IETE Technical Review

ISSN: 0256-4602 (Print) 0974-5971 (Online) Journal homepage: http://www.tandfonline.com/loi/titr20

Cloud Computing in VANETs: Architecture,

Taxonomy, and Challenges

Ahmed Aliyu, Abdul Hanan Abdullah, Omprakash Kaiwartya, Yue Cao,

Mohammed Joda Usman, Sushil Kumar, D. K. Lobiyal & Ram Shringar Raw

To cite this article: Ahmed Aliyu, Abdul Hanan Abdullah, Omprakash Kaiwartya, Yue Cao,
Mohammed Joda Usman, Sushil Kumar, D. K. Lobiyal & Ram Shringar Raw (2017): Cloud
Computing in VANETs: Architecture, Taxonomy, and Challenges, IETE Technical Review, DOI:
10.1080/02564602.2017.1342572

To link to this article: https://doi.org/10.1080/02564602.2017.1342572

Published online: 23 Aug 2017.

Submit your article to this journal

Article views: 86

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=titr20
IETE TECHNICAL REVIEW, 2017
https://doi.org/10.1080/02564602.2017.1342572

Cloud Computing in VANETs: Architecture, Taxonomy, and Challenges

Ahmed Aliyu1,5, Abdul Hanan Abdullah1, Omprakash Kaiwartya 1
, Yue Cao2, Mohammed Joda Usman1,5,
Sushil Kumar 3, D. K. Lobiyal3 and Ram Shringar Raw4
1
Faculty of Computing, Universiti Teknologi Malaysia, Johor Bahru 81310, Malaysia; 2Department of Computer and Information Sciences,
Northumbria University, Newcastle upon Tyne NE1 8ST, UK; 3School of Computer and Systems Sciences, Jawaharlal Nehru University, New Delhi
110067, India; 4Department of Computer Science, Indira Gandhi National Tribal University, Amarkatnak 484886, India; 5Department of
Mathematics, Bauchi State University Gadau, PMB 068 Bauchi State, Nigeria

ABSTRACT KEYWORDS
Cloud computing in VANETs (CC-V) has been investigated into two major themes of research Architecture; Cloud
including vehicular cloud computing (VCC) and vehicle using cloud (VuC). VCC is the realization of computing; Taxonomy;
autonomous cloud among vehicles to share their abundant resources. VuC is the efficient usage Vehicle using cloud;
of conventional cloud by on-road vehicles via a reliable Internet connection. Recently, a number of Vehicular ad hoc networks;
Vehicular cloud
advancements have been made to address the issues and challenges in VCC and VuC. This paper
qualitatively reviews CC-V with the emphasis on layered architecture, network component,
taxonomy, and future challenges. Specifically, a four-layered architecture for CC-V is proposed
including perception, coordination, artificial intelligence and smart application layers. Three
network components of CC-V, namely, vehicle, connection, and computation are explored with
their cooperative roles. A taxonomy for CC-V is presented considering major themes of research in
the area including design of architecture, data dissemination, security, and applications. Related
literature on each theme is critically investigated with comparative assessment of recent advances.
Finally, some open research challenges are identified as future issues. The challenges are the
outcome of the critical and qualitative assessment of the literature on CC-V.

1. INTRODUCTION realizing smart ITS applications including real-time traf-

fic prediction and web-based services [5].
Recently, cloud computing has witnessed significant
attention in vehicular communication. It is because of
The investigations are in initial stages in both VCC and
the realization of smart intelligent transport system
VuC. There are opportunities and challenges consider-
(ITS) applications, and architectural similarity between
ing a new area for cloud computing. In fact, huge resour-
mobile cloud computing (MCC) and vehicular ad hoc
ces of vehicles are underutilized at various places (where
networks (VANETs) [1–3]. Cloud computing in
a group of vehicles assemble), which is needed to be
VANETs (CC-V) has been investigated in two major
explored and tap into [6]. The framework of CC-V
dimensions of research in the area. First, vehicular cloud
improves the usability, reliability, and efficiency of ITS
computing (VCC) has been explored. In VCC, a group
applications [7]. Subsequently, it enhances safety in
of connected vehicles forms a cloud with the aim of
transportation, reduces traffic congestion, decreases air
sharing their abundant resources.
pollution, and augments comfort in driving [8].
VCC increases resource utilization in vehicular commu-
Some of the cloud computing services realized in CC-V
nication [4]. The abundant resources include storage,
include network as a service (NaaS), storage as a service
computing capability, sensing, and communication
(STaaS), cooperation as a service (CaaS), computing as a
capability. The resources became available for sharing,
service (COaaS), and sensing as a service (SEaaS) [9–11].
when several vehicles assemble at the places including
Since vehicles are considered to be able to connect to the
parking lots, garage, canteens, highway hold-up, and
Internet, NaaS can be used by passengers to connect to
traffic lights. At these places, the idea of VCC has been
the Internet in VuC. Applications and data of vehicles,
realized. VCC dynamically allocates abundant resources
which require more memory for storage, can utilize
to authorized users. Second, vehicle using cloud (VuC)
STaaS of VCC [12]. CaaS can be utilized to share traffic
has been explored in which the connected vehicles are
information among vehicles in case of accident, and
considered. VuC enables the access to the services of
road maneuverer. Vehicles can realize smart ITS
conventional cloud at vehicles via Internet. VuC helps in
© 2017 IETE
2 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

applications using COaaS to fulfill higher computation

power requirement of smart applications. SEaaS can be
utilized for monitoring real-time condition of car as well
as driver’s behaviour [13]. Whaiduzzaman et al. [14]
have explored VCC by presenting a taxonomy based on
strategic management, security and privacy, cloud for-
mations, inter-cloud communication, and applications
issues. An architecture for VCC, comparison with cloud
computing, and open research issues have been pre-
sented. However, the architecture has not been explicitly
defined considering layer-wise function, representation,
and protocol. The network components related to VCC
are also not explored.

Although various contributions have been made for real-

izing these cloud services in CC-V, yet there are a num-
ber of issues that need to be addressed in near future. In
this paper, a qualitative review of CC-Vs has been pre-
sented. The review focuses on layered architecture, net-
work component of CC-V, taxonomy of recent Figure 1: The layered architecture of CC-V
advances, and open challenges and issues for future
direction of investigation in the area. A four-layered
architecture for CC-V is designed including perception, structural point of view, CC-V has four major compo-
coordination, artificial intelligence, and smart applica- nents, namely, perception physical devices, coordination,
tion layers. The representative components, responsibili- artificial intelligence, and smart application services.
ties, and protocols of each layer are described. The three Therefore, the architecture consists of four layers includ-
network components of CC-V, namely, vehicle, commu- ing perception, coordination, artificial intelligence, and
nication, and computation have been explored with their smart-application (see Figure 1). The perception layer is
cooperative roles. From the best of our knowledge, there correlated with physical devices. The coordination layer is
is no layered architecture available for CC-V. The major linked with communication networks. The artificial intel-
network components with their cooperative roles have ligence layer is associated with computation, and the
not been explored previously. The taxonomy of CC-V is smart application layer is correlated with the services.
presented considering four major issues including design The layers effectively divide the functionalities of CC-V
of architecture, data dissemination, security, and appli- into four groups. The layers are unambiguously differen-
cation development. Each issue has been qualitatively tiable in terms of both functions and representations.
and critically reviewed with comparative assessments of Each layer is described below considering representative
recent advancements. Finally, some future challenges are components and functionalities.
identified according to the qualitative and critical review
of related literature.
2.1 Perception Layer
The rest of this paper is organized as follows. Section 2 The perception layer of the architecture represents in-
presents four-layered architecture of CC-V. Section 3 vehicle devices including sensors, actuators, display
describes three network components of CC-V. Section 4 unit, smart mobile devices, global positioning system
presents a taxonomy of CC-V with qualitative and criti- (GPS) receiver, actuators, etc. The two main functions
cal review of related literature. Some future challenges in of the layer include gathering traffic data through
CC-V are identified in Section 5, followed by conclusion sensing and delivering information at end level. The
made in Section 6. layer defines vehicle-to-physical world interactions
through smart devices. Apart from the two functional-
ities, a number of services are provided by the percep-
2. LAYERED ARCHITECTURE
tion layer to the next layer as interface services. The
In this section, a four-layered architecture for CC-V is interface services include error detection and correc-
designed in terms of representation, functionalities, and tion on sensed data, result verification on inferred
protocol perspective. In a broader operational and information, etc.
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 3

In terms of protocol perspective, physical layer part of considered. The protocols include IEEE 1609.6 of
the protocols including IEEE 802.11p (PHY-802.11p) of WAVE, COaaS, STaaS, picture as a service (PICaaS),
WAVE [15], 802.11a/b/g of WLAN [16], Wi-Max [17], infrastructure as a service (INaaS), CaaS, NaaS, and gate-
and 4G/LTE [18] are considered. This is due to the con- way as a service (GaaS). The design of protocols for BDA
sideration of heterogeneous in-vehicle devices contain- and VCC is an open research theme in CC-V due to the
ing different types of sensors. growing volume of traffic data.

2.2 Coordination Layer 2.4 Smart Application Layer

The coordination layer represents network devices and The smart application layer represents smart cloud-
internetworking technologies. The network coordination based applications of three categories including safety,
is significant due to the consideration of different types efficiency, and infotainment. The layer is responsible
of networks including VANETs, Wi-Fi, and 4G/LTE. for delivering end-user smart services. The layer
The main functionalities of the layer include transfer of operations include application management, service
data packet and efficient handoff among these networks. management, application- and service-based data man-
The heterogeneous network architecture makes both the agement, application-based authentication, and autho-
tasks quite challenging. The other functionalities of the rization. Although the applications for safety and
layer include data dissemination in heterogeneous net- efficiency are also implemented in VANETs, yet the
works, location-based networking support, network level cloud-based operation significantly enhances the intel-
authentication, and authorization. ligence, and usefulness of these applications. This is due
to the advancement of computing power in applications
In terms of protocol perspective, two sub-layers are con- operating through cloud support, for computation and
sidered. In the first sub-layer, medium access control traffic data. Some of the smart applications include
(MAC) layer part of the protocols including IEEE 1609.4 real-time traffic forecasting, smart toll collection
[19] in WAVE, 802.11p, and LLC are considered. In the through car payment, smart challan for traffic rule vio-
second sub-layer, network and transport protocols lation, ad hoc multimedia sharing, smart black-box,
including fast application and communication enabler and smart emergency call.
(FAST) [20–23] in CALM, car-to-car (C2C-net) [24]
and short message protocol (SMP) [19] in WAVE are In terms of protocols perspective, resource handler pro-
considered apart from traditional IP and TCP/UDP tocol IEEE 1609.1 of WAVE is considered for efficient
combinations. resource management among smart applications. Apart
from this, the business models related to advertisement,
2.3 Artificial Intelligence Layer sale, service, and insurance are considered. The protocol
development to support traffic data-based business mod-
The artificial intelligence layer represents cloud-based els is an open research theme in CC-V.
computing infrastructure. It includes data centres, and
servers of conventional cloud in case of VuC. The com-
puting resources of vehicles are also included in case of 3. ELEMENT
VCC. The functionalities of the layer include big data In this section, three major elements which formed CC-
computation, analysis, and inferring intelligent decisions V are explored focusing on networking aspect of the
for real-time applications. Apart from cloud computing proposed layered architecture (see Figure 2). In a
and decision-making, the other operations of the layer broader physical structural point of view, the compo-
include mining traffic data for new business models, nents including vehicle, connection, and computation
ensuring computing efficiency, scheduling and paralle- represent the networking aspects of CC-V. The vehicles
ling computation, etc. The operations of the layer are are end-users, and, thus, the services are delivered at
very much similar to the operations of cloud-based serv-
ices. The layer significantly enhances the computing
capability of vehicles through VuC architecture. It also
improves the utilization of resources of vehicles through
VCC architecture.

In terms of protocol perspective, the protocols related to

service, big data analysis (BDA) and VCC are Figure 2: The network component of CC-V
4 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

vehicles. This component is closely related to the percep- considered in CC-V. The network devices are utilized to
tion layer of the proposed architecture. This is due to establish reliable communication between vehicles and
different types of sensors for monitoring speed, direc- cloud. In the case of VuC, the connection is either a
tion, and position which are considered to be attached to direct communication between vehicles and cloud infra-
the vehicles. The second component, namely, connec- structure, or a multi-hop communication using road
tion includes network or communication devices of the side units (RSUs) with vehicles. In the case of VCC, con-
heterogeneous networks. A heterogeneous network nection is mostly direct communication between a vehi-
including VANETs, Wi-Fi, and 4G/LTE is primarily cle and vehicular cloud infrastructure. The connection
considered. This component is closely related with the also defines services level agreement (SLA) between a
coordination layer of the proposed architecture. The last vehicular client and cloud infrastructure. It is due to the
component named computation refers to the computing fact that SLA represents the level of Quality of Service
and storage devices, responsible for the efficient process- (QoS) of an application based on the architecture of
ing of big traffic data, and inferring intelligent decisions. CC-V. Therefore, connection determines delivery and
This last component is closely related to the artificial acceptances of the services.
intelligence layer of the proposed layered architecture.
This is due to the related functionalities of the layer. The
3.3 Computation
relationship among these network components is poten-
tial due to the collective operation requirement for the The computation refers to the computing and storage
services offered by the architecture. The vehicle-to-con- devices where big traffic data are processed and intelli-
nection relationship determines the efficiency of delivery gent decisions are inferred. The computation is broadly
and acceptance of the services provided. The connection divided into three models including computation pro-
and computation determine the quality of information vider, computation consumer, and hybrid. In computa-
of the services. Each element and their role in CC-V are tion provider model, the vehicles are required to register
defined below. with service provider for availing their vehicular resour-
ces to cloud computing architecture using SLA. In com-
putation consumer model, vehicles need to register for
3.1 Vehicle
utilizing the cloud computing services as their own
The vehicles have several in-built technologies and resources. In the case of hybrid model, vehicles need to
resources, which are underutilized in traditional vehicular register for both as a provider and as a consumer. The
communication architecture. The technologies and provider model would generate revenue by improving
resources include telematics wire devices (TWDs), GPS, resources utilization. The consumer model would
sensors, dedicated short-range communication (DSRC), enhance computing capability of vehicles by utilizing
actuators, computer, and smartphones. The utilization of cloud infrastructure. The hybrid model would be more
these resources can be improved through a cooperative complex due to the need for maintaining provider or
collaborative network architecture. The underutilization consumer state for each vehicle.
of the resources motivates the concept of CC-V. This is
due to the fact that the vehicles do not have some of the
4. CLOUD COMPUTING IN VANETs
challenges of mobile devices in MCC, such as lower com-
puting capability, smaller memory size, and battery life. In this section, CC-V is qualitatively reviewed on the
The resources of vehicles can be shared under the frame- basis of a taxonomy depicted in Figure 3. Cloud comput-
work of VCC at many places while travelling. It includes ing is an emerging research theme in VANETs. It is
parking lot, garage, and traffic light. In case of extensive evolving due to the growing need of higher computing
computing need for longer duration, vehicle’s in-built capability at vehicles to expand the various commercial
technologies can be utilized for establishing durable con- services currently available in the Internet to VANETs.
nection to a conventional cloud. The communication is CC-V has four major issues, namely, design of architec-
known as VuC. For both the cases, VCC and VuC, ture, data dissemination, security, and applications. Each
vehicles equipped with modern communication technolo- issue has been investigated in several directions which are
gies are one of the most important constituents. qualitatively and critically reviewed in following sections.

3.2 Connection 4.1 Design of Architecture

The connection refers to the network or communication In this section, related literature on designing architec-
devices of the heterogeneous networks architecture ture to access cloud computing services in VANETs is
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 5

Cloud Computing in VANETs (CC-V)

Design of Architecture Security

Services Computation Communication Privacy Intrusion Detection Authentication

MVCC [25] GCCM [28] M-EVMVC [30] TIaaS [45] ICDI [47] PVCM[49]
VNVC [26] VMIA [29] VCN [31] TCBI [46] VCC-SSF [48]
TVC [27]
Data Dissemination Privacy Preservation
CPAS [50]
DPK-LSV [51]
Distributed Clustering Non-Distributed Clustering PASS [52]
DMCNF [32] RSAP [38]
COHORT [33] ClouDiV [39] Applications
OC-DSG [34] GaaS [40]
CARD [35] PORP [41]
RADD [36] IP6-VCN [42] Safety Efficiency Infotainment
LTE-A [37] RAR-MCV [43]
SMDP [44] EDRS [56] SIAV [60] CITS [63]
WATVSA [57] SCST [61] MSCVN [64]
AEB-ACD [58] RTCV [62]
CVSSs [59]

Figure 3: The taxonomy of CC-V

reviewed. The challenges of lower utilization of VANETs

resources and unstandardized architecture of CC-V are
the main focus in the literature. The design of architec-
ture has been divided into three categories including
services, computation, and communication. Architec-
tures focusing on how to utilize vehicular resources as
cloud resources have been categorized in services. Archi-
tectures focusing on how to utilize traditional cloud
computing services in vehicles, and how to improve
communication between traditional cloud and vehicular
cloud, are categorized in computation and communica-
tion, respectively. Figure 4: The three computing architecture for cloud computing
in vehicular communication
4.1.1 Service
In [25], a concept that merges VANETs with cloud com-
puting (MVCC) has been suggested to address underuti- applications, network coordination, and intelligence
lization of vehicles’ on-board devices and lack of layer of the proposed CC-V-layered architecture.
standard architecture for VCC. The on-board devices
have the capabilities including computation, communi- The paradigm shift from VANETs to VANET-based
cation, and storage of information. An architecture for clouds (VNVC) [26] is an extension of the work in
cloud computing in vehicular communication has been MVCC. The VNCN proposed an architecture including
defined and further divided into three architectural communication paradigm, cloud services, computation,
frameworks including VCC, VuC, and hybrid vehicular and traffic information dissemination through cloud
cloud (see Figure 4). In addition, security and privacy architecture. The major contribution of VNCN is traffic
issues related to VCC are outlined. However, elements information as a service. It handles complex traffic infor-
of the cloud computing in vehicular communication are mation computations using cloud. It also provides serv-
not explicitly defined. The protocols and functions of ices including big traffic data analysis, remote
the architecture are not considered in terms of configuration, car performance checking, smart
6 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

Figure 5: The network model for traffic information as a service

location-based advertisements, and vehicle witnesses; its

application can be related to smart application and artifi-
cial intelligence layer of the proposed CC-V layered
architecture. Figure 5 shows a network model for the
traffic information as a service. In traffic information Figure 6: The network as a service and storage as a service for
dissemination through cloud, moving vehicles serve as cloud computing in vehicular communication
cooperative forwarder to send coarse-grained informa-
tion to the cloud, and to receive fined-grained informa-
tion from the cloud. The functional modules including close to the functions considered under artificial intelli-
cloud processing module (CPM), cloud knowledge base gence and smart application layers. However, the techni-
(CKB), and cloud decision module (CDM) and authenti- cal details of the implementations have not been
cators have been considered at the cloud layer, which are presented.
also applicable to artificial intelligence layer of the CC-
V-layered architecture. However, some performance 4.1.2 Computation
metrics have been measured, yet the distance between In [28], a generic cloud computing model (GCCM) for
communicating vehicles and infrastructure, and the VANETs has been suggested to enhance the availability
delay in connectivity, are not considered. of on-board system for other client vehicles. VANETs
cloud model has been unveiled which consists of two
Another idea of taking VANETs to the clouds (TVC) has concepts including permanent and temporary cloud
been suggested by Olariu et al. [27] to address the prob- layer model. The permanent cloud layer serves as con-
lem of standardization of VANETs to cloud integration ventional cloud, and the temporary cloud serves as
architecture. Furthermore, privacy and security issues VCC. The permanent cloud layer is the same as the tra-
related to integration of VANETs to the cloud have been dition cloud which handles related functions of smart
outlined. The architecture includes two different serv- applications and artificial intelligence layer of the CC-V-
ices, namely, NaaS and STaaS. The cloud computing layered architecture, whereas the temporary cloud can
vehicular communication scenarios for the TVC have be related to the perception layer in the proposed CC-V-
been depicted in Figure 6. It shows the NaaS and STaaS layered architecture. The infrastructure components are
concepts in vehicular environment. From the authors’ grouped into three layers including client, communica-
deduction, NaaS is more suitable for cloud computing in tion, and cloud. The novel intelligent transportation sys-
vehicular communication due to the consideration of tem applications supported by VANETs cloud include
3G- or Wi-Fi-based Internet connection at vehicles. business and research applications, vehicular software,
Therefore, Internet connections can be used by passen- web services and processing cloud backup, and safety
gers on-board to surf Internet. SaaS is also applicable in applications. However, protocols and functions of the
cloud computing over vehicular communication, since architecture are not defined.
vehicles are considered interconnected to form local
clouds. These vehicles can share their memory, and pro- Yan et al. [29] proposed a vehicular cyber-physical sys-
cessor for storage, and computation of large data. The tems (VCPS) based on mobile integrated architecture
implementation of services of traditional cloud comput- (VMIA). The system addresses the increasing demand
ing has been theoretically presented for the vehicular from MCC users to access VCC services. VMIA consists
cloud. Thus, the investigation can be considered very of the conceptual architecture for VCPS with MCC
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 7

presented. The problem has been formulated as a

mixed-integer quadratic programming problem. The
programming problem is based on four constrains
including path selection, virtual machine placement,
resource capacity, and link capacity. The virtual machine
placement issue has been addressed using static offline
placement approach. Virtual machine migration and
system model of cloud computing in vehicular commu-
nication has been explored. The performance of the two
algorithms has been evaluated to test the effect of cost
on the density and resource requirement of virtual
machines, and roadside cloudlets, and traffic rate
requirement of virtual machine. Although implementa-
tion has been carried out, yet well-known network simu-
Figure 7: The VMIA lator has not been used for implementing the
architecture in order to evaluate its effectiveness. The
virtual machine migration is closely related to the men-
capabilities (see Figure 7). VMIA is the integration of tioned CC-V-layered architecture in terms of artificial
VCPS and MCC to provide mobility support to users. intelligence and perception layer, since it deals with
Cloud-supported components are grouped into traffic- migration technology of traditional cloud.
aware mobile geographic information system and
dynamic vehicle routing algorithm. In order to provide Vehicular cloud networking (VCN) and design princi-
driving assistance, a paradigm called traffic-aware ples have been suggested to address the need for intelli-
mobile geographical information system with traffic gent computation for safety and comfort applications in
cloud support system has been discussed. The functions vehicular environment [31]. The VCN architectural for-
of mobile geographical information system can be used mation can be seen in relation to smart applications and
for traffic cloud support by incorporating traffic dynam- artificial intelligence layer of the CC-V-layered architec-
ics with base map management. A decentralized and ture. VCN architecture and design principles have been
proactive dynamic vehicle routing algorithm has been discussed. The VCN architecture is constituted based on
developed to enable drivers to self-organize the traffic cloud computing for vehicular communication and
and shift the system state from either dynamic all-or- information-centric networking for handling safety,
nothing or dynamic user equilibrium to dynamic system comfort, and privacy. In VCN routing, it does not need
optimal. VCPS based on MCC support architecture have to know who sent the information. Furthermore, new
been explained. The CC-V elements are related to model for application and networking has been dis-
VMIA in terms of mobile phone devices which are situ- cussed. However, the elements are not clearly defined in
ated in the vehicle, the communication between mobile terms of the CC-V. Also, functions and protocols are
device and the cloud is aided by connection, and the not considered.
computation is carried out at the MCC layer. The VMIA
findings can be considered very close to the functions of 4.1.4 Comparative Discussion on Design of
smart applications and artificial intelligence layer of the Architecture
CC-V-layered architecture. Nevertheless, frequent inter- The aforementioned literature review on design of archi-
mittent connection might arise due to high mobility of tecture is summarized in Table 1. The summary is based
vehicles. Security-related issues and implementation of on the parameters including contribution, type of archi-
the architecture have not been dealt with. tecture, technique, implementation, and remarks. The
contribution points out the progressive impact of the
4.1.3 Communication articles on the research theme of architecture design for
In [30], an idea of whether to migrate or not has been CC-V. The architecture determines the category from
investigated, by exploring virtual machine migration in the three types including VuC, VCC, and hybrid vehicu-
roadside cloudlet-based vehicular cloud (M-EVMVC). lar cloud (HVC). The techniques identify the approach
The concepts are to address the challenges of sharing followed for addressing the raised issue. The implemen-
resources with high-mobility vehicles. The roadside tation shows implementation tools and performance
cloudlet-based vehicular cloud architecture and two- metric. The critical remarks have been also made. A
phase polynomial heuristic algorithm have been comparison is also presented in Table 2. The comparison
8 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

Table 1: The summary of related literatures on design of architecture

Protocols Contribution Architectures Techniques Implementation Remarks
MVCC [25] VANETs and cloud VuC, VCC, and HVC Architecture design Not considered Intermittent connectivity not addressed
VNVC [26] TIaaS VuC Model and NS-2.34, TranNS, and Communication delay not considered
architecture SUMO
TVC [27] NaaS, SaaS VCC Architecture design Not considered Practically not tested
GCCM [28] Computing model VuC Model design Not considered Model is not associated with any
architecture
VMIA [29] VMIA support VuC Architecture design Not considered Security issue is idea not outlined
architecture
M-EVMVC [30] Algorithms for cloudlet VCC Algorithms Own network simulator Standard simulation tool not considered
development
VCN [31] VCN design VCC Architecture design Not considered Practically not tested

is based on three parameters including technique, archi- layered architecture. It is a routing scheme that uses pro-
tecture, and implementation. The technique is defined active and shortest path methods. An algorithm based
using conceptual model, complete framework, and algo- on one-hop neighborhood follow strategy has been sug-
rithm. The architecture is defined using VuC, VCC, and gested. The algorithm considers three factors for choos-
HVC. The summary and comparative assessment attests ing a follow vehicle including relative mobility, current
that M-EVMVC [30] has more viable architecture than number of follows, and history of cluster. In proactive
other proposed architectures when related to the pro- clustering scheme, high signalling load overhead might
posed CC-V-layered architecture, CC-V elements, better occur due to the frequent update of its follow informa-
algorithm with complexity analysis, and implementation tion, and dynamic change of node’s state. In this work,
proof with a wide range of performance metrics quality of network (QoN) at each node has not been
considered during the selection of follow node. The
aforementioned elements of CC-V is the constituent of
4.2 Data Dissemination the DMCNF, but has not been defined in this study.
In CC-V, clustering is the preferred option for data dis-
semination among vehicles. This is due to the higher A cluster-based vehicular cloud system with learning-
possibility of cloud-based resource sharing while dissem- based resource management (COHORT) has been pre-
inating data in case of clustering. In this section, related sented. COHORT deals the challenges faced during
literature on designing clustering schemes for transmit- deployment of new applications and advancement of
ting data in CC-V is critically explored and compara- ITS services [33]. An example of cluster-based VCC sys-
tively assessed. tem and case scenario for resource management issue is
depicted in Figure 8. A VCC system, q-learning tech-
nique, and queuing strategies have been discussed. VCC
4.2.1 Distributed Clustering system has been designed to conform to clustering pro-
A distributed multi-hop clustering algorithm based on cedures. It then further presents the resource limitation
neighborhood follow (DMCNF) has been presented to difficulties, by grouping vehicles and cooperatively
enhance robustness in clustering algorithms [32]. The
DMCNF basically focuses on how the vehicular cloud
nodes communicate among each other to improve effi-
ciency. Hence, its functional suitability is in coordination
and artificial intelligence layer of the proposed CC-V-

Table 2: A comparative assessment on design of architecture

Techniques Architectures
Protocol CM CF AL VuC VCC HVC IM
MVCC [25] @ @ @ @
VNVC [26] @ @ @ @
TVC [27] @ @
GCCM [28] @ @
VMIA [29] @ @
M-EVMVC [30] @ @ @ @
VCN [31] @ @
CM, conceptual model; CF, complete framework; AL, algorithm; IM, imple-
mentation. @ = Yes. Figure 8: The cluster-based VCC architecture
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 9

providing the resources to the needy vehicles. In essence, game. These players are used for taking adaptive deci-
the clustering structure technique has made flexible sions with regard to effective and reliable data forward-
using fuzzy logic in the cluster head (CH) selection pro- ing. Each player monitors the moves of the other players
cess. Frequent broadcasting and updating of routing in the game for reliable data forwarding. Additionally,
table information might cause high signalling loads on security issues have not been considered. The implemen-
the network. Security challenges with respect to the clus- tation environment has not been discussed.
ter-based architecture have not been considered.
COHORT can be best related to artificial intelligence In [36], a replication-aware data dissemination (RADD)
and smart applications layer of the CC-V-layered archi- is presented for VANETs using location prediction to
tecture considered in terms of functions and its address the challenges of disconnection due to high
protocols. node mobility. The new replication-aware scheme has
been suggested to estimate the location of nodes. An
COHORT resource management issues. Case 1: vehicle algorithm for position estimation, accessing, and routing
requester sends a request to the CH. Four helpers’ messages from remote vehicles to the destination has
vehicles are available and the CH should select one of been developed. The bloom filters are used for searching
them. Case 2: Four requester vehicles have different suitable vehicles for replica assignment. It makes search-
requests and the CH should set up a priority queue to ing faster and improves the total performance of RADD.
allocate limited resources to them, with only one helper. Additionally, radio frequency identification (RFID) tags
are employed on the vehicles and RSUs serve as RFID
Kumar et al. [34] has suggested an optimized clustering reader to gather data from these tags. The tags data serve
for data forwarding using stochastic coalition game in as location information for short-range communication
VCPS (OC-DSG) to address the issue of lesser contact in case of global positioning system failure. The RADD
time for vehicles with access points. A stochastic coali- is closely related to the coordination and perception
tion game for an optimized clustering and an algorithm layer of the CC-V-layered architecture because it deals
for data dissemination in VCPS environment have been with connectivity of vehicular nodes. The security issues
developed. Stochastic coalition game is employed as a of RADD have not been explored.
selection strategy in VCPS. The vehicles are represented
as players in the coalition game. A vehicle accesses a Enabling cooperative relaying in VANETs cloud over
fixed number of resources from the cloud. Learning LTE-A networks has addressed connectivity and device
automata techniques are utilized in vehicles to gather heterogeneity issues in highly populated urban area [37].
and process information from the surrounding based on A cooperative vehicular relaying transmission scheme
pre-stated policies. The optimization clustering scheme has been designed. The scheme contributes towards the
is typically related to coordination layer of the CC-V- formation of an advanced heterogeneous telecommuni-
layered architecture, since it deals with communication cation network. It provides increased networking capa-
between clustered node and RSUs. However, the authors bilities for heavily populated urban areas. This scheme
have claimed, through the evaluation of some perfor- made use of vehicles equipped with low-elevation anten-
mance metrics, that the scheme outperformed the exist- nas and short- and medium-range wireless communica-
ing state-of-the-art schemes. Implementation of the CC- tion technologies. The authors claimed that reasonable
V-layered architecture will enhance performance of the diversity gains and minimized error rate were achiev-
clustering techniques employed. able. Furthermore, there is a significant reduction in the
required transmitting energy when compared to the
A Bayesian coalition game for contention-aware reliable existing transmission scheme, and also improvement in
data (CARD) forwarding in vehicular mobile cloud distance area coverage. Despite enhancing connectivity
addresses the issue of performance degradation due to and heterogeneity, low density and rural environment
the unicast sender-based data forwarding [35]. The for VANETs setting have not been considered. The
problem of reliable data forwarding is formulated as investigations can be considered very close to the func-
Bayesian coalition game using an adaptive learning tions considered under coordination and perception
automata concept. An adaptive learning automata-based layer of the CC-V-layered architecture for cooperative
contention aware data forwarding algorithms for critical relaying transmission in VANETs cloud.
applications in the vehicular mobile cloud has been
developed. The approach is based on coordination and 4.2.2 Non-Distributed Clustering
artificial intelligence layer of the proposed CC-V-layered A data dissemination model for cloud-enabled VANETs
architecture. The vehicles represent the players in the using in-vehicular resource system based on road side
10 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

access point (RSAP) has been suggested to handle con-

nectivity and resources availability issues [38]. The tech-
nology of RSAP is closely applicable to the coordination
and perception layer of the CC-V-layered architecture
since it functions as networking support layer. Different
services and applications of VANETs including connec-
tivity have been pointed through RSAP. The deployment
of cloud, need of VCC service provider, and classifica-
tion of VCC have been discussed. The classification of
VCC includes private or public, dynamic or static, in-
vehicular or out-of-vehicle, NaaS or communication as a
service. It has been further divided as data centric or
address centric, distributed or hybrid. The classification
is based on the factors including participation, mobility,
Figure 10: The GaaS system model
integration point, content management criteria, and
integration with backbone networks. However, security
issues with regard to the outlined model are not
discussed. supported gateway model, GaaS, and a link lifetime pre-
diction scheme have been developed. GaaS model is
Cloud computing-based message dissemination protocol depicted in Figure 10. GaaS has been divided into two
for VANETs (ClouDiv) has been presented to deal with gateways including mobile gateway and static gateway.
the issue of intermittent connectivity due to the higher In mobile gateway, access point is mounted on high-
speed of vehicles and their restricted capacity in terms of mobility vehicles such as public bus in the city. It serves
bandwidth [39]. Figure 9 shows data centre and as a gateway for other vehicles to connect to the cloud.
VANETs node’s routing table in ClouDiV. ClouDiv has In static gateway, the conventional RSUs serve as access
provided an adaptive dissemination of safety and non- point, and RSUs are used as gateway for vehicles to con-
safety messages through cloud computing architecture. nect to the Internet. Link life time prediction scheme
In ClouDiV dissemination, a proactive routing approach considers time of entering or exiting from the gateway
for data centres and reactive approach for vehicle have coverage. In this scheme, qualitative packet delivery has
been adopted. Stochastic routing method has been used been achieved since link lifetime is considered. However,
during dissemination. Hence, since both proactive fault detection has become complex due to large num-
and reactive methods are employed, high signalling load bers of gateways in the network setup. The GaaS model
might occur due to the adoption of proactive routing and link lifetime prediction scheme are nearly related to
table discovery. The data centre has to update the table the functions considered under coordination and per-
frequently. ClouDiV is closely related to functions of ception layer of the CC-V-layered architecture due to
coordination layer of the proposed CC-V-layered archi- the handoff operation and location-based networking
tecture. QoN at each vehicular node has not been support.
considered.
Ikeda et al. [41] have suggested a Performance of Opti-
Cloud-supported seamless Internet access in intelligent mized link state Routing Protocol (PORP) for video
transportation system has been recommended for streaming application in VCC. PORP has addressed
accessing high-quality ITS services [40]. Cloud- the challenges of advancement in communication and
the need of efficient connection. NaaS architecture for
VANET cloud computing has been considered to
investigate the performance of Optimized Link State
x1 Source Node
D D
Routing (OLSR) protocol for video streaming applica-
x3 x7
C1 C2 x9 Destination Node tion. OLSR has used proactive and shortest path
x1 D Data Center i approaches which might cause high signalling load. It
x4 x5 Ci
x2
x6
x8 x9 is required to measure signal strength at each node to
x2 to x8 Intermediate
Nodes choose the best path for routing. The PORP protocol
Wireless Link
can be considered under the functions of coordination
Wired Link
and artificial intelligence layer in terms of the CC-V-
Figure 9: Data centre view for ClodDiV layered architecture.
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 11

IPv6-based VCN (IP6-VCN) has been suggested in [42]. implementation of the concept have not been consid-
The flooding techniques adopted in most of the recent ered. The findings can be closely related under the func-
studies tend to increase the cost of content acquisition tions of coordination and artificial intelligence layer of
due to the content-centric approach for dissemination in the proposed CC-V-layered architecture. The basic ele-
VCC. The vehicular cloud domain system, VCN, and ments of the RAR-MCV have not been adequately
performance evaluation have been discussed. VCN explored. However, our proposed CC-V elements can be
includes addressing structure, vehicular cloud construc- applicable in this study.
tion, vehicular cloud management, and content acquisi-
tion. The addressing structure creates the relationship Zheng et al. [44] proposed a semi-Markov decision pro-
between IP and content data for effective routing. The cess (SMDP)-based resource allocation in VCC system in
vehicular cloud domain system effectively minimizes the order to address underutilization of vehicular resource.
cost of content acquisition. However, in this scheme, The SMDP is closely suitable in relation to the aforemen-
communication between vehicle-to-infrastructure (V2I) tioned CC-V-layered architecture in terms of artificial
has not been considered for assisting the content sharing intelligence and perception layer. A computation resource
and acquisition. The IP-VCN can be related to the afore- allocation scheme has been presented. Furthermore, the
mentioned functions of coordination layer of the CC-V- resource allocation problem has been formulated as an
layered architecture. infinite horizon problem for SMDP. SMDP defines state
space, action space, reward model, and transition proba-
In [43], a Reliable Adaptive Resource Management for bility distribution of the VCC system. In order to develop
cognitive Cloud Vehicular networks (RAR-MCV) has optimal scheme, iteration algorithm has been used to
been suggested to address limited computing capability define the action taken under a specific state. In addition,
and energy of smartphones in car in order to utilize the resource allocation and decision-making schemes, and a
available V2I Wi-Fi connections for traffic data offload- reward system were developed. Authors claimed that rea-
ing [43]. An optimal joint controller and related sup- sonable performance gain has been achieved by the
porting access protocol have been discussed. The SMDP-based scheme within the permissible complexity.
protocol has been claimed to be adaptive, scalable, and However, effects of tolerance parameter to the optimal
distributive. The developed optimal controller dynami- scheme have not been investigated.
cally manages the access time windows at the serving
RSUs. It also manages the access rates and traffic flows 4.2.3 Comparative Discussion on Data Dissemination
at the served VCC system in a distributed and scalable The above reviewed literature on data dissemination in
way. Its implementation complexity is fully independent CC-V focusing on clustering approach is summarized in
from the number of serving RSUs and served VCC sys- Table 3. The summary considers the parameters includ-
tem. Nevertheless, optimized routing management and ing contribution, type of architecture, technique,

Table 3: The summary of related literatures on data dissemination

Protocols Contribution Architecture Techniques Implementation Remarks
DMCNF [32] Multi-hop clustering HVC Follow algorithm NS-2 and Signalling overhead
VanetMobiSim
COHORT [33] Q-learning, Queuing HVC Fuzzy logic OMNet++, SUMO and Framework security issue
Veins
OC-DSCGV [34] Algorithm for system VuC Stochastic coalition game NS-2 with SUMO Urban scenario not considered
CARD [35] Contention-aware forwarding VuC Learning automata VANET MobiSim Security in forwarding issue
RADD [36] Replication-aware communication VCC Algorithm design NS-2 and SUMO Replication overhead
LTE-A [37] Relay node selection VCC Pre-coding transmission Monte–Carlo simulator Standard simulator not
considered
RSAP [38] Optimizing resource sharing VCC Mathematical model OPNET 17.5 Theoretical model validation not
performed
ClouDiV [39] Dissemination protocol VuC Proactive and reactive NS-2 QoS of each routing path issue
routing
GaaS [40] Link life time prediction VuC Mathematical prediction NS-2 Complexity not analyzed
model
PORP [41] Impact of OLSR in video VuC OLSR routing protocol Scrangie network Signalling overhead
dissemination
IP6-VCN [42] Architecture with IPv6 VCC System level design NS-2 IP6 enabled access point issue
RAR-MCV [43] Distributed and adaptive resource VuC Markovian random Mathematical tool Routing optimization issue
management model
SMDP [44] Intelligent decision-aided resource HVC Algorithm design Modelling tool Impact of parameters on decision
allocation issue
12 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

Table 4: A comparative assessment on data dissemination identity of each other, is privacy in CC-V. The identifica-
Techniques Architectures tion of uncooperative neighbour vehicle is intrusion
Protocol CM CF AL VuC VCC HVC IM
detections. The verification of neighbour vehicle base on
DMCNF [32] @ @ @
COHORT [33] @ @ @ @ some reputation is authentication in CC-V. All these
OC-DSG [34] @ @ @ @ major CC-V security issues can be improved on, if the
CARD [35] @ @ @ @
RADD [36] @ @ @
CC-V-layered architecture is critically analyzed and
LTE-A [37] @ @ @ @ incorporated with the security model. Security chal-
RSAP [38] @ @ @ @ lenges of smart applications are basically authentication
ClouDiV [39] @ @ @
GaaS [40] @ @ @ @ issues including automatic safety, information manage-
PORP [41] @ @ @ ment, and services authentication. For artificial intelli-
IP6-VCN [42] @ @ @
RAR-MCV [43] @ @ @ @ @ gence, the major issues with security is in intrusion
SMDP [44] @ @ @ detection of servers and data centres or vehicles which
CM, conceptual model; CF, complete framework; AL, algorithm; IM, imple- form cloud when a computing request from external cli-
mentation. @ = Yes.
ent is being sent to the machines. Coordination layer
can be used as a channel for intrusion and bridging of
implementation, and remarks. The contribution repre- authentication process. Hence, intrusion detection and
sents the progressive impact of the articles on the design authentication issues can be tackled effectively at this
of data dissemination technique for CC-V. layer. The perception layer constitutes in-vehicle devices
and the vehicle nodes, and, thus, the major challenge in
The architecture determines the category from the three this layer is privacy and intrusion detection. With imple-
types including VuC, VCC, and HVC. The technique mentation of the CC-V-layered architecture, a more
defines the approach followed for addressing the raised secured CC-V can be achieved.
issue. The implementation shows simulation tools and
performance metric. The critical remarks have also been
made. A comparative study is also presented in Table 4. 4.3.1 Privacy
The comparative study is based on three parameters Hussain et al. [45] suggested a secure and privacy-aware
including technique, architecture, and implementation. traffic information as a service (TIaaS) for VNVC to
The technique is defined using conceptual model, com- address the vehicle user privacy issue. The privacy issue
plete framework, and algorithm. The architecture is is one of the major concerns why many vehicle users do
defined using VuC, VCC, and HVC. The summary and not want to take part in the information sharing among
comparative study of data dissemination techniques sug- vehicles. Hence, a privacy-aware TIaaS has been dis-
gest that RAR-MVC [43] is a more feasible and better cussed. The network model of TIaaS is shown in
data dissemination technique. RAR-MCV considered Figure 11. The privacy issues cause lower usage of
Markovian random walk as its mathematical modelling VANETs infrastructure including communication, com-
technique which effectively optimizes routing objectives. putation, and on-board storage. A revocation mecha-
The performance evaluation is widely explored with a nism, TIaaS thin-client concept for vehicles, and
range of metrics and environments. Both the distributive efficient mobility vectors framework have been designed.
and non-distributive clustering can be efficiently
enhanced based on coordination, artificial intelligence,
Cloud Level
and perception layer of our proposed CC-V-layered
architecture. The coordination layer will handle cluster- CCP CKB
ing and connectivity issues. The artificial intelligence
Authenticator CDM
layer will handle vehicle clustering in order to perform
cloud computing operations. The perception layer is
linked with the in-vehicular devices and vehicular node Communication Level Wired

itself.
Road Side Unit (RSU), Base Station
(BS), and Access Point (AP)

4.3 Security Wireless VANETs Level Wireless

The major challenges of security in CC-V include pri- Vehicles with Terminal Gateway
vacy, intrusion detection, and authentication. The com- Vehicles connect to RSUs
munication with neighbour vehicles using location,
speed, and direction information, without unfolding the Figure 11: The TIaaS network model
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 13

Transportation Manager

Vehicles’ Parking Lot

RSUs

Vehicles on the Road

Distributed Vehicular Cloud

Figure 12: The network model for vehicular DTN

TIaaS model has used VuC framework. It provided fine-

grained traffic information to vehicles from the cloud
due to the subscriber cooperation with the cloud in
secure and privacy-preserving way. Despite its strength
in concealing location information, it might take longer
processing time due to the encryption scheme adopted. Figure 13: The network model for ICDI
The TIaaS is a security scheme that is best applicable
under the functions of perception and coordination layer
of the CC-V-layered architecture. Implementation is The system is based on clustering, standard crypto-
required to test the feasibility of the scheme. graphic techniques, and reward penalty stochastic
scheme. The system takes intelligent decision, and uses
In [46], a secure and privacy-preserving protocol for pseudo-dynamic clustering technique to select the CH,
cloud-based vehicular delay tolerance networks (DTNs) and then determine the cluster structure. The CH han-
has been presented to address the issue of privacy in dles well-organized dissemination of information and
incentive system and packet forwarding protocol. Net- storage through cloud-based infrastructure. However, to
work model for vehicular DTNs is depicted in Figure 12. secure the learning automata system from malicious
A threshold credit-based incentive (TCBI) mechanism vehicles, a standard cryptographic technique has been
has been designed for privacy preserving packet for- employed. Nevertheless, a lower processing speed and
warding. TCBI encourages vehicles to cooperate with complexity issue might occur due to the complex crypto-
each other by calculating security and privacy and shar- graphic approach and dynamic nature of the system.
ing resources with a certain rule. The privacy-preserving The intrusion detection approach in ICDI is relevant to
packet forwarding protocol has been used to address the the functions under artificial intelligence and coordina-
challenges of layer attack by contracting out privacy-pre- tion layer of the CC-V-layered architecture in terms of
serving transmission proof generation for resource-con- efficient intrusion detection implementation.
strained vehicles. In TCBI, the vehicular privacy is well
secured from both the cloud and transportation manager Kang et al. [48] suggested a VCC service-oriented secu-
for performing any one-way trap-door function. The rity framework (VCC-SSF) to solve the challenges of
failure at either cloud or transportation manager side insufficient internal or external security in vehicles’
might hinder communication of the vehicles. As dis- infrastructure, and information leakage from sensors
cussed in the proposed CC-V-layered architecture, the attached to the vehicles. Framework for VCC-SSF has
layer functions closely related to the TCBI are perception been shown in Figure 14. The suggested framework has
and artificial intelligence layer. been used to handle user-oriented payment and accident
avoidance management services. Furthermore, the
4.3.2 Intrusion Detection framework has provided encryption, authentication,
An intelligent clustering scheme for distributed intrusion access control, confidentiality, integrity, and privacy pro-
(ICDI) detection in VCC has been suggested to address tection of personal information related to users and
the issue of security, i.e. alteration and misuse of infor- vehicles. Accident avoidance management service uses
mation in VCC [47]. A network module of ICDI is the VCC model. The architecture consists of two models:
depicted in Figure 13. A learning automata-assisted dis- one for before accident and another for after accident.
tributive intrusion detection system has been developed. The before accident model has utilized sensors attached
14 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

Application Server Security Core Technology

secured framework that uses key management and revo-
cation technique to secure VCC from malicious nodes
Payment Service
Authentication V2V has been presented. The framework is a decentralized
Product Management one. It has used multiple zone authorities. Each zone
V2I
authority manages an area called zone which consist of
Reservation RSUs, vehicles, and the clients at that zone. Every zone
Encryption
authority serves as a gateway which authenticates actions
IVN
Accident of that zone. It manages the service requests and the data
Management Service flow and preserves the privacy of the cloud entities
Access Control
V2N including vehicles and the client. In this kind of node
Before Accident
protection, too many authentications occur while mov-
Privacy Protection CC
ing from one zone to another. This might degrade the
After Accident
communication between vehicles. Implementation is
also required to examine suitability of the framework.
The investigation can be considered to be close to the
Figure 14: The VCC-SSF framework
functions considered under coordination and smart
application layer of the CC-V-layered architecture. The
PVCM framework can be improved based on the CC-V-
with vehicles to monitor the status of driver’s health and
layered architecture.
driving capabilities. However, this framework might not
be economically viable due to the higher requirements
4.3.3.1 Privacy Preserving Authentication. The pri-
and complexity. The findings can be considered very
vacy-preserving authentication is also one of the major
close to the functions employed under smart applica-
issues in CC-V under privacy and security issues. Pri-
tions, artificial intelligence, and coordination layers of
vacy-preserving authentication is broadly divided into
the CC-V-layered architecture. They can be applied in
three categories including pseudonym changing, silent
implementation of the VCC-SSF.
period, and mix zone. Some of the recent advances in
these categories are critically reviewed. In [50], a condi-
4.3.3 Authentication tional privacy-preserving authentication scheme (CPAS)
Protecting vehicular cloud against malicious (PVCM) for vehicular sensor networks has been suggested to
nodes using zone authorities has been suggested to secure communication between vehicle and infrastruc-
address the issues of weak protection of VCC, and ture in VANETs. CPAS uses pseudo-identity-based sig-
threats to data, resources, and services [49]. A zone nature to secure V2I communication. It enables RSU to
authority framework has been depicted in Figure 15. A validate numerous received signatures in parallel. It sig-
nificantly reduces the total verification duration. The
performance of the CPAS has been investigated in terms
of verification delay. It shows great potentials when
compared with identity based batch verification (IBV).
Privacy-preserving authentication of vehicle-to-vehicle
(V2V) communication has not been considered in this
work. The coordination and perception layer are best
suitable for the CPAS scheme when related to functions
and its protocol of the CC-V-layered architecture.

Lu et al. [51] suggested a Dynamic Privacy-preserving

Key management scheme for Location-based Services in
VANETs (DPK-LSV) to preserve privacy of vehicles,
while improving efficiency of location-based services in
VANETs [51]. DPK-LSV provides anonymous authenti-
cation to vehicles and enables dual registration detection.
The efficient location-based service sessions have been
used. The service sessions are based on several time slots
to hold the session key. An integration of dynamic
Figure 15: The zone authority framework threshold technique with V2V and V2I communication
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 15

has been performed to accomplish the session key’s Table 6: A comparative assessment on security
backward secrecy. The authors claimed about the effec- Techniques Architectures
Protocol CM CF EN VuC VCC HVC IM
tiveness and efficiency of the scheme in relation to fast
TIaaS [45] @ @
key update ratio and low key update delay. A pseudony- TCBI [46] @
mous authentication scheme with strong (PASS) privacy ICDI [47] @ @ @
VCC-SSF [48] @ @
preservation for vehicular communications has been PVCM [49] @ @
suggested to enhance privacy preservation in vehicle CPAS [50] @ @ @ @
communication [52]. PASS has applied pseudonymous DPK-LSV [51] @ @ @ @
PASS [52] @ @ @ @
authentication in preserving vehicles’ privacy. It also CM, conceptual model; CF, complete framework; EN, encryption; IM,
supports RSU-aided distributed certificate services that implementation. @ = Yes.
allow the vehicles to update the information on road. It
has been claimed that PASS outperformed previous
schemes in relation to certificate updating overhead and defined using the model for security, complete frame-
revocation cost. However, only highway scenario was work, and security algorithm. The architecture is defined
considered during implementation. From the findings, it using VuC, VCC, and HVC. The summary and compar-
demonstrates that PASS is very close to the functions ative investigation of the security techniques affirms that
considered under coordination and perception layer of ICDI [47] provides better security and privacy preserva-
the CC-V-layered architecture. In DPK-LSV privacy- tion in CC-V environments. The distributed security
preserving scheme, the functions of coordination and model is presented in ICDI. All the major CC-V security
perception layer of the CC-V-layered architecture are challenges including privacy, intrusion detection,
suitable layers that can enhance authentication of the authentication, and privacy-preserving authentication
system. can be efficiently enhanced based on smart applications,
coordination, artificial intelligence, and perception layer
of our proposed CC-V-layered architecture. Authentica-
4.3.4 Comparative Discussion on Security tion and intrusion detection will be best implemented
The aforementioned literature review on security in CC- on smart applications layer to attain reliability and effi-
V is summarized in Table 5. The summary is based on ciency. Intrusion detection, authentication, and privacy-
the parameters including contribution, type of architec- preserving authentication are more applicable to coordi-
ture, security technique, implementation, and remarks nation layer. For the artificial intelligence layer, intrusion
as security holes. The contribution points out the level of detection and privacy-preserving authentication are
security enhancement provided by the articles in CC-V. applicable. The perception layer is linked with in-vehicu-
The architecture determines the applicability of the secu- lar devices and vehicular nodes, and is thus applicable to
rity technique in the categories including VuC, VCC, the privacy and privacy-preserving authentication in
and HVC. The technique tells the novel method used for relation to CC-V-layered architecture.
providing security. The implementation shows experi-
mental tools and metric for security attestation. The crit-
ical remarks in terms of security holes have also been
4.4 Applications
identified. A comparative investigation is also presented
in Table 6. The comparative investigation is based on CC-V applications are much more essential for realiza-
three parameters including security technique, architec- tion of operational and effective VANET communica-
ture, and implementation. The security technique is tion based on cloud computing [53]. CC-V applications

Table 5: The summary of related literatures on security

Protocols Contribution Architectures Techniques Implementation Remarks
TIaaS [45] Location privacy scheme VuC Location-based encryption No Cryptography overhead
TCBI [46] Incentive mechanism VCC Framework design Custom simulator using System complexity issue
JAVA
ICDI [47] Adaptive intrusion detection VCC Pseudo-dynamic clustering NS-2 and SUMO Adaptivity processing overhead
VCC-SSF [48] Authentication scheme VuC Cryptography based design No Cryptography overhead
PVCM [49] Zone-based revocation system VCC Algorithm development No Complexity analysis issue
CPAS [50] Delay reduction scheme VCC Signature-based framework Matlab V2V communication not
implemented
DPK-LSV [51] Dual registration detection VCC Privacy-preserving Custom simulator Privacy level not defined
detection
PASS [52] Pseudonymous authentication VCC Mathematical framework NGSIM project tool Only highway scenario was
scheme design considered
16 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

Applications in
CC-V C2C C2I C2X Public Interface
Safety Efficiency Infotainment Transport
Smartphone
Emergency call Real time traffic information Wi-Fi in vehicle Authorities Interface
Wrong-way warning Locating parking space Music downloading
Lane change warning Speeding evidence Online streaming
Automatic breaking Navigation area extension SMS using car’s display
Overtaking warning Multi-modal transportation Online radio IP, Transport
Automatic speed control Traffic sign recognition Advertisements
Infrastructures and Administrators
Interface
Figure 16: The applications of CC-V Gateways

System Interface Layer

can be divided into three categories: safety, efficiency, Information and
Data Collection
and infotainment. Safety applications are created to Data Dispatch
enhance vehicle’s behaviour awareness, so as to eradicate
or reduce vehicle crashes via V2V communication. Its Intelligence Layer

Optimization
applications include control loss warning, emergency

Modeling

Analysis
electronic brake lights, blind spot/lane change warning,
etc. V2I communication applications include oversize Knowledge Decision
vehicle warning, railroad crossing warning, curve speed Base Making
warning, etc. Vehicle-to-pedestrian (V2P) communica-
tion includes transit pedestrian indication. Some other
applications in these categories are listed in Figure 16. Cloud Infrastructure as a Service (IaaS) Layer

The major CC-V applications including safety, effi-

ciency, and infotainment can be improved on, if the CC- Figure 17: The emergency response application design
V-layered architecture is critically analyzed and incorpo-
rated with the application models. Safety applications are
most suitable at the smart applications, artificial intelli-
introduced. ERDS system consists of the three main
gence, and perception layer in terms of functions and
layers as shown in Figure 17. The layers include infra-
protocols. The efficiency applications are best linked to
structure as a service, intelligence, and system interface
smart applications and perception layer of the layered
layer. Infrastructure as a service layer consists of base
CC-V. Infotainment can be represented at smart appli-
platform and environment intelligent emergency
cations, artificial intelligence, and perception layer of the
response system. Intelligent layer provides computa-
CC-V layered architecture.
tional model and algorithms. The system interface layer
acquires data from gateways such as Internet, roadside
In a generic term, safety application includes warning
masts, mobile smart phones, and social networks. Light-
and support advisories, and infrastructure and vehicle
hill–Whitham–Richards (LWR) model has been adopted
controls [54].
for modelling the disaster system. Its effectiveness has
been demonstrated in terms of improved disaster evacu-
The efficiency applications provide information on
ation characteristics. Security issues related to the archi-
vehicles’ and drivers’ condition for passenger’s comfort
tecture have not been discussed. The result of the
and health. The applications monitor both car and driv-
findings demonstrates that the functions considered are
er’s performance during journey [55]. Safety applications
closely related under smart applications and coordina-
can be best implemented by considering the functions
tion layer of the CC-V-layered architecture. The CC-V
and its protocols under smart applications, artificial
elements are suitable for emergency disaster response
intelligence, and perception layers of the CC-V-layered
system, since it requires computation, connection, and
architecture.
vehicular node for its communication.

4.4.1 Safety A wireless access technology for vehicular network safety

Transportation and communication plays a critical role applications (WATVSA) has been suggested to address
in disaster response and management in order to combat the issue of non-reliable broadcast of safety messages, in
or reduce loss of life and property [56]. An emergency order to realize standard road safety applications [57].
disaster response system (EDRS) model has been The adopted wireless technologies include time division
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 17

multiple access called vehicle MAC (VeMAC), latest cel- nodes affect VANETs communication. The channel
lular network standards, and IEEE 802.11p standard. In occupancy or busy ratio can be used as feedback mea-
addition, performance of VeMAC protocol has been sure that quantifies the success of information broad-
compared with that of IEEE 802.11p standard, through casted. Consequently, these outcomes are used to
simulation considering both urban and highway scenar- develop feedback control system for transmission range
ios. It considers traffic problems caused by emergency adaptation. The findings can be closely related to the
parking of vehicles in highway scenarios. Authors functions and its protocols considered under smart
claimed that VeMAC has better potentials compared application and coordination layer of the CC-V-layered
with IEEE 802.11p in terms of broadcasting of safety architecture. The major constituent elements of CVSSs
messages in VANETs. The functions and its protocols related to the CC-V elements are including DSRC net-
under smart applications and c-ordination layer of the work and vehicular node. However, effects of some
CC-V-layered architecture are suitable for implementa- parameters such as contention window size on informa-
tion of the WATVSA applications system. The funda- tion dissemination rate have not been derived.
mental components of the application system including
WAVE technology and vehicular node are related to the
4.4.2 Efficiency
proposed CC-V elements. However, the optimal values
A secured incentive-based architecture for vehicular
of VeMAC parameters such as number of time slots per
(SIAV) cloud has been suggested to address the issues of
frame and slot duration are not considered.
underutilization of computational, communication, and
storage capabilities of vehicles because of non-participa-
Segata and Renato [58] presented an automatic emer-
tion in vehicular cloud [60]. A summary of SIAV is
gency braking with realistic analysis of car dynamics
shown in Figure 18. Two design architecture approaches
(AEB-ACD) and network performance as one of the
including system model and secure token reward system
important applications for VANETs safety. The simula-
have been developed for encouraging the vehicles to par-
tion and analysis of driver behaviour awareness have
ticipate in computation and sharing of information.
been conducted. Emergency braking application has
Major components of system model include service pro-
been simulated by embedding mobility, cars’ dynamic,
vider manager, reward token system, revocation author-
and driver behaviour models in to the network simula-
ity, trusted authority, RSUs, and on-board units. Secure
tor. Furthermore, a simpler message aggregation mecha-
token reward system has three major phases, namely,
nism has been presented to enhance message re-
searching resources, requesting reward tokens, and using
propagation during peak load. The complete system per-
the token for cloud services. The efficiency of the system
mits capturing the interactions of the communications
model has not been evaluated. Efficiency applications
with vehicle’s automated break mechanism and driver’s
can be optimally achieved by considering functions
behaviour.
under smart applications and artificial intelligence of the
CC-V-layered architecture. The proposed CC-V
The system yields detailed information on the communi-
cation level during experimentation as claimed by
authors. However, there is a need for refinement of com-
munication channel model and development of the SIAV
vehicular dynamics models. The car dynamics and net-
work analysis can be linked to the CC-V elements and
archive better performance. It represents the basic com- System Model Secure Token Reward
ponents of AEB-ACD including road side infrastructure, System
vehicle node, and cloud computation.
Service Provider Searching Resources
In [59], an analysis of information dissemination in Management(SPM) Requesting Reward
VANETs with application has been presented. The anal- Reward Token Tokens
System(RTS) Using Tokens for Cloud
ysis considers cooperative vehicle safety systems
Revocation Authority Services
(CVSSs) to demonstrate functionalities and viability of (RA)
the systems. Thus, analysis of the effects of different Trusted Authority (TA)
communication ranges and rates has been conducted. RoadSide Unit (RSU)
The novel models that measure network performance in On-board Units (OBUs)
terms of their ability to broadcast tracking information
are presented. The study demonstrates that hidden Figure 18: The system model and secure token reward system
18 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

SaaS Client 2
Service API, Interface Agent and
Environmental Monitoring

Transportation Monitoring
Service Agent Wifi
Application Layer

Health Monitoring Internet

Support Layer
PaaS

Perceptual Layer
Resource Management, Network

Network Layer
Management, Load Balancing,
Real Time Scheduling and End-
to-End Performance Evaluation
IaaS
Identity unit, Security Audit,
Access Control, SLA, Storage
Security and Policy Management
Client 1

Figure 19: The cloud computing module for STS system

elements can be closely related to computation, vehicular

nodes, and RSUs.

Singh et al. [61] suggested a secure and reliable cloud

networks for smart transportation (SCST) services for
accident prevention, monitoring, and controlling system Figure 20: The three-tier system for vehicle cloud
[58]. A smart transportation system and security issues
related to the transportation system have been discussed.
Smart transportation system has four functional layers introduced as road traffic and health care monitoring,
including application layer, support layer, network layer, and other customized services. The context information
and perceptual layer (see Figure 19). Application layer is classified into low- and high-level context. In another
represents various user applications. Support layer cov- point of view, it can be classified into driver, car, and
ers the cloud services, and network layer entails the road traffic contexts.
Internet for connection. And perceptual layer is the cli-
ent layer. An algorithm for vehicle detouring procedure 4.4.3 Infotainment
in smart transportation system has been developed. The In [63], cloud-based ITS (CITS) has been suggested to
algorithm has been used to solve cloud computing address the increasing transportation problem with the
down-time routing problem. The functions of smart help of infotainment applications. A system for multi-
transportation system include preventing accident, find- layered vehicular data cloud has been presented. The
ing destination, and transfer of accident information to system employs cloud computing and Internet of things
the vehicles using cloud. Implementation of vehicle (IoT) technologies. The system has three modules
detouring has not been conducted. The investigation including intelligent parking cloud service, communica-
and implementation can be considered to be close to the tion from VANETs to cloud, and vehicular data mining
functions under smart application and perception layer cloud. Intelligent parking cloud module handles the
of the CC-V-layered architecture. CC-V elements can be decision process of selecting an available parking space
linked to the major constituent of the smart transporta- for vehicles, and the mobile device with android applica-
tion services for accident and emergency prevention, tion service for communication with the cloud. The sys-
controlling, and monitoring. tem has higher interdependence between layers, which
might degrade the performance of the system. The CITS
A Real Time services concept for future Cloud comput- is closely related to the functions under smart applica-
ing-enabled Vehicle (RTCV) networks has been sug- tions, artificial intelligence, and perception layer of the
gested to ensure real-time performance as well as to CC-V-layered architecture. The component which
improve accuracy and comfort degree for drivers [62]. A serves as elements of CITS are also applicable for the ele-
cloud computing system, real-time vehicular cloud serv- ments of CC-V and CC-V-layered architecture.
ices, and context classifications have been presented.
Vehicular cloud system is partitioned into three tiers Multimedia services have become one of the major
including device, communication, and service levels (see research areas of interest in both cloud computing and
Figure 20). The real-time vehicular cloud services are VANETs because of its relevance in both infotainment
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 19

and safety. Thus, multimedia services in cloud-based Table 8: A comparative assessment on applications
vehicular networks (MSCVNs) have been employed to Techniques Architectures
Protocol CM CF AL VuC VCC HVC IM
integrate cloud computing and storage with vehicles, in
EDRS [56] @ @ @
order to increase the accessibility to multimedia services WATVSA [57] @ x x
[64]. Different systems including LTE system for net- AEB-ACD [58] @ @ @ @ @
CVSSs [59] @ @ @ @
work access, and multimedia cloud computing system, SIAV [60] @ @ @
have been suggested. Three-layered cloud-based vehicu- SCST [61] @ @ @
lar network model, which includes cluster layer, physical RTCV [62] @ @ @
CITS [63] @ @
layer, and perception layer, has also been presented. A MSCVN [64] @ @ @
dynamic road monitoring system has been discussed. In CM, conceptual model; CF, complete framework; AL, algorithm; IM, imple-
video up-linking scenario, the MSCVN performs closer mentation. @ = Yes.

to the optimum when compared with two well-known

schedulers including maximum largest weighted delay
first, and exponential. However, delay in connectivity application. A comparative analysis is also presented in
might arise due to the large audio and video files that Table 8. The comparative analysis is based on three
need to be transmitted. parameters including application design technique, suit-
able system, and implementation performed. The appli-
Infotainment applications are best represented by con- cation design technique is defined using the basic
sidering the CC-V-layered architectures’ functions and concept of the application, overall framework, and algo-
protocols in relation to smart applications, artificial rithm of the application operations. The application sys-
intelligence, and perception layers. tem is defined using the suitability in the cloud
environment for vehicular communications including
4.4.4 Comparative Discussion on Applications VuC, VCC, and HVC. The summary and comparative
The aforementioned literature review on application- analysis of the application-based developments in CC-V
based developments in CC-V is summarized in Table 7. suggest that MSCVN [64] is a more practical application
The summary considers the parameters including con- concept with greater user friendly services. The imple-
tribution as service, type of systems suitable for the mentation plan of MSCVN is widely acceptable in CC-V
application, the technique followed in application devel- environments. The application model is more scalable
opment, implementation detail of the application, and due to the plug-in-based service concepts. The CC-V
remarks as strength and weaknesses of the application. applications can be enhanced in terms of the representa-
The contribution highlights the signification of the ser- tions, functions and protocols of the four layers of the
vice provided by the application. The architecture tells aforementioned CC-V-layered architecture. The pro-
the suitability of the application in the cloud-based posed basic elements of CC-V are closely related to the
vehicular communication categories including VuC, MSCVN elements.
VCC, and HVC. The techniques are the development
methods followed in application. The implementation
5. FUTURE RESEARCH CHALLENGES
shows the process and metric for quality attestation
regarding the services of the application. The critical The CC-V is a new paradigm that combines the idea of
remarks are also made in terms of limitations of the cloud computing and VANETs. Many research issues

Table 7: The summary of related literatures on applications

Protocols Contribution Architectures Techniques Implementation Remarks
EDRS [56] Disaster management system VuC System design Mathematical tool Security issues of the system
WATVSA [57] Reliable broadcasting of safety VCC Time division multiple NS-2 and VISSIM Time slots per frame not considered
access
AEB-ACD [58] Automatic break system VCC Algorithms design NS-3 Refinement of communication channel
issue
CVSSs [59] Cooperative vehicle safety VCC Markov model OPNET and SHIFT Contention window size issue
framework (IDR)
SIAV [60] Incentive-based security VCC System-level design No Practically not tested
SCST [61] Vehicle monitoring system VuC Algorithm design No Distributed monitoring issue
RTCV [62] Real-time safety services VuC On-demand approach No Time constraint issue
CITS [63] Architecture for safety VuC Framework design No Multi-layered architecture issue
application
MSCVN [64] Multimedia content classification HVC Taxonomy-based NS-2 Standard mobility model issue
investigation
20 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

need to be addressed for realizing CC-V. These research (5) Standardization and interoperability. The vehicles
issues are discussed below: have heterogeneous devices, such as sensors, GPS,
and smartphones. How to make the varied devices
(1) Architecture design. Due to the fact that CC-V is to work together efficiently is highly required for
still a new area of research, there is no generalized data gathering. Standardization is also required for
standard architecture for this new idea. Although these on-board devices, by looking into compati-
many initial architectures for CC-V have been sug- bility, quality, and implementation of guidelines,
gested, yet standard architectures with implemen- interoperability, and repeatability of on-board
tation details are unavailable [25–31]. Therefore, vehicle equipment and software.
the issue needs thorough exploration. (6) Security and privacy. Security and privacy are also
(2) Data dissemination. Efficient data dissemination in the major challenges in CC-V, because the partici-
CC-V is a challenging issue, due to the high pating users are always mindful of their privacy
dynamicity of vehicles in changing their positions. and whereabouts at any given time. Also, since
The design and how to transmit data in CC-V vehicles rely on information from the cloud or
need to be critically addressed. Even though some other vehicle for their navigations, this might cause
work has been done in [32–44] by suggesting solu- serious havoc; if there is information distortion by
tions to handle various type of data dissemination an intruder, it can lead to waste of time, fuel, or
issues using clustering, data-centric, and other even loss of life. Several authors [45–52] have
routing approaches, yet many issues have not cov- made attempt to see how these issues including
ered, such as location verification, data dissemina- intrusion detection, authentication, and location
tion, and video-centric routing. The conventional privacy could be handled. But the reality is that,
VANETs routing [65–67] may not be suitable in extensive security-related research work need to be
the case of CC-V due to the connectivity chal- done considering security, privacy, authentication,
lenges in mobility. and their efficiency in CC-V.
(3) Data offloading. The complex unrefined data need (7) Delay in cloud--client communication. It is one of
to be offloaded to conventional cloud or to vehicu- the fundamental issues in any cloud-based service
lar cloud for processing. After processing, this due to the dynamic network environment and the
refined data is accessed by vehicles and other consideration of cloud infrastructure in high
related organizations. However, the issue of how mobile vehicular environment. The sparse distri-
to offload unrefined data and access the refined bution of vehicles and the dynamic nature of den-
data in a high-mobility vehicular scenario is sity of vehicles in the network environment are
required to be looked into. Data aggregation and unavoidable. CC-V requires real-time communica-
computation are also required since vehicles use tion decision in safety applications, which is quite
sensors and other on-board devices for collecting challenging considering the delay issue in network
information relating to vehicle, environment, and access [70–73].
traffic. Hence, the issue of data aggregation and (8) Autonomous driving. It is one aspect of ITS which
computation are attached clearly, both at vehicle requires artificial intelligence, learning capabilities
and cloud level. The data need to be aggregated and storage, for making computational decision
and refined for users. on possible route for achieving fast and safe navi-
(4) Application design and deployment. Some potential gation [74,75]. Cloud computing needs to be
VANETs applications are yet to be designed or applied for computation of intelligent data and,
deployed at vehicles for efficient usage, such as pic- subsequently, for analysis, which should be
ture and video coverage as a service (PVCaaS) in accessed by autonomous vehicle.
CC-V. PVCaaS is a service whereby vehicles on (9) Learning-based data storage. It is another aspect of
the road take pictures and video covering of their ITS which is required in order to achieve distrib-
surroundings. The picture and video are automati- uted ITS. It needs some initial storage of informa-
cally offloaded to the cloud for storage and analy- tion in VANETs without any provision for device
sis. This service is useful for road safety and setup [76,77]. The device setup could be RSU or
security organization. Some of the recent works by external access point. Hence, there is need for a
[56–64] tackle some application designs for incen- robust strengthening learning-based, dynamic,
tive-based disaster management, highway traffic and adaptive data storage techniques for VANETs
control [68,69], and multimedia applications. in order to achieve distributive ITS.
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 21

6. CONCLUSION REFERENCES
We have reviewed related work and put forth a frame- 1. A. K. A. Abdalla and A. S. K. Pathan. (2014). “On protect-
work in vehicular communication for CC-V. The frame- ing data storage in mobile cloud computing paradigm,”
work suffices from the merging of cloud computing, IETE Tech. Rev. 31 (1), pp. 82–91.
pervasive sensing, improved network mobility, and in- 2. W. Wu He, G. Yan and L. D. Xu. (2014). “Developing
build vehicle resources. The large volume of underutil- vehicular data cloud services in the IoT environment,”
ized on-board vehicle resources, such as computing IEEE Trans. Ind. Inf. 10 (2), pp. 1587–1595.
power, Internet, and storage, could be utilized by various
clients Internet, in relation to the conventional cloud 3. J. M. De Fuentes, L. Gonz
alez-Manzano, A. I. Gonz
alez-
Tablas and J. Blasco. (2014). “Security models in vehicular
resources. Numerous of these resources can provide sup-
ad-hoc networks: A survey,” IETE Tech. Rev. 31 (1), pp.
port for handling traffic incident. CC-V can generate 47–64.
income, and enhance security and safety. CC-V can also
help minimize the losses in different types of emergency 4. H. Zhang, Q. Zhang and X. Du. (2015). “Toward vehicle-
incidences including fire outbreak, flood, or accidents. assisted cloud computing for smartphones,” IEEE Trans.
Veh. Technol. 64 (12), pp. 5610–5618.
In this work, layered architecture for the CC-V is pre- 5. S. Olariu, T. Hristov and G. Yan, “The next paradigm
sented. Three elements of CC-V including vehicle, com- shift: From vehicular networks to vehicular clouds,” in
putation, and connection are identified, and their Cutting Edge Directions. Mobile Ad Hoc Networking, 2nd
relations are discussed in the aspect of CC-V. Several ed., New Jersey: John Wiley & Sons, 2013, pp. 645–700.
application scenarios, security and privacy, and forma-
6. M. Gerla, “Vehicular cloud computing,” in IEEE the 11th
tion of CC-V have been identified and discussed.
Annual Mediterranean. Ad Hoc Networking Workshop
We present a taxonomy for CC-V and explored related (Med-Hoc-Net), 2012, pp. 152–155.
literature based on the taxonomy. We have also identi-
fied open research issues and challenges with regard 7. J. K. Lee, Y. S. Jeong and J. H. Park. (2015). “s-ITSF: A ser-
to CC-V. vice based intelligent transportation system framework for
smart accident management,” Hum. Cent. Comput. Inf.
Sci. 5 (1), p. 1.
However, a number of issues are still not been thor-
oughly explored by researchers, which include data dis- 8. K. S. Nesamani, L. Chu, M. G. McNally and R.
semination to cloud, data offloading, data-centric Jayakrishnan. (2007). “Estimation of vehicular emissions
routing, intermittent connection, VCN, VANETs appli- by capturing traffic variations,” Atmos. Environ. 41 (14),
cation and deployment, security, and privacy-aware data pp. 2996–3008.
sharing. In conclusion, CC-V could not be fully imple-
9. A. Ghose, P. Biswas, C. Bhaumik, M. Sharma, A. Pal and
mented, except if governments and automobile indus- A. Jha, “Road condition monitoring and alert application:
tries would come together to utilize the benefit of Using in-vehicle smartphone as Internet-connected sen-
the framework of CC-V. Thus, CC-V could be the next sor,” in IEEE International Conference, Pervasive Comput-
paradigm shift that offers feasible and technologically ing and Communications Workshops (PERCOM
viable and smart solutions for traffic issues with eco- Workshops), 2012, pp. 489–491.
nomical gains. 10. N. N. Qadri, M. Fleury, M. Altaf and M. Ghanbari. (2010).
“Multi-source video streaming in a wireless vehicular ad
hoc network,” IET Commun. 4 (11), pp. 1300–1311.
FUNDING
11. M. Hemalatha, S. Nanthini, D. Dhivya and A. Vaishnavi.
The research is supported by Ministry of Higher Education “Intelligent parking system using vehicular data cloud
Malaysia (MOE) and conducted in collaboration with services,” in IEEE International Conference, Innovations
Research Management Center (RMC) at University Teknologi in Information Embedded and Communication Systems
Malaysia (UTM) [VOT number Q.J130000.2528.14H81]. The (ICIIECS), 2015, pp. 1–5.
research is also supported in part by the Jawaharlal Nehru
University, New Delhi, India, under grant UPE-II. 12. F. Dressler, P. Handle and C. Sommer, “Towards a vehic-
ular cloud-using parked vehicles as a temporary network
and storage infrastructure,” in ACM International Work-
ORCID shop. Proceedings of the 2014 on Wireless and Mobile
Technologies for Smart Cities, 2014, pp. 11–18.
Omprakash Kaiwartya http://orcid.org/0000-0001-9669-
8244 13. R. Hussain and H. Oh. (2014). “Cooperation-aware
Sushil Kumar http://orcid.org/0000-0001-9113-2890 VANET clouds: Providing secure cloud services to
22 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

vehicular ad hoc networks,” J. Inf. Process. Syst. 10 (1), pp. 26. R. Hussain, Z. Rezaeifar, and H. Oh. (2015). “A paradigm
103–118. shift from vehicular ad hoc networks to VANET-based
clouds,” Wireless Pers. Commun. 83 (2), pp. 1131–1158.
14. M. Whaiduzzaman, M. Sookhak, A. Gani and R. Buyya.
(2014). “A survey on vehicular cloud computing,” J. Net- 27. S. Olariu, I. Khalil and M. Abuelela. (2011). “Taking
work Comput. Appl. 40, pp. 325–344. VANET to the clouds,” Int. J. Pervasive Comp. Comm.
7 (1), pp. 7–21.
15. IEEE 802.11p Published Standard. (2010). Amendment 6:
Wireless Access in Vehicular Environments (WAVE). 28. S. Bitam, A. Mellouk and S. Zeadally. (2015). “VANET-
Available: http://standards.ieee.org/getieee802/download/ cloud: A generic cloud computing model for vehicular ad
802.11p-2010.pdf hoc networks,” IEEE Wireless Commun. 22 (1), pp. 96–
102.
16. IEEE Working Group for Wireless Standard. IEEE
802.11TM Wireless Local Area Networks. Available: 29. H. Yan, J. Wan, Y. Wang, Z. Wang and Z. Song, Vehicular
http://www.ieee802.org/11/ Cyber-Physical Systems with Mobile Cloud Computing
Support. Cloud Computing. Springer International Pub-
17. IEEE 802.16 based Standards. WiMax forum-technology. lishing, 2013, pp. 27–35.
Available: http://www.wimaxforum.org/
30. H. Yao, C. Bai, D. Zeng, Q. Liang and Y. Fan. (2015).
18. C. Cox, An Introduction to LTE: LTE, LTE-Advanced, SAE “Migrate or not? Exploring virtual machine migration in
and 4G Mobile Communications. 1st ed., New Jersey: John roadside cloudlet based vehicular cloud,” Concurrency
Wiley & Sons, 2012. Comput. Pract. Experience 27 (18), pp. 5780–5792.

19. U.S. Department of Transportation. (2013). IEEE 1609 - 31. E. Lee, E. K. Lee, M. Gerla and S. Y. Oh. (2014). “Vehicu-
Family of standards for Wireless Access in Vehicular lar cloud networking: Architecture and design principles,”
Environments (WAVE). Available: https://www.stand IEEE Commun. Mag. 52 (2), pp. 148–155.
ards.its.dot.gov/Factsheets/Factsheet/80
32. Y. Chen, M. Fang, S. Shi, W. Guo and X. Zheng. (2015).
20. B. Richard and B. Roberto. (2009). “European ITS commu- “Distributed multi-hop clustering algorithm for VANETs
nication architecture: Overall framework-proof of concept based on neighborhood follow,” EURASIP J. Wireless
implementation,” Deliverable. 31, COMeSafety Project. Commun. Netw. 2015 (1), pp. 98–110.

21. R. Khan, S. U. Khan, R. Zaheer and S. Khan, “Future inter- 33. H. R. Arkian, R. E. Atani, A. Diyanat and A. Pourkhalili.
net: The internet of things architecture, possible applica- (2015). “A cluster-based vehicular cloud architecture with
tions and key challenges,” in Proceedings of the learning-based resource management,” J. Supercomput.
International Conference on Frontiers of Information Tech- 71 (4), pp. 1401–1426.
nology (FIT), IEEE, Islamabad, Pakistan, 2012, pp. 257–260.
34. N. Kumar, R. S. Bali, R. Iqbal, N. Chilamkurti and S. Rho.
22. Z. Yang, Y. Peng, Y. Yue, X. Wang, Y. Yang and W. Liu, (2015). “Optimized clustering for data dissemination
“Study and application on the architecture and key tech- using stochastic coalition game in vehicular cyber-physi-
nologies for IOT,” in Proceedings of the International cal systems,” J. Supercomput. 71 (9), pp. 3258–3287.
Conference on Multimedia Technology (ICMT), IEEE,
Hangzhou, China, 2011, pp. 747–751. 35. M. Ikeda, T. Honda and L. Barolli. (2015). “Performance
of optimized link state routing protocol for video stream-
23. O. Brickley, M. Koubek, S. Rea and D. Pesch, “A network ing application in vehicular ad hoc networks cloud com-
centric simulation environment for CALM-based cooper- puting,” Concurrency Comput. Pract. Experience 27 (8),
ative vehicular systems,” in Proceedings of the Interna- pp. 2054–2063.
tional ICST Conference on Simulation Tools and
Techniques ICST, Brussels, Belgium, 2010, pp. 70–75. 36. N. Kumar, J. J. Rodrigues, J. Lloret and A. Dua. (2015).
“Replication-aware data dissemination for vehicular ad
24. R. Baldessari, B B€odekker, M. Deegener, A. Festag, hoc networks using location determination,” Mobile
W. Franz, C. C. Kellum, T. Kosch, A. Kovacs, M. Lenardi, Netw. Appl. 20 (2), pp. 251–267.
C. Menig and T. Peichl. (2007). Car-2-car communication
consortium-manifesto. Available: https://www.car-2-car. 37. M. F. Feteiha and H. S. Hassanein. (2015). “Enabling
org/index.php?id=31 cooperative relaying VANET clouds over LTE-A net-
works,” IEEE Trans. Veh. Technol. 64 (4), pp. 1468–
25. R. Hussain, J. Son, H. Eun, S. Kim and H. Oh, “Rethinking 1479.
vehicular communications: Merging VANET with cloud
computing,” in IEEE 4th International Conference, Cloud 38. M. U. Farooq, M. Pasha and K. U. R. Khan, “A data dis-
Computing Technology and Science (CloudCom), 2012, pp. semination model for cloud enabled VANETs using in-
606–609. vehicular resources,” in IEEE International Conference.
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 23

Computing for Sustainable Global Development (INDIA- based services in VANETs,” IEEE Trans. Intell. Transport.
Com), 2014, pp. 458–462. Syst. 13 (1), pp. 127–139.

39. S. Bitam and A. Mellouk, Cloud Computing-Based Mes- 52. Y. Sun, R. Lu, X. Lin, X. Shen and J. Su. (2010). “An effi-
sage Dissemination Protocol for Vehicular Ad Hoc Net- cient pseudonymous authentication scheme with strong
works. Wired/Wireless Internet Communications. privacy preservation for vehicular communications,” IEEE
Springer International Publishing, 2015, pp. 32–45. Trans. Veh. Technol. 59 (7), pp. 3589–3603.

40. Y. W. Lin, J. M. Shen and H. C. Weng. (2013). “Cloud- 53. D.V. Duren, S. Cadzow, J. Petit, W. Whyte and R. Rausch,
supported seamless Internet access in intelligent transpor- “Connected Vehicle Pilot Deployment Program Phase 2,”
tation systems,” Wireless Pers. Commun. 72 (4), pp. 2081– Technical Report No. FHWA-JPO-17-453, Department
2106. of Transportation, USA, 2016.

41. M. Ikeda, T. Honda and L. Barolli. (2015). “Performance 54. J. Yin, T. ElBatt, G. Yeung, B. Ryu, S. Habermas, H.
of optimized link state routing protocol for video stream- Krishnan and T. Talty, “Performance evaluation of safety
ing application in vehicular ad hoc networks cloud com- applications over DSRC vehicular ad hoc networks,” in
puting,” Concurrency Comput. Pract. Experience 27 (8), Proceedings of the 1st ACM international workshop on
pp. 2054–2063. Vehicular ad HOC Networks, ACM, Oct. 2004.

42. X. Wang. (2015). “IPv6-based vehicular cloud network- 55. H. Hartenstein, VANET: Vehicular Applications and Inter-
ing,” IEEE Commun. Lett. 19 (6), pp. 933–936. Networking Technologies. Vol. 1. Kenneth Laberteaux, Ed.
Chichester: Wiley, Jan. 2010.
43. N. Cordeschi, D. Amendola and E. Baccarelli. (2015).
“Reliable adaptive resource management for cognitive 56. Z. Alazawi, S. Altowaijri, R. Mehmood and M. B. Abdlja-
cloud vehicular networks,” IEEE Trans. Veh. Technol. 64 bar, “Intelligent disaster management system based on
(6), pp. 2528–2537. cloud-enabled vehicular networks,” in IEEE 11th Interna-
tional Conference, ITS Telecommunications (ITST), Aug.
44. K. Zheng, H. Meng, P. Chatzimisios, L. Lei and X. Shen. 2011, pp. 361–368.
(2015). “An SMDP-based resource allocation in vehicular
cloud computing systems,” IEEE Trans. Ind. Electron. 62 57. H. A. Omar, Hassan, L. Ning and Z. Weihua. (2016).
(12), pp. 7920–7928. “Wireless access technologies for vehicular network safety
applications,” IEEE Netw. 30 (4), pp. 22–26.
45. R. Hussain, Z. Rezaeifar, Y. H. Lee and H. Oh. (2015).
“Secure and privacy-aware traffic information as a service 58. M. Segata and L. C. Renato. (2013). “Automatic emer-
in VANET-based clouds,” Pervasive Mob. Comput. 24, pp. gency braking: Realistic analysis of car dynamics and net-
194—209. work performance,” IEEE Trans. Veh. Technol. 62 (9), pp.
4150–4161.
46. J. Jun Zhou, X. Dong, Z. Cao and A. V. Vasilakos. (2015).
“Secure and privacy preserving protocol for cloud-based 59. Y. P. Fallah, C. L. Huang, R. Sengupta and H. Krishnan.
vehicular DTNs,” IEEE Trans. Inform. Forensic Secur. 10 (2011). “Analysis of information dissemination in vehicu-
(6), pp. 1299–1314. lar ad-hoc networks with application to cooperative vehi-
cle safety systems,” IEEE Trans. Veh. Technol. 60 (1), pp.
47. N. Kumar, J. P. Singh, R. S. Bali, S. Misra and S. Ullah. 233–247.
(2015). “An intelligent clustering scheme for distributed
intrusion detection in vehicular cloud computing,” Clus- 60. K. Lim, I. M. Abumuhfouz, and D. Manivannan, Secure
ter. Comput. 18 (3), pp. 1263–1283. Incentive-Based Architecture for Vehicular Cloud. Athens:
Ad-hoc, Mobile, and Wireless Networks. Springer Inter-
48. W. M. Kang, J. D. Lee, Y.-S. Jeong and J. H. Park. (2015). national Publishing, Jun. 2015, pp. 361–374.
“VCC-SSF: Service-oriented security framework for vehic-
ular cloud computing,” Sustainability 7 (2), pp. 2028– 61. D. Singh, M. Singh, I. Singh and H. J. Lee, “Secure and
2044. reliable cloud networks for smart transportation services,”
in IEEE 17th International Conference, Advanced Commu-
49. I. M. Abumuhfouz and K. Lim. (2015). “Protecting vehic- nication Technology (ICACT), Jul. 2015, pp. 358–362.
ular cloud against malicious nodes using zone authori-
ties,” IEEE SoutheastCon., pp. 1–2. 62. J. Wang, J. Cho, S. Lee and T. Ma, “Real time services for
future cloud computing enabled vehicle networks,” in
50. K. A. Shim. (2012). “An efficient conditional privacy-pre- IEEE International Conference, Wireless Communications
serving authentication scheme for vehicular sensor net- and Signal Processing (WCSP), Nov. 2011, pp. 1–5.
works,” IEEE Trans. Veh. Technol. 61 (4), pp. 1874–1883.
63. K. Ashokkumar, B. Sam and R. Arshadprabhu. (2015).
51. R. Lu, X. Lin, X. Liang and X. Shen. (2012). “A dynamic “Cloud based intelligent transport system,” Procedia Com-
privacy-preserving key management scheme for location- put. Sci. 50, pp. 58–63.
24 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES

64. M. K. Jiau, S. C. Huang, J. N. Hwang and A. V. Vasilakos. 71. A. Mehmood, H. Song and J. loret. (2014). “Multi-agent
(2015). “Multimedia services in cloud-based vehicular net- based framework for secure and reliable communication
works,” IEEE Intell. Transport. Syst. Mag. 7 (3), 62–79. among open clouds,” Netw. Protocols Algorithms 6 (4), pp.
60–76.
65. O. Kaiwartya, S. Kumar, D. K. Lobiyal, A. H. Abdullah
and A. N. Hassan. (2014). “Performance improvement in 72. K. Shafiee and C. L. Leung. (2011). “Connectivity-aware
geographic routing for vehicular ad hoc networks,” Sen- minimum-delay geographic routing with vehicle tracking
sors 14 (12), pp. 22342–22371. in VANETs,” Ad Hoc Netw. 9 (2), pp. 131–141.

66. W. Ren. (2015). “uLeepp: An ultra-lightweight energy- 73. J. Li and C. Chunxiao, “Delay-aware transmission
efficient and privacy-protected scheme for pervasive and range control for VANETs,” in Global Telecommunica-
mobile WBSN-cloud communications,” Adhoc Sensor tions Conference (GLOBECOM 2010), 2010 IEEE, 2010,
Wireless Netw. 27, pp. 173–195. pp. 1–6.

67. C. Wu, S. Ohzahata, Y. Ji and T. Kato. (2016). “How to 74. M. Gerla, E. K. Lee, G. Pau and U. Lee, “Internet of
utilize inter flow network coding in VANETs: A backbone vehicles: From intelligent grid to autonomous cars and
based approach,” IEEE Trans. Intell. Transport. Syst. vehicular clouds,” in Internet of Things (WF-IoT), 2014
17 (8), pp. 2223–2237. IEEE World Forum on. IEEE, Mar. 2014, pp. 241–246.

68. O. Kaiwartya and S. Kumar. (2015). “Guaranteed geocast 75. A. Briones, R. de Pozuelo, J. Navarro and A. Zaballos.
routing protocol for vehicular adhoc networks in highway (2016). “Resource allocation on a hybrid cloud for smart
traffic environment,” Wireless Pers. Commun. 83 (4), grids,” Netw. Protocols Algorithms 8 (1), pp. 7–25.
pp. 2657–2682.
76. C. Wu, T. Yoshinaga, Y. Ji, T. Murase and Y. Zhang,
69. R. Chang and C. Chuang. (2014). “A service-oriented “Reinforcement learning based data storage scheme in
cloud computing network management architecture for vehicular ad hoc networks,” Proceedings of IEEE ICC,
wireless sensor networks,” Ad-hoc Sensor Wireless Netw. 2016, pp. 1–6.
22 (1–2), pp. 65–90.
77. C. Yang, J. Ma and J. Ren. (2015). “Provable ownership of
70. Y. Liu, J. Niu, J. Ma, L. Shu, T. Hara and W. Wang. (2013). encrypted files in de-duplication cloud storage,” Ad Hoc
“The insights of message delivery delay in VANETs with a Sensor Wireless Netw. 26 (1–4), pp. 43–72.
bidirectional traffic model,” J. Netw. Comput. Appl. 36 (5),
pp. 1287–1294.
 A. ALIYU ET AL.: CLOUD COMPUTING IN VANETS: ARCHITECTURE, TAXONOMY, AND CHALLENGES 25

Authors
Ahmed Aliyu is a PhD candidate in Fac- Mohammed Joda Usman is a PhD can-
ulty of Computing at Universiti of Tech- didate in Faculty of Computing at Uni-
nologi Malaysia. He received Master’s in versiti of Technologi Malaysia. He
Computer Science and Technology received Master’s in Computer Science
degree in July 2014 from LUT, China. and Technology degree in July 2014
His main research areas are VANETs from LUT, China. His main research
and cloud computing. areas are Cloud Computing, Wireless
Sensor Network, and Artificial
E-mail: ahmedaliyu8513@gmail.com Intelligence.

Abdul Hanan Abdullah received his E-mail: umjoda@gmail.com

PhD degree from Aston University in
Birmingham, United Kingdom in 1995. Sushil Kumar received his PhD degree in
He is currently working as a Professor at Computer Science from School of Com-
Faculty of Computing, Universiti Tekno- puter and Systems Sciences, Jawaharlal
logi Malaysia, Johor Bahru, Malaysia. He Nehru University, New Delhi, India in
was the Dean at the faculty from 2004 to 2014. He is currently working as Assis-
2011. Currently, he is heading Pervasive tant Professor at School of Computer
Computing Research Group. His and Systems Sciences, Jawaharlal Nehru
research interests include WSNs, VANETs, and next genera- University, New Delhi, India. His
tion networks. research interest includes VANETs,
MANETs, and WSNs.
E-mail: hanan@utm.my
E-mail: skdohare@mail.jnu.ac.in
Omprakash Kaiwartya received his PhD
degree in Computer Science from School D. K. Lobiyal received his PhD degree in
of Computer and Systems Sciences, Computer Science from School of Com-
Jawaharlal Nehru University, New Delhi, puter and Systems Sciences, Jawaharlal
India in 2015. He is currently a Postdoc- Nehru University, New Delhi, India in
toral Research Fellow at Faculty of Com- 1996. He is currently working as Profes-
puting, Universiti Teknologi Malaysia, sor at School of Computer and Systems
Johor, Malaysia. His research interests Sciences, Jawaharlal Nehru University,
focus on IoV, EVs, VANETs, and WSNs. New Delhi, India. His research interest
includes Mobile Ad hoc Networks, VoD,
E-mail: omprakash@utm.my Bioinformatics, and Natural Language Processing.

Yue Cao received his PhD degree from E-mail: dlk@mail.jnu.ac.in

the Institute for Communication Sys-
tems (ICS) formerly known as Centre for Ram Shringar Raw received his PhD
Communication Systems Research, at degree in Computer Science from School
University of Surrey, Guildford, UK in of Computer and Systems Sciences,
2013. He was a Research Fellow at the Jawaharlal Nehru University, New Delhi,
ICS. Currently, he is Lecturer in Depart- India in 2011 He is currently working as
ment of Computer Science and Digital Associate Professor at Department of
Technologies, at Northumbria Univer- Computer Science, Indira Gandhi
sity, Newcastle upon Tyne, UK. His research interests focus on National Tribal University, Amarkantak,
delay/disruption tolerant networks, electric vehicle (EV) India. His research interest includes
charging management, information centric networking (ICN), VANETs, MANETs, and WSNs.
device-to-device (D2D) communication and mobile edge com-
puting (MEC). E-mail: rsrao08@yahoo.in

E-mail: yue.cao@northumbria.ac.uk