Content uploaded by Usman Tariq
Author content
All content in this area was uploaded by Usman Tariq on Oct 17, 2023
Content may be subject to copyright.
Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
!.53#2)04
/24)&9).'/!'!).34#2)-0,).'#9"%2!44!#+3!3934%-!4)#2%6)%7/24)&9).'/!'!).34#2)-0,).'#9"%2!44!#+3!3934%-!4)#2%6)%7
3-!.!2)1
2&!.(-%$
5(!--!$44)15%(!.
,)!3()&!3()2
/,,/74()3!.$!$$)4)/.!,7/2+3!4(4403+)*/-35/+%2"!,!%$5)1(/-%
!24/&4(%00,)%$!4(%-!4)#3/--/.3)/).&/2-!4)#3/--/.3)/,/'9/--/.3(%-)3429/--/.3
/-054%2#)%.#%3/--/.3(93)#3/--/.3!.$4(%934%-3!.$/--5.)#!4)/.3/--/.3
/24)&9).'/!'!).34#2)-0,).'#9"%2!44!#+3!3934%-!4)#2%6)%7/24)&9).'/!'!).34#2)-0,).'#9"%2!44!#+3!3934%-!4)#2%6)%7
"342!#4"342!#4
(%2!0)$'2/74(!.$).#2%!3).'$%-!.$&/2.4%2.%4/&().'3/$%6)#%3)./52%6%29$!9,)6%3#2%!4%
%8#)4).'/00/245.)4)%3&/2(5-!.).6/,6%-%.4$!4!).4%'2!4)/.!.$3%!-,%33!54/-!4)/.()3&5,,9
).4%2#/..%#4%$%#/3934%-#/.3)$%2!",9)-0!#43#25#)!,!30%#43/&/52,)6%335#(!342!.30/24!4)/.
(%!,4(#!2%%.%2'9-!.!'%-%.4!.$52"!.).&2!3425#452%/7%6%2!,/.'3)$%4(%)--%.3%"%.%;434(%
7)$%302%!$!$/04)/./&/!,3/"2).'3!#/-0,%87%"/&3%#52)494(2%!434(!4#!.).<5%.#%3/#)%49
0/,)#9!.$).&2!3425#452%#/.$)4)/.3/$%6)#%3!2%0!24)#5,!2,965,.%2!",%4/3%#52)496)/,!4)/.3!.$
).$5342)!,2/54).%3&!#%0/4%.4)!,,9$!-!').'65,.%2!"),)4)%3/%.352%!425347/24(9!.$2/"5343%#52)49
&2!-%7/2+)4)3#25#)!,4/4!#+,%4(%$)6%23%#(!,,%.'%3).6/,6%$()33526%90!0%2!)-34/!)$
2%3%!2#(%23"9#!4%'/2):).'!44!#+3!.$65,.%2!"),)4)%3"!3%$/.4(%)24!2'%43402/6)$%3!$%4!),%$
!.!,93)3/&!44!#+-%4(/$3!.$02/0/3%3%&&%#4)6%#/5.4%2-%!352%3&/2%!#(!44!#+#!4%'/29(%0!0%2
!,3/()'(,)'(43#!3%345$)%3/)4)#!,/!00,)#!4)/.33(/7#!3).'3%#52)493/,54)/.3.!$$)4)/.4/
42!$)4)/.!,#2904/'2!0()#!002/!#(%34()37/2+%80,/2%3%-%2').'4%#(./,/')%3,)+%5!.45-2904/
(93)#!,.#,/.!",%5.#4)/.33!.$",/#+#(!).$)3#533).'4(%)202/3!.$#/.3).3%#52).'/
%.6)2/.-%.43(%2%3%!2#()$%.4);%3!.$%8!-).%3!44!#+365,.%2!"),)4)%3!.$3%#52)49-%!352%3!.$
%.$%!6/234/)-0!#44(%/6%2!,,5.$%234!.$).'/&/3%#52)49(%).3)'(43!.$;.$).'302%3%.4%$(%2%
7),,3%26%!3!6!,5!",%2%3/52#%&/22%3%!2#(%23'5)$).'4(%$%6%,/0-%.4/&2%3),)%.43%#52)49
-%#(!.)3-34/%.352%4(%425347/24(9!.$3!&%/0%2!4)/./&/%#/3934%-3
%97/2$3%97/2$3
.4%2.%4/&().'39"%2%#52)49./-!,9%4%#4)/.934%-!4)#)4%2!452%%6)%7!#().%%!2.).'
,/#+#(!).
2%!4)6%/--/.3)#%.3%2%!4)6%/--/.3)#%.3%
()37/2+)3,)#%.3%$5.$%2!2%!4)6%/--/.3442)"54)/./.#/--%2#)!,/%2)6!4)6% /2+3
)#%.3%
REVIEW ARTICLE
Fortifying IoT Against Crimpling Cyber-attacks: A
Systematic Review
Usman Tariq
a,
*, Irfan Ahmed
b
, Muhammad Attique Khan
c
, Ali Kashif Bashir
d
a
Department of Management Information Systems, College of Business Administration, Prince Sattam Bin Abdulaziz University, Al-
Kharj, 16278, Saudi Arabia
b
Department of Computer Science, College of Engineering, Virginia Commonwealth University, Richmond, VA, 23284, USA
c
Department of Computer Science, HITEC University, Taxila, Pakistan
d
Department of Computing and Mathematics, Manchester Metropolitan University, Manchester, M156BH, UK
Abstract
The rapid growth and increasing demand for Internet of Things (IoT) devices in our everyday lives create exciting
opportunities for human involvement, data integration, and seamless automation. This fully interconnected ecosystem
considerably impacts crucial aspects of our lives, such as transportation, healthcare, energy management, and urban
infrastructure. However, alongside the immense benefits, the widespread adoption of IoT also brings a complex web of
security threats that can influence society, policy, and infrastructure conditions. IoT devices are particularly vulnerable to
security violations, and industrial routines face potentially damaging vulnerabilities. To ensure a trustworthy and robust
security framework, it is crucial to tackle the diverse challenges involved. This survey paper aims to aid researchers by
categorizing attacks and vulnerabilities based on their targets. It provides a detailed analysis of attack methods and
proposes effective countermeasures for each attack category. The paper also highlights case studies of critical IoT ap-
plications, showcasing security solutions. In addition to traditional cryptographic approaches, this work explores
emerging technologies like Quantum Crypto Physical Unclonable Functions (QC-PUFs) and blockchain, discussing their
pros and cons in securing IoT environments. The research identifies and examines attacks, vulnerabilities, and security
measures and endeavors to impact the overall understanding of IoT security. The insights and findings presented here
will serve as a valuable resource for researchers, guiding the development of resilient security mechanisms to ensure the
trustworthy and safe operation of IoT ecosystems.
Keywords: Internet of things, Cyber security, Anomaly detection, Systematic literature review, Machine learning (ML),
Blockchain
1. Introduction
The realm of IoT security encompasses a broad
range of strategies, tools, processes, systems,
and methods aimed at safeguarding the entirety of
the Internet of Things. It involves protecting phys-
ical components, applications, data, and network
fixtures to guarantee the availability, integrity, and
confidentiality of IoT ecosystems. Security chal-
lenges are abundant due to the continuous discov-
ery of numerous vulnerabilities within IoT systems.
Robust IoT security entails a holistic approach to
protection, encompassing measures such as
component hardening, continuous monitoring,
firmware updates, access management, proactive
threat response, and active vulnerability remedia-
tion. The significance of IoT security cannot be un-
derstated, as these sprawling and vulnerable
systems represent highly attractive attack vectors.
IoT security vulnerabilities are pervasive across a
wide range of domains, including vehicles, smart
grids, watches, and smart home devices. For
instance, researchers have identified webcams with
glaring security flaws [1], easily exploitable for un-
authorized network access. Similarly, smartwatches
have been found to harbor vulnerabilities enabling
Received 6 July 2023; revised 21 August 2023; accepted 24 August 2023.
Available online 18 October 2023
*Corresponding author.
E-mail address: u.tariq@psau.edu.sa (U. Tariq).
https://doi.org/10.33640/2405-609X.3329
2405-609X/©2023 University of Kerbala. This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
hackers to track wearers'locations and eavesdrop on
their conversations [2]. These examples underscore
the pressing need to comprehensively and proac-
tively address IoT security concerns.
1.1. IoT security challenges
Securing IoT environments presents many chal-
lenges owing to the distinctive characteristics of IoT
devices and systems. The absence of standardized
regulations and a lack of awareness regarding
inherent risks significantly compound the
complexity of IoT security. Key challenges encom-
pass limited visibility and control over deployed
devices, the intricate integration of diverse IoT de-
vices into existing security frameworks, vulnerabil-
ities stemming from open-source code in firmware,
the sheer volume of data generated by IoT systems,
inadequate vulnerability testing practices, unre-
solved vulnerabilities, susceptible APIs, and the
prevalence of weak passwords. Effectively address-
ing these challenges necessitate implementing spe-
cific security measures, including API security,
meticulous device inventory management, contin-
uous software updates, robust encryption for data at
rest and in transit, multi-factor authentication,
comprehensive network security provisions, and
diligent vulnerability patching.
1.2. IoT adoption use case
In today's world, the widespread adoption of IoT
devices necessitates implementing robust measures
for connectivity, management, and control. To ach-
ieve this, three fundamental steps are essential.
Firstly, the process of discovering and classifying
each connected object enables rapid identification
and automated provisioning based on device type
and the application of appropriate configuration
policies. Secondly, the network infrastructure can be
effectively segmented into dedicated virtual net-
works, ensuring the separation of services and ap-
plications to optimize functionality and enhance
security. Lastly, continuous monitoring of device
behaviors enables real-time inventory management
and prompt response in the event of deviations. By
adhering to these steps, cybersecurity researchers
can significantly enhance the usability of their IoT
devices, promoting efficient operation, timely
detection of anomalies, and proactive device man-
agement and security practices.
Multiple fundamental elements are essential for
enabling the functionality of IoT. Identification, a
crucial element, plays a significant role in naming and
matching services to their respective demands. IoT
devices utilize sensing capabilities to capture data and
transmit it to the cloud or databases for analysis.
Communication serves as a binding force, enabling
seamless interaction among diverse objects to provide
targeted digital services. Various communication
protocols [3] such as Wireless Fidelity (WiFi), Blue-
tooth, Zigbee, Message Queuing Telemetry Transport
(MQTT), Institute of Electrical and Electronics Engi-
neers (IEEE) 802.15.4, Object Linking and Embedding
for Process Control Unified Architecture (OPC-UA),
Near Field Communication (NFC), Z-wave, Long
Range Wide Area Network (LoRaWAN), SigFox, and
Long-Term Evolution Advanced (LTE-Advanced) are
utilized for facilitating these interactions. Hardware
components, including microcontrollers, micropro-
cessors, Field-programmable gate arrays (FPGAs),
and system-on-chip (SoCs) handle processing tasks,
while software functions and processing systems form
the intelligent core of IoT. The eventual objective of
IoT is to render services accessible anytime, any-
where, and to anyone.
1.3. Impact of device specification in IoT anomaly
detection
With reference to Table 1, it is evident that the
specifications of IoT devices and network infra-
structure significantly impact the capability of
anomaly detection and the effectiveness of cyber
defense in IoT systems. The CPU clock speed and
cache size of devices determine their processing
power and ability to manage real-time anomaly
detection algorithms. A higher clock speed and
larger cache enable faster processing and analysis of
data, improving the responsiveness of anomaly
detection systems. The availability of sufficient RAM
and flash memory allows for storing and processing
large volumes of data, facilitating comprehensive
anomaly detection, and enhancing the system's de-
fense capabilities. The presence of cameras and
audio/video support enables the capture and anal-
ysis of multimedia data, enriching the anomaly
detection process. Supported protocols play a vital
role in facilitating communication and data ex-
change between IoT devices and the detection sys-
tem, enhancing the system's ability to monitor and
identify anomalies. The instruction size, available
registers, memory access type, and instruction set
architectures influence the execution efficiency and
computational capabilities of anomaly detection al-
gorithms. Compliance with applicable IoT stan-
dards ensures interoperability and compatibility,
enabling seamless integration of different devices
and systems for a robust cyber defense mechanism
in IoT environments. Hence, careful consideration
666 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
of IoT device and network specifications is crucial
for developing effective anomaly detection systems
and strengthening the overall cybersecurity posture
in IoT deployments.
1.4. Generalized IoT layered-architecture
The generalized architecture of an IoT system
entails four layers: Perception, Network, Processing,
and Application. In the Perception layer, nodes such
as conveyor systems, surveillance cameras, Global
Positioning System (GPS) modules, Radio-fre-
quency Identification (RFID) scanners, and
manufacturing robots are liable for supervising the
settings and aggregation of sensory data. The
Network layer comprises communication systems
[4] like WiFi, Bluetooth, Zigbee, Long-Term Evolu-
tion (LTE), and protocols like Internet Protocol
version 4 (IPv4) and Internet Protocol version 6
(IPv6), facilitating the transfer of data to the subse-
quent Processing layer. Within the Processing layer,
cloud servers and databases handle tasks such as
data analysis, computation, decision-making, and
the storage of vast amounts of information. Ulti-
mately, the Application layer caters to the distinctive
requirements of end-users, delivering tailored ser-
vices based on their requirements and preferences.
1.5. Deviations or anomalies in IoT setting
Anomaly detection focuses on data, device &
network changes, revealing previously unknown
threats and communication patterns that have not
yet been documented in threat databases or operate
covertly, causing gradual shifts. By analyzing exist-
ing IoT devices, network communication and
infrastructure, an effective anomaly detection can
provide administrators with a comprehensive
network mapping that offers valuable insights. This
includes identifying devices and clients within the
network, establishing connections and hierarchies
between devices (such as master/slave relation-
ships), monitoring data packets and their content,
identifying utilized protocols, and analyzing the
frequency of specific communication patterns.
Through this approach, IoT enabler-establishments
can become eligible to effectively detect and address
deviations or anomalies, enabling proactive mitiga-
tion of potential security risks and ensuring the
stability and integrity of the network.
1.6. Research contributions
This research paper significantly contributes to
the domain knowledge of IoT vulnerabilities,
Table 1. IoT Device cataloging with reinforced computer peripherals.
Devices CPU Clock Cache RAM Flash Supported Protocols Instruction Size Available
Registers
Memory
Access Type
Instruction Set
Architectures
Raspberry Pi 4 Broadcom BCM2711 1.5 GHz 512 KB 4 GB 16 GB Wi-Fi, Bluetooth 32 bits 16 Memory Mapped ARMv8-A
Arduino Uno Atmel ATmega328P 16 MHz 2 KB 2 KB 32 KB UART, I2C 16 bits 32 Memory Mapped AVR
ESP32 Tensilica Xtensa LX6 240 MHz 512 KB 520 KB 4 MB Wi-Fi, Bluetooth 32 bits 16 Memory Mapped Xtensa LX6
BeagleBone
Black
Texas Instruments
AM335x
1 GHz 256 KB 512 MB 4 GB Ethernet, UART 32 bits 32 DDR3 ARMv7-A
NVIDIA
Jetson Nano
Quad-core ARM
Cortex-A57
1.43 GHz 2 MB L2 4 GB 16 GB eMMC Ethernet, USB 64 bits 32 LPDDR4 ARMv8-A
Intel Edison Intel Atom 500 MHz 512 KB L2 1 GB 4 GB eMMC Wi-Fi, Bluetooth 32 bits 16 DDR3 x86
Particle Argon Nordic Semiconductor 64 MHz 256 KB 128 KB 1 MB Wi-Fi, Bluetooth 32 bits 16 Memory Mapped ARM Cortex-M4
Microchip PIC32 MIPS32 M4K 80 MHz 32 KB 128 KB 512 KB UART, SPI, I2C 16 bits 32 Memory Mapped MIPS32
Adafruit Feather M0 Atmel SAMD21 48 MHz 256 KB 32 KB 256 KB UART, SPI, I2C 32 bits 16 Memory Mapped ARM Cortex-M0þ
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 667
anomalies, risks, threats, and security features.
Firstly, it provides a comprehensive review of the
existing literature, synthesizing the current under-
standing of IoT security issues and highlighting the
key vulnerabilities and threats that can conciliate
the integrity and confidentiality of IoT systems. This
review serves as a valuable resource for researchers
and practitioners seeking to gain a holistic under-
standing of the security challenges in the IoT
domain.
Secondly, the paper presents the results of a
comprehensive system generated anomaly spread
and identification survey conducted among experts
and practitioners in the field. The survey data offers
insights into real-world experiences and practices
regarding IoT security, shedding light on the most
common anomalies and risks encountered. By
analyzing the survey responses, the research paper
identifies trends, patterns, and emerging concerns
in the IoT security backdrop, informing future
research directions and best practices.
Furthermore, the research paper proposes a sys-
tematic framework for assessing and mitigating IoT
vulnerabilities and risks. It provides a structured
approach that the cybersecurity scientific commu-
nity can adopt to identify potential threats, evaluate
their impact, and implement appropriate security
measures. The framework considers both technical
aspects (such as encryption, authentication, and
access control) and non-technical factors (such as
policy and governance) to create a comprehensive
security strategy.
Lastly, the paper evaluates the effectiveness of
existing security features and protocols in mitigating
IoT risks. It examines the strengths and limitations
of commonly used security mechanisms and pro-
poses enhancements to address the identified gaps.
By critically assessing the current state of security
features, the research paper guides the develop-
ment and implementation of more robust and
resilient security solutions for IoT environments.
We, the authors, believe that the findings of this
research paper will empower IoT infrastructure
handlers and researchers to better understand,
address, and mitigate the vulnerabilities and risks
associated with IoT deployments.
1.7. Paper organization
Section 2provides a detailed explanation of the
‘Procedural Research Method’and its rationale for
utilizing a systematic methodology in evaluating
security challenges within the jurisdiction of IoT.
This approach aims to establish a comprehensive
threat taxonomy to thoroughly understand the
subject matter. In Section 3, the research delves into
the ‘Security background, terminology, and objec-
tives', offering valuable insights into the contextual
aspects of the study. Section 4presents ‘Key
Applicable Recommendations', proposing light-
weight, scalable, and effective cyber-shields for IoT
cyber-defense. Finally, Section 5serves as the
conclusion, highlighting the study's limitations and
providing recommendations for future research di-
rections. This sequential organization of sections
ensures a coherent and logical flow of information,
enhancing the overall structure and readability of
the research paper.
2. Procedural research method
This research paper employed a systematic
methodology to assess the security challenges in the
realm of IoT and construct a comprehensive threat
taxonomy. A meticulous literature search was con-
ducted using pertinent keywords, such as “IoT”and
“security,”across renowned publication databases
including Elsevier, IEEE, ACM, Springer, IET,
MDPI, Wiley and etc. This process yielded a wide
array of survey papers, providing a solid foundation
for comprehending the IoT security landscape.
Leveraging their own expertise in the security field
[5], the authors carefully examined and selected
pertinent topics crucial to network security. Fig. 1
illustrates the research dispersal data with respect to
subject area. Investigating from diverse sources of
research journals and considering diverse subjects
is important for IoT cyber security research survey
as it allows for a comprehensive and multidisci-
plinary understanding of the complex challenges
and potential solutions. Drawing from various fields
such as computer science, decision science, engi-
neering, material science, mathematics, machine
learning, artificial intelligence (AI), and more en-
sures a holistic approach to address the multifaceted
aspects of IoT security, fostering innovative ideas
and robust methodologies to mitigate risks and
safeguard IoT systems effectively. In our examina-
tion of dissimilar procedural research methods,
including Scoping Review, Integrative Review,
Realist Review, and Quantitative Synthesis, we
aimed to ensure the effectiveness and accuracy of
our furnished analysis. Upon careful consideration,
we found that the gold standard in evidence syn-
thesis is a systematic review. This method provides
a rigorous and transparent approach to gather,
assess, and analyze existing literature, following a
predefined protocol to minimize bias and ensure
reproducibility. Contrasting the Scoping Review,
which focuses on mapping available literature to
668 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
identify key concepts and gaps but lacks the depth
and comprehensive analysis, Integrative Review
synthesizes diverse research methodologies but
may not exhibit the same level of methodological
rigor. Realist Review, while exploring underlying
mechanisms, may be limited in generalizability
compared to the broad scope of a systematic review.
Whereas, Quantitative Synthesis may lack the
qualitative depth and context provided by the
systematic review, which thoroughly examines both
quantitative and qualitative evidence. Overall, the
meticulous and comprehensive approach of the
systematic review ensures the generation of robust
and reliable findings.
To gain a deeper understanding of the primary
studies and identify shared patterns, an analysis was
conducted to examine the prevalence, frequency, and
occurrence of keywords across the complete set of
Fig. 1. Research distribution statistics.
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 669
studies. This comprehensive examination aimed to
identify recurring themes and topics within the
selected studies and shed light on the focus and
emphasis of the research. The findings, illustrated in
Table 2, revealed a growing interest and emphasis on
safeguarding IoT devices against cyber-attacks,
providing valuable insights into the prevailing trends
and interests in this area of research. This analysis
provides a broader perspective and highlights signif-
icant aspects related to the primary topic of interest.
It is evident that the IoT has witnessed a
remarkable growth trajectory, with a substantial
number of connected devices already in use and an
anticipated doubling by 2025 (i.e., as illustrated in
Fig. 2)[6,7]. This rapid expansion has resulted in
diverse applications across various domains, such as
the Industrial Internet of Things (IIoT), facilitating
enhanced communication and optimization of pro-
duction processes among machines. Another note-
worthy domain is the Internet of Medical Things
(IoMT), focusing on healthcare applications like
remote patient monitoring and personalized health
tracking. Likewise, IoT plays a pivotal role in the
evolution of Smart Cities, enabling efficient man-
agement of traffic and waste disposal while
unlocking the potential of data-driven governance.
The integration of IoT devices in smart homes has
also revolutionized daily life, with interconnected
appliances such as thermostats, televisions, and se-
curity systems. However, the widespread deploy-
ment of IoT devices raises legitimate concerns
regarding security and privacy, necessitating the
implementation of robust measures to mitigate risks
and safeguard sensitive data.
3. Security background, terminology, and
objectives
In the perspective of IoT security, several key
objectives play a crucial role in safeguarding IoT
devices and systems. These objectives include
integrity, authentication, confidentiality, privacy,
availability, authorization, non-repudiation, identi-
fication, reliability, freshness, access control
methods, and soundness. Ensuring data integrity is
essential to prevent unauthorized modification or
destruction of data during transmission, storage,
and processing. Authentication and authorization
are crucial in verifying the identities of entities
within the IoT system and ensuring that they have
the appropriate permissions to access resources.
Confidentiality protects sensitive information from
unauthorized disclosure, both during transference
and storage. Privacy becomes a significant concern
in handling and processing data, ensuring that the
rights of individuals regarding the use of personal
information are respected. Availability focuses on
the system's operational state and capability to
deliver required services. Additionally, non-repu-
diation prevents entities from denying their actions,
enabling the resolution of potential conflicts within
the system. These security objectives highlight the
comprehensive measures required to address the
unique challenges and vulnerabilities posed by IoT
systems. Table 3 provides an overview of IoT secu-
rity objectives, corresponding attack types, and
anomaly detection techniques. It offers insights into
the layers involved, ML methods utilized, detection
accuracy, datasets used, and relevant references for
further exploration. Incorporating quantitative
analysis into our research on IoT security was
crucial to providing a data-driven and objective
assessment of the identified vulnerabilities and po-
tential solutions. By utilizing a relevant dataset
encompassing various aspects of IoT security, such
as the frequency and types of cyber-attacks, the
effectiveness of different cryptographic algorithms,
and the performance of existing security protocols,
we could derive valuable insights and meaningful
conclusions. This quantitative analysis enabled us to
measure the impact and significance of security
measures, identify trends and patterns in cyber-at-
tacks, and assess the overall effectiveness of IoT
security strategies.
Table 4 consolidates and expands crucial infor-
mation on ML models employed in the context of
IoT security for detecting and mitigating cyber-at-
tacks. It is evident that the ML models leverage
various techniques like anomaly detection, behavior
analysis, pattern recognition, and signature-based
detection to identify and classify abnormal activities
and malicious patterns within the IoT environment.
The advantages of these models include high accu-
racy, real-time detection, scalability, and adapt-
ability to evolving attack patterns. However, they
also face challenges such as false positives, compu-
tational complexity, dataset requirements, and sus-
ceptibility to adversarial attacks.
3.1. Machine learning for IoT security
In reference to Tables 3 and 4, the machine
learning and deep learning techniques, derived
from artificial intelligence, play a vital role in
detecting malware and malevolent network traffic
within IoT systems. Traditional attack discovery
systems rely on predefined strategies and feature
sets to identify and classify network attacks, result-
ing in limitations when it comes to detecting new
attack types and being restricted to specific
670 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
scenarios [14]. However, this limitation can be
overcome by employing ML algorithms that learn
from previous experiences rather than relying on
predetermined rules. Recent research [4,8,11e13]
has successfully applied and supported the efficacy
of ML algorithms in bolstering IoT security. These
studies demonstrate that machine learning algo-
rithms can adapt to the dynamic behaviors of IoT
Table 2. Analysis of keywords in primary studies.
Keywords Count Keywords Count
Internet of Things (IoT) 724 Cyber-attacks 609
Wireless Sensor Networks (WSNs) 427 Denial-of-Service (DoS) 578
Secure Routing 255 Trustworthiness 651
Edge Computing 310 Cloud Computing 318
Scalability 645 Machine Learning 595
Artificial Intelligence 607 Deep Learning 487
Neural Networks 496 Supervised Learning 477
Unsupervised Learning 545 Reinforcement Learning 268
Dimensionality Reduction 271 Feature Extraction 542
Transfer Learning 732 Model Selection &Evaluation 445
Autoencoder 273 Ensemble Learning 257
Convolutional Neural Networks 335 Adversarial Networks 620
Data Augmentation 421 Active Learning 715
Anomaly Detection 732 Semi-Supervised Learning 319
Lightweight Cryptography 246 Encryption and Decryption 715
Authentication 543 Key Exchange 615
Secure Communication 677 Resource-Constrained Devices 609
Energy Efficiency 405 Memory Efficiency 307
Hardware &Software Implementation 246 Lightweight Authentication Protocol 294
Resistance to Differential Power Analysis (DPA) 212 Performance Analysis of Lightweight Cryptographic Algorithms 215
Blockchain 725 Distributed Ledger 494
Consensus Mechanisms 483 Decentralization 357
Peer-to-Peer (P2P) Network 570 Immutable 369
Transparency 231 Private/Public/Permissioned Blockchain 699
Proof of Work (PoW) 377 Proof of Stake (PoS) 573
Proof of Authority (PoA) 716 Byzantine Fault Tolerance (BFT) 287
Scalability 398 Interoperability 624
Cross-Chain Communication 606 Zero-Knowledge Proofs 501
Blockchain Governance 316 Blockchain Intermediaries 492
Consensus Algorithms 547 Blockchain Adoption 623
Blockchain IoT Security Use Cases 298 Applied Protocols 674
Fig. 2. IoT market size (year 2018e2025).
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 671
systems without compelling manual intrusion. By
continuously monitoring network behavior, ma-
chine learning algorithms can swiftly detect various
IoT attacks at an early stage, making them well-
suited for IoT devices with limited resources.
3.1.1. Supervised ML
Supervised ML algorithms play a crucial role in
accomplishing specific tasks by training ML models
using a learning procedure and a training dataset.
These algorithms classify the output based on the
acquired training knowledge. Supervised learning
involves two primary processes: classification and
regression. Classification algorithms excel in catego-
rizing output based on input data, enabling tasks such
as determining the authenticity of information or
distinguishing between real and fake entities. Promi-
nent supervised ML classifiers include Support Vector
Machine (SVM) [15], Naive Bayes (NB) [16], K-Nearest
Neighbor (KNN) [17], and Random Forest (RF) [18].
SVM, for example, has gained significant adoption in
the field of IoT security, effectively classifying diverse
attacks such as DoS/DDoS, privacy fortification, IoT
botnet recognition, and encoding attacks. Although
SVM demonstrates high classification accuracy, it has
limitations, such as a propensity for over-
generalization, deliberate convergence rapidity, and
sensitivity to local extrema.
Another widely used supervised ML-based classi-
fication algorithm in IoT security is Random Forest
(RF). RF constructs a collection of decision trees, and
the classification and prediction accuracy improve as
the number of trees in the model increases. RF has
been successfully employed in various IoT security
tasks, including irregularity recognition, user to root
intrusion detection, and remote to local risk strike
discovery. Nevertheless, it is important to note that
exceeding a certain number of trees can adversely
impact RF's performance, rendering it slower and
less suitable for real-time classification operations.
The K-Nearest Neighbor (KNN) algorithm calculates
the Euclidean distance between nodes, allowing the
prediction of unknown nodes based on the average
value of their k-nearest neighbors. In IoT applica-
tions, KNN has found utility in tasks such as malware
detection, anomaly detection, and intrusion detec-
tion. Although KNN offers advantages in terms of
ease, cost-efficacy, and compliant execution, its per-
formance may be compromised with larger datasets,
and it proves to be overly vulnerable to outliers and
overlooked values.
Regression algorithms, including Decision Trees
(DTs) [19], Linear Regression (LR) [20], and Neural
Networks (NNs) [21], play a vital role in investi-
gating relationships between independent features
Table 3. IoT security objectives and anomaly detection techniques.
Security
Objectives [5]
Layer [7] Anomaly and
Attacks
Attack
Type
Anomaly
Detection
ML Method to
Detect Anomaly [8]
Detection
Accuracy
Dataset
Integrity Application Data modification Active Statistical analysis Support Vector Machines 95% IoT-23 [9]
Authentication Network Unauthorized access Active Rule-based analysis Decision Trees 90% e
Confidentiality Perception Eavesdropping Passive Encryption analysis Neural Networks 92% e
Privacy Sensing Data leakage Passive Behavior-based analysis Random Forest 88% CIC IoT
Dataset 2022 [10]
Availability Application Denial of Service Active Traffic analysis K-nearest neighbors 96% e
Authorization Network Unauthorized
resource access
Active Role-based analysis Naive Bayes 87% e
Non-repudiation Application Transaction dispute Passive Signature analysis Hidden Markov Models 93% e
Identification Perception Device spoofing Active Pattern recognition Convolutional Neural Networks 91% IoT-23 [9]
Reliability Sensing Data corruption Active Outlier analysis Isolation Forest 94% e
Freshness Network Replay attack Passive Time-stamp analysis Long Short-Term Memory 89% e
Access Control Methods Application Unauthorized privilege escalation Active Rule-based analysis Decision Trees 90% e
Soundness Perception Impersonation attack Active Behavioral profiling Support Vector Machines 93% IoT-23 [9]
672 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
Table 4. IoT cyber attack detection and mitigation with machine learning models.
IoT Cyber
Attack Type [5]
ML Model
Name [11e13]
Category Brief Description Detection Mechanism Advantages Disadvantages and
Limitations
Denial of Service
(DoS)
LSTM-based Classifier Anomaly
Detection
Long Short-Term Memory
(LSTM)
model to detect DoS
attacks
Analyzes network traffic
patterns and behavior
anomalies
Ability to capture sequen-
tial dependencies in data,
high detection accuracy
May require large
amounts of training data,
potential false positives/
negatives
Man-in-the-Middle
(MitM)
Random Forest
(RF)
Supervised
Learning
RF classifier to identify
and prevent
MitM attacks
Analyzes network traffic
and identifies suspicious
activities
High accuracy, can handle
large feature spaces,
interpretability
Limited in handling dy-
namic or evolving attack
patterns, may require
frequent retraining
Device Spoofing Support Vector Ma-
chine
(SVM)
Supervised
Learning
SVM model for detecting
device
spoofing based on behav-
ioral analysis
Analyzes device behavior
and compares with known
patterns
Can handle complex
feature spaces, good
generalization capability
Vulnerable to noise in
training data, may strug-
gle with detecting sophis-
ticated spoofing
techniques
Data Tampering Deep Belief Networks
(DBN)
Unsupervised
Learning
DBN model for detecting
anomalies and
identifying data
tampering
Compares data patterns
and identifies deviations
Good at detecting un-
known attacks, can handle
complex data structures
Requires significant
computational resources,
may have high training
time and complexity
Eavesdropping Convolutional Neural
Network
(CNN)
Deep
Learning
CNN-based model to
detect eavesdropping
by analyzing network
traffic
Extracts features from
network data and detects
anomalies
Effective in capturing
spatial patterns, high
detection accuracy
Requires large, labeled
datasets, may struggle
with detecting advanced
eavesdropping techniques
Replay Attacks Recurrent Neural
Network
(RNN)
Deep
Learning
RNN-based model to
detect and
prevent replay attacks
Analyzes message time-
stamps and detects
replayed messages
Can capture temporal de-
pendencies, effective in
detecting repeated mes-
sage patterns
Reliance on accurate
timestamp synchroniza-
tion, may require contin-
uous monitoring and
synchronization of devices
Malware Injection Decision Tree Supervised
Learning
Decision tree model for
identifying and
blocking malware
injection
Analyzes network traffic
and identifies malicious
patterns
Interpretable model, can
handle both numeric and
categorical data, relatively
low computational
requirements
May struggle with com-
plex data relationships,
may have limitations in
handling unknown or
evolving malware variants
Insider Threats Recurrent Neural
Network
(RNN)
Deep
Learning
RNN model to detect
anomalous behavior
and identify insider
threats
Analyzes user activity
patterns and identifies
anomalies
Ability to capture sequen-
tial dependencies, effec-
tive in detecting subtle
insider behavior
Dependence on accurate
and representative
training data, may require
continuous monitoring
and profiling of user
behavior
(continued on next page)
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 673
and dependent variables for predictive modeling.
DTs utilize simple decision rules derived from
extracted features to predict target variable values
but are prone to instability and struggle with
continuous variables. LR models estimate accurate
constraints by minimizing the error between pre-
dicted and actual values, yet they are sensitive to
outliers and assume linear relationships. NNs,
inspired by human intelligence, control complex
and nonlinear information efficiently, but their
computational complexity poses challenges for
implementation in resource-constrained IoT sys-
tems. To overcome limitations, Ensemble Learning
(EL) [22] combines multiple algorithms to improve
performance, making it a valuable tool for complex
IoT problems such as network monitoring, attack
detection, and anomaly detection. Careful consid-
eration of model selection, optimization techniques,
and ensemble strategies is necessary to harness the
potential of these algorithms in IoT applications.
3.1.2. Unsupervised ML algorithms
Unsupervised ML algorithms can realize hidden
models and examine unlabeled datasets without
relying on training data. By evaluating relationships
between dataset models and input variables, these
algorithms group samples into clusters, thus
enhancing the discretion and protection of IoT de-
vices. Among the extensively sourced unsupervised
ML procedures, K-means [23] effectively clusters ob-
jects into distinct groups based on their nearest mean,
making it suitable for IoT systems. Principal Compo-
nent Analysis (PCA) [24] serves as a dimensionality
reduction technique, improving computational speed
and feature selection for attack detection in IoT,
although it assumes linearity and is sensitive to out-
liers. Hierarchical clustering [25] creates a hierarchy of
clustered data samples, eliminating the need for a
predefined number of clusters, but it struggles with
mixed data types and large-scale datasets. Fuzzy K-
means Clustering (FCM) [26] utilizes fuzzy logic to
assign data points probabilities for cluster member-
ship, offering a more flexible approach than tradi-
tional clustering techniques. Gaussian Mixture
Models (GMMs) [27] assume that data models are
spawned from a mixture of Gaussian distributions,
employing a probabilistic methodology for soft clus-
tering, where each cluster corresponds to a probability
distribution in a multidimensional space.
3.1.3. Reinforcement learning (RL)
RL algorithms enable autonomous learning and
decision-making in systems through interaction
with the environment. The incorporation of Quality-
learning mechanisms [28] (i.e., Distributed Q-
Table 4. (continued)
IoT Cyber
Attack Type [5]
ML Model
Name [11e13]
Category Brief Description Detection Mechanism Advantages Disadvantages and
Limitations
Physical Attacks Support Vector Ma-
chine
(SVM)
Supervised
Learning
SVM model for detecting
physical attacks
on IoT devices
Analyzes sensor data and
identifies abnormal phys-
ical events
Ability to handle complex
feature spaces, good
generalization capability
Limited by the availability
of labeled physical attack
data, may require fine-
tuning for specific phys-
ical attack scenarios
Traffic Manipulation Deep Reinforcement
Learning
(DRL)
Reinforcement
Learning
DRL-based model for
detecting and mitigating
traffic manipulation
Analyzes network traffic
and learns optimal defen-
sive strategies
Adaptive and self-learning
model, can respond to
changing attack patterns
High computational re-
quirements, complex
training process, potential
for suboptimal policy
convergence
674 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
learning, Double Q Network, Dueling Q Network
(DQN), etc.) in RL models allows for automatic de-
cision-making without prior knowledge. RL oper-
ates dynamically, employing a trial-and-error
approach to identify optimal actions for maximizing
rewards. RL algorithms, such as Quality-learning
and Deep Quality Network (DQN) [29], are utilized
for security attack detection in IoT systems. These
RL algorithms address limitations in conventional
ML techniques, including high computational time,
large parameter requirements, lower accuracy, and
the inability to oversee complex problems.
Conversely, RL encounters challenges associated
with computational overload due to the significant
amount of data required for computation.
3.1.4. Future directions for utilizing ML in anomaly
detection
This section addresses the existing research gap
by presenting potential areas of further investiga-
tion, focusing on enhancing the intelligence and
dynamism of protocols by utilizing ML techniques.
(a) In the context of anomaly detection in IoT, future
research should focus on developing ML models
that are explainable and interpretable. This will
enable stakeholders to understand the underlying
reasons for anomaly detections and build trust in
the system. Techniques such as rule extraction,
feature importance analysis, and model visuali-
zation can be explored to provide meaningful ex-
planations for anomaly predictions.
(b) As IoT systems become more vulnerable to
adversarial attacks, potential research should
investigate the application of adversarial ML
techniques for robust anomaly detection.
Adversarial training, defensive distillation, and
anomaly detection in adversarial settings are
potential research directions to enhance the
resilience of ML-based anomaly detection
models against sophisticated attacks.
(c) With the continuous stream of data produced by
IoT devices, there is a need for real-time
anomaly detection techniques. Imminent
research should explore ML algorithms and
frameworks that can handle high-velocity data
streams and detect anomalies in real-time. In-
cremental learning, adaptive models, and online
feature selection methods are potential ap-
proaches to address the challenges of online and
streaming anomaly detection.
(d) Preserving data privacy is critical in IoT envi-
ronments. Forthcoming research should focus on
developing privacy-preserving ML techniques
for anomaly detection in IoT. Reliable multi-
party data processing, homomorphic encoding,
and federated learning approaches can enable
anomaly detection while ensuring data privacy
and compliance with privacy regulations.
(e) IoT devices repeatedly have inadequate
computational resources and energy constraints.
Future research should focus on developing ML
models and algorithms that are lightweight and
energy-efficient, enabling anomaly detection
directly on resource-constrained devices. Model
compression, quantization, and knowledge
distillation techniques can be explored to reduce
the computational and memory requirements of
ML models deployed on IoT devices.
Table 5 supports a structured and broad assess-
ment of projected research and previously con-
ducted surveys on cyber-attack detection using ML
in the IoT network. Applied and considered termi-
nologies are explained as follow.
i. ‘Structured’indicates that the assessments
and comparisons are presented in a well-
organized and coherent format, making it
easier to understand and analyze the
information.
ii. ‘Pros/Cons’highlight the advantages and
weaknesses of the research and surveys that
furnish a balanced view by discussing both the
positive aspects and potential limitations of the
approaches taken.
iii. ‘Disparaging’suggests that certain assess-
ments or comparisons in the survey might be
critical or unfavorable in nature. It implies that
there may be findings that are less favorable or
that highlight shortcomings in the research or
survey methodologies.
iv. ‘Assessment’examines the procedure of
evaluating or analyzing the research and sur-
vey that involves forming judgments, identi-
fying patterns, and drawing conclusions based
on the collected information.
v. ‘Coverage of Other Technique’assesses the
extent to which the research and surveys have
explored and considered various techniques
other than ML for detecting cyber-attacks in
IoT networks.
vi. ‘Technical Difficulty’indicates the degree of
expertise and resources required to success-
fully apply evaluated techniques in real-world
scenarios.
vii. ‘Performance Comparison’involves
comparing and evaluating the effectiveness
and efficiency of different ML procedures in
distinguishing cyber-attacks in IoT networks.
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 675
It focuses on measuring and analyzing factors
such as accuracy, speed, false positives, and
false negatives to determine the performance
levels of various approaches.
3.2. Confidentiality in IoT
The purpose of cryptography in IoT security is to
protect sensitive information from unauthorized
access, interception, and tampering. By employing
encryption algorithms, data can be transformed into
ciphertext, making it unreadable to adversaries
without the corresponding decryption key. This
ensures confidentiality, preventing unauthorized
entities from extracting meaningful information
from intercepted data. In this context, the perfor-
mance assessment of lightweight cryptographic al-
gorithms is of utmost importance to determine their
suitability for secure communication in IoT envi-
ronments. Factors such as encryption type, signa-
ture schemes, communication latency, gate density,
power consumption, and microcontroller platform
significantly impact the overall performance and
resource utilization. Additional look after features
for adopting a suitable encoding algorithm are as
follows [38e45].
a) The frequency of operation directly affects the
processing speed, with higher frequencies
enabling faster encryption and decryption
operations.
b) The available RAM and ROM in kilobytes (kB)
play a vital role in determining the memory re-
quirements of cryptographic algorithms.
c) Power consumption, measured in milliamperes
(mA), is a crucial consideration due to the
limited power resources in IoT devices.
d) The choice of encryption algorithm, along with
the key size, block size, and number of rounds,
determines the cryptographic strength and
efficiency.
e) The selection of an appropriate cipher and
network type ensures secure communication
and compatibility with IoT protocols.
f) Cyphering throughput, measured in megabits
per second (Mbps), indicates the data processing
speed, while latency in cycles reflects the
responsiveness and real-time capabilities of IoT
systems.
The evaluation of encryption features (i.e., illus-
trated in Table 6) provides insights into the practical
applicability of lightweight cryptographic algo-
rithms for securing IoT networks.
Table 5. Comparative representation of assessment between projected research and previously conducted surveys on cyber-attack detection in the IoT network utilizing ML techniques.
Survey Year Structured Pros./Cons. Disparaging
Assessment
Coverage of Other Techniques Technical Difficulty Performance Comparison Similar Research
a
Projected Survey 2023 Not Available
[30] 2022 11 papers
(From 1989 to 2022)
[31] 2021 1335 Papers
(From 2000 to 2021)
[32] 2022 776 Papers
(From 1983 to 2022)
[33] 2022 291 Papers
(From 1998 to 2022)
[34] 2021 1057 Papers
(From 2009 to 2021)
[35] 2021 971 Papers
(From 2000 to 2021)
[36] 2023 228 Papers
(From 2000 to 2007)
[37] 2023 Not Available
a
Similar Research data was evaluated using ‘ResearchRabbit’tool. https://www.researchrabbitapp.com/collection/public/MLPEDN35ZG.
676 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
3.2.1. Quantum Crypto Physical Unclonable
Functions (QC-PUF)
QC-PUF combines the principles of quantum
cryptography and physical unclonable functions to
enhance the security of cryptographic systems.
PUFs exploit the inherent variations in hardware
devices to generate unique cryptographic keys. In
the context of quantum cryptography, PUFs can be
used to generate and store quantum keys, providing
a higher level of security against attacks.
In a Quantum Crypto PUF system, the bit pattern
for plain text and ciphered data can be determined
by the encryption algorithm used. The encryption
algorithm operates on the plain text data using the
cryptographic key, resulting in ciphered data. The
key bits, generated by the PUFs, serve as the input
for the encryption algorithm, ensuring the unique-
ness and security of the key.
Key exchange criteria in Quantum Crypto PUFs
involve securely exchanging the cryptographic keys
between communicating parties. This can be ach-
ieved through protocols like Quantum Key Distri-
bution (QKD) that utilize quantum properties to
establish secure key exchange. The number of
cryptographic rounds depends on the specific
implementation and security requirements,
ensuring the desired level of encryption strength.
In terms of infrastructure limitations for IoT, the use
of Quantum Crypto PUFs faces challenges such as
limited distance for transmitting photons, the need for
specialized quantum devices that may be bulky and
expensive, and scalability concerns when dealing with
large-scale IoT networks. Yet, advancements in tech-
nology and research aim to address these limitations
and enable the practical implementation of Quantum
Crypto PUFs in IoT environments.
Nevertheless, by leveraging the inherent varia-
tions in hardware devices and incorporating quan-
tum cryptographic principles, QC-PUFs enhance
the resilience of cryptographic systems in Industrial
IoT and IoMT, mitigating the risks associated with
unauthorized access, data breaches, and tampering.
This advanced security measure contributes to
safeguarding critical industrial processes, sensitive
medical data, and the integrity of interconnected
devices in these evolving and interconnected
ecosystems.
3.2.1.1. Novel technical considerations for QC-PUFs
implementation. We have encountered the following
considerations while investigating the QC-PUF.
a) QC-PUF solutions need to be designed to
seamlessly integrate with diverse IoT, IIoT, and
IoMT hardware platforms. Factors such as
Table 6. Comparative analysis of cryptographic algorithms for IoT devices.
Lightweight Encryption Algorithm Gate Density
(kGEs/mm)
Power Consumption
(nW/MHz/GE)
Microcontroller
Platform
Frequency
(MHz)
RAM
(kB)
ROM
(kB)
Power
(mA)
Key Size,
Block Size
Rounds Throughput
(Mbps)
Latency
(Cycle)
GRAIN-128-AEADv2 [38] 5.2 40 Raspberry Pi 4 1500 4096 4096 650 128 bits 12 200 1000
PHOTON-beetle [39] 7.8 30 Arduino Uno 16 2 32 20 128 bits 12 10 100
FPGA Romulus [40] 4.6 35 ESP32 240 520 4096 80 128 bits 10 80 200
Sparkle [41] 8.3 45 BeagleBone Black 1 512 4096 200 128 bits 16 500 500
TinyJambu [42] 3.9 25 Particle Photon 120 128 128 70 128 bits 12 50 500
Ascon [43] 5.1 35 NVIDIA Jetson Nano 1600 8192 8192 1200 128 bits 12 600 800
GIFT-COFB [44] 4.8 40 Intel Edison 500 1024 4096 150 128 bits 12 400 300
Grain-128AEADv2 [45] 5.2 40 Particle Argon 120 256 4096 100 128 bits 12 300 400
SPARKLE-PIC32 [41] 8.3 45 Microchip PIC32 80 128 512 50 128 bits 16 100 1000
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 677
limited resources, low power consumption, and
compatibility with various microcontroller ar-
chitectures must be considered to ensure effi-
cient and practical deployment.
b) To establish secure communication channels,
post-quantum key exchange protocols should be
implemented. These protocols address the
vulnerability of classical cryptographic protocols
to quantum attacks and guarantee the confi-
dentiality and authenticity of data transmitted
between devices.
c) The physical limitations of IoT, IIoT, and IoMT
devices must be considered, including process-
ing capabilities, memory constraints, and energy
resources. Lightweight and efficient QC-PUF
implementations are necessary to minimize
computational overhead and power consump-
tion while maintaining adequate security
measures.
d) Environmental factors, such as temperature
variations, electromagnetic interference, and
physical vibrations, can affect the performance
and reliability of QC-PUF enabled devices.
Robust designs that account for these factors is
essential to ensure consistent operation in real-
world deployment scenarios.
e) Comprehensive lifecycle management strategies
for QC-PUF qualified devices are vital for secure
key generation, distribution, storage, rotation,
and revocation processes. These strategies
ensure the efficient management of crypto-
graphic keys throughout the device lifecycle,
minimizing the risk of key compromise.
3.2.2. Correlation of ML with lightweight
cryptography
Machine learning modeling is crucial in fortifying
the IoT against crippling cyber-attacks by collabo-
rating with cryptographic implementation. ML al-
gorithms provide powerful capabilities for data
analysis, pattern recognition, and anomaly detec-
tion, which are vital in identifying and mitigating
security threats. By integrating ML algorithms with
lightweight cryptographic techniques, IoT systems
can benefit from enhanced threat detection, robust
authentication mechanisms, and secure communi-
cation protocols. ML algorithms contribute to the
overall defense of IoT against cyber-attacks by
strengthening key management, facilitating real-
time threat monitoring, and enabling proactive se-
curity measures. The correlation between ML and
cryptographic algorithms forms a comprehensive
approach to strengthen IoT devices and networks,
ensuring resilience and safeguarding against
evolving threats.
Here it is worth highlighting that the centralized
nature of IoT systems poses vulnerabilities such as
data tampering, unauthorized access, and single
points of failure. These challenges can be effectively
addressed by leveraging Blockchain (BC) technology,
which provides decentralized consensus, immuta-
bility, and transparency, ensuring a robust and secure
IoT ecosystem. Likewise, by leveraging ML algo-
rithms, the blockchain can analyze massive amounts
of IoT data, detect anomalies, and identify potential
security threats in real-time. Lightweight cryptog-
raphy ensures efficient and secure communication
between IoT devices, while the blockchain acts as a
trusted distributed ledger, facilitating secure data
sharing and authentication. The need for blockchain
in IoT lies in its ability to establish a decentralized and
tamper-resistant infrastructure, ensuring data integ-
rity, privacy, and resilience against sophisticated
cyber-attacks, ultimately fortifying IoT ecosystems.
3.3. Blockchain solution for IoT security
Literature review assessment [46e51] revealed
that the Blockchain, whether it is implemented as a
private, public, or federated network, plays a crucial
role in strengthening IoT against anomalous cyber-
attacks (i.e., illustrated in Table 7). By granting a
dispersed and absolute ledger, blockchain ensures
the integrity and transparency of IoT transactions
and data. Cryptological algorithms and consensus
procedures establish trust and enhance security,
making it exceptionally complicated for malicious
actors to fiddle with or compromise IoT devices and
their associated data.
The blockchain paradigm utilizes cryptographic
hash functions to calculate the data hash, providing a
unique digital fingerprint. This hash, along with
other transaction details, is stored in blocks, forming
a chain. The blockchain's decentralized nature and
consensus mechanisms make it particularly tough for
adversary to alter or manipulate data stored on the
BC, ensuring data authentication and integrity. A
block in the BC comprises of a list of transactions that
record data exchanges between IoT devices. The
impact of blockchain on secure communication pro-
tocols like Hypertext Transfer Protocol (HTTP),
Message Queuing Telemetry Transport (MQTT),
Constrained Application Protocol (CoAP), or Exten-
sible Messaging and Presence Protocol (XMPP) lies in
providing an additional layer of security and trust
through the decentralized nature of the blockchain,
ensuring secure and reliable data transmission.
Similarly, blockchain can impact IoT-specific routing
protocols by enhancing the security, privacy, and
reliability of data routing within IoT networks.
678 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
Considering IoT device resource constraints, block-
chain functionality can be tailored to accommodate
heterogeneous devices by employing lightweight
consensus algorithms, optimizing data storage &
processing, and leveraging off-chain solutions to
minimize resource consumption while maintaining
the core benefits of blockchain technology.
3.3.1. Challenges of blockchain
Integrating Blockchain technology into IoT pre-
sents domain-specific challenges. Table 8 outlines
technical challenges and corresponding solutions.
3.3.1.1. Scalability limitations of blockchain. The scal-
ability limitations of blockchain technology in large-
scale IoT networks present considerable challenges
in ensuring efficient and reliable transaction pro-
cessing. As the number of IoT devices and trans-
actions increases, network congestion becomes a
significant concern, leading to delays and increased
transaction fees. The transaction throughput of
traditional blockchain networks, such as Bitcoin and
Ethereum, is limited, making it difficult to handle
the vast number of transactions generated by IoT
devices in real-time. Moreover, the consensus
mechanisms employed in blockchain, such as Proof
of Work (PoW) and Proof of Stake (PoS), can exac-
erbate scalability issues. PoW requires extensive
computational resources, leading to slower trans-
action confirmation times, while PoS has its limita-
tions in handling high transaction volumes. To
address these challenges, various approaches are
being explored, including sharding, sidechains, and
off-chain processing. Sharding divides the block-
chain network into smaller partitions, allowing
parallel processing of transactions, while sidechains
enable the execution of specific smart contracts off
the main blockchain, reducing congestion. Off-chain
processing moves non-critical transactions outside
the main blockchain, alleviating the burden on the
network. Achieving scalability in large-scale IoT
blockchain networks requires a careful balance be-
tween transaction volume, consensus mechanisms,
and innovative scaling solutions to ensure efficient
and seamless data processing for IoT applications.
4. Ethical and legal considerations
Ethical considerations, privacy concerns, and legal
frameworks play a pivotal role in shaping the
Table 7. Comprehensive Comparison of Public, Private, and Federated Blockchains for IoT Devices.
Public [47] Private [48] Federated [49,50]
Access Publicly accessible to any
IoT entity
Restricted access to
authorized entities
Restricted access to a
group of trusted entities
Speed Moderate to Slow Fast Fast
Efficiency Moderate to Low High High
Security High High High
Immutability Immutable once
confirmed
Immutable once
confirmed
Immutable once
confirmed
Consensus Process Decentralized consensus Permissioned consensus Permissioned consensus
Consensus Mechanism Proof of Work, Proof of
Stake (Ethereum)
Various (e.g., Practical
Byzantine Fault Tolerance,
Raft)
Various (e.g., Federated
Byzantine Agreement)
Network Type Publicly shared network Private or consortium
network
Private or consortium
network
Open Source Yes Yes Yes
Smart Contracts Type Turing Complete
(Ethereum)
Turing Complete (Hyper-
ledger Fabric)
Turing Complete
(Multichain)
Particular Hardware
Requisite
High computational
power required
No specific requirements No specific requirements
Avg. Transactions per
Second
Varies (e.g., Ethereum:
15 TPS)
Varies (depends on
network infrastructure)
Varies (depends on
network infrastructure)
Hashing Algorithm Various (e.g., SHA-256,
Ethash)
Various (e.g., SHA-256,
SHA-3)
Various (e.g., SHA-256,
SHA-3)
Key Administration Yes (through public-pri-
vate key pairs)
Yes (through access
controls)
Yes (through access
controls)
Data Confidentiality No (Transparent) Yes Yes
Scalability Limited Flexible Flexible
Governance Decentralized Centralized Consortium-based
Customization Limited High High
Network Overhead Higher Lower Lower
Interoperability Limited Limited Limited
Cost Competence Lower Higher Higher
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 679
landscape of IoT security. As IoT devices become
more pervasive in our daily lives and critical infra-
structure, safeguarding user privacy and adhering to
ethical principles becomes paramount. The collection,
storage, and processing of vast amounts of personal
data by IoT devices raise significant ethical concerns
regarding consent, data ownership, and potential
misuse. Addressing these issues requires robust
technical solutions that prioritize data protection and
user control over their information. Implementing
privacy by design principles, encryption, and secure
data transmission protocols can help mitigate privacy
risks and ensure data confidentiality.
In addition to technical measures, adhering to
legal frameworks is essential to establish clear
guidelines and responsibilities for all stakeholders
involved in the IoT ecosystem. Compliance with
existing data protection regulations, such as the
General Data Protection Regulation (GDPR) and
HIPPA, ensures that user data is handled lawfully
and transparently. Legislative bodies worldwide
must work collaboratively to create comprehensive
and adaptive IoT-specific regulations to address
emerging challenges. These legal frameworks must
encompass device security standards, data breach
notification requirements, and liability allocation to
promote accountability among IoT manufacturers,
service providers, and users. Moreover, ethical
considerations extend beyond data privacy to
encompass the potential societal impact of IoT
technologies. Striking a balance between innovation
and ethical use is crucial to prevent unintended
consequences and potential harm. Robust risk
assessment and ethical impact assessments should
be integrated into the development and deployment
of IoT systems. Responsible innovation in IoT se-
curity involves not only technical expertise but also
a deep understanding of ethical principles, user
perspectives, and social implications. By addressing
ethical considerations, privacy concerns, and legal
frameworks, the IoT security community can pave
the way for a safer, more secure, and ethically
conscious IoT ecosystem.
5. Key applicable recommendations
A rigorous review of survey analysis triggered the
following recommendations to nominate light-
weight, scalable, and effective cyber-shield in IoT
cyber-defense.
a) Robust device authentication mechanisms, such
as X.509 certificates and mutual authentication
protocols like EAP-TLS (Extensible Authentica-
tion Protocol-Transport Layer Security), help
prevent unauthorized access to IoT devices.
Implementing strong authentication mitigates
the risk of impersonation attacks.
b) Deploying protocols like MQTT with TLS for
end-to-end encrypted communication ensures
the confidentiality and integrity of data
exchanged between IoT devices and the server.
Proper protocol selection and configuration are
crucial for secure IoT communication.
c) Establishing a well-defined process for timely
security updates and patches, following stan-
dards like ISO/IEC 27001, and leveraging
vulnerability management frameworks like
CVSS (Common Vulnerability Scoring System),
ensures that known vulnerabilities are promptly
addressed and reduces the likelihood of suc-
cessful attacks.
d) Employing network segmentation using VLANs
(Virtual Local Area Networks) or SDN (Soft-
ware-Defined Networking) techniques helps
isolate IoT devices into separate security zones,
limiting the lateral movement of threats and
minimizing the potential impact of attacks.
e) Employing security information and event
management (SIEM) algorithms integrated with
threat intelligence feeds, such as STIX/TAXII
(Structured Threat Information Expression/
Trusted Automated Exchange of Indicator In-
formation), facilitates proactive monitoring,
timely incident detection, and response to
emerging threats.
f) Employing techniques such as data-at-rest
encryption, using algorithms like XTS-AES
(XEX-based Tweaked CodeBook Mode with Ci-
pher-Text Stealing and Advanced Encryption
Standard), and secure protocols like HTTPS
(Hypertext Transfer Protocol Secure) or SFTP
(Secure File Transfer Protocol), ensures data
confidentiality and integrity throughout its
lifecycle.
g) Exploring the integration of blockchain tech-
nology, such as Ethereum or Hyperledger Fab-
ric, in IoT infrastructure enhances security,
transparency, and trust among participants.
Employing smart contracts and distributed led-
ger technology ensures tamper-resistant data
integrity, decentralized consensus, and audit-
able transactions.
h) Considering the future threat of quantum com-
puters, exploring post-quantum cryptographic
algorithms like lattice-based or code-based
cryptography, along with physical unclonable
functions (PUFs) for device authentication, en-
hances the resistance of IoT infrastructure
against potential quantum-based attacks.
680 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
Table 8. The Integration of Blockchain Technology in IoT: Addressing Multi-Domain Challenges and Solutions.
BC Challenges Description Solution Hindrance Issues Applied Technologies
and Protocols
Ref.
Interoperability Ensuring seamless
integration and
compatibility be-
tween heteroge-
neous IoT devices
and Blockchain
networks
Standardized
data formats and
protocols
Diverse device architec-
tures and protocols
IoT protocols: MQTT,
CoAP, HTTP
[52]
Scalability Addressing the
scalability limita-
tions of Blockchain
to accommodate
the growing num-
ber of IoT devices
Sharding, side-
chains, off-chain
processing
Performance impact,
network congestion
Sharding: Ethereum 2.0,
Hyperledger Fabric
[53]
Privacy and
Security
Protecting sensitive
IoT data from un-
authorized access,
tampering, and
privacy breaches
Cryptographic al-
gorithms, access
control
Privacy breaches, data
leakage, key management
Encryption: AES, RSA;
Access control: ACL,
RBAC
[54]
Smart Contracts Developing secure
and efficient execu-
tion environments
for automated
transactions and
agreements
Secure contract
coding, auditing
Vulnerabilities, smart
contract bugs
Ethereum Virtual Ma-
chine (EVM), Solidity
[55]
Power
Consumption
Optimizing energy
efficiency in IoT
Blockchain net-
works to minimize
power
consumption
Energy-efficient
consensus
mechanisms
Limited device resources,
battery life
Proof of Stake (PoS), Proof
of Authority (PoA)
[56]
Protocol
Standardization
Establishing stan-
dardized commu-
nication protocols,
data formats, and
interfaces for IoT-
Blockchain
integration
IoT protocol
harmonization
Lack of consensus,
compatibility challenges
IETF standards, ISO/IEC
standards
[57]
Infrastructure
Compatibility
Ensuring compati-
bility between
Blockchain infra-
structure and
diverse IoT device
architectures
Middleware solu-
tions, IoT gateway
integration
Resource constraints, con-
nectivity limitations
MQTT brokers, IoT gate-
ways, Blockchain APIs
[58]
Legal and
Compliance
Addressing legal
and regulatory
frameworks for
data protection,
privacy, and cyber-
security in IoT-
Blockchain
Compliance frame-
works, regulatory
guidelines
Jurisdictional issues,
cross-border data
transfers
GDPR, HIPAA, ISO/IEC
standards
[59]
Security Attacks Mitigating security
threats such as
Sybil attacks, 51%
attacks, and dou-
ble-spending
attacks
Consensus mecha-
nisms, Byzantine
fault tolerance
Attack complexity,
network vulnerability
Proof of Work (PoW),
Practical Byzantine Fault
Tolerance (PBFT)
[60]
Trust
Establishment
Establishing trust
among IoT stake-
holders through
identity manage
ment and reputa-
tion systems
PKI, decentralized
identity systems
Trustworthiness verifica-
tion, identity theft
Blockchain-based identity
solutions, DIDs
[61]
(continued on next page)
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 681
i) The emergence of quantum-based attacks
poses significant implications for the security
of IoT-driven blockchain systems. Traditional
cryptographic algorithms, such as RSA and
ECC, are vulnerable to being broken by
powerful quantum computers, jeopardizing
the confidentiality and integrity of data
exchanged in blockchain networks. To address
this threat, post-quantum cryptographic algo-
rithms have been proposed and evaluated for
their applicability in the context of IoT-driven
blockchain. These algorithms, based on lattice-
based cryptography, code-based cryptog-
raphy, multivariate polynomials, and other
mathematical structures, offer resistance
against quantum attacks due to their under-
lying mathematical complexity. However, the
adoption of post-quantum cryptographic al-
gorithms in IoT-driven blockchain introduces
challenges related to performance and
resource constraints. The higher computa-
tional overhead and memory requirements of
these algorithms must be carefully balanced
with the limited processing capabilities and
energy constraints of IoT devices. In addition,
ensuring interoperability and compatibility
with existing blockchain frameworks and
smart contract execution environments be-
comes crucial to achieve seamless integration.
Further research and empirical validation are
required to determine the most suitable post-
quantum cryptographic algorithms for IoT-
driven blockchain, striking a balance between
security, performance, and resource efficiency
in the face of quantum-based threats.
j) Human behavior and user interactions with IoT
devices play a significant role in introducing
potential vulnerabilities (e.g., misconfiguration,
insecure network connections, lack of device
firmware updates, etc.) to IoT security.
Table 8. (continued)
BC Challenges Description Solution Hindrance Issues Applied Technologies
and Protocols
Ref.
Skyline Query
Processing
Designing efficient
techniques for pro-
cessing complex
queries on Block-
chain data in real-
time
Distributed query
processing
algorithms
Query complexity, real-
time response
requirements
Distributed databases,
MapReduce algorithms
[62]
Decentralized
Cooperation
Enabling collective
decision-making,
consensus, and
governance in IoT-
Blockchain
ecosystems
Decentralized
governance models
Governance conflicts,
scalability challenges
DAO (Decentralized
Autonomous Organiza-
tion), Voting mechanisms
[63]
Consensus Protocol Evaluating and
selecting consensus
protocols that meet
IoT application
requirements
Lightweight
consensus
algorithms
Scalability, latency,
consensus fault tolerance
Proof of Stake (PoS),
Delegated Proof of Stake
(DPoS)
[64]
Big Data and Ma-
chine Learning
Integrating secure
and privacy-pre-
serving big data
analytics and ML
algorithms in IoT-
Blockchain
Homomorphic
encryption, feder-
ated learning
Data privacy, model accu-
racy, computational
overhead
Secure Multiparty
Computation (MPC), Dif-
ferential Privacy
[65]
SDN and Cellular
Network
Integrating Block-
chain with SDN
and cellular net-
works for enhanced
security, scalability,
and device
management
Blockchain-based
network
management
Network interoperability,
latency, network overhead
Software-Defined
Networking (SDN), 5G,
Cellular IoT protocols
[66]
Energy
Management
Designing energy-
efficient mecha
nisms and pro
tocols for IoT de
vices in Blockchain
networks
Low-power
consensus
algorithms
Limited device resources,
energy consumption
Proof of Stake (PoS), Proof
of Authority (PoA)
[60]
682 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
Understanding the impact of human factors, us-
ability considerations, and user-centric security
design principles is crucial in addressing these
risks effectively. User-centered approaches that
prioritize intuitive interfaces, clear instructions,
and simplified security measures can enhance
the overall security posture of IoT devices.
Moreover, incorporating user education and
awareness programs can empower users to make
informed security decisions and adopt safe
practices while interacting with IoT technology.
By emphasizing user-centric security measures,
the IoT ecosystem can mitigate potential security
risks and create a more resilient and secure
environment for users and their interconnected
devices.
6. Enhancements from existing reviews/
surveys
This review paper distinguishes itself from related
surveys/reviews [31,32],[34e37],[47],[59],[67,68]
through its adoption of a comprehensive research
method, which rigorously explores procedural
approach ‘Systematic Review’. Unlike conventional
surveys, this review delves deeper into the literature
on IoT security, offering a comprehensive under-
standing of the subject matter. It not only identifies
the vulnerabilities, threats, and challenges posed by
interconnected devices but also proposes a robust
framework for vulnerability assessment and mitiga-
tion. Moreover, the paper critically evaluates the ef-
ficacy of existing security features and protocols,
providing a thorough analysis of their strengths and
limitations. It goes beyond mere summarization of
findings and recommends the integration of inno-
vative technologies like blockchain and machine
learning algorithms to fortify IoT security. This
comprehensive and well-rounded analysis ensures a
reliable and authoritative approach to evidence syn-
thesis in the ever-evolving domain of IoT security.
7. Conclusion and future work
The 21st century has witnessed the widespread
adoption of IoT in various domains, including smart
homes, industries, and healthcare facilities, bringing
numerous benefits and advancements in efficiency,
automation, and convenience. Nonetheless, the
growing reliance on IoT infrastructure necessitates a
robust security framework due to the inherent vul-
nerabilities and threats associated with inter-
connected devices. These include potential
unauthorized access, data breaches, device
manipulation, and network disruptions, emphasizing
the criticality of implementing effective security
measures. This research paper on IoT security review
makes significant contributions by providing a
comprehensive understanding of the existing litera-
ture, conducting a systematic survey to identify
anomalies and risks, proposing a framework for
vulnerability assessment and mitigation, and evalu-
ating the effectiveness of current security features and
protocols. While the paper offers valuable insights
and recommendations, it is important to acknowledge
some limitations. The presented systematic review
may not encompass all possible security aspects, and
the proposed enhancements to security mechanisms
require further empirical validation to ensure their
efficacy in real-world IoT environments.
Future work. To pave the way for future research, it
is imperative to explore various avenues that can
enhance IoT security beyond the existing capabilities.
Firstly, investigating the integration of quantum
systems, 6G, Federated Learning (FL), and artificial
intelligence (AI) into IoT infrastructure holds prom-
ise in significantly enhancing data processing, pri-
vacy preservation, and overall security measures.
Moreover, evaluating the potential benefits and
feasibility of implementing Named Data Network
(NDN) as an alternative to IP-based systems is
crucial. NDN's inherent data-centric approach can
potentially improve data integrity, confidentiality,
and resilience in IoT environments, warranting
further exploration and experimentation.
Likewise, to address the ever-evolving threat land-
scape, continuous monitoring, and adaptation to
emerging security regulations and standards are
essential. This involves actively staying abreast of
evolving policies, industry guidelines, and best prac-
tices to ensure that IoT security measures remain
effective and up todate. Ultimately, the exploration of
novel cryptographic algorithms, lightweight authen-
tication protocols, and secure firmware update
mechanisms specifically designed for IoT devices
should be researched and investigated to significantly
enhance the IoT ecosystem's security posture.
Funding
This study was sponsored by Prince Sattam bin
Abdulaziz University via project number 2023/RV/8.
Institutional review board statement
The study was conducted according to the
guidelines of the Declaration of Deanship of Scien-
tific Research, Prince Sattam Bin Abdulaziz Uni-
versity, Saudi Arabia.
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 683
Informed consent statement
Not applicable.
Data availability statement
Not applicable.
Conflicts of interest
The authors declare no conflict of interest.
Acknowledgments
This study was sponsored by Prince Sattam bin
Abdulaziz University through project number 2023/
RV/8.
References
[1] S. Ashraf, A proactive role of IoT devices in building smart
cities, Int. Things and Cyber-Physical Systems 1 (2021) 8e13,
https://doi.org/10.1016/j.iotcps.2021.08.001.
[2] A.G. Silva-Trujillo, M.J. Gonz
alez Gonz
alez, L.P. Rocha
P
erez, L.J. García Villalba, Cybersecurity analysis of wear-
able devices: smartwatches passive attack, Sensors 23 (2023)
1e18, https://doi.org/10.3390/s23125438.
[3] C. Bayõlmõs
¸, M.A. Ebleme, Ü. Çavus
¸o
glu, K. Küçük,
A. Sevin, A survey on communication protocols and per-
formance evaluations for Internet of Things, Digital Com-
munications and Networks 8 (2022) 1094e1104, https://
doi.org/10.1016/j.dcan.2022.03.013.
[4] L.J.S. Kumar, P. Krishnan, B. Shreya, S. MS, Performance
enhancement of FSO communication system using machine
learning for 5G/6G and IoT applications, Optik 252 (2022)
1e9, https://doi.org/10.1016/j.ijleo.2021.168430.
[5] U. Tariq, I. Ahmed, A.K. Bashir, K. Shaukat, A critical
cybersecurity analysis and future research directions for the
internet of things: a comprehensive review, Sensors 23 (2023)
1e46, https://doi.org/10.3390/s23084117.
[6] N. Priya, Cybersecurity considerations for industrial IoT in
critical infrastructure sector, Int. J. Comput. Organ. Trends
12 (2022) 27e36, https://doi.org/10.14445/22492593/ijcot-
v12i1p306.
[7] E. Schiller, A. Aidoo, J. Fuhrer, J. Stahl, M. Zi€
orjen, B. Stiller,
Landscape of IoT security, Computer Science Review 44
(2022) 1e18, https://doi.org/10.1016/j.cosrev.2022.100467.
[8] S. Fraihat, S. Makhadmeh, M. Awad, M.A. Al-Betar, A. Al-
Redhaei, Intrusion detection system for large-scale IoT
NetFlow networks using machine learning with modified
Arithmetic Optimization Algorithm, Internet of Things 22
(2023) 1e22, https://doi.org/10.1016/j.iot.2023.100819.
[9] S. Garcia, A. Parmisano, M.J. Erquiaga, IoT-23: a labeled
dataset with malicious and benign IoT network traffic, Zen-
odo (2020), https://doi.org/10.5281/zenodo.4743746.
[10] S. Dadkhah, H. Mahdikhani, P.K. Danso, A. Zohourian,
K.A. Truong, A.A. Ghorbani, Towards the development of a
realistic multidimensional IoT profiling dataset, 19th annual
international conference on privacy, Security &Trust (PST)
(2022) 1e11, https://doi.org/10.1109/PST55820.2022.9851966.
[11] A. Pinto, L.-C. Herrera, Y. Donoso, J.A. Gutierrez, Survey on
intrusion detection systems based on machine learning
techniques for the protection of critical infrastructure, Sen-
sors 23 (2023) 1e18, https://doi.org/10.3390/s23052415.
[12] M. Hasan, Md.M. Islam, M.I.I. Zarif, M.M.A. Hashem, Attack
and anomaly detection in IoT sensors in IoT sites using
machine learning approaches, Internet of Things 7 (2019)
1e14, https://doi.org/10.1016/j.iot.2019.100059.
[13] K. He, D.D. Kim, M.R. Asghar, Adversarial machine learning
for network intrusion detection systems: a comprehensive
survey, IEEE Communications Surveys &Tutorials. 25 (2023)
538e566, https://doi.org/10.1109/comst.2022.3233793.
[14] V. Gugueoth, S. Safavat, S. Shetty, Security of Internet of
Things (IoT) using federated learning and deep learning d
recent advancements, issues and prospects, ICT Express 9
(2023) 1e20, https://doi.org/10.1016/j.icte.2023.03.006.
[15] M. Arunkumar, K.A. Kumar, GOSVM: gannet optimization-
based support vector machine for malicious attack detection
in cloud environment, Int. J. Inf. Technol. 15 (2023)
1653e1660, https://doi.org/10.1007/s41870-023-01192-z.
[16] R. Yadav, I. Sreedevi, D. Gupta, Augmentation in perfor-
mance and security of WSNs for IoT applications using
feature selection and classification techniques, Alex. Eng. J.
65 (2023) 461e473, https://doi.org/10.1016/j.aej.2022.10.033.
[17] M. Mohy-eddine, A. Guezzaz, S. Benkirane, M. Azrour, An
efficient network intrusion detection model for IoT security
using K-NN classifier and feature selection, Multimed. Tool.
Appl. 82 (2023) 23615e23633, https://doi.org/10.1007/s11042-
023-14795-2.
[18] J.B. Awotunde, F.E. Ayo, R. Panigrahi, A. Garg, A.K. Bhoi,
P. Barsocchi, A multi-level random forest model-based
intrusion detection using fuzzy inference system for internet
of things networks, Int. J. Comput. Intell. Syst. 16 (2023) 1e22,
https://doi.org/10.1007/s44196-023-00205-w.
[19] M. Douiba, S. Benkirane, A. Guezzaz, M. Azrour, An
improved anomaly detection model for IoT security using
decision tree and gradient boosting, J. Supercomput. 79 (2022)
3392e3411, https://doi.org/10.1007/s11227-022-04783-y.
[20] Y. Zhou, L. Song, Y. Liu, P. Vijayakumar, B.B. Gupta,
W. Alhalabi, H. Alsharif, A privacy-preserving logistic
regression-based diagnosis scheme for digital healthcare,
Future Generat. Comput. Syst. 144 (2023) 63e73, https://
doi.org/10.1016/j.future.2023.02.022.
[21] N.A. Bajao, J. Sarucam, Threats Detection in the Internet of
Things Using Convolutional neural networks, long short-
term memory, and gated recurrent units, Mesopotamian J.
Cyber Security 2023 (2023) 22e29, https://doi.org/10.58496/
mjcs/2023/005.
[22] O. Abu Alghanam, W. Almobaideen, M. Saadeh, O. Adwan,
An improved PIO feature selection algorithm for IoT
network intrusion detection system based on ensemble
learning, Expert Syst. Appl. 213 (2023) 1e16, https://doi.org/
10.1016/j.eswa.2022.118745.
[23] L. Prathibha, K. Fatima, A novel high-speed data encryption
scheme for internet of medical things using modified elliptic
curve diffieehellman and advance encryption standard, Int.
J. Image Graph. 22 (2022) 1e15, https://doi.org/10.1142/
s0219467823400041.
[24] M. Alhanaya, K. Hamdi Ateyeh Al-Shqeerat, Performance
analysis of intrusion detection system in the IoT environ-
ment using feature selection technique, Intelligent Automa-
tion &Soft Computing 36 (2023) 3709e3724, https://doi.org/
10.32604/iasc.2023.036856.
[25] M. Asad, M. Aslam, S.F. Jilani, S. Shaukat, M. Tsukada,
SHFL: K-Anonymity-Based secure hierarchical federated
learning framework for smart healthcare systems, Future
Internet 14 (2022) 1e16, https://doi.org/10.3390/fi14110338.
[26] S.V.N. Santhosh Kumar, Y. Palanichamy, M. Selvi,
S. Ganapathy, A. Kannan, S.P. Perumal, Energy efficient
secured K means based unequal fuzzy clustering algorithm
for efficient reprogramming in wireless sensor networks,
Wireless Network 27 (2021) 3873e3894, https://doi.org/
10.1007/s11276-021-02660-9.
[27] C.-S. Shieh, W.-W. Lin, T.-T. Nguyen, C.-H. Chen, M.-
F. Horng, D. Miu, Detection of unknown DDoS attacks with
deep learning and Gaussian mixture model, Appl. Sci. 11
(2021) 1e13, https://doi.org/10.3390/app11115213.
[28] G. Sharma, J. Grover, A. Verma, QSec-RPL: detection of
version number attacks in RPL based mobile IoT using Q-
Learning, Ad Hoc Netw. 142 (2023) 1e16, https://doi.org/
10.1016/j.adhoc.2023.103118.
684 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
[29] X. Chen, X. Liu, Y. Chen, L. Jiao, G. Min, Deep Q-Network
based resource allocation for UAV-assisted Ultra-Dense
Networks, Comput. Network. 196 (2021) 1e10, https://
doi.org/10.1016/j.comnet.2021.108249.
[30] Z. Chen, J. Liu, Y. Shen, M. Simsek, B. Kantarci,
H.T. Mouftah, P. Djukic, Machine learning-enabled IoT se-
curity: open issues and challenges under advanced persis-
tent threats, ACM Comput. Surv. 55 (2022) 1e37, https://
doi.org/10.1145/3530812.
[31] R. Ahmad, I. Alsmadi, Machine learning approaches to IoT
security: a systematic literature review, Internet of Things 14
(2021) 1e42, https://doi.org/10.1016/j.iot.2021.100365.
[32] M. Abdullahi, Y. Baashar, H. Alhussian, A. Alwadain,
N. Aziz, L.F. Capretz, S.J. Abdulkadir, Detecting cyberse-
curity attacks in internet of things using artificial intelligence
methods: a systematic literature review, Electronics 11 (2022)
1e27, https://doi.org/10.3390/electronics11020198.
[33] U. Inayat, M.F. Zia, S. Mahmood, H.M. Khalid,
M. Benbouzid, Learning-based methods for cyber attacks
detection in IoT systems: a survey on methods, analysis, and
future prospects, Electronics 11 (2022) 1e20, https://doi.org/
10.3390/electronics11091502.
[34] L. Aversano, M.L. Bernardi, M. Cimitile, R. Pecori,
A systematic review on Deep Learning approaches for IoT
security, Comput. Sci. Review 40 (2021) 1e18, https://doi.org/
10.1016/j.cosrev.2021.100389.
[35] S.H. Haji, S.Y. Ameen, Attack and anomaly detection in IoT
networks using machine learning techniques: a review,
Asian J. Res. Comput. Sci. 9 (2021) 30e46, https://doi.org/
10.9734/ajrcos/2021/v9i230218.
[36] A. Ahmed Jamal, A.-A. Mustafa Majid, A. Konev,
T. Kosachenko, A. Shelupanov, A review on security analysis
of cyber physical systems using Machine learning, Mater.
Today: Proc. 80 (2023) 2302e2306, https://doi.org/10.1016/
j.matpr.2021.06.320.
[37] L.G.F. da Silva, D.F.H. Sadok, P.T. Endo, Resource opti-
mizing federated learning for use with IoT: a systematic re-
view, J. Parallel Distr. Comput. 175 (2023) 92e108, https://
doi.org/10.1016/j.jpdc.2023.01.006.
[38] B. Li, H. Zhang, D. Lin, Efficient (masked) hardware imple-
mentation of grain-128AEADv2, Secur. Commun. Network.
2023 (2023) 1e16, https://doi.org/10.1155/2023/8044164.
[39] A.T. Mozipo, J.M. Acken, Residual vulnerabilities to power
side channel attacks of lightweight ciphers cryptography
competition finalists, IET Comput. Digital Tech. 17 (2023)
1e14, https://doi.org/10.1049/cdt2.12057.
[40] S. Jin, W. Yuanzhi, S. Yining, Design and implementation of
wireless multimedia sensor network node based on FPGA
and binocular vision, EURASIP J. Wirel. Commun. Netw.
2018 (2018) 1e8, https://doi.org/10.1186/s13638-018-1172-8.
[41] C. Beierle, A. Biryukov, L. Cardoso dos Santos, J. Großsch€
adl,
L. Perrin, A. Udovenko, V. Velichkov, Q. Wang, Lightweight
AEAD and hashing using the sparkle permutation family,
IACR Transactions on Symmetric Cryptology 2020 (2020)
208e261, https://doi.org/10.46586/tosc.v2020.is1.208-261.
[42] H.W. Sun, B.-B. Cai, S.-J. Qin, Q.-Y. Wen, F. Gao, Quantum
attacks on beyond-birthday-bound macs, Phys. Stat. Mech.
Appl. 625 (2023) 1e17, https://doi.org/10.1016/
j.physa.2023.129047.
[43] P. Joshi, B. Mazumdar, SSFA: subset fault analysis of
ASCON-128 authenticated cipher, Microelectron. Reliab. 123
(2021) 114155, https://doi.org/10.1016/j.microrel.2021.114155.
[44] S. Banik, A. Chakraborti, A. Inoue, T. Iwata, K. Minematsu,
M. Nandi, T. Peyrin, Y. Sasaki, S.M. Sim, Y. Todo, GIFT-
COFB, Cryptology ePrint Archive, 2022, pp. 1e30. https://
eprint.iacr.org/2020/738.
[45] T. Oder, T. Schneider, T. P€
oppelmann, T. Güneysu, Practical
CCA2-secure and masked ring-LWE implementation, IACR
Transactions on Cryptographic Hardware and Embedded
Systems 2018 (2018) 142e174, https://doi.org/10.46586/
tches.v2018.i1.142-174.
[46] U. Tariq, Rampant Smoothing (RTS) Algorithm: an opti-
mized consensus mechanism for private Blockchain enabled
technologies, EURASIP J. Wirel. Commun. Netw. 2022 (2022)
1e22, https://doi.org/10.1186/s13638-022-02123-5.
[47] A. Attkan, V. Ranga, Cyber-physical security for IoT net-
works: a comprehensive review on traditional, blockchain
and artificial intelligence based key-security, Complex &
Intelligent Systems 8 (2022) 3559e3591, https://doi.org/
10.1007/s40747-022-00667-z.
[48] M. Dibaei, X. Zheng, Y. Xia, X. Xu, A. Jolfaei, A.K. Bashir,
U. Tariq, D. Yu, A.V. Vasilakos, Investigating the prospect of
leveraging blockchain and machine learning to secure
vehicular networks: a survey, IEEE Trans. Intell. Transport.
Syst. 23 (2022) 683e700, https://doi.org/10.1109/
tits.2020.3019101.
[49] S. Liao, J. Wu, A.K. Bashir, W. Yang, J. Li, U. Tariq, Digital twin
consensus for blockchain-enabled intelligent transportation
systems in smart cities, IEEE Trans. Intell. Transport. Syst. 23
(2022) 22619e22629, https://doi.org/10.1109/tits.2021.3134002.
[50] R. Arul, Y.D. Al-Otaibi, W.S. Alnumay, U. Tariq, U. Shoaib,
M.D.J. Piran, Multi-modal secure healthcare data dissemi-
nation framework using blockchain in IoMT, Personal
Ubiquitous Comput. 2021 (2021) 1e13, https://doi.org/
10.1007/s00779-021-01527-2.
[51] G.P. Joshi, E. Perumal, K. Shankar, U. Tariq, T. Ahmad,
A. Ibrahim, Toward blockchain-enabled privacy-preserving
data transmission in cluster-based vehicular networks,
Electronics 9 (2020) 1e15, https://doi.org/10.3390/
electronics9091358.
[52] M.S. Rahman, M.A.P. Chamikara, I. Khalil, A. Bouras,
Blockchain-of-blockchains: an interoperable blockchain
platform for ensuring IoT data integrity in smart city, Journal
of Industrial Information Integration 30 (2022) 1e11, https://
doi.org/10.1016/j.jii.2022.100408.
[53] P. Pabitha, J.C. Priya, R. Praveen, S. Jagatheswari, ModChain:
a hybridized secure and scaling blockchain framework for
IoT environment, Int. J. Inf. Technol. 15 (2023) 1741e1754,
https://doi.org/10.1007/s41870-023-01218-6.
[54] E.H. Abualsauod, A hybrid blockchain method in internet of
things for privacy and security in unmanned aerial vehicles
network, Comput. Electr. Eng. 99 (2022) 1e13, https://
doi.org/10.1016/j.compeleceng.2022.107847.
[55] Q. Zhou, K. Zheng, K. Zhang, L. Hou, X. Wang, Vulnerability
analysis of smart contract for blockchain-based IoT applica-
tions: a machine learning approach, IEEE Internet Things J. 9
(2022) 24695e24707, https://doi.org/10.1109/jiot.2022.3196269.
[56] O.A. Khashan, N.M. Khafajah, Efficient hybrid centralized
and blockchain-based authentication architecture for het-
erogeneous IoT systems, Journal of King Saud University -
Computer and Information Sciences 35 (2023) 726e739,
https://doi.org/10.1016/j.jksuci.2023.01.011.
[57] A. Dixit, A. Trivedi, W.W. Godfrey, A survey of cyber at-
tacks on blockchain based IoT systems for industry 4.0, IET
Blockchain 3 (2022) 1e15, https://doi.org/10.1049/
blc2.12017.
[58] E.E.-D. Hemdan, W. El-Shafai, A. Sayed, Integrating digital
twins with IoT-based blockchain: concept, architecture,
challenges, and future scope, Wireless Pers. Commun. 131
(2023) 1e24, https://doi.org/10.1007/s11277-023-10538-6.
[59] V. Wylde, N. Rawindaran, J. Lawrence, R. Balasubramanian,
E. Prakash,A. Jayal, I. Khan, C. Hewage,J. Platts, Cybersecurity,
Data privacy and blockchain: a review, SN Computer Science 3
(2022) 1e12, https://doi.org/10.1007/s42979-022-01020-4.
[60] A. Alsirhani, M.A. Khan*, A. Alomari, S. Maryam,
A. Younas, M. Iqbal, M.H. Siqqidi, A. Ali, Securing low-
power blockchain-enabled IoT devices against energy
depletion attack, ACM Trans. Internet Technol. 23 (2022)
1e17, https://doi.org/10.1145/3511903.
[61] T. Hewa, A. Braeken, M. Liyanage, M. Ylianttila, Fog
computing and blockchain-based security service architec-
ture for 5G industrial IoT-enabled cloud manufacturing,
IEEE Trans. Ind. Inf. 18 (2022) 7174e7185, https://doi.org/
10.1109/tii.2022.3140792.
[62] Z. Wang, L. Zhang, X. Ding, K.-K.R. Choo, H. Jin, A dynamic-
efficient structure for secure and verifiable location-based
U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686 685
skyline queries, IEEE Trans. Inf. Forensics Secur. 18 (2023)
920e935, https://doi.org/10.1109/tifs.2022.3224666.
[63] H. Xue, D. Chen, N. Zhang, H.-N. Dai, K. Yu, Integration of
blockchain and edge computing in internet of things: a sur-
vey, Future Generat. Comput. Syst. 144 (2023) 307e326,
https://doi.org/10.1016/j.future.2022.10.029.
[64] A. Singh, G. Kumar, R. Saha, M. Conti, M. Alazab,
R. Thomas, A survey and taxonomy of consensus protocols
for blockchains, J. Syst. Architect. 127 (2022) 1e27, https://
doi.org/10.1016/j.sysarc.2022.102503.
[65] A. Mitra, B. Bera, A.K. Das, S.S. Jamal, I. You, Impact on
blockchain-based AI/ML-enabled big data analytics for
Cognitive Internet of Things environment, Comput. Com-
mun. 197 (2023) 173e185, https://doi.org/10.1016/
j.comcom.2022.10.010.
[66] B. Sellami, A. Hakiri, S. Ben Yahia, Deep Reinforcement
Learning for energy-aware task offloading in join SDN-
Blockchain 5G massive IoT edge network, Future Generat.
Comput. Syst. 137 (2022) 363e379, https://doi.org/10.1016/
j.future.2022.07.024.
[67] A.O. Bang, U.P. Rao, A. Visconti, A. Brighente, M. Conti, An
IOT inventory before deployment: a survey on IOT pro-
tocols, communication technologies, vulnerabilities, attacks,
and future research directions, Comput. Secur. 123 (2022)
1e14.
[68] B.K. Mohanta, D. Jena, U. Satapathy, S. Patnaik, Survey on
IOT security: challenges and solution using machine
learning, artificial intelligence and Blockchain technology,
Internet of Things 11 (2020) 1e27, https://doi.org/10.1016/
j.iot.2020.100227.
686 U. Tariq et al. / Karbala International Journal of Modern Science 9 (2023) 665e686
