphrase This Document
Need a fresh take? Get an instant paraphrase of this document with our AI Paraphraser
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
along with security awareness that must consider secure data transmission, end-to-end security, secure and private839
storage, threats resistant UEs, and valid network and software access.840
6 Methodologies and Technologies for 5G Networks841
It is already mentioned in Section 1 that the development of 5G networks requires the design and implementation of842
new methodologies, techniques, and architectures. We have reviewed some of the methodologies and technologies843
in the previous sections, such as: (i) full duplex radios in Sections 3 and 5.1, (ii) CRNs in Section 4.2, (iii)844
D2D communication in Section 4.3, (iv) multi-tier heterogeneous deployment or dense-deployment techniques845
in Sections 4.1, 4.2, and 4.3, (v) C-RANs in Section 4.4, (vi) ‘green’ communication systems in Section 4.5, and846
(vii) techniques related to interference, QoS, handoff, channel access, load balancing in Section 5.847
In this section, we will briefly describe some techniques that are mentioned but not explained in earlier sections.848
Self-Interference Cancellation (SIC). When a full duplex radio receives signals from another radio, it also849
receives interference signals by its own transmission, resulting in self-interference. Hence, a full duplex radio850
has to implement techniques to cancel self-interference [27, 56, 122]. SIC techniques are classified into passive851
and active cancellations; see [122]. As advantages, the implementation of SIC enables seamless global roaming,852
high-throughput services, and low-latency applications in a cost effective manner [59].853
Downlink and Uplink Decoupling (DUD). In the current cellular networks, a UE is associated with a BS based854
on the received signal power in its DL channel, and then, uses the same BS for UL channel transmission [100].855
DUD allows a UE to select a DL channel and a UL channel from two different BSs, based on the link quality, the856
cell load, and the cell backhaul capacity [42, 43]. Therefore, a UE may have the DL channel connected through857
a BS and the UL channel connected through a different BS, resulting in a user-centric 5G architecture [64] and858
improving the capacity of UL channels, which is a prime concern.859
Network Function Virtualization (NFV). NFV [55] implements network functions such as network address860
translation, firewalls, intrusion detection, domain name service, the traffic load management, and caching through861
software running on commodity servers. However, the conventional networks implement these functions on862
dedicated and application specific servers. Hence, NFV decreases the burden on network operators by not updating863
dedicated servers/hardware, thereby saves cost.864
Software-Defined Networking (SDN). SDN architectures [68, 45] partition network control functions and data865
forwarding functions, thereby the network control functions are programmable, and the network infrastructure866
handles applications and network services. SDN architectures can be divided into three parts, as: (i) the software867
controller: holds network control functions such as the network manager, APIs, network operating system, and868
maintaining the global view of the network; (ii) the southbound part: provides an interface and a protocol between869
the controller and SDN-enable infrastructure, where OpenFlow [84] is the most famous protocol that provides870
communication between the controller and the southbound part; (iii) the northbound part: provides an interface871
between SDN applications and the controller [57]. Interested readers may refer to [31, 117, 70, 45, 23, 19] to find872
details of SDN, challenges in SDN, and applications of SDN.873
Note that SDN, NFV, and C-RANs offload functionalities to software running on commodity servers. However,874
SDN separates network control functions from data forwarding functions, while NFV implements network875
functions in software [108]. Besides that C-RANs integrate both SDN and NFV to meet the scalability and876
flexibility requirements in the future mobile networks [108].877
Millimeter Waves (mmWave). The current wireless bandwidth is not able to support a huge number of UEs in 5G878
networks. Hence, researchers are looking at 30 GHz to 300 GHz frequency bands, where mmWave communication879
is proposed for achieving high-speed data transfer. The current research focuses on 28 GHz band, 38 GHz band,880
60 GHz band, and the E-band (71–76 GHz and 81–86 GHz) [91]. However, mmWave has several challenges at881
the physical, MAC, and network layers. There are a number of papers about mmWave, and hence, we are not882
discussing mmWave in details. Interested readers may refer to [95, 29, 91].883
Machine-to-Machine (M2M) communication. M2M communication refers to the communication between884
(network) devices (e.g., sensors/monitoring-devices with a cloud) without human intervention. Some examples885
of M2M communication are intelligent transport systems, health measurement, monitoring of buildings, gas and886
oil pipelines, intelligent retail systems, and security and safety systems. However, the development of M2M887
communication involves several challenges to be handled in the future, as follows: connectivity of massive devices,888
19
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
along with security awareness that must consider secure data transmission, end-to-end security, secure and private839
storage, threats resistant UEs, and valid network and software access.840
6 Methodologies and Technologies for 5G Networks841
It is already mentioned in Section 1 that the development of 5G networks requires the design and implementation of842
new methodologies, techniques, and architectures. We have reviewed some of the methodologies and technologies843
in the previous sections, such as: (i) full duplex radios in Sections 3 and 5.1, (ii) CRNs in Section 4.2, (iii)844
D2D communication in Section 4.3, (iv) multi-tier heterogeneous deployment or dense-deployment techniques845
in Sections 4.1, 4.2, and 4.3, (v) C-RANs in Section 4.4, (vi) ‘green’ communication systems in Section 4.5, and846
(vii) techniques related to interference, QoS, handoff, channel access, load balancing in Section 5.847
In this section, we will briefly describe some techniques that are mentioned but not explained in earlier sections.848
Self-Interference Cancellation (SIC). When a full duplex radio receives signals from another radio, it also849
receives interference signals by its own transmission, resulting in self-interference. Hence, a full duplex radio850
has to implement techniques to cancel self-interference [27, 56, 122]. SIC techniques are classified into passive851
and active cancellations; see [122]. As advantages, the implementation of SIC enables seamless global roaming,852
high-throughput services, and low-latency applications in a cost effective manner [59].853
Downlink and Uplink Decoupling (DUD). In the current cellular networks, a UE is associated with a BS based854
on the received signal power in its DL channel, and then, uses the same BS for UL channel transmission [100].855
DUD allows a UE to select a DL channel and a UL channel from two different BSs, based on the link quality, the856
cell load, and the cell backhaul capacity [42, 43]. Therefore, a UE may have the DL channel connected through857
a BS and the UL channel connected through a different BS, resulting in a user-centric 5G architecture [64] and858
improving the capacity of UL channels, which is a prime concern.859
Network Function Virtualization (NFV). NFV [55] implements network functions such as network address860
translation, firewalls, intrusion detection, domain name service, the traffic load management, and caching through861
software running on commodity servers. However, the conventional networks implement these functions on862
dedicated and application specific servers. Hence, NFV decreases the burden on network operators by not updating863
dedicated servers/hardware, thereby saves cost.864
Software-Defined Networking (SDN). SDN architectures [68, 45] partition network control functions and data865
forwarding functions, thereby the network control functions are programmable, and the network infrastructure866
handles applications and network services. SDN architectures can be divided into three parts, as: (i) the software867
controller: holds network control functions such as the network manager, APIs, network operating system, and868
maintaining the global view of the network; (ii) the southbound part: provides an interface and a protocol between869
the controller and SDN-enable infrastructure, where OpenFlow [84] is the most famous protocol that provides870
communication between the controller and the southbound part; (iii) the northbound part: provides an interface871
between SDN applications and the controller [57]. Interested readers may refer to [31, 117, 70, 45, 23, 19] to find872
details of SDN, challenges in SDN, and applications of SDN.873
Note that SDN, NFV, and C-RANs offload functionalities to software running on commodity servers. However,874
SDN separates network control functions from data forwarding functions, while NFV implements network875
functions in software [108]. Besides that C-RANs integrate both SDN and NFV to meet the scalability and876
flexibility requirements in the future mobile networks [108].877
Millimeter Waves (mmWave). The current wireless bandwidth is not able to support a huge number of UEs in 5G878
networks. Hence, researchers are looking at 30 GHz to 300 GHz frequency bands, where mmWave communication879
is proposed for achieving high-speed data transfer. The current research focuses on 28 GHz band, 38 GHz band,880
60 GHz band, and the E-band (71–76 GHz and 81–86 GHz) [91]. However, mmWave has several challenges at881
the physical, MAC, and network layers. There are a number of papers about mmWave, and hence, we are not882
discussing mmWave in details. Interested readers may refer to [95, 29, 91].883
Machine-to-Machine (M2M) communication. M2M communication refers to the communication between884
(network) devices (e.g., sensors/monitoring-devices with a cloud) without human intervention. Some examples885
of M2M communication are intelligent transport systems, health measurement, monitoring of buildings, gas and886
oil pipelines, intelligent retail systems, and security and safety systems. However, the development of M2M887
communication involves several challenges to be handled in the future, as follows: connectivity of massive devices,888
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
bursty data, the zero latency, scalability in terms of supporting devices, technologies and diverse applications,889
fast and reliable delivery of messages, and cost of devices. In addition, the development of efficient algorithms890
for location, time, group, priority, and multi-hop data transmission management for M2M communication is891
needed [13, 103]. Interested readers may refer to [22, 44, 82, 36, 96].892
Massive MIMO (mMIMO). mMIMO systems are also known as large-scale antenna systems, very large MIMO,893
hyper-MIMO, and full-dimension MIMO [71, 87]. mMIMO systems use antenna arrays with hundreds of antennas894
at MBS for simultaneously serving many UEs with a single antenna in identical time and frequency. Hence,895
expensive equipment are mounted on a MBS [81]. A mMIMO system relies on spatial multiplexing, which in turn896
relies on the channel knowledge at MBS, on both UL and DL channels. A mMIMO system reduces latency and897
energy, simplifies the MAC layer, shows robustness against intentional jamming, and increases the capacity due to898
spatial multiplexing.899
Visual Light Communication (VLC). VLC is a high-speed data transfer medium for short range LOS optical links900
in the future cellular networks [116]. The light-emitting diodes (LEDs) provide VLC using amplitude modulation at901
higher frequencies and achieve higher data rates while keeping the LED’s primary illumination function unaffected.902
VLC can be used for outdoor applications, where high power laser-based equipment provide transmission links,903
and for indoor application, where LEDs provide short distance transmission links. VLC is an energy-efficient904
technology, works on a wider range of unregulated frequency bands, shows high spatial reuse, and inherits security905
due to LOS. However, VLC is sensitive to sunlight and not able to work for a long range without LOS, and hence,906
confined coverage [54, 66]. The implementation of VLC has still some unanswered questions, such as: how VLC907
will work in a long range without LOS and will it work for backhaul data transfer in multihop?908
Fast caching. Caching is a way for storing temporary data for reducing data access from slow memory or the909
network. In a network, content caching5 is popular and answers the request while responding in place of the910
application servers as a proxy, thereby reducing the amount of hits that are directly sent to the ultimate backend911
server. In caching, three decisions are prominent as: what to cache, where to cache, and how to cache? The912
authors [112, 29] suggested that UEs will have enough memory in the future, and they can work as a cache for913
any other UE, since a small amount of popular data requires to be cached. Wang et al. [112] provided a caching914
mechanism based content centric networking (CCN), assuming 5G networks will include CCN-capable gateways,915
routers, and MBSs. CCN provides in-network data storage, also known as universal caching, at every node in the916
network. The cached data is uniquely identified at each node. Accordingly, a user can request for a particular917
content from the content cache of any device within the network, or the request is forwarded to the actual source of918
content storage.919
Interested readers may refer to [75, 29] for finding details about some of the above mentioned methodologies920
and technologies.921
7 Applications of 5G Networks922
The zero latency, high speed data transfer, and ubiquitous connectivity are the salient features of 5G networks that923
are expected to serve a wide range of applications and services. In this section, we enumerate the most prominent924
applications of 5G networks, as follows:925
Personal usages. This domain of 5G networks would be capable of supporting a wide range of UEs, from scalable926
to heterogeneous devices. Also the data demands (e.g., multimedia data, voice communication, and Web surfing),927
would be satisfied while keeping the QoS requirements.928
Virtualized homes. Due to C-RAN architectures, users may have only low cost UEs (e.g., set-top box for TVs929
and residential gateways for accessing the Internet) with services of the physical and data link layers. All the other930
higher layers’ applications may move to the cloud for universal access and outsourced computation services [55].931
Smart societies. It is an abstract term for connected virtualized homes, offices, and stores. Accordingly, every932
digital and electronic services/appliances, e.g., temperature maintenance, warning alarms, printers, LCDs, air933
conditioners, physical workout equipment, and door locks, would be interconnected in a way that the collaborative934
actions would enhance the user experience. Similarly, smart stores would assist in filtering out irrelevant product935
details, sale advertisements, and item suggestions on the go.936
5https://developer.akamai.com/stuff/Caching/Content_Caching.html
20
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
bursty data, the zero latency, scalability in terms of supporting devices, technologies and diverse applications,889
fast and reliable delivery of messages, and cost of devices. In addition, the development of efficient algorithms890
for location, time, group, priority, and multi-hop data transmission management for M2M communication is891
needed [13, 103]. Interested readers may refer to [22, 44, 82, 36, 96].892
Massive MIMO (mMIMO). mMIMO systems are also known as large-scale antenna systems, very large MIMO,893
hyper-MIMO, and full-dimension MIMO [71, 87]. mMIMO systems use antenna arrays with hundreds of antennas894
at MBS for simultaneously serving many UEs with a single antenna in identical time and frequency. Hence,895
expensive equipment are mounted on a MBS [81]. A mMIMO system relies on spatial multiplexing, which in turn896
relies on the channel knowledge at MBS, on both UL and DL channels. A mMIMO system reduces latency and897
energy, simplifies the MAC layer, shows robustness against intentional jamming, and increases the capacity due to898
spatial multiplexing.899
Visual Light Communication (VLC). VLC is a high-speed data transfer medium for short range LOS optical links900
in the future cellular networks [116]. The light-emitting diodes (LEDs) provide VLC using amplitude modulation at901
higher frequencies and achieve higher data rates while keeping the LED’s primary illumination function unaffected.902
VLC can be used for outdoor applications, where high power laser-based equipment provide transmission links,903
and for indoor application, where LEDs provide short distance transmission links. VLC is an energy-efficient904
technology, works on a wider range of unregulated frequency bands, shows high spatial reuse, and inherits security905
due to LOS. However, VLC is sensitive to sunlight and not able to work for a long range without LOS, and hence,906
confined coverage [54, 66]. The implementation of VLC has still some unanswered questions, such as: how VLC907
will work in a long range without LOS and will it work for backhaul data transfer in multihop?908
Fast caching. Caching is a way for storing temporary data for reducing data access from slow memory or the909
network. In a network, content caching5 is popular and answers the request while responding in place of the910
application servers as a proxy, thereby reducing the amount of hits that are directly sent to the ultimate backend911
server. In caching, three decisions are prominent as: what to cache, where to cache, and how to cache? The912
authors [112, 29] suggested that UEs will have enough memory in the future, and they can work as a cache for913
any other UE, since a small amount of popular data requires to be cached. Wang et al. [112] provided a caching914
mechanism based content centric networking (CCN), assuming 5G networks will include CCN-capable gateways,915
routers, and MBSs. CCN provides in-network data storage, also known as universal caching, at every node in the916
network. The cached data is uniquely identified at each node. Accordingly, a user can request for a particular917
content from the content cache of any device within the network, or the request is forwarded to the actual source of918
content storage.919
Interested readers may refer to [75, 29] for finding details about some of the above mentioned methodologies920
and technologies.921
7 Applications of 5G Networks922
The zero latency, high speed data transfer, and ubiquitous connectivity are the salient features of 5G networks that923
are expected to serve a wide range of applications and services. In this section, we enumerate the most prominent924
applications of 5G networks, as follows:925
Personal usages. This domain of 5G networks would be capable of supporting a wide range of UEs, from scalable926
to heterogeneous devices. Also the data demands (e.g., multimedia data, voice communication, and Web surfing),927
would be satisfied while keeping the QoS requirements.928
Virtualized homes. Due to C-RAN architectures, users may have only low cost UEs (e.g., set-top box for TVs929
and residential gateways for accessing the Internet) with services of the physical and data link layers. All the other930
higher layers’ applications may move to the cloud for universal access and outsourced computation services [55].931
Smart societies. It is an abstract term for connected virtualized homes, offices, and stores. Accordingly, every932
digital and electronic services/appliances, e.g., temperature maintenance, warning alarms, printers, LCDs, air933
conditioners, physical workout equipment, and door locks, would be interconnected in a way that the collaborative934
actions would enhance the user experience. Similarly, smart stores would assist in filtering out irrelevant product935
details, sale advertisements, and item suggestions on the go.936
5https://developer.akamai.com/stuff/Caching/Content_Caching.html
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Smart grids. The smart grids would decentralize the energy distribution and better analyze the energy937
consumption. This would allow the smart grids to improve efficiency and economic benefits. The 5G networks938
would allow a rapid and frequent statistical data observation, analysis, and fetching from remote sensors and would939
adjust the energy distribution accordingly.940
The tactile Internet. The tactile Internet improves the user experience in a virtual environment to an extent941
of only milliseconds of interaction latency. The futuristic applications such as automated vehicle platooning,942
self-organizing transportation, the ability to acquire a virtual sense for physically challenged patients, synchronized943
remote smart-grids, remote robotics, and image processing with customized/panoramic view would use the tactile944
Internet protocols [47].945
Automation. Self-driving vehicles would take place in the near future, and as a requirement, vehicles would946
communicate with each other in real-time. Moreover, they would communicate with other devices on the roads,947
homes, and offices with a requirement of almost zero latency. Hence, an interconnected vehicular environment948
would provide a safe and efficient integration with other information systems.949
Healthcare systems. A reliable, secure, and fast mobile communication can strengthen medical services, e.g.,950
frequent data transfer from patients’ body to the cloud or health care centers. Therefore, the relevant and urgent951
medical services could be predicted and delivered to the patients very fast.952
Logistics and tracking. The future mobile communication would also assist in inventory or package tracking using953
location based information systems. The most popular way would be to embed a radio frequency identification954
(RFID) tag and to provide a continuous connectivity irrespective of the geographic locations.955
Industrial usages. The zero latency property of 5G networks would help robots, sensors, drones, mobile devices,956
users, and data collector devices to have real-time data without any delay, which would help to manage and operate957
industrial functions quickly while preserving energy.958
8 Real Demonstrations and Testbeds for 5G Networks959
In this section, we present some real demonstrations and testbeds for 5G Networks. DOCOMO [8] is developing960
a real-time simulator for evaluating and simulating mMIMO, small-cell, and mmWave. The simulator has shown961
1000-times increase in the system capacity, and 90% users achieved 1Gbps data rate. DOCOMO also performed a962
real experiment in the year 2012, where data was uploaded at the speed of 10Gbps.963
Samsung performed data transmission experiment using mmWave at 28 GHz frequency band and achieved the964
world’s first highest data rate of 1.2Gbps on a vehicle running at the speed of> 100km/h. Further, when the vehicle965
was nearby a stop, the data transmission speed was achieved up to 7.5Gbps. In the experiment, the peak data rate966
was more than 30-times faster as compared to the state-of-the-art 4G technology. Samsung is also developing array967
antennas that have nearly zero-footprint and reconfigurable antenna modes [11]. Ericsson has achieved the speed968
of 5Gbps in a demonstration. 6 Huawei at University of Surrey in Guildford is developing a testbed, which would969
be used for developing methodologies, validating them, and verifying a C-RAN for an ultra-dense network.7
970
The European Commission funded METIS [6] has developed more than 140 technical components, including:971
air interface technologies, new waveforms, multiple access and MAC schemes, multi-antenna and mMIMO972
technologies, multi-hop communications, interference management, resource allocation schemes, mobility973
management, robustness enhancements, context aware approaches, D2D communication, dynamic reconfiguration974
enablers, and spectrum management technologies. They have also implemented some testbeds, as: three975
D2D communication related testbeds, one massive machine communications related testbed, and one related to976
waveform design. There are other projects funded by the European Commission working on the development of977
5G networks, as: 5GNOW (http://www.5gnow.eu/), TROPIC (www.ict-tropic.eu), MCN ( www.978
mobile-cloud-networking.eu ), COMBO ( www.ict-combo.eu), MOTO ( www.fp7-moto.eu),979
and PHYLAWS (www.phylaws-ict.org).980
Small-cells: Sprint, Verizon, and AT&T in the United States, Vodafone in Europe, and Softbank in Japan981
are developing femtocells [111]. Alcatel-Lucent, Huawei, and Nokia Siemens Networks has been involved in982
the development of plug-and-play SBSs [88]. mMIMO: TitanMIMO by Nutaq is a testbed for 5G mMIMO.983
6http://www.ericsson.com/research-blog/5g/ericsson-research-hits-5gbps-5g-labs-demo/
7http://www.fiercewireless.com/tech/story/huawei-invests-5g-test-bed-university-surrey/2014-11-04
21
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
Smart grids. The smart grids would decentralize the energy distribution and better analyze the energy937
consumption. This would allow the smart grids to improve efficiency and economic benefits. The 5G networks938
would allow a rapid and frequent statistical data observation, analysis, and fetching from remote sensors and would939
adjust the energy distribution accordingly.940
The tactile Internet. The tactile Internet improves the user experience in a virtual environment to an extent941
of only milliseconds of interaction latency. The futuristic applications such as automated vehicle platooning,942
self-organizing transportation, the ability to acquire a virtual sense for physically challenged patients, synchronized943
remote smart-grids, remote robotics, and image processing with customized/panoramic view would use the tactile944
Internet protocols [47].945
Automation. Self-driving vehicles would take place in the near future, and as a requirement, vehicles would946
communicate with each other in real-time. Moreover, they would communicate with other devices on the roads,947
homes, and offices with a requirement of almost zero latency. Hence, an interconnected vehicular environment948
would provide a safe and efficient integration with other information systems.949
Healthcare systems. A reliable, secure, and fast mobile communication can strengthen medical services, e.g.,950
frequent data transfer from patients’ body to the cloud or health care centers. Therefore, the relevant and urgent951
medical services could be predicted and delivered to the patients very fast.952
Logistics and tracking. The future mobile communication would also assist in inventory or package tracking using953
location based information systems. The most popular way would be to embed a radio frequency identification954
(RFID) tag and to provide a continuous connectivity irrespective of the geographic locations.955
Industrial usages. The zero latency property of 5G networks would help robots, sensors, drones, mobile devices,956
users, and data collector devices to have real-time data without any delay, which would help to manage and operate957
industrial functions quickly while preserving energy.958
8 Real Demonstrations and Testbeds for 5G Networks959
In this section, we present some real demonstrations and testbeds for 5G Networks. DOCOMO [8] is developing960
a real-time simulator for evaluating and simulating mMIMO, small-cell, and mmWave. The simulator has shown961
1000-times increase in the system capacity, and 90% users achieved 1Gbps data rate. DOCOMO also performed a962
real experiment in the year 2012, where data was uploaded at the speed of 10Gbps.963
Samsung performed data transmission experiment using mmWave at 28 GHz frequency band and achieved the964
world’s first highest data rate of 1.2Gbps on a vehicle running at the speed of> 100km/h. Further, when the vehicle965
was nearby a stop, the data transmission speed was achieved up to 7.5Gbps. In the experiment, the peak data rate966
was more than 30-times faster as compared to the state-of-the-art 4G technology. Samsung is also developing array967
antennas that have nearly zero-footprint and reconfigurable antenna modes [11]. Ericsson has achieved the speed968
of 5Gbps in a demonstration. 6 Huawei at University of Surrey in Guildford is developing a testbed, which would969
be used for developing methodologies, validating them, and verifying a C-RAN for an ultra-dense network.7
970
The European Commission funded METIS [6] has developed more than 140 technical components, including:971
air interface technologies, new waveforms, multiple access and MAC schemes, multi-antenna and mMIMO972
technologies, multi-hop communications, interference management, resource allocation schemes, mobility973
management, robustness enhancements, context aware approaches, D2D communication, dynamic reconfiguration974
enablers, and spectrum management technologies. They have also implemented some testbeds, as: three975
D2D communication related testbeds, one massive machine communications related testbed, and one related to976
waveform design. There are other projects funded by the European Commission working on the development of977
5G networks, as: 5GNOW (http://www.5gnow.eu/), TROPIC (www.ict-tropic.eu), MCN ( www.978
mobile-cloud-networking.eu ), COMBO ( www.ict-combo.eu), MOTO ( www.fp7-moto.eu),979
and PHYLAWS (www.phylaws-ict.org).980
Small-cells: Sprint, Verizon, and AT&T in the United States, Vodafone in Europe, and Softbank in Japan981
are developing femtocells [111]. Alcatel-Lucent, Huawei, and Nokia Siemens Networks has been involved in982
the development of plug-and-play SBSs [88]. mMIMO: TitanMIMO by Nutaq is a testbed for 5G mMIMO.983
6http://www.ericsson.com/research-blog/5g/ericsson-research-hits-5gbps-5g-labs-demo/
7http://www.fiercewireless.com/tech/story/huawei-invests-5g-test-bed-university-surrey/2014-11-04
21
Secure Best Marks with AI Grader
Need help grading? Try our AI Grader for instant feedback on your assignments.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
The TitanMIMO-4 testbed provides a realistic throughput by aggregating the entire mMIMO channel into a984
central baseband processing engine. 8 Lund University and National Instruments, Austin are also involved in the985
development of testbeds for mMIMO [106]. DOCOMO [8] is developing a real-time simulator for mMIMO.986
C-RANs: China Mobile Research Institute has developed C-RANs [14]. The European Commission supported987
iJoin ( http://www.ict-ijoin.eu/) is also involved in developing C-RAN architectures, especially,988
RANaaS [97]. IBM, Intel, Huawei, and ZTE are developing C-RANs [115]. OpenAirInterface (OAI) [89] is an989
open experimentation and prototyping platform created by the Mobile Communications Department at EURECOM.990
OAI used for two purposes as: C-RANs and M2M communication. Full duplex radio: groups at Stanford and991
Rice Universities are focusing on the development of full duplex radios [27]. CRNs: Vodafone Chair Mobile992
Communications Systems at the TU Dresden has created a testbed for studying CRNs in the future cellular993
networks [37]. 9 mmWave: DOCOMO [8], Huawei, the European Commission funded projects, and New York994
University [95] are developing methodologies and testbeds for mmWave and have successfully performed various995
experiments.996
9 Concluding Remarks997
In this survey, we discussed salient features, requirements, applications, and challenges behind the development998
of the next and the fifth generation (5G) of cellular mobile communication that is expected to provide very high999
speed data transfer and ubiquitous connectivity among different types of devices. We reviewed some architectures1000
for 5G networks based on the inclusion of small-cells, cognitive radio networks, device-to-device communication,1001
and cloud-based radio access networks. We find out that energy consumption by the infrastructure will be a major1002
concern in 5G networks, and hence, reviewed energy-efficient architectures. We figured out several open issues,1003
which may drive the future inventions and research, in all the architectures.1004
The development of new architectures is not only a solution to 5G networks; there will be a need for1005
handling other implementation issues in the context of users, e.g., interference removal, handoff management, QoS1006
guarantee, channel accessing, and in the context of infrastructures, e.g., load balancing. During our illustration,1007
we included several new techniques, e.g., full duplex radios, dense-deployment techniques, SIC, DUD, mmWave,1008
mMIMO, and VLC. We also discussed the current trends in research industries and academia in the context of 5G1009
networks based on real-testbed and experiments for 5G networks.1010
We conclude our discussion with a resonating notion that the designing of 5G infrastructure is still under1011
progress. The most prominent issues are enlisted below, also providing elegant solutions to these issues would1012
contribute in early deployment and long run growth of 5G networks.1013
• Security and privacy of devices, infrastructures, communication, and data transfer is yet to be explored. We1014
believe that the current solutions based on encryption would not work in the future due to a huge number of1015
devices. Intuitively, a solution that would use an authenticated certificate may be feasible [39].1016
• The development of network devices, infrastructures, and algorithms must be self-healing, self-configuring,1017
and self-optimizing to preform dynamic operations as per the need, for example, dynamic load balancing, QoS1018
guarantee, traffic management, and pooling of residual resources.1019
• The cloud computing is an attractive technique in the current trend for various applications. We have reviewed1020
C-RANs; however, the current solutions do not consider an impact of virtualization for backhaul data transfer,1021
the trust of the cloud, inter-cloud communication, ubiquitous service guarantee, and real-time performance1022
guarantee with zero-latency. Thus, the development of C-RANs must address the major question that how much1023
virtualization is good?1024
• Multi-RATs are attractive solutions to access different RATs. However, would it be possible for devices to use1025
more than one RAT at an identical time for uplink and downlink channels? Further, the network densification1026
must quantify that how much network density is good?1027
• Designs, developments, and usages of user devices, service-application models, and, especially, the network1028
devices must be affordable to cater the needs of overwhelming users, service providers, and network providers.1029
• Zero latency is a primary concern in most of the real-time applications and services, especially, in the tactile1030
Internet. However, all the existing architectures and implementations of 5G networks are far from achieving1031
8http://www.kanecomputing.co.uk/pdfs/Nutaq_TitanMIMO-4.pdf
9https://mns.ifn.et.tu-dresden.de/Chair/Pages/The-Chair.aspx
22
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
The TitanMIMO-4 testbed provides a realistic throughput by aggregating the entire mMIMO channel into a984
central baseband processing engine. 8 Lund University and National Instruments, Austin are also involved in the985
development of testbeds for mMIMO [106]. DOCOMO [8] is developing a real-time simulator for mMIMO.986
C-RANs: China Mobile Research Institute has developed C-RANs [14]. The European Commission supported987
iJoin ( http://www.ict-ijoin.eu/) is also involved in developing C-RAN architectures, especially,988
RANaaS [97]. IBM, Intel, Huawei, and ZTE are developing C-RANs [115]. OpenAirInterface (OAI) [89] is an989
open experimentation and prototyping platform created by the Mobile Communications Department at EURECOM.990
OAI used for two purposes as: C-RANs and M2M communication. Full duplex radio: groups at Stanford and991
Rice Universities are focusing on the development of full duplex radios [27]. CRNs: Vodafone Chair Mobile992
Communications Systems at the TU Dresden has created a testbed for studying CRNs in the future cellular993
networks [37]. 9 mmWave: DOCOMO [8], Huawei, the European Commission funded projects, and New York994
University [95] are developing methodologies and testbeds for mmWave and have successfully performed various995
experiments.996
9 Concluding Remarks997
In this survey, we discussed salient features, requirements, applications, and challenges behind the development998
of the next and the fifth generation (5G) of cellular mobile communication that is expected to provide very high999
speed data transfer and ubiquitous connectivity among different types of devices. We reviewed some architectures1000
for 5G networks based on the inclusion of small-cells, cognitive radio networks, device-to-device communication,1001
and cloud-based radio access networks. We find out that energy consumption by the infrastructure will be a major1002
concern in 5G networks, and hence, reviewed energy-efficient architectures. We figured out several open issues,1003
which may drive the future inventions and research, in all the architectures.1004
The development of new architectures is not only a solution to 5G networks; there will be a need for1005
handling other implementation issues in the context of users, e.g., interference removal, handoff management, QoS1006
guarantee, channel accessing, and in the context of infrastructures, e.g., load balancing. During our illustration,1007
we included several new techniques, e.g., full duplex radios, dense-deployment techniques, SIC, DUD, mmWave,1008
mMIMO, and VLC. We also discussed the current trends in research industries and academia in the context of 5G1009
networks based on real-testbed and experiments for 5G networks.1010
We conclude our discussion with a resonating notion that the designing of 5G infrastructure is still under1011
progress. The most prominent issues are enlisted below, also providing elegant solutions to these issues would1012
contribute in early deployment and long run growth of 5G networks.1013
• Security and privacy of devices, infrastructures, communication, and data transfer is yet to be explored. We1014
believe that the current solutions based on encryption would not work in the future due to a huge number of1015
devices. Intuitively, a solution that would use an authenticated certificate may be feasible [39].1016
• The development of network devices, infrastructures, and algorithms must be self-healing, self-configuring,1017
and self-optimizing to preform dynamic operations as per the need, for example, dynamic load balancing, QoS1018
guarantee, traffic management, and pooling of residual resources.1019
• The cloud computing is an attractive technique in the current trend for various applications. We have reviewed1020
C-RANs; however, the current solutions do not consider an impact of virtualization for backhaul data transfer,1021
the trust of the cloud, inter-cloud communication, ubiquitous service guarantee, and real-time performance1022
guarantee with zero-latency. Thus, the development of C-RANs must address the major question that how much1023
virtualization is good?1024
• Multi-RATs are attractive solutions to access different RATs. However, would it be possible for devices to use1025
more than one RAT at an identical time for uplink and downlink channels? Further, the network densification1026
must quantify that how much network density is good?1027
• Designs, developments, and usages of user devices, service-application models, and, especially, the network1028
devices must be affordable to cater the needs of overwhelming users, service providers, and network providers.1029
• Zero latency is a primary concern in most of the real-time applications and services, especially, in the tactile1030
Internet. However, all the existing architectures and implementations of 5G networks are far from achieving1031
8http://www.kanecomputing.co.uk/pdfs/Nutaq_TitanMIMO-4.pdf
9https://mns.ifn.et.tu-dresden.de/Chair/Pages/The-Chair.aspx
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
the zero latency. Therefore, the latest real-time and ultra-reliable network configurations must be improved to a1032
latency free environment.1033
Acknowledgements1034
Authors are thankful to Shlomi Dolev and anonymous referees for remarks that improved the quality of the paper.1035
References1036
[1] More than 50 billion connected devices, Ericsson, white paper, 2011. Available at: http://www.1037
akos-rs.si/files/Telekomunikacije/Digitalna_agenda/Internetni_protokol_Ipv6/1038
More-than-50-billion-connected-devices.pdf .1039
[2] Ericsson Mobility Report: On the Pulse of the Networked Society, Ericsson, 2015. Available at:http://www.ericsson.com/1040
res/docs/2015/ericsson-mobility-report-june-2015.pdf .1041
[3] 5G Rause cases and requirements, Nokia Networks, white paper. Available at: http://networks.nokia.com/sites/1042
default/files/document/5g_requirements_white_paper.pdf .1043
[4] The next generation of communication networks and services, The 5G Infrastructure Public Private Partnership1044
(5GPPP), European Commission, 2015. Available at: http://5g-ppp.eu/wp-content/uploads/2015/02/1045
5G-Vision-Brochure-v1.pdf .1046
[5] Analysis Understanding 5G: Perspectives on future technological advancements in mobile, GSMA intelligence, 2014. Available at:1047
https://gsmaintelligence.com/research/?file=141208-5g.pdf&download .1048
[6] METIS final project report, Deliberable D8.4, METIS, 2015. Available at: https://www.metis2020.com/wp-content/1049
uploads/deliverables/METIS_D8.4_v1.pdf .1050
[7] 5G is coming: Are you prepared?, Alcatel-Lucent, strategic white paper 2015. Available at:http://www2.alcatel-lucent.1051
com/landing/5g/.1052
[8] 5G Radio Access: Requirements, Concept and Technologies, DOCOMO, 5G paper. Available at: https://www.nttdocomo.1053
co.jp/english/binary/pdf/corporate/technology/whitepaper_5g/DOCOMO_5G_White_Paper.pdf .1054
[9] 5G: A Technology Vision, Huawei, white paper, 2013. Available at:http://www.huawei.com/5gwhitepaper/ .1055
[10] 5G: Vision for the next generation of connectivity, Qualcomm, 2015. Available at: https://www.qualcomm.com/media/1056
documents/files/whitepaper-5g-vision-for-the-next-generation-of-connectivity.pdf .1057
[11] 5G Vision, Samsung Electronics Co., Ltd, 2015. Available at: http://www.samsung.com/global/business-images/1058
insights/2015/Samsung-5G-Vision-0.pdf .1059
[12] Available at: http://research.microsoft.com/en-us/projects/spectrum/fcc_dynamic_spectrum_1060
access_noi_comments.pdf.1061
[13] Requirements and vision for NG-Wireless, Outlook: Visions and research directions for the Wireless World, Wireless1062
World Research Forum, 2011. Available at: http://www.wwrf.ch/files/wwrf/content/files/publications/1063
outlook/Outlook7.pdf.1064
[14] C-RAN: The Road Towards Green RAN, China Mobile Research Institute, white paper, 2011. Available at: http://labs.1065
chinamobile.com/cran/wp-content/uploads/CRAN_white_paper_v2_5_EN.pdf .1066
[15] P. K. Agyapong, M. Iwamura, D. Staehle, W. Kiess, and A. Benjebbour.Design considerations for a 5G network architecture. IEEE1067
Communications Magazine, 52(11):65–75, 2014.1068
[16] I. F. Akyildiz, W.-Y. Lee, and K. R. Chowdhury. CRAHNs: Cognitive radio ad hoc networks. Ad Hoc Networks, 7(5):810–836, 2009.1069
[17] I. F. Akyildiz, W.-Y. Lee, M. C. Vuran, and S. Mohanty. NeXt generation/dynamic spectrum access/cognitive radio wireless networks:1070
A survey. Computer Networks, 50(13):2127–2159, 2006.1071
[18] I. F. Akyildiz, P. Wang, and S.-C. Lin. SoftAir: A software defined networking architecture for 5G wireless systems. Computer1072
Networks, 85(0):1 – 18, 2015.1073
[19] I. Alsmadi and D. Xu. Security of software defined networks: A survey. Computers Security, 53:79–95, 2015.1074
[20] J. G. Andrews, S. Buzzi, W. Choi, S. V. Hanly, A. E. Lozano, A. C. K. Soong, and J. C. Zhang. What will 5G be? IEEE Journal on1075
Selected Areas in Communications, 32(6):1065–1082, 2014.1076
[21] J. G. Andrews, S. Singh, Q. Ye, X. Lin, and H. S. Dhillon. An overview of load balancing in hetnets: old myths and open problems.1077
IEEE Wireless Commun., 21(2):18–25, 2014.1078
[22] C. Anton-Haro and M. Dohler. Machine-to-machine (M2M) Communications: Architecture, Performance and Applications. Elsevier,1079
2014.1080
[23] M. Y. Arslan, K. Sundaresan, and S. Rangarajan. Software-defined networking in cellular radio access networks: potential and1081
challenges. IEEE Communications Magazine, 53(1):150–156, 2015.1082
[24] A. Asadi, Q. Wang, and V. Mancuso. A survey on device-to-device communication in cellular networks. IEEE Communications1083
Surveys and Tutorials, 16(4):1801–1819, 2014.1084
[25] K. Au, L. Zhang, H. Nikopour, E. Yi, A. Bayesteh, U. Vilaipornsawai, J. Ma, and P. Zhu. Uplink contention based SCMA for 5G1085
radio access. In IEEE GLOBECOM Workshops, pages 900–905, 2014.1086
[26] S. Barbarossa, S. Sardellitti, and P. D. Lorenzo. Communicating while computing: Distributed mobile cloud computing over 5G1087
heterogeneous networks. IEEE Signal Process. Mag., 31(6):45–55, 2014.1088
[27] D. Bharadia, E. McMilin, and S. Katti. Full duplex radios. In ACM SIGCOMM, pages 375–386, 2013.1089
[28] N. Bhushan and et al. Network densification: the dominant theme for wireless evolution into 5G. IEEE Communications Magazine,1090
52(2):82–89, 2014.1091
23
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
the zero latency. Therefore, the latest real-time and ultra-reliable network configurations must be improved to a1032
latency free environment.1033
Acknowledgements1034
Authors are thankful to Shlomi Dolev and anonymous referees for remarks that improved the quality of the paper.1035
References1036
[1] More than 50 billion connected devices, Ericsson, white paper, 2011. Available at: http://www.1037
akos-rs.si/files/Telekomunikacije/Digitalna_agenda/Internetni_protokol_Ipv6/1038
More-than-50-billion-connected-devices.pdf .1039
[2] Ericsson Mobility Report: On the Pulse of the Networked Society, Ericsson, 2015. Available at:http://www.ericsson.com/1040
res/docs/2015/ericsson-mobility-report-june-2015.pdf .1041
[3] 5G Rause cases and requirements, Nokia Networks, white paper. Available at: http://networks.nokia.com/sites/1042
default/files/document/5g_requirements_white_paper.pdf .1043
[4] The next generation of communication networks and services, The 5G Infrastructure Public Private Partnership1044
(5GPPP), European Commission, 2015. Available at: http://5g-ppp.eu/wp-content/uploads/2015/02/1045
5G-Vision-Brochure-v1.pdf .1046
[5] Analysis Understanding 5G: Perspectives on future technological advancements in mobile, GSMA intelligence, 2014. Available at:1047
https://gsmaintelligence.com/research/?file=141208-5g.pdf&download .1048
[6] METIS final project report, Deliberable D8.4, METIS, 2015. Available at: https://www.metis2020.com/wp-content/1049
uploads/deliverables/METIS_D8.4_v1.pdf .1050
[7] 5G is coming: Are you prepared?, Alcatel-Lucent, strategic white paper 2015. Available at:http://www2.alcatel-lucent.1051
com/landing/5g/.1052
[8] 5G Radio Access: Requirements, Concept and Technologies, DOCOMO, 5G paper. Available at: https://www.nttdocomo.1053
co.jp/english/binary/pdf/corporate/technology/whitepaper_5g/DOCOMO_5G_White_Paper.pdf .1054
[9] 5G: A Technology Vision, Huawei, white paper, 2013. Available at:http://www.huawei.com/5gwhitepaper/ .1055
[10] 5G: Vision for the next generation of connectivity, Qualcomm, 2015. Available at: https://www.qualcomm.com/media/1056
documents/files/whitepaper-5g-vision-for-the-next-generation-of-connectivity.pdf .1057
[11] 5G Vision, Samsung Electronics Co., Ltd, 2015. Available at: http://www.samsung.com/global/business-images/1058
insights/2015/Samsung-5G-Vision-0.pdf .1059
[12] Available at: http://research.microsoft.com/en-us/projects/spectrum/fcc_dynamic_spectrum_1060
access_noi_comments.pdf.1061
[13] Requirements and vision for NG-Wireless, Outlook: Visions and research directions for the Wireless World, Wireless1062
World Research Forum, 2011. Available at: http://www.wwrf.ch/files/wwrf/content/files/publications/1063
outlook/Outlook7.pdf.1064
[14] C-RAN: The Road Towards Green RAN, China Mobile Research Institute, white paper, 2011. Available at: http://labs.1065
chinamobile.com/cran/wp-content/uploads/CRAN_white_paper_v2_5_EN.pdf .1066
[15] P. K. Agyapong, M. Iwamura, D. Staehle, W. Kiess, and A. Benjebbour.Design considerations for a 5G network architecture. IEEE1067
Communications Magazine, 52(11):65–75, 2014.1068
[16] I. F. Akyildiz, W.-Y. Lee, and K. R. Chowdhury. CRAHNs: Cognitive radio ad hoc networks. Ad Hoc Networks, 7(5):810–836, 2009.1069
[17] I. F. Akyildiz, W.-Y. Lee, M. C. Vuran, and S. Mohanty. NeXt generation/dynamic spectrum access/cognitive radio wireless networks:1070
A survey. Computer Networks, 50(13):2127–2159, 2006.1071
[18] I. F. Akyildiz, P. Wang, and S.-C. Lin. SoftAir: A software defined networking architecture for 5G wireless systems. Computer1072
Networks, 85(0):1 – 18, 2015.1073
[19] I. Alsmadi and D. Xu. Security of software defined networks: A survey. Computers Security, 53:79–95, 2015.1074
[20] J. G. Andrews, S. Buzzi, W. Choi, S. V. Hanly, A. E. Lozano, A. C. K. Soong, and J. C. Zhang. What will 5G be? IEEE Journal on1075
Selected Areas in Communications, 32(6):1065–1082, 2014.1076
[21] J. G. Andrews, S. Singh, Q. Ye, X. Lin, and H. S. Dhillon. An overview of load balancing in hetnets: old myths and open problems.1077
IEEE Wireless Commun., 21(2):18–25, 2014.1078
[22] C. Anton-Haro and M. Dohler. Machine-to-machine (M2M) Communications: Architecture, Performance and Applications. Elsevier,1079
2014.1080
[23] M. Y. Arslan, K. Sundaresan, and S. Rangarajan. Software-defined networking in cellular radio access networks: potential and1081
challenges. IEEE Communications Magazine, 53(1):150–156, 2015.1082
[24] A. Asadi, Q. Wang, and V. Mancuso. A survey on device-to-device communication in cellular networks. IEEE Communications1083
Surveys and Tutorials, 16(4):1801–1819, 2014.1084
[25] K. Au, L. Zhang, H. Nikopour, E. Yi, A. Bayesteh, U. Vilaipornsawai, J. Ma, and P. Zhu. Uplink contention based SCMA for 5G1085
radio access. In IEEE GLOBECOM Workshops, pages 900–905, 2014.1086
[26] S. Barbarossa, S. Sardellitti, and P. D. Lorenzo. Communicating while computing: Distributed mobile cloud computing over 5G1087
heterogeneous networks. IEEE Signal Process. Mag., 31(6):45–55, 2014.1088
[27] D. Bharadia, E. McMilin, and S. Katti. Full duplex radios. In ACM SIGCOMM, pages 375–386, 2013.1089
[28] N. Bhushan and et al. Network densification: the dominant theme for wireless evolution into 5G. IEEE Communications Magazine,1090
52(2):82–89, 2014.1091
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
[29] F. Boccardi, R. W. Heath, A. E. Lozano, T. L. Marzetta, and P. Popovski. Five disruptive technology directions for 5G. IEEE1092
Communications Magazine, 52(2):74–80, 2014.1093
[30] Y. Cai, F. R. Yu, and S. Bu. Cloud computing meets mobile wireless communications in next generation cellular networks. IEEE1094
Network, 28(6):54–59, 2014.1095
[31] B. Cao, F. He, Y. Li, C. Wang, and W. Lang.Software defined virtual wireless network: framework and challenges. IEEE Network,1096
29(4):6–12, 2015.1097
[32] A. Checko, H. L. Christiansen, Y. Yan, L. Scolari, G. Kardaras, M. S. Berger, and L. Dittmann.Cloud RAN for mobile networks - A1098
technology overview. IEEE Communications Surveys and Tutorials, 17(1):405–426, 2015.1099
[33] B. Chen, J. Zheng, Y. Zhang, and H. Murata. SARA: A service-aware resource allocation scheme for device-to-device communication1100
underlaying cellular networks. In IEEE GLOBECOM, pages 4916–4921, 2014.1101
[34] S. Chen and J. Zhao. The requirements, challenges, and technologies for 5G of terrestrial mobile telecommunication. IEEE1102
Communications Magazine, 52(5):36–43, 2014.1103
[35] R. Chávez-Santiago and et al. 5G: The convergence of wireless communications. Wireless Personal Communications, pages 1–26,1104
2015.1105
[36] M. Condoluci, M. Dohler, G. Araniti, A. Molinaro, and K. Zheng. Toward 5G densenets: architectural advances for effective1106
machine-type communications over femtocells. IEEE Communications Magazine, 53(1):134–141, 2015.1107
[37] M. Danneberg, R. Datta, and G. Fettweis. Experimental testbed for dynamic spectrum access and sensing of 5G GFDM waveforms.1108
In IEEE 80th Vehicular Technology Conference, pages 1–5, 2014.1109
[38] N. Deng, W. Zhou, and M. Haenggi. A heterogeneous cellular network model with inter-tier dependence. In IEEE GLOBECOM,1110
pages 1522–1527, 2014.1111
[39] S. Dolev, Ł. Krzywiecki, N. Panwar, and M. Segal. Vehicle authentication via monolithically certified public key and attributes.1112
Wireless Networks, pages 1–18, 2015.1113
[40] X. Duan and X. Wang. Authentication handover and privacy protection in 5G hetnets using software-defined networking. IEEE1114
Communications Magazine, 53(4):28–35, 2015.1115
[41] H. ElSawy, E. Hossain, and D. I. Kim. HetNets with cognitive small cells: user offloading and distributed channel access techniques.1116
IEEE Communications Magazine, 51(6), 2013.1117
[42] H. Elshaer, F. Boccardi, M. Dohler, and R. Irmer. Downlink and uplink decoupling: A disruptive architectural design for 5G networks.1118
In IEEE GLOBECOM, pages 1798–1803, 2014.1119
[43] H. Elshaer, F. Boccardi, M. Dohler, and R. Irmer. Load & backhaul aware decoupled downlink/uplink access in 5G systems. CoRR,1120
abs/1410.6680, 2014.1121
[44] Z. Fan, R. J. Haines, and P. Kulkarni.M2M communications for e-health and smart grid: an industry and standard perspective.IEEE1122
Wireless Commun., 21(1):62–69, 2014.1123
[45] H. Farhadi, H. Lee, and A. Nakao. Software-defined networking: A survey. Computer Networks, 81:79–95, 2015.1124
[46] G. Fettweis and S. M. Alamouti. 5G: Personal mobile internet beyond what cellular did to telephony. IEEE Communications1125
Magazine, 52(2):140–145, 2014.1126
[47] G. P. Fettweis. The tactile Internet: Applications and challenges. Vehicular Technology Magazine, IEEE, 9(1):64–70, 2014.1127
[48] K. Gai, M. Qiu, L. Tao, and Y. Zhu. Intrusion detection techniques for mobile cloud computing in heterogeneous 5G. Security and1128
Communication Networks, 2015.1129
[49] D. Gale and L. S. Shapley. College admissions and the stability of marriage. American mathematical monthly, pages 9–15, 1962.1130
[50] L. Gavrilovska, V. Rakovic, and V. Atanasovski. Visions towards 5G: Technical requirements and potential enablers. Wireless1131
Personal Communications, pages 1–27, 2015.1132
[51] X. Ge, H. Cheng, M. Guizani, and T. Han. 5G wireless backhaul networks: challenges and research advances. IEEE Network,1133
28(6):6–11, 2014.1134
[52] F. Giust, L. Cominardi, and C. J. Bernardos. Distributed mobility management for future 5G networks: overview and analysis of1135
existing approaches. IEEE Communications Magazine, 53(1):142–149, 2015.1136
[53] L. Goratti, S. Savazzi, A. Parichehreh, and U. Spagnolini. Distributed load balancing for future 5G systems on-board high-speed1137
trains. In 5GU, pages 140–145, 2014.1138
[54] L. Grobe, A. Paraskevopoulos, J. Hilt, D. Schulz, F. Lassak, F. Hartlieb, C. Kottke, V. Jungnickel, and K. Langer.High-speed visible1139
light communication systems. IEEE Communications Magazine, 51(12):60–66, 2013.1140
[55] B. Han, V. Gopalakrishnan, L. Ji, and S. Lee. Network function virtualization: Challenges and opportunities for innovations. IEEE1141
Communications Magazine, 53(2):90–97, 2015.1142
[56] S. Han, C. I, Z. Xu, C. Pan, and Z. Pan. Full duplex: Coming into reality in 2020? In IEEE GLOBECOM, pages 4776–4781, 2014.1143
[57] E. J. Hernandez-Valencia, S. Izzo, and B. Polonsky. How will NFV/SDN transform service provider opex? IEEE Network,1144
29(3):79–108, 2015.1145
[58] T. D. Hoang, L. B. Le, and T. Le-Ngoc. Resource allocation for D2D communications under proportional fairness. In IEEE1146
GLOBECOM, pages 1259–1264, 2014.1147
[59] S. S. Hong, J. Brand, J. I. Choi, M. Jain, J. Mehlman, S. Katti, and P. Levis. Applications of self-interference cancellation in 5G and1148
beyond. IEEE Communications Magazine, 52(2):114–121, 2014.1149
[60] X. Hong, J. Wang, C. Wang, and J. Shi. Cognitive radio in 5G: a perspective on energy-spectral efficiency trade-off. IEEE1150
Communications Magazine, 52(7):46–53, 2014.1151
[61] E. Hossain, M. Rasti, H. Tabassum, and A. Abdelnasser. Evolution toward 5G multi-tier cellular wireless networks: An interference1152
management perspective. IEEE Wireless Commun., 21(3):118–127, 2014.1153
[62] R. Q. Hu and Y. Qian. An energy efficient and spectrum efficient wireless heterogeneous network framework for 5G systems. IEEE1154
Communications Magazine, 52(5):94–101, 2014.1155
[63] L. Huang, G. Zhu, and X. Du. Cognitive femtocell networks: an opportunistic spectrum access for future indoor wireless coverage.1156
IEEE Wireless Commun., 20(2):44–51, 2013.1157
24
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
[29] F. Boccardi, R. W. Heath, A. E. Lozano, T. L. Marzetta, and P. Popovski. Five disruptive technology directions for 5G. IEEE1092
Communications Magazine, 52(2):74–80, 2014.1093
[30] Y. Cai, F. R. Yu, and S. Bu. Cloud computing meets mobile wireless communications in next generation cellular networks. IEEE1094
Network, 28(6):54–59, 2014.1095
[31] B. Cao, F. He, Y. Li, C. Wang, and W. Lang.Software defined virtual wireless network: framework and challenges. IEEE Network,1096
29(4):6–12, 2015.1097
[32] A. Checko, H. L. Christiansen, Y. Yan, L. Scolari, G. Kardaras, M. S. Berger, and L. Dittmann.Cloud RAN for mobile networks - A1098
technology overview. IEEE Communications Surveys and Tutorials, 17(1):405–426, 2015.1099
[33] B. Chen, J. Zheng, Y. Zhang, and H. Murata. SARA: A service-aware resource allocation scheme for device-to-device communication1100
underlaying cellular networks. In IEEE GLOBECOM, pages 4916–4921, 2014.1101
[34] S. Chen and J. Zhao. The requirements, challenges, and technologies for 5G of terrestrial mobile telecommunication. IEEE1102
Communications Magazine, 52(5):36–43, 2014.1103
[35] R. Chávez-Santiago and et al. 5G: The convergence of wireless communications. Wireless Personal Communications, pages 1–26,1104
2015.1105
[36] M. Condoluci, M. Dohler, G. Araniti, A. Molinaro, and K. Zheng. Toward 5G densenets: architectural advances for effective1106
machine-type communications over femtocells. IEEE Communications Magazine, 53(1):134–141, 2015.1107
[37] M. Danneberg, R. Datta, and G. Fettweis. Experimental testbed for dynamic spectrum access and sensing of 5G GFDM waveforms.1108
In IEEE 80th Vehicular Technology Conference, pages 1–5, 2014.1109
[38] N. Deng, W. Zhou, and M. Haenggi. A heterogeneous cellular network model with inter-tier dependence. In IEEE GLOBECOM,1110
pages 1522–1527, 2014.1111
[39] S. Dolev, Ł. Krzywiecki, N. Panwar, and M. Segal. Vehicle authentication via monolithically certified public key and attributes.1112
Wireless Networks, pages 1–18, 2015.1113
[40] X. Duan and X. Wang. Authentication handover and privacy protection in 5G hetnets using software-defined networking. IEEE1114
Communications Magazine, 53(4):28–35, 2015.1115
[41] H. ElSawy, E. Hossain, and D. I. Kim. HetNets with cognitive small cells: user offloading and distributed channel access techniques.1116
IEEE Communications Magazine, 51(6), 2013.1117
[42] H. Elshaer, F. Boccardi, M. Dohler, and R. Irmer. Downlink and uplink decoupling: A disruptive architectural design for 5G networks.1118
In IEEE GLOBECOM, pages 1798–1803, 2014.1119
[43] H. Elshaer, F. Boccardi, M. Dohler, and R. Irmer. Load & backhaul aware decoupled downlink/uplink access in 5G systems. CoRR,1120
abs/1410.6680, 2014.1121
[44] Z. Fan, R. J. Haines, and P. Kulkarni.M2M communications for e-health and smart grid: an industry and standard perspective.IEEE1122
Wireless Commun., 21(1):62–69, 2014.1123
[45] H. Farhadi, H. Lee, and A. Nakao. Software-defined networking: A survey. Computer Networks, 81:79–95, 2015.1124
[46] G. Fettweis and S. M. Alamouti. 5G: Personal mobile internet beyond what cellular did to telephony. IEEE Communications1125
Magazine, 52(2):140–145, 2014.1126
[47] G. P. Fettweis. The tactile Internet: Applications and challenges. Vehicular Technology Magazine, IEEE, 9(1):64–70, 2014.1127
[48] K. Gai, M. Qiu, L. Tao, and Y. Zhu. Intrusion detection techniques for mobile cloud computing in heterogeneous 5G. Security and1128
Communication Networks, 2015.1129
[49] D. Gale and L. S. Shapley. College admissions and the stability of marriage. American mathematical monthly, pages 9–15, 1962.1130
[50] L. Gavrilovska, V. Rakovic, and V. Atanasovski. Visions towards 5G: Technical requirements and potential enablers. Wireless1131
Personal Communications, pages 1–27, 2015.1132
[51] X. Ge, H. Cheng, M. Guizani, and T. Han. 5G wireless backhaul networks: challenges and research advances. IEEE Network,1133
28(6):6–11, 2014.1134
[52] F. Giust, L. Cominardi, and C. J. Bernardos. Distributed mobility management for future 5G networks: overview and analysis of1135
existing approaches. IEEE Communications Magazine, 53(1):142–149, 2015.1136
[53] L. Goratti, S. Savazzi, A. Parichehreh, and U. Spagnolini. Distributed load balancing for future 5G systems on-board high-speed1137
trains. In 5GU, pages 140–145, 2014.1138
[54] L. Grobe, A. Paraskevopoulos, J. Hilt, D. Schulz, F. Lassak, F. Hartlieb, C. Kottke, V. Jungnickel, and K. Langer.High-speed visible1139
light communication systems. IEEE Communications Magazine, 51(12):60–66, 2013.1140
[55] B. Han, V. Gopalakrishnan, L. Ji, and S. Lee. Network function virtualization: Challenges and opportunities for innovations. IEEE1141
Communications Magazine, 53(2):90–97, 2015.1142
[56] S. Han, C. I, Z. Xu, C. Pan, and Z. Pan. Full duplex: Coming into reality in 2020? In IEEE GLOBECOM, pages 4776–4781, 2014.1143
[57] E. J. Hernandez-Valencia, S. Izzo, and B. Polonsky. How will NFV/SDN transform service provider opex? IEEE Network,1144
29(3):79–108, 2015.1145
[58] T. D. Hoang, L. B. Le, and T. Le-Ngoc. Resource allocation for D2D communications under proportional fairness. In IEEE1146
GLOBECOM, pages 1259–1264, 2014.1147
[59] S. S. Hong, J. Brand, J. I. Choi, M. Jain, J. Mehlman, S. Katti, and P. Levis. Applications of self-interference cancellation in 5G and1148
beyond. IEEE Communications Magazine, 52(2):114–121, 2014.1149
[60] X. Hong, J. Wang, C. Wang, and J. Shi. Cognitive radio in 5G: a perspective on energy-spectral efficiency trade-off. IEEE1150
Communications Magazine, 52(7):46–53, 2014.1151
[61] E. Hossain, M. Rasti, H. Tabassum, and A. Abdelnasser. Evolution toward 5G multi-tier cellular wireless networks: An interference1152
management perspective. IEEE Wireless Commun., 21(3):118–127, 2014.1153
[62] R. Q. Hu and Y. Qian. An energy efficient and spectrum efficient wireless heterogeneous network framework for 5G systems. IEEE1154
Communications Magazine, 52(5):94–101, 2014.1155
[63] L. Huang, G. Zhu, and X. Du. Cognitive femtocell networks: an opportunistic spectrum access for future indoor wireless coverage.1156
IEEE Wireless Commun., 20(2):44–51, 2013.1157
24
Paraphrase This Document
Need a fresh take? Get an instant paraphrase of this document with our AI Paraphraser
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
[64] C. I, C. Rowell, S. Han, Z. Xu, G. Li, and Z. Pan. Toward green and soft: a 5G perspective. IEEE Communications Magazine,1158
52(2):66–73, 2014.1159
[65] C. Jiang, Y. Chen, K. J. R. Liu, and Y. Ren. Cognitive femtocell market: How to price? In IEEE GLOBECOM, pages 4514–4519,1160
2014.1161
[66] A. Jovicic, J. Li, and T. Richardson. Visible light communication: opportunities, challenges and the path to market. IEEE1162
Communications Magazine, 51(12):26–32, 2013.1163
[67] R. Kantola, J. L. Santos, and N. Beijar. Policy-based communications for 5G mobile with customer edge switching. Wiley Security1164
Comm. Networks, 2015.1165
[68] H. Kim and N. Feamster. Improving network management with software defined networking. IEEE Communications Magazine,1166
51(2):114–119, 2013.1167
[69] J. Kim and A. F. Molisch. Quality-aware millimeter-wave device-to-device multi-hop routing for 5G cellular networks. In IEEE1168
International Conference on Communications, ICC, pages 5251–5256, 2014.1169
[70] D. Kreutz, F. M. V. Ramos, P. J. E. Veríssimo, C. E. Rothenberg, S. Azodolmolky, and S. Uhlig. Software-defined networking: A1170
comprehensive survey. Proceedings of the IEEE, 103(1):14–76, 2015.1171
[71] E. G. Larsson, O. Edfors, F. Tufvesson, and T. L. Marzetta. Massive MIMO for next generation wireless systems. IEEE1172
Communications Magazine, 52(2):186–195, 2014.1173
[72] C. Lee and C. Shih. Coverage analysis of cognitive femtocell networks. IEEE Wireless Commun. Letters, 3(2):177–180, 2014.1174
[73] A. Leu, B. Mark, and S. Tang. Analysis of handoff interference and outage along arbitrary trajectories in cellular networks. IEEE1175
Transactions on Wireless Communications, 7(9):3597–3607, 2008.1176
[74] P. Li, S. Guo, T. Miyazaki, and W. Zhuang. Fine-grained resource allocation for cooperative device-to-device communication in1177
cellular networks. IEEE Wireless Commun., 21(5):35–40, 2014.1178
[75] Q. Li, H. Niu, A. Papathanassiou, and G. Wu. 5G network capacity: Key elements and technologies. Vehicular Technology Magazine,1179
IEEE, 9(1):71–78, March 2014.1180
[76] Y. Li, B. Kaur, and B. Andersen. Denial of service prevention for 5G. Wireless Personal Communications, 57(3):365–376, 2011.1181
[77] Y. Li, T. Wu, P. Hui, D. Jin, and S. Chen. Social-aware D2D communications: qualitative insights and quantitative analysis. IEEE1182
Communications Magazine, 52(6):150–158, 2014.1183
[78] J. Liu, T. Zhao, S. Zhou, Y. Cheng, and Z. Niu. CONCERT: a cloud-based architecture for next-generation cellular systems. IEEE1184
Wireless Commun., 21(6):14–22, 2014.1185
[79] J. K. Liu, M. H. Au, W. Susilo, K. Liang, R. Lu, and B. Srinivasan. Secure sharing and searching for real-time video data in mobile1186
cloud. IEEE Network, 29(2):46–50, 2015.1187
[80] Y. Liu, Y. Zhang, R. Yu, and S. Xie. Integrated energy and spectrum harvesting for 5G wireless communications. IEEE Network,1188
29(3):75–81, 2015.1189
[81] L. Lu, G. Y. Li, A. L. Swindlehurst, A. E. Ashikhmin, and R. Zhang. An overview of massive MIMO: benefits and challenges. J.1190
Sel. Topics Signal Processing, 8(5):742–758, 2014.1191
[82] G. C. Madueño, C. Stefanovic, and P. Popovski. Reliable reporting for massive M2M communications with periodic resource pooling.1192
IEEE Wireless Commun. Letters, 3(4):429–432, 2014.1193
[83] C. X. Mavromoustakis and et al. An energy-aware scheme for efficient spectrum utilization in a 5G mobile cognitive radio network1194
architecture. Telecommunication Systems, 59(1):63–75, 2015.1195
[84] N. McKeown, T. Anderson, H. Balakrishnan, G. M. Parulkar, L. L. Peterson, J. Rexford, S. Shenker, and J. S. Turner. OpenFlow:1196
enabling innovation in campus networks. Computer Communication Review, 38(2):69–74, 2008.1197
[85] P. Mell and T. Grance. The NIST definition of cloud computing. 2011.1198
[86] W. Nam, D. Bai, J. Lee, and I. Kang. Advanced interference management for 5G cellular networks. IEEE Communications Magazine,1199
52(5):52–60, 2014.1200
[87] Y. Nam, B. L. Ng, K. Sayana, Y. Li, J. Zhang, Y. Kim, and J. Lee. Full-dimension MIMO (FD-MIMO) for next generation cellular1201
technology. IEEE Communications Magazine, 51(6), 2013.1202
[88] W. Ni, I. B. Collings, X. Wang, and R. P. Liu. Multi-hop point-to-point FDD wireless backhaul for mobile small cells. IEEE Wireless1203
Commun., 21(4):88–96, 2014.1204
[89] N. Nikaein, M. K. Marina, S. Manickam, A. Dawson, R. Knopp, and C. Bonnet. OpenAirInterface: A flexible platform for 5G1205
research. Computer Communication Review, 44(5):33–38, 2014.1206
[90] H. Nikopour, E. Yi, A. Bayesteh, K. Au, M. Hawryluck, H. Baligh, and J. Ma. SCMA for downlink multiple access of 5G wireless1207
networks. In IEEE GLOBECOM, pages 3940–3945, 2014.1208
[91] Y. Niu, Y. Li, D. Jin, L. Su, and A. V. Vasilakos.A survey of millimeter wave (mmWave) communications for 5G: Opportunities and1209
challenges. CoRR, abs/1502.07228, 2015.1210
[92] E. Oh, K. Son, and B. Krishnamachari. Dynamic base station switching-on/off strategies for green cellular networks. IEEE1211
Transactions on Wireless Communications, 12(5):2126–2136, 2013.1212
[93] A. Orsino, G. Araniti, A. Molinaro, and A. Iera. Effective RAT selection approach for 5G dense wireless networks. CoRR,1213
abs/1502.01482, 2015.1214
[94] A. Osseiran and et al. Scenarios for 5G mobile and wireless communications: the vision of the METIS project. IEEE Communications1215
Magazine, 52(5):26–35, 2014.1216
[95] T. S. Rappaport, S. Sun, R. Mayzus, H. Zhao, Y. Azar, K. Wang, G. N. Wong, J. K. Schulz, M. Samimi, and F. G. Jr. Millimeter wave1217
mobile communications for 5G cellular: It will work! IEEE Access, 1:335–349, 2013.1218
[96] R. Ratasuk, A. Prasad, Z. Li, A. Ghosh, and M. A. Uusitalo. Recent advancements in M2M communications in 4G networks and1219
evolution towards 5G. In ICIN 2015, pages 52–57, 2015.1220
[97] P. Rost, C. J. Bernardos, A. D. Domenico, M. D. Girolamo, M. Lalam, A. Mäder, D. Sabella, and D. Wübben. Cloud technologies1221
for flexible 5G radio access networks. IEEE Communications Magazine, 52(5):68–76, 2014.1222
[98] S. Sharma and A. K. Singh. On detecting termination in cognitive radio networks. Int. Journal of Network Management,1223
25
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
[64] C. I, C. Rowell, S. Han, Z. Xu, G. Li, and Z. Pan. Toward green and soft: a 5G perspective. IEEE Communications Magazine,1158
52(2):66–73, 2014.1159
[65] C. Jiang, Y. Chen, K. J. R. Liu, and Y. Ren. Cognitive femtocell market: How to price? In IEEE GLOBECOM, pages 4514–4519,1160
2014.1161
[66] A. Jovicic, J. Li, and T. Richardson. Visible light communication: opportunities, challenges and the path to market. IEEE1162
Communications Magazine, 51(12):26–32, 2013.1163
[67] R. Kantola, J. L. Santos, and N. Beijar. Policy-based communications for 5G mobile with customer edge switching. Wiley Security1164
Comm. Networks, 2015.1165
[68] H. Kim and N. Feamster. Improving network management with software defined networking. IEEE Communications Magazine,1166
51(2):114–119, 2013.1167
[69] J. Kim and A. F. Molisch. Quality-aware millimeter-wave device-to-device multi-hop routing for 5G cellular networks. In IEEE1168
International Conference on Communications, ICC, pages 5251–5256, 2014.1169
[70] D. Kreutz, F. M. V. Ramos, P. J. E. Veríssimo, C. E. Rothenberg, S. Azodolmolky, and S. Uhlig. Software-defined networking: A1170
comprehensive survey. Proceedings of the IEEE, 103(1):14–76, 2015.1171
[71] E. G. Larsson, O. Edfors, F. Tufvesson, and T. L. Marzetta. Massive MIMO for next generation wireless systems. IEEE1172
Communications Magazine, 52(2):186–195, 2014.1173
[72] C. Lee and C. Shih. Coverage analysis of cognitive femtocell networks. IEEE Wireless Commun. Letters, 3(2):177–180, 2014.1174
[73] A. Leu, B. Mark, and S. Tang. Analysis of handoff interference and outage along arbitrary trajectories in cellular networks. IEEE1175
Transactions on Wireless Communications, 7(9):3597–3607, 2008.1176
[74] P. Li, S. Guo, T. Miyazaki, and W. Zhuang. Fine-grained resource allocation for cooperative device-to-device communication in1177
cellular networks. IEEE Wireless Commun., 21(5):35–40, 2014.1178
[75] Q. Li, H. Niu, A. Papathanassiou, and G. Wu. 5G network capacity: Key elements and technologies. Vehicular Technology Magazine,1179
IEEE, 9(1):71–78, March 2014.1180
[76] Y. Li, B. Kaur, and B. Andersen. Denial of service prevention for 5G. Wireless Personal Communications, 57(3):365–376, 2011.1181
[77] Y. Li, T. Wu, P. Hui, D. Jin, and S. Chen. Social-aware D2D communications: qualitative insights and quantitative analysis. IEEE1182
Communications Magazine, 52(6):150–158, 2014.1183
[78] J. Liu, T. Zhao, S. Zhou, Y. Cheng, and Z. Niu. CONCERT: a cloud-based architecture for next-generation cellular systems. IEEE1184
Wireless Commun., 21(6):14–22, 2014.1185
[79] J. K. Liu, M. H. Au, W. Susilo, K. Liang, R. Lu, and B. Srinivasan. Secure sharing and searching for real-time video data in mobile1186
cloud. IEEE Network, 29(2):46–50, 2015.1187
[80] Y. Liu, Y. Zhang, R. Yu, and S. Xie. Integrated energy and spectrum harvesting for 5G wireless communications. IEEE Network,1188
29(3):75–81, 2015.1189
[81] L. Lu, G. Y. Li, A. L. Swindlehurst, A. E. Ashikhmin, and R. Zhang. An overview of massive MIMO: benefits and challenges. J.1190
Sel. Topics Signal Processing, 8(5):742–758, 2014.1191
[82] G. C. Madueño, C. Stefanovic, and P. Popovski. Reliable reporting for massive M2M communications with periodic resource pooling.1192
IEEE Wireless Commun. Letters, 3(4):429–432, 2014.1193
[83] C. X. Mavromoustakis and et al. An energy-aware scheme for efficient spectrum utilization in a 5G mobile cognitive radio network1194
architecture. Telecommunication Systems, 59(1):63–75, 2015.1195
[84] N. McKeown, T. Anderson, H. Balakrishnan, G. M. Parulkar, L. L. Peterson, J. Rexford, S. Shenker, and J. S. Turner. OpenFlow:1196
enabling innovation in campus networks. Computer Communication Review, 38(2):69–74, 2008.1197
[85] P. Mell and T. Grance. The NIST definition of cloud computing. 2011.1198
[86] W. Nam, D. Bai, J. Lee, and I. Kang. Advanced interference management for 5G cellular networks. IEEE Communications Magazine,1199
52(5):52–60, 2014.1200
[87] Y. Nam, B. L. Ng, K. Sayana, Y. Li, J. Zhang, Y. Kim, and J. Lee. Full-dimension MIMO (FD-MIMO) for next generation cellular1201
technology. IEEE Communications Magazine, 51(6), 2013.1202
[88] W. Ni, I. B. Collings, X. Wang, and R. P. Liu. Multi-hop point-to-point FDD wireless backhaul for mobile small cells. IEEE Wireless1203
Commun., 21(4):88–96, 2014.1204
[89] N. Nikaein, M. K. Marina, S. Manickam, A. Dawson, R. Knopp, and C. Bonnet. OpenAirInterface: A flexible platform for 5G1205
research. Computer Communication Review, 44(5):33–38, 2014.1206
[90] H. Nikopour, E. Yi, A. Bayesteh, K. Au, M. Hawryluck, H. Baligh, and J. Ma. SCMA for downlink multiple access of 5G wireless1207
networks. In IEEE GLOBECOM, pages 3940–3945, 2014.1208
[91] Y. Niu, Y. Li, D. Jin, L. Su, and A. V. Vasilakos.A survey of millimeter wave (mmWave) communications for 5G: Opportunities and1209
challenges. CoRR, abs/1502.07228, 2015.1210
[92] E. Oh, K. Son, and B. Krishnamachari. Dynamic base station switching-on/off strategies for green cellular networks. IEEE1211
Transactions on Wireless Communications, 12(5):2126–2136, 2013.1212
[93] A. Orsino, G. Araniti, A. Molinaro, and A. Iera. Effective RAT selection approach for 5G dense wireless networks. CoRR,1213
abs/1502.01482, 2015.1214
[94] A. Osseiran and et al. Scenarios for 5G mobile and wireless communications: the vision of the METIS project. IEEE Communications1215
Magazine, 52(5):26–35, 2014.1216
[95] T. S. Rappaport, S. Sun, R. Mayzus, H. Zhao, Y. Azar, K. Wang, G. N. Wong, J. K. Schulz, M. Samimi, and F. G. Jr. Millimeter wave1217
mobile communications for 5G cellular: It will work! IEEE Access, 1:335–349, 2013.1218
[96] R. Ratasuk, A. Prasad, Z. Li, A. Ghosh, and M. A. Uusitalo. Recent advancements in M2M communications in 4G networks and1219
evolution towards 5G. In ICIN 2015, pages 52–57, 2015.1220
[97] P. Rost, C. J. Bernardos, A. D. Domenico, M. D. Girolamo, M. Lalam, A. Mäder, D. Sabella, and D. Wübben. Cloud technologies1221
for flexible 5G radio access networks. IEEE Communications Magazine, 52(5):68–76, 2014.1222
[98] S. Sharma and A. K. Singh. On detecting termination in cognitive radio networks. Int. Journal of Network Management,1223
25
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
24(6):499–527, 2014.1224
[99] S. Singh and J. G. Andrews. Joint resource partitioning and offloading in heterogeneous cellular networks. IEEE Transactions on1225
Wireless Communications, 13(2):888–901, 2014.1226
[100] K. Smiljkovikj, P. Popovski, and L. Gavrilovska. Analysis of the decoupled access for downlink and uplink in wireless heterogeneous1227
networks. IEEE Wireless Commun. Letters, 4(2):173–176, 2015.1228
[101] I. K. Son, S. Mao, Y. Li, M. Chen, M. X. Gong, and T. T. S. Rappaport. Frame-based medium access control for 5G wireless1229
networks. Mobile Networks and Applications, pages 1–10, 2015.1230
[102] H. Song, X. Fang, and L. Yan. Handover scheme for 5G C/U plane split heterogeneous network in high-speed railway. IEEE T.1231
Vehicular Technology, 63(9):4633–4646, 2014.1232
[103] P. F. Stenumgaard, J. Chilo, J. Ferrer-Coll, and P. Ängskog. Challenges and conditions for wireless machine-to-machine1233
communications in industrial environments. IEEE Communications Magazine, 51(6), 2013.1234
[104] M. N. Tehrani, M. Uysal, and H. Yanikomeroglu. Device-to-device communication in 5G cellular networks: challenges, solutions,1235
and future directions. IEEE Communications Magazine, 52(5):86–92, 2014.1236
[105] N. Ulltveit-Moe, V. A. Oleshchuk, and G. M. Køien. Location-aware mobile intrusion detection with enhanced privacy in a 5G1237
context. Wireless Personal Communications, 57(3):317–338, 2011.1238
[106] J. Vieira, S. Malkowsky, K. Nieman, Z. Miers, N. Kundargi, L. Liu, I. Wong, V. Öwall, O. Edfors, and F. Tufvesson. A flexible1239
100-antenna testbed for massive MIMO. In IEEE GLOBECOM Workshops, pages 287–293, 2014.1240
[107] C. Wang and et al. Cellular architecture and key technologies for 5G wireless communication networks. IEEE Communications1241
Magazine, 52(2):122–130, 2014.1242
[108] R. Wang, H. Hu, and X. Yang.Potentials and challenges of C-RAN supporting multi-rats toward 5G mobile networks.IEEE Access,1243
2:1187–1195, 2014.1244
[109] W. Wang and V. K. N. Lau. Delay-aware cross-layer design for device-to-device communications in future cellular systems. IEEE1245
Communications Magazine, 52(6):133–139, 2014.1246
[110] W. Wang, G. Yu, and A. Huang.Cognitive radio enhanced interference coordination for femtocell networks. IEEE Communications1247
Magazine, 51(6), 2013.1248
[111] W. Wang and Q. Zhang. Local cooperation architecture for self-healing femtocell networks. IEEE Wireless Commun., 21(2):42–49,1249
2014.1250
[112] X. Wang, M. Chen, T. Taleb, A. Ksentini, and V. C. M. Leung.Cache in the air: exploiting content caching and delivery techniques1251
for 5G systems. IEEE Communications Magazine, 52(2):131–139, 2014.1252
[113] L. Wei, R. Q. Hu, Y. Qian, and G. Wu. Enable device-to-device communications underlaying cellular networks: challenges and1253
research aspects. IEEE Communications Magazine, 52(6):90–96, 2014.1254
[114] G. Wu, C. Yang, S. Li, and G. Y. Li. Recent advances in energy-efficient networks and their application in 5G systems. IEEE Wireless1255
Commun., 22(2):145–151, 2015.1256
[115] J. Wu, Z. Zhang, Y. Hong, and Y. Wen.Cloud radio access network (C-RAN): a primer. IEEE Network, 29(1):35–41, 2015.1257
[116] S. Wu and H. Wang. Visible light communications for 5G wireless networking systems: from fixed to mobile communications. IEEE1258
Network, 28(6):41–45, 2014.1259
[117] J. Xie, D. Guo, Z. Hu, T. Qu, and P. Lv. Control plane of software defined networks: A survey. Computer Communications, 67:1–10,1260
2015.1261
[118] N. Yang, L. Wang, G. Geraci, M. Elkashlan, J. Yuan, and M. D. Renzo. Safeguarding 5G wireless communication networks using1262
physical layer security. IEEE Communications Magazine, 53(4):20–27, 2015.1263
[119] V. Yazici, U. C. Kozat, and M. O. Sunay. A new control plane for 5G network architecture with a case study on unified handoff,1264
mobility, and routing management. IEEE Communications Magazine, 52(11):76–85, 2014.1265
[120] X. Zhang, W. Cheng, and H. Zhang. Heterogeneous statistical QoS provisioning over 5G mobile wireless networks. IEEE Network,1266
28(6):46–53, 2014.1267
[121] X. Zhang, J. Zhang, W. Wang, Y. Zhang, C. I, Z. Pan, G. Li, and Y. Chen. Macro-assisted data-only carrier for 5G green cellular1268
systems. IEEE Communications Magazine, 53(5):223–231, 2015.1269
[122] Z. Zhang, X. Chai, K. Long, A. V. Vasilakos, and L. Hanzo. Full duplex techniques for 5G networks: self-interference cancellation,1270
protocol design, and relay selection. IEEE Communications Magazine, 53(5):128–137, 2015.1271
[123] X. Zhou, Z. Zhao, R. Li, Y. Zhou, T. Chen, Z. Niu, and H. Zhang. Toward 5G: when explosive bursts meet soft cloud. IEEE Network,1272
28(6):12–17, 2014.1273
[124] Y. Zhou, D. Li, H. Wang, A. Yang, and S. Guo.QoS-aware energy-efficient optimization for massive MIMO systems in 5G.In Sixth1274
International Conference on Wireless Communications and Signal Processing, WCSP, pages 1–5, 2014.1275
26
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
24(6):499–527, 2014.1224
[99] S. Singh and J. G. Andrews. Joint resource partitioning and offloading in heterogeneous cellular networks. IEEE Transactions on1225
Wireless Communications, 13(2):888–901, 2014.1226
[100] K. Smiljkovikj, P. Popovski, and L. Gavrilovska. Analysis of the decoupled access for downlink and uplink in wireless heterogeneous1227
networks. IEEE Wireless Commun. Letters, 4(2):173–176, 2015.1228
[101] I. K. Son, S. Mao, Y. Li, M. Chen, M. X. Gong, and T. T. S. Rappaport. Frame-based medium access control for 5G wireless1229
networks. Mobile Networks and Applications, pages 1–10, 2015.1230
[102] H. Song, X. Fang, and L. Yan. Handover scheme for 5G C/U plane split heterogeneous network in high-speed railway. IEEE T.1231
Vehicular Technology, 63(9):4633–4646, 2014.1232
[103] P. F. Stenumgaard, J. Chilo, J. Ferrer-Coll, and P. Ängskog. Challenges and conditions for wireless machine-to-machine1233
communications in industrial environments. IEEE Communications Magazine, 51(6), 2013.1234
[104] M. N. Tehrani, M. Uysal, and H. Yanikomeroglu. Device-to-device communication in 5G cellular networks: challenges, solutions,1235
and future directions. IEEE Communications Magazine, 52(5):86–92, 2014.1236
[105] N. Ulltveit-Moe, V. A. Oleshchuk, and G. M. Køien. Location-aware mobile intrusion detection with enhanced privacy in a 5G1237
context. Wireless Personal Communications, 57(3):317–338, 2011.1238
[106] J. Vieira, S. Malkowsky, K. Nieman, Z. Miers, N. Kundargi, L. Liu, I. Wong, V. Öwall, O. Edfors, and F. Tufvesson. A flexible1239
100-antenna testbed for massive MIMO. In IEEE GLOBECOM Workshops, pages 287–293, 2014.1240
[107] C. Wang and et al. Cellular architecture and key technologies for 5G wireless communication networks. IEEE Communications1241
Magazine, 52(2):122–130, 2014.1242
[108] R. Wang, H. Hu, and X. Yang.Potentials and challenges of C-RAN supporting multi-rats toward 5G mobile networks.IEEE Access,1243
2:1187–1195, 2014.1244
[109] W. Wang and V. K. N. Lau. Delay-aware cross-layer design for device-to-device communications in future cellular systems. IEEE1245
Communications Magazine, 52(6):133–139, 2014.1246
[110] W. Wang, G. Yu, and A. Huang.Cognitive radio enhanced interference coordination for femtocell networks. IEEE Communications1247
Magazine, 51(6), 2013.1248
[111] W. Wang and Q. Zhang. Local cooperation architecture for self-healing femtocell networks. IEEE Wireless Commun., 21(2):42–49,1249
2014.1250
[112] X. Wang, M. Chen, T. Taleb, A. Ksentini, and V. C. M. Leung.Cache in the air: exploiting content caching and delivery techniques1251
for 5G systems. IEEE Communications Magazine, 52(2):131–139, 2014.1252
[113] L. Wei, R. Q. Hu, Y. Qian, and G. Wu. Enable device-to-device communications underlaying cellular networks: challenges and1253
research aspects. IEEE Communications Magazine, 52(6):90–96, 2014.1254
[114] G. Wu, C. Yang, S. Li, and G. Y. Li. Recent advances in energy-efficient networks and their application in 5G systems. IEEE Wireless1255
Commun., 22(2):145–151, 2015.1256
[115] J. Wu, Z. Zhang, Y. Hong, and Y. Wen.Cloud radio access network (C-RAN): a primer. IEEE Network, 29(1):35–41, 2015.1257
[116] S. Wu and H. Wang. Visible light communications for 5G wireless networking systems: from fixed to mobile communications. IEEE1258
Network, 28(6):41–45, 2014.1259
[117] J. Xie, D. Guo, Z. Hu, T. Qu, and P. Lv. Control plane of software defined networks: A survey. Computer Communications, 67:1–10,1260
2015.1261
[118] N. Yang, L. Wang, G. Geraci, M. Elkashlan, J. Yuan, and M. D. Renzo. Safeguarding 5G wireless communication networks using1262
physical layer security. IEEE Communications Magazine, 53(4):20–27, 2015.1263
[119] V. Yazici, U. C. Kozat, and M. O. Sunay. A new control plane for 5G network architecture with a case study on unified handoff,1264
mobility, and routing management. IEEE Communications Magazine, 52(11):76–85, 2014.1265
[120] X. Zhang, W. Cheng, and H. Zhang. Heterogeneous statistical QoS provisioning over 5G mobile wireless networks. IEEE Network,1266
28(6):46–53, 2014.1267
[121] X. Zhang, J. Zhang, W. Wang, Y. Zhang, C. I, Z. Pan, G. Li, and Y. Chen. Macro-assisted data-only carrier for 5G green cellular1268
systems. IEEE Communications Magazine, 53(5):223–231, 2015.1269
[122] Z. Zhang, X. Chai, K. Long, A. V. Vasilakos, and L. Hanzo. Full duplex techniques for 5G networks: self-interference cancellation,1270
protocol design, and relay selection. IEEE Communications Magazine, 53(5):128–137, 2015.1271
[123] X. Zhou, Z. Zhao, R. Li, Y. Zhou, T. Chen, Z. Niu, and H. Zhang. Toward 5G: when explosive bursts meet soft cloud. IEEE Network,1272
28(6):12–17, 2014.1273
[124] Y. Zhou, D. Li, H. Wang, A. Yang, and S. Guo.QoS-aware energy-efficient optimization for massive MIMO systems in 5G.In Sixth1274
International Conference on Wireless Communications and Signal Processing, WCSP, pages 1–5, 2014.1275
26
Nisha Panwar is a PhD student at Department of Computer Science, Ben-Gurion University of
the Negev, Israel. She received her Master of Technology (M.Tech.) degree in Computer
Engineering from National Institute of Technology, Kurukshetra, India in 2011. Her research
interests include security and privacy issues in vehicular networks, computer network and
communication, and distributed algorithms.
Shantanu Sharma is a PhD student at Ben-Gurion University of the Negev, Israel. He received
his Master of Technology (M.Tech.) degree in Computer Science from National Institute of
Technology, Kurukshetra, India, in 2011. He was awarded a gold medal for the first position in
his M.Tech. degree. His research interests include designing models for MapReduce
computations, distributed algorithms, mobile computing, and wireless communication.
Awadhesh Kumar Singh received his Bachelor of Technology (B.Tech.) degree in Computer
Science from Madan Mohan Malaviya University of Technology, Gorakhpur, India, in 1988, and
his Master of Technology (M.Tech.) and PhD degrees in Computer Science from Jadavpur
University, Kolkata, India, in 1998 and 2004, respectively. He joined the Department of
Computer Engineering at National Institute of Technology, Kurukshetra, India, in 1991, where
he is presently a Professor and serving as the chairman of the department. Earlier, he also served
as the chairman of the department during 2007-2009. He is currently chief vigilance officer of
the institute. His research interests include distributed algorithms, mobile computing, and fault
tolerance.
*Brief Biography
Click here to download Brief Biography: Author bio.docx
the Negev, Israel. She received her Master of Technology (M.Tech.) degree in Computer
Engineering from National Institute of Technology, Kurukshetra, India in 2011. Her research
interests include security and privacy issues in vehicular networks, computer network and
communication, and distributed algorithms.
Shantanu Sharma is a PhD student at Ben-Gurion University of the Negev, Israel. He received
his Master of Technology (M.Tech.) degree in Computer Science from National Institute of
Technology, Kurukshetra, India, in 2011. He was awarded a gold medal for the first position in
his M.Tech. degree. His research interests include designing models for MapReduce
computations, distributed algorithms, mobile computing, and wireless communication.
Awadhesh Kumar Singh received his Bachelor of Technology (B.Tech.) degree in Computer
Science from Madan Mohan Malaviya University of Technology, Gorakhpur, India, in 1988, and
his Master of Technology (M.Tech.) and PhD degrees in Computer Science from Jadavpur
University, Kolkata, India, in 1998 and 2004, respectively. He joined the Department of
Computer Engineering at National Institute of Technology, Kurukshetra, India, in 1991, where
he is presently a Professor and serving as the chairman of the department. Earlier, he also served
as the chairman of the department during 2007-2009. He is currently chief vigilance officer of
the institute. His research interests include distributed algorithms, mobile computing, and fault
tolerance.
*Brief Biography
Click here to download Brief Biography: Author bio.docx
Secure Best Marks with AI Grader
Need help grading? Try our AI Grader for instant feedback on your assignments.
Nisha Panwar
Shantanu Sharma
*Author Photo
Click here to download Author Photo: Author pic.docx
Shantanu Sharma
*Author Photo
Click here to download Author Photo: Author pic.docx
Awadhesh Kumar Singh
1 out of 30
Your All-in-One AI-Powered Toolkit for Academic Success.
+13062052269
info@desklib.com
Available 24*7 on WhatsApp / Email
![[object Object]](/_next/static/media/star-bottom.7253800d.svg)
Unlock your academic potential
© 2024 | Zucol Services PVT LTD | All rights reserved.