4278 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO.

5, MAY 2020

A Blockchain-SDN-Enabled Internet of Vehicles

Environment for Fog Computing and 5G Networks
Jianbin Gao , Kwame Opuni-Boachie Obour Agyekum , Emmanuel Boateng Sifah ,
Kingsley Nketia Acheampong, Qi Xia , Xiaojiang Du , Senior Member, IEEE,
Mohsen Guizani , Fellow, IEEE, and Hu Xia

Abstract—The goal of intelligent transport systems (ITSs) Index Terms—Blockchain, fifth generation (5G), fog comput-
is to enhance the network performance of vehicular ad hoc ing, Internet of Things (IoT), Internet of Vehicles (IoV), software-
networks (VANETs). Even though it presents new opportunities defined networks (SDNs), trust, vehicular ad hoc network
to the Internet of Vehicles (IoV) environment, there are some (VANET).
security concerns including the need to establish trust among
the connected peers. The fifth-generation (5G) communication
system, which provides reliable and low-latency communication
services, is seen as the technology to cater for the challenges in I. I NTRODUCTION
VANETs. The incorporation of software-defined networks (SDNs)
HE ADVANCEMENT and wide deployment of wireless
also ensures an effective network management. However, there
should be monitoring and reporting services provided in the
IoV. Blockchain, which has decentralization, transparency, and
T communication technologies have seen the emergence of
a world of connected things. The Internet of Things (IoT) has
immutability as some of its properties, is designed to ensure trust recently sprung up as a novel computing paradigm that pro-
in networking platforms. In that regard, this article analyzes the
combination of blockchain and SDN for the effective operation of
vides Internet connectivity to a plenitude of devices with com-
the VANET systems in 5G and fog computing paradigms. With munication and sensing abilities. These devices are connected
managerial responsibilities shared between the blockchain and to the Internet via a host of communication technologies,
the SDN, it helps to relieve the pressure off the controller due to such as WiFi, ZigBee, long-term evolution (LTE), Bluetooth,
the ubiquitous processing that occurs. A trust-based model that and fifth generation (5G). These technologies enhance a wide
curbs malicious activities in the network is also presented. The
range of applications, including smart homes, smart cities,
simulation results substantially guarantee an efficient network
performance, while also ensuring that there is trust among the smart grid, etc., [1]–[3]. Recently, it has been predicted that a
entities. whopping 25 billion devices will be connected to the Internet
by 2020, with the majority of them being vehicles. The
vehicular ad hoc network (VANET) provides connectivity to
millions of vehicles with the aim of improving traffic effi-
Manuscript received July 31, 2019; revised October 31, 2019; accepted ciency, and this is a very important component of IoT and
November 23, 2019. Date of publication November 27, 2019; date of cur- smart cities. The peers (vehicles, mostly) in a VANET commu-
rent version May 12, 2020. This work was supported in part by the National
Key Research and Development Project under Grant 2017YFB0802900, in
nicate among themselves by sharing road-related information
part by the Program of International Science and Technology Cooperation and conditions, to upgrade passenger and street safeguards
and Exchange of Sichuan Province under Grant 2017HH0028, Grant and, furthermore, course traffic through densely populated
2018HH0102, and Grant 2019YFH0014, in part by the Sichuan Science and
Technology Program under Grant 2017CC0071, in part by the Natural Science
areas.
Foundation of China under Grant 61572115, in part by the CCF-Tencent Open Enormous efforts have been spent on the advancement of the
Research Fund WeBank Special Funding, and in part by the Research on road condition-related systems, such as traffic view systems,
Key Technologies of Energy Internet Mobile and Interconnection Security.
(Corresponding author: Hu Xia.)
message sharing, impact shirking, and secure accident report-
J. Gao is with the School of Resources and Environment, University of ing [4]–[7]. In VANET, the vehicles are equipped with onboard
Electronic Science and Technology of China, Chengdu 610054, China (e-mail: sensors that intelligently exchange messages with other con-
gaojb@uestc.edu.cn).
K. O.-B. Obour Agyekum, E. B. Sifah, and H. Xia are with the
nected peers by virtue of vehicle-to-vehicle (V2V) and/or
School of Computer Science and Engineering, University of Electronic vehicle-to-infrastructure (V2I) communication. 5G communi-
Science and Technology of China, Chengdu 610054, China (e-mail: cation promises to be a very strong technology that allows
obour539@yahoo.com; emmanuelsifah@yahoo.com; xiahu@uestc.edu.cn). vehicular connections in a secured, fast, and reliable way,
K. N. Acheampong is with the School of Information and Software
Engineering, University of Electronic Science and Technology of China, thereby proving the Internet of Vehicles (IoV) as well as
Chengdu 610054, China (e-mail: kingsley@std.uestc.edu.cn). vehicle-to-everything (V2X) integration. It is expected that
Q. Xia is with the Center for Cyber Security, University of Electronic 5G communication technology will see to the requirements
Science and Technology of China, Chengdu 610054, China (e-mail:
xiaqi@uestc.edu.cn). of future IoV applications and also offer intelligent transport
X. Du is with the Department of Computer and Information Sciences, systems (ITSs) in several scenarios that involve high mobility,
Temple University, Philadelphia, PA 19122 USA (e-mail: dxj@ieee.org). dynamic network topology, and high data volume [8].
M. Guizani is with the Department of College of Engineering, Qatar
University, Doha, Qatar (e-mail: mguizani@ieee.org). Because of the sensitive nature of VANET, it is manda-
Digital Object Identifier 10.1109/JIOT.2019.2956241 tory that there always be a secured, trustworthy, and an
2327-4662 
c 2019 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See https://www.ieee.org/publications/rights/index.html for more information.
Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4279

attack-free environment that easily facilitates the dissemi- 3) On the issue or trust, we establish a trust-based model
nation of information throughout the network. Due to how that uses information collated from peers to make
crucial every relayed information is, it is paramount that the decisions on messages sent in the network.
information is authentic and generated from legitimate entities. The organization of this article is as follows. Section II
Also, the network should be capable of handling and satisfy- introduces some of the related literature to this article, while
ing security requirements [9]–[12]. Aside from these security Section III describes the relevant definitions utilized in this
requirements, there is also the need to effectively and effi- article. Sections IV and V introduce the system model and the
ciently manage, control and operate the system, besides the implementation, respectively. Section VI evaluates and dis-
high assurances and capacity VANET brings. Software-defined cusses the implementation, and Section VII concludes this
networking (SDN) has therefore being proposed [13] as an article.
effective network technology that is capable of supporting the
dynamic nature of VANETs and intelligent applications, with
the aim of lowering operation costs through the utilization of
simplified hardware, software, and management. The combina- II. R ELATED W ORKS
tion of SDN and IoV brought about the software-defined IoV IoV provides Internet services to everything related to
(SD-IoV) [14], a technology established to solve challenges vehicular communication: vehicles, drivers, and passengers
in VANETs, such as identifying the efficient exploitation alike. The addition of SDN to the IoV paradigm brings
of the heterogeneity of the network, meeting the diverse about the SD-IoV that ensures an innovation in the support
quality-of-service (QoS) requirements, etc. of IoV services, applications, and communications. Some of
Numerous implementations have focused on the dependable the challenges in the IoV environment, including increase
conveyance of messages among connected entities, and as a in capacity, management, and control in a scalable manner,
result, there has been a minimal focus on assessing the qual- QoS requirements, and efficiency in resource usage, have
ity of information delivered by peers in order to cope with the been illustrated by Chen et al. [15]. Several basic issues
possible compromise of the network by malicious peers. Most relating to vehicular communication, such as low capacity,
systems, like the SDN, incorporate a centralized authority that poor scalability, and connectivity issues, brought about the
oversees the whole operation of the system, and this will bring proposal of the long-term evolution (LTE) mobile communica-
about several security and privacy challenges. The controller of tion technologies in support of vehicular communication [16].
the SDN requires girth descriptions on the statuses of all con- Moreover, there were some open problems with LTE that were
nected items, and if the controller cannot be trusted or there is actively discussed to ensure that there are potential solutions
a breach in the database, severe privacy issues will arise. With for future vehicular communications. Karagiannis et al. [17]
the SDN being centrally controlled, it becomes vulnerable to presented an overview of the applications and associated
the denial-of-service (DOS) attacks and also suffers from a requirements of VANETs, and challenges were also addressed.
single point of failure, such that a failed response by the SDN Zhang et al. [18] also presented an analytical model that
controller limits connectivity in the network. The control func- supports multihop relay of infrastructure-based VANETs. The
tionality can be crippled by a large number of fake requests model was used to analyze uplink and downlink connectivity
that may be received within a short time, if there is any malice probabilities, and the simulation results illustrated the trade-
found. off between system parameters and performance metrics. The
In this article, however, we incorporate blockchain and SDN performance of LTE vehicular networks is cut down when
in VANETs, to effectively and efficiently manage and con- there is the integration of LTE communication technologies
trol the network. The introduction of the blockchain, which into VANETs [19]. The millimeter-wave transmission technol-
is a distributed system, eliminates the single point of failure ogy was therefore proposed to provide the connection between
(SPoF) in a centralized system. Also, the blockchain provides users inside vehicles.
a trust-based message propagation module which supports the In relation to QoS requirements of the network, an adap-
effective evaluation of information sent by connected peers tive edge computing solution that utilizes SDN was proposed
and enhances an immediate control of false information in to monitor and react to network QoS changes in varying
VANETs for fog computing and 5G communication systems. vehicular network scenarios [20]. The computing solution was
The main contributions of this article are outlined as based on fuzzy weighted queuing and regressive admission
follows. control. Enhancing data dissemination in vehicular networks
1) We introduce a vehicular network architecture that brought about the introduction of a cooperative data schedul-
incorporates a blockchain, SDN, and fog computing, ing algorithm that is integrated in roadside units (RSUs) [21].
with the blockchain establishing trust in the system, However, it was realized that there were frequent handover
SDN guaranteeing the control processes are adequately problems due to the limitation of the SDN in RSUs when a
accomplished, and the fog computing technology avoid- lot of vehicles are connected to the RSUs. This, therefore,
ing frequent handovers. It also ensures an effective inhibits the performance of the SDN [22]. With the introduc-
bandwidth utilization in the network. tion of fog computing, Ge et al. [8] were able to oversee an
2) We introduce a communication model that tests the improvement in the performance of the transport management
performance of our system by analyzing the packet system in vehicular networks, after it was realized that there
delivery ratio (PDR) and transmission delay of the is a minimum transmission delay subject to different vehicular
system. densities. However, the problems with routing retransmission
Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
4280 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 5, MAY 2020

and gateway still persist, and those directly affect the system
performance.
On the issue of enabling trust in the network, there has
been a focus on the eviction of malicious peers in the network
through certificate revocation [23], [24]. It was assumed that
the quality of data only depended on the honesty of the sender
without considering the opinions of the peers. It, therefore,
showed that their mode of mitigation was entity oriented, and
it usually takes a considerable amount of time for a central
authority to provide and distribute up-to-date revocation list
before malicious peers are timely identified. Golle et al. [25]
proposed a module that detects and corrects maliciousness in
vehicular networks. They assumed each vehicle has a model
with complete knowledge of the peers in the network, and the
data are trusted if it has a high probability of agreeing with
the model. This is not feasible in practice.
Sharma et al. [26] proposed a distributed architecture for Fig. 1. Simplified IoV environment.
vehicular networks based on the blockchain. They examined
the architecture and its operations, and also made some discus-
sions on its service scenarios and design principles. The focus to mobility constraints and the number of connected devices
of this article was on vehicular information and networking (vehicles). Traffic jams, tall buildings that cause multipath
so that there will be an effective resource sharing to pro- signal propagation, and driver behavior, which are all char-
vide value-added services. However, no SDN component was acteristics of urban centers, inhibit its use. The complexity of
accounted for in this article. Xie et al. [27] designed a trust the network also hinders its applications.
management system which utilized blockchain technology. VANETs, therefore, cannot provide global and sustainable
This article also investigated the security and privacy concerns applications for its customers since most of the objects are
in the transportation system and the vehicular IoT environ- random and unstable, temporary and have a local or discrete
ment, and it made use of SDN. In this article, the proof use of the range. However, the IoV’s technology directions
of work (PoW) and proof of stake (PoS) consensus proto- are twofold: 1) vehicle intelligence and 2) vehicle networking.
cols were used to conduct regular elections in the network. VANET, vehicle telematics, and mobile Internet are the three
However, these protocols incur high energy costs and long main components of vehicle networking. The intelligence of
delays on the network. Furthermore, handover problems in the vehicles encompasses the intelligence of the vehicle and
vehicular networks were not discussed. the driver as unity, which is sometimes referred to as deep
Therefore, in this article, we hope to achieve a decentralized learning, swarm computing, cognitive computing, etc., and is
network management by the utilization of blockchain, and also more powerful than using network topology.
ensuring trust in the network, by introducing a trust model that Therefore, the focus of IoV is on the integration of the intel-
counters malice in the network. A consortium blockchain is ligence of humans, vehicles, objects, and the environment, and
used because it offers a high level of security in data shar- this is a larger network providing services to cities. IoV offers
ing environments. Also, mining costs are less compared to high manageability, operation, and control, and it comprises
the public blockchain. Due to the SDN being centralized in multiple users, vehicles, things, and networks. The goal of
nature, the inclusion of the blockchain alleviates the SPoF IoV is to provide high efficiency in transportation, improve
associated with the centralized systems. The addition of fog services, reduce cost, and also realize the in-depth integration
computing also ensures an effective utilization of bandwidth, of human-vehicle-object-environment, as this is a very impor-
and the avoidance of frequent handovers. tant component of smart cities. Fig. 1 illustrates a simplified
IoV topology.
III. BACKGROUND
In this section, we introduce the major technologies that are B. Software-Defined Networking
involved in this article.
The SDN is a relatively new approach to enhancing network
management in typical communication networks. It is a pro-
A. Internet of Vehicles grammable network that changes the manner in which the
In the typical VANET systems, all participating vehicles are networks are designed and managed by introducing a concep-
turned into either wireless routers or mobile nodes, and that tion that separates the control and data planes, since there is the
in turn creates a wide-ranged network. Other vehicles outside bottleneck for extending and updating data networks due to the
the network can always join when a vehicle in a network gets complex structure of traditional data networks where the data
out of range or drops out. So there is always a connection and control planes are coupled together and embedded in each
of vehicles that creates a mobile Internet. However, VANETs element. The decoupling of these planes by the SDN removes
only cover a very small mobile network that is always subject control functionality from the nodes of the network and turns

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4281

them into simple data/packet forwarding nodes. It also changes

decisions on data and route forwarding from destination based
to flow based.
The SDN controller is an external entity that controls the
affairs of the network. A centralized controller with whole
network awareness, it is possible to manage the entire network
through intelligent orchestration and provisioning systems.
This expands opportunities for implementing services and
applications on top of the network abstraction layer and facili-
tates optimization of routing and switching equipment. It also
implements optimal forwarding policies based on the entire
network. The controllers interact with the controller agent in
the physical network devices via an industry standard proto-
col, OpenFlow [28], which is a common and widely accepted
protocol.
Fig. 2. Blockchain-SDN IoV architecture.
C. Blockchain
Blockchain is a decentralized infrastructure and a distributed
computing paradigm which uses cryptographic chained block
overhead among the vehicles is managed in the fog zones by
structures to validate and store data, consensus algorithms to
the RSU hubs (RSUHs). The overhead is not sent to the SDN
generate and update data, and smart contracts to program and
controller, unless requested for. Hence, RSUHs play an impor-
manipulate data. Each block of data in the blockchain consists
tant role in minimizing the network overhead. They connect
of a block header and a block body. The header comprises the
multiple fog zones with fronthaul links. RSUHs, therefore, act
current version number, the previous block address, the target
as a bridge that connects the VANET infrastructure with the
hash value of the current block, a nonce, a Merkle root, and
SDN controller. They collect state information of the differ-
finally, a timestamp.
ent fog cells connected to it and make forwarding decisions
Traditional databases are centrally deployed and managed
based on its own local intelligence. This therefore reduces the
by a single organization. Blockchain, on the other hand, is
centralized controller’s overhead. Interzonal communication is
decentralized, and management is achieved by multiple partic-
established between the RSUHs and the BBUs. RSUHs are
ipants, with each participant providing nodes and storing data
therefore both data and control devices.
on the chain. This enables an easy sharing of data. The use of
The SDN controller is a core component of the architecture
a time-stamped chain of data makes for easy verification and
and is responsible for resource allocation, mobility manage-
traceability. Each modification to the data is recorded and can
ment, and rule generation. Furthermore, advanced network
be tracked on the blockchain. The utilization of Merkle trees
functionalities, such as data preprocessing and analysis can
and blockchain hash tables enable an irreversible modification
be handled by the controller. The functionalities of the SDN
of the data. Information recorded in the blockchain cannot be
are shared among the RSUs in a hierarchical manner. The
tampered with, after reaching a consensus.
SDN’s logical structure comprises of the data plane, the con-
trol plane, and the application plane, as shown in Fig. 3. The
IV. S YSTEM A RCHITECTURE data plane consists of vehicles, BSs, and RSUs, and its pur-
Fig. 2 illustrates our proposed architecture. Considering the poses are mainly made up of data collection, quantization, and
quick mobility of vehicles and the massive wireless traffic forwarding the data to the control plane. The control plane
between the RSUs and the vehicles, frequent handover should decides the flow rules for the network. It consists of the SDN
be avoided. To solve this problem, the fog computing plat- controller, the RSUHs, and a blockchain. The SDN controller
form is configured at the edge of the network. The platform will not totally control the network. However, the respon-
consists of infrastructures, such as RSUs and base stations sibilities are shared among the RSUHs and the blockchain
(BSs), which are connected to the broadband unit (BBU) con- nodes as well. The RSUHs will give a global view map of the
trollers via broadband connections and vehicles. Most of the network, depending on the information received from the data
data exchanged in the network are processed and saved at the plane.
edges. The vehicles, which are equipped with SDN-enabled The network monitoring module is responsible for monitor-
on-board units (OBUs), act as end users. The potential func- ing the links of the architecture, with the resource allocation
tionalities of these OBUs include packet forwarding, taking providing resources based on the information received. The
vehicular information, such as speed, direction, vehicle type, application plane is responsible for generating rules based
etc., and also information regarding the environment. Power on different application requirements of vehicles and users
control mechanisms, transmission modes (either V2V or V2I), alike, and also providing an efficient network. The application
and channel selection are also the part of their functions. requirements are then forwarded to the control plane.
Moreover, the devices in the fog zones are SDN enabled, The SDN model is made up of forwarding devices, the
meaning they are controlled by the SDN controller. The control controller, and network applications. The forwarding devices

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
4282 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 5, MAY 2020

Fig. 4. Block structure.

The blockchain is an open database that maintains an

immutable distributed ledger and enables registration and
update of transactions in a decentralized form securely. The
Fig. 3. Network’s logical view. blockchain is utilized in this wireless network to reduce admin-
istrative expenses for the dynamic access systems that increase
spectral efficiency. Blockchain is integrated to provide a secure
receive packets, take actions on these packets, and update network orchestration, and more importantly, to ensure trust
counters. The types of actions include dropping packets, modi- among the vehicles in the domain of messages sent to the
fying packet headers, and sending packets out of either a single network. It records all transactions generated in the network
port or multiple ports. Instructions on how to handle packets and maintains a distributed ledger to increase security and pri-
originate from the SDN controller. The forwarding devices vacy. Transactions can be resource sharing, energy trade, etc.
will parse the packet headers and either already know what Since mining processes incur higher costs and long delays
to do with the packet or query the controller. The network in the public blockchain, a consortium blockchain is used.
applications on the SDN controller determine what action to With the provision of a high level of security, a consortium
take on the packet, and push information down to the for- blockchain can provide a secured environment for data sharing.
warding devices. Using the SDN controller as the translator, A blockchain usually consists of three necessitous compo-
the forwarding devices will then take the assigned action on nents: 1) transactions; 2) transactional records (blocks); and
the packet. The forwarding devices will also cache instruc- 3) a consensus algorithm. The transaction includes data types,
tions so that future packets do not require checking with the metadata, an encrypted link to the blocks, and a timestamp.
controller. This lasts until flow entries on forwarding devices Digital signatures are used to guarantee the authentication of
expire. the encrypted transactions. After the transactions are signed,
With less costs involved through software, hardware, and they are packed into a cryptographically tamper-proof structure
management, the SDN is able to handle various topolog- called a block, and these blocks are linked in a chronolog-
ical changes in the vehicular networks. This significantly ical manner by hash pointers to form a chain, hence the
ameliorates the services provided by the IoT in a VANET envi- blockchain. A consensus algorithm is then designed to gen-
ronment. Also, due to constraints in bandwidth, scalability, and erate an agreement on the order of the blocks, and also for
connectivity, gNodeBs are deployed in VANETs to provide validation purposes.
broadband wireless Internet access. RSUs and gNodeBs are The practical Byzantine fault-tolerance (PBFT) consensus
also controlled by the SDN controller through the OpenFlow algorithm is adopted. In this scenario, the RSUHs serve as the
protocol by the use of high capacity fiber optic backhaul miners of the blocks. A leader is selected among them and
links. This helps in achieving consistency in policy and global it creates a block. A small group of preselected nodes, after
management of the network. receiving the block participate in the voting process before
The design of the 5G access network is based on either consensus is reached. On the contrary, every node in a public
the environment and/or the coverage. The environmental sce- blockchain engages in solving a PoW puzzle before partic-
nario can be outdoor or indoor, while the coverage can also ipating in a voting process. PBFT is used in verifying the
be wide area, micro or macro coverage. In smart cities, the correctness of a block. Due to its decentralized nature, each
deployment of the 5G will be in urban, dense populated RSUH has access to the whole transaction record and has the
areas. This area will usually have large cells, continuous and opportunity to become a leader. The leader is chosen before the
quotidian coverage. However, being densely populated will block generation process in a consortium blockchain, and does
amount to high traffic loads, and outdoor-to-indoor coverage. not change until after the consensus process. Fig. 4 illustrates
The network will therefore be complex, and several advanced the structure of a block.
features will be expected to increase the efficiency in these The block structure comprises the block ID, which speci-
networks. The improvement in the efficiency will come from fies the block height, the block header whose components are
technologies, such as cloud and fog computing, SDN, etc. the version, used to keep track of software upgrades, previous
while the blockchain is also utilized to ensure a trustworthy block hash which is a reference to the hash of the previous
and a secured communicating environment. (parent) block, a Merkle root which is a hash of the root of

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4283

the Merkle tree of the block’s transactions, and a timestamp motivation behind the introduction of the trust model in
indicating the time of creation of the block. The transaction the network.
counter keeps track of the number of transactions in the chain The advantages of this system include a secured and intel-
and the transaction field specifies the transactions recorded in ligent resource management and flexible networking. The
the chain. Similarly, the transaction field consists of several blockchain provides a tamper-resistant block generation to aid
elements, such as the transaction ID which uniquely identi- in secure resource sharing. SDN can perceive complicated
fies a specific transaction in the chain, the transaction type wireless networks and the diverse requirements of emerging
which determines the kind of transaction being initiated, a services, and thereby provide an optimal resource alloca-
transaction hash that is the hash value created for that partic- tion policy in a secure communication environment. Also,
ular transaction, a transaction time determining the time the the coexistence of the various network components provides
transaction was made and finally, the transaction details that an avenue for a flexible networking. SDN can make anal-
provides a brief description of the transaction. ysis of the topology of the network, channel assignment,
The nodes in the blockchain consists of an authentication and packet forwarding, and then make decision regarding
server, an access controller, a data management server, and a the selection of appropriate wireless modes to enhance com-
policy management server. The blockchain works concurrently munication, reduce energy consumption, and enhance user
with the SDN to ensure an effective and efficient network experience. Lastly, the incorporation of blockchain enables
management. synchronization and replication of network configurations in a
1) Authentication Server: The authentication server pro- decentralized manner, and that will facilitate network orches-
vides authentication services to the entities in the tration and diagnosis. In addition, computations become easier
network. Before an entity becomes a member of the for the blockchain-enabled systems.
network, it needs to be registered and cryptographic keys
are issued to that entity. After successful registration, and V. S YSTEM I MPLEMENTATION
to establish any form of connection on the network, the
In the implementation of our system, we discuss certifi-
entity needs to be authenticated. The server authenticates
cate issuance and revocation, and also introduce two models,
authorized users of the network to prevent malicious
communication and trust models, to test its effectiveness and
activities, as a compromise of the network will cause
efficiency.
severe security and privacy concerns.
2) Data Management Server: The data management server
takes some of the management duties off the SDN con- A. Certificate Issuance and Revocation
troller. In such a huge, vast environment, there will be The cryptographic algorithms of the public key infrastruc-
access and processing of several requests by the peers ture (PKI) enable secure communication between entities in a
in the network. With the SDN being a centralized one, network, as long as the private keys remain intact [29]. The
there is the tendency of an SPoF, and that will greatly authentication server is responsible for this in order to ensure
inhibit the performance of the network. The decentral- the authenticity and integrity of sent messages. This unit also
ized nature of the blockchain ensures that the problem keeps in its database a list of vehicles’ public keys and their
of SPoF is alleviated. When incoming requests are too identities so as to preserve the anonymity of the vehicles, and
many for the SDN controller to handle, it redirects some also to guarantee legitimacy of vehicles. Furthermore, the unit
of these requests to the data management server for the can trace vehicles in case of disputes for evidence gathering.
necessary actions to be done. The blockchain in itself is equipped with hardware that
3) Access Control: In every network, there is an access con- makes it a perfect candidate to handle high computations,
trol list that ensures the right entities are the ones using including data processing, storage, and transmission, which
the data generated. This is the work of the access con- are the limitations in RSUs and vehicles. Each vehicular mes-
troller. Before data are issued to a peer in the network, sage that is broadcast contains a timestamp, the address of
there is a need to crosscheck whether the entity is associ- the sender, and digital signatures of the authentication server
ated with the access list defined by the data owner. When and an entity called the certificate issuance authority (CIA).
the check returns successful, the access controller grants The CIA is a subentity of the authentication server responsi-
the entity access to the data. For unsuccessful requests, ble for issuing certificates, and revoking public keys under the
they are dropped. adjuration of the authentication server. A certificate transac-
4) Policy Management Server: Finally, there is a policy tion contains the date of expiry and the authorized public key.
management server that penalizes defaulting members of A revoked transaction contains the revoked public key.
the network. A paramount component of the blockchain Each entity generates a public and private key pair
is the smart contract, which is a programmable script that {Ppub , Ppriv }. Upon the entry of Vehicle A into the network, it
evokes actions based on stipulated rules in the contract. submits its public key and other materials for identification to
In a scenario, where a vehicle provides information that the authentication server through a secured channel to prove
turns out to be false, and may have caused an accident its identity. The server will send a signed clearance to the CIA
in the network, the policy management server penalizes if the materials provided are valid. The CIA then issues a cer-
the vehicle, and this can be in the form of toll tickets tificate to the vehicle. The materials provided by Vehicle A are
issued to the vehicle. Such misleading information is the only privy to the authentication server as they contain private

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
4284 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 5, MAY 2020

information which will be used for tracking the identity of the is the mean value in several security studies [32]. Also, a
vehicle in case of disputes. packet can only be correctly received if it arrives after 10 μs
Vehicle A’s certificate comprises the following tuple: CA = on the reception of the previous packet. Requiring a minimal
{PpubV , PpubCIA , PpubAS , XCIA , XAS , RScVA , ExTVA }, where X is signal-to-noise (SNR) ratio, an SNR value of 5 dB is set.
A
the signature, RSc is the vehicle’s reputation score, AS is Furthermore, the probabilistic Nakagami distribution [33]
the authentication server, and ExT is the expiration time for is utilized due to its suitability in the estimation of fading
the certificate. To obtain a certificate, Vehicle A generates phenomena experienced in the mobile communication chan-
its public and private key pair {PpubV , PprivV }, and sends a nels. The successful reception of a signal in the Nakagami
A A
certificate request to the authentication server. This request distribution is a decreasing function of the distance. Nakagami
contains PpubV and the signature XVA , and it will be encrypted covers the entire span from Rayleigh to additive white
A
with the server’s public key PpubAS . The authentication server Gaussian noise (AWGN) channels. The probability density
will then forward this signed clearance to the CIA after a suc- function (pdf) of the Nakagami-m fading is given by the
cessful verification process. After inspecting and verifying the expression
signature in the clearance, the CIA issues the certificate that  m m vm−1  v 1
contains the vehicle’s reputation score, RScVA , and the time of fv (v) = exp m , m≥ (1)
expiry, ExTVA . This information is recorded in the blockchain. v (m) v 2
For a revoked user due to malicious actions, the authenti- where v is the received signal and v is the average received
cation server sends a signed revocation warrant to the CIA, SNR ratio. m is the Nakagami fading figure, and (·) is the
which contains the revoked public key Ppubrvk , and the time of Euler gamma function, represented as (α) = (α − 1)! The
revocation, Trvk . The CIA will then broadcast this transaction Nakagami fading model is known to accurately characterize a
containing the timestamp, revoked public key, and the signa- range of multipath mobile signals, both indoor and outdoor. As
tures of both the CIA and authentication server. All connected m tends to infinity, the channel approaches an AWGN channel.
entities will verify the revocation transaction and update their The Nakagami model is used in this article to estimate the
databases. A revoked transaction is made up of the following capacity of the fading channel, and also to realize which ranges
tuple: RV = {Ppubrvk , PpubCIA , PpubAS , XCIA , XAS , Trvk }. will there be a high probability of message reception, in order
to base our experiment on. m is characterized by the expression
B. Communication Model  2
m=  2  (2)
Essentially, vehicles are equipped with dedicated short- E V2 − 
range communication (DSRC) devices (IEEE 802.11p) [30]
that make up mobile ad hoc networks (MANETs) on the road. where  = E[V 2 ].
These enhance the ITS applications in instances where latency Given an average transmit power, the capacity of the fading
and reliability are the crucial components of communication. channel with the pdf function is given by
The IEEE 802.11p or wireless access in vehicular networks  +∞
(WAVE) is licensed in the 5.9-GHz frequency band and uti- v
C=B log2 fv (v)dv (3)
lizes orthogonal frequency division multiplexing (OFDM) in v0 v0
the provision of V2V and V2I connections up to distances
of 1000 m. The WAVE spectrum is divided into channels where B is the bandwidth of the channel in Hertz (Hz), and
which are labeled from 172 to 184, each with an interval v0 is the cutoff level of the SNR, below which there is a
of 2. Channel 178 is the control channel (CCH), in which suspension in data transmission. Optimal cutoff should satisfy
all the devices listen to for information as it is used to com-  +∞
v − v0
municate safety messages. OFDM has up to seven 10-MHz fv (v)dv = 1. (4)
v0 vv0
channels and a 5-MHz guard band, totaling a 75-MHz band-
width. Out of these seven channels, the first channel (CH 172) In order to achieve maximum capacity (3), tracking should be
is an unused channel, and the last one (CH 184) is reserved made of the channel fade at both transmitter and receiver, and
for future use [high availability low latency (HALL) chan- there should also be an adaptation of rate and power at the
nel]. The other channels are service channels (SCHs) which transmitter. High power allocation should be provided for good
are used to broadcast the Internet protocol (IP)-based services. channel conditions (a large v), and vice versa. The probability
Each communication zone utilizes the CCH and at least one of no transmission is given by
of the 4 SCHs for communication.  v0  +∞
To increase its tolerance to multipath propagation and Pnt = fv (v)dv = 1 − fv (v)dv (5)
effects of Doppler spread on the network, each channel uses 0 v0
a 10-MHz frequency bandwidth, which is in contrast to stemming from the fact that at a level where v < v0 , there is
the 20 MHz used in IEEE 802.11a standard. Binary phase- no data sent.
shift keying (BPSK) is also utilized due to its robustness to Substituting (1) into (4), v0 must meet the condition
interference, and having the lowest coding rate (approximately  
0.5) [31]. The signals are transmitted at a data rate of 3 Mb/s, v m, m vv0  v0 
and the size of each packet is configured to 500 bytes, which − m m − 1, m = v(m). (6)
v0 v

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4285

When m = 1, the Nakagami distribution reduces to the

Rayleigh fading [suitable for non-line-of-sight (nLOS) wire-
less channels]. Equation (6) therefore reduces to
ve−v0 /v v 
0
− E1 =v (7)
v0 v
where E1 (·) represents the first-order exponential integral.
Defining the integral ζr as
 +∞
ζr (ϕ) = ur−1 ln(u)e−ϕu du (8)
1
and substituting (1) into (3), the channel capacity can be
rewritten as
B log2 (e)
C= (mκ)m ζr (mκ) (9)
(m)
where κ = v0 /v. From this, the channel capacity per unit Fig. 5. Probability of message reception against distance.
bandwidth can be deduced as
m−1
C (i, mκ)
= log2 (e) . (10)
B i! C. Network Trust Model
i=0
Unlike MANETs, several challenges, including high mobil-
This can also be written as
ity, rapid topological changes in the network, and an eavesdrop
m−1
C Pi (mκ) on the broadcast message(s) from vehicles make it difficult
= log2 (e) E1 (mκ) + (11) to implement an effective trust model in VANETs. In several
B i
i=0 instances, vehicles may exchange a smattering amount of mes-
where Pi (·) is a Poisson distribution, given by sages due to the challenges mentioned above and therefore
building a long-term interaction between vehicles becomes
i−1
ϕj impractical. Due to the changes in the network topology, a
Pi (ϕ) = e−ϕ . (12)
j! trust model should be established in the presence of conges-
j=0
tion and sparse environments. To verify the trustworthiness of
Using the Rayleigh fading channel, and from (7), the channel the received data, cooperative information from various vehi-
capacity becomes cles and/or RSUs is needed and this information needs to be
immutable, credible, and authentic. In the blockchain, mes-
e−κ
C = B log2 (e) −v . (13) sages from various senders and all the verdicts on the message,
κ delivered by other vehicles, should be recorded as evidence.
The probability of no transmission from (5) can be rewritten In cases where the reputation of a vehicle is warranted, the
using (1) as blockchain can be inferred. This helps in an active and honest
information sharing among the entities.
(m, mκ)
Pnt = 1 − = 1 − Pm (mκ). (14) We deploy the need of a trusted network in order to ensure
(m) that information provided by peers (vehicles) is trustworthy
When no fading is considered, the probability of a suc- enough to be adhered to. Peers in the network also need
cessful message reception is less than 1 in the considered to provide their verdicts on every information (deemed to
communication range. Also, for an increase in the fading be useful) provided by the vehicles. In order to detect a
figure, there is a decrease in the probability of signal recep- falsified information provided by the peer, the leader of a
tion, as shown in Fig. 5. m = 3, 5, and 7 are referred to as clustered vehicular network weighs the various verdicts on the
gruff, medium, and mild fading figures, respectively. However, information provided by the sender, and if the weight is above
OFDM receivers mitigate the time and frequency selectiv- a threshold (also set by the leader), a decision is made on the
ity concerns in wireless environments. From the figure, it is trustworthiness of the message.
evident that there is a high probability of message reception We propose these types of messages in our network: a
within the ranges of 200 and 500 m. Even outside the 500- sender message, which is sent by a vehicle, the verdict on the
m communication range, there are chances of a successful trustworthiness of the message, and finally the sum total mes-
reception (less than 0.3). sage. The sender’s message should comprise the event being
The simulation scenario consists of an 8-km-long highway reported, the confidence the sender has in the message, the
which is bidirectional, and has four lanes in each direction. time of event reporting, and the location the event happened
Averagely, there are ten vehicles per kilometer in each lane, at. When it comes to the verdict given by the other vehicles
and an average traveling speed of 80 km/h. All simulations on the message, the vehicles should decide whether they have
are carried out in MATLAB and NS-3 simulator [34]. assurance in the message and provide their reaction to it. The

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
4286 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 5, MAY 2020

Algorithm 1 Clustering Module Algorithm 2 Message Forwarding Module

Require: Initialization of Parameters: Require: Initialization of Parameters:
Input: ClusterID, LeaderID, Input: SetOfPeers, LeaderID, Verdict, VerdictStatus,
ClusterMember,Message,Signature,Verdict,LeaderStatus Threshold, MessageWeight
Output: MST is the sum total of the message Output: Permission given to relay a message
Ensure: MST is generated and signed Ensure: Message forwarded or dropped
if numberofClusters ≤ MaxClusterNum then if numberofClusters ≤ MaxClusterNum then
for each ClusterMember ∈ ClusterID do Specify Message as M
Specify Message as M Specify Verdict as Vi
Specify LeaderID as LID end if
end for for each i ∈ V do
for i ∈ VerdictNumber do if V == trust then
Specify Verdict as Vi Specify SetofPeers as R
end for else
end if Specify SetofPeers as R
if LeaderStatus == true then end if
Specify XL ← Leader’s Signature Specify Leader as L
Sign Final message, Mi = {M||Vi ||IDi ||Xi } L computes Weight of Vi as W
Set Sum total message, MST ← M1 + Mi − 1 Specify threshold α
Update members as members + 1 if W ≥ α then
else Forward Message
do nothing else
end if Drop Message
end if
end for

reaction to a message is either to trust or mistrust a reported the message, the message is forwarded to other clus-
event. The sum total message comprises the combination of ters, else the message is dropped. When a message is
the sender’s message and the various verdicts provided by the broadcast onto the network, we denote the set of peers
connected peers. Our model is organized into three algorithms that trust the message as R, and those that do not trust
and they are explained as follows. the message as R . The cluster leader then calculates the
1) Cluster Model: For the cluster model, vehicles weight W of the trust and mistrust outcomes. It then
are grouped into clusters geographically. Clustering sets a threshold, α, with which the weight is compared.
enhances network scalability, as message relay is done If W ≥ α, the message is forwarded. If W < α, the
between the leaders of a cluster instead of two neighbor- message is dropped.
ing peers. Each cluster chooses one vehicle to become 3) Judgment Model: The judgmental model concerns deci-
the leader, L, of that group. When a peer, say R, sions taken over a sender’s message after several ver-
broadcasts a message, M, to the group, the members dicts have been obtained from it. With the aggregated
have to provide a verdict V on the message. As an trustworthiness represented as TA over a message M, the
example, a message from a vehicle, say, Ri , will be peers in the network get to decide whether they adhere
read as M0 = [M, ID0 , X0 ], where ID0 is the ID of to the information in the message or not. Algorithm 3
the sender, and X0 is the signature on the message M0 . illustrates the judgment model.
Each vehicle in the cluster will provide a verdict on
the message. The cluster leader then takes the vari-
ous verdicts of the members, Mi = [M, Vi , IDi , Xi ], VI. P ERFORMANCE E VALUATION AND R ESULTS
and provides a sum total message MST , i.e., MST = A highway traffic scenario was designed using MATLAB
[M, Vi , . . . , Vβ , IDi , . . . , IDβ , X ∗ ], i ∈ [1, β], where X ∗ to evaluate the performance under varying distances according
is the aggregation of all the signatures. Algorithm 1 to the IEEE 802.11p standard. To validate the results, NS-3
illustrates the cluster model. was used as the network simulator. Carrier sense multiple
2) Message Forwarding Model: After all the verdicts on a access/collision avoidance (CSMA/CA), which is a method
message have been taken and aggregated, the message used for accessing shared medium, is utilized by the enhanced
needs to be forwarded to other clusters in the network distributed channel access (ECDA), a medium access control
to create awareness. Depending on the trustworthiness (MAC) protocol. A node starts transmitting if it listens to the
of the message, it has to be relayed to other peers, else channel and detects it is free, but backs off if the channel is
the message is dropped. Algorithm 2 describes the for- busy.
warding of a message depending on how trustworthy it With forwarding mechanism and robust routing being
is. When majority of the vehicles in the network trust important issues in VANET, the ad hoc on-demand distance

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4287

Algorithm 3 Judgment Model

Require: Initialization of Parameters:
Input: Sum Total Message, Original Sender, Peers,
Verdicts, Weight, Trustworthiness
Output: TA is the aggregated Trustworthiness on Sum Total
Message MST
Ensure: Message adherence or ignore
Specify Original Sender as S
Specify Peers as R
Specify Verdicts as Vi
Compute Weight of s as W
for each i ∈ R do
Compute trustworthiness of message M as Ti
if Ti == 1 then
adhere
if Ti == −1 then
ignore
end if Fig. 6. PDR against transmission range.
end if
end for

TABLE I
S IMULATION PARAMETERS

Fig. 7. PDR to packet flows.

All numerical results are carried out on an Intel i7 desktop

computer with 3.0-GHz processing speed, 16-GB 1600-MHz
DDR3, using an Ubuntu 16.04.1 to simulate the global
blockchain. Hyperledger Fabric (v1.0.2) [36] is used as the
vector (AODV) routing protocol [35] is utilized as the routing blockchain platform. The peer nodes were operated with a
protocol due to its efficiency because the conventional rout- virtualization technique based on Ubuntu containers.
ing protocols are not suitable in VANETs due to the specific Fig. 6 depicts the performance of the routing protocol
properties of VANET, including speed, mobility, and delay in in varying transmission ranges. There were optimal results
packet delivery, among others. With routes created on demand, obtained for ranges between 200 and 500 m (appreciable range
each node broadcasts packets periodically to maintain the due to a higher message reception in the Nakagami-m fading
topology. The storage of precomputed routes in the route table model), but a steady decline in the packet delivery after the
is only done whenever communication is established. 500-m range. Within the range where the delivery ratio was
PDR and transmission delay are the two parameters consid- appreciable, there was sufficient connectivity which ensured
ered in exploring the performance of the system. The PDR is that every packet arrived at the destination. The number of
a ratio of the total number of packets that are exchanged in the hops was also reduced. However, packet contention affected
communication between a source and destination node. On the the delivery ratio with ranges between 600 and 1000 m.
transmission delay, we consider the time it takes for a packet Also, at a range of 100 m from the sender, the PDR was
to be received when sent through the distances. Simulation low due to an increasing number of hops between source
parameters are given in Table I, and the results are given in and destination. Furthermore, it was because usually for a
Figs. 6–14. first packet to be transmitted, the protocol needs to perform

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
4288 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 5, MAY 2020

Fig. 8. PDR versus the packet number.

Fig. 10. Transmission delay against range.

Fig. 11. Transmission delay at each packet flow.

Fig. 9. PDR against the number of vehicles.

collisions from packet contentions. Other factors that affect the
performance stem from the fact that in some distances, vehi-
a route discovery process before packets can be successfully cles may not be in each other’s transmission range and there
exchanged. An increase in the number of hops is as a result will also be a high vehicular density to maintain connectivity.
of a decrease in the transmission range, and this sometimes Also, PDR decreases with an increasing number of vehicles
leads to disconnectivity in transmission. CSMA/CA rules limit due to queuing and buffering effects as the network becomes
nodes from communicating, and this restricts the reuse of dense, which results in packet losses. With an increase in
channel bandwidth. mobility, the probability that a vehicle leaves the network
The optimal PDR ranges were then investigated for the flow becomes high. This is seen in Fig. 9.
of the packets. It was realized in Fig. 7 that there was a direct Fig. 10 depicts the performance of the protocol in terms of
proportion between the performance and an increase in the the transmission delay with respect to the different transmis-
transmission distance for each round of the packet flow. This sion ranges. The transmission delay basically decreases as the
was due to disconnectivities in the early stages of the connec- number of hops increases. A very high delay was experienced
tion at the various distances due to an increase in the number of at 100 m due to MAC and propagation delays. Sufficient con-
hops between the source and the destination in packet delivery. nectivity and a high vehicular density within ranges of 200
Also, the performance was affected due to the scanty vehicular and 500 m resulted in a decrease in the delay, but then there
density. was a steady rise in the delay after the 500-m range due to
Similarly, Fig. 8 shows the average PDR for a number nodes sending contended data which produce collisions and
of packets as they are transmitted through the various dis- interference.
tances. As noted from the figure, the transmission range shows Again in Fig. 11, we analyze the performance of the routing
a positive correlation with the protocol’s performance. As protocol for each packet flow during the optimal transmission
the number of packets increase, the PDR decreases due to ranges in terms of the delay in transmission. It was realized

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4289

Fig. 12 analyzes the performance of the protocol by con-

sidering the average delay experienced as the number of
packets is increased. It was realized that increasing the num-
ber of packets increases the average transmission delay due to
interference that occurs as a result of a high packet contention
among the nodes.
As the number of vehicles increase, the transmission delay
increases because the network becomes congested and the time
it takes for the packet to reach the destination increases. This
is depicted in Fig. 13.
Fig. 14 shows the processing time of blocks in the
blockchain network. Latency is a paramount parameter when
blockchain systems are implemented in any structure. In this
article, however, the use of blockchain introduces time delays
in the SDN. The miners (RSUHs) generate blocks periodically,
nevertheless, there is little overhead.
Fig. 12. Transmission delay versus the packet number.

VII. C ONCLUSION
This article has presented a blockchain-SDN-enabled archi-
tecture for the IoV in the 5G and fog computing systems.
Even though the SDN handles the management processes
for the VANET systems in 5G networks, the addition of the
blockchain enhances the efficiency of the network. Also, the
blockchain is incorporated to enhance trust among vehicles as
every information provided by a vehicle is deemed very impor-
tant. Fog computing, to alleviate handover problems among
the vehicles, was also presented. In this article, we proposed a
trust system whereby reputation scores were given to vehicles
who provide messages which were deemed trustworthy based
on the verdicts given by the connected peers. Furthermore,
we exploited the performance of our system by analyzing the
PDR and transmission delay as information was transmitted
among the vehicles in the network. According to the results
Fig. 13. Transmission delay against the number of vehicles.
obtained, our proposal presents a viable solution for improv-
ing VANET communication, and also a platform where all
connected entities can trust each other.
For our future work, we will explore in detail other routing
protocols which will be suitable for VANET communications,
and also consider some leader election protocols in the selec-
tion of a cluster head among the vehicles. Other performance
metrics to improve the performance will also be addressed.

R EFERENCES
[1] I. Yaqoob, I. A. T. Hashem, Y. Mehmood, A. Gani, S. Mokhtar, and
S. Guizani, “Enabling communication technologies for smart cities,”
IEEE Commun. Mag, vol. 55, no. 1, pp. 112–120, Jan. 2017.
[2] S. Chen, H. Xu, D. Liu, B. Hu, and H. Wang, “A vision of IoT:
Applications, challenges, and opportunities with China perspective,”
IEEE Internet Things J., vol. 1, no. 4, pp. 349–359, Aug. 2014.
[3] Y. Mehmood, F. Ahmad, I. Yaqoob, A. Adnane, M. Imran, and
S. Guizani, “Internet-of-Things-based smart cities: Recent advances and
challenges,” IEEE Commun. Mag., vol. 55, no. 9, pp. 16–24, Sep. 2017.
[4] T. Nadeem, S. Dashtinezhad, C. Liao, and L. Iftode, “TrafficView:
Fig. 14. Block processing time. Traffic data dissemination using car-to-car communication,” ACM
SIGMOBILE Mobile Comput. Commun. Rev., vol. 8, no. 3, pp. 6–19,
that as the distance increased, the delay decreased due to a 2004.
[5] Q. Xu, T. Mak, J. Ko, and R. Sengupta, “Vehicle-to-vehicle safety mes-
decrease in the number of hops. This decrease enhances an saging in DSRC,” in Proc. ACM Int. Workshop Veh. Ad Hoc Netw.,
efficient connectivity among vehicles. Philadelphia, PA, USA, 2004, pp. 19–28.

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
4290 IEEE INTERNET OF THINGS JOURNAL, VOL. 7, NO. 5, MAY 2020

[6] T. ElBatt, S. K. Goel, G. Holland, H. Krishnan, and J. Parikh, [30] IEEE. (Apr. 2009). IEEE P802.11p/D5.0 Draft Amendments for
“Cooperative collision warning using dedicated short range wireless Wireless Access in Vehicular Environments (WAVE). [Online]. Available:
communications,” in Proc. ACM Int. Workshop Veh. Ad Hoc Netw., http://ieeexplore.ieee.org/servlet/opac?punumber=4810961
Los Angeles, CA, USA, 2006, pp. 1–9. [31] J. Maurer, T. Fugen, and W. Wiesbeck, “Physical layer simulations of
[7] S. U. Rahman and U. Hengartner, “Secure crash reporting in vehicular IEEE802.11a for vehicle-to-vehicle communications,” in Proc. IEEE
ad hoc networks,” in Proc. 3rd Int. Conf. Security Privacy Commun. 62nd Veh. Technol. Conf. (VTC-Fall), vol. 23. Dallas, TX, USA,
Netw. Workshops (SecureComm), Nice, France, 2007, pp. 443–452. Sep. 2005, pp. 1849–1853.
[8] X. Ge, Z. Li, and S. Li, “5G software defined vehicular networks,” IEEE [32] M. Raya and J.-P. Hubaux, “The security of vehicular ad hoc networks,”
Commun. Mag., vol. 55, no. 7, pp. 87–93, Jul. 2017. in Proc. 3rd ACM Workshop Security Ad Hoc Sensor Netw. (SASN),
[9] H. Zhong, B. Huang, J. Cui, Y. Xu, and L. Liu, “Conditional privacy- Alexandria, VA, USA, Nov. 2005, pp. 11–21.
preserving authentication using registration list in vehicular ad hoc [33] M. Nakagami, “The m-distribution: A general formula of intensity dis-
networks,” IEEE Access, vol. 6, pp. 2241–2250, 2018. tribution of the rapid fading,” in Statistical Methods in Radio Wave
[10] Y. Sun et al., “Security and privacy in the Internet of Vehicles,” in Propagation, W. C. Hoffman, Ed. Oxford, U.K.: Pergamon, 1960.
Proc. Int. Conf. Identification Inf. Knowl. Internet Things (IIKI), Beijing, [34] W. R. Braun and U. Dersch, “A physical mobile radio channel model,”
China, Oct. 2015, pp. 116–121. IEEE Trans. Veh. Technol., vol. 40, no. 2, pp. 472–482, May 1991.
[11] Z. Lu, G. Qu, and Z. Liu, “A survey on recent advances in vehicular [35] R. K. Jaiswal and C. Jaidhar, “An applicability of AODV and OLSR pro-
network security, trust, and privacy,” IEEE Trans. Intell. Transp. Syst., tocols on IEEE 802.11p for city road in VANET,” Internet of Things,
vol. 20, no. 2, pp. 760–776, Feb. 2019. Smart Spaces, and Next Generation Networks and Systems. Cham,
[12] J. Cui, W. Xu, H. Zhong, J. Zhang, Y. Xu, and L. Liu, “Privacy- Switzerland: Springer, 2015, p. 286–298.
preserving authentication using a double pseudonym for Internet [36] E. Androulaki et al., “Hyperledger fabric: A distributed operating
of Vehicles,” Sensors, vol. 18, no. 5, 2018, Art. no. E1453, system for permissioned blockchains,” in Proc. 13th EuroSys Conf.,
doi: 10.3390/s18051453. Apr. 2018, p. 30.
[13] S. Sezer et al., “Are we ready for SDN? Implementation challenges
for software-defined networks,” IEEE Commun. Mag., vol. 51, no. 7,
pp. 36–43, Jul. 2013. Jianbin Gao received the Ph.D. degree in computer
[14] K. Zheng, Q. Zheng, P. Chatzimisios, W. Xiang, and Y. Zhou, science from the University of Electronic Science
“Heterogeneous vehicular networking: A survey on architecture, chal- and Technology of China (UESTC), Chengdu,
lenges, and solutions,” IEEE Commun. Surveys Tuts., vol. 17, no. 4, China, in 2012.
pp. 2377–2396, 4th Quart., 2015. He was an Visiting Scholar with the University of
[15] J. Chen, H. Zhou, N. Zhang, P. Yang, L. Gui, and X. Shen, “Software Pennsylvania, Philadelphia, PA, USA, from 2009 to
defined Internet of Vehicles: Architecture, challenges and solutions,” 2011. He is currently an Associate Professor with
J. Commun. Inf. Netw. vol. 1, no. 1, pp. 14–26, 2016. UESTC.
[16] G. Araniti, C. Campolo, M. Condoluci, A. Iera, and A. Molinaro, “LTE
for vehicular networking: A survey,” IEEE Commun. Mag., vol. 51,
no. 5, pp. 148–157, May 2013.
[17] G. Karagiannis et al., “Vehicular networking: A survey and tutorial on
requirements, architectures, challenges, standards and solutions,” IEEE Kwame Opuni-Boachie Obour Agyekum received
Commun. Surveys Tuts., vol. 13, no. 4, pp. 584–616, 4th Quart., 2011. the B.Sc. degree in telecommunications engineering
[18] W. Zhang et al., “Multi-hop connectivity probability in infrastructure- from the Kwame Nkrumah University of Science
based vehicular networks,” IEEE J. Sel. Areas Commun., vol. 30, no. 4, and Technology, Kumasi, Ghana, in 2014, and the
pp. 740–747, May 2012. M.Eng. degree in communication and information
[19] T. Taleb and A. Ksentini, “VECOS: A vehicular connection steering engineering from the University of Electronic
protocol,” IEEE Trans. Veh. Technol., vol. 64, no. 3, pp. 1171–1187, Science and Technology of China, Chengdu, China,
Mar. 2015. in 2017, where he is currently pursuing the Ph.D.
[20] M. Jutila, “An adaptive edge router enabling Internet of Things,” IEEE degree in computer science and technology.
Internet Things J., vol. 3, no. 6, pp. 1061–1069, Dec. 2016. He is a Researcher with RGlobal, Chengdu. His
[21] K. Liu, J. K. Y. Ng, V. C. S. Lee, S. H. Son, and I. Stojmenovic, research interests include blockchain technologies,
“Cooperative data scheduling in hybrid vehicular ad hoc networks: big data security and privacy, and wireless communication.
VANET as a software defined network,” IEEE/ACM Trans. Netw.,
vol. 24, no. 3, pp. 1759–1773, Jun. 2016.
Emmanuel Boateng Sifah received the B.Sc. degree
[22] T. Taleb and K. B. Letaief, “A cooperative diversity based handoff in telecommunications engineering from Ghana
management scheme,” IEEE Trans. Wireless Commun., vol. 9, no. 4, Technology University College, Accra, Ghana, in
pp. 1462–1471, Apr. 2010. 2014, and the M.Eng. degree in computer sci-
[23] J. J. Haas, Y.-C. Hu, and K. P. Laberteaux, “Design and analysis of a ence and technology from the School of Computer
lightweight certificate revocation mechanism for VANET,” in Proc. 6th Science and Engineering, University of Electronic
ACM Int. Workshop Veh. Internetw. (VANET), New York, NY, USA, Science and Technology of China, Chengdu, China,
2009, pp. 89–98. in 2017, where he is currently pursuing the Ph.D.
[24] M. Raya, P. Papadimitratos, I. Aad, D. Jungels, and J.-P. Hubaux, degree in computer science and technology.
“Eviction of misbehaving and faulty nodes in vehicular networks,” His current research interests include blockchain
IEEE J. Sel. Areas Commun., vol. 25, no. 8, pp. 1557–1568, technologies, big data security, and privacy.
Oct. 2007.
[25] P. Golle, D. Greene, and J. Staddon, “Detecting and correcting malicious
data in VANETs,” in Proc. ACM Int. Workshop Veh. Ad Hoc Netw., Kingsley Nketia Acheampong received the B.S.
Philadelphia, PA, USA, 2004, pp. 29–37. degree in computer engineering from the Kwame
[26] P. K. Sharma, S. Y. Moon, and J. H. Park, “Block-VN: A distributed Nkrumah University of Science and Technology,
blockchain based vehicular network architecture in smart city,” J. Inf. Kumasi, Ghana, in 2014, the M.Eng. degree in soft-
Process. Syst., vol. 13, no. 1, pp. 184–195, 2017. ware engineering from the University of Electronic
[27] L. Xie, Y. Ding, H. Yang, and X. Wang, “Blockchain-based secure Science and Technology of China (UESTC),
and trustworthy Internet of Things in SDN-enabled 5G-VANETs,” IEEE Chengdu, China, in 2017, where he is currently
Access, vol. 7, pp. 56656–56666, 2019. pursuing the Ph.D. degree with the School of
[28] N. McKeown et al., “OpenFlow: Enabling innovation in campus Information and Software Engineering.
networks,” ACM SIGCOMM Comput. Commun. Rev., vol. 38, no. 2, He is a Team Lead with RGlobal, Chengdu. and
pp. 69–74, 2008. his current field is with the SunData Group, UESTC.
[29] M. Al-Bassam, “SCPKI: A smart contract-based PKI and iden- His research interests include natural language processing, data analysis and
tity system,” in Proc. ACM Workshop Blockchain Cryptocurrencies analytics, and optimization of data structure, and algorithms using artificial
Contracts, 2017, pp. 35–40. intelligence.

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.
GAO et al.: BLOCKCHAIN-SDN ENABLED IoV ENVIRONMENT FOR FOG COMPUTING AND 5G NETWORKS 4291

Qi Xia received the B.Sc., M.Sc., and Ph.D. Mohsen Guizani (S’85–M’89–SM’99–F’09)
degrees in computer science from the University received the bachelor’s (with distinction) and
Electronic Science and Technology of China master’s degrees in electrical engineering, and
(UESTC), Chengdu, China, in 2002, 2006, and 2010, the master’s and doctoral degrees in computer
respectively. engineering from Syracuse University, Syracuse,
She was a Visiting Scholar with the University NY, USA, in 1984, 1986, 1987, and 1990,
of Pennsylvania, Philadelphia, PA, USA, from 2013 respectively.
to 2014, and has published over 20 papers. She is He is currently a Professor with the Department
the Vice Dean of the Center for Cyber Security, of College of Engineering, Qatar University, Doha,
where he is currently an Associate Professor with Qatar. He has over 400 publications in refereed
UESTC. She is PI of the National Key Research journals and conferences.
and Development Program of China in Cyber Security. Prof. Guizani was the Chair of the IEEE Communications Society Wireless
Dr. Xia won the National Scientific and Technological Progress Second Technical Committee and the TAOS Technical Committee. He has served
Prize in 2012. various positions for several academic institutions and currently serves on
the editorial boards for several international technical journals. He is the
Founder and the Editor-in-Chief of Wireless Communications and Mobile
Computing (Wiley).

Xiaojiang Du (SM’09) received the B.S. and M.S.

degrees in electrical engineering from Tsinghua
University, Beijing, China, in 1996 and 1998, respec-
tively, and the Ph.D. degree in electrical engineering
from the University of Maryland at College Park,
College Park, MD, USA, in 2003. Hu Xia received the Ph.D. degree from the
He is a Tenured Professor with the Department University of Electronic Science and Technology of
of Computer and Information Sciences, Temple China (UESTC), Chengdu, China, in 2012.
University, Philadelphia, PA, USA. His research He was a Visiting Scholar with the University of
interests are wireless communications, wireless Minnesota, Minneapolis, MN, USA, from 2010 to
networks, security, and systems. He has over 200 2011. He is currently an Associate Research Fellow
journal and conference papers in the above areas, as well as a book published with UESTC.
by Springer.
Dr. Du has been awarded over $5 million in research grants. He is a Life
Member of ACM.

Authorized licensed use limited to: Jatiya Kabi Kazi Nazrul Islam University. Downloaded on February 20,2024 at 12:39:22 UTC from IEEE Xplore. Restrictions apply.