Privacy Regulations Challenges on Data-centric and IoT Systems:
A Case Study for Smart Vehicles
Lelio Campanile, Mauro Iacono, Fiammetta Marulli
and Michele Mastroianni
Dipartimento di Matematica e Fisica, Universit
`
a degli Studi della Campania ”L.Vanvitelli”,
Viale A. Lincoln, 5, Caserta, Italy
Keywords:
Internet of Things, Data Centric System, Internet of Vehicles, GDPR, Data Privacy, Data Security,
Cyber-physical Systems, Blockchain.
Abstract:
Internet of Things (IoTs) services and data-centric systems allow smart and efficient information exchang-
ing. Anyway, even if existing IoTs and cyber security architectures are enforcing, they are still vulnerable
to security issues, as unauthorized access, data breaches, intrusions. They can’t provide yet sufficiently ro-
bust and secure solutions to be applied in a straightforward way, both for ensuring privacy preservation and
trustworthiness of transmitted data, evenly preventing from its fraudulent and unauthorized usage. Such data
potentially include critical information about persons’ privacy (locations, visited places, behaviors, goods,
anagraphic data and health conditions). So, novel approaches for IoTs and data-centric security are needed. In
this work, we address IoTs systems security problem focusing on the privacy preserving issue. Indeed, after
the European Union introduced the General Data Protection Regulation (GDPR), privacy data protection is a
mandatory requirement for systems producing and managing sensible users’ data. Starting from a case study
for the Internet of Vehicles (IoVs), we performed a pilot study and DPIA assessment to analyze possible miti-
gation strategies for improving the compliance of IoTs based systems to GDPR requirements. Our preliminary
results evidenced that the introduction of blockchains in IoTs systems architectures can improve significantly
the compliance to privacy regulations.
1 INTRODUCTION
The Internet of Things (IoTs) has become a naturally
included part of people’s daily life; in the last decade,
it introduced an effectively revolutionary system of
technologies able to modify people’s life, in sectors
ranging from communication and services to trans-
port, to health, from entertainment (Marulli et al.,
2016),(Marulli and Vallifuoco, 2017) to our interac-
tions with public institutions and government.
Anyway, as any relevant opportunity, also IoTs in-
troduces a number of significant challenges, most of
which are strictly related to users’ security and pri-
vacy rights. Data are easily produced, exchanged and
consumed by a multitude of services and devices but
users don’t know exactly where their data are physi-
cally stored and who can effectively access.
Indeed, both trustworthiness and security of data
management systems have become issues of primary
relevance; privacy and private data rights manage-
ment has become so crucial to the point that govern-
Corresponding Author
ments were obliged to introduce specific regulations
to ensure privacy preservation in any kind of transac-
tion involving sensible information.
In May 2018, the European Union (EU) perma-
nently introduced the General Data Protection Reg-
ulation (GDPR) (The European Union, 2016), a
mandatory regulation on privacy data management
applied across all the Members of EU. It has been ap-
plied directly to processing activities of personal data
having relationships with any markets or territory in
the European Union; from the date of its official entry
to force, any breach of its provisions has been punish-
able by heavy sanctions, imposed by all Data Protec-
tion Authorities (DPAs) (Albrecht, 2016).
The GDPR was the latest step in the ongoing
global recognition of the value and importance of
personal information and it arised from the necessity
to have a clear and unified regulation and protection
measures in front of a fragmented digital market and
the lack of enforcement in the field of data protection
provisions. It represents a unified and directly ap-
plicable data protection law for the European Union
Campanile, L., Iacono, M., Marulli, F. and Mastroianni, M.
Privacy Regulations Challenges on Data-centric and IoT Systems: A Case Study for Smart Vehicles.
DOI: 10.5220/0009839305070518
In Proceedings of the 5th International Conference on Internet of Things, Big Data and Security (IoTBDS 2020), pages 507-518
ISBN: 978-989-758-426-8
Copyright
c
2020 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
507
which has replaced almost all of the existing Mem-
ber States’ provisions and which have been applied by
businesses, individuals, courts and authorities without
transposition into national law (Team, 2017).
The GDPR has conditioned not only the European
data protection laws but the whole world, thus evi-
dencing that the GDPR is currently one of the most
significant data security law in the world.
So far, if the importance of respecting the GDPR
regulations has been well understood by organiza-
tions,several are the concerns on its effective and cor-
rect implementation. It suggests a set of reference
guidelines but it does not provide precise explanations
or effective practices for transactions in order to be
privacy preserving compliant.
For massive digital systems and services, as the
IoTs based ones, the GDPR requirements introduced
further criticalities, ranging from the design and im-
plementation of novel robust and effective solutions
for secure data management to the proof of com-
pliance to mandatory regulations (Di Martino et al.,
2019). So, data-centric and IoTs systems have been
strongly impacted by these regulations and fully ap-
propriate solutions haven’t been still designed.
In this work, the impacts of GDPR privacy regu-
lations on data-centric and IoTs systems design and
applications are investigated. In particular, we fo-
cused on analyzing possible strategies to assess the
level of compliance to specific privacy regulation, as
the GDPR, that are exhibited by service infrastruc-
tures and applications, based on IoTs technologies.
To this aim, we considered a case study from the
Internet of Vehicles (IoVs), describing the scenario
of an accident occurring among smart vehicles. Such
case study was functional to perform a pilot study for
analyzing mitigation strategies to better meet trust-
worthiness and GDPR privacy constraints. When an
accident occurs, privacy concerning data are typically
exchanged among several stakeholders, ranging from
the people involved in the accident (e.g., the vehi-
cles drivers or witnesses) to the insurance companies
and/or the law enforcements.
In such a scenario, data are quickly exchanged by
multiple stakeholders and no one can exactly ensure
who will be enabled to access to that data, referring
people’s locations, involved people’s identities and
personal data. So, data can’t be exchanged by com-
mon communication services and means but need to
be managed in appropriate ways, able to ensure the
freedom and the right of anonymity besides the effec-
tive stakeholders.
To the aim of assessing any possible mitigation
strategy, we discussed two variant scenarios for the
proposed case study, considering, in turn the ab-
sence and the presence in introducing the use of a
blockchain in the data exchanging and managing loop
of an IoVs application.
Indeed, we performed a comparison between the
two scenarios (with and without adopting the use of
the blockchain), evidencing the benefits obtained by
using blockchains, in terms of an increased com-
pliance to the privacy preserving goal prescribed by
GDPR. The degree of compliance was defined as the
estimation of the risk level occurring for an IoTs ap-
plication to be in privacy regulation violation region,
with or without introducing the blockchain in the data
managing loop.
Finally, the pilot study provided in this work is
functional to further studies for defining a systematic
metrics framework for assessing privacy regulations
compliance for practical IoTs service applications.
The rest of the paper is organized as follows: Sec-
tion 2 presents related works, section 3 describes the
considered case study for an internet of vehicles ap-
plication; in section 4 a blockchain based and GDPR-
compliant solution we proposed is described; in sec-
tion 5 the results obtained by the proposed approach
are discussed and section 6 concludes the work by de-
scring the insights of our study and the further direc-
tions for improving this investigation.
2 RELATED WORKS
Internet of Things (Atzori et al., 2010) primarily
means a world-spanning information fabric built on
a wide range network computing environment made
of smart devices, able to provide smart solutions, us-
ing internet for communicating automatically among
each other without requiring any human intervention.
Such intelligent devices perform actions basing on
the information they get from other connected de-
vices anytime and anywhere; this is how they perform
their designated tasks intelligently by deciding in real
time (Solangi et al., 2018). As the next generation of
the Wireless Sensor Networks, IoTs systems and in-
frastructures behaviors and performances can be also
simulated by advanced network simulators as the NS-
3 (Campanile et al., 2020), provide with tailored li-
braries for supporting the various IoTs scenarios.
However, beyond several fantastic opportunities,
IoTs also presents a number of significant challenges.
The growth in the number of devices and the speed
of that growth presents challenges to our security
and freedoms that require to develop policies, stan-
dards, and governance able to shape this development
without stifling innovation. Various applications of
IoTs, like smart home and buildings, smart health-
AI4EIoTs 2020 - Special Session on Artificial Intelligence for Emerging IoT Systems: Open Challenges and Novel Perspectives
508
care, smart vehicles and smart appliances, e.g., have
anticipated yet security issues due to authentication
and data integrity and, in the recent past, an increas-
ing number of researches have focused on security,
data privacy and trust concerns related to IoTs in-
frastructures. One of the main factor causing secu-
rity and privacy pitfalls and faults in the IoTs ecosys-
tem, as discussed in (Maple, 2017), can be found in
the lack of standardisation and regulation, since the
IoT has evolved using a wide range of core technolo-
gies from a number of key visions, through develop-
ments by distinct, often disparate communities. Fur-
ther, these developments have been made in different
application areas, often using specific and proprietary
standards. This diffuse nature of development has led
to an inevitable lack of harmonisation and shared vi-
sion, hampering standardisation and effective regula-
tion, which has left technicians and users without the
necessary information and control to, service, update,
and address problems created with devices and ser-
vices. The lack of coherence, oversight, understand-
ing, and protocols means that security risk analysis,
risk assessment, and countermeasure implementation
(Cimino et al., 2020) are much more difficult tasks
than they would be with a more directed and coor-
dinated development path. (Maple, 2017) provides
a deep discussion on the security and privacy chal-
lenges in the IoT, illustrated through a number of key
applications. In this discussion, the securing prin-
ciples of systems are also discussed, from the CIA
of information security (confidentiality, integrity, and
availability), to the five keys of information assur-
ance (confidentiality, integrity, availability, authentic-
ity, and non- repudiation) (Frattolillo et al., 2009) and
the Parkerian Hexad (confidentiality, integrity, avail-
ability, authenticity, possession, and utility) (Parker,
1983). Research works discussing security consider-
ations relating to complex cyber-physical (as opposed
to information) and IoT systems vary in which prin-
ciples they adopt. The majority of researchers restrict
consideration to the CIA. The Parkerian Hexad use-
fulness remains the subject of debate among security
professionals (Sattarova Feruza and Kim, 2007). Oth-
ers studies also include robustness, reliability, safety,
resilience, performability, and survivability (Sterbenz
et al., 2010). challenges.
As for the authentication within the IoT, it is a
critical point too, since without appropriate authen-
tication the confidentiality, integrity, and availability
of systems could compromised. This is the reason
for an adversary to be able to authenticate as a legiti-
mate user, to access to any data that the user has, and
be able to see (compromising confidentiality), modify
(compromising integrity), and delete or restrict avail-
ability (compromising availability) in the same way
that the user can. The authentication and identifica-
tion of users in the IoT remains a significant chal-
lenge. Currently, username/password pairs are the
most common form of authentication and identifica-
tion of users in electronic systems. However, the
vision of the IoT as ubiquitous will eliminate many
of the physical interaction interfaces through which
usernames and passwords are passed.
Furthermore mobility, privacy, and anonymity re-
quire further analysis and research because those IoTs
systems featuring mobile services will have users
passing through different architectures and infrastruc-
tures owned by different providers. Managing the
identity of users in mobile, multiply owned and het-
erogeneous environments is strongly challenging.
As for the security and privacy challenges in the
IoTs environments, in (Mena et al., 2018) is provided
a survey from the perspective of the adopted technolo-
gies and architecture. This research is focused on IoT
intrinsic vulnerabilities and their implications to the
fundamental information security challenges in con-
fidentiality, integrity, and availability. From a wide
range exploration of the IoT architecture security re-
lated issues, concluding that all layers of IoT archi-
tectures are involved in security issues: data access
and authentication, phishing attacks, malware attacks
(Mercaldo and Santone, ),(Casolare et al., 2019), ma-
licious scripts are all examples of threats for the IoT
application layer are; in the IoT network layer, net-
work congestions, devices interference, eavesdrop-
ping attack, routing attack, denial of service,attack
node jumping in WSN and heterogeneity problem are
examples of security issues.
2.1 The GDPR
As for the privacy issues (YANG et al., 2019), the sce-
nario was furtherly complicated by the introduction
of mandatory regulations for data protection in all the
world countries. As for the European Union, it intro-
duced the General Data Privacy Regulation (GDPR)
(Voigt and Von dem Bussche, 2017) that currently re-
quires all data controllers and processors that handle
the personal information of EU residents to adopt and
implement appropriate technical and organisational
measures to ensure the ongoing confidentiality, in-
tegrity, availability and resilence of processing sys-
tems and services or face heavy financial penalties.
The GDPR is the latest step in the ongoing global
recognition of the value and importance of personal
information. Such a recognition has obtained only
recently the appropriate attention, because of the in-
creasing number of cyber thefts and the misuse of per-
Privacy Regulations Challenges on Data-centric and IoT Systems: A Case Study for Smart Vehicles
509
sonal data by governments and corporations, that ex-
pose EU citizens to significant personal risks. Thus,
the need of new law clarifying the data rights of EU
citizens to ensure the appropriate level of EU-wide
protection for personal data, led to the composition of
the GDPR, that applies across all the Member Coun-
tries of the EU but involves any organisation any-
where in the world providing services into the EU
that involve processing and managing personal data
(Team, 2017). Consequently, technological smart ser-
vices and infrastructures (e.g., smart cities) dealing
with sensitive data have been strongly impacted since
they must ensure individual privacy and security in or-
der to ensure that people will going on participating.
Sensitive data (special categories of personal data, as
defined in Article 9 of GDPR), which are generated
by a plurality of subjects or connected to the inter-
est of a plurality of subjects, have to be collected,
maintained and kept unaltered but inaccessible until
legally authorized. Internet-of-Things (IoT) has be-
come a naturally included part of our daily life with
billions of IoT devices collecting data through wire-
less technology and able to interoperate within the ex-
isting Internet infrastructure. Moreover, the new fog
computing paradigm allows storing and processing
data at the network edge or anywhere along the cloud-
to-endpoint continuum, thus overcoming the limita-
tions of IoT devices. IoT, fog computing and smart
cities are all representative of Data Centric Systems
and paradigms, needing of being provided by people’s
partecipation and data in order to be working. If peo-
ple are reluctant to participate, the core advantages of
all data centric infrastructures will dissolve. Security
and privacy issues, such as secure communication,
authentication and authorization, information confi-
dentiality and integrity, introduce destabilizing and
costly disruptions for data centric systems.Moreover,
the widely distribution of millions of smart devices,
located in different areas, increases the risk of being
compromised by some malicious parties. Actual chal-
lenges that Data Centric Systems are obliged to issue
and manage includes privacy preservation with high
dimensional data, securing networks with large attack
surface, establishing trustworthy data sharing prac-
tices, properly utilizing artificial intelligence, and mit-
igating failures cascading through the smart network,
in order to avoid people deceiving and data protection
regulations penalties.
2.2 Blockchain Applications to IoTs
Systems
Blockchain, or distributed ledger technology, is an
emerging platform that is designed to support transac-
tions services within a multi-party business network,
hopefully ensuring for all involved partners signifi-
cant cost and risk reductions. In the last few years,
this technology has started to be used in a wider range
of application, including Internet of Things (IoT). A
blockchain enables IoT devices to send data for inclu-
sion in a shared transaction repository with tamper-
resistant records, and enables business parties to ac-
cess and supply IoT data without the need for cen-
tral control and management Many researchers have
addressed the problem of integrating blockchain with
IoT.
Blockchain (BC) technologies are well known for
their feature of data immutability (Zheng et al., 2019).
A BC, in fact, consists of an open and distributed
ledger running over a peer-to-peer (P2P) network that
can manage transactions for multiple entities effi-
ciently and in a verifiable and traceable way, thus
without recurring to the intervention of a middleman.
A blockchain is made of a list of blocks record-
ing a transaction and linked using cryptography. Each
block contains a timestamp represented by a crypto-
graphic hash of the previous block, where the data
of the transaction are represented as a Merkle tree.
The tamperproof nature of blockchain comes from
the fact that, once consensus is reached and a block
is committed, the data in the block cannot be al-
tered retroactively without alteration of all subsequent
blocks, which requires the consensus of the majority.
Furthermore, BC can follow both a centralized ar-
chitecture and a decentralized one but the fundamen-
tal problem of scalability keeps the blockchains oper-
ating in beta and alpha mode in business and industry.
All these security and performance concerns
are due to the “Scalability Trilemma”, where a
blockchain system tries to offer scalability, decen-
tralization and security, without compromising any
of them. Decentralization is the core property that
enables the censorship-resistance and permissionless
features. Scalability is the ability to process transac-
tions on a network with increased size.
Security is an important component that guaran-
tees the immutability of the ledger and its resistance
to general cyber attacks.
As for the Internet of Things (IoTs) applications
(Wang et al., 2019), BC technologies have the po-
tential to address the data security concern in IoT
networks. While providing data security to the IoT,
Blockchain also encounters a number of critical chal-
lenges inherent in the IoT, such as a huge number
of IoT devices, non-homogeneous network structure,
limited computing power, low communication band-
width, and error-prone radio links. Traditional solu-
tions are based on registries owned and managed by
AI4EIoTs 2020 - Special Session on Artificial Intelligence for Emerging IoT Systems: Open Challenges and Novel Perspectives
510
trusted third-party or Authorities representing public
regulators with generally complex, anyway certified
insertion or updating procedures. The regulator can
actually manage this registry as a whole, if the fre-
quency of operations can be actually handled, as a
system of partial registries, articulating the distribu-
tion of operations on the basis of proper criteria (e.g.,
according to a well defined geographic partitioning),
or acting as a top authority of a tree-organized del-
egated registries that are provided by third parties
according to specific agreements and duties, that al-
lows online legal identification of persons by using
tokens issued by third parties that can verify creden-
tials. However, these solutions represent a logical
single point of failure, when non distributed, a cost
(including technical, organizational and legal com-
ponents), when maintained by public regulators, or
a strong delegation choice, when involving commer-
cial third parties that may be also providing other
non-regulated services, potentially exposed to secu-
rity breaches. Furthermore, the spreading of the In-
ternet of Vehicles, that enables a large number of ser-
vices ranging from vehicle-to-vehicle communication
to autonomous drive and platooning, to the interac-
tion with smart roads (Karpiriski et al., 2006), makes
accessible the use of a blockchain for traffic monitor-
ing, security and safety support and the solution of le-
gal issues because of the native integration of sensing
and computing on board and the possibility of cap-
turing and cross compare the local state of the traf-
fic. The fact that different stakeholders have a legit-
imate interest in ensuring the availability of all rele-
vant information to support their internal processes,
to discharge responsibilities, to certify operations and
to disambiguate critical situations with a lower level
of risk and reducing the need for unilateral preventive
solutions makes sharing such a blockchain infrastruc-
ture viable and convenient. In the more specific field
of integrating Blockchain technology and Internet of
Vehicles (IoV) (Odiete et al., 2018), propose a fully
decentralized architecture that combines blockchain
and traditional distributed database to gain additional
features such as efficient query and retrieval of meta-
data stored on the blockchain. (Truong et al., 2018)
assert that before making any transactions it is cru-
cial to evaluate trust between participants for reduc-
ing the risk of dealing with malicious peers, and is
proposed a trust-based IoV model including a sys-
tem architecture, components and features. (Arora
and Yadav, 2018) present an authentication and secure
data transfer algorithm, in the IoV framework using
blockchain technology, to ensure secure information
communication and to improve the efficiency. Fur-
thermore, Regarding blockchain and IoT integration,
(Reyna et al., 2018) analyze the challenges emerging
from the integration of IoT and blockchain, present-
ing possible ways of integration and platforms that are
integrating IoT and blockchain in a general context.
(Christidis and Devetsikiotis, 2016) describe how a
blockchain-IoT combination may facilitates the shar-
ing of services and resources leading to the creation of
a marketplace of services between devices and allows
to automate in a cryptographically verifiable manner
several existing, time-consuming work ows.
3 THE PROPOSED PILOT STUDY
FOR PRIVACY COMPLIANCE
The rise of the Internet of Vehicles, that enables a
large number of services ranging from vehicle-to-
vehicle communication to autonomous drive and pla-
tooning, to the interaction with smart roads, makes ac-
cessible the use of a blockchain for traffic monitoring,
security and safety support and the solution of legal
issues because of the native integration of sensing and
computing on board and the possibility of capturing
and cross compare the local state of the traffic. More-
over, the integration of production and maintenance
information may be soon seamlessly available with
the same technology, as literature reports the appear-
ance of blockchains to support supply chain in manu-
facturing sectors like aerospace and automotive. The
fact that different stakeholders (insurances, manufac-
turers, garages, black box service providers, smart
road concessionaires) have a legitimate interest in en-
suring the availability of all relevant information to
support their internal processes, to discharge respon-
sibilities, to certify operations and to disambiguate
critical situations with a lower level of risk and reduc-
ing the need for unilateral preventive solutions makes
sharing such a blockchain infrastructure viable and
convenient. Moreover, a careful design of the data
management and signature mechanism may also pro-
vide the possibility of implementing different levels
of privacy, so to have non-personal but commercially
significant data certified and accessible by the autho-
rized party to produce additional information-driven
services in the interest of the final users.
3.1 The Proposed Case Study for the
Internet of Vehicles
The case study we considered for the aim to inves-
tigate mitigation strategies for major compliance of
IoTs based systems to privacy regulations, was taken
from the Internet of Vehicles (IoVs) application do-
Privacy Regulations Challenges on Data-centric and IoT Systems: A Case Study for Smart Vehicles
511
main and the specific privacy regulation we consid-
ered is the European Union GDPR. The main idea
behind our pilot study consists is the fact that sev-
eral are the stakeholders interested in both manag-
ing the information about the behavior of the vehi-
cles and defending themselves from possible privacy
rights violations deriving from data detention. Indeed,
the uncontrolled tracking of data concerning events
related to personal or commercial information can’t
be allowed,while the availability of a non-contestable
registration and partial access to anonymized infor-
mation with sharing of responsibility is needed. In
the proposed case study, we adopted asymmetric key
pairs both to sign information and assure the require-
ment of non repudiation, and to encrypt sensitive in-
formation, in order to allow the authorized access
only to the legitimate involved stakeholders or to dis-
close information to the authorities when a law judge-
ment procedure occurs. Furthermore, we adopted the
usage of blockchains to share responsibility, guaran-
tee and verify immutability of information, beyond a
lower level of risk and protection costs, without dis-
closing sensible information.
3.2 Internet of Vehicles Scenario
The set of stakeholders we considered in our study is
depicted in Figure 1and briefly characterized as fol-
lows:
vehicles drivers;
vehicles manufacturers;
insurance companies;
smart road dealerships;
car parking garages;
public registry authorities;
Drivers are the effective users of vehicles. Users rep-
resent citizens that own or drive private vehi- cles,
while company users represent people that drive ve-
hicles belonging to the farm they are working for; fi-
nally, vehicle rental customers represent people that
drive vehicles belonging to a company that rented
them for a short or long period to them. In the first
case, a user is not expected to be a participant of
a blockchain on his own, but he interacts with the
blockchain by his vehicle, and the vehicle is expected
to be equipped with a black box and a black box ser-
vice provider acting as a proxy; the user may also
interact with the blockchain for bureaucracy through
another participating blockchain service provider, and
the black box will directly provide services to a vehi-
cle as well. In the second case, vehicles are under the
responsibility of a fleet manager whose company par-
ticipates the blockchain to track vehicle events that
will be disclosed by the authority if needed, to en-
act defensive strategies and build a non repudiable
log to prove drivers’ responsibilities, and to collect
anonymized data about own fleet for management,
maintenance and further optimization reasons.
A vehicle manufacturer produces vehicles and
performs maintenance in the warranty period. So, it
can be considerate a legitimate interest allowing this
stakeholder to collect diagnostic data about the circu-
lating vehicles it has produced in order to collect de-
fects affecting safety of drivers, circulation and other
people or items. Anyway, data should anonimyzed in
order to be not linkable to the owners identities. No
other stakeholder should be able to obtain this infor-
mation, while its correlation to the sense status of the
car should be noncontestable, to let authorities under-
stand as better as possible if any accident might occur.
Such a stakeholder has the financial and technical ca-
pability needed to participate in a blockchain, that can
be used to track all data related to its supply chain and
production processes.
A insurance company has no right to access any
information about the vehicle until no events cov-
ered by an insurance policy occur. This stakeholder
has a legitimate interest in holding a copy of all en-
crypted data to be assure the veracity of information
used in case of disputes or trials. In this perspective, it
can take advantage in participating in the blockchain,
even if data are anonimyzed, if the authority pro-
vides plain data and a policy to check them against
the blockchain without violating the users’ privacy.
A smart road dealership tracks data about vehicles
passing by road it grants for, for limited time usage
or based on statistical aggregations over long periods.
It aims for improving safety, allowing autonomous
drive, traffic management and implementing strate-
gies for improving efficiency. It gets systematically
in control of a large quantity of data about drivers’
behavior, vehicle movement and traffic patterns, pos-
sibly with a chance for managing enough data to be
potentially used to violate privacy and mine sensitive
information: thus, it is interested both in a third-party
certification of the veracity of information, eventually
requested by authorities when incidents occur and in
sharing responsibilities. Both the services can be pri-
vided by participating in the blockchain, without dis-
closing data and with the possibility of processing
anonymized data with the aid of law authorities as
a compensation for the production of public interest
data.
A vehicle parking garage works on vehicles to per-
form scheduled or extraordinary maintenance, poten-
AI4EIoTs 2020 - Special Session on Artificial Intelligence for Emerging IoT Systems: Open Challenges and Novel Perspectives
512
Figure 1: Stakeholders in the Blockchain Based IoVs.
tially impacting the safety of the vehicle. This stake-
holder, in the person of the owners, has a legitimate
interest in reliably tracking operations to limit own
responsibility, while the manufacturer could benefit
from being updated with maintenance events occur-
ring outside the boundaries of its own branches. Any-
way, participating the blockchain may be not tech-
nologically possible for small business organizations,
like garages, so a trusted proxy could participate the
blockchain and provide services to garages.
Finally, the public registry represents the authority
managing accesses to sensitive data. It participates
the blockchain on a peer role, evenly provided with
the rights to access all data on behalf law authorities
or when insurance companies signal the starting of
a procedure involving their ordinary activities. Par-
ticipating the blockchain allows the public registry to
retain a copy of all data without the duty of directly
managing all registrations from all parties and with-
out the responsibility of being the only official repos-
itory of all data. The public registry is here consid-
ered a trusted party operating by applying all legal
requirements, including systems security and enforc-
ing all law constraints over the blockchain, that im-
pacts over all participants. The public registry has
the responsibility of keeping all public cryptographic
keys used in conjunction with the blockchain, the pri-
vate key needed to access to sensitive data logged by
the blockchain and the mapping between identities of
users and vehicles and between vehicles and black-
boxes.
4 THE GDPR-COMPLIANT
BLOCKCHAIN BASED
SOLUTION
The GDPR is always applicable when personal data
are managed and processed. The more significant
rights for users include several rights, among which
the ones to access own data or data about oneself, to
change or fix wrong and missing information about
oneself, to request the erasure of data about oneself
(”right to be forgotten”). Some of these rights seem
to be forged to shape the features of blockchain sys-
tems, as the data added to a digital ledger, without
the possibility to change or remove it after its writ-
ing. Furthermore, blockchain-based solutions may
ensure more security features when compared with
other solutions: there are no single point of failure,
since it can improve fault tolerance and makes hack-
ing data very difficult. So, a blockchain-based solu-
tion may be profitably used in a GDPR-compliant en-
vironment and taking into account some remarkable
constraints. In order to ensure this constraint, all par-
ticipants (nodes) holding personal data has be known
by all the users; all of them need to be defined as joint
controllers (see Articles 24 and 26, GDPR). The most
suitable model of blockchain fitting these constraints
follows a private/permissioned paradigm (van Geelk-
erken and Konings, 2017). Moreover, in order to en-
sure full GDPR compliance, one of the main chal-
lenges for blockchain developers is to comply with
Privacy Regulations Challenges on Data-centric and IoT Systems: A Case Study for Smart Vehicles
513
the right to rectification (Article 16) and the right to
be forgotten (Article 17). Users have the right to re-
quest for rectification of their own personal data, and,
in some conditions, may request the erasure of own
personal data. Personal data must also be erased when
the personal data are no longer necessary in relation
to the purposes for which they were collected or oth-
erwise processed (Article 17), even if the user did not
request for it. Accordingly to these constraints, the
proposed blockchain based solution fits the following
requirements:
it is private and permissioned;
data modifications are allowed but each update is
recorded in an additive block;
data erasing is allowed by encrypting blocks and
deleting the key.
4.0.1 Use Cases for the Compliance Analysis
The main use cases considered in the study for dis-
cussing the compliance analysis to GDPR are the
listed and briefly described in the following:
new vehicle registration: users can interact with
the system for registering a new vehicle before its
first usage, to activate all required procedures en-
abling the tracking by preserving his/her privacy,
to link the vehicle with its black box and to bind
his own identity to the vehicle and the black box.
new maintenance registration: a vehicle manufac-
turer could register maintenance when the vehicle
is maintained while the warranty period or at con-
trolled garages,and should be able to access all di-
agnostic information about each single produced
car.
a smart road concessionaire monitors events oc-
curring while vehicles ride its granted road and
should be able to register them officially, includ-
ing information that relates events with the in-
volved vehicles.
an insurance company interacts with its customers
outside the system for the most of the cases, be-
cause all commercial activities, communications
and service requests happen in presence, by phone
or through the internet.
The public registry needs to access data on au-
thority request with full availability of contents,
to register pseudo identities, being the only entity
that keeps the correspondence between pseudo-
identities and vehicles, handles insurance requests
by providing data about involved parties about a
limited time frame and verification means in case
of incidents, encrypts obsolete data to delete them
automatically when re- quested by GDPR or en-
crypts data about a user to delete them on user
request as allowed and prescribed by GDPR.
5 THE GDPR COMPLIANCE
ANALYSIS
Ensuring privacy preservation and data protection is
mandatory for Government agencies and enterprise
organizations that require personal data of their cus-
tomers for providing IT services. Article 35 of the
GDPR refers to new technologies and prescribes data
processing by these novel technological could turn
into a high risk to the rights and freedoms of nat-
ural persons. So, each controller shall carry out a
Data Protection Impact Assessment (DPIA), that aims
to conduct a systematic risk assessment in order to
identify system vulnerabilities and privacy threats,
thus prescribing and imposing technical and organi-
zational controls to mitigate those threats. Article 35
prescribes that DPIA shall in particular be required in
the case of:
a systematic and extensive evaluation of personal
aspects relating to natural persons which is based
on automated processing, including profiling, and
on which decisions are based that produce legal
effects concerning the natural person or similarly
significantly affect the natural person;
processing on a large scale of special categories
of data referred to in Article 9(1), or of personal
data relating to criminal convictions and offenses
referred to in Article 10;
a systematic monitoring of a publicly accessible
area on a large scale. The case we are study-
ing, a car tracking system, relates both to item
a. and c. cited before; so, according to GDPR,
DPIA must be carried out. In the following sub-
sections, a DPIA is performed on our system com-
pared with an hypothetical state-of-the-art cloud-
based system to emphasize advantages of our de-
sign choice. In this paper we do not wish to draw
up a complete impact assessment, as it would need
more technical and organizational details. Our
aim is to show which advantages may be obtained
by using a blockchain- based framework instead
of a simple cloud-based solution, and only some
significant threats are evaluated.
5.1 DPIA and Risk Analysis
Risk is defined, according to standard ISO
31000:2009, as the effect of uncertainty on ob-
AI4EIoTs 2020 - Special Session on Artificial Intelligence for Emerging IoT Systems: Open Challenges and Novel Perspectives
514
jectives, and it is often quantified as the probability
of occurrence of an adverse event, times its negative
impact (severity). Information Security Risk Assess-
ment (ISRA) is commonly classified according to
three criteria: Quantitative, Qualitative, and Hybrid
(Semi-Quantitative) (Shameli-Sendi et al., 2016).
Quantitative risk assessment is based on ob-
jective measurements; its results are expressed in
management-specific metrics (i.e. monetary value,
percentages and probabilities). The main problem
with quantitative assessment is the heavy effort re-
quired in terms of time-consuming and it is biased by
the quality of information. Due to these limitations
too, this approach is not so easy to be practically im-
plemented. Alternatively, the qualitative approach is
the most common into information security risk as-
sessment processes, where many organizations find it
adequate for their needs (Landoll and Landoll, 2005).
In a qualitative approach, classes are introduced to
show the impact and probability of a particular sce-
nario. The qualitative approach is widely used, be-
cause of the lack of very accurate historical data for
computing the impact and probability of occurrence
of risks; it results to be also much easier to understand
and implement, the calculations involved are simpler
and the evaluation of assets,threats and vulnerabilities
is simplified (Landoll and Landoll, 2005). Hybrid ap-
proach represents a combination of the two previous
approaches (Shameli-Sendi et al., 2016).
5.2 Risk Assessment Metrics
In this study we adopt a qualitative approach based
methodology, as the one adopted by CNIL, the French
Data Protection Authority; this methodology is also
recommended by many EU Data Protection Authori-
ties, including the Italian Authority for Personal Data
Protection. In addition to the methodology, CNIL also
developed a software tool (PIA) to support managers
in drawing up DPIA.
The CNIL-PIA software provides classifications
accord to risk sources:
internal human resources (employees, managers,
e.g.);
external human resources (business partners, visi-
tors, maintenance staff);
non-human sources (malicious code such as
viruses, worms etc., earth-quake, malicious
users);
and according to feared events:
disappearance of personal data.
illegitimate access to personal data;
unwanted modification of personal data;
Severity and likelihood are the qualitative metrics es-
timated for performing the risk assessment. Sever-
ity represents the magnitude of a risk and it is pri-
marily estimated in terms of the extent of potential
impacts on data subjects, taking account of existing,
planned or additional controls. Its value is estimated
by considering the potential damage occurred to the
user/data subject. The scale for estimating severity
ranges among the following levels:
1. negligible: not affected/few inconveniences
(loss of time, unsolicited mail);
2. limited: significant inconveniences (denial of
access in services, additional costs);
3. significant: significant consequences (fraud,
damages to property);
4. maximum: significant/irreversible conse-
quences (inability to work, loss of evidence in the
context of litigation, loss of access to vital infras-
tructure).
The severity level may be lowered by including
additional factors to contrast identification of per-
sonal data, such as encryption, pseudoanonymization,
anonymization, and so on.
Likelihood represents the feasibility of a risk to
occur and it is primarily estimated as the level of
vulnerabilities concerned to the supporting assets and
the level of capabilities of the risk sources to exploit
them, evenly considering existing or planned controls.
The classification scale for likelihood is related to the
feasibility of the occurrence of a risk, for each se-
lected risk sources that is able to exploit the properties
of supporting assets. As for the severity metric, also
the likelihood can be estimated according four ranged
values:
1. negligible: not (apparently) possible (e.g. theft
of paper documents stored in a room protected by
a badge reader and access code);
2. limited: difficult (e.g. theft of paper documents
stored in a room protected by a badge reader);
significant: possible (e.g. theft of paper docu-
ments stored in offices);
4. maximum: easy to breach (e.g. theft of paper
documents stored in the public lobby).
The likelihood level may be decreased by including
security enforcement components, as firewalls, log-
ging and monitoring services and so on.
5.3 DPIA Results and Discussion
In order to perform an impact assessment analysis and
compare the proposed blockchain solution, we con-
Privacy Regulations Challenges on Data-centric and IoT Systems: A Case Study for Smart Vehicles
515
sidered a second system, consisting of a hypothetical
IoT and cloud-based solution, accomplishing the fol-
lowing specifications:
backup and business continuity features, more
than two different physical sites;
cryptography of all data;
use of digital signature;
operation security compliant with a well-known
standard (eg. ISO 27001);
archiving and tracking feature for all transactions.
We also considered some potential impacts burden-
ing to users/data subjects. The choice of impacts is
related to the specific kind of data involved. The im-
pacts taken into account, and the related severity ac-
cording to CNIL-PIA, are represented by:
cost rise (severity = 2);
targeted online advertising on a confidential as-
pect (severity = 2);
fraud (severity = 3);
loss of evidence in the context of litigation (sever-
ity = 4).
Threats selection was mainly based on the data re-
trieved on Verizon 2019 Data breach Investigation re-
port (Ashraf, 2019) and EY Global Information secu-
rity Survey 2018-19 (Feng and Wang, 2019). These
threats can lead to illegitimate access (I), unwanted
modification (U) of or disappearing (D) of Personal
Data and are listed as follows:
hacking: it is the most frequent threat, circa 54%
(Verizon Enterprise (2019)) (I), (U);
use of stolen credentials: almost 30% of threats
(Verizon Enterprise (2019)) (I), (U);
privilege abuse: circa 10% of threats (Verizon En-
terprise (2019)) (I), (U), (D);
DDoS attacks: almost 60% of attacks in incidents,
usually this kind of attack does not lead to a data
modification or illegitimate access, but leads to
loss of availability (data disappearing), which is
considered a breach event in GDPR (D);
natural disasters: although they are infrequent
events (EY appraises circa 2% of total breaches
(EY (2019))), they are taken into account because
they can lead to a severe data loss (D).
In case of the considered cloud based solution, the
value of Risk Severity and Likelihood are estimated
as follows:
illegitimate access to personal data: the involved
impacts are cost rise, confidential aspects and
fraud, so severity is placed at 3 (significant), be-
cause this is the maximum level of severity in this
set. Likelihood is placed to 3 (significant), due to
the relatively consistent probability of occurrence
of the event;
unwanted modification of personal data: all im-
pacts are involved, so severity is 4 (maximum).
Likelihood is placed to 3 (significant), due to the
relatively consistent probability of occurrence of
the event;
disappearance of personal data: the impact in-
volved are cost rise and loss of evidence, so sever-
ity is 4 (maximum), but we think that state-of-the-
art countermeasures such as business continuity
may lower the severity to 3. Likelihood is placed
to 2 (limited), due to the low probability of occur-
rence of the event.
The insights of our DPIA analysis evidenced in are
shown in Fig. 2. In this picture we used the green
region and the red region to represent compliance and
non compliance regions. For the cloud based solu-
tion we observed that DPIA leads to a ”High Risk”
evaluation and, according to Article 36 of GDPR, it
is mandatory for the data controller to consult the su-
pervisory authority before processing. In the case of
the assessment for the blockchain-based solution, the
values were estimated as follows:
illegitimate access to personal data: the involved
impacts are cost rise, con
dential aspects and fraud, so the severity is placed
at 3 (significant). In this case, severity is low-
ered due to the large-scale use of pseudonymiza-
tion. Users’ data are all exchanged using pseudo
identities; vehicle data are anonymized and de-
coupled from user identfiers and pseudo identity
data. Likelihood is also lowered to 2 (limited): the
risk due to stolen credential and privilege abuse is
significantly lower, because the only credentials
which may see/modify personal data are those as-
signed to the public registry and no sysadmin or
black box manager has any access. Hacking prob-
ability is also decreased, due the extremely low
probability of a success in enacting a hacking at-
tack that overtakes 50% + 1 of the blockchain
nodes;
unwanted modification of personal data: all im-
pacts are involved, so severity is 4 (maximum).
Also in this case, the severity is mitigated due to
the large-scale use of pseudonymization. Users’
data are all exchanged using pseudo identities; ve-
hicle data are anonymized and decoupled from
user and pseudo identity data. In this case, the
severity and the likelihood values are lowered to
AI4EIoTs 2020 - Special Session on Artificial Intelligence for Emerging IoT Systems: Open Challenges and Novel Perspectives
516
2 (limited). The risk due to the fact that stolen
credential and privilege abuse events likelihood is
significantly lower, because the only credentials
which may see/modify personal data are those as-
signed to the public registry and no sysadmin or
black box manager has any access. Hacking prob-
ability is also lower, as for the previous point;
disappearance of personal data: the involved im-
pacts are cost rise and loss of evidence, so severity
is 4 (maximum), but we think that state-of-the-art
countermeasures such as business continuity may
lower the severity to 3. Likelihood is also lowered
to 1 (negligible): the event has a low probability
to occur, and the large number of nodes involved
make almost impossible to lead to a permanent
data loss.
Summarizing up the DPIA results, in Figure 2 red
dots represent the cloud-based solution while blue
dots represent the blockchain-based solution. Is no-
ticeable that in the case of the blockchain-based so-
lution all the tree values (I, U and D) are placed in
the ”green” area. This means that, according to Arti-
cle 26, DPIA leads to ”medium risk” evaluation and
there is no need to consult the supervisory authority
before processing.
Figure 2: Risk Metrics Estimation for the Blockchain-based
solution.
6 CONCLUSIONS
In this article we have discussed a pilot study aim-
ing to perform a preliminary analysis on the impacts
of data privacy regulations, on data-centric and IoTs
service infrastructures and applications. Indeed, one
of the major challenge for security and privacy as-
surance in the IoT, posed by the introduction of pri-
vacy regulations is still represented by the definition
and the standardization of tangible specifications that
could be transformed in software and hardware re-
quirements. Arguably the most fundamental chal-
lenge, in the close future, is to encourage standardi-
sation and coordination in the IoT systems design, by
taking into account all conflicting views of the sevearl
stakeholders on the IoT and the difficulties related in
gaining consensus and trust between parties with dif-
ferent interests, goals and visions. In this perspec-
tive, we investigated any mitigation strategies for as-
sessing the compliance of IoTs and data-centric sys-
tems to privacy regulations, as the European Union
GDPR, putting also the basis for further studies aim-
ing to identify a systematic metrics framework to
guide the practical design and implementation of IoTs
and data-centric systems and applications to be effec-
tively compliant to data privacy regulations and for
measuring the compliance level, in order to prevent
violations. We performed our study by starting from a
case study taken from the Internet of Vehicles and we
aimed to observe if the introduction of a blockchain
in the data managing loop could mitigate the impact
of data privacy regulations on the system, in terms of
compliance to its mandatory requirements. The de-
gree of compliance was defined as the estimation of
the risk level occurring for an IoTs application to be
in privacy regulation violation region, with or without
introducing the blockchain in the data managing loop
and was performed by the means of a DPIA. As a pre-
liminary result of this study, we evidenced that the in-
troduction of a blockchain in the data loop could mit-
igate part of these problems and improve the position
of systems and application from the boundaries of the
region in which a regulations violation could occur.
We explored this possibility in this work and we ob-
tained, as we expected, promising insights that will be
further investigate in the proceeding of the study we
have proposed in this work.
ACKNOWLEDGEMENTS
This work is part of the research activities realized
within the project ”Attrazione e Mobilit
`
a dei Ricerca-
tori” Italian PON Programme (PON AIM 2018 num.
AIM1878214-2).
REFERENCES
Albrecht, J. P. (2016). How the gdpr will change the world.
Eur. Data Prot. L. Rev., 2:287.
Arora, A. and Yadav, S. K. (2018). Block chain based se-
curity mechanism for Internet of Vehicles (IoV). In
Privacy Regulations Challenges on Data-centric and IoT Systems: A Case Study for Smart Vehicles
517
Proceedings of 3rd International Conference on Inter-
net of Things and Connected Technologies (ICIoTCT),
2018, pages 267–272.
Ashraf, C. (2019). 5g and mixed reality glasses help change
how we see the world. verizon enterprise tech.
Atzori, L., Iera, A., and Morabito, G. (2010). The internet of
things: A survey. Computer networks, 54(15):2787–
2805.
Campanile, L., Gribaudo, M., Iacono, M., Marulli, F., and
Mastroianni, M. (2020). Computer network simula-
tion with ns-3: A systematic literature review. Elec-
tronics, 9(2):272.
Casolare, R., Martinelli, F., Mercaldo, F., and Santone, A.
(2019). A model checking based proposal for mo-
bile colluding attack detection. In 2019 IEEE Inter-
national Conference on Big Data (Big Data), pages
5998–6000. IEEE.
Cimino, M. G., De Francesco, N., Mercaldo, F., San-
tone, A., and Vaglini, G. (2020). Model checking
for malicious family detection and phylogenetic anal-
ysis in mobile environment. Computers & Security,
90:101691.
Christidis, K. and Devetsikiotis, M. (2016). Blockchains
and Smart Contracts for the Internet of Things. IEEE
Access, 4:2292–2303.
Di Martino, B., Mastroianni, M., Campaiola, M., Morelli,
G., and Sparaco, E. (2019). Semantic techniques for
validation of gdpr compliance of business processes.
In Conference on Complex, Intelligent, and Software
Intensive Systems, pages 847–855. Springer.
Feng, C. Q. and Wang, T. (2019). Does cio risk appetite
matter? evidence from information security breach in-
cidents. International Journal of Accounting Informa-
tion Systems, 32:59–75.
Frattolillo, F., Landolfi, F., and Marulli, F. (2009). A novel
approach to drm systems. In 2009 International Con-
ference on Computational Science and Engineering,
volume 3, pages 492–497. IEEE.
Karpiriski, M., Senart, A., and Cahill, V. (2006). Sensor
networks for smart roads. In Fourth Annual IEEE In-
ternational Conference on Pervasive Computing and
Communications Workshops (PERCOMW’06), pages
5 pp.-310
Landoll, D. J. and Landoll, D. (2005). The security risk as-
sessment handbook: A complete guide for performing
security risk assessments. CRC Press.
Maple, C. (2017). Security and privacy in the internet of
things. Journal of Cyber Policy, 2(2):155–184.
Marulli, F., Pareschi, R., and Baldacci, D. (2016). The inter-
net of speaking things and its applications to cultural
heritage. In IoTBD, pages 107–117.
Marulli, F. and Vallifuoco, L. (2017). Internet of things for
driving human-like interactions: a case study for cul-
tural smart environment. In Proceedings of the Second
International Conference on Internet of things, Data
and Cloud Computing, pages 1–9.
Mena, D. M., Papapanagiotou, I., and Yang, B. (2018). In-
ternet of things: Survey on security. Information Se-
curity Journal: A Global Perspective, 27(3):162–182.
Mercaldo, F. and Santone, A. Deep learning for image-
based mobile malware detection. Journal of Computer
Virology and Hacking Techniques, pages 1–15.
Odiete, O., Lomotey, R. K., and Deters, R. (2018). Using
blockchain to support data and service management
in IoV/IoT. In Peng, S.-L., Wang, S.-J., Balas, V. E.,
and Zhao, M., editors, Security with Intelligent Com-
puting and Big-data Services, pages 344–362, Cham.
Springer International Publishing.
Parker, D. B. (1983). Fighting computer crime. Scribner
New York, NY.
Reyna, A., Mart
´
ın, C., Chen, J., Soler, E., and D
´
ıaz,
M. (2018). On blockchain and its integration with
IoT. Challenges and opportunities. Future Genera-
tion Computer Systems, 88:173–190.
Sattarova Feruza, Y. and Kim, T. (2007). It security review:
Privacy, protection, access control, assurance and sys-
tem security. International journal of multimedia and
ubiquitous engineering, 2(2):17–32.
Shameli-Sendi, A., Aghababaei-Barzegar, R., and Cheriet,
M. (2016). Taxonomy of information security risk as-
sessment (isra). Computers & security, 57:14–30.
Solangi, Z. A., Solangi, Y. A., Chandio, S., bt. S. Abd. Aziz,
M., bin Hamzah, M. S., and Shah, A. (2018). The
future of data privacy and security concerns in internet
of things. In 2018 IEEE International Conference on
Innovative Research and Development (ICIRD), pages
1–4.
Sterbenz, J. P., Hutchison, D., C¸ etinkaya, E. K., Jabbar, A.,
Rohrer, J. P., Sch
¨
oller, M., and Smith, P. (2010). Re-
silience and survivability in communication networks:
Strategies, principles, and survey of disciplines. Com-
puter Networks, 54(8):1245–1265.
Team, I. P. (2017). EU general data protection regulation
(GDPR): an implementation and compliance guide.
IT Governance Ltd.
The European Union (2016). Regulation (EU) 2016/679
on the protection of natural persons with regard to the
processing of personal data and on the free movement
of such data, and repealing Directive 95/46/EC (Gen-
eral Data Protection Regulation). Official Journal of
the European Union.
Truong, N. B., Um, T., Zhou, B., and Lee, G. M. (2018).
Strengthening the blockchain-based Internet of Value
with trust. In 2018 IEEE International Conference on
Communications (ICC), pages 1–7.
van Geelkerken, F. W. J. and Konings, K. (2017). Using
blockchain to strengthen the rights granted through
the GDPR. In 7th International youth science forum
”Litteris et Artibus”, pages 458–461.
Voigt, P. and Von dem Bussche, A. (2017). The eu general
data protection regulation (gdpr). A Practical Guide,
1st Ed., Cham: Springer International Publishing.
Wang, X., Zha, X., Ni, W., Liu, R. P., Guo, Y. J., Niu, X.,
and Zheng, K. (2019). Survey on blockchain for inter-
net of things. Computer Communications, 136:10–29.
YANG, J., KIM, C., and ONIK, M. M. H. (2019). Aggre-
gated risk modelling of personal data privacy in inter-
net of things. In 2019 21st International Conference
on Advanced Communication Technology (ICACT),
pages 425–430.
Zheng, X., Zhu, Y., and Si, X. (2019). A survey on chal-
lenges and progresses in blockchain technologies: A
performance and security perspective. Applied Sci-
ences, 9(22):4731.
AI4EIoTs 2020 - Special Session on Artificial Intelligence for Emerging IoT Systems: Open Challenges and Novel Perspectives
518