Secure IoT: An Improbable Reality
Nayana Mannilthodi and Jinesh M. Kannimoola
Amrita Center for Cybersecurity Systems & Networks, Amrita School of Engineering, Amritapuri, India
Amrita Vishwa Vidyapeetham, Amrita University, Kollam, Kerala, India
Keywords:
Internet of Things (IoT), Security, Attacks, Privacy, Vulnerability.
Abstract:
Internet of Things(IoT) has been the buzzword for the past decade. Apart from its hype over opportunities,
the security implications of IoT are unsolvable with current technologies. There is a wide range of security
challenges in each layer of IoT conceptual model. We discuss the security challenges caused by the complex
structures and integration of different techniques from diverse domains. By analysing attacks at the various
layers we argue that the current standards are not enough to provide a secure framework for IoT. The econom-
ical and practical reasons make it impossible to puzzle out the various security challenges in IoT stack. From
this perspective, we should think twice before connecting a device to the network of things.
1 INTRODUCTION
Internet of things (IoT) is an expansion of Internet to
the real world which interconnects the physical de-
vices to the virtual world. The term IoT was first pro-
posed by Kevin Ashton in 1999 at MIT labs (Ash-
ton, 2009) with a vision for a smarter planet. The
vital part of IoT stack are the sensors, which collects
data from the environment and shares with computing
cloud through the Internet (Sathyadevan et al., 2014).
The concept of IoT is revolutionary, which
changes the way people live, work, entertain, and
travel, as well as how governments and businesses
interact with the world (Insider, 2016). Everything
around us from daily life appliances such as wash-
ing machines, refrigerators, mobile phones to cars,
buildings and traffic cameras would gather data and
then stream it to concerned computing platform au-
tonomously. The analysis of this data can trigger the
automated action using actuators. In other words, ma-
chines would start to replace human decision-making.
Gartner has predicted that there will be 25 billion IoT
devices by the year 2020 (Gartner, 2014).
IoT opens new security challenges to the world
due to the wide use of sensors and smart devices. As
the model is tightly coupled with the real world in real
time, its security flaws cause more drastic results than
the Internet security flaws. The sensors, gateways and
repositories with full of interesting information are at-
tractive targets for hackers. A network of the poorly
secured device affects the security and resilience of
the IoT globally. In this paper, we will look into vari-
ous security implications of IoT and its reasons. The
security challenges of IoT is never ending. There are
many more attacks, which are beyond the scope of
this paper. IoT can never be secure even in the future.
New vulnerabilities are being invented and fixing it
each time is not economically and practically feasi-
ble. This is because of the diversity of IoT. Bringing
IoT to our daily activities is an invasion of our privacy
and leaves us more exposed to the attackers.
The remainder of this paper is organised as fol-
lows. Section 2 discusses conceptual layers in IoT
framework, and sections 3 and 4 describe the security
challenges and current standards with their limitations
respectively; section 5 describes why it is impossible
to solve IoT challenges; section 6 outlines the related
works and section 7 presents the conclusion.
2 CONCEPTUAL MODEL
ITU recommends(Kurakova, 2013) a five layer con-
ceptual model for Internet of Things. It offers a mod-
ular structure to IoT stack by separating the respon-
sibilities between different layers as described in Fig-
ure 1. We can summarise each layer in the conceptual
model as follows.
Perception Layer: This layer is responsible for
collecting the information about the things which
include sensors data, building parameters, loca-
tion etc. It also provides this information to the
338
Mannilthodi, N. and Kannimoola, J.
Secure IoT: An Improbable Reality.
DOI: 10.5220/0006352903380343
In Proceedings of the 2nd International Conference on Internet of Things, Big Data and Security (IoTBDS 2017), pages 338-343
ISBN: 978-989-758-245-5
Copyright © 2017 by SCITEPRESS Science and Technology Publications, Lda. All r ights reserved
upper layer for transmission over the Internet.
Access Layer: This layer enables the communi-
cation and information exchange through Internet
by various networks such as WiFi, mobile net-
works, satellites.
Internet Layer: It defines the connection estab-
lishment and the infrastructure for the upper lay-
ers along with the management and analysis of
data.
Service Management Layer: This layer is respon-
sible for providing the data models and interoper-
ability of different IoT services.
Application Layer: The application layer takes the
information and uses it for the different applica-
tions such as traffic signals, health care appliances
and disaster monitors.
Figure 1: Conceptual Layers of IoT.
3 SECURITY CHALLENGES
IoT devices interact very closely with human lives.
The security challenges of IoT devices has severe ef-
fects because it is the safety critical systems that are
connected to the Internet and these devices can per-
form machine to machine communication (Suo et al.,
2012; Sathyadevan et al., 2015). Securing IoT means
securing all the sensor data, access points, the net-
work infrastructure, the cloud, and the storage. Inter-
net of things is a collection of various technologies,
devices, and standards, thus ensuring security in this
diverse environment is one of the biggest challenges.
Few security challenges on each layer are mentioned
below.
3.1 Perception Layer
The perception layer mainly consists of sensor nodes
which have low computational power and are oper-
ated by battery. Most sensors would be disposable so
applying a software updateis a tedious task and would
not be cost effective.
Illegal access: The tags in the sensors use RFID
which does not use any authentication mechanisms
(Uttarkar and Kulkarni, 2014). An attacker can hack
into the sensors just by following the steps of an au-
thorised user and gather the information. The attacker
can modify or access the sensor data without much
difficulty.
Unauthorised tag cloning: An attacker can easily get
the tag identity and conduct integrity attacks by ma-
nipulating tag information. This vulnerability can be
exploited to overcome counterfeit protection in pass-
ports and drug labels. This vulnerability also ques-
tions the verification step which uses the RFID tags
for security procedures (Burmester and De Medeiros,
2007).
Unauthorised tag tracking: The attacker can trace and
track the tags then find its activity and location. These
are privacy attacks, which can also be exploited by
hackers or adversarial organisations.
Relay attacks: RFID is used as a contact-less identi-
fication card due to its read at a distance feature. But
the RFID identification information can be hacked by
the attacker. In this attack, the attacker borrows the
victims card information without his/her knowledge
or any physical access. This attack exploits the tags
response to a rogue readers challenge to impersonate
the tag (Hancke, 2005).
Eavesdropping: Eavesdropping attacks are one of the
most prominent risks of RFID devices. It is very easy
to sniff the information flow from reader to tag or tag
to reader (Hancke et al., 2008). The majority of the
communication use radio frequency spectrum. The
signals that are broadcasted can be simply intercepted
with a receiver tuned to the same frequency.
Spoofing: Spoofing is a variant of cloning that does
not require RFID tag to be physically replicated. The
attackers impersonate a valid RFID tag or broad-
cast fake information to the RFID (Mitrokotsa et al.,
2010).
Active jamming: An attacker can exploit the fact that
RFID tags listen to all radio signals in its range, he
can conduct a DoS attack by creating signals in the
same range that node uses to communicate with the
reader (Li, 2012).
Secure IoT: An Improbable Reality
339
3.2 Network and Access Layer
Network layer consists of the Wireless Sensor Net-
work (WSN), WiFi capabilities, mobile networks,
satellites, access points, gateways and all other net-
work infrastructure. The function of this layer is to
reliably transmit information between the perception
layer and service management layer. The network
layer has the challenges inherited from the Internet
and the additional ones from the specific features of
IoT (Patton et al., 2014).
Sybil attack: Sybil attack is a type of attack in which
a node illegitimately claims multiple identities. The
system can be compromised by this attack and the in-
formation from the perception layer would be mis-
interpreted. The attack can alter too many impor-
tant functions of the sensor network such as routing,
resource allocation, misbehaviour detection (Messai,
2014).
Sinkhole attack: Sinkhole is an attack in which an at-
tacker makes a node the most attractive one for the
other nodes to transfer the data. Thereby all the data
traffic will be diverted to this particular node which
can be rerouted to an attacker or made to drop the
packets to cause a DoS attack (Kibirige and Sanga,
2015).
Sleep deprivation attack: The ability of the wireless
node to sleep when the battery is down can be ex-
ploited by an adversary. An attacker can exploit the
clustering algorithm to drain the node’s energy sav-
ings and make it sleep. There are two varieties of
such attack one is sleep deprivation attack and the
other is barrage attack (Pirretti, 2006). Barrage attack
is noticeable as it aggressively bombards the node.
Whereas in the sleep deprivation attack the adversary
sends packets with an interval by which adversary can
keep the victim awake (Pirretti, 2006).
Location disclosure attacks: An attacker can get the
node location, structure of the network, a route map
by analysing the traffic. Thus the attacker can fig-
ure out the identities, learn from the network traffic
and track changes in the traffic pattern (Miller and
Valasek, 2015).
Other common network attacks: As mentioned before
the IoT also has the inherited vulnerabilities from the
Internet. Attackers can poison the routing tables, con-
trol the traffic flow, flood the network, drop the pack-
ets by interfering the network. An attacker can in-
ject malicious code into the network by compromis-
ing a node; by doing so, he can take down the entire
network or get full control over the network (Farooq
et al., 2015).
3.3 Service Management Layer
Insider attack: This attack is very common in cloud
infrastructure where an attacker is someone who is
inside the system and has access to cloud data. An
attacker can tamper the data on the cloud for personal
or third party benefits.
Weak authentication: In a test conducted by Syman-
tec on cloud services on IoT, it was found that most
of the IoT cloud services allow users to choose weak
passwords. In some other services, users are pre-
vented from using complex passwords. For example,
one service was found with password restriction of a
PIN code with maximum four digits. Even after many
failed login attempts, many services do not lock the
users out of their accounts. Also, none of the tested
back-end cloud services provides two-factor authenti-
cation.
3.4 Application Layer
The application layer is also vulnerable to security
attacks. The application contains lots of user data.
The protocols used in application layer such as HTTP,
SMTP, FTP have known vulnerabilities and are entry
points for attackers.
Malware attacks: Malware such as virus, spyware
and Trojan horses can steal the user information and
do malicious activities.
Code injection: The vulnerabilities in the applications
due to the mixing up of code and data can be exploited
by hackers. An attacker can inject malicious code into
the system and steal user data (Barcena and Wueest,
2015).
3.5 Multi-layer Attacks
Some attacks are done by exploiting the vulnerabili-
ties that are evolved from integrating multiple layers.
Botnets: Botnet is a network of nodes remotely con-
trolled by an adversary to perform DDoS attacks, to
distribute malware, to steal private data and to sent
spam or phishing emails. Thingbots are botnets of
connected objects. These botnets have a variety of
connected devices. Wireless routers and modems are
the main targets for Thingbots due to their ubiquity
and direct connectivity to the Internet. Other devices
such as network cameras, network storage system,
and all the smart devices that have Internet connec-
tivity and are able to transfer data are also targets of
Thingbots (Sabanal, 2016).
Man in middle attack: This is one of the old at-
tack concepts in the network where an attacker tries
IoTBDS 2017 - 2nd International Conference on Internet of Things, Big Data and Security
340
to interrupt and breach communication between two
nodes. As the attacker has the original communica-
tion, they can trick the recipient into thinking that they
are still getting a legitimate message.
Data and identity theft: Data available from the In-
ternet, social media, smart devices that the person
owns, the smart meters and much more can provide an
overview of the person’s identity, location and other
personal information. These data can be utilised for
commercial purpose or to create personalised attacks.
Denial of service (DoS) attack: A DoS attack can be
initiated from different layers. At the physical layer,
an attacker can jam the signals. At link layer, the at-
tacker can do a capture effect. At the network layer,
the attacker can do selective dropping, routing table
poisoning, table overflow. At transport and session
layer, the attacker can hijack the session, or do a SYN
flood. At the application layer, there are malware to
do DoS attack (Messai, 2014).
4 CURRENT STANDARDS AND
THEIR LIMITATIONS
The core need of IoT is to have a full interoperability
among IoT devices. Practically a full interoperability
is very complex. The concept of connecting any IoT
object to the Internet is one of the biggest standardisa-
tion challenges. An IoT device needs to communicate
with any other IoT device or IoT infrastructure from
various layers of communication protocols at differ-
ent degrees (Rose et al., 2015).
There have been many industrial and academic
attempts to standardise IoT environment and to cre-
ate a secure architecture for IoT. CISCO proposed
a secure framework for securing different layers of
IoT(CISCO, 2013). Ning et al. (Ning et al., 2012)
proposed a secure framework called IPM considering
the informational, physical and management models.
Another example for a secure framework is PubNub
which consist of a secure global Data Stream Network
and APIs that are easy to use for customers(PubNub,
2015). The Internet Engineering Task Force (IETF)
is currently leading the standardisation process of
communication protocols for resource constrained de-
vices, being developed. There are other Internet pro-
tocols such as Routing Protocols for Low Power and
Lossy Networks (RPL) and Constrained Application
Protocol (CoAP) (Group, 2016). From the available
standards choosing the best framework is not easy.
There is no accepted reference architecture among
vendors. As a solution for this many vendors build
their IoT platform from the start. They customise and
design services according to their platform. However,
this approach ends up with more non-interoperable
IoT systems. Such independent development of IoT
platforms leads to less secure choices, gaps in the
standards and lack of an agreed-uponmethods(Group,
2016).
5 WHY IT IS IMPOSSIBLE TO
SOLVE IOT CHALLENGES
The technological shift promised by IoT is very excit-
ing; at the same time, the security vulnerabilities are
worse than what the Internet currently faces. A study
by HP Security unit found that 70 percentage of IoT
devices are hackable (HP, 2014). An IoT device like
smart fridge is as vulnerable as laptops and mobiles.
Users ignorance: Most of the consumers do not know
or do not care to know how IoT devices operate. Even
for a tech savvy person interested inthe security of de-
vice is unable to find more about device’s operations.
Most of the ‘things’ do not provide access to system
information such as its operating system or the soft-
ware version, hardware configuration or any details
about its last update. A person is unable to manu-
ally update, in case he discovers vulnerabilities in his
device. Unlike the laptop and mobile phones, there
is little documentation and tools publicly available to
the check the device security.
Productions: One of the most important question to
solve is, who makes the ‘things’. In a study con-
ducted by Jay Schulman, (Spring, 2016), in search of
an activity tracker on Amazon, he found that there are
thousands of brand names on Amazon. Also, there
were the things that are high rated and perfect prod-
ucts from unknown producers. Later on the investiga-
tion, it was found that these were products from pro-
ducers not much concerned about the security of the
IoT things. Most of the producers are concerned only
about customer service and satisfaction. Many prod-
ucts today are produced by a third party and shipped
to customers.
Unknown vulnerability: Integrating many complex
technologies together bring new unknown vulnera-
bilities. The common man being excited about the
possibilities of IoT use the smart appliances know-
ing nothing about the security implications. In recent,
researchers could show that Samsung’s smart fridge
RF28HMELBSR can be exploited by Man in Middle
attack and could access house owners social media
credentials through the fridges touch screen display
(Schulman, 2016). Hack on Jeep Cherokee well ex-
plains the possibility of a remote attack (Zunnurhain,
Secure IoT: An Improbable Reality
341
2016). The hackers Charlie Miller and Chris Valasek
could remotely control the entire car systems like air-
conditioning, radio, and windshield wipers, accelera-
tor and even brakes.
Cost of production and Patch: Most of the RFID
tags in detecting and identifying things’ are dispos-
able. It is economically infeasible to call back all
the RFID tags on finding a security flaw and to fix
it. On the other hand some devices have vulnerabili-
ties that could not be fixed because they did not have
the ability to be updated. Industries also do not pro-
vide long-term support and a patching solution for
Internet-connected devices that have to be updated
and patched from the upcoming security bugs in the
future.
Architecture: IoT has to be secure by design, but un-
fortunately, security has turned out to be an add-on
feature for IoT (Zunnurhain, 2016). In the current
state of IoT, the developers are trying to solve the
problem by adding patches. This can only increase
the complexity of the system. There are different or-
ganisations striving to make a secure architecture for
IoT. But these attempts are not going together to have
a common standard.
Limited capability of devices: Most IoT devices have
limited computational power and run on batteries.
There is only limited space available in the devices
for additional codes and data. A security researcher
Maxim Rupp found a vulnerability on ESCs 8832
Data Controller, and it was mentioned that the device
had no available code space to add the security patch
according to ICS-CERT 2016.
Limitation of lightweight encryption: Lightweight en-
cryption was suggested as a solution for security on
IoT devices. Symmetric algorithms provide confi-
dentiality, integrity. These algorithms have small
keys and low complexity. But authentication and
key distribution still pose a major problem. Some
symmetric algorithms used for IoT are AES, which
was found to have man in middle attack vulnerabil-
ity (Drozhzhin, 2015). Another algorithm, High se-
curity and lightweight (HIGHT) that use basic opera-
tions such as addition mod 28 or XOR work for Feis-
tel network was found to have saturation vulnerabil-
ity (Luhach et al., 2016). PRESENT is lightweight
encryption based on SPN and is used as an ultra
lightweight algorithm for security, it is vulnerable to
differential attack on 26 of 31 rounds (Derbez and
Fouque, 2013). RC5 that work as a lightweight algo-
rithm for wireless sensors are also vulnerable to dif-
ferential attack. Tiny Encryption Algorithm (TEA)
used in constrained environment are also found vul-
nerable to many attacks (Lee et al., 2010). The asym-
metric algorithms have very large key sizes which
make it more complex to be calculated on devices that
have energy and computation constraints.
Convenience and Ease of usage: Since it is not just
the laptops and mobiles connected to the Internet,
users find it difficult to be concerned about security all
the time. A user would think that a microwave con-
nected to IoT is still a microwave but he does not re-
alise that it can have same vulnerabilities as his laptop
which is connected to the Internet. Daily things ask-
ing for security updates, permissions and passwords
will take out the cool quotient of IoT.
Vertical integration: The integration of technologies
are really complex. Small devices that have their
own software, are integrated into one bigger device.
For example, a car would have several devices from
different manufactures. It will be more complicated
when the device becomes more complex. Its a big
challenge to make sure that all the sub-devices are se-
cure. Even if the individual devices would be secure
it would be difficult to say that its integrated version
is secure.
Lack of standards: Some of the layers of IoT stack
have no standards and on the other side, there are nu-
merous standards from different organisations com-
peting to win. Some of these standards are incompat-
ible with each other. For example, in the connectiv-
ity there are Bluetooth, ZigBee, LTE category 0 stan-
dards. Even if agreed upon a common networking
protocol, then there is an issue of software standards
to contend with. These make the devices impossible
to share a common security protocol. Gartner says
that the divergent number of approaches to solve the
problems will only create security gaps.
6 RELATED WORKS
The security of Internet of Thing has started to get at-
tention in recent years. Kaspersky has regularly writ-
ten about how unexpectedly vulnerable a connected
device can be (Miller and Valasek, 2015). Allen
Storey,Intercede has criticised IoT as “Theres nothing
‘smart’ about insecurely connected devices” in which
he has mentioned about the possible cyber crimes
through IoT (Storey, 2014). Mark et al. did a study
on the vulnerable devices on Internet of Things and
found vulnerability rates ranging from a low of 0.44%
to a high of 40% at various IoT domains (Patton et al.,
2014).
IoTBDS 2017 - 2nd International Conference on Internet of Things, Big Data and Security
342
7 CONCLUSIONS
Though the Internet of things seems to be a revolu-
tionary concept it can’t come to reality without solv-
ing its security issues. The security challenges of IoT
is so vast, and it cannot be solved by current technolo-
gies. Integration of multiple technologies will only
make the system more complex and hence less secure.
We explained various reasons why IoT cannot be se-
cure practically. There are no satisfying solutions for
the security issues of IoT. Even the suggested solu-
tions like lightweight encryption and standardisations
face many criticisms for its efficiency. Hence we con-
clude that IoT can never be secure.
REFERENCES
Ashton, K. (2009). That internet of things thing. RFiD
Journal, 22:97–114.
Barcena, M. B. and Wueest, C. (2015). Insecurity in the
internet of things. Security Response, Symantec.
Burmester, M. and De Medeiros, B. (2007). Rfid security:
attacks, countermeasures and challenges. In The RFID
Journal.
CISCO (2013). Securing the internet
of things: A proposed framework.
http://www.cisco.com/c/en/us/about/security-
center/secure-iot-proposed-framework.html.
Derbez, P. and Fouque, P.-A. (2013). Exhausting demirci-
selc¸uk meet-in-the-middle attacks against reduced-
round aes. In International Workshop on Fast Soft-
ware Encryption, pages 541–560. Springer.
Drozhzhin, A. (2015). Internet of crappy things.
https://kas.pr/at2F.
Farooq, M., Waseem, M., Khairi, A., and Mazhar, S. (2015).
A critical analysis on the security concerns of internet
of things (iot). International Journal of Computer.
Gartner, I. (2014). Gartner says 4.9 billion con-
nected ”things” will be in use in 2015.
http://www.gartner.com/newsroom/id/2905717.
Group, I. W. (2016). Future-proofing the connected world:
13 steps to developing secure iot products.
Hancke, G. et al. (2008). Eavesdropping attacks on high-
frequency rfid tokens. In 4th Workshop on RFID Se-
curity (RFIDSec), pages 100–113.
Hancke, G. P. (2005). A practical relay attack on iso 14443
proximity cards. Technical report, University of Cam-
bridge Computer Laboratory, 59:382–385.
HP (2014). Hpe fortify and the internet of things.
http://go.saas.hpe.com/fod/internet-of-things.
Insider, B. (2016). The master key to understanding the iot
revolution. http://read.bi/29mOi16.
Kibirige, G. W. and Sanga, C. (2015). A survey on detection
of sinkhole attack in wireless sensor network. arXiv
preprint arXiv:1505.01941.
Kurakova, T. (2013). Overview of internet of things. Pro-
ceedings of the Internet of things and its enablers
(INTHITEN), pages 82–94.
Lee, J., Kapitanova, K., and Son, S. H. (2010). The price
of security in wireless sensor networks. In Computer
Networks. Elsevier.
Li, L. (2012). Study on security architecture in the internet
of things. In Measurement, Information and Control
(MIC), 2012 Int. Conference on, volume 1, pages 374
377. IEEE.
Luhach, A. K. et al. (2016). Analysis of lightweight crypto-
graphic solutions for internet of things. Indian Journal
of Science and Technology, 9.
Messai, M.-L. (2014). Classification of attacks in wireless
sensor networks. arXiv preprint arXiv:1406.4516.
Miller, C. and Valasek, C. (2015). Remote exploitation of
an unaltered passenger vehicle. Black Hat USA.
Mitrokotsa, A., Rieback, M. R., and Tanenbaum, A. S.
(2010). Classifying rfid attacks and defenses. Infor-
mation Systems Frontiers, 12(5):491–505.
Ning, H., Liu, H., et al. (2012). Cyber-physical-social based
security architecture for future internet of things. Ad-
vances in Internet of Things.
Patton, M., Gross, E., Chinn, R., Forbis, S., Walker, L., and
Chen, H. (2014). Uninvited connections: a study of
vulnerable devices on the internet of things (iot). In
JISIC, 2014, pages 232–235. IEEE.
Pirretti, M., Z. S. V. N. M. P. K. M. B. R. (2006). The
sleep deprivation attack in sensor networks: Analysis
and methods of defense. International Journal of Dis-
tributed Sensor Networks, 2.
PubNub (2015). A new approach to iot security.
Rose, K., Eldridge, S., and Chapin, L. (2015). The internet
of things: An overview. The Internet Society (ISOC),
pages 1–50.
Sabanal, P. (2016). Thingbots: The future of botnets in the
internet of things. RSA Conference.
Sathyadevan, S., Achuthan, K., and Poroor, J. (2014). Ar-
chitectural recommendations in building a network
based secure, scalable and interoperable internet of
things middleware. In Advances in Intelligent Sys-
tems and Computing, volume 327, pages 429–439.
Springer.
Sathyadevan, S., Kalarickal, B., and Jinesh, M. K. (2015).
Security, trust and implementation limitations of
prominent iot platforms. In Advances in Intelligent
Systems and Computing, volume 328, pages 85–95.
Springer.
Schulman, J. (2016). Why we will never secure the internet
of things. buildingacareerinsecurity.com.
Spring, T. (2016). Iot insecurity pinpointing the problems.
Storey, A. (2014). There’s nothing smartabout insecure con-
nected devices. Network Security, pages 9–12.
Suo, H., Wan, J., Zou, C., and Liu, J. (2012). Security in
the internet of things: a review. In Computer Science
and Electronics Engineering (ICCSEE), Int. Confer-
ence on, volume 3, pages 648–651. IEEE.
Uttarkar, M. R. and Kulkarni, R. (2014). Internet of things:
Architecture and security. International Journal of
Computer Application, pages 12–19.
Zunnurhain, K. (2016). Vulnerabilities with internet of
things. In Proceedings of the Int. Conference on SAM,
page 83.
Secure IoT: An Improbable Reality
343