ArticlePDF Available

Abstract and Figures

Vehicular Ad hoc NETworks (VANET) are becoming popular due to the emergence of the Internet of Things and ambient intelligence applications. In such networks, secure resource sharing functionality is accomplished by incorporating trust schemes. Current solutions adopt peer-to-peer technologies that can cover the large operational area. However, these systems fail to capture some inherent properties of VANETs, such as fast and ephemeral interaction, making robust trust evaluation of crowdsourcing challenging. In this article, we propose MobileTrust – a hybrid trust-based system for secure resource sharing in VANETs. The proposal is a breakthrough in centralized trust computing that utilizes cloud and upcoming 5G technologies in order to provide robust trust establishment with global scalability. The ad hoc communication is energy-efficient and protects the system against threats that are not countered by the current settings. To evaluate its performance and effectiveness, MobileTrust is modelled in the SUMO simulator and tested on the traffic features of the small-size German city of Eichstatt. Similar schemes are implemented in the same platform in order to provide a fair comparison. Moreover, MobileTrust is deployed on a typical embedded system platform and applied on a real smart car installation for monitoring traffic and road-state parameters of an urban application. The proposed system is developed under the EU-founded THREAT-ARREST project, to provide security, privacy, and trust in an intelligent and energy-aware transportation scenario, bringing closer the vision of sustainable circular economy.
Content may be subject to copyright.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
MobileTrust: Secure Knowledge Integration in VANETs
GEORGE HATZIVASILIS,
FORTH and Hellenic Mediterranean University
OTHONAS SOULTATOS,
FORTH and City University of London
SOTIRIS IOANNIDIS,
FORTH
GEORGE SPANOUDAKIS,
City University of London
VASILIOS KATOS,
Bournemouth University
GIORGOS DEMETRIOU,
Ecole des Ponts Business School
Vehicular Ad hoc NETworks (VANET) are becoming popular due to the emergence of the Internet of
Things and ambient intelligence applications. In such networks, secure resource sharing functionality is
accomplished by incorporating trust schemes. Current solutions adopt peer-to-peer technologies that can
cover the large operational area. However, these systems fail to capture some inherent properties of
VANETs, such as fast and ephemeral interaction, making robust trust evaluation of crowdsourcing
challenging. In this article, we propose MobileTrust – a hybrid trust-based system for secure resource
sharing in VANETs. The proposal is a breakthrough in centralized trust computing that utilizes cloud and
upcoming 5G technologies in order to provide robust trust establishment with global scalability. The ad hoc
communication is energy-efficient and protects the system against threats that are not countered by the
current settings. To evaluate its performance and effectiveness, MobileTrust is modelled in the SUMO
simulator and tested on the traffic features of the small-size German city of Eichstatt. Similar schemes are
implemented in the same platform in order to provide a fair comparison. Moreover, MobileTrust is deployed
on a typical embedded system platform and applied on a real smart car installation for monitoring traffic and
road-state parameters of an urban application. The proposed system is developed under the EU-founded
THREAT-ARREST project, to provide security, privacy, and trust in an intelligent and energy-aware
transportation scenario, bringing closer the vision of sustainable circular economy.
CCS Concepts:• Security and privacy Trust frameworks Networks Network security Networks
Network privacy and anonymity • Networks Network mobility • Networks Cyber-physical
networks • Networks Cloud computing
KEYWORDS
VANET, IoT, CPS, Trust, Reputation, Privacy, Mobility, Cloud, MEC, Parallel computing, GPU, Circular
Economy
ACM Reference format:
George Hatzivasilis, Othonas Soultatos, Sotiris Ioannidis, George Spanoudakis, Vasilis Katos and Giorgos
Demetriou. 2018. MobileTrust: Secure Knowledge Integration in VANETs. ACM Trans. Cyber-Phys. Syst.
4, XXXX. 2 (October 2017), 24 pages.
https://doi.org/124564
1 INTRODUCTION
In this era of the 4
th
Industrial revolution, Internet-of-Things (IoT) applications influence our daily
activities, with the cyber-physical systems (CPS) market value reaching the 8,120 million dollars
in 2021 [1]. Vehicular Ad hoc NETworks (VANETs) constitute a special type of mobile
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee
provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the
full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses,
contact the owner/author(s).
© 2017 Copyright held by the owner/author(s). 0730-0301...$15.00
https://doi.org/124564
39
39:2 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
networking in this domain [2], [3], [4], [5]. They include vehicle-to-vehicle (V2V) and vehicle-to-
roadside infrastructure (V2I) communication for providing, among others, traffic information,
navigation, safety, or other services [6], [7].
In the recent year, several circular economy (CE) initiatives are emerging, penetrating also in
the VANET domain. Quality-of-Service (QoS) and energy-aware solutions that serve numerous
users and devices, enhance the design perspectives towards CE, creating looping assets and
maximizing the utilization of resources, like ridesharing [8]. Other interesting proposals for
intelligent transportation include smart traffic control, ride service hailing, or even car-sharing.
The USA standard IEEE 1609 Wireless Access in Vehicular Environments (WAVE) is built on
IEEE 802.11p [9], [10]. It operates in 5.9 GHz frequency and supports multi-channel
communication, security, and lightweight application layer protocols. The relevant European
standard ETSI ITS G5 [11] operates in the same band, implementing multi-radio multi-channel
functionality, security, and a complex hierarchy of higher layer protocols that integrate a broad
range of main services.
This paper focuses on the security aspects of VANET ecosystems. The aforementioned
standards cover coarse and generic security serviceswhere cryptographic primitives are utilized to
accomplish the properties of authentication, authorization, confidentiality, and integrity.However,
the VANET problem domain constitutes a field of security challenges arising from the violation
of the fair use by selfish or malicious participants [12]. Selfish entities may attemptto avoid the
additional computational or communication effort from contributing to the collective operations
of the social network. Malicious entities exploit the system’s vulnerabilities and rating
mechanisms to launch attacks, like Denial of Service (DoS), Byzantine failures, or interception of
normal interaction by injecting fake information.
As such, trust-based computing approach must be adopted [13], [14]. The underlying
techniques aim toevaluatethe participants’ activity in order to detect security policy violation
attempts and in which case countermeasures would beenforced that restrict or banthe involving
entities. The goal is to encourage the legitimate nodes in keeping up withtheir positive
contribution, discourage and penalise the selfish and malicious participants from misbehaving,
and protect the wholeinfrastructurefromattacks.
This paper presents a trust-based resource sharing system for VANETS, called MobileTrust.
The system is built upon the aforementioned main security functionality (i.e. IEEE 802.11p[10],
ETSI ITS G5[11]) and its main contribution is to evaluate the entities’ behavioronce the
cryptographically secure communication is achieved. As a case-study, we considered the scenario
where vehicles and smart city infrastructure components exchange information regarding traffic
congestion and other road events. A simulation study is conducting in the Simulation of Urban
MObility (SUMO)
1
for the evaluation of the main protection and performance properties of our
proposal, and the comparison with relevant systems under the same setting. A real
implementation is alsodeployed on embedded devices and Android smart phones that emulate the
sensors and the communication equipment of a smart vehicle.
The remainder of the paper is organized as follows. Section 2 discusses the related workin the
domain. Section 3 presents MobileTrust. Section 4 describes the theoretic analysis. Sections 5
shows the simulation analysis and the comparison of our proposal with the current solutions.
Section 6 details the real implementation of MobileTrust on embedded devices. Finally, Section 7
concludes and refers future work.
2 RELATED WORK
1
SUMO: http://www.dlr.de/ts/en/desktopdefault.aspx/tabid-9883/16931_read-41000/
MobileTrust: Secure Knowledge Integration in VANETs 39:3
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
In the context of VANETs, trust-based interactions occur at high speeds and short time periods
[13]. The knowledge integration algorithm should exhibit low complexity constraints, such as
effective throughput, fast memory access, and cost-efficient processing. Scalability is a main
concern, as incidents or road events could be reported either by single mobile nodes or by
multiple vehicles.
General attacks on mobile ad hoc networks are surveyed in [15] and [16]. Furthermore,
VANETs are vulnerable to false positioning, message modification, false message sending, and
DoS [14]. Trust-based schemes are deployed as a main countermeasure against most of these
threats. Inevitably, the formal trust evaluation processes are targeted by more sophisticated
attackers [17]. The involved challenges of trust management in VANETs are reviewed in [18] and
[19].
2.1 Trust-Based Resource Management
Several trust-based schemes have been proposed in the literature attempting to tackle resource
management. For this study, we focus on these systems that collect and process information in
real-time for a VANET setting.
The Vehicle Ad hoc network Reputation System (VARS) [20] is one of the first attempts to
incorporate trust-based computing in the field of VANETs. It forms a modular and peer-to-peer
(P2P) reputation system that separates direct and indirect information when integrating
knowledge. The system rests its confidence on the distributed content itself rather on the node
behavior. The forwarding nodes attach in the transmitted message their opinion about the content
(e.g. traffic congestion on a road segment or blocked route). The receiver utilizes the
contextualized data to evaluate the trustworthiness of the message. A node’s opinion about an
event can be formed by direct knowledge, indirect knowledge sent by a known sender, partially
by the attached opinions on the message by the forwarding nodes, or by a combination of all these
interactions. A main performance drawback of VARS is the fact that the communication
bandwidth is substantially affected due to the overheads from the information that is cumulatively
appended along the communication path when a high number of forwarding nodes is involved.
Furthermore, the proposed approach does not enforce authentication, thus making the system
vulnerable to attacks from simple modification to complete deletion of messages.
The Event-based Reputation System (ERS) [21] attempts to detect inaccurate traffic events and
deter attackers from spreading false messages. Traffic information is collected from the vehicles
and the road infrastructure. The system utilizes distributed observations and tries to figure out if a
traffic event really exits and how long it lasts. The information of an event is maintained in a table
by every vehicle that becomes aware of the incident, either by direct interaction or indirect
notifications. The event table stores data regarding the event’s identity, type, occurrence, location,
message transmission range, reputation, and indirect confidence list. One main scalability
problem that arises is the difficulty in managing the confidence lists for all events that are
reported in the network. Also, the simple evaluation process of indirect notifications makes the
system vulnerable to attackers that perform badmouthing (advertise bad recommendations for
legitimate nodes) and ballot-stuffing (gain good reports from colluding malicious entities) [17].
Moreover, as ERS requires a significant amount of contributions in order to designate that an
event has occurred, it fails when it comes to the early and timely notification of drivers regarding
the existence of an accident on the road ahead. However, the safety of drivers is an essential goal
towards the real deployment of VANETs ([18], [19]).
The Trust Based Security Enhancement (TBSE) for VANETs [22] is designed upon the core
cryptographic techniques and acts as a second defense mechanism against inside attackers. Both
direct and indirect knowledge are utilized for the trust estimations. Direct interaction is evaluated
based on the Bayesian rule. For the rating of the third-party recommendations, the Dempster-
Shafer Theory (DST) [23] is employed, allowing the probabilistic assessment of the truthfulness
39:4 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
of an event in the presence of uncertainty. Then, the calculated trust values are stored in a
repository where upper layer applications can gain access and utilize this information in order to
achieve their security goals, like secure resource sharing or routing. However, the attacker model
is restricted to a low number of malicious entities that do not collude with each other. Passive
attacks are also not studied. Moreover, TBSE is vulnerable to information cascading and
oversampling [24]. In cascading, a decision is madeby the nodes deciding sequentially after
observing the behavior of the other participants. Therefore, a node’s choice is highly influenced
by the previous decisions that can overrule its own observation. Oversampling occurs when a
node receives the opinion of nodes i and j, but j’s opinion has been initially influenced by the
decision of i. Thus, the contributing information by j is oversampled.
A Social Network approach to Trust Management (SNTM) in VANETs is presented in [24].
The main contribution of this study is the limitation of the bad effects from information cascading
and oversampling in P2P communication. SNTM tackles directly cascading by assigning higher
voting weight to the vehicles that are closer to the event. As the influence from cascading is
eliminated and the result becomes more accurate, the potential oversampling is also limited
indirectly. The nodes also attach their opinion regarding an event in the message. The receivers
contrast this knowledge with their own opinion. Then, they rate the other contributors and
estimate the message’s quality.
2.2 Discussion
Centralized trust evaluation was not considered an appropriate architecture for VANETs by the
research community due to scalability issues attributed to the high computational and
communicational burden on a central entity, as well as dependability concerns as this central
entity would be a single point of failure [24]. The design of P2P architectures seemed as a one-
way approach towards the establishment of mobile trust. All the aforementioned related works
promote the P2P paradigm. However, in order to build effective and trustworthy P2P
relationships, one would need a significant and frequent amount of interactions between the peers.
Thus, these studies fail to capture some inherent properties of VANETs. The ephemeral nature of
such networks does not veritably facilitate the P2P trust. Moreover, trust is not necessarily
transitive and the gathering of information from past interaction, by each peer for every other
participant, is computationally expensive and even impossible under realistic assumptions. As the
vehicles drive in high speeds, when the selection and processing of the proper knowledge for an
event are completed, the outcome is no longer required, as the vehicle has either faced, dealt with
or bypassed the reported problem. P2P trust will inevitably become vulnerable to information
cascading and oversampling in an actual setting.
Authentication and cryptographic security is another important issue. A number of studies (e.g.
[20], [21]) ignored or overlooked the implications of cryptographic communication and the
overheads from supporting this essential functionality in P2P networks. On the other hand, when
peer authentication is implemented, privacy issues arise as a vehicle publishes data regarding its
driving route or other user sensitive information [25].
The evolution of cloud computing and the upcoming 5G technologies, like Multi-access Edge
Computing (MEC) [26], [27] can help towards overcoming the restrictions of both centralized and
P2P trust in the VANETs domain. This study proposes a hybrid trust algorithm where both types
of interaction are utilized. Theoretical results regarding the scalability of hybrid P2P systems for
data distribution are reported in [28]. The core trust calculation and maintenance is performed in a
centralized manner, focusing equally on the data correctness and the evaluation of the nodes’
behavior. Nevertheless, the obstacles of centralized knowledge integration are overcome by
distributing and parallelizing the related computations in the cloud and MEC. The internal
computations are further parallelized in Graphics Processing Units (GPUs), enhancing the overall
performance.
MobileTrust: Secure Knowledge Integration in VANETs 39:5
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
Authenticated users contribute with first-hand observations. Only the central authority
integrates this knowledge and reports the existing road events, eliminating the bad effects of
cascading and oversampling. In order to decrease the centralized communication, the peers can
broadcast this timely- and spatial-sensitive data that are digitally signed by the authority, without
requiring to authenticate each other, and thus, preserving their privacy between the vehicle nodes.
Fig. 1 illustrates the proposed setting. In short, the proposed MobileTrust system aims to offer:
Efficient centralized knowledge integration, exploiting the capabilities of cloud
computing, MEC, and GPUs
Effective trust computation of authenticated users in a centralized manner, capturing the
inheriting properties of VANETs (e.g. ephemeral environment, time-sensitive
information, different sensing capabilities, and safety response)
High accuracy in estimating the presence of a road event, even in the presence of
malicious entities
Preservation of P2P privacy
Secure cryptographic communication in all transactions
DoS resilience
The overall setting provides efficient, effective, and timely accurate information, countering a
high variety of attacks and offering secure and privacy-preserving communication.
Fig. 1. The MobileTrust design approach.
3 MobileTrust
This section details the MobileTrust design. The system integrates first-hand information from the
involving participants. The trustworthiness of each contributing node is taken into consideration
during this process. The contributors are authenticated while privacy is preserved in the P2P
interaction. The overall setting presupposes that there exists a safe procedure to register/identify
39:6 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
the participating vehicles in order to prevent the attackers from introducing virtual vehicles that
transmit bogus data. The potential towards global scalability is also presented. The assignment of
the constant coefficients for the reputation and trust computations (e.g. the reputation rating
values or the trust integrating weights) is based on relevant studies ([20], [21], [22], [24]) and in a
previous analysis that the authors have conducted under a similar setting [37].
3.1 Knowledge Integration
Three reputation resources are considered by direct interaction, notices send by trustworthy public
authorities, and notifications made by a centralized entity. The following event types (E) are
modelled: i) Road accidents, ii) Traffic congestion, iii) Bad road conditions, iv) Working
activities, v) Protest, ceremony, military, religious, athletic, or other march, vi) Route diversion by
police, vii) Blocked road (i.e. by physical obstacle), viii) General alarm to draw the drivers’
attention and be cautious.
When a major event is reported, the nodes that are located closer and reckoned sooner can
contribute with data having a higher level of accuracy. These advantageous geo location reporters
gain higher reputation values, if the transmitted data is confirmed.
Based on the evaluation result, a message can be rejected, accepted but not integrated, or
accepted and integrated. Messages originating from untrusted nodes or those who report events
that are not close to the nodes’ location are rejected and the involving nodes are rated negatively.
For the traffic report application, the vehicles must collect information regarding their speed.
The road routes are segmented, for example every 500m. A vehicle calculates its average speed in
each crossed segment. If the speed falls under a threshold, the vehicle indicates the segment as
congested. When a decent amount of similar indications are received by the backend component
(see the text below), the application reports the traffic congestion. If there are no incoming
indications, the congestion report for a segment fades as time elapses, until it is assigned as
normal again. MobileTrust bounds the misbehavior of a malicious vehicle from arbitrarily
reporting segments as congested. The attacker will be rated negatively as long as his indications
will not be supported by other participants.
The events of working activities, protest or other marches, route diversion by the police,
blocked road segment, or general alarm are manually designated by the driver. Moreover, the
backend application can be informed in advance for such events by the involving authorities,
which are considered trustworthy. For example, the municipality can report the scheduled
working activities or athletic events, which are then imported into the system.
The centralized trust management component performs the following steps:
1. The first notification of an event on a road section rs creates a new event entry with
neutral confidence (the value depends on the node’s overall trustworthiness and its
reputation in observing the specific event type).
2. As other nodes located in the same road section advocate that the event is ongoing (by
sending the same notification type as in step 1), the confidence that the event is actually
happening increases.
3. The confidence C
rs,e
is calculated as the ratio of the contributing nodes (contr
rs,e
) to the
nodes that enter the road section (passing
rs
). It is defined in (1) as:








!

"
#

#
$
(1)
4. Incoming vehicles are notified about the active events in advance, before entering the
road section. The reported event’s confidence fades as time elapses and no new
MobileTrust: Secure Knowledge Integration in VANETs 39:7
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
notifications are made, to designate that the event is smoothing out (i.e. 10% decrement
for every minute that passes while the incoming vehicles do not notice the event).
5. The event entry is terminated after a fixed-time period (i.e. 5 minutes) to prevent the case
where an attacker sends the same notification over-and-over in order to keep the entry
active.
6. Then, the nodes that cross the road section are evaluated. The nodes gain a reputation
rate according to the events maximum level of confidence that was achieved during the
active period.
a. The contributing nodes gain negative (-4), neutral (0), and positive (+1)
reputation values r for low, moderate, and high level of confidence,
respectively. In the last case, the first responders (2% of the total contributors)
are awarded with an extra bonus score (+1).
b. Similarly, the non-contributing nodes are rated negatively (-4) if the event
gained moderate or high confidence (based on equation (1)) in order to detect
selfish nodes, otherwise they are ranked neutrally (0).
c. Also, the system ranks negatively (-4) the nodes that remain constantly in a
road segment and report events (i.e. traffic jam) while the majority of the
entering nodes continue their route, in an attempt to force the malicious nodes
to be in continuous move and be unable to target distinct road segments.
7. If notifications are sent after the active period, a new event entry is created (step 1).
3.2 The Trust Framework
For every user, MobileTrust evaluates his/hers trustworthiness based on the past interaction. All
new users start with neutral trust and reputation values (0), which are continuously updated as
new pieces of knowledge are integrated.
MobileTrust maintains a small history of the latest evaluation results for each event type. The
history size (h
e
) is determined by the event’s frequency, ranging from 500 ranks for traffic
monitoring to 30 ranks for the user defined reports (e.g. blocked road segments or protest march).
Reputation fading is applied for each event type, based on a beta distribution [29], by assigning
higher weights (fading_w
i
) to the most recent interactions. The reputation value of the user for a
specific event type, R
(User, e)
, is the average weighted summation of the relevant historical values,
as defined in (2):
%
&
'
'
'
(
)
*
+
,!
-
*
.
/
*
0
1
#
2
(2)
This information determines the effectiveness of a user in reporting a specific type of event.
Thus, the system can discriminate the different sensing capabilities of the various nodes and
exploit this knowledge (if the reputation value is for example >0.5), especially when an event has
been just reported or in road segments with few contributors. Moreover, the system can detect a
malfunction in the sensing equipment or a malicious activity that tries to manipulate the rating
mechanism for a specific event type (i.e. when the reputation value decreases beyond a threshold
of 0.4) and not integrate the relevant contributions.
Trust estimates the overall cooperativeness of a user. Each event reputation gains a weight (w
e
)
based on the event’s frequency and severity. Traffic monitoring and road accident reports are
assigned with a weight of 0.30 and the rest types are assigned with a weight of 0.1 (the weights
summation is equal to 1.0). The trustworthiness for a user, T
User
, is then calculated by aggregating
the underlying reputation values, as defined in (3):
39:8 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
3
&
)
%
&
'
'
'
(
+
456789:66;<
0
=
0
1
(3)
A user with a trust level that exceeds 0.8 or 0.6 is considered trusted and legitimate,
respectively. When the trust falls beyond a threshold (i.e. 0.5), the user is notified for his/hers
misbehavior and is categorized as suspicious. If the cooperation level keeps degrading (i.e. 0.4),
the node is finally reported as malicious and is expelled from the network.
3.3 Authentication and Privacy
In order to maintain the reputation records and operate in a trustworthy manner, the users that
contribute information to MobileTrust must be authenticated. Each driver creates an account with
a username/password pair (or other mean of authentication, like near-field communication (NFC),
Bluetooth pairing, etc.) and logs on the system through the vehicle’s infotainment system
(SSL/TLS communication). Only authenticated users can send information to the application
through a secure channel. The data are encrypted with a symmetric session key (such as AES) and
through an appropriate protocol such as SSL/TLS.
Furthermore, passive anonymous use is also supported. The notifications that are reported by
MobileTrust are signed with the authority’s private asymmetric key (RSA). To reserve resources,
all active notifications for a road segment are included in the signed message. Thus, every node
can verify the central authority, as its public asymmetric key is known to all participants.
However, it should be stated that a high volume of anonymous users who do not contribute with
any information, could degrade the effectiveness of the system in the presence of attackers (as
detailed in the theoretic analysis subsection 4.2).
A main contribution of the overall setting is the privacy protection of the V2V interaction. In
order to preserve privacy, the nodes exchange only the publicly-signed notifications. Thus, if a
vehicle requests information regarding some road segments, the other vehicles or the road
infrastructure will respond with the relevant messages that have been previously retrieved by the
central authority, without requiring to authenticate each other. Both the request and the response
are broadcasted messages in a P2P communication of the nearby nodes.
The notification format, as described in (4), is consisted of the road segment’s ID, the event
type, the date (timestamp) where the event occurred, and a sequence number (bounded to a
specific segment ID and event type, and initialized every day when the date is changed). This
information is both included in the encrypted data and also sent in plaintext along with the rest of
the message. To preserve resources, the nodes maintain only the latest notifications of each event
type for a road segment. Moreover, a node deletes the older messages of passed segments as time-
elapses and the events’ active period terminates (i.e. after 5 minutes) in order to keep the memory
usage profile low.
>?@ABACD@A?E
'
F?GHD@
I
J?DK
'
LMNHME@
'
OP
'
Q
'
'
@TUM
'
Q
'
PD@M
'
Q
'
VMWXMECM
'
E?
Y
(4)
The plaintext notification identifier is utilized as an automatic access control tag to degrade the
effectiveness of DoS attacks at the responder and requester ends. A request message is
unencrypted and contains road segment IDs. The responder examines the included road segment
IDs and broadcasts the latest related notifications. If irrelevant segments are included, the node
discards the message and does not proceed in any further actions. The node also adheres to a
threshold for a maximum number of request responds in the current time window (i.e. at most ten
responds per minute) as an additional measure against DoS by bogus request messages.
MobileTrust: Secure Knowledge Integration in VANETs 39:9
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
A response message replays the public notifications, as they were retrieved by the central
authority. For each incoming response, the requester examines the plaintext notification identifier.
The node automatically recognizes the relevant notifications by examining the road segment ID
and the timestamp. Then, it checks the sequence number and evaluates the newest notification
version. The node decrypts the message with the authority’s public key. If the decryption and the
integrity check (SHA512) are successful, the node integrates the notification. Otherwise, the node
proceeds to the older received versions of the notification. However, an attacker can still send
erroneous messages with a valid notification identifier and a high sequence number in order to
enforce the requester to perform the decryption operation. Thus, if several decryption attempts fail
in a time interval (e.g. five fails in the last minute), the node turns in back-off mode for a while,
until the vehicle drives away from the region. The user can later login the system and report the
road segments where the malicious activity was observed.
3.4 Vehicle Registration
Except from authenticating the active users through the proposed application, we must also
enforce a procedure to examine that he/she is driving a real vehicle. Otherwise, a malicious entity
could create several user accounts for virtual vehicles, which would later launch coordinated
sophisticated attacks on-line by sending fake positioning information along with erroneous road
events. This is a general problem that affects all VANET settings (e.g. [20], [21], [22], [24]). An
indicative case is the money laundering scam in Uber
2
. Collaborating drivers log in the system
through a virtual machine (VM) which emulates a fake-GPS service. Users/payers find these
taxicabs via specific marketplaces in the dark-web and hire services that they do not actually use,
laundering money in the process.
Identification and authentication of a vehicle is out of scope for this study and MobileTrust
presupposes that the smart vehicle industry will deploy such an identification mechanism.
Nonetheless, a basic protection approach is sketched. Each manufacturer nowadays assigns a
unique identifier for each produced vehicle, known as the Vehicle Identification Number (VIN)
3
.
VIN is composed of 17 digits and acts as the vehicle’s fingerprint. For the modern smart
functionality, the manufacturer (or another trusted party) can sign and provide the vehicle owner
with a digital certificate. The certificate will contain VIN and other vehicle-related information.
Then, MobileTrust or other smart applications can obtain this data through the user’s equipment
and check the existence of the mobile node. The application can also interact with the signing
entity in order to evaluate the certification’s validity (e.g. if it has been revoked).Except from
VIN, if the vehicle is equipped with a modern Electronic License Plate (ELP)
4
, the related data
could be also utilized for the same purpose.
The overall process can be further enhanced with the analysis that is performed by popular
Internet services and social media (i.e. Gmail and Facebook).For security purposes, they can
record the set of devices with which the user usually logs onto the system and try to keep out
intruders that have disclosed the user’s credentials.
3.5 Towards Global Scalability
Centralized architectures were considered, in general, disadvantageous in comparison to P2P
settings, as they constitute a single point of processing and a bottleneck [13], [24]. However,
cloud and fog computing is now providing several solutions regarding efficiency and
dependability [30].
2
Uber laundering money scam: https://bdtechtalks.com/2018/03/21/fraudsters-uber-money-laundering/
3
Vehicle Identification Number (VIN): https://en.wikipedia.org/wiki/Vehicle_identification_number
4
Electronic License Plate (ELP): https://en.wikipedia.org/wiki/Electronic_license_plate
39:10 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
The component of MobileTrust that performs all the computational intensive operations is the
central authority. Nevertheless, VANETs and the proposed trust algorithm exhibit some inherited
design properties that enable distribution, parallelization and global scalability. First of all, each
authenticated user/driver is located in a specific region at each time point, and, usually, drives
around a particular geographic location most of the time. Secondly, the processed data are region
dependent. Thus, the knowledge integration of each road segment can be parallelized in a fine-
grained manner.
The IT industry that provides cloud services installs several datacenters in each continent
around the globe. The application traffic is routed to the nearest datacenter based on the IP
address. Load balancing among the datacenters is delivered by the cloud provider [31]. Thus, the
traffic from a country is transmitted to specific datacenters. The user data can be also maintained
in an efficient and distributed manner in the cloud, similarly to an email service, a social network,
or other global-scale applications [32].
The MobileTrust data gathering and integration algorithm can be further parallelized based on
the road segments. Each hired parallel processing unit in a datacenter can perform the trust
computations of a single or a group of nearby road segments. The processing units can read data
from a pool and serve the incoming traffic completely independent from each other. Each core
processes data buckets, where each bucket contains the information for a distinct event entry. As
mentioned earlier, the event entries in a road segment remain active for a fixed-time period, for
example 5 minutes. Then, the core updates the trust profiles of the contributing users. GPU-
processing can decrease the overall processing time.
Additionally, MEC technologies [26] can further reduce the communication delay. The parallel
processing can be installed closer to the smart road infrastructure and drastically reduce the
round-trip time (Fig. 1).
4 THEORETIC ANALYSIS
This section details the theoretical analysis of the proposed approach and its effectiveness in
countering the attacker models that are detailed in the simulation study (Section 5). The overall
analysis is based on the ordinary assumption that the mainstream cryptographic primitives (i.e.
TLS, RSA, AES, and SHA512) are implemented and configured properly, and thus, their usage is
considered theoretically secure. At first, we examine the security aspects that are provided by the
deployed communication protocols. Then, we give proofs regarding the capabilities of
MobileTrust in supporting the legitimate functionality in presence of attackers.
4.1 Protocol Analysis
The theoretic security analysis of the communication links between the involved entities and the
MobileTrust component is modelled in the verification tool ProVerif [33] (the code is not
included for brevity). It is a widely-used automatic symbolic protocol verifier that proves the
security properties of the examined protocol, like authentication, secrecy, and adversary
equivalence aspects. The examined protocol is modelled in a process calculus and is automatically
translated in Horn clauses [33]. The tool resolves these clauses and determines if the security
properties hold or not. In case where all properties are validated, ProVerif returns “true”.
Otherwise, it outputs the properties that could not be satisfied. Three communication protocols are
examined:
1. The user/vehicle login and knowledge transfer towards the Central Authority
(CA)
2. Road segment notification requests from a vehicle to the CA
3. Road segment notification requests from a vehicle to another vehicle (P2P)
MobileTrust: Secure
Knowledge Integration in VANETs
ACM Transactions on Cyber-
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October
4.1.1 User login and contribution.
community, the driver has to authenticate him/herself and enter the system. This includes an
ordinary login process where the user inputs a pair of username/
connection via HTTPS [34
]. The CA safely receives the data and verifies the user.
The communication is initiated by the user. The user application and the CA perform the
handshake phase
. The application obtains the CA’s
two entities agree upon a randomly generated symmetric session key.
From this point on, the transmitted data are encrypted with the common key as in the
record phase
. The user is prompted to provide his
the user’s credentials (e.g. username, password’s digest, etc.). The CA decrypts it, checks the
message’s integrity (HMAC-SHA256)
and uniqueness, and contrasts the received information
with the credentials
that are maintained in an internal database. If the process is successful, the
user is authenticated and can upload the evidence for ongoing road events. Then, the CA performs
MobileTrust and evaluates the user’s cooperation. Fig.
Fig. 2.
User’s login and knowledge contribution.
Knowledge Integration in VANETs
39:
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October
2019
In order to actively participate in the MobileTrust
community, the driver has to authenticate him/herself and enter the system. This includes an
ordinary login process where the user inputs a pair of username/
password through a SSL
/TLS
]. The CA safely receives the data and verifies the user.
The communication is initiated by the user. The user application and the CA perform the
TLS
. The application obtains the CA’s
digital certificate (X.509) and validates it. The
two entities agree upon a randomly generated symmetric session key.
From this point on, the transmitted data are encrypted with the common key as in the
TLS
. The user is prompted to provide his
/hers data (e.g. [35])
. The login request contains
the user’s credentials (e.g. username, password’s digest, etc.). The CA decrypts it, checks the
and uniqueness, and contrasts the received information
that are maintained in an internal database. If the process is successful, the
user is authenticated and can upload the evidence for ongoing road events. Then, the CA performs
MobileTrust and evaluates the user’s cooperation. Fig.
2 depicts the communication steps.
User’s login and knowledge contribution.
11
2019
.
In order to actively participate in the MobileTrust
community, the driver has to authenticate him/herself and enter the system. This includes an
/TLS
TLS
digital certificate (X.509) and validates it. The
TLS
. The login request contains
the user’s credentials (e.g. username, password’s digest, etc.). The CA decrypts it, checks the
and uniqueness, and contrasts the received information
that are maintained in an internal database. If the process is successful, the
user is authenticated and can upload the evidence for ongoing road events. Then, the CA performs
39:12
ACM Transactions on Cyber-
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019
The two entities began with the establishment of the secure channel (
Then, they utilized the derived session key and encrypt the rest exchanged messages (
record). ProV
erif evaluates the abovementioned protocol steps
of the sensed data from the user to the CA)
authentication, confidentiality, integrity, and immunity to replay attacks.
4.1.2 Notification requests.
Every user can ask the CA for the active event notification list of
specific neighboring road segments. If the user is logged in, then the communication can be
encrypted by the session key. Otherwise, the user can make unauthenticate
his/her privacy.
The request contains the required road segment ids and is encrypted with the CA’s public key.
The digest of the data is also computed
(SHA512)
verifies the message’s integrity
. The respon
nonce. The data digest is computed and then signed with CA’s private key. The user re
the digest of the received data and decrypts the signature with the CA’s public key. Fig.
illustrates the exchanged messages.
George Hatzivasilis
et al.
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019
.
The two entities began with the establishment of the secure channel (
HTTPS/TLS
handshake
Then, they utilized the derived session key and encrypt the rest exchanged messages (
TLS
erif evaluates the abovementioned protocol steps
(TLS phases and the transmission
of the sensed data from the user to the CA)
and validates that the overall setting offers
authentication, confidentiality, integrity, and immunity to replay attacks.
Every user can ask the CA for the active event notification list of
specific neighboring road segments. If the user is logged in, then the communication can be
encrypted by the session key. Otherwise, the user can make unauthenticate
d requests and retain
The request contains the required road segment ids and is encrypted with the CA’s public key.
(SHA512)
and the message is sent to the CA. The CA
. The respon
se
contains the relevant notification list and a random
nonce. The data digest is computed and then signed with CA’s private key. The user re
-
computes
the digest of the received data and decrypts the signature with the CA’s public key. Fig.
et al.
handshake
).
TLS
(TLS phases and the transmission
and validates that the overall setting offers
Every user can ask the CA for the active event notification list of
specific neighboring road segments. If the user is logged in, then the communication can be
d requests and retain
The request contains the required road segment ids and is encrypted with the CA’s public key.
and the message is sent to the CA. The CA
contains the relevant notification list and a random
computes
the digest of the received data and decrypts the signature with the CA’s public key. Fig.
3
MobileTrust: Secure Knowledge Integration in VANETs 39:13
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
Fig. 3. Notification requests and anonymous use.
ProVerif validates that the communication is safe and provides CA authentication, integrity,
and immunity to replay attacks. As described in the previous sections, full protection is not
required here, as the road event notifications are meant to be publicly available.
4.1.3 P2P interaction. Nevertheless, a vehicle can request information about the neighboring
road segments by the nearby vehicles. The requester sends the segment id list in plaintext along
with the message’s digest. The contributors check the message’s integrity and respond with the
related active notifications that they possess. The requester verifies the received notifications’
legitimacy as in the previous case. The protocol’s phases are similar with the relevant steps that
are depicted in the second part of Fig. 3 for the anonymous communication between the user and
the CA, with the initial request being an unencrypted broadcast message.
As before, the tool verifies that the interaction accomplishes the CA authentication, integrity,
and immunity to replay attacks. Moreover, the contributors’ privacy is retained as no user-
/vehicle-related data is disclosed.
4.2 Attack Mitigation
The theoretical security analysis of the trust computing mechanisms and its effectiveness in
constraining the malicious activity is performed in two steps. First, we prove that the integration
of malicious knowledge is bounded (Theorem 4.1). Then, we argue that MobileTrust’s efficacy is
strongly related with the active participation of the legitimate users (Lemma 4.1 and Theorem
4.2). The proposed defense mechanism increases significantly the effort that must be devoted by
the attacker in order to exploit the system (Lemma 4.2). Finally, we evince that the system can
provide protection even without the full participation of all legitimate users that may want to
retain their privacy (Lemma 4.3).
For the system model, we form a VANET with N nodes, such as V={1, 2, …, N}. It is assumed
that all links are bidirectional. If node i can receive packets that are directly transmitted by node j,
then node j can receive packets that are directly transmitted by node i. The nodes are either
vehicles or smart road infrastructure. Since the majority of these nodes are vehicles, the network
dynamics are characterized by confined mobility, high speeds, and thus, short connection times
between the nodes. The operational area can be very large, but the V2V communications are
mainly local and the currently processed information for each node concerns specific road
segments and not the whole network. Information regarding the examined road events is collected
and exchanged by each node, either with other nodes or with the CA. Vehicles can travel freely,
but in most cases, they have a relatively fixed route, such as buses.
Due to the complexity of a VANET system, several types of attacks can be performed. For the
adversary model, we consider that attackers are internal nodes (malicious or compromised) that
send out fraudulent event messages: they notify unreal events (i.e. traffic jam while the road
segment is not congested) and if an event exists, they do not send any message to the authority
denoting the incident (mislead the trust mechanism into producing false negativesin the case
where a smaller number of legitimate nodes have reported the events). They are global attackers
(have an overall coverage of the network and can eavesdrop every message diffused by any
vehicle), who can be either active or passive. Legitimate nodes can also malfunction with a low
probability and perform the same actions for a while (i.e. Byzantine faults [36]).
We start by providing some definitions and then the theorems and the proofs.
Definition 3.1. Let 
Z
be the total number of correct event notifications that are
reported to the CA for a specific event e in a road segment rs.
39:14 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
Definition 3.2. Let 
[
be the total number of erroneous event notifications that are
reported to the CA for a specific event in a road segment.
Definition 3.3. Let Tcontr
rs,e
be the total number of received event notifications for a specific
event in a road segment, determined as the sum of 
Z
and 
[
.
Definition 3.4. Let p be the success rate of the total received event notifications Tcontr
rs,e
.
4.2.1 The bounded malicious effect.
T
HEOREM
4.1. As the contributing users are authenticated, they can send information for a
specific region at each time point. The ideal network exhibits \
[]^_
`
\
Z]^_
a. For up to an additive constant, ignoring a bounded number φ of
erroneous notifications, it holds that the number of erroneous notifications is a p-fraction of the
number of received notifications. Specifically, there exists an upper bound φ, as described in (5):
)
[


]^
_
`
)
Z


]^
_
a
b
(5)
P
ROOF
.
Suppose that there exist N nodes in the network V, M of which are malicious with
M<N. Let MRS denote the set of road segments RS that contain malicious nodes. The maximum
value for MRS is RS (if there is at least one M in every %c).
Let β denote the number of erroneous notifications that exposes a node as malicious. The
number of convictions conv
rs
is at least \
[d]^_
e
f. Thus,
\
[


d]^
_
e
)
g


d]^
h
(6)
Similarly, the number of rewards rew
rs
is at most
\i5:6
jkl/klmno/p
q <
f
. Thus,
)



d]^
\
Z


d]^
_
e
f
h
(7)
Therefore,
\
[


d]^
_
e
\
Z


d]^
_
e
f
'
a
)
g



(

d]^
(8)
By combining (5) and (8), we derive:
)
[

d]^_
`
)
Z

d]^_
a
e
)
g



(

d]^
a
e
`
r
`
r%c
(9)
Therefore, the malicious activity which affects the network is bounded. If there are no
MobileTrust: Secure Knowledge Integration in VANETs 39:15
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
malicious nodes (M=0), then (9) describes the ideal case of
\

%c
`
\

s%c
a.
4.2.2 Correlation between the legitimate user’s participation and the robust deduction
procedure of MobileTrust.
L
EMMA
4.1. The MobileTrust’s evaluation mechanism can decrease the attack rate.
P
ROOF
.
Consider M
up
and M
mt
as the number of malicious nodes per window for an unprotected
(up) network and a setting that is protected by MobileTrust (mt), respectively. As the
malfunctioning is detected with MobileTrust in place, the malicious nodes (M
mt
) are denoted as
untrustworthy, punished, and eventually expelled from the network based on (3). Thus, it holds
that r
t6
r\u#'3
&
aYv( . Based on (9), it is concluded that the
malfunctioning which is caused by the malicious nodes for an unprotected setting, where M
up
=M,
can be higher as it holds that M
up
M
mt
.
T
HEOREM
4.2. The MobileTrust’s mitigation rate is directly related with the user’s active
participation in the crowdsourcing operations.
P
ROOF
.
Let

wr denote the malicious nodes that exist in a road segment r%c. The
maximum volume of malicious notifications for a specific event entry e in rs is


.
Consider a faulty event e for a rs. The event’s confidence C
rs,e
is given by (10):




'
`
Z

"
(
`
Z




(10)
Based on (10), the malicious activity will be punished, if the erroneous notifications are below
50%, thus if"(`
Z
s

(aY`

. As the malfunction ratio (1-p) of
the legitimate entities will be low, the higher the user participation the higher the difficulty and
the effort for the attacker to avoid punishment.
On the contrary, this also reflects the general limitation of the recommendation systems ([37],
[38]). If the majority of the nodes are compromised (50%), such approaches cannot provide
guarantees against colluding malicious attacks. Networks with so high volume of attackers are in
general discarded ([20], [21], [22], [24], [34]).
4.2.3 Attacker’s effort and user’s participation.
L
EMMA
4.2. MobileTrust increases the effort that is required in order to exploit the
crowdsourcing system.
P
ROOF
.
Based on T
HEOREM
4.2, it is concluded that in order to perform an attack, one would
need to marshal a high portion of colluding malicious user-accounts that cross a r%c. As
MobileTrust forces the colluding nodes to move around in order to avoid detection (step 6.c in
subsection 3.1), the attacker cannot know in advance the passing
rs
nodes from the rs, and thus,
he/she could not know if the malfunctioning will be detected by the CA during the evaluation of
an event entry, or not. Therefore, the attacker could be burdened with a number of failed
notifications. This fact will make the attacker to devote a significant effort in order to ensure that
this single rs will be mistreated, straining the wile resources. On the other hand, the falsely
accused legitimate users can retain their trustworthiness by their rewards from non-faulty
segments.
L
EMMA
4.3. MobileTrust does not require the absolute cooperation from all users in order to
report events with high confidence and effectively mitigate the attackers.
39:16 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
P
ROOF
.
Taking into account privacy issues and the user’s convenience, MobileTrust
preconceives that users will not be always participating actively in the crowdsourcing activities.
From T
HEOREM
4.2 it is derived that if the low volume of periodically malfunctioning legitimate
modes along with the malicious ones do not constitute the majority, there exists a threshold of
non-contributing users (NC-User
thr
) where the system continues deducing the actual events’ status
while retaining its capability in detecting/mitigating the erroneous functionality.
5 SIMULATION STUDY
This section presents the simulation evaluation of MobileTrust and other related trust
schemes(Vehicle Ad hoc network Reputation System (VARS) [20], Event-based Reputation
System (ERS) [21], Trust Based Security Enhancement (TBSE) [22], and Social Network
approach to Trust Management (SNTM) [24]) over a common setting for VANET
communication, under normal operation and two attack scenarios. As the theoretical analysis
results provoke, we test the MobileTrust’s defense effectiveness for different ratios of
contributing/malicious nodes.
5.1 Network Assumptions and Simulation Setup
Regarding the network assumptions, this study concentrates on VANETs and other similar mobile
ad hoc networks. Attacks on resource management are mainly studied. MobileTrust is modelled
in the SUMO and is implemented in C++. As aforementioned, the assignment of the constant
coefficients for the reputation and trust elements is based on relevant studies ([20], [21], [22],
[24]) and in a previous analysis by the authors [37].
Fig. 4 illustrates the simulation test for the map of Eichstatt, a small size city with around
13,000 citizens. The simulated area covers 4.5km×3.0km. For the evaluation of the different
schemes, a VANET of 1500 nodes is modelled. The vehicles’ top speed varies from 18-80 km/h.
There are utilized the propagation model two-ray ground reflection and the IEEE 802.11p [10] as
the MAC layer. Each node has a 2Mbps raw bandwidth with 100m physical radio range. Table 1
summarizes the simulation parameters.
MobileTrust: Secure Knowledge Integration in VANETs 39:17
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
Fig. 4. i) Simulation map of the German city Eichstatt in scale 1:1000m, ii) simulation snapshot of a road
segment in scale 1:10m and iii) its projection on Google Maps.
Table 1. Simulation Parameters of MobileTrust
Parameter Value
Number of Vehicles 1500
Vehicle top speed (20% of vehicles for each speed value) [18, 36, 50, 60, 80] km/h
Frequency 2400 MHz
Path loss model Two ray ground
MAC layer IEEE 802.11p
Antenna height 1.5 m
Transmission power 15 dbm
Transmission range 100 m
Type of antenna 360 degrees
TxGain of antenna 1
RxGain of antenna 1
Fading channel Ricean
Width of road 20 m
The experiments include an initialization and an evaluation phase. At first, nodes start
communicating with the default values of each scheme. Then, the normal operation and two
attack scenarios are evaluated, measuring the overall performance and security, respectively. The
39:18 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
simulated period lasts 5 mins for each phase. Every experiment is performed10 times for each
setting and the average measurements are reported. The randomly located nodes for the first
evaluated system are recorded in order to use the exact same environment in all schemes.
The first case examines the normal operation of each scheme, with no attacks taking place. Fig.
4 depicts a snapshot from the simulation and its projection on Google maps. The different colors
illustrate the maximum speed of the vehicles and the current average speed in each road segment.
In general, as the processed data are spatial, the performance of MobileTrust can be affected by
the density of the mobile nodes. Nevertheless, one of the main contributions of this paper is that
the trust calculation can be distributed. Thus, if the same computational resources are devoted for
every 13,000 citizens, the similar results will be concluded for medium-sized (50-100k residents)
or even bigger (200-500k citizens) cities.
Then, malicious nodes are introduced. For each different attack, 5 experiments are performed
for 10, 20, 30, 40, and 50 percentage of malicious nodes respectively (networks with higher
volume of malicious entities are generally getting discarded in practice).
5.2 Normal Operation
In order to evaluate the trust systems under normal operation, we emulate around 100 traffic
congestion events and 10 car crashes in each experiment. Two metrics estimate the performance
and the energy consumption of each system. M
1-1
measures the number of trust messages that are
sent during normal operation, representing a rough estimation of the involving communication
and computational overheads. M
1-2
calculates the number of messages that are required in order to
identify a road event, revealing the early notification capabilities of each scheme. For both
metrics, the lower the value, the better. Fig. 5 illustrates the relevant metrics.
Fig. 5. Performance evaluation results for the metrics M
1-1
and M
1-2
.
For M
1-1
, the MobileTrust authority receives one message by each involved vehicle. The rest
P2P systems require a much higher number of messages in order to form the trust parameters,
based on the information propagation scope. The SNTM’s closest node strategy has moderate
performance while TBSE and VARS are less efficient. ERS requires the highest amount of
interaction.
For M
1-2
, the MobileTrust publishes instantly a new road event when it is reported by a trusted
or a legitimate user with high event reputation along with the current confidence level. Then, it
takes a decent amount of incoming contributions in order to notify the event with high confidence.
The rest systems take several messages until a node obtains sufficient knowledge regarding the
event’s occurrence. Moreover, as direct interaction gains higher weight when integrating direct
and indirect knowledge, the nodes in the P2P schemes usually have to detect the event by direct
observation before being able to reason about it with high confidence. Among these systems,
MobileTrust: Secure Knowledge Integration in VANETs 39:19
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
SNTM provides the earliest notification as indirect knowledge gives higher weight to the nodes
that are closest to the event. TBSE, VARS, and ERS exhibit similar low performance.
As is evident by the two metrics, MobileTrust is the most efficient system with effective early
notification of road events. Nonetheless, MobileTrust’s performance is normally even better as
public authorities can report several types of scheduled events, helping the drivers to avoid the
relevant segments without requiring any further contribution from them. The actual performance
of MobileTrust is measured in a preliminary implementation on real devices and is detailed in the
next Section.
5.3 Attack Scenarios
In the first attack, the malicious nodes launch a social-based on-off attack. They cooperate at the
initialization phase to gain high trust values, while at the evaluation phase they start injecting false
notifications. We evaluate the effectiveness of the knowledge integration mechanism of each
system in detecting bogus data and producing accurate notifications regarding the road state by
measuring the volume of the false observations that are integrated by the trust system, defined as
M
2-1
, and the number of false notifications that are made by the system, defined as M
2-2
. For both
metrics, the lower the value, the better. Fig. 6 illustrates the evaluation metrics M
2-1
and M
2-2
.
Fig. 6. Security evaluation results for the metrics M
2-1
, M
2-2
, and M
3
.
For M
2-1
, the single and centralized authority of MobileTrust collects a high volume of
legitimate votes and thus is able to recognize the malicious entities and reject their contributions.
As the attackers increase, the system may integrate a small number of bogus observations at the
beginning of the attack phase. Nevertheless, the authority can detect the misbehaving nodes fast,
even for a high volume of attackers (50%). All relevant schemes perform poorly for a large
number of attackers (>40%). TBSE rejects a high volume of bogus observations for a low and
moderate percent of attackers (10%-30%) due to its enhanced trust evaluation process. SNTM,
ERS, and VARS exhibit the worst results in all cases.
For M
2-2
, the central authority of MobileTrust reports only a small number of false notifications
for a high volume of malicious nodes (50%). When many attackers participate in the attack
(>40%) the relevant schemes perform badly, as they integrate a large number of bogus
39:20 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
observations (M
2-1
). TBSE produces moderate results for low to moderate attackers’ volumes
(10%-30%) as it processes less malicious messages. SNTM, ERS, and VARS have the lowest
performance across all attack setups.
In the second scenario, the attackers exploit the ad hoc communication in order to perform
rushing attacks by flooding messages and exhaust the resources of the other nodes. We estimate
the robustness of the ad hoc communication against rushing attacks by calculating the number of
flooding request messages that are processed by the legitimate nodes, defined as M
3
, the lower the
value the better. We consider that all messages contain valid data (not erroneous). Fig. 6also
illustrates the metric M
3
.
In MobileTrust, the nodes can instantly recognize the involved road segments from the
notification format and ignore irrelevant requests. As the nodes exchange only public information,
they can reason when a specific notification has been just sent and they will not retransmit it as
long as they remain inside the same road segment. Moreover, each node imposes a maximum
number of responses in the current time-window. Thus, MobileTrust nodes will process a low
volume of different requests, constraining the effectiveness and propagation of flooding. The P2P
systems serve a high number of flooding requests under many colluding attackers. TBSE can
degrade the flooding attack for a low to moderate volume of malicious participants (10%-20%) as
it detects selfish behavior and ranks negatively the relevant nodes. SNTM and ERS also provide
adequate defense against a small amount of misbehaving nodes (10%). VARS offers little
protection.
5.4 Security and Comparison with Other Systems
As is evident by the above analysis, MobileTrust implements efficient and robust mechanisms for
knowledge integration and ad hoc communication that outperform the current state of the art. The
proposed system offers immunity-by-design against some types of threats and attacks. It is not
vulnerable to information cascading and oversampling as only first-hand observations are
evaluated by a single centralized entity. Moreover, badmouthing and ballot-stuffing attacks are
not applicable as no third-party recommendations about trust are supported. As nodes either
advertise their observations directly to the authority or broadcast the public notifications to the
other cars, the deletion of transmitted messages is not applicable, in contrast to the P2P trust
schemes. Regarding privacy, P2P communication does not disclose any vehicle information,
while restricted anonymous usage is supported for the centralized interaction. Eavesdropping is
also ineffective as the transmitted messages are either encrypted communication between the
authority and the vehicles or publicly available data.
The cryptographic solutions provide the mainstream security properties. Identity-based attacks
(e.g. impersonation, newcomer, HELLO flooding, and Sybil attacks) are countered as the creator
of each message is properly authenticated and evaluated. Session symmetric keys encrypt the
transmitted data between the authority and the contributors, offering confidentiality and
preventing information disclosure. The HMAC primitive enables integrity checks and detects
message modification attempts. Random nonces are also included in the message data to avoid
replay attacks. The system accomplishes non-repudiation regarding the actions of the authority
and the knowledge contributors. Each entity performs a specific set of actions and role-based
authorization is imposed for the central authority, the authenticated contributors, and the passive
anonymous users.
Moreover, automatic access control techniques (i.e. the plaintext notification identifiers) and
back-off mechanisms degrade the effectiveness of DoS attacks. The low
computational/communicational overheads for each task and edge computing further enhances
protection.
Table 2 summarizes the comparison results of MobileTrust and the 4 related schemes.
MobileTrust: Secure
Knowledge Integration in VANETs
ACM Transactions on Cyber-
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October
Table 2.
Comparison of Trust
Feature
MobileTrust
Architecture
Hybrid
Trust scope
Hybrid
Authentication YES
Privacy YES
Early notification YES
Secure Communication YES
6 IMPLEMENTATION
This section presents the implementation details of MobileTrust on a smart city setting with a real
vehicle.
We evaluate the system’s performance under normal operation and measure the
runtime overhead, which was sketched in the simulation study (subsection 5.2 Normal Operation).
6.1 The Car Installation
A real smart car setting is deployed by the authors
infotainment, collects information of the vehicles electronic control units (ECU) via a Bluetooth
enabled OBD scan tool, and communicates to Internet, via a Global System for Mobile
Communications (GSM) conne
ction, with a command
cloud. The research platform o
f GRNET Virtual MAchines (ViMA)
computing. The VMs of the cloud service are monitored through a laptop. The emulated scenario
implements the
European Union’s road safety system eCall [
accidents and automatically informs the involved emergency services. Fig.
setting.
Fig. 7.
In this paper, we further extend this deployment in order to implement MobileTrust and model
the event types that are described in Section 3. The in
application, running in the smart phone, and the
6.2 Testbed
5
GRNET ViMA: https://vima.grnet.gr/about/info/en/
Knowledge Integration in VANETs
39:
21
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October
2019
Comparison of Trust
-Based Schemes for VANETs
MobileTrust
VARS ERS TBSE SNTM
Hybrid
P2P P2P P2P P2P
Hybrid
Data Node Node Data
NO NO NO NO
NO NO NO NO
NO NO NO NO
NO NO NO NO
This section presents the implementation details of MobileTrust on a smart city setting with a real
We evaluate the system’s performance under normal operation and measure the
actual
runtime overhead, which was sketched in the simulation study (subsection 5.2 Normal Operation).
A real smart car setting is deployed by the authors
in [39
]. A smart phone, acting as the vehicles
infotainment, collects information of the vehicles electronic control units (ECU) via a Bluetooth
enabled OBD scan tool, and communicates to Internet, via a Global System for Mobile
ction, with a command
-and-
control (C&C) center, located in the
f GRNET Virtual MAchines (ViMA)
5
is utilized for cloud
computing. The VMs of the cloud service are monitored through a laptop. The emulated scenario
European Union’s road safety system eCall [
40
], where the system detects car
accidents and automatically informs the involved emergency services. Fig.
7
depicts the smart car
Smart car deployment.
In this paper, we further extend this deployment in order to implement MobileTrust and model
the event types that are described in Section 3. The in
-
vehicle component is an Android
application, running in the smart phone, and the
CA is implemented as a web
service in the C&C.
21
2019
.
This section presents the implementation details of MobileTrust on a smart city setting with a real
actual
runtime overhead, which was sketched in the simulation study (subsection 5.2 Normal Operation).
]. A smart phone, acting as the vehicles
infotainment, collects information of the vehicles electronic control units (ECU) via a Bluetooth
-
enabled OBD scan tool, and communicates to Internet, via a Global System for Mobile
control (C&C) center, located in the
is utilized for cloud
computing. The VMs of the cloud service are monitored through a laptop. The emulated scenario
], where the system detects car
depicts the smart car
In this paper, we further extend this deployment in order to implement MobileTrust and model
vehicle component is an Android
service in the C&C.
39:22
ACM Transactions on Cyber-
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019
Under this setting, a testbed was evaluated on a VM (Intel Core i7 at 2.1 GHz CPU, 8GB RAM,
64-bit OS Windows 8.1 Pro), where
1000
the response time of MobileTrust in
ms for the smart phone client application and the central
authority at the backend, with spikes depicting TLS processing. The additional communication
delay through GSM was on average 1-4
internal compon
ents is summarized in Table 3.
cryptographic services and exhibit similar results.
Fig. 8.
Response time of MobileTrust applications in ms.
Table 3.
Resource Consumption of MobileTtrust components
Component
Cryptographic service
TLS handshake
TLS session
RSA-2048 sign
RSA-2048 verify
AES-256 in CBC
SHA512
Trust scheme
Reputation estimation (per node)
Trust computation (per node)
Event evaluation
Event confidence calculation (per 500 nodes in a road segment)
6.3 Parallelization in GPUs
As aforementioned, the core knowledge integration mechanism of MobileTrust can be
parallelized in a fine-grained
manner. Except from the CPU parallelization on multiple core, the
GPUs is a well-
tried practice that can drastically reduce the processing time in computation
intensive operations. Recommendation systems perform well in this setting with
speedup [38
]. In this study, we utilize the CUDA General Purpose GPU (GPGPU) in order to
parallelize the MobileTrust algorithm. The GPU
George Hatzivasilis
et al.
Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019
.
Under this setting, a testbed was evaluated on a VM (Intel Core i7 at 2.1 GHz CPU, 8GB RAM,
1000
event notifications where advertised. Fig. 8
illustrates
ms for the smart phone client application and the central
authority at the backend, with spikes depicting TLS processing. The additional communication
seconds. The processing resource consumption of the
ents is summarized in Table 3.
The V2V interactions also
include the referred
cryptographic services and exhibit similar results.
Response time of MobileTrust applications in ms.
Resource Consumption of MobileTtrust components
RAM [KB] CPU [ms]
1400 692.8
65 30.45
6.1 3.229
3.2 0.117
3.4 0.004
2.4 0.009
0.08 – 1.1 3.4 - 54.48
1.4 0.4
Event confidence calculation (per 500 nodes in a road segment)
1.5 0.14
As aforementioned, the core knowledge integration mechanism of MobileTrust can be
manner. Except from the CPU parallelization on multiple core, the
tried practice that can drastically reduce the processing time in computation
intensive operations. Recommendation systems perform well in this setting with
30-32
times
]. In this study, we utilize the CUDA General Purpose GPU (GPGPU) in order to
parallelize the MobileTrust algorithm. The GPU
-
accelerated deployment is implemented on the
et al.
Under this setting, a testbed was evaluated on a VM (Intel Core i7 at 2.1 GHz CPU, 8GB RAM,
illustrates
ms for the smart phone client application and the central
authority at the backend, with spikes depicting TLS processing. The additional communication
seconds. The processing resource consumption of the
include the referred
As aforementioned, the core knowledge integration mechanism of MobileTrust can be
manner. Except from the CPU parallelization on multiple core, the
tried practice that can drastically reduce the processing time in computation
times
]. In this study, we utilize the CUDA General Purpose GPU (GPGPU) in order to
accelerated deployment is implemented on the
MobileTrust: Secure Knowledge Integration in VANETs 39:23
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
same machine as the CPU version, equipped with the NVIDIA GeForce GTX 1050 GPU (640
cores, 2GB buffer, 6Gbps memory speed, 1.4 GHz clock).
Consider the most frequent and resource demanding traffic congestion events for road
segments with 500 vehicles on average. After the event’s active period, the event’s confidence
must be evaluated (see Table 3). Then, the reputation histories and the trust structures of the 500
users must be updated. Fig. 9 depicts the average processing time of evaluating a road event for
the CPU, CPU-accelerated, and GPU-accelerated versions of MobileTrust. With the parallel
processing on 4 CPU cores (8 threads), each thread processes one event for a specific road
segment and updates the users’ data. The CPU parallelism busts performance by 23% in
comparison with the single CPU version. This CPU-accelerated implementation is further
improved by parallelizing the users’ data update in the GPU (500 cores of 560KB each). This
setting is around 32 times faster than the single CPU one, and can serve 1.16 events per second.
The optimized proposal can be deployed by every VM in both cloud and MEC architectures,
providing an overall efficient and globally scalable system with real-time response.
Fig. 9. Processing time of MobileTrust versions in ms.
4 CONCLUSIONS
In this work MobileTrust – a trust-based resource sharing system for VANETs is presented. The
system is modelled in the SUMO simulator along with 4 relevant schemes and a comparative
analysis is conducted. MobileTrust is an energy efficient solution that provides the higher level of
protection. On real embedded devices in the context of an emulated VANET and smart city CPS,
the overhead of the trust scheme is low and acceptable for the added security and the overall
protection level that it is achieved.
One limitation for MobileTrust and all the relevant systems concerns the accuracy of the
location positioning system [41]. For example, the discrimination of the different lanes in a road
section may not be possible. Thus, if only one lane faces a specific event (i.e. traffic jam), the
notification and rating mechanisms can produce erroneous results. Nevertheless, these issues will
be contained as the Global Positioning System (GPS) technology is evolved and integrated with
5G pioneering frequency communications (such as 26 and 60GHz).
As future extensions of our work, we consider the further inclusion of modern smart
transportation technologies and circular economy business models. The goal is to extend the trust
mechanism in order to evaluate application specific features for ecosystems that provide
operations like ride service hailing and car-sharing.
ACKNOWLEDGMENTS
39:24 George Hatzivasilis et al.
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
This work has received funding from the European Union Horizon’s 2020 research and
innovation programme H2020-DS-SC7-2017, under grant agreements No. 786890 (THREAT-
ARREST) and No. 830927 (CONCORDIA), as well as the Marie Skodowska-Curie Actions
(MSCA) Research and Innovation Staff Exchange (RISE), H2020-MSCA-RISE-2017, under
grant agreements No. 777855 (CE-IoT) and No. 778228 (Ideal-Cities).
REFERENCES
[1]
N. Trent, “Cyber physical system 2017 market segmentation, application, technology & market analysis research
report to 2021,” Wiseguyreports.com, Maharashtra, India, October 2017, [On-line]
http://www.abnewswire.com/pressreleases/cyber-physical-system-2017-market-segmentationapplicationtechnology-
market-analysis-research-report-to-2021_150981.html .
[2]
A. Dahiya and R. K. Chauhan, “A comparative study of MANET and VANET environment,” Journal of Computing,
SCRIBD, vol. 2, issue 7, pp. 87-92, July 2010.
[3]
F. J. Martinez, C. K. Toh, J.-C. Cano, C. T. Calafate, and P. Manzoni, “A survey and comparative study of simulators
for vehicular ad hoc networks (VANETs),” Wireless Communications & Mobile Computing, Wiley, vol. 11, issue 7,
pp. 813-828, July 2011.
[4]
P. Chouhan, G. Kaushal, and U. Prajapath, “Comparative study MANET and VANET,” International Journal of
Engineering and Computer Science, vol. 5, issue 04, pp. 16079-16083, April 2016.
[5]
N. Saxena, S. Grijalva, V. Chukwuka, and A. V. Vasilakos, “Network security and privacy challenges in smart
vehicle-to-grid,” IEEE Wireless Communications, IEEE, vol. 24, issue 4, pp. 88-98, August 2017.
[6]
C. K. Toh, “Future application scenarios for MANET-based intelligent transportation systems,” Future Generation
Communication and Networking (FGCN), IEEE, Jeju, South Korea, pp. 1-4, January 2008.
[7]
F. J. Martinez, C. K. Toh, J.-C. Cano, C. T. Calafate, and P. Manzoni, “Emergency services in future intelligent
transportation systems based on vehicular communication networks,” IEEE Intelligent Transportation Systems
Magazine, IEEE, vol. 2, issue 2, pp. 6-20, Summer 2010.
[8]
B. Cici, A. Markopoulou, and N. Laoutaris, “SORS: a scalable online ridesharing system,” 9
th
International Workshop
on Computational Transportation Science (IWCTS), ACM SIGSPATIAL, Burlingame, California, pp. 13-18, October
2016.
[9]
S. Grafling, P. Mahonen, and J. Riihijarvi, “Performance evaluation of IEEE 1609 WAVE and IEEE 802.11p for
vehicular communications,” 2
nd
International Conference on Ubiquitous and Future Networks (ICUFN), IEEE, Jeju
Island, South Korea, pp. 344-348, August 2010.
[10]
Y. Wang, A. Ahmed, B. Krishnamachari, and K. Psounis, “IEEE 802.11p performance evaluation and protocol
enhancement,” IEEE International Conference on Vehicular Electronics and Safety (ICVES), IEEE, Columbus, OH,
USA, pp. 317-322, September 2008.
[11]
D. Eckhoff, N. Sofra, and R. German, “A performance study if cooperative awareness in ETSI ITS G5 and IEEE
WAVE,” 10
th
Annual Conference on Wireless On-demand Network Systems and Services (WONS), IEEE, Banff, AB,
Canada, pp. 196-200, August 2013.
[12]
X. Gong and N. B. Shroff, “Trustful mobile crowdsensing for strategic users with private qualities,” 15
th
International
Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt), IEEE, Paris, France,
May 2017.
[13]
M. Mejia and R. haparro-Vargas, “Distributed trust and reputation mechanisms for vehicular ad-hoc networks,”
Vehicular Technologies – Deployment and Applications, INTECH, Chapter 6, pp. 117-140, 2013.
[14]
N. J. Patel and R. H. Jhaveri, “Trust based approaches for secure routing in VANET: a survey,” International
Conference on Advanced Computing Technologies and Applications (ICACTA), Elsevier, Procedia Computer
Science, vol. 45, pp. 592-601, March 2015.
[15]
M. A. Ngadi, R. H. Khokhar, and S. Mandala, “A review of current routing attacks in mobile ad-hoc networks,”
International Journal of Computer Science and Security, vol. 2, issue 3, pp. 18-29, 2008.
[16]
D. Airehrour, J. Gutierrez, and S. K. Ray, “Secure routing for Internet of things,” Journal of Network and Computer
Applications, Elsevier, vol. 66, pp.198-213, May 2016.
[17]
E. Carrara and G. Hogden, “ENISA: reputation-based system: a security analysis,” European Network and
Information Security Agency (ENISA) Position Papre, no. 2, October 2007.
[18]
J. Zhang, “Trust management for VANETs: challenges, desired properties and future directions,” International Journal
of Distributed Systems and Technologies, vol. 1, issue 1, pp. 48-62, January-March 2012.
[19]
S. A. Soleymani, A. H. Abdullah, W. H. Hassan, M. H. Anisi, S. Goudarzi, M. A. R. Baee, and S. Mandala, “Trust
management in vehicular ad hoc network: a systematic review,” EURASIP Journal on Wireless Communications and
Networking, Springer, vol. 2015, article 146, pp. 1-22, December 2015.
MobileTrust: Secure Knowledge Integration in VANETs 39:25
ACM Transactions on Cyber-Physical Systems, Vol. 4, No. 2, Article 39. Publication date: October 2019.
[20]
F. Dotzer, L. Fischer, and P. Magiera, “VARS: a Vehicle Ad-hoc network Reputation System,” 6
th
IEEE International
Symposium on a World of Wireless Mobile and Multimedia Networks (WoWMoM), IEEE, Taormina-Giardini
Naxos, Italy, pp. 454-456, June 2005.
[21]
N.-W. Lo and H.-C. Tsai, “A reputation system for traffic safety event on vehicular ad hoc networks,” EURASIP
Journal on Wireless Communications and Networking – Special Issue on enabling Wireless Technologies for Green
Pervasive Computing, Springer, vol. 2009, no. 9, article ID 125348, pp. 1-10, February 2009.
[22]
Z. Wei, F. R. Yu, and A. Boukerche, “Trust based security enhancements for vehicular ad hoc networks,” 4
th
ACM
International Symposium on Development and Analysis of Intelligent Vehicular Networks and Applications
(DIVANet), ACM, Montreal, QC, Canada, pp. 103-109, September 2014.
[23]
G. Shafer and J. Pearl, “Readings in Uncertain Reasoning,” Morgan Kaufman Publishers, pp. 1-768, July 1990.
[24]
Z. Huang, S. Ruj, M. A. Cavenaghu, M. Stojmenovic, and A. Nayak, “A social network approach to trust
management,” Peer-to-Peer Networking and Applications, Springer, vol.7, issue 3, pp. 229-242, September 2014.
[25]
Q. Kong, R. Lu, M. Ma, and H. Bao, “A privacy-preserving sensory data sharing scheme in Internet of Vehicles,”
Future Generation Computer Systems, Elsevier, vol. 92, pp. 644-655, March 2019.
[26]
A. Machen, S. Wang, K. K. Leung, B. J. Ko, and T. Salonidis, “Migrating running applications across mobile edge
clouds: poster,” 22
nd
Annual International Conference on Mobile Computing and Networking (MobiCom), ACM, New
York, USA, pp. 435-436, October 2016.
[27]
L. Gkatzikis and I. Koutsopoulos, “Mobiles on cloud nine: efficient task migration policies for cloud computing
systems,” 3rd International Conference on Cloud Networking (CloudNet), IEEE, Luxembourg, pp. 204-210, October
2014.
[28]
S. Ioannidis and P. Marbach, “On the design of hybrid peer-to-peer systems,” International Conference on
Measurements and Modeling of Computer Systems (SIGMETRICS), ACM, Annapolis, MD, USA, pp. 157-168, June
2008.
[29]
Y. Zhang, L. Xu, and X. Wang, “A cooperative secure routing protocol based on reputation system for ad hoc
networks,” Journal of Communications, vol. 3, no. 6, pp. 43-50, November 2008.
[30]
M. Armbrust, A. Fox, R. Griffith, A. D. Joseph, R. Katz, A. Konwinski, G. Lee, D. Patterson, A. Rabkin, I. Stoica,
and M. Zaharia, “A view of cloud computing,” Communications of the ACM, ACM, vol. 53, no. 4, pp. 50-58, April
2010.
[31]
W. Tian, Y. Zhao, Y. Zhong, M. Xu, and C. Jing, “A dynamic and integrated load-balancing scheduling algorithm for
Cloud datacenters,” IEEE International Conference on Cloud Computing and Intelligence Systems (CCIS), IEEE,
Beijing, China, pp. 311-315, September 2011.
[32]
J. Kubiatowicz, D. Bindel, Y. Chen, S. Czerwinski, P. Eaton, D. Geels, R. Gummadi, S. Rhea, H. Weatherspoon, W.
Weimer, C. Wells, and B. Zhao, “OceanStore: An Architecture for Global-Scale Persistent Storage,” ACM SIGPLAN
Notices, ACM, vol. 35, issue 11, pp. 190-201, November 2000.
[33]
B. Blanchet, “Automatic verification of security protocols in the symbolic model: the verifier ProVerif,” Foundations
of Security Analysis and Design (FOSAD) VII, Springer, LNCS, vol. 8604, pp. 54-87, 2014.
[34]
I. Ristic, “OpenSSL cookbook,” Feisty Duck, pp. 1-94, March 2016.
[35]
S. Yang, S. Ji, and R. Beyah, “DPPG: A Dynamic Password Policy Generation System,” IEEE Transactions on
Information Forensics and Security, IEEE, vol. 13, no. 3, pp. 545-558, March 2018.
[36]
Y. S. Han, H.-T. Pai, R. Zheng, and W. H. Mow, “Efficient exact regenerating codes for Byzantine fault tolerance in
distributed networked storage,” IEEE Transactions on Communications, IEEE, vol. 62, issue 2, pp. 385-397, February
2014.
[37]
G. Hatzivasilis, I. Papaefstathiou, and C. Manifavas, “ModConTR: a Modular and Configurable Trust and Reputation-
based system for secure routing in ad-hoc networks,” 11
th
ACS/IEEE International Conference on Computer Systems
and Applications (AICCSA), IEEE, Doha, Qatar, November 2014.
[38]
A. V. Rodrigues, A. Jorge, and I. Dutra, “Accelerating recommender systems using GPUs,” 30
th
Annual ACM
Symposium on Applied Computing (SAC), ACM, Salamanca, Spain, pp. 879-884, April 2015.
[39]
K. Fysarakis, G. Hatzivasilis, I. Papaefstathiou, and C. Manifavas, “RtVMF: A secure real-time vehicle management
framework with critical incident response,” IEEE Pervasive Computing Magazine (PVC), Special Issue on Smart
Vehicle Spaces, IEEE, vol. 15, issue 1, pp. 22-30, January 2016.
[40]
J. Hayes, “Calling all cars,” IET Engineering & Technology, IET, vol. 9, issue 3, pp. 58-61, April 2014.
[41]
Q. Pei, B. Kang, L. Zhang, K.-K. R. Choo, Y. Zhang, and Y. Sun, “Secure and privacy-preserving 3D vehicle
positioning schemes for vehicular ad hoc network,” EURASIP Journal on Wireless Communications and Networking,
Springer, vol. 2018, issue 1, article 271, pp. 1-12, December 2018.
Received December 2018; revised July 2019; accepted September 2019
... 5G communication is therefore important for the implementation of flexible, efficient, and energy-aware solutions that underpin the IoT and realize the vision of a sustainable CE (Hatzivasilis et al., 2018). The merging of 5G and computing technologies can help to build trust and scalability in IoT-based applications such as smart transportation, ride-service hailing, and car sharing (Hatzivasilis et al., 2020). ...
... We categorized the articles' keywords that signify various ways in which the IoT enables the CE into three themes; product recovery strategies (i.e., reuse, remanufacturing, recycling), servitization, and CE levers (i.e., circular business models, interindustry cooperation, collaboration, and synergy) (see Table A). Impact of augmented reality on CE process performance Facilitate effective real-time interaction and data sharing between field technicians and expert engineers situated remotely Role of augmented reality to shift existing CE systems into data-driven and smart systems Enable product customization and the design of more efficient and cost-effective manufacturing networks Willingness of employees and stakeholders (e.g., customers) to engage in augmented reality-based CE systems Support prototyping processes by increasing design process visualization, testing, manufacturing redesign, and reducing material use Impact of augmented reality on economic, environmental, and social sustainability in manufacturing Role of 5G to support repair and maintenance decisions (Demestichas and Daskalakis, 2020;Hatzivasilis et al., 2018Hatzivasilis et al., , 2020Mboli et al., 2020;Miaoudakis et al., 2020) Support the development of smart factories Investigation of the security and privacy risks raised by the adoption of 5G in the CE Increase industrial automation in the CE Evaluation of the specific requirements of CE use cases when deploying 5G Streamline manufacturing processes and respond to their needs in terms of network throughput, efficiency, and scalability 5.2.1. Support product recovery strategies Reuse. ...
Article
Full-text available
In recent years, the concept of a circular economy (CE) has gained importance and attracted significant attention among scholars and practitioners. Research that examines the role of modern technologies in supporting the transition from a linear economy to the CE is therefore highly needed. This article analyzes and classifies existing research at the intersection of the CE and the Internet of Things (IoT), as an enabling technology. While studies on both concepts have proliferated, there is a lack of research that systematizes the literature and clarifies the relationship between the IoT and the CE. In order to achieve this, we reviewed a total of 170 academic articles published between 2007 and 2021 from the Scopus and Web of Science databases. Based on the coding of keywords, four categories can be identified: (1) IoT-related technologies in the CE context, (2) enablers of IoT in the CE, (3) barriers to IoT adoption in the CE, and (4) the impacts of the IoT on the sustainability of (circular) economies. The current study is the first attempt to use a keyword coding approach to better understand IoT research in the CE domain. The review findings identify important drivers and enablers and provide a structured framework for research in this field. Finally, this study highlights several research directions that may provide valuable insights for researchers and practitioners.
... To address the above limitations, this research introduces the need of a unified trust management framework that evaluates service providers' provisioned services in the MEC network considering various parameters. Trust evaluation is performed in a centralized manner, by a fully trusted third party known as cloud service manager (CSM), to grantee trust results credibility (Felix & Ricardo, 2012;Hatzivasilis et al., 2020). This also promotes for a secure and successful transactions between service users and providers in the MEC environment (Rathee et al., 2017). ...
Article
Full-text available
Mobile edge computing (MEC) is introduced as part of edge computing paradigm, that exploit cloud computing resources, at a nearer premises to service users. Cloud service users often search for cloud service providers to meet their computational demands. Due to the lack of previous experience between cloud service providers and users, users hold several doubts related to their data security and privacy, job completion and processing performance efficiency of service providers. This paper presents an integrated three-tier trust management framework that evaluates cloud service providers in three main domains: Tier I, which evaluates service provider compliance to the agreed upon service level agreement; Tier II, which computes the processing performance of a service provider based on its number of successful processes; and Tier III, which measures the violations committed by a service provider, per computational interval, during its processing in the MEC network. The three-tier evaluation is performed during Phase I computation. In Phase II, a service provider total trust value and status are gained through the integration of the three tiers using the developed overall trust fuzzy inference system (FIS). The simulation results of Phase I show the service provider trust value in terms of service level agreement compliance, processing performance and measurement of violations independently. This disseminates service provider’s points of failure, which enables a service provider to enhance its future performance for the evaluated domains. The Phase II results show the overall trust value and status per service provider after integrating the three tiers using overall trust FIS. The proposed model is distinguished among other models by evaluating different parameters for a service provider.
... For performance evaluation of suggested system and for comparison with existing systems, we utilized the software simulator tool named NS-2 [51]; in Table 2, the suggested system is evaluated with different significant challenges such as CH duration, Average, Stability convergence, cluster member, control overhand by speed and vehicle, energy consumption and throughput of various parameters. Secure Stable-CA [51][52][53][54][55][56][57] and Moth-Flame-CA [58] were used to perform the comparison of stab Trust. Once the operation on different elements of trust, the value of stab Trust is between 0.0 to 1.0. ...
Article
Full-text available
Many algorithms use clustering to improve vehicular ad hoc network performance. The expected points of many of these approaches support multiple rounds of data to the roadside unit and constantly include clustering in every round of single-hop data transmission towards the road side unit; however, the clustering in every round maximizes the number of control messages and there could be the possibility of collision and decreases in network energy. Multi-hop transmission prolongs the cluster head node’s lifetime and boosts the network’s efficiency. Accordingly, this article proposes a new fuzzy-clustering-based routing algorithm to benefit from multi-hop transmission clustering simultaneously. This research has analyzed the limitation of clustering in each round, different algorithms were used to perform the clustering, and multi-hop routing was used to transfer the data of every cluster to the road side unit. The fuzzy logic was used to choose the head node of each cluster. Three parameters, (1) distance of each node, (2) remaining energy, and (3) number of neighbors of every node, were considered as fuzzy criteria. The results of this research were compared to various other algorithms in relation to parameters like dead node in every round, first node expire, half node expire, last node expire, and the network lifetime. The simulation results show that the proposed approach outperforms other methods. On the other hand, the vehicular ad hoc network (VANET) environment is vulnerable at the time of data transmission. The NS-2 software tool was used to simulate and evaluate the proposed fuzzy logic opportunistic routing’s performance results concerning end-to-end delay, packet delivery, and network throughput. We compare to the existing protocols, such as fuzzy Internet of Things (IoT), two fuzzy, and Fuzzy-Based Driver Monitoring System (FDMS). The performance comparison also emphasizes an effective utilization of the resources. Simulations on the highway environment show that the suggested protocol has an improved Quality of Service (QoS) efficiency compared to the above published methods in the literature.
... To share resources in a secure way among different network entities, Hatzivasilis et al. proposed a new combined trust model, namely, MobileTrust [32]. This proposal takes into account the advantages of the back-end centralised cloud and 5G technologies to provide an environment where trusted information can be disseminated in VANET. ...
Article
Internet-of-Vehicles (IoV) is a novel technology to ensure safe and secure transportation by enabling smart vehicles to communicate and share sensitive information with each other. However, the realization of IoV in real-life depends on several factors, including the assurance of security from attackers and propagation of authentic, accurate and trusted information within the network. Further, the dissemination of compromised information must be detected and vehicle disseminating such malicious messages must be revoked from the network. To this end, trust can be integrated within the network to detect the trustworthiness of the received information. However, most of the trust models in the literature relies on evaluating node or data at the application layer. In this study, we propose a novel hybrid trust management scheme, namely, NOTRINO, which evaluates trustworthiness on the received information in two steps. First step evaluates trust on the node itself at network layer, while second step computes trustworthiness of the data at application layer. This mechanism enables the vehicles to efficiently model and evaluate the trustworthiness on the received information. The performance and accuracy of NOTRINO is rigorously evaluated under various realistic trust evaluation criteria (including precision, recall, F-measure and trust). Furthermore, the efficiency of NOTRINO is evaluated in presence of malicious nodes and its performance is benchmarked against three hybrid trust models. Extensive simulations indicate that NOTRINO achieve over 75% trust level as compared to benchmarked trust models where trust level falls below 60% for a network with 35% malicious nodes. Similarly, 92% precision and 87% recall is achieved with NOTRINO for the same network, comparing to benchmark trust models where precision and recall falls below 87% and 85% respectively.
... The physics of the mobile ecosystem are different from the web [37]- [38]. A mobile application runs on movable devices, like laptops, smart phones or tablets. ...
Conference Paper
Full-text available
Pairs of usernames and passwords are widely used nowadays by mobile and web applications to identify users. The exposure of this data harms both users and vendors. The client-server model is the most common. The provided services implement front-end interfaces that run on the client's side and back-end interfaces that run on the server side. A proper password management policy administrates the password creation, storage, processing, and transmission in both ends. This article overviews the theory and provides a practical guide for password management and implementation of a safe login process for mobile and web application developers, and IT organizations. An empirical research and several case studies are surveyed for the password habits of three universities, an army school, an IT company, and two accounting offices in the province of Crete in Greece. Moreover, a software benchmark analysis is conducted for the computational demanding primitives of the secure login operations .
... Smart packaging design: As recycled or renewable packaging materials often come with different properties, e.g., related to stability or appearance, smart designs are the prerequisite further to replace virgin material in Henkel's product packaging-this is why the company will put an even stronger focus (Continues) (2020) The implementation of Industry 4.0 and the CE also includes an in-depth look at business models. It is interesting to note that positions emerge that enhance the sharing economy of vehicles (Pham et al., 2019) and traffic and road status monitoring systems with updated consumption indicators that make drivers aware of a more linear view of driving styles (Hatzivasilis et al., 2020). The academic analysis also includes the field of food; the results of Sarti, Corsini, Gusmerotti, and Frey (2017) show business models based on social sharing platforms, at company level and food alert. ...
Article
The union between Industry 4.0 and the circular economy (CE) appears relatively recent. In this sense, new trading zones for sharing a common scenario among academics and practitioners are needed. The paper aims to investigate the link between Industry 4.0 and the CE by understanding how Industry 4.0 can foster the impact of the CE on companies. The study proposes a broader perspective that includes thematic and content analysis gathering data on professional documents based on business cases, newspaper articles, press releases and specialised blogs, as well as scientific papers. The joint academic-practitioners view highlights how Industry 4.0 has the potential to impact on the CE through countless actions: increasing waste disposal; promoting remanufacturing; enhancing the efficiency of critical resources such as water, energy, gas and CO 2 ; and improving business models and the mission of companies. However, barriers still exist in its adoption, stressing the need for holis-tic and integrated design and a proactive environment of collaboration among stake-holders. Results lead to practical as well as research implications. K E Y W O R D S business models, circular economy, Industry 4.0, remanufacturing, resource efficiency, waste management
... Every day, highly interconnected systems, utilizing not just the ordinary computer technologies but also the Internet of Things (IoT) and the cloud, exchange high volumes of data and user-related information [1,2]. This complex ecosystem cannot be safeguarded easily, as the attack surface is continuously increasing, while the security of the deployed primitives is not always retained [3][4][5]. Therefore, successful attacks have been demonstrated by researchers or have been actually performed by hackers, exploiting the underlying vulnerabilities (e.g., [6,7]). The risk still remains high, not only for large organizations, but for small-to-medium enterprises (SMEs) and individuals as well. ...
Article
Full-text available
Nowadays, more-and-more cyber-security training is emerging as an essential process for the lifelong personnel education in organizations, especially for those which operate critical infrastructures. This is due to security breaches on popular services that become publicly known and raise people’s security awareness. Except from large organizations, small-to-medium enterprises and individuals need to keep their knowledge on the related topics up-to-date as a means to protect their business operation or to obtain professional skills. Therefore, the potential target-group may range from simple users, who require basic knowledge on the current threat landscape and how to operate the related defense mechanisms, to security experts, who require hands-on experience in responding to security incidents. This high diversity makes training and certification quite a challenging task. This study combines pedagogical practices and cyber-security modelling in an attempt to support dynamically adaptive training procedures. The training programme is initially tailored to the trainee’s needs, promoting the continuous adaptation to his/her performance afterwards. As the trainee accomplishes the basic evaluation tasks, the assessment starts involving more advanced features that demand a higher level of understanding. The overall method is integrated in a modern cyber-ranges platform, and a pilot training programme for smart shipping employees is presented.
Article
Full-text available
Due to the mobility of devices, device-to-device (D2D) communication is a promising fifth-generation (5G) technology in dynamic environments for improving message transmission efficiency for group communication. Additionally, all services in an ad hoc network are current Vehicle Ad Hoc Network (VANET) applications. Therefore, D2D communication has been introduced in ad hoc environments to reduce latency during vehicle conversations, such as autonomous vehicle solutions and drone fleet management for cellular vehicle-to-everything (C-V2X) modules and Internet of Drones (IoDs) networks. However, providing secure and effective group communication is an urgent challenge. To solve these problems, we propose a dynamic group management solution based on distributed ledger technology. This study demonstrates that a distributed ledger-based hierarchical architecture for dynamic group management is faster and more adaptable without compromising security and performance. Furthermore, the proposed method can facilitate the transfer of direct communication data without a centralized database, thereby reducing the chance of a single point of failure. In addition, the research was tested by a third party that has established close cooperation with world-leading automotive electronics suppliers in Taiwan.
Conference Paper
Full-text available
A main type of Mobile Ad hoc Networks (MANET) and essential infrastructure to provide a wide range of safety applications to passengers in vehicles (VANET) are established. VANETs are more popular today as they connect to a variety of invisible services. VANET protection is crucial as its potential use must not endanger the safety and privacy of its users. The safety of these VANETs is essential to safe and efficient safety systems and facilities and uncertainty continues and research in this field continues to grow rapidly. We will explain the characteristics and problems of VANETs in this paper. Also, all threats and attacks that affect integrity and authentication in VANETs will be defined. Description of researchers' work was consequently addressed as the table with the problems of the suggested method and objective. Keywords-VANETs, V2V, Integrity, Authentication, RSU. I. In t r o d u c t i o n The ad hoc network of vehicles (VANETS) has increasingly called attention to both academic and industrial considerations. It is a crucial element of the intelligent transport system aimed at improving driving safety through vehicle communication and roadside infrastructure communication^, 2]. With the help of the roadside and web, this VANETS Base improves the early access of traffic accident data for end-users. VANETs can be used to provide reliable indications of prevention systems in the early stage of a potential transport incident as a monitoring and early warning framework. In the past few decades, emerging developments have developed into the intelligent Transportation System (ITS) and been data-driven ITS as massive quantities of data are gathered from a wide range of sources. VANET s are processes that are self-organized and decentralized[3]. VANET provides Dedicated short-range communication through the IEEE 802.11p protocol of 5.9 GHz band (DSRC). In addition, VANET provides another scenario for networking such as V2V and V2I communications for local and remote environments. VANET also has a different communications scenario. Wherever VANET s have been one of the most exciting and growing mobile ad hoc network subsets (MANETs). These include autonomous devices, roadside systems, and onboard units (OBUs) that are connected by unreliable wireless media[4]. They are also available on the Internet. Besides the lack of infrastructure, the equipment manufacturers travel at different speeds[5, 6]. The delay is the establishment and effective transfer of secure end-to-end communications ways. VANETs thus faced different systems issues and security challenges when it comes to open, safe communication, communication, and reputation management systems. This affects the trust of the portable system in collaboration and arrangement. They are genuinely defenseless to attacks by the fluctuating performance of the brand, which can lead to life-threatening conditions as seen in figure 1. Fig. 1. Vehicular ad hoc network (VANETS). In this article, we discuss the challenges of VANET privacy and security threats. We discuss the architecture, privacy, and security problems in the performance of VANET's and cloud computing. The number of private vehicles on the road has also increased with an increase in trade and increased progress[7]. The probability of accidents and road accidents has been increased. The ad-hoc vehicle networks (VANETs) have shown promise to reduce road accidents and deaths by allowing vehicles to be communicated. Furthermore, VANETs allow road users to control and track rapid rush and relief vehicles. However, design, routing, communication, and safety problems face these networks[8]. The use of wireless networks has opened these networks to various types of security attacks. The possibility that a malicious vehicle or RSU has several identities is an important weakness in such a network. The attack is called an attack by Sybil[9, 10]. VANETs allow for a malicious vehicle to send inaccurate traffic signals, a forward collision, a closed lane, etc. This can cause the driver 978-1-6654-4058-5/21/$31.00 ©2021 IEEE Authorized licensed use limited to: Karabuk University. Downloaded on June 28,2021 at 10:45:35 UTC from IEEE Xplore. Restrictions apply.
Chapter
Traffic info about the vehicular community is able to assist within the setup of smart site traffic managing, like car crash alerts, street building notices, as well as car owner path modifications to lessen site traffic congestion a bit earlier. Real-time site traffic info could effectively mirror the visitor’s state within urbanized scenarios. Nevertheless, how you can develop effective real-time info discussing mechanism to accomplish the real-time going dynamic path and time estimation preparation continue to be a difficult issue. With this newspaper, we very first suggest a highly effective real-time site traffic info discussing mechanism according to a sent out commuter routes process, with reduced computing intricacy as well as a reduced amount of redundancy. Next, talking about today’s site traffic info, a method is presented by us to calculate the going period via a spot within the highway system, referred to as traveling time estimation (TTE) that functions being a metric for track preparation. We have achieved the accuracy level of information about the traffic in real time on comparing with existing models.
Article
Full-text available
Abstract Industrial wireless networks (IWNs) have applications in areas such as critical infrastructure sectors and manufacturing industries such as car manufacturing. In car manufacturing, IWNs can facilitate manufacturers to improve the design of the vehicles by collecting vehicular status and other related data (such an IWN is also known as vehicular ad hoc networks—VANETs). Vehicle positioning is a key functionality in VANETs. Most existing vehicle positioning systems are capable of providing accurate 2D positioning, but the demand for accurate 3D positioning has increased sharply in recent times (e.g., due to the building of more elevated roads). There are, however, security and privacy concerns relating to 3D positioning systems in VANET. In this paper, we propose two secure and privacy-preserving 3D positioning schemes based on vehicle-to-roadside (V2R) and vehicle-to-vehicle (V2V) communications, respectively. Our schemes are based on the round trip time ranging technique which is used to achieve 3D position. The security and the privacy of vehicles in our schemes are guaranteed through a newly designed one-pass authenticated key agreement protocol. Using experiments, we show that a vehicle can determine whether it is on or under an elevated road in a short period of time.
Article
Full-text available
Smart vehicle-to-grid (V2G) involves intelligent charge and discharge decisions based on user operational energy requirements, such as desired levels of charging and waiting time. V2G is also supported by information management capabilities enabled by a secure network, such as a reliable privacy-preserving payment system. In this article, we describe the network security and privacy requirements and challenges of V2G applications. We present a new network security architecture to support V2G. We propose a scheme with the following security and privacy-preserving features: anonymous authentication, fine-grained access control, anonymous signatures, information confidentiality, message integrity, remote attestation, and a payment system. This article is oriented toward practitioners interested in designing and implementing secure and privacy-preserving networks for smart V2G applications.
Conference Paper
Full-text available
Mobile edge clouds (MECs) are small cloud-like infras-tructures deployed in close proximity to users, allowing users to have seamless and low-latency access to cloud services. When users move across different locations, their service applications often need to be migrated to follow the user so that the benefit of MEC is maintained. In this paper, we propose a layered framework for migrating running applications that are encapsulated either in virtual machines (VMs) or containers. We evaluate the migration performance of various real applications under the proposed framework.
Article
Internet of Vehicles (IoV), which enables information gathering and disseminating among vehicles, roadside infrastructures and surrounding environments, has received considerable attention recently. However, the flourishing of IoV still faces several challenges in terms of location privacy preservation, vehicular sensory data collection and vehicular sensory data acquisition. Aiming at these challenges, in this paper, we propose a novel efficient and location privacy-preserving data sharing scheme with collusion resistance in IoV, which enables the collection and distribution of the data captured by vehicular sensors. During the vehicular sensory data collection phase, each vehicle structures the multi-dimensional sensory data captured at different locations and exploits the modified Paillier Cryptosystem to achieve the location privacy-preserving sensory data aggregation, while the data aggregation result of multiple vehicles can be recovered by the trusted central entity. During the vehicular sensory data acquisition phase, the proposed scheme exploits the proxy re-encryption technique to achieve the location privacy-preserving data querying at the network edge, i.e., the vehicles query the RSU instead of the trusted central entity. Numerical analysis are performed to validate the effectiveness of the proposed scheme, i.e., low data querying failure probability. Security analysis and performance evaluations are also carried out to validate the security properties and show the computation and communication efficiency of the proposed scheme.
Article
To keep password users from creating simple and common passwords, major websites and applications provide a password-strength measure, namely a password checker. While critical requirements for a password checker to be stringent have prevailed in the study of password security, we show that regardless of the stringency, such static checkers can leak information and actually help the adversary enhance the performance of their attacks. To address this weakness, we propose and devise the Dynamic Password Policy Generator, namely DPPG, to be an effective and usable alternative to the existing password strength checker. DPPG aims to enforce an evenly-distributed password space and generate dynamic policies for users to create passwords that are diverse and that contribute to the overall security of the password database. Since DPPG is modular and can function with different underlying metrics for policy generation, we further introduce a diversity-based password security metric that evaluates the security of a password database in terms of password space and distribution. The metric is useful as a countermeasure to well-crafted offline cracking algorithms and theoretically illustrates why DPPG works well.
Conference Paper
In this paper, we design and evaluate SORS- a scalable online ridesharing system, where drivers and passengers send their requests for a ride in advance, possibly on a short notice. SORS is modular and consists of two main, loosely coupled, components: the Constraint Satisfier and the Matching Module. The Constraint Satisfier takes as input information about the desired trajectories and spatio-temporal constraints of drivers and passengers and it returns a list of feasible (driver, passenger) pairs. We use a road networks data structure, optimized for the specific spatio-temporal queries in the context of ridesharing, and we show that our Constraint Satisfier has a 4.65x more scalable query time than a general-purpose database. We represent the feasible pairs of drivers and passengers as a weighted bipartite graph with edge weight being the length of the shared trip of the pair, which captures the revenue in real-world ridesharing systems, such as Lyft Carpool. The Matching Module then takes as input this weighted bipartite graph and returns the maximum weighted matching (MWM), using an algorithm that solves the problem online and efficiently, by incrementally updating the matching solution in real-time. We show that our algorithm achieves 51% larger weight (i.e., total revenue) compared to greedy heuristics used by many real systems today. We also evaluate the SORS system as a whole, using mobile datasets to extract driver trajectories and passenger locations in urban environments. We show that SORS can provide a ridesharing recommendation to individual users within a sub-second query response time, even at high workloads.
Article
MANET and VANET are the active research areas and lots of routing protocols have been proposed for use in these areas.. In MANET, nodes are connected through wireless channels in a network and each node acts as a router and as a host. One of the scenario of MANET is Vehicular ad-hoc networks. For communication in VANET, efficient Routing Protocols are needed. Because of highly changing network topology and frequent disconnection it's strenuous to design an efficient routing protocol for vehicles, there can be two types of VANET that are V2V(Vehicle to Vehicle) and V2RSU(Vehicle to Road Side Unit). Because of daily happening of accidents VANET is one of the affecting areas for the refinement of Intelligent Transportation System (ITS) which can insure passengers and road safety. The Intelligent Transport Systems gives information if there exists any emergency and tells about traffic density. the traffic and traffic density. The existing routing protocols for VANET are not efficient enough to meet all traffic scenarios. Worthy routing protocols are required to initiate communication between vehicles in future for passengers and road safety. This paper shows literature survey related to Reactive and Proactive Routing Protocols of MANET as AODV, DSDV, OLSR, and DSR. Analysis and characterization of these protocols is shown in the paper which helps in further improvement of existing routing protocols.
Conference Paper
Vehicular Ad hoc Networks (VANETs) have been becoming a popular research topic in both academia and industry due to the demanding requirements of safe and efficient transportation systems. The security of VANETs is a significant issue for the implementation of most types of useful applications based on VANETs. In this paper, a trust-based framework is proposed to solve the security issue in VANETs. The trus-based mechanism can provide a second protection for VANETs to defense the inside attackers, which are hard to be thwarted by cryptographic techniques. The proposed scheme establishes trust with direct interactions and recommendations. The DempsterShafer Theory is used to tackle the uncertainty of thirdparty recommendations in order to obtain accurate trust values. The effectiveness of the proposed scheme is validated by simulations.