Journal of Network and Computer Applications 78 (2017) 9–29

Contents lists available at ScienceDirect

Journal of Network and Computer Applications

journal homepage: www.elsevier.com/locate/jnca

Review

A survey on device-to-device (D2D) communication: Architecture and MARK

security issues
⁎,1
Pimmy Gandotraa, Rakesh Kumar Jhaa, , Sanjeev Jainb
a
Department of Electronics and Communication Engineering, Shri Mata Vaishno Devi University, J & K, India
b
Department of Computer Science Engineering, Shri Mata Vaishno Devi University, J & K, India

A R T I C L E I N F O A BS T RAC T

Keywords: The number of devices is expected to radically increase in near future, with an estimate of above 50 billion
Device-to-device (D2D) communication connected devices by 2020. The subscribers demand improved data rates, with reduced latency and increased
Spectral efficiency system capacity. To endure the rising demands, cellular networks need to undergo suitable changes. For
Resource allocation fulfillment of the rising needs of users and efficient utilization of the available scarce resources, device-to-device
Security
(D2D) communication is being looked upon as an important emerging technology for present and future cellular
InternetProtocol Security (IP Sec)
networks. It allows peer-to-peer communication between users, with improved spectral efficiency, energy
efficiency and system throughput. In this paper, a detailed survey on device-to-device (D2D) communications
has been offered, along with the challenges which exist for D2D (like resource allocation, security, interference
management etc.) to become a successful paradigm of wireless networks. In order to fulfill the subscriber needs,
architecture has been proposed which assures overcoming the various implementation challenges of D2D
communication. The paper largely focuses on security in D2D communication and the possible attacks to which
the direct links are susceptible to. For ensuring a secure D2D communication, solution has been proposed,
based upon Internet Protocol Security (IP Sec).

1. Introduction four generations of wireless communication (1 G, 2 G, 3 G, 4 G), the

network operators nowadays are getting attracted towards the D2D
The current cellular network technologies tend to be incapable of technology, due to the advancing trends (Astely et al., 2013). It also
meeting the rising subscriber demands. This explosion in demand has provides an efficient offloading solution to the mobile network opera-
been due to popularity of various large bandwidth demanding applica- tors (MNOs).
tions like mobile computing, video streaming etc. It is predicted that by For a low energy direct communication between the transmitter
2020, trillions of wireless devices shall be serving billions of people and receiver, traditionally, Wireless Personal Area Networks (WPAN),
(Doppler et al., 2011). The overcrowded spectrum triggers research for and Wireless local area network (WLAN) technologies have been used.
new technologies for efficient utilization of the spectral resources (Hu These technologies operate in the license exempt band, providing local
and Qian, 2014). As a result, the notion of the fifth generation (5 G) services at a low cost. device-to-device (D2D) communication, how-
networks is being extensively explored, which is essentially an aggrega- ever, allows a direct transmission in cellular spectrum, greatly enhan-
tion of a number of technologies. These networks are expected to cing the spectrum utilization. Considering the pitfall of interference,
support the new usage scenarios. One of the promising technologies, the licensed band is much better than the unlicensed band since
device-to-device (D2D) Communication, is expected to meet the interference is controllable due to the involvement of a central
technical goals of the next generation networks (NGNs). controlling entity in the licensed spectrum, the Evolved NodeB
D2D communication refers to direct transmission between prox- (eNB). In case of uncontrollable interference levels, the Quality of
imate devices, without relaying information through the base station Service (QoS) of the network deteriorates, which is not desirable. Apart
(BS). Such a direct transmission improves spectral efficiency, system of uncontrollable interference levels in the license exempt bands, the
capacity (Chai et al., 2013) and reduces latency. It facilitates proximity traditional direct communication technologies also consume more
aware services in the cellular networks. Although the functionality of energy, causing faster battery drainage. Thus, D2D communication is
Device to Device (D2D) communication has been neglected in the first a desirable technique for direct communication, under a controlled

⁎
Corresponding author.
E-mail addresses: [email protected] (P. Gandotra), [email protected] (R. Kumar Jha), [email protected] (S. Jain).
1
Assistant Professor, DoECE, Shri Mata Vaishno Devi University, Katra, J & K, India.

http://dx.doi.org/10.1016/j.jnca.2016.11.002
Received 26 July 2016; Received in revised form 13 October 2016; Accepted 4 November 2016
Available online 12 November 2016
1084-8045/ © 2016 Elsevier Ltd. All rights reserved.
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

C-RAN INTERNET OF VEHICLES (IoV)

WIRED LINK
PUBLIC
SAFETY CELLULAR LINK
SERVICES DEVICE RELAYING
DIRECT LINK
COMMUNICATION LINK FROM
SERVER SERVER
CELLULAR USER
D2D USER
D2D PAIR
BASE
STATION
SMALL CELL

SMALL
CRAN ~ Cloud Radio Access Network
CELL

D2D PAIR

Fig. 1. A General D2D Scenario.

environment, for better network performance. network. Resources must be available to the D2D users for carrying out
3GPP is investigating Proximity Service (ProSe) (Jin et al., 2014) communication over direct links. Transmission range can be enhanced
for enabling direct communication between proximate devices. Various with the placement of relays in a D2D scenario. Assuring security in the
use cases of D2D and its feasibility is discussed in 3GPP TR 22.803 D2D communication thus becomes essential, as relays and rest of the
(2012). Enabling D2D communication in cellular networks requires communicating devices, are susceptible to eavesdropping attack, jam-
certain architectural enhancements, which are investigated upon ming attack, etc. All these issues need special attention.
in 3GPP TR 23.703 (2013). FlashLinQ (Wu et al., 2010) from
Qualcomm propose PHY/MAC architecture for D2D networks. A
comprehensive survey on the standardization of D2D communication 1.1. Contributions
(3GPP Release 12) has been provided in Lien et al. (2016), which
includes architectures, discovery procedures, mobility, channels and Existing surveys on D2D communication have been complemented,
signals. and the missing pieces of information have been brought up in this
Numerous services can be associated with D2D communication, survey. An inclusive overview of 5 G technology is provided in Dohler
like, mobile cloud computing, cellular data offloading, disaster manage- and Nakamura (2016). In Asadi et al. (2014a), an extensive literature is
ment, etc. A general scenario of device-to-device (D2D) communication available on device-to-device (D2D) communication, with focus on
in cellular networks is depicted in Fig. 1. The figure depicts various use various under-explored and over-explored areas of research in D2D
cases of device-to-device (D2D) communication in a single cell, with communication. The authors of Liu et al. (2014) bring about a detailed
the base station at the centre. Vehicle-to-vehicle (V2V) communication survey of the existing research in this field, mentioning about the
with Internet of Vehicles (IoV), public safety services, offloading of various experimental prototypes, applications of D2D communication
cellular traffic from the base station, etc. are depicted in the figure. The and its open research areas. The authors in Mach et al. (2015) present a
Cloud Radio Access Network (CRAN) in the 5 G networks supports detailed survey on device-to-device (D2D) communication, primarily
flexible network architecture. It centralizes the network functionalities, focusing on Orthogonal Frequency Division Multiplexing Access
as per the application to be serviced. For 5 G cellular networks and (OFDMA) which is being considered for the future 5 G networks.
beyond, small cells in conjunction with D2D communication, can serve Architectural, business and technical aspects of D2D communication
as effective means of enhancing coverage and enabling traffic off- have been explained in Mumtaz (2014) comprehensively. Network
loading. As D2D communication allows sharing of local content, assisted multi hop, full duplex and multi-antenna D2D have been
without multiple base station requests, proactive caching (Pappalardo reviewed in Fodor et al. (2016). The authors argue that the challenges
et al., 2016) has been introduced, to reduce load on the backhaul of the information society beyond 2020 can be effectively met through a
networks. Caching can be performed at the base station, or at the user proper combination of cellular and adhoc technologies. Architecture
terminals, communicating over direct links (Gregori et al., 2016). A has been proposed for the next generation networks in Pimmy and
careful caching strategy in device-to-device (D2D) communication can Rakesh (2016).
greatly enhance Quality of Experience (QoE) (Bai et al., 2016). In this paper, an exhaustive literature on device-to-device (D2D)
Though D2D communication allows direct communication between communication has been studied and thoroughly reviewed. There is a
proximate devices and supports improved network performance, its lot of scope in this field for propelling the researchers all over the globe
implementation involves number of concerns. For setting up of direct to switch toward the D2D technology. A few important algorithms,
links, the users need to discover each other. The discovery mechanisms pertaining to the various issues in device-to-device (D2D) communica-
are being widely studied. There is a very high probability of inter- tion have been discussed. This paper has also shown a suitable network
ference between D2D users and cellular users, within the cellular model which is an integral part of basic resource allocation strategy.
But the prime focus of the paper is on the proposed architecture for

10
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 1 Direct communication came with the introduction of wireless personal

Comparison of wireless technologies. area network (WPAN) and wireless local area network (WLAN)
technologies in the third generation (3 G) networks. These technologies
S No. Feature 3G 4G 5G
Considered allowed content sharing in the unlicensed band only, at low cost and
low energy consumption. However, interference levels were uncontrol-
1. Approach Used Network Network Device Centric lable and no Quality of Service (QoS) guarantees could be provided,
centric centric
due to the license exempt band operation. Also, it involved higher
2. 3GPP Releases Release 4,5,6 Release Release 12,13
8,10,11 power consumption. Due to the massive growth in the traffic and
3. Access CDMA OFDMA/SC- SCMA, BDMA, changing trends, direct communication in the licensed band started
Methodology FDMA FBMC gaining popularity for the 4 G networks (Long Term Evolution
4. Approx Latency 100–150 ms Upto less Less than Advanced, LTE-A), and beyond. A detailed discussion on the evolution
than 5 ms 1 ms(Zero Latency)
of the generations of wireless communication has been provided
5. Carrier 0.8/ 0.85/ 0.9/ 1.8/2.3/2.5/ 1.8/2.6 GHz, and
Frequency 1.8/1.9/ 2.1/ 2.6/3.5 GHz expected 30– in Akhil and Rakesh Jha (2015).
2.6/3.5/ 300 GHz The need for the next generation networks arouse given that the
5.8 GHz network society of 2020 and beyond is expected to undergo an
6. Mobility (in 300–350 300–350 500 (Alliance,
explosion in the data traffic (Cisco, 2014). Mobile video traffic is
Km/h) 2015)
7. Offloading Bluetooth, LIPA, SIPTO, Small Cells, D2D expected to account for more than 70% of the total data traffic by 2018.
Technique WLAN Small Cells communication A key component technology of the fifth generation (5 G) networks is
(Yang et al., device-to-device (D2D) communication. The first four generations have
2013) been network centric, but 5 G is heading towards device-centric net-
8.. Resource Poor Better Highly efficient
work architecture. Network centric means base station was the sole
Utilization
9. Support for No No Yes controlling entity in the network. With the advancement in wireless
Energy industry, there has been an increase in the small cell deployments. The
Harvesting trend is thus shifting towards user-centric (device centric) networks, in
10. Relay Support No Yes Yes
which the user is expected to actively perform storage, relaying,
11. Carrier No Yes Yes
Aggregation
computation and content delivery, which was earlier being performed
by the network (the base station (BS)). D2D communication enhances
coverage, cost efficiency, reliability, capacity density and spectrum
D2D communication in the next generation networks (NGNs). This efficiency. A brief comparison of the generations of wireless commu-
paper has also contributed towards security by proposing a solution nication supporting direct communication has been presented in
concept for securing D2D communication using Internet Protocol Table 1. Efforts are on the run to integrate D2D communication to
Security (IP Sec). Various security attacks on D2D have been briefed the existing cellular networks.
in the survey, with particular emphasis on jamming attack (Section 5).
The paper is organized as follows. Following the introduction, a 3. Fundamentals of device-to-device (D2D) communicaion
brief journey through the generations of wireless communication has
been presented in Section 2, describing the onset of device-to-device Traditional cellular communication allows communication between
(D2D) communication in the cellular networks. Fundamentals of user equipments (UEs) through the base station (BS), but the UEs may
device-to-device (D2D) communication have been discussed in be in range to carry out direct communication. It means communica-
Section 3, which includes the essential basic architecture, different tion between the UEs is possible without traversing the core network.
types of D2D communication, channel models and performance Such a communication results in improved spectral efficiency, energy
metrics for D2D communication. In this section, architecture has been efficiency, reduced delay and enhanced throughput. Additionally, data
proposed for D2D communication in next generation networks traffic is drastically increasing due to the content being shared by the
(NGNs), and it aims to optimally fulfill all the demands of the users, demanding large bandwidth, and high data rates. To serve the
subscribers and the requirements of the service providers in near requirements of the subscribers, D2D communication plays an essen-
future. The key issues in D2D communication are discussed in Section tial role. D2D enables merging ad hoc and centralized networking.
4. Security in D2D communication, and a proposed solution for Taking benefit of ad hoc network, D2D communication can be used in
securing the direct links is discussed in Section 5. D2D is expected to conjunction with other technologies like cooperative communication,
support a large number of applications in the next generation cellular cognitive radio, Internet of Things (IoT), to enhance the spectral
networks, which are discussed in Section 6. In spite of being such an efficiency. With centralized networking, D2D communication helps in
important technology, few research challenges exist in its implementa- overall enhancement of the network performance, having the control
tion, as discussed in Section 7. Eventually, the paper concludes in from the operator.
Section 8. From the architecture point of view, D2D networks look similar to
Mobile Ad-hoc Networks (MANETs), and Cognitive Radio Networks
(CRNs), but in actuality, there exist some differences. Numerous issues
2. Evolution of device-to-device (D2D) communication related to MANETs and CRNs can be resolved in D2D communication,
through the generations as a result of the presence of a central controlling entity, the base
station (BS). Distinguishing users as primary or secondary is possible
Wireless communication has evolved from the first generation (1 G) with D2D communication and CRNs (Cheng et al., 2012), as a result of
to the fourth generation (4 G), and to counter the demands of the which both can be treated as similar technologies but, CRNs function
subscribers, the fifth generation (5 G) networks will be soon available. autonomously and use cognitive sensing (Ejaz et al., 2011), which is
Cellular communication began with the introduction of the first not supported by D2D communication. Machine-to-machine (M2M)
generation (1 G) of wireless communication. In 1 G, only analog communication also seems to be similar to D2D communication but it
cellular technology was used. Digital cellular came into existence in is application oriented, unlike D2D, which primarily focuses on
the Second Generation (2 G). Until this time, no direct communication proximity connectivity services. The M2M communication does not
between users had been introduced. It was due to lesser number of contribute towards enhancement of spectral efficiency, however D2D
subscribers and their demands being limited only to voice and data. communication does.

11
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

works are considered as a two-tier network, where the tiers are the
macro cell tier and device tier. The macro cell tier involves the
conventional cellular communication, through the base station (BS).
The device tier, on the other hand, involves direct communication
between the devices (D2D communication). The base station can have
different levels of control in the device tier network, which maybe full
or partial, or no control at all. As a result, device to device (D2D)
communication can be categorized into four types, depending upon the
degree of involvement of the base station in assisting the communica-
tion between devices (Tehrani et al., 2014), and are discussed below.
Case (i) Device Relaying with base station assisted con-
trolled link.
A device in an area of poor coverage or at the cell edge can
communicate with the base station by relaying information through
other devices. The allocation of resources and setting up of call in such
a scenario is done by the base station (BS). Also, interference can be
Fig. 2. Fundamental Network Architecture for D2D communication.
avoided between devices as a result of the centralized control. This
enhances the battery life of the device.
With D2D communication underlaying cellular networks, content
Case (ii) Direct communication between devices with base
sharing by a large number of users is possible by setting up of local
station assisted controlled link.
caching and sharing zones. This supports network management in a
The source and the destination devices communicate with each
distributive manner. Also, relaying of data is possible with device-to-
other directly, with the base station (BS) assisting them for link
device (D2D) communication. Device-to-device (D2D) communication
establishment by providing the control links. The BS primarily deals
has been identified as a key enabler for various 5 G services, by the
with access authentication, connection control, resource allocation and
METIS research project. As identified by METIS (Li et al., 2014),
interaction between devices.
device-to-device (D2D) communication can be utilized in a number of
Case (iii) Relaying device with device assisted controlled
scenarios, depending upon the requirements. These have been listed
link.
below as per the needs:
The source and destination communicate with each other via relays.

•
The entire communication is controlled and managed by the devices
For backhaul applications, D2D-B
•
themselves, thus eliminating the role of the base station in the
For critical applications, D2D-C
•
communication process.
For direct D2D, without involvement of the core network D2D-D
•
Case (iv) Direct D2D with device assisted controlled link.
For direct M2M communication, D2D-M
•
The source and destination devices communicate directly, without
For non-critical applications, D2D-N
provision of any control links from the base station at all. Thus, these
devices must be careful with respect to the management of interference
A complete overview of the METIS project is available in Dohler
within the device tier and, also with the macro cell tier. Before starting
and Nakamura (2016) and Fodor et al. (2016). METIS 5 G concepts are
the transmission of data, the devices need to find each other (device
evaluated in detail and their significance with D2D communication is
discovery) and then communicate. All these types are depicted in
presented.
Fig. 3.
Another classification of D2D communication is on the basis of the
3.1. Fundamental network architecture spectrum (licensed/unlicensed) accessed by the D2D users. When
device-to-device (D2D) communication occurs on the licensed spec-
The basic network architecture of D2D communication is divided trum, it is referred to as inband D2D and on the unlicensed spectrum,
into three parts (Alkurd et al., 2014): area network, network manage- outband D2D (Asadi et al., 2014a, 2014b). The inband D2D can be
ment and D2D applications. The fundamental architecture is depicted categorized as underlay or overlay. Underlay D2D uses the cellular
in Fig. 2. A large number of devices, communicating with each other via spectrum for cellular communication as well as for device-to-device
direct links form a part of the D2D area network. Aggregators are (D2D) communication, while, in overlay, a dedicated portion of the
available in the network architecture, which collect the data from all the entire spectrum is used for device-to-device (D2D) communication.
D2D devices and after aggregation, connect to the core network. Outband D2D communication can be autonomous or controlled. It is
Thereafter, the data is sent to the gateway, which connects to the said to be controlled when the radio interfaces are coordinated by the
access network. The access network can be wired or wireless. The base station, even though the direct communication is occurring in the
devices are connected to the service providers from the core network. unlicensed band. For autonomous D2D, these interfaces are coordi-
In order to enable working of D2D communication in the cellular nated by the users themselves. For outband D2D communication,
networks, new functionalities need to be added to the existing devices thus need two wireless interfaces, like LTE and Wi-Fi. The
structures. Certain enhancements in the LTE-Advanced (LTE-A) interfaces can be Wi-Fi Direct (Alliance, 2010), Bluetooth (Bluetooth,
architecture (Technical Specification Group Services, 2014) can enable 2001), Zigbee (Alliance, 2006). This categorization of D2D is shown in
the functioning of D2D communication in the existing cellular net- Fig. 4. Coexistence of number of modes increases complexity of the
works. Architecture supports for the next generation networks is system (Asadi et al., 2014b). As a result, the mode i.e. spectrum
described in Lien et al. (2016) and (Yang et al., 2013). Another accessible by the D2D devices should be decided, taking into con-
modified LTE-A network, to support D2D communication has been sideration the overall system conditions. Different types of D2D
discussed in Raghothaman et al. (2013), with the introduction of D2D communication (inband/outband) have their pros and cons. A compar-
server in the system. But security of communication must be ensured. ison between these has been given in Table 2.
Majority of the available literature on device-to-device (D2D)
3.2. D2D types communication is based on underlay inband D2D. It is efficient for
enhancing spectral efficiency, coverage area and other targets for better
Device-to-device (D2D) Communication enabled 5 G cellular net- performance of cellular networks.

12
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Fig. 3. Types of D2D Communication, on the basis of involvement of base station.

requires more number of channels proportionate to the coverage area.

Cells are split into sectors, with the use of directional antennas at the
base station. Sectoring may divide the cell into three 120° sectors or six
60° sectors. This makes them more efficient and greater number of
calls can then be handled at the same time. It is an effective technique
for initiating improved signal-to-interference-plus-noise ratio (SINR)
and reducing co-channel interference. Architecture consisting of direc-
tional antennas at the base station, dividing the coverage area into
Fig. 4. Types of D2D Communication (based on spectrum access). three sectors of 120° each is proposed (Fig. 5). This architecture is
proposed in view of the rising traffic scenario and deployment
Table 2 considerations of the next generation networks (NGNs).
Comparison between inband and outband D2D. It is a suitable architecture for servicing users that are located in
congested areas or at the cell edges. The users within each sector are
Inband D2D Outband D2D distinguished as cellular users (CUEs) and D2D users (DUEs). The
Exploits spatial diversity, resulting in Does not result in improved spectral
distance constraint, D≤d0 determines whether two user equipments
enhanced spectral efficiency efficiency (UEs) form a valid D2D pair or not. Here, D is the distance between the
Base station control results in better No control from the base station, thus, D2D transmitter and D2D receiver and d0 is the threshold distance.
QoS management no support for better QoS Resource allocation to the D2D users and cellular users is addressed, in
management
this proposal.
High level of interference between D2D Interference problem overcome due to
communication and cellular operation of cellular and D2D As shown in the figure, we explain its working by considering a
communication and among D2D communication in different bands single sector only. The total scheduling period of the base station is
users given by Ts, in which time slots shall be allocated to both, the D2D
No extra interface like Wi-Fi needed by Devices require an extra interface,
users and the cellular users. Prior to allocating time slots to the users
the device thus two interfaces needed by a device
for its successful operation
within a sector, the number of D2D users and cellular users are
High possibility of resource wastage Easier resource allocation, and no determined. As the number of D2D users will be less than the number
wastage caused of cellular users within a network, time slots are first allocated to the
No coding/decoding involved, as only a Coding and decoding of packets is D2D users. In the proposed architecture, we consider three D2D pairs
single interface is used essential as different interfaces use
and six cellular users in a single sector, which are being considered for
different protocols
time slot allocation. The resource allocation is adaptive. Adaptive time
slots are allocated to the D2D users, on the basis of the application
1) Proposed Architecture demanded by them, which may be video, voice or text application. Of
these applications, maximum length of time slot will be required by
Due to increase in the number of subscribers and their demand for video application, followed by voice and text, respectively. As depicted
high data rates, channels available in a cell become insufficient. This in Fig. 5, the adaptive time slot allocation is made to the D2D pairs, on

13
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Fig. 5. Proposed aArchitecture.

the basis of the applications demanded by them. Pair 1 demands a on the basis of these types, different channel models exist. Channel
video application, being allocated the largest size time slot of all the modeling and measurements includes three steps: channel measure-
three. Pair 2 demands voice application, allocated a comparatively ment, raw data processing and channel modeling and simulations. The
smaller sized time slot. Pair 3 demands a text application, allocated existing channel models of cellular networks cannot be used in D2D
with the minimum sized time slot. This time, used by D2D pairs is communication directly. The suitability of channel models, in D2D
denoted by TD2D. At another instant, these pairs may demand a scenarios are discussed in Cheng et al. (2015) and different channel
different application, thus, will be allocated suitable sized time slots models for D2D have been presented in Table 3. These consider
then. Such an adaptive time slot allocation is performed using the indoor-to-indoor (I2I), outdoor-to-outdoor (O2O), line-of-sight (LOS)
Hidden Markov Model (HMM). After allocating time slots to the D2D and non-line-of-sight (NLOS) scenarios. The authors categorize the
users, time slots are adaptively allocated to the cellular users from the channel models as deterministic or stochastic. The need for new
remaining time duration. This allocation of time slots to the cellular channel models in D2D communication is the fact that in D2D
users is also adaptive, based on the applications (video, voice and text) communication, both the transmitter and receiver can be mobile and
they demand. The simulations for such a resource allocation can be they may be shadowed by human body.
performed in MATLAB. The time allocated to cellular users is denoted D2D channel modeling is affected by path loss and shadow fading.
by TCellular. A fast fading channel generation method has been proposed in Kim and
The architecture targets throughput maximization, latency mini- Oh (2014), based on Doppler frequency response and, the
mization, and better Quality of Service (QoS). It is an effective solution InternationalTelecommunication Union-Radio(ITU-R) channel model
for overcoming the various open issues in D2D communication. With is extended to the D2D channel. The Doppler effects and scattering are
the proposed approach, resources are shared between the two types of taken into consideration on both the sides, the D2D transmitter as well
users (CUEs and DUEs). Since D2D communication allows direct the D2D receiver. Double ring channel models have been proposed for
communication between proximate users, power consumed is less. As direct communication in Patel et al. (2005), and depicted in Fig. 6. This
a result, this architecture can optimize the overall power of the model assumes use of an Omni directional antenna at the transmitter
network. This proposed architecture is capable of performing direct as well as the receiver. The model considers the existence of uniformly
communication between devices efficiently, thereby improving the spaced scatters at both the ends, i.e. transmitter as well as receiver.
overall network performance and satisfying subscriber demands. The This is a mathematically convenient model. Such a model is particu-
primary focus of the architecture remains optimal resource allocation, larly useful for simulating scattering environments, but computational
so as to meet the subscriber demands in the most efficient manner. No load is high. A D2D model with Rician fading channel has been
single possible architecture can be clearly defined for D2D commu- investigated in Peng et al. (2014). In Li et al. (2015), a three-
nication. This fact becomes clear from Jin et al. (2014), Asadi et al. dimensional two-cylinder model has been proposed for D2D commu-
(2014b) and Rêgo and da (2012). Choice of architecture is dependent nication. From this model, more realistic results can be achieved. New
on a number of factors like, coverage extension, energy efficiency, channel models need to be evaluated, in order to realize the complete
Quality of Service, and other architecture related issues. potential of D2D communication. A number of challenges related to
channel modeling for D2D communication have to be dealt with, and is
an area of open research.
3.3. Channel models for D2D communication

As discussed, there are different types of D2D communication and

14
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 3 performance of a network. It can be seen as analogous to the

D2D channel models (Cheng et al., 2015). traditional channel capacity, subject to secrecy constraints. Secrecy
rate maximization is a prime agenda in cellular networks.
S. No Scenarios/Properties Authors/ Channel Models
Organizers 5. Mean Opinion Score (MOS): It is used to evaluate the perfor-
mance of a cellular network, in terms of the Quality of Experience.
1. Indoor to Indoor (I2I) 3GPP InH Mode lDual stripe The range of MOS lies between 1 and 5. Values close to 5 are
model
desirable. It signifies excellent quality of information.
COST (MBPGM)
IEEE 802.11 COST 231 model 6. Latency: It is an indicator of the delay between transmission and
(MBPGM) reception of information. D2D communication generally results in
Berg Recursive 802.11 model (MBPGM) lower latency, due to transmission over a small distance.
I2I model (MBPGM) 7. Energy Efficiency: It is the ratio of the system throughput to the
power consumption per unit area. It is an indicator of the efficient
2. Outdoor to Outdoor 3GPP ITU UMi Model Xia
(O2O) model (MBPGM) energy (Battery power) utilization within a cellular network.
ITU-R O2O model (MBPGM) 8. Spectral Efficiency: The number of bits transmitted per unit
P.1411-6 bandwidth indicates the spectral efficiency. It is essential for
quantification of the D2D network performance.
Berg recursive O2O model (MBPGM)
I2O model (MBPGM)
This section touches upon various essential aspects of D2D com-
3. Vehicle to Vehicle (V2V) D. W. Matolak, MBPGM munication. These fundamentals are essential for any researcher, for
M. A. Ingram Geometry-based investigating various research aspects of D2D communication. This
etc. deterministic
new emerging technology is paving way for a very efficient cellular
J. Maurer etc. models
X. Cheng etc. Regular-shaped architecture, but, redressing some of the issues is critical, for its
practical implementation. These have been discussed in the next
A. F. Molisch geometry-based section. Of all the open challenges, security in D2D communication is
etc. stochastic
a rising concern at present, and has been extensively studied in Section
models
Irregular-shaped
5.
geometry-based
stochastic 4. Key issues in D2D communication
models

In order to enhance the network performance by using D2D

4. Line of Sight/Non Line of 3GPP ITU UMi Model InH
Sight (LOS/NLOS) model communication, a number of issues have to be dealt with. Some
important issues have been discussed in this section, each using a
ITU-R (MBPGM) different analytical tool and evaluation method to support efficient D2D
P.1411-6 O2O model (MBPGM)
communication in cellular networks. The methods proposed for resol-
ving the open issues mostly have a limited scope and achievable
performance. The challenges which have been listed below include
peer discovery and mode selection, radio resource management,
interference management, power control, and security.

4.1. Peer discovery and mode selection

To establish potential direct links between the devices, the user

equipments (UEs) need to identify their neighbors. Exploring oppor-
tunities for peer discovery within the cellular networks is essential for
the sustenance of D2D communication underlaying cellular networks.
Fig. 6. Double Ring channel model for D2D communication. Prior to setting up of the communication link, the UEs can detect each
other by transmission of beacons. A number of approaches exist in
3.4. Performance metrics literature, for device discovery and link set up between the D2D users.
In Feng et al. (2014), the authors propose two techniques of device
In this section, a set of commonly used performance metrics for discovery: restricted discovery and open discovery. In case of restricted
D2D communication have been briefly discussed. These are mostly discovery, the UEs cannot be detected without their prior explicit
related to the SINR (Signal-to-interference-plus-noise). permission thus maintaining user privacy. In case of open discovery,
UEs can be detected during the duration for which they lie in proximity
1. System throughput: It is defined as the overall throughput of all of other UEs. Device discovery may be controlled by the base-station
D2D pairs and cellular users within a cellular system. A higher value tightly or lightly as discussed in Fodor et al. (2012) and Lei et al.
of throughput signifies better performance of the algorithm. It is (2012). Peer discovery can furthermore be either distributed, or
measured in bits/sec. It is an indicator of successful information centralized. In distributed approach, a UE broadcasts beacons to the
transfer between the pairs. other devices in the network (Corson et al., 2010). In Centralized
2. D2D User Capacity: For any resource allocation algorithm, D2D approach, the process of discovery is managed by a single entity or
user capacity is the number of DUEs that can be accommodated for a many entities, which check for devices in proximity and authenticate
given set of cellular users in the network, subject to the maximum the setup of direct links (Lei et al., 2012). Both have their merits and
data rate constraint. A higher user capacity value is always desirable. limitations.
3. Fairness: This is a very critical indicator of evaluating how fair a From the perspective of the User Equipment (UE), energy require-
given scheme is (generally resource allocation), for D2D commu- ment for supporting D2D communication is more. The authors of Hong
nication. Measurement of fairness is decided by the fairness index. et al. (2013) propose an energy-efficient technique for discovering
4. Secrecy Rate: It is an important parameter to evaluate the secrecy devices in proximity and establishing D2D links. The simulation results

15
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 4
Different methods of peer discovery for D2D communication and their comparison.

Reference No. Technique for device discovery Objectives

Easy discoverability Reduced latency Energy efficiency/ power Throughput enhancement

efficiency

(Feng et al., 2014) Restricted or Open Discovery × ✓ × ×

(Hong et al., 2013) Energy-efficient device discovery × × ✓ ×
(Tang et al., 2014) Sound Referencing Signal for neighbor ✓ ✓ × ×
discovery
(Zou et al., 2014) • Bluetooth Discovery ✓ × ✓ ×
• WiFi Device Discovery
• WiFi Direct Device Discovery
• IrDA Device Discovery
• Network Assisted Discovery
• Packet and Signature-based Discovery
• Request Based Discovery

(Nguyen et al., 2014)

• Direct Discovery
Network Assisted Discovery Technique ✓ × × ×
(Chao et al., 2013) Bio-inspired Proximity Discovery ✓ ✓ × ×
(Pratas and Popovski, Network assisted proximity discovery × ✓ × ✓
2015)
(Ngo and Kim, 2015) Time Advance proximity discovery ✓ × ✓ ×

show trade-off between the delay and energy consumption of the

network. The energy consumption in idle state of the UEs is less than in
the connected state. After discovery, though, the delay across the link is
longer in the idle state, than in the connected state, with a maximum
value of 5 msec in the connected state. The authors of Tang et al.
(2014) address the neighbor discovery problem. Using the sounding
reference signal (SRS) channel, neighbors are discovered. Uplink
transmissions of cellular users play an important role in finding the
neighbors. Neighbor discovery under unknown channel statistics is also
considered. Maximum delay in a single discovery cycle is computed as
4 μsec. The authors in Zou et al. (2014) provide a review of the Fig. 7. Steps Included in Direct Link Setup and direct communication.
techniques for proximity discovery. Existing techniques have been
discussed like Bluetooth discovery, Wi-Fi (AD Hoc) Device Discovery, depicted in Fig. 7.
IrDA Device Discovery, following which Request based discovery, After the completion of peer discovery, the UEs need to set up
Direct Discovery, Packet and Signature-based Discovery and sessions. There are two methods for session set up, named IP based
Network-Assisted discovery have been identified as different discovery detection and dedicated D2D signaling. D2D users require optimal
mechanisms for device-to-device (D2D) communication in cellular amount of resources for supporting the communication over direct
networks. An effective network-assisted technique for device discovery links. The links must not interfere with the ongoing cellular commu-
has been proposed in Nguyen et al. (2014) for the support of device-to- nication also. As seen, in this section, a number of peer discovery
device (D2D) communication in LTE networks. The results show that a techniques for device-to-device communication have been discussed.
very high probability of device discovery exists in this technique for a The objective of discovery not only remains finding the communicating
certain discovery interval. Yet, network assisted D2D communication partners, but energy efficiency, security, interference should also be
can take place without device discovery at the device side, as network taken care of. Optimal resource allocation and interference manage-
itself can detect the presence of a device in proximity. ment are discussed next.
A bio-inspired discovery technique is proposed in Chao et al.
(2013), where proximity signals are transmitted and detected by the 4.2. Radio resource allocation
devices to accomplish proximity discovery. Firefly spanning tree
algorithm has also been proposed in Chao et al. (2013), in order to The objective of introducing D2D communication in cellular net-
overcome problems associated with various other topologies, for device works has been to improve the spectrum utilization. Resource alloca-
discovery. The authors in Pratas and Popovski (2015) propose a tion deals with allocation of frequency resources to the D2D pairs,
network assisted proximity discovery protocol in which cellular users within the cellular networks. This allocation must be optimal for
act as monitors or announcers to listen to, or broadcast the information functioning of D2D networks underlaying cellular networks, and meet
of interest within the proximity region. This protocol is highly efficient the objectives of this emerging technology. Once device discovery is
in terms of energy. Another device discovery mechanism has been complete, using a comprehensive discovery mechanism (Mumtaz,
proposed in Bagheri et al. (2015), considering the near far effect. An 2014), resource allocation has to be performed. Different modes of
efficient time advance method for peer discovery has been proposed resource allocation exist for D2D communication: cellular mode,
in Ngo and Kim (2015) which compute the distance between the UEs in dedicated mode, mode supporting reuse of resources of a single cellular
the cellular network. As a result, this method is based on simple user (CUE) and mode supporting reuse of resources of multiple cellular
geometry. A novel peer selection approach to assign a direct link to a users. There is no interference problem in the first and second mode as
D2D pair has been proposed in Namvar et al. (2015), exploiting context they involve utilization of orthogonal resources, but these two modes
information about user mobility and size of data demanded. A maybe inefficient to maximize the overall network throughput. The
summary of these methods is with a comparative analysis is given in other two modes suffer interference because of resource sharing with
Table 4. A sequence of steps involved in the setup of direct links is cellular users. Yet, maximum resource utilization is achieved through

16
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

reuse mode only. The radio resources are allocated using centralized or technique used in Kim et al. (2006). The second phase uses a heuristic
distributed techniques, in D2D communication. Centralized techniques sub channel allocation scheme, which takes into consideration, the
result in increased complexity in case of large networks, supporting a minimum data rate constraint for resource allocation. The incentive
large number of UEs. In these techniques of resource allocation, resource allocation scheme is proposed in Sun et al. (2015). It
resources are allocated to the D2D users and cellular users by the base introduces energy efficiency, as an incentive parameter and stimulates
station (BS). Distributed techniques, on the other hand, decrease the cooperation among users. For the users needing cooperation, DUE
device complexity, and reduce interference with the cellular networks. relays maximize the received signal-to-noise ratio (SNR) first. Then,
Some distributed techniques of resource allocation in D2D commu- the resource allocation problem is formulated for two users and finally
nication is studied in Zhang et al. (2013), Song et al. (2014), Kuang solved using a two-dimensional search method. A semi-control re-
et al. (2015), Zhang et al. (2013), and Mohammed and Hossain (2014). source scheduling algorithm has been proposed in Wang et al. (2013),
In cellular networks, information like channel gains, SINR measure- in which the boundary of the resource-assignment and resource-
ment etc. are uncertain and time-varying. Robust resource allocation contention area is adjusted dynamically. A proportional fair scheduling
can then deal with optimization problems under such circumstances, (PFS) technique is used for resource allocation, and maintaining
and can model the uncertainties by ellipsoids, polyhedrons and D-norm overall fairness of allocation. The simulation results show that the
methods (Kai et al., 2008). A number of resource allocation algorithms performance of this algorithm is far better than round-robin schedul-
have been proposed by the researchers for efficient utilization of the ing. The authors of Zheng et al. (2015) propose an adaptive time
spectrum. Hybrid solutions can also be provided and are an open area division scheduling algorithm which also is based on proportional
of research. An effective hybrid technique is discussed in Lee et al. fairness algorithm for adaptively scheduling D2D pairs. For reduction
(2014). of delay, wireless caching with D2D is a smart choice. A resource
In Zhou et al. (2013), intra cluster retransmission scheme has been allocation scheme, based on social interaction is addressed in Wu et al.
proposed for efficient resource utilization A Heuristic Location (2016), for improvement of the system efficiency. Heuristic algorithms
Dependent Resource Allocation Algorithm has been proposed for dynamic resource allocation have been investigated upon
in Botsov et al. (2014) and resource pooling in Fodor et al. (2012). in Belleschi et al. (2015).
Taking into consideration the spectral efficiency of LTE-A networks, A radio resource management scheme for V2X communications has
(Phunchongharn et al., 2013) proposes a resource allocation technique. been formulated in Sun et al. (2016). As V2X communications adhere
It minimizes the transmission length of the D2D links on increased to strict latency and reliability requirements, an NP hard optimization
power consumption. An NP-Complete problem is formulated (Garey problem is formulated, to enable robust resource allocation. Optimal
and Johnson, 1979) as a Mixed Integer Programming problem and a resource allocation for improving the delay performance of VANETs
column generation method allocated resources optimally, supporting has been discussed in Cao et al. (2016). Packet lifetime is checked
low complexity. In Koskela et al. (2010) and (Hakola et al., 2010), the periodically by each vehicle, and D2D links are setup, if needed. An
authors target the Signal-to-Interference-plus-Noise ratio (SINR). The optimal receiver vehicle is selected for the establishment of D2D links.
authors (Feng et al., 2013) provide a method for overall throughput The effectiveness of the proposed algorithm is validated by simulation
improvement and enhancement in spectral efficiency through power results. Resource allocation techniques for applying D2D communica-
allocation and admission control. A general resource scheduling tion in VANETs have been discussed in Weijun et al. (2014) also, which
problem is investigated in Asadi et al. (2014b). support high density of vehicles.
In Gjendemsjo et al. (2006), Chandrasekhar and She (2008), non- The authors of Ciou et al. (2015) propose a Greedy Throughput
orthogonal resource sharing is discussed, considering the maximum Maximization plus (GTM+) algorithm for resource allocation. It
transmit power constraint of the user devices. The authors of Wang considers multi-sharing resource allocation in cellular networks con-
et al. (2011) propose an optimal resource allocation technique that is taining D2D links. GTM+, which is an iterative algorithm, maximizes
able to significantly improve the sum throughput of D2D communica- throughput by taking into consideration conflict graph and maximal
tion as well as the cellular communication, in a network. In order to weight independent set. The system performance is better with the
improve the overall network throughput and user satisfaction ratio, the GTM+ algorithm, compared to other algorithms, in terms of system
authors of Chen et al. (2015) introduce a time-division scheduling throughput, percent D2D pairs permitted, and running time. The
algorithm for efficient utilization and allocation of resources, using authors of Chen et al. (2014) propose a distributed resource allocation
non-orthogonal sharing mode. In Liu et al. (2013), the authors propose and power control strategy for resource allocation to enhance the
a semi-persistent resource reuse algorithm in inter cell as well as intra system capacity. A brief overview of some of the mentioned algorithms
cell scenarios for improvement of throughput of the network. The has been given in Table 5. Different algorithms proposed by the various
authors of Yu and Tirkkonen (2012) introduce a rate splitting based researchers target optimization of various parameters. This has been
resource sharing algorithm, in which the message being transmitted is depicted in Table 6, which clearly depicts the particular parameter
divided into a public part and a private part. The private part of the maximized for every algorithm.
message can be decoded only by the receiver for which it is intended. On the basis of a comparative analysis of the various algorithms
The public part is available to all receivers. In comparison to other discussed so far, we have observed that compared to other well known
algorithms, it is possible to achieve higher sum rate with rate splitting schemes, (Lin et al., 2013; Liu et al., 2013) (Particle Swarm optimiza-
mode. A dynamic resource allocation scheme has been proposed in Liu tion and Semi-Persistent Scheduling) have provided the best perfor-
et al. (2015) with joint beamforming, routing and flow control. It is an mance. Taking throughput as the parameter of comparison, a compar-
efficient method for reducing the signaling overhead, and the overall ison of some of the discussed algorithms has been shown in Fig. 8.
latency. Other parameters for comparison of the algorithms can be: D2D access
The authors of Lin et al. (2013) aim at maximization of the system rate, user satisfaction ratio, spectral efficiency, energy efficiency etc (as
throughput, considering the minimum data rate requirement. To shown in Table 6). From the graph, it is clear that for the same
obtain the solution, particle swarm optimization (Kennedy, 2010) simulation parameters, maximum system throughput is achieved
method is used. D2D communication underlying cellular network is by Lin et al. (2013).
considered, and the results show that the system performance can Hence, it is one of the most efficient resource allocation algorithms
greatly improve with this technique of mode selection and resource of available in literature.
allocation. A two-phase resource allocation scheme for D2D commu- Any resource allocation strategy needs to follow a sequence of steps,
nication underlying cellular networks has been proposed in Le (2012). for optimal allocation of resources and throughput maximization.
The first phase performs optimal resource allocation, on the basis of the These have been depicted below, by using a simple network model

17
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 5 Table 5 (continued)

Algorithms for resource allocation in device-to-device (D2D) communication.
Ref. No. Algorithm Description of Objective met
Ref. No. Algorithm Description of Objective met algorithm
algorithm
Tirkkonen, Based Resource splitting is used and spectrum
(Zhou et al., Intra cluster A D2D Optimal resource 2012) Sharing the algorithm efficiency is
2013) D2D retransmission utilization and Algorithm coordinates achieved.
retransmission scheme (intra enhancement of communication
scheme cluster) where retransmission between the UEs
cooperative relays throughput when the distance
are adaptively get between them is
selected through large. The
multicast interfering links in
retransmissions, the transmission
exploiting the between DUEs are
multichannel coordinated using
diversity Successive
(Botsov et al., Heuristic A cellular offloading Reduced Interference
2014) Location and capacity signaling cancellation (SIC)
Dependent enhancing solution overhead and (Lin et al., 2013) Particle swarm Joint mode To maximize
Resource for D2D minimal optimization selection and system
Allocation communication interference with resource resource allocation throughput,
algorithm underlaying cellular the primary allocation for enabling flexible while
networks, by network reuse of cellular guaranteeing
application of resources by DUEs, minimum data
persistent resource based on particle rate requirement
allocation a swarm optimization
vehicular network (PSO-MSRA),
supporting D2D guaranteeing
communication; the minimum rate
algorithm involves requirements of
strict QoS UEs
requirements (Le, 2012) Two-phase Takes into Optimal resource
(Phunchongharn Column A heuristic Enhanced resource consideration sub allocation with
et al., 2013) generation algorithm that spectral allocation channel allocations efficient
method Detection of efficiency and Scheme for cellular flows in exploitation of
maximum number reduced the first phase and the spectrum
of active D2D links, transmission D2D flows in the
using a heuristic length of the D2D second phase
approach; The links links on (Sun et al., 2015) Incentive Three step resource Optimization of
must be capable of increased power resource allocation scheme: energy efficiency,
transmitting consumption allocation relay selection, with cooperation
simultaneously in algorithm setup of two-user among UEs
every time slot and case, and finally
satisfy the solving resource
constraints of the allocation problem
access pattern by two-dimensional
(Feng et al., 2013) Admission Admission control To improve search
control and performed followed spectral (Wang et al., Proportional Fair A scheme To increase
power allocation by power allocation efficiency, and 2013) Scheduling considering central- system
to admissible D2D overall Resource controlled D2D throughput and
pairs and its cellular throughput of the Allocation network in which maintain fairness
user partners. Then network Scheme boundaries of among D2D pairs
maximum weight resource-
bipartite scheme for assignment and
finding suitable contention areas are
D2D pair for each dynamically
CU adjusted
(Chen et al., Time Division Base station's entire Improve system (Zheng et al., Adaptive Time Fixed length time Higher system
2015) Scheduling scheduling time throughput and 2015) division are considered in throughput and
(TDS) Algorithm divided into fixed D2D user Scheduling which D2D pairs are better fairness in
number of time satisfaction ratio Algorithm scheduled comparison to
slots, with a adaptively on the other existing
balanced allocation basis of algorithms.
of the D2D pairs in proportional
the slots using a fairness algorithm
location dispersion (Ciou et al., 2015) GTM+ Algorithm Multi-sharing To maximize
principle in order to for resource resource allocation system
avoid interference allocation algorithm which throughput, and
(Liu et al., 2013) Semi-Persistent Uplink resources are Reduced allows any CUE to number of
Scheduling (SPS) utilized by the D2D interference share its resources permitted D2D
Resource Reuse users, to minimize levels with with multiple DUEs pairs
Algorithm interference within improved system (Chen et al., Distributed Resource allocation Significant
the cellular throughput 2014) resource and power control improvement in
networks(inter cell allocation and are jointly system capacity,
as well as intra cell power control optimized. A with high
interference) algorithm, with stackelberg game convergence of
(Yu and Rate Splitting Han-Kobayashi rate High per user stackelberg game model is used to the distributed
(continued on next page) (continued on next page)

18
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 5 (continued)

Ref. No. Algorithm Description of Objective met

algorithm

framework solve the algorithm

optimization
problem
(Fodor et al., Resource Pooling Allows resource Enhanced
2012) Algorithm reuse between D2D throughput and
and cellular users, better spectrum
taking advantage of utilization
hop gain

Table 6
Comparative analysis of resource allocation algorithms (Comparing Performance
Metrics).

Ref No. Parameter optimized

System Spectral Energy Fairness ratio

throughput efficiency efficiency

(Fodor et al., 2012) ✓ ✓ × ×

(Zhou et al., 2013) ✓ × × ×
(Botsov et al., 2014) ✓ × × ×
(Phunchongharn × ✓ ✓ ×
et al., 2013)
(Feng et al., 2013) ✓ ✓ × × Fig. 9. Network Model.
(Chen et al., 2015) ✓ × × ×
(Liu et al., 2013) ✓ × × ×
considered with fast fading as a result of multipath propagation and
(Yu and Tirkkonen, × ✓ × ×
2012) slow fading as a result of shadowing. All D2D pairs have a cellular reuse
(Lin et al., 2013) ✓ × × × partner, for sharing of resources. Rij represents the resource reuse
(Le, 2012) × ✓ × × indicator, having a value 0 or 1. The system is assumed to have the
(Sun et al., 2015) × × ✓ × same number of resource blocks as the number of cellular users.
(Wang et al., 2013) ✓ × × ✓
(Zheng et al., 2015) ✓ × × ✓
Taking the channel gain between base station and ith cellular user
(Ciou et al., 2015) ✓ × × × as Gb, cu(i); channel gain between a D2D transmitter and receiver (pair
(Chen et al., 2014) ✓ × × × j) as GD(j); the interference link gain from the base station (BS) to jth
D2D pair as Ib,D(j), and gain of interference link from jth D2D pair's
receiver to ith cellular user be ID(j), cu(i), the network model is
characterized by the following set of equations. The transmission
power of the ith cellular user and jth D2D pair are denoted by Pcu(i)
and PD(j), respectively.
For setting up of D2D pairs in a cellular network, the minimum
SINR requirements need to be fulfilled. The interference caused to the
cellular users must be less than the threshold, to ensure a balanced
cellular network performance. The SINRs of ith cellular user and jth
D2D pair are given as
Pcu (i ) Gb, cu (i )
cu (i)=
σ0 + PD (j ) Rij ID (j ), cu (i ) (1)

PD (j ) GD (j )
D (j )=
σ0 + Pcu (i ) Rij ID (j ), cu (i ) (2)
where Rij € {0,1}, such that.
Rij=1 when jth pair shares resources with ith cellular user,
otherwise 0.
The minimum SINR requirements for the D2D pair and cellular
user within a network are stated as,
Fig. 8. Comparison of Throughput of Different resource allocation algorithms.
min
D (j ) ≥ έ D (3)
and mathematical equations. min
cu (i ) ≥ έ cu (4)
1) Network model In order to ensure fulfillment of these SINR conditions, when no
resources are shared between D2D users and cellular users, minimum
A single cell scenario with the base station at the centre is transmission power of the cellular users must be
considered having n number of D2D pairs and m number of cellular
min
users (Fig. 9). In D2D underlaying cellular communication networks, έ cu σ0
Pcumin(i ) =
interference between the CUEs and DUEs is high. This is controlled by Gb, cu (i ) (5)
the base station, acting as a controlling entity. A path loss model is
Similarly, for the D2D user, the minimum transmission power is

19
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 7 algorithm. Efficient spectrum utilization is not addressed in this

Symbols used in network model. technique and needs further investigation. For context aware resource
allocation, an energy efficient algorithm has been proposed in Zhou
Symbol Meaning
et al. (2016), based on nonlinear fractional programming and Lagrange
Gb, cu(i) Channel gain between the BS and cellular user dual composition. A number of power control strategies, applicable to
GD(j) Channel gain between the D2D pair D2D communication are studies in Fodor et al. (2013). LTE power
Ib,D(j) Channel gain of interference link between BS and D2D pair
control scheme is applied to a hybrid cellular D2D network and then
ID(j), cu(i) Channel gain of interference link between ith Cellular user and jth
D2D pair compared with a distributed power control scheme. The LTE power
έcu(i) SINR at ith CU control scheme is highly flexible, as stated by the simulation results. A
έD(j) SINR at jth D2D pair distributed power control scheme, aiming utility maximization, is
min
D
Minimum SINR required for a D2D pair proposed in da Silva et al. (2014). Mode selection and resource
cu (i ) Minimum SINR required for a cellular user allocation constraints are taken into account for balancing spectral
min
Pcu (i )
Cellular user's minimum transmission power efficiency and energy efficiency. It is an important algorithm for
PDmin
(j )
D2D pairs minimum transmission power maximization of proximity, reuse and hop gains, as well as for
Γcu(i)D(j) Throughput of ith cellular user after sharing resources with jth D2D extending cellular coverage. The proposed algorithm is efficient for
pair D2D networks which do not employ traditional mode selection
Γcu(i) Throughput of ith cellular user not sharing resources with D2D pair
ΓD2D Overall System throughput
schemes and power control schemes.
σ0 Noise Power on each channel A near optimal low complexity scheme for power control has been
proposed in Song et al. (2015). In Lee et al. (2015), two types of
algorithms are proposed for power control: centralized and distributed.
given as These are effective for coordinating the interference in D2D underlaid
min cellular networks. A distributed scheme for controlling power in D2D
έ D σ0
PDmin
(j ) = communications has been proposed in Fodor and Reider (2011). It
GD (j ) (6) considers a D2D underlay scenario and is an efficient technique for
When a cellular user does not share any resources with the D2D minimizing the overall power consumption. Using Augmented
pair, then throughput is represented as, Lagrangian Penalty Function (ALPF) method, optimal SINR targets
are achieved in this algorithm. For energy efficient D2D communica-
Γcu (i ) = log2 (1 + έ cu(i) ) (7)
tion, a power control algorithm has been framed as a non-convex
th optimization problem in Jiang et al. (2015). Remarkable improvements
When ith cellular user share resources with j D2D pairs, the
throughput is given by, are attained with this algorithm, in terms of energy efficiency. In Tang
et al. (2015), power control is considered in a D2D scenario in which
Γcu (i ), D (j ) = log2 (1 + έ D(j) )+ log2 (1 + έ cu(i) ) (8)
the entire problem is divided into two parts: mitigating co-channel
Thus, overall system throughput is given by interference by allocating channel from graph perspective; link assign-
ment (LA) and power control (PC) are jointly optimized. For optimiza-
ΓD2D = Γcu (i ), D (j )−Γcu (i ) (9)
tion of LA, genetic algorithm (GA) is used. A single cell scenario
Optimal resource allocation to the D2D pairs underlying cellular resulting in an optimal power control is considered and evaluated in Yu
network can be achieved by maximizing the throughput gain of the et al. (2009). A binary power control (BPC) scheme is proposed in da
system, which can be stated as the following optimization problem Silva and Fodor (2015), in which the resources of a single CUE are
shared by multiple DUEs. BPC does not involve any iteration and is
argmaxΓD2D (10)
much simpler, providing optimal transmits powers. A power control
for all i € m and j € n. scheme for supporting vehicle-to-vehicle (V2V) communication in D2D
The various symbols used in the described model are given in networks has been proposed in Ren et al. (2015). Mobility of devices
Table 7. affects the efficiency of the system, in terms of energy, and is addressed
in Wu et al. (2014). For maximizing energy efficiency, a distance based
4.3. Power control and optimization mobile association algorithm is proposed in Xiao et al. (2016), for D2D
enabled HetNets. This scheme effectively meets the trade-off between
As stated in the network model, controlling power of UEs is energy efficiency and complexity. A hybrid power control scheme has
essential for efficient resource utilization. Blindly applying D2D com- been proposed in Belleschi et al. (2015) which interference to the
munication to cellular networks can result in degradation of system cellular users, by the D2D users is limited by utility optimal scheme.
performance. The D2D users need to limit their power so that the SINR The sum transmission power is minimized, while maximizing the
requirements of the cellular users can be effectively met. Controlling spectral efficiency. A summary of different power control techniques
power is the key to mitigating interference between the users in a in D2D communication is given in Table 8.
cellular network (Min et al., 2011). The overall system capacity can be Since the UEs are battery operated, energy efficiency is also a major
enhanced by limiting the transmit power, as discussed in Nguyen et al. concern. A survey of energy optimization in the next generation
(2014), in case of D2D communication in MIMO networks. Due to networks has been presented in Abrol and Kumar Jha (2016), focusing
limited battery capacity, energy efficiency is critical to be achieved. For upon the rising need for green communication. The tradeoff between
improved system performance, few power control methods have been saving the device battery and achievable QoS of the network must be
reviewed in Wen et al. (2012), Dongyu Wang and Wang (2013). A low maintained. Efficient power optimization can be achieved using
complexity algorithm has been proposed in Wang et al. (2014), which massive MIMO (Gupta and Kumar Jha, 2016). A fast converging
considers game-theory as its basis. The source selection and controlling power control scheme for D2D Massive MIMO systems has been
power levels of the links are decided using the game theory. In this, the evaluated in Xu et al. (2016). Some guidelines for energy efficient
Stackelberg game model is used to demonstrate the impact of D2D architectures have been provided in Zhang et al. (2015), in which
improvement in the quality of D2D transmission quality. For a multi- energy efficient technologies have been discussed from different
cast D2D communication, joint power and channel allocation is aspects, and provide a good platform for researchers to ahead in the
proposed in Meshgi and Zheng (2015), using a maximum weight direction of energy efficient D2D networks.
bipartite matching scheme along with a low complexity heuristic In Fig. 10, the graph depicts a comparison between the transmis-

20
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 8 sion power of UEs in cellular as well as D2D modes, with respect to
Power control algorithms in D2D communication. their positions within a cellular system. With D2D communication
underlaying cellular networks, there is considerable power saving in
Reference Algorithm Description Objective
No. the cellular network, as depicted. This benefit of D2D communication
has been investigated upon by the various algorithms discussed above.
(Wang Game theory based Stackelberg game Improvement in Thus, overcoming the challenge of power control in device-to-device
et al., power control model is used for the D2D
(D2D) communication is essential for the overall power optimization in
2014) scheme selecting source and transmission
controlling the power quality the network.
(Song et al., Joint Power and Maximize the sum rate Coexistence of
2015) Rate control for of the cellular users in both cellular and 4.4. Interference control
D2D the network, D2D users
communication in guaranteeing Quality of contributing
Device-to-device (D2D) communication, a competent technology
cellular networks Service (QoS) of D2D towards
communication at the improving system for the next generation networks (NGNs) promises lower delay, lower
same time throughput transmission power, increased system capacity, etc. Integration of D2D
(Lee et al., Centralized and Centralized power Improved technology with cellular networks results in interference, due to the
2015) distributed power control ensures throughput
sharing of same resources between them. When the cellular downlink
control algorithms sufficient coverage performance of
using stochastic probability of cellular cellular users resources are shared by the D2D users, the cellular users and their
geometry users and distributed achieved neighbors suffer interference. The receiver of the D2D pair is also
method maximizes sum susceptible to counter interference by the base station.
rate of D2D links When sharing cellular uplink resources, the base station suffers
(Fodor and Distributed Power ALPF method is used , Minimize the
interference by the D2D users. Interference between D2D users and
Reider, Control Scheme with optimal SINR overall power
2011) target, to control the consumption of cellular users may be intercell or intracell and needs to be mitigated for
user equipment power the network proper functioning of D2D communication in cellular networks. The
(Jiang Joint resource Penalty function Maximizing existing interference management schemes are not capable of inter-
et al., allocation and approach is adopted for energy efficiency
ference mitigation in the next generation networks. A survey, along
2015) power control power control in the of the network
technique network; the problem
with a comparison of various existing schemes has been provided
is formulated as a non- in Hossain et al. (2014), suggesting guidelines for modification of the
convex optimization various schemes and efficiently deal with the next generation networks.
problem, and solved Interference mitigation is possible through proper mode selection,
using a two-layer
optimum resource allocation and power control of the devices. The
scheme
(da Silva Binary Power A simple algorithm Improve spectral various approaches for interference mitigation are categorized into
and Control Scheme achieving near optimal efficiency and three types:
Fodor, utility , efficiently power efficiency
2015) supporting a large
number of D2D pairs in
• Interference Cancellation Techniques, which use advanced coding
the cellular network
and decoding methods for cancellation of interference signals at the
(Xiao et al., Distance based Access point selection, Better tradeoff CUE or DUE. Interference cancellation techniques have the ability to
2016) mobile association mode switching, relay achieved between enhance the capacity of cellular networks.
Scheme selection and power
control, all are
energy efficiency
and complexity, in
• Interference Avoidance Techniques, which avoid interference be-
tween cellular links and D2D links by using orthogonal time-
considered and basis of comparison to
this technique is the other techniques frequency resource allocation techniques.
location of UEs in the
network
• Interference Coordination Techniques, which mitigate interference
between D2D and cellular links with power control schemes and
(Xu et al., Revised Graph A fast converging Considerable proper scheduling.
2016) coloring based pilot algorithm proposed to reduction in pilot
allocation overcome pilot overhead
contamination by A number of algorithms have been proposed to overcome the
limiting D2D problem of interference in D2D underlaid cellular networks.
transmission power. It Interference management is studied in Guo et al. (2015), by consider-
is an optimal choice for
ing the Interference suppression area's range. Adjusting the interfer-
practical
considerations. ence suppression area adequately provides an optimal system perfor-
mance. Interference aware interference mitigation methods in D2D
communication have been discussed in Kwon et al. (2014), where the
interference information is known to the UE by the base station or
through a blind estimation. It is an efficient technique for improving
the system throughput. Another interference aware scheme has been
proposed in Janis et al. (2009), based on Hungarian algorithm and
aims at maximization of the number of permissible D2D pairs. The
authors in Yin et al. (2015) propose an interference coordination
scheme by joint spectrum allocation and power control. This scheme
guarantees Quality of service to the D2D users as well cellular users. An
interference cancellation technique has been proposed in Zhou et al.
(2015), where authors maximize the utility of the network by con-
sidering transmission powers of users in the network. In Xiao et al.
(2011), the authors propose a technique for interference reduction by a
Fig. 10. Power saving in D2D mode.
power optimization scheme.
Most of the existing literature on D2D communication exclusively
focuses on single-antenna systems. However, in the next generation

21
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

networks, multi-antenna deployment will gain importance. MIMO Table 9

transmission schemes are introduced for interference avoidance, Possible security attacks in D2D communication.
in Jänis et al. (2009), which results in a considerable SINR improve-
S.No Security attack Description
ment. In Ni et al. (2016) D2D system underlaying cellular networks is
analyzed for interference cancellation, as well as beamforming, with the 1. Eavesdropping Unauthorized interception of the confidential
base station equipped with multiple number of antennas. The users data between the DUEs
2. Denial of Service Attempt to make resources unavailable for
however are equipped with a single antenna. Ergodic process is
(DoS) authorized UEs
adopted for analysis of such a system, along with some suitable 3. Man-in-the-Middle The malicious node acts as a client for the server
approximations. The results of ergodic achievable rates are very close Attack (MITMA) and a server for the client, i.e. a D2D receiver to
to Monte-Carlo results. the D2D transmitter and vice-versa
In order to achieve the required cell edge capacity in the next 4. Free-riding attack Selfish UEs receive data from their pairs, and at
the same time are not willing to share their own
generation networks, advanced interference management schemes are
resources with other UEs because of energy
crucial. With inter cell inference becoming common in D2D underlaid consumption in the process, resulting in reduced
cellular networks, a promising approach to overcome this are coopera- system availability for D2D communication
tive communication. This fact is exploited in Shin et al. (2013), which 5. Node Impersonation Malicious node acts as a legitimate node, posing
threat to D2D communication
propose an interference mitigation scheme, D2D communication
6. Malware Attacks Hostile software, targeting the D2D pairs to
assisted interference alignment. A realistic multi-cell scenario has been gather sensitive information
considered in Batista et al. (2015), to propose an interference mitiga- 7. IP Spoofing Manipulation of IP packets, particularly the
tion technique for D2D communication. A large amount of literature is headers, by the malicious nodes
thus available on interference management in D2D communication. An 8. Bandwidth Spoofing Unauthorized access to the bandwidth of the
legitimate user, by the malicious node
extensive survey on interference management in device-to-device
9. Inference Attack Privacy is attacked, by studying either logically or
(D2D) communication has been presented in Mach et al. (2015), statistically, transmission patterns between
Noura and Nordin (2016). If the interference is not managed properly, devices
it can hamper the performance of cellular networks. Effective inter- 10. Trust Forging Attack Intruder forges its trust value for attracting D2D
transaction requests
ference management strategy can result in enhancement of the overall
11. Location Spoofing Fake GPS Signals sent by the attackers, causing
capacity of cellular networks (Min et al., 2011). Advanced D2D (GPS Spoofing) arbitrary location s to be chosen by a subscriber
receivers can be used to mitigate interference. However, some chal-
lenges exist for interference cancellation (Halperin et al., 2007).
Forward error correction (FEC) decoding and error recovery can cancel Possible attacks on the direct links and some algorithms for securing
interference in D2D networks and are an efficient technique for them are discussed herein.
interference management in the future networks. The heuristics related
to interference need to be explored. An overview of some of the 1) Possible Attacks in D2D
important existing algorithms for managing interference in D2D
scenarios has been provided in this section. Still, this issue remains In wireless communication, D2D is susceptible to a number of
an open challenge for the researchers. There is a promising future security attacks, like, eavesdropping, data fabrication or alternation,
research direction in this field. denial of service attack (DoS), free-riding attack. Some possible attacks
are briefed in Table 9. The eavesdropping attack in D2D communica-
4.5. Security in D2D tion is depicted in Fig. 11, where an eavesdropper is present in the
network, trying to take up information from the senders. This figures
In conventional cellular communication, the UE is first identified, depicts a general attack on a D2D network.
authenticated and then encryption of the radio link occurs. The core
network is a trusted party. But this is not the Case in D2D commu-
nication as transmission occurs without the assistance of the core
network. Also, wireless channels are broadcast in nature. As a result,
they are susceptible to a number of attacks, making security an
important concern. Cryptographic solutions are needed to secure the
information, when it passes through the wireless channels. The security
schemes provided by the cellular operators can be used by the D2D
users if they are under their coverage. But, users outside the coverage
of the operators can’t be secured. In such a case, security signals may
be passed on through relays. It is well known that relays are highly
susceptible to malicious attacks; therefore, designing security schemes
for D2D communication is an important challenge to be addressed. Cha
et al. (2009), Yue et al. (201), Perrig et al. (2004), Zhou et al. (2008),
Muraleedharan and Lisa (2006) discuss about the security concerns in
D2D communication. Security in D2D has been discussed in detail in
the following section (Section 5).

5. Securing D2D communication

Facilitating D2D communication in cellular networks is beneficial,

in terms of throughput enhancement, energy consumption, etc. In spite
of the extensive ongoing research in the field of D2D communication,
security aspects are less addressed. An attacker can attack the core
network, or the user applications, and extract useful information,
thereby hindering the information exchange over the direct links. Fig. 11. A D2D scenario, with eavesdropper.

22
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

2) Secrecy Algorithms in D2D Communication trusted relay, via a secure beam forming design, which provides high
level of secrecy. The DUEs are not powerful enough to prevent attacks
To make D2D communication secure, physical layer security plays a from eavesdroppers, and fighting for attacks upon them. Wang et al.
key role. The concept of physical layer security came into picture to (2015) present a heuristic genetic algorithm to select jamming partners
describe the physical characteristics of wireless channels. Sensitive for DUEs in the network. The secrecy rate has been maximized, by a
information may be retrieved by the eavesdropper, in a communication two step approach: power allocation for worst case eavesdropper and
system. Physical layer security employs techniques which exploit the best jamming partner selection, which is clearly depicted in the
characteristics of the wireless channel, modulation, coding, and multi- simulation results. Using a cryptographic approach, Zhang et al.
ple antennas to avoid eavesdropping. Physical layer security is inves- (2015) introduce a secure data sharing protocol. This uses symmetric
tigated in Zhang et al. (2016) to improve secrecy performance of the encryption, along with public key based signature to achieve security.
system. The authors in Zhang et al. (2015) investigate physical layer Security goals fulfilled by this protocol are data confidentiality,
security in Cooperative D2D networks. Cooperation among cellular integrity, transmission and reception non-repudiation, entity authenti-
users and D2D users is formulated as a coalition game, based on Max- cation and free riding resistance. Accounting social relationships and
Coalition order for an efficient cooperation, along with guaranteed proximity between devices, a game theoretic approach is used in the
secrecy in the network. The physical layer security is proposed, proposed solution to improve the system performance, in terms of
considering an energy-constrained D2D transmitter. Wireless power throughput and latency. It is also efficient in enhancing network
transfer (WPT) model is used to ensure secrecy of transmissions. In coverage. The work in Wang and Wu (2015) investigates upon selection
order to ensure secure D2D communications, certain security require- of jamming partners for D2D users by taking into consideration full
ments need to be fulfilled. and partial channel state information (CSI). Heuristic genetic algo-
Another concept is secrecy capacity, which quantifies the security of rithm and low complexity optimization problems have been proposed
transmission. It defines the maximum rate of faithful information sent for maximizing the worst case eavesdropping in D2D communications.
from transmitter to receiver, in the presence of threat from eaves- Trust and social aware cooperation is mandatory for adoption of
dropper. In Zhang et al. (2014), secrecy capacity in cellular networks D2D communication on a large scale (Ometovy et al., 2016). For
with underlaying D2D communication has been explored. A graph- ensuring secure D2D communications, sociality of users and trust
based model is used for radio resource allocation problem and factors play an essential role. The trustworthiness of a device is given in
formulated as a matching problem in a weighted bipartite graph. To terms of its trust value, adjusted by the BS, as per the last transaction
obtain maximum secrecy capacity for DUEs and CUEs, the Kuhn over the D2D link. For any user, mth transaction of D2D pair, it is given
Munkres (KM) algorithm has been considered. This algorithm provides by using the Gompertz function (Kenney and Sydney Keeping, 1954),
better performance, in comparison to random or greedy spectrum as,
sharing.
ta (m)= ( −1t
For ensuring secure D2D transmissions, encryption of messages by a (m −1) )+∆ta (m ) (11)
cryptographic techniques is required. A secure key agreement protocol
where ᶃ(z) denotes the Gompertz function, for a vector z, denoted by
has been presented in Shen et al. (2014), enabling establishment of a
shared secret key between UEs, and is based on Diffie-Hellman key (z )= e−e
−CG (z )
(12)
agreement protocol. The main advantage of this protocol is low
computation and communication overhead. The simulation results where
show that a higher level of usability is possible with this protocol. CG(z) determines sensitivity of the trust value variation, along time.
With the increase in the number of eavesdropper, the D2D perfor- Also ∆ta (j ) denotes the device's behavior at mth transaction.
mance degrades. A solution is proposed in Abd-Elrahman et al. (2015) Relationships based on trust among users have been extensively
to secure D2D discovery and D2D communication in cellular networks, studied, of which an important one is discussed in Urama et al.
based on Identity Based Encryption (IBE). Ensuring security in cellular (2016). D2D communications require a sophisticated control to resist
networks with D2D communications is an open research area. This eavesdropping and other security attacks. Malicious behavior of devices
scheme is not useful, for multi-domain networks. As a result, Abd- in a D2D environment resulting in data loss is studied in Militano et al.
Elrahman et al. (2015) investigate upon ensuring security in discovery (2016) and 86% reduced data loss is achieved with the proposed trust
and communication phases of D2D communication in multicast group bases coalition game. Another technique to limit the impact of
communication. Group Key management has been studied, for ensur- malicious nodes has been proposed in Militano et al. (2016), and is
ing the security. This mechanism ensures high security, in comparison based on trust solutions. Simulation results show how filtering of
to other methods. In Liu et al. (2016), security in cognitive D2D malicious nodes is done with the proposed scheme, thereby increasing
networks is considered. Authors in Wang and Yan (2015) propose successful cooperative interactions in small-scale IoT environments.
security architecture for D2D communication and the requirements to Security of direct communication is a key concern, which determines
ensure security have also been suggested. Various applications and use ultimate adoption of this technology. Spatial and social notions of
cases of D2D communication have also been discussed. proximate users are addressed in Ometov et al. (2016) for improving
Since security is a major concern for device-to-device (D2D) the system performance using game-theoretic mechanisms. Although
communication, for D2D assisted cellular networks, secure throughput there was increased overhead, but seamless connectivity was provided
optimization solution has been proposed in Ometovy and Orsino to users outside the network coverage as well. This clearly depicts the
(2016). Public safety D2D communication over LTE Hetnets have been impact of social and spatial notions, on the performance of D2D
investigated in Tata and Kadoch (2014) for addressing the issue of networks.
secure network coding. Using data splitting mechanism, an algorithm The various security algorithms for device-to-device (D2D) com-
is presented, and named as Secure Network Coding based Data munication have been listed in Table 10. Security algorithms allow
Splitting Algorithm. It avoids passing of information to the eaves- transmission and reception of the information among intended users
dropper, without increase in overhead of the system. The sequence of only. Algorithms in Zhang et al. (2015) and Wang et al. (2015) prove to
the sent packets is known only to the source and destination, thus, not be highly efficient for securing D2D links. A comparison of achievable
allowing any useful data to be known to the attackers. A technique to secrecy capacity by different algorithms has been shown in Fig. 13.
prevent eavesdropping in MIMO D2D communication is investigated From the above discussion, it is quite clear that a very challenging task
in Jayasinghe et al. (2015). A relay in the network performs physical for the research community still remains securing the direct links.
layer network coding (PNC). The devices communicate through this Possible attacks on the links are becoming more threatening with the

23
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Table 10 Table 10 (continued)

Security algorithms for D2D communication.
Reference Algorithm Description Objective
Reference Algorithm Description Objective No.
No.
allocation ; based on in better system
(Shen et al., Diffie-Hellman A key agreement Secure key setup the social trust policy performance
2014) Secure Key protocol is designed, between two mobile of the jammer with
Establishment based on Diffie- users , with low minimum social
using Hellman key computation cost trust index threshold
agreement protocol, and overhead; being 0.4
and is integrated to maximum (Zhang et al., Secure Data Detects free-riding Maintain data
Wi-Fi Direct probability of attack 2015) Sharing Protocol attacks by maintain confidentiality,
protocol. It is also is 2−k, where k is the record of the UEs in integrity, provide
implemented using no. of bits used for the network with the free-riding
Smartphone authentication use of digital resistance, entity
strings signature and authentication
(Tata and Secure Network Data Splitting No important symmetric
Kadoch, Coding based mechanism is information encryption
2014) Data Splitting applied over a received by the (Urama Proof of Concept In absence of Helpful in
Algorithm butterfly network, to eavesdropper; low et al., Application availability of a evaluating social
ensure computation 2016) Development cellular link to the trust in D2D
confidentiality of the overhead D2D users, secrecy communication
transmitted can be ensured , the
information proposed technique
(Jayasinghe Securing MIMO Using a beam A fast converging assures security to
et al., D2D forming technique, algorithm, providing the D2D users by
2015) Communication information passing high secrecy levels establishment of
through a relay is for D2D dynamic connections
secured by ensuring communication
the presence of a
trusted relay in the growth of technology. The next generation networks are becoming
network. The trusted
denser, making easier for intruders to spoof information and the
relay is safe as it
performs PNC disrupt overall functioning of the networks. A proposal for securing
(Abd- Securing D2D Investigates security Less overhead with direct links using IP Sec has been given in the following sub-section.
Elrahman communication in D2D discovery the use of a single
et al., using Identity- and communication hash function,
3) Security concern for UDNs and Proposed Solution for Securing
2015) based Encryption phases to enhance thereby enhancing
the level of security overall security of
direct links
of the network the network
(Abd- Group Key Dynamic approach Ensure high security In the next generation networks, densification of networks is
Elrahman Management to Secure messages essential for meeting the needs of the subscribers. This densification
et al., (GKM) on direct links,
results in ultra dense networks (UDNs). In such heterogeneous net-
2015) Mechanism during discovery and
communication works (HetNets), with UDNs and D2D communication, there are high
phases by building a chances of jamming attack (Grover et al., 2014). The jammer emits a
strong protocol for radio signal causing a continuous interference mostly in the form of
group key
noise, in the network. A scenario of jamming in UDNs with D2D
management
(Zhang et al., Merge and split Physical layer Improved secrecy
communication underlaying the cellular networks is depicted in
2015) based coalition security is rate of the system Fig. 12. When devices communicate through direct links, the distance
formation investigated In D2D and social welfare between devices for information exchange is less (proximity devices).
algorithm communication from As a result, it becomes easier for the jammer to hamper the commu-
a cooperative
nication. The jammer (intruder) can attack the network by inserting an
perspective, with the
formation of a interference signal (single-carrier jamming) in the entire system or
coalition formation jamming selective carriers (multi-carrier jamming). This issue needs a
game, to ensure special attention.
cooperation between
In order to ensure integrity, confidentiality and authentication of
cellular users and
D2D pairs
data, security mechanisms are applied at the internet protocol (IP)
(Liu et al., Wireless Power Investigate physical Improved values of layer. This is referred to as IP Security (IP Sec). It supports two
2016) Transfer (WPT) layer security in D2D secrecy throughput, encryption modes: Transport and Tunnel. The transmitted data con-
Policy for securing communication secrecy outage tains the header and payload. Transport mode encrypts only the
Cognitive D2D using three wireless probability and
payload, while Tunnel mode encrypts data as well as the payload.
communication power transfer power outage
policies: cooperative probability, but Thus, Tunnel mode is of more importance. A solution for securing D2D
power beacons, best demands more networks from eavesdropper has been proposed and depicted in
power beacon and overhead Fig. 14, which is based on IP Security. The figure depicts ongoing
nearest power communication between a D2D pair, which is hindered by an intruder
beacon, to ensure
secure cognitive D2D
by spoofing the bandwidth. Without a security mechanism applied to
communication in the communication, the intruder would obtain the information from
cellular networks the channel, between the D2D transmitter and D2D receiver. Using IP
(Wang et al., Genetic Algorithm An optimal jammer Increase in secrecy security, the information can be prevented to be sent to the unauthor-
2015) for maximizing selection scheme is rate with increase in
ized user, with both transmitter and receiver sharing a public key.
secrecy rate proposed with the number of
optimized power jammers, resulting Thus, the transmitted information can be safeguarded. The D2D pairs
(continued on next page) have a small distance between D2D transmitter and D2D receiver.

24
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Fig. 12. Jamming in Ultra dense networks with D2D communication.

Fig. 13. Comparison of secrecy capacity from different security algorithms in D2D.

Thus, IP Sec can prove to be an efficient technique for securing the

short direct links. The effect of mobility on security in D2D commu-
nication is yet to be further studied. D2D security concerns require
significant research, for assuring secrecy over direct links.
The various aforementioned issues in D2D need to be critically
Fig. 14. Proposed solution for security in D2D, using IP Sec.
addressed, prior to the implementation of device-to-device (D2D)
communication in cellular networks.
As a result, device-to-device (D2D) communication has the ability to
support a large number of use cases like cellular offloading (Bao et al.,
6. Use cases of D2D communication 2010), multicasting (Zhou et al., 2013; Du et al., 2012), video
dissemination (Golrezaei et al., 2012), (Li et al., 2012), M2M commu-
In D2D communication, two devices can communicate either nication (Pratas and Popovski, 2013), public safety services (Usman
directly or via relays, independent of the control from the base station. et al., 2015), V2V communication, location-aware services, social

25
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

networking, smart grids (Fey et al., 2012), D2D for e-health, D2D for networks (UDNs) is posing serious threat to the integrity, confidenti-
smart city etc. Some use cases of D2D communication have been ality and authenticity of the information transfer over direct links. This
addressed in Lei et al. (2012). D2D communication has the ability to needs special attention from the research community. Measures have
support a number of telecommunication applications as well. been proposed, yet, it remains a vital challenge, seeking further
In case of multiuser cooperative communication (MUCC), device- investigation. After this extensive survey on D2D communication, it
to-device (D2D) communication can provide innovative ideas for is quite eminent that security remains an exclusive research field for
enhancement of the network performance. Device-to-device (D2D) practicable implementation of D2D communication. The proposed
communication is an optimal choice for provision of energy efficient solutions and architectures may not be the only one, but are an
communication within cellular networks. If a group of users request the efficient means to target and solve the research challenges.
same services from the base station (BS), they can collectively form a
cluster. The content requested will be served by the base station to the 8. Conclusion
cluster head only, which can then be multi casted to the rest of the
users. This supports energy efficiency within the network and also In this paper, device-to-device (D2D) communication has been
reduces unnecessary overhead. comprehensively outlined. It is a promising technology of the next
D2D communication, in conjunction with Internet of Things (IoT) generation networks, which ensures increased system capacity, en-
results in a massive interconnected wireless network (Bello and hanced throughput, reduced latency and efficient spectrum utilization.
Zeadally, 2014) which has the ability to support numerous applica- Fundamentals of D2D communication have been discussed in depth,
tions. Achieving intelligent D2D communication in the IoT is over- including the basic D2D architecture. A number of key issues in D2D
viewed in Bello and Zeadally (2016). Device-to-device (D2D) commu- communication have been discussed. Architecture for D2D commu-
nication assures low energy consumption, thus supporting energy nication has been proposed, considering the scenario of NGNs, using
efficiency in the Internet of Things (IoT). The remarking potential of directional antennas at the base station, dividing the cell into three
D2D communication in IoT, for the next generation networks is studied sectors. Such a scenario provides better coverage, high spectral
in Militano et al. (2015). A distinctive application is Vehicle-to-Vehicle efficiency and enhanced network capacity, overcoming the various
(V2V) communication, playing a vital role in car crash warning systems open challenges of D2D communication. Possible security attacks on
(Hayami et al., 2015; Bohmländer et al., 2015), brake coordinating D2D communication have been described. The future networks contain
systems, intelligent transport systems (Warabino et al., 2005). Device- dense deployments, resulting in Ultra dense networks (UDNs), causing
to-device (D2D) communication has the ability to serve as an im- high risk of jamming attack in them. A solution for securing D2D
portant component for assuring public protection and disaster relief communication has also been proposed, which is based on IP security
(PPDR) and national security and public safety services (NSPS). The (IP Sec). A few examples have been quoted, where D2D communication
network requirements to support these are discussed in Fodor et al. plays a crucial role. Thus, D2D communication is an integral technol-
(2014). Therefore, D2D communication is an important emerging ogy of the future networks, motivating the researchers to overcome the
technology of the future networks. associated challenges in order to completely take advantage of its
utility.
7. Research challenges in D2D communication
References
Although D2D communication is an open research field for
academicians, as well as industrialists, but still there exists wide range Muraleedharan, Rajani, Lisa A. Osadciw, 2006. Jamming attack detection and
of challenges in its implementation. Modification is needed on the part countermeasures in wireless sensor network using ant system. In: Proceedings of the
Defense and Security Symposium. International Society for Optics and Photonics.
of the user equipments (UEs). This includes: UEs supporting device 3GPP TR 22.803, 2012. Feasibility study for proximity services (ProSe) (Release 12), v.
discovery; Proximity Services (ProSe); and relay functionality in the 12. 2. 0, June
relay scenario. The protocols for device discovery in D2D must consider 3GPP TR 23.703, 2013. Study on architecture enhancements to support proximity
services (ProSe) (Release 12), v. 1. 0. 0 , December.
reduction of energy consumption by the UEs. The issue of UE discovery Gjendemsjo, A., Gesbert, D., Oien, G.E., Kiani, S.G., 2006. Optimal power allocation and
and its power consumption is an open research problem. For efficient scheduling for two-cell capacity maximization. In: Proceedings of the International
utilization of the limited available spectrum, optimal resource alloca- Symposium on Modeling and Optimization in Mobile, Ad Hoc and Wireless
Networks.
tion is required. While allocating the resources to the DUEs in the Abd-Elrahman, Emad, Hatem Ibn-khedher, Hossam Afifi. D2D group communications
cellular networks, for enabling them to communicate, care must be security. In: Proceedings of the 2015 International Conference on Protocol
taken that the same resources are not shared by the nearby CUEs. If it Engineering (ICPE) and International Conference on New Technologies of
Distributed Systems (NTDS), . IEEE, 2015.
is so, chances of interference are very high. Interference management is
Abd-Elrahman, Emad, et al., 2015. Fast group discovery and non-repudiation in D2D
being studied by the researchers but the nature of the interference is communications using IBE. In: Proceedings of the 2015 International Wireless
not completely exploited and understood. Interference management Communications and Mobile Computing Conference (IWCMC).
and avoidance involves addressing issues like collision issues, conges- Abrol, Akshita, Kumar Jha, Rakesh, 2016. Power optimization in 5G networks: a step
towards GrEEn communication. IEEE Access 4, 1355–1374.
tion, and power control. All these issues need to be addressed in an Akhil, Gupta, Kumar Jha, Rakesh, 2015. A survey of 5 G network: architecture and
accurately balanced manner. emerging technologies. IEEE Access 3, 1206–1232.
For enabling any type of communication, channel information must Alkurd, R., Shubair, R., Abualhaol, I., 2014. Survey on device-to-device communications:
Challenges and design issues. In: Proceedings of the 2014 IEEE 12th International
be precisely available. This is a challenging perspective for implemen- New Circuits and Systems Conference (NEWCAS). IEEE.
tation of D2D communication, as, for the direct communication, extra Alliance, N.G.M.N., 2015. 5G white paper. Gener. Mob. Netw. White Pap..
channel information is required, resulting in an increased overhead. Alliance, W., 2010. Wi-Fi Peer-to-Peer (P2P) Specification v1. 1, WI-FI ALLIANCE
SPECIFICATION, vol. 1, pp. 1–159.
Controlling and charging for D2D services is a vital concern for the Alliance, Z., 2006. Zigbee specification, Document 053474r06, Version, vol. 1
service providers. The service providers may charge for security. The Asadi, Arash, Wang, Qing, Mancuso, Vincenzo, 2014a. A survey on device-to-device
modulation format for D2D communication is yet to be addressed. communication in cellular networks. IEEE Commun. Surv. Tutor. 16.4, 1801–1819.
Asadi, Arash, Jacko, Peter, Mancuso, Vincenzo, 2014b. Modeling multi-mode D2D
Addressing these critical problems thoughtfully will help in practical communications in lte. arXiv Prepr. arXiv 1405., 6689.
implementation of D2D communication, in cellular networks. Astely, D., et al., 2013. LTE release 12 and beyond. IEEE Commun. Mag. 51 (7),
A critical research challenge, considering the practicality of D2D 154–160.
B. Raghothaman et al., 2013. Architecture and protocols for LTE-based device to device
communication, is security of the direct links. The D2D links are highly
communication. In: Proceedings of the ICNC, pp. 895–899
susceptible to various security attacks (listed in Table 8). As a result of Bagheri, H.. et al. 2015. Device-to-device proximity discovery for LTE systems. In:
densification of the cellular architecture, jamming in Ultra dense Proceedings of the 2015 IEEE International Conference on Communication

26
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Workshop (ICCW), . IEEE. components in METIS. IEEE Access 4, 3288–3299.

Bai, Bo, et al., 2016. Caching based socially-aware D2D communications in wireless Fodor, Gabor, et al., 2013. A comparative study of power control approaches for device-
content delivery networks: a hypergraph framework. IEEE Wirel. Commun. 23 (4), to-device communications. In: Proceedings of the 2013 IEEE International
74–81. Conference on Communications (ICC). IEEE, 2013.
Bao, X., Lee, U., Rimac, I., Choudhury, R.R., 2010. DataSpotting: offloading cellular Garey, M.R., Johnson, D.S., 1979. Computers and Intractability: A Guide to the Theory of
traffic via managed device-to-device data transfer at data spots. ACM SIGMOBILE NP-Completeness. W. H. Freeman and Co..
Mob. Comput. Commun. 14 (3), 37–39. Golrezaei, N., Molisch, A.F., Dimakis, A.G., 2012. Base-station assisted device-to-device
Batista, R.. et al. 2015. Interference Mitigation Using Band Selection for Network- communications for high-throughput wireless video networks. Proc. IEEE ICC,
Assisted D2D Communications. In: Proceedings of the 2015 IEEE 81st Vehicular 7077–7081.
Technology Conference (VTC Spring). Gregori, Maria, et al., 2016. Wireless content caching for small cell and D2D networks.
Belleschi, Marco, et al., 2015. Benchmarking practical RRM algorithms for D2D IEEE J. Sel. Areas Commun. 34.5, 1222–1234.
communications in LTE advanced. Wirel. Pers. Commun. 82.2, 883–910. Grover, Kanika, Lim, Alvin, Yang, Qing, 2014. Jamming and anti–jamming techniques in
Bello, Oladayo, Zeadally, Sherali, 2016. Intelligent Device-to-Device communication in wireless networks: a survey. Int. J. Ad Hoc Ubiquitous Comput. 17 (4), 197–215.
the internet of things. IEEE Syst. J. 10. Guo, Bin, Sun, Shaohui, Gao, Qiubin, 2015. Interference Management for D2D
Bello, Oladipupo, Zeadally, Sherali, 2014. Intelligent Device-to-Device communication in Communications Underlaying Cellular Networks at Cell Edge. ICWMC.
the internet of things. IEEE Syst. J.. Gupta, Akhil, Kumar Jha, Rakesh, 2016. Power optimization using massive MIMO and
Bluetooth, S., Bluetooth specification version 1.1, Available HTTP: 〈http://www. small cells approach in different deployment scenarios. Wirel. Netw., 1–15.
bluetooth.com〉, 2001. Halperin, Daniel, et al., 2007. Interference Cancellation: Better Receivers for a New
Bohmländer, D.. et al. 2015. Advantages in Crash Severity Prediction using Vehicle to Wireless MAC. HotNets.
Vehicle Communication. In: Proceedings of the 2015 IEEE International Conference Hayami et al., 2015. Crash warning for intersection and head-on car collision in vehicle –
on Dependable Systems and Networks Workshops (DSN-W). IEEE. to-vehicle communication. In: Proceedings of the International Conference on
Botsov, Mladen, et al., 2014. Location dependent resource allocation for mobile device- Connected Vehicles, (ICCVE), (2015).
to-device communications. In: Proceedings of the 2014 IEEE Wireless Hong, J.. et al. 2013. Analysis of Device-to-Device discovery and link setup in LTE
Communications and Networking Conference (WCNC). networks. In: Proceedings of the 24th International Symposium on Personal Indoor
Cao, Xianghui, et al., 2016. On optimal device-to-device resource allocation for and Mobile Radio Communications (PIMRC), 2013 IEEE. IEEE.
minimizing end-to-end delay in VANETs. IEEE Trans. Veh. Technol.. Hossain, Ekram, et al., 2014. Evolution toward 5G multi-tier cellular wireless networks:
Cha, Inhyok, et al., 2009. Trust in M2M communication. IEEE Veh. Technol. Mag. 4.3, an interference management perspective. IEEE Wirel. Commun. 21.3, 118–127.
69–75. Hu, Rose, Qian, Yi, 2014. An energy efficient and spectrum efficient wireless
Chai, Yingqi, Qinghe Du, Pinyi Ren. 2013. Partial time-frequency resource allocation for heterogeneous network framework for 5G systems. Commun. Mag. 52.5, 94-101
device-to-device communications underlaying cellular networks. In: Proceedings of IEEE.
the 2013 IEEE International Conference on Communications (ICC). Kennedy, J., 2010. Particle Swarm optimization. In: Encyclopedia of Machine Learning,
Chandrasekhar, V., She, Z., 2008. Optimal uplink power control in two cell systems with Springer. pp. 760–766.
rise-over-thermal constraints. IEEE Commun. Lett. 12 (3). Janis, P.. et al. 2009. "Interference-aware resource allocation for device-to-device radio
Chao, Shih-Lung, et al., 2013. Bio-inspired proximity discovery and synchronization for underlaying cellular networks. In: Proceedings of the IEEE 69th Vehicular
D2D communications. IEEE Commun. Lett. 17 (12), 2300–2303. Technology Conference, 2009. VTC Spring 2009. IEEE.
Chen, Biwei, Jun Zheng, Yuan Zhang, 2015. A time division scheduling resource Jänis, P.. et al. 2009. Interference-avoiding MIMO schemes for device-to-device radio
allocation algorithm for D2D communication in cellular networks. In: Proceedings of underlaying cellular networks. In: Proceedings of the IEEE 20th International
the 2015 IEEE International Conference on Communications (ICC). IEEE. Symposium on Personal, Indoor and Mobile Radio Communications.
Chen, Xue, Rose Qingyang Hu, Yi Qian, 2014. Distributed resource and power allocation Jayasinghe, K.. et al. 2015. Physical layer security for relay assisted MIMO D2D
for device-to-device communications underlaying cellular network. In: Proceedings communication. In: Proceedings of the IEEE 2015 IEEE International Conference on
of the 2014 IEEE Global Communications Conference (GLOBECOM). Communication Workshop (ICCW).
Cheng, Xiang, et al., 2015. Device-to-device channel measurements and models: a Jiang, Yanxiang, et al., 2015. Energy efficient joint resource allocation and power control
survey. Communications, IET 9.3, 312–325. for D2D communications. IEEE Trans. Veh. Technol. (99).
Cheng, Peng, et al., 2012 Resource allocation for cognitive networks with D2D Jin, X., Andrews, J.G., Ghosh, A., Ratasuk, R., 2014. An overview of 3GPP device-to-
communication: an evolutionary approach. In: Proceedings of the Wireless device proximity services. IEEE Commun. Mag., (April).
Communications and Networking Conference (WCNC), 2012 IEEE. IEEE. Kai, Y.. et al. 2008. Distributed robust optimization for communication networks.
Ciou, S.. et al. 2015. Multi-sharing resource allocation for device-to-device INFOCOM 2008. In: Proceedings of the 27th Conference on Computer
communication underlaying 5 G mobile networks. In: Proceedings of the 26th Communications. IEEE. IEEE.
Annual International Symposium on Personal, Indoor, and Mobile Radio Kenney, John Francis, Ernest Sydney Keeping, 1954. Mathematics of statistics-part one.
Communications (PIMRC), 2015 IEEE . IEEE. Kim, Inhyoung, Park, In-Soon, Lee, Yong H., 2006. Use of linear programming for
Cisco, V.N.I., 2014. The Zettabyte Era: Trends and Analysis. Updated (29/05/2013), dynamic subcarrier and bit allocation in multiuser OFDM. IEEE Trans. Veh.
〈http://www.cisco.com/c/en/us/solutions/collateral/serviceprovider/visual- Technol. 55.4, 1195–1207.
networking-index-vni/VNI_Hyperconnectivity_WP〉. html. Kim, Kyung-Won, Oh, Seong-Jun, 2014. An efficient implementation of the ITU-R
Corson, M. Scott, et al., 2010. Toward proximity-aware internetworking. IEEE Wirel. channel model for device-to-device simulation. IEEE Commun. Lett. 18.9,
Commun. 17.6, 26–33. 1633–1636.
Dohler, Mischa, Nakamura, Takehiro, 2016b. In: Osseiran, Afif (Ed.), 5G Mobile and Kuang, Y., Lei, L., Zhong, Z., 2015. Delay-Optimal Distributed Resource Allocation for
Wireless Communications Technology. Cambridge University Press. Device-to-Device Communications. In: Proceedings of the 2015 IEEE 82nd
Dongyu Wang, Wang, X., 2013. An interference management Scheme for device-to- Vehicular Technology Conference (VTC Fall). IEEE.
device multicast spectrum sharing hybrid network. In: Proceedings of the 2013 IEEE Kwon, H., Lee, J., Kang, I., 2014. Interference-Aware Interference Mitigation for Device-
24rd International Symposium Personal Indoor and Mobile Radio Communications to-Device Communications. In: Proceedings of the 2014 IEEE 79thVehicular
(PIMRC). Technology Conference (VTC Spring). IEEE.
Doppler, K., Ribeiro, C., Kneckt, J., 2011. Advances in D2D communications: Energy Militano, L., Orsino, A., Araniti, G., Nitti, M., Atzori, L., Iera, A., 2016. Trusted D2D-
efficient service and device discovery radio. In: Proceedings of the 2nd International based Data Uploading in In-band Narrowband-IoT with Social Awareness. In:
Conference on Wireless Communication, Vehicular Technology, Information Theory Proceedings of the IEEE Workshop PIMRC.
and Aerospace & Electronic Systems Technology (Wireless VITAE). IEEE. Le, L., 2012. Fair resource allocation for device-to-device communications in wireless
Du, J.. et al. 2012. A compressed HARQ feedback for device-to-device multicast cellular networks. Global Communications Conference (GLOBECOM), 2012 IEEE.
communications. In: Proceedings of the 2012 IEEE Vehicular Technology IEEE.
Conference (VTC Fall). Lee, Dong Heon, et al., 2014. Two-stage semi-distributed resource management for
Ejaz, Waleed, ul Hasan, Najam, Seok Kim, Hyung, 2011. Spectrum sensing in cognitive device-to-device communication in cellular networks. 1908–1920IEEE Trans.
radio mobile Ad hoc networks: a survey. J. Korea Inst. Commun. Sci. Conf., Wirel. Commun. 13.4.
376–377. Lee, Namyoon, et al., 2015. Power control for D2D underlaid cellular networks:
Feng, Daquan, et al., 2013. Device-to-device communications underlaying cellular modeling, algorithms, and analysis. IEEE J. Sel. Areas Commun. 33.1, 1–13.
networks. IEEE Trans. Commun. 61.8, 3541–3551. Lei, L., et al., 2012. Operator controlled device-to-device communications in LTE-
Feng, Daquan, et al., 2014. Device-to-device communications in cellular networks. IEEE advancednetworks. IEEE Wirel. Commun. 19 (3), 96–104.
Commun. Mag. 52.4, 49–55. Li, J.C., Lei, M., Gao, F., 2012. Device-to-device (D2D) communication in MU-MIMO
Fey, Simon, et al., 2012. Device-to-device communication for Smart Grid devices. In: cellular networks. Proc. IEEE GLOBECOM, 3583–3587.
Proceedings of the 2012 3rd IEEE PES International Conference and Exhibition on Li, Y.. et al. 2015. Three-dimensional modeling, simulation and evaluation of Device-to-
Innovative Smart Grid Technologies (ISGT Europe). IEEE Device channels. In: Proceedings of the 2015 IEEE International Symposium on
Fodor, G., Reider, N., 2011. A distributed power control scheme for cellular network Antennas and Propagation & USNC/URSI National Radio Science Meeting.
assisted D2D communications. In: Proceedings of the Global Telecommunications Li, Z.. et al. 2014. Overview on initial METIS D2D concept. 5 G for Ubiquitous
Conference (GLOBECOM 2011), 2011 IEEE. IEEE. Connectivity (5 GU), 2014 In: Proceedings of the 1st International Conference on.
Fodor, G.ábor, et al., 2012. Design aspects of network assisted device-to-device IEEE.
communications. IEEE Commun. Mag. 50.3, 170–177. Lien, Shao-Yu, et al., 2016a. 3GPP device-to-device communications for beyond 4G
Fodor, G.ábor, et al., 2014. Device-to-device communications for national security and cellular networks. IEEE Commun. Mag. 54.3, 29–35.
public safety. IEEE Access 2, 1510–1520. Lin, S.. et al. 2013. Resource allocation using particle swarm optimization for D2D
Fodor, G.ábor, et al., 2016a. An overview of device-to-device communications technology communication underlay of cellular networks. In: Poceedings of the Wireless

27
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Communications and Networking Conference (WCNC), 2013 IEEE. IEEE. (ICCE).

Liu, An, et al., 2015. Two timescale joint beamforming and routing for multi-antenna da Silva, J.M.B.. et al., 2014. Performance analysis of network-assisted two-hop D2D
D2D networks via stochastic cutting plane. IEEE Trans. Signal Process. 63.18, communications, Globecom Workshops, December.
4854–4865. da Silva, Jose Mairton B., Fodor, G.ábor, 2015. A binary power control scheme for D2D
Liu, Jia, et al., 2013. A resource reuse scheme of D2D communication Underlaying LTE communications. IEEE Wirel. Commun. Lett. 4.6, 669–672.
network with Intercell interference. Commun. Netw. 5.03, 187. Song, Hojin, et al., 2015. Joint power and rate control for device-to-device
Liu, Jiangchuan, et al., 2014. Device-to-device communication in LTE-advanced communications in cellular systems. IEEE Trans. Wireless Commun. 14.10,
networks: a survey. IEEE Commun. Surv. Tutor. 17. 5750–5762.
Liu, Yuanwei, et al., 2016. Secure D2D communication in large-scale cognitive cellular Song, Lingyang, et al., 2014. Game-theoretic resource allocation methods for device-to-
networks: a wireless power transfer model. IEEE Trans. Commun. 64. device communication. IEEE Wirel. Commun. 21.3, 136–144.
Mach, Pavel, Becvar, Zdenek, Vanek, Tomas, 2015. In-band device-to-device Sun, Q.. et al. 2015. Energy efficient incentive resource allocation in D2D cooperative
communication in OFDMA cellular networks: a survey and challenges. IEEE communications. In: Proceedings of the 2015 IEEE International Conference on
Commun. Surv. Tutor. 17.4, 1885–1922. Communications (ICC). IEEE.
Meshgi, Hadi, Dongmei Zhao, Rong Zheng. Joint channel and power allocation in Sun, Wanlu, et al., 2016. Cluster-based radio resource management for D2D-supported
underlay multicast device-to-device communications. In: Proceedings of the 2015 safety-critical V2X communications. IEEE Trans. Wirel. Commun. 15.4, 2756–2769.
IEEE International Conference on Communications (ICC). IEEE, 2015. Koskela, T., Hakola, S., Chen, T., Lehtomaki, J., 2010. Clustering concept using device-
Militano, L., Orsino, A., Araniti, G., Nitti, M., Atzori, L., Iera, A., 2016a. Trust-based and to-device communication in cellular system. In: Proceedings of the IEEE Wireless
social-aware coalition formation game for multihop data uploading in 5G systems. Communications and Networking Conference.
Comput. Netw., (ISSN 1389-1286). Tang, Huan, Ding, Zhi, Levy, Bernard C., 2014. Enabling D2D communications through
Militano, L., et al., 2015. Device-to-Device Communications for 5G Internet of Things.", neighbor discovery in LTE cellular networks. IEEE Trans. Signal Process. 62.19,
IOT, EAI. 5157–5170.
Min, Hyunkee, et al., 2011. Capacity enhancement using an interference limited area for Tang, Rui, Zhao, Jihong, Qu, Hua, 2015. Joint optimization of channel allocation, link
device-to-device uplink underlaying cellular networks. IEEE Trans. Wirel. Commun. assignment and power control for device-to-device communication underlaying
10.12, 3995–4000. cellular network. Communications 12.12, 92–100.
Mohammed, Hasan, Hossain, Ekram, 2014. Distributed resource allocation for relay- Tata, C.H.A.F.I.K.A., Kadoch, M.I.C.H.E.L., 2014. Secure network coding based data
aided device-to-device communication: a message passing approach. IEEE Wirel. splitting for public safety D2D communications over LTE heterogeneous networks.
Commun. 13.11, 6326–6341. In: Proceedings of the 14th IEEE International Conference on Computer and
Mumtaz, S., Huq, K.M.S., Rodriguez, J., 2014. Direct mobile-to-mobile communication: Information Technology (CIT’14)..
Paradigm for 5G. IEEE Wireless Commun. 21 (5), 14–23. Technical Specification Group Services and System Aspects, 2014 Proximity-Based
Namvar, N., Bahadori, N., Afghah, F., 2015. Context-aware D2D peer selection for load services (ProSe); Stage 2 (Release 12), 3GPP TS 23.303 V12.1.0,Jun.
distribution in LTE networks. In: Proceedings of the 49th Asilomar Conference on Tehrani, Mohsen, Nader, Mustafa, Uysal, Yanikomeroglu, Halim, 2014. Device-to-device
Signals, Systems and Computers. IEEE. communication in 5G cellular networks: challenges, solutions, and future directions.
Ngo, T., Kim, Y., 2015. Using Timing Advance to support proximity discovery in network- IEEE Commun. Mag. 52.5, 86–92.
assisted D2D communication. In: Proceedings of the Seventh International Urama, Jani, et al., 2016. Dynamic Social Trust Associations over D2D Communications:
Conference on Ubiquitous and Future Networks (ICUFN), 2015. IEEE. An Implementation Perspective. In: Proceedings of the 5th IEEE International
Nguyen, P.. et al. 2014. Network-assisted device discovery for LTE-based D2D Conference on Mobile Services. San Francisco: InterContinental Mark Hopkins San
communication systems. In: Proceedings of the 2014 IEEE International Conference Francisco. s. 186–189
on. Communications (ICC) IEEE. Usman, Muhammad, et al., 2015. A software-defined device-to-device communication
Ni, Yiyang, et al., 2016. Beamforming and interference cancellation for D2D architecture for Public safety applications in 5GNetworks. IEEE Access 3,
communication underlaying cellular networks. IEEE Trans. Commun. 64.2, 1649–1654.
832–846. Wang, H., Chen, J., Liu, Z., 2013. Resource allocation in central-controlled Device-to-
Noura, Mahda, Nordin, Rosdiadee, 2016. A survey on interference management for Device communications networks. Global Communications Conference
device-to-device (D2D) communication and its challenges in 5G networks. J. Netw. (GLOBECOM), 2013 IEEE. IEEE.
Comput. Appl.. Wang, L.. et al. 2015. Secrecy-oriented partner selection based on social trust in device-
Ometov, Aleksandr, et al., 2016. A novel security-centric framework for D2D connectivity to-device communications. In: Proceedings of the 2015 IEEE International
based on spatial and social proximity. Comput. Netw.. Conference on Communications (ICC).
Ometovy, A., et al., 2016. Towards trusted, social-aware D2D connectivity: bridging Wang, Li, Wu, Huaqing, 2015. Jamming partner selection for maximising the worst D2D
across technology and sociality realms. (1-1)IEEE Wirel. Commun. PP (99). secrecy rate based on social trust. Trans. Emerg. Telecommun. Technol., (Oct).
Ometovy, A., Orsino, A., 2016. Validating Information Security Framework for Wang, Qijie, et al., 2014. Quality-optimized joint source selection and power control for
Offloading from LTE onto D2D Links. FRUCT In: Proceedings of the 18th wireless multimedia D2D communication using stackelberg game. IEEE Trans. Veh.
Conference, /FRUCT-ISPIT. Technol. 64 (8).
Pappalardo, Irene, et al., 2016. Caching strategies in heterogeneous networks with D2D, Wang, Bin, et al., 2011. Resource allocation optimization for device-to-device
small BS and macro BS communications. In: Proceedings of the 2016 IEEE communication underlaying cellular networks. Vehicular Technology Conference
International Conference on Communications (ICC). (VTC Spring), 2011 IEEE 73rd. IEEE, 2011.
Patel, C.S., Stuber, G.L., Pratt, T.G., 2005. Simulation of Rayleigh faded mobile-to-mobile Wang, Mingjun, Zheng Yan, 2015. Security in D2D Communications: A Review.
communication channels. (1876–1844, Nov)IEEE Trans. Commun. 53 (11), (1876– Trustcom/BigDataSE/ISPA, 2015 IEEE. Vol. 1. IEEE.
1844, Nov). Warabino, T.. et al. 2005. Adaptive media switching for future vehicle-to-vehicle
Peng, Mugen, et al., 2014. Device-to-device underlaid cellular networks under Rician communication. In: Proceedings of the 16th International Symposium on Personal,
fading channels. IEEE Trans. Wirel. Commun. 13.8, 4247–4259. Indoor and Mobile Radio Communications, 2005. PIMRC. Vol. 2.
Perrig, Adrian, Stankovic, John, Wagner, David, 2004. Security in wireless sensor Weijun, X.. et al. 2014. Resource allocation schemes for D2D communication used in
networks. Commun. ACM 47.6, 53–57. VANETs.In: Proceedings of the 2014 IEEE 80th Vehicular Technology Conference
Phunchongharn, Phond, Hossain, Ekram, In Kim, Dong, 2013. Resource allocation for (VTC2014-Fall). IEEE.
device-to-device communications underlaying LTE-advanced networks. IEEE Wirel. Wen, S., Zhu, X., Lin, Z., Zhang, X., Yang, D., 2012.Optimization of interference
Commun. 20.4, 91–100. coordination schemes in device -to-device (D2D) communication. In: Proceedings of
Pratas, N.K., Popovski, P., 2013. Low-rate machine-type communication via wireless the 7th International ICST Conference on Communications and Networking in China
device-to-device (D2D)links. arXiv Prepr. arXiv 1305., 6783. (CHINACOM). pp. 542–547.
Pimmy, Gandotra, Rakesh, Kumar Jha, 2016. Device-to-Device Communication in Wu, Dan, et al., 2014. The role of mobility for D2D communications in LTE-advanced
Cellular Networks: A Survey. J. Netw. Comput. Appl.. networks: energy vs bandwidth efficiency. IEEE Wirel. Commun. 21.2, 66–71.
Pratas, Nuno K., Petar Popovski, 2015 "Network-Assisted Device-to-Device (D2D) Direct Wu, H.. et al. 2016. Resource allocation for wireless caching in socially-enabled D2D
Proximity Discovery with Underlay Communication. In: Proceedings of the IEEE communications. In: Proceedings of the 2016 IEEE International Conference on
GLOBECOM. Communications (ICC). IEEE.
Rêgo, Márzio G. da S., et al., 2012 "Performance analysis of power control for device-to- Wu, X., Tavildar, S., Shakkottai, S., Richardson, T., Li, J., Laroia, R., Jovicic, A., 2010.
device communication in cellular MIMO systems. In: Proceedings of the 2012 FlashLinQ: A synchronous distributed scheduler for peer-to-peer ad hoc networks.
International Symposium on Wireless Communication Systems (ISWCS). IEEE. In: Proceedings of the IEEE Allerton Conference on Communication, Control and
Ren, Yi, et al., 2015. Power control in D2D-based vehicular communication networks. Computing, pp. 514–521.
IEEE Trans. Veh. Technol. 64.12, 5547–5562. Xiao, Sa, et al., 2016. Energy-efficient mobile association in heterogeneous networks with
Hakola, S., Chen, T., Lehtom¨aki andki, J., Koskela, T., 2010. Device-todevice (D2D) device-to-device communications. IEEE Trans. Wirel. Commun. 15, 5260–5271.
communication in cellular network–performance analysis of optimum and practical Xiao, X., Tao, X., Lu, J., 2011. A QoS-aware power optimization scheme in OFDMA
communication mode selection. In: Proceedings of the IEEE Wireless systems with integrated device-to-device (D2D) communications. In: Proceedings of
Communications and Networking Conference, 2010 the 2011 IEEE Vehicular Technology Conference (VTC Fall)
Shen, Wenlong, et al., 2014 "Secure key establishment for device-to-device Xu, Hao, et al., 2016. Pilot allocation and PowerControl in D2D underlay massive MIMO
communications. In: Proceedings of the 2014 IEEE Global Communications systems. IEEE Commun. Lett..
Conference (GLOBECOM). Yang, Mi. Jeong, et al., 2013. Solving the data overload: device-to-device bearer control
Shin, Wonjae, KyungHun Jang, Hyun-Ho Choi, 2013. Device-to-device communication architecture for cellular data offloading. IEEE Veh. Technol. Mag. 8.1, 31–39.
assisted interference mitigation for next generation cellular networks. In: Yin, Rui, et al., 2015. Pricing-based interference coordination for D2D communications
Proceedings of the 2013 IEEE International Conference on Consumer Electronics in cellular networks. IEEE Trans. Wirel. Commun. 14.3, 1519–1532.

28
P. Gandotra et al. Journal of Network and Computer Applications 78 (2017) 9–29

Yu, C.-H., Doppler, K., Ribeiro, C., Tirkkonen, O., 2009. Power optimization of device-to- Dr. Rakesh Kumar Jha (S′10-M′13-SM’15) received
device communication underlaying cellular communication systems. In: Proceedings the B.Tech. degree in electronics and communication
of the IEEE International Conference on Communications, June. engineering in Bhopal, India, the M.Tech. degree from
Yu, Chia-Hao, Olav Tirkkonen, 2012. Device-to-device underlay cellular network based NIT Jalandhar, India, and the Ph.D. degree from NIT
on rate splitting. In: Proceedings of the IEEE Wireless Communications and Surat, India, in 2013. He is currently an Assistant
Networking Conference (WCNC). Professor with the School of Electronics and
Yue, Jianting, et al., 2013. Secrecy-based access control for device-to-device Communication Engineering, Shri Mata Vaishno Devi
communication underlaying cellular networks. IEEE Commun. Lett. 17.11, University, Jammu and Kashmir, India. He is carrying
2068–2071. out his research on wireless communication, power opti-
Zhang, Aiqing, et al., 2015. SeDS: secure data sharing strategy for D2D communication mizations, wireless security issues, and optical communi-
in LTE-AdvancedNetworks. IEEE Trans. Veh. Technol.. cations. He has authored over 30 international journal
Zhang, Bo, et al., 2015. Energy-efficient architecture and technologies for device to device papers and more than 20 international conference papers.
(D2D) based proximity service. China Commun. 12.12, 32–42. His area of interest is wireless communication, optical fiber
Zhang, H.. et al. 2014. Radio resource allocation for physical-layer security in D2D communication, computer networks, and security issues. Dr. Jha’s concept related to
underlay communications. In: Proceedings of the 2014 IEEE International router of wireless communication has been accepted by the International
Conference on Communications (ICC). IEEE. Telecommunication Union (ITU) in 2010. He received the Young Scientist Author
Zhang, J.. et al. 2013. Utility-maximization resource allocation for device-to-device Award from ITU in 2010, the APAN Fellowship in 2011 and 2012, and the Student Travel
communication underlaying cellular networks. Globecom Workshops (GC Wkshps), Grant from COMSNET in 2012. He is a Senior Member of the Global ICT Standardization
2013 IEEE. IEEE. Forum for India, Society for Industrial and Applied Mathematics, the International
Zhang, R., Cheng, X., Yang, L., 2015. Cooperation via Spectrum Sharing for Physical Association of Engineers, and the Advance Computing and Communication Society. He is
Layer Security in Device-to-DeviceCommunications Underlaying Cellular Networks. also a member of Association for Computing Machinery (ACM).
In: Proceedings of the 2015 IEEE Global Communications Conference
(GLOBECOM). IEEE.
Zhang, R., Cheng, X., Yang, L., 2016. Joint power and access control for physical layer Prof. Sanjeev Jain born at Vidisha in Madhya Pradesh in
security in D2D communications underlaying cellular networks. In: Proceedings of 1967, obtained his Post Graduate Degree in Computer
the IEEE International Conference on Communications (ICC). IEEE. Science and Engineering from Indian Institute of
Zhang, Rongqing, et al., 2013. Distributed resource allocation for device-to-device Technology, Delhi, in 1992. He later received his
communications underlaying cellular networks. In: Proceedings of the 2013 IEEE Doctorate Degree in Computer Science & Engineering
International Conference on Communications (ICC). IEEE, 2013. and has over 24 years’ experience in teaching and research.
Zheng, J., Chen, B., Zhang, Y., 2015. An Adaptive Time Division Scheduling Based He has served as Director, Madhav Institute of Technology
Resource Allocation Algorithm for D2D Communication Underlaying Cellular and Science (MITS), Gwalior. Presently, he is working as a
Networks. In: Proceedings of the 2015 IEEE Global Communications Conference vice chancellor at Shri Mata Vaishno Devi University,
(GLOBECOM). IEEE. Katra. Besides teaching at Post Graduate level Professor
Zhou, Bin, et al., 2013. Intracluster device-to-device relay algorithm with optimal Jain has the credit of making significant contribution to R
resource utilization. IEEE Trans. Veh. Technol. 62.5, 2315–2326. & D in the area of Image Processing and Mobile Adhoc
Zhou, L., Ruttik, K., Tirkkonen, O., 2015. Interference canceling power optimization for Network. He has guided Ph.D. Scholars and has under-
device to device communication. In: Proceedings of the 2015 IEEE 81st Vehicular taken a number of major R & D projects sponsored by the Government and Private
Technology Conference (VTC Spring). Agencies. His work on Digital Watermarking for Image Authentication is highly valued in
Zhou, Yun, Fang, Yuguang, Zhang, Yanchao, 2008. Securing wireless sensor networks: a the research field. He is also a member of Association for Computing Machinery (ACM).
survey. IEEE Commun. Surv. Tutor. 10.3, 6–28.
Zhou, Zhenyu, et al., 2016. Iterative energy-efficient stable matching approach for
context-aware resource allocation in D2D communications. IEEE Access.
Zou, Kingsley Jun, et al., 2014. Proximity discovery for device-to-device communications
over a cellular network. IEEE Commun. Mag. 5 (2.6), 98–107.

Miss. Pimmy Gandotra (S′16) received the B.E. degree

in Electronics and Communication Engineering from
Jammu University, Jammu and Kashmir, India, in 2013.
She pursued the M.Tech degree in Electronics and
Communication Engineering at Shri Mata Vaishno Devi
University, Katra, Jammu and Kashmir, India. She is
currently Ph. D research scholar at Shri Mata Vaishno
Devi University, Katra, Jammu and Kashmir. Her research
interest includes the emerging technologies of 5 G wireless
communication network. Currently she is doing research
work in Resource Allocation in Device to Device commu-
nication. She is working on Qualnet simulation and Matlab
tools for Wireless Communication. She has received teach-
ing assistantship from MHRD. She is a student member of Institute of Electrical and
Electronics Engineers (IEEE).

29