A Comprehensive Risk Assessment Framework for IoT-Enabled
Healthcare Environment
Mofareh Waqdan
a
, Habib Louafi
b
and Malek Mouhoub
c
Department of Computer Science, University of Regina, Regina, SK, Canada
Keywords:
Internet of Things (IoT), Risk Assessment, Risk Parameters, IoT Attacks.
Abstract:
The significance of risk assessment in medical sectors, particularly in emergency rooms, is crucial due to
the criticality of the service. We present a comprehensive risk assessment framework for analyzing the risks
associated with deploying and using Internet of Things (IoT) technologies in a healthcare environment. In this
context, we improve upon the existing methodologies by dynamically calculating the risk score for different
devices profiles, considering their number along with other parameters, such as network protocols, device
heterogeneity, device security updates, device physical security status, device history status, layer history
status, and device criticality. The framework helps healthcare organizations identify, assess, and manage
the risks of IoT, which can range from data privacy and confidentiality to system integrity, availability, and
performance.
1 INTRODUCTION
The overwhelming spread and use of the Internet
have enabled hyper-connectivity, where data and in-
formation can instantly be shared anywhere across the
globe. This phenomenon of hyper-connectivity has
further been fueled by the concept of the IoT, which
has transformed our daily lives in a wide variety of
ways (Koohang et al., 2022). IoT applications are
now extended to roughly all facets of human life, e.g.,
transportation, e-governance, sustainable cities, smart
agriculture, power grids, smart homes, e-healthcare,
etc. This hyper-connectivity through IoT networks
is not always safe or secure because attackers are
always on a hunt to find weaknesses, misconfigura-
tions, protocol flaws, or hardware failures in order to
subvert the Confidentiality, Integrity, and Availability
(CIA) of these IoT networks. Moreover, the hetero-
geneous nature of IoT devices along with their lim-
ited computing and processing power make the prob-
lem quite intense. This becomes way more serious
in IoT-enabled healthcare facilities, especially emer-
gency rooms, critical care units, and operation the-
aters (Razdan and Sharma, 2022). IoT has revolution-
ized healthcare service delivery by connecting medi-
a
https://orcid.org/0000-0003-4916-3911
b
https://orcid.org/0000-0002-3247-3115
c
https://orcid.org/0000-0001-7381-1064
cal devices and systems that help monitor and manage
patient health more accurately, and efficiently, which
has never been possible before.
However, this advancement also carries risks that
must be addressed prior to IoT applications imple-
mentation as cyber-attacks might directly affect hu-
man lives in critical care and otherwise. The compro-
mise of CIA objectives taking place as a result of a
highly expected cyber-attack definitely results in im-
pacting normal and routine business operations of any
organization. In these circumstances, the organiza-
tion is considered to be prone to risks. A security
risk is usually evaluated as a function of the likelihood
of a certain threat agent exploiting any potential vul-
nerability and the consequent negative business im-
pact on the system or network (Roy, 2020). Organi-
zations adopt multiple ways to thwart such risks and
ensure protection against them with the implementa-
tion of certain security controls. The entire process of
managing and mitigating risks is called Information
Assurance (IA) (McIlwraith, 2021). Risk Manage-
ment is the fundamental component of any organiza-
tion’s IA program, as it helps in identifying the cur-
rent threats, vulnerabilities, and their associated risks,
with an aim to provide a cost-effective defensive and
information protective regime. The risk management
programs as shown in Figure 1 are meant to address
all the four types of security events (Interruption, In-
terception, Modification, and Fabrication).
Waqdan, M., Louafi, H. and Mouhoub, M.
A Comprehensive Risk Assessment Framework for IoT-Enabled Healthcare Environment.
DOI: 10.5220/0012060900003555
In Proceedings of the 20th International Conference on Security and Cryptography (SECRYPT 2023), pages 667-672
ISBN: 978-989-758-666-8; ISSN: 2184-7711
Copyright
c
2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
667
Risk Management
Risk Assessment
Identification of Risk
Analysis of Risk
Prioritization of Risk
Risk Control
Risk Management
Planning
Resolution of Risk
Monitoring of Risk
Figure 1: Risk management components.
In Table 1, we summarize the reviewed solutions
and framework related to IoT risk assessment. Over-
all, the existing research work focuses on Risk As-
sessment in an IoT-enabled healthcare environment
using contemporary approaches, where only the Risk
and Likelihood impacts are taken into account. They
do not take into account cases, where the number of
devices is variable and differs significantly. This was
the key factor that led us to develop the proposed
Risk Assessment framework for IoT-enabled health-
care environments, taking into consideration the im-
portant parameter of the number of devices thereby
giving more insightful risk scoring.
IoT devices in a healthcare environment are highly
prone to a number of cyber risks. The most significant
of these risks involve the security and privacy of pa-
tient data. Due to the high degree of interconnectivity,
there is an increased risk of someone gaining unau-
thorized access to patient data and affecting its CIA.
Moreover, due to the centralization of data, it has be-
come increasingly difficult for healthcare providers to
ensure the privacy of patient data. Another major risk
in IoT-based healthcare environments is their poten-
tial for malfunction or service disruption. As compo-
nents are interconnected, a disruption in one compo-
nent may affect the entire system, resulting in unin-
tended consequences. This is especially concerning
in critical care settings, where any error could have
dire consequences. Moreover, a potential risk is the
production of large amounts of data that are difficult
to interpret and use. The complexity of the data col-
lected may overwhelm healthcare providers if they are
not trained to properly interpret and use the data. This
becomes more serious when the number of devices is
increasing (Raghuvanshi et al., 2022).
In this paper, we propose a comprehensive risk as-
sessment framework for IoT-enabled healthcare envi-
ronments. Our framework takes into account all the
aforementioned concerns, and is meant to be used by
researchers and practitioners in cyber security related
to healthcare systems.
2 PROPOSED RISK
ASSESSMENT FRAMEWORK
The proposed risk assessment framework, for IoT
devices in an emergency healthcare environment, is
based on Risk Likelihood, Impact Likelihood, and the
Device Profile and Threshold level. Figure Figure 2
explains the overall flow of the proposed risk assess-
ment methodology.
Figure 2: Proposed Framework.
2.1 Risk Evaluation
Risk is not just a technical problem, it is also a man-
agerial one. Any asset in the organization, be it hard-
ware, software, services, human resource, etc., face
risks that challenge its confidentiality, integrity, and
availability. Risk is always an amalgamation of threat
to the asset, the likelihood of occurrence of that threat,
and the corresponding impact on business. From
these perspectives, the threat’s impact and its likeli-
hood are of immense importance as the business as-
pect in our case is linked to the healthcare sector’s
emergency facility. The broader risk evaluation for-
mula is given by Equation 1.
R
a
= W
a
× S
a
(1)
where, a represents any asset, W represents the risk
impact due to vulnerabilities and cybersecurity issues,
and S is the likelihood of that risk-taking place.
2.2 Risk Impact Evaluation (W)
Since the risk impact class W is a key factor in
determining the risk potential, and its overall impact
on the organization and asset, it is therefore funda-
mentally important to study the very causes of this
factor. The proposed risk evaluation framework takes
into account the following key factors that play a part
in the risk impact: - Network factor (N) - Protocol
factor (P) - Network design (D) - Device Level (L)
SECRYPT 2023 - 20th International Conference on Security and Cryptography
668
Table 1: Existing work on IoT security risk assessment.
Research
Work
Risk Detection Technique(s)
IoT Envi-
ronment
Strengths Weaknesses
(Jasour
et al., 2022)
Profile justification, risk as-
sessment model selection, se-
curity monitoring, model eval-
uation, validation, and update
Connected
and au-
tonomous
vehicles
Dynamic risk evaluation, Making use of
ML techniques
Not accounting for the number of IoT-
enabled devices. Validated only on the
network of vehicles
(Zhao et al.,
2021)
IoT-based threat situation
awareness architecture using
edge computing
Normal IoT
equipment
Use of ML for dynamic evaluation. Use
of dedicated device for threat analysis
The framework has only been tested for
threats generated as a result of network
traffic
(Matsuda
et al., 2021)
Penetration Testing
Industry 4.0
components
Use of AI and OPC Unified Architec-
ture along with penetration test for real
world monitoring of threats
The size of the IoT network has not been
accounted for in Risk evaluation
(Bahizad,
2020)
Risk Assessment considera-
tions for developing IoT de-
vices
IoT commu-
nication de-
vices
The proposed framework did discuss
risks due to increasing IoT devices in the
network
No new or novel risk assessment and
evaluation framework solution provided
(Wang et al.,
2020)
IoT device recommendation
mechanism for selecting
trusted participants
Smart city
environ-
ments
An intelligent device selective recom-
mendation mechanism along with game
theory based validation method.has
been proposed
Not very scalable for more IoT devices
dynamically being added and removed
from network
(Matheu
et al., 2020)
Real-world Risk assessment
strategy
Industry 4.0
IoT cyber-security certification frame-
work, integrating research and technical
tools have been discussed
Survey of existing frameworks; no novel
solution has been provided
(Oser et al.,
2020)
Helps selecting optimal/secure
IoT devices through user
awareness
IoT de-
vice(s)
selection
Focused on improved users’ ability to
assess the security of connected devices
Findings are based on survey/ interview
of one organization with only limited
number and types of IoT devices
(Datta,
2020)
SIEM based risk assessment
framework
Focuses on
five major
IoT attacks
An end-to-end IoT Platform with inte-
grated SIEM has been tested against five
types of cyber attacks
Focuses on DDoS attacks in a simulated
docker container environment instead of
a real IoT network
(Radanliev
et al., 2020)
Real-time intelligence, AI/ML
Industrial
IoT
Dynamic supply chain system inte-
grated with ML based real-time analyt-
ics has been proposed
Proposed system is suitable for only SME
with limited IoT devices
(Chen et al.,
2020)
Probability response and utility
attenuation behavioral model
Power grids
Behavioral model is proposed and tested
on IEEE RTS-79 system, making use of
probability responses
The proposed framework has only been
tested only for cyber-attacks on Power
grids
(Radanliev
et al., 2018)
Functional Dependency,
Network-based Linear Depen-
dency Modelling
Generic IoT
Use of mathematical formalism for IoT
Risk assessment along with the eco-
nomic impact of IoT attacks on the net-
works has been presented
No use of modern analytics, such as the
use of AI/ML
(Samad
et al., 2018)
Flexible four-step Risk eval-
uation model (justification,
model selection, training,
security monitoring)
Autonomous
Vehicles
(Child
Seats)
The proposed framework presents an
autonomous system, working at multi-
level of mobile cloud infrastructure
Tested on a Baby car seat, connecting
with cloud; thus limited numbers and
types of devices
(Nurse
et al., 2017)
Analysing dynamic and unique
IoT characteristics
CPS
A comprehensive survey of IoT charac-
teristics used in Risk Assessment
No new/novel framework presented
- Attack attributes (A), which are detailed in the
following:
Network Attacks/Issues (N): All the IoT networks
placed in a healthcare facility are potentially con-
nected to the Internet and are part of some network
to offer speedy, robust e-healthcare services. Being
part of the network(s) automatically makes them
susceptible to a wide variety of issues, which if com-
promised directly impacts the device’s functionality
and indirectly affects the overall IoT network. There
can be many network-related issues and attacks that
can generate and amplify the risk impact. Table 2
discusses a few network attacks, which may affect
IoT devices in a healthcare facility.
Protocol Issues (P): Another issue that directly af-
fects the risk likelihood is the protocol type used by
the IoT devices. Many IoT devices make use of
various protocols to communicate and perform their
network-related tasks. For instance, MQTT, ZigBee,
Bluetooth, and RFID are common protocols that are
used by IoT devices and, to some extent, have some
issues and are prone to certain protocol attacks.
Network Design Issues (D): The way the network
is designed and configured also directly impacts the
overall security. For instance, more intermediate
systems would result in more windows of opportunity
for the attacker, which significantly increases the risk.
In this case, not just the number of compromised
devices will increase but also the exposure factor of
the IoT network will augment. Therefore, this factor
is also fundamentally important in computing the
overall risk impact.
Device Level Security (L): The security at the device
level is also important, as different types of devices
have different security issues. A vulnerable device is
more susceptible to attacks as compared to a patched
and/or hardened one.
Attack Attribute (A): We are also quite interested
in the actual security attribute that has been com-
promised because, under different circumstances, the
importance of security attributes vary significantly.
For instance, Dos or DDoS attacks affect availability,
while replay attacks target confidentiality or integrity,
and ransomware targets availability. The violation of
CIA objectives has different repercussions under dif-
ferent scenarios.
Given all the aforelisted factors, we propose to
evaluate the risk impact for a given asset a, as
shown in Equation 2.
W
a
= (N + P + D + L + A)/5 (2)
A Comprehensive Risk Assessment Framework for IoT-Enabled Healthcare Environment
669
Table 2: Examples of Known Attacks and Their Impact.
Attacks Impact
Denial of Service
(DoS)
A DoS attack can directly bombard any IoT device with a large number of packets if not coun-
tered at the network’s gateway. Since DoS attacks deplete the ability of a device to respond to
normal/routine network probes, such as those related to a life-critical healthcare facility.
Distributed Denial
of Service (DDoS)
DDoS attacks are more serious as compared to DoS attacks because in this case, the victim is
targeted by a large number of machines (bots) making it tougher for the network gateway or
border gateway firewall to detect and stop. A successful DDoS attack can make the device in
particular and the network in general totally unable to perform its normal routine tasks.
Web Attacks
Since all IoT devices are part of a larger network, they do have an individual web component
from where they are configured, operated, and hardened. This makes them vulnerable to a
wide variety of web-based attacks, such as Remote Code Execution (RCE), Code/Command
Injection, and Malicious File Inclusion. Successful execution of these attacks severely impacts
the ability of the device to perform normally. Another major issue with web attacks is that
they are not detected at the network gateway by the border firewall because they operate at
the application layer of the TCP/IP stack. Therefore, a Web Application Firewall (WAF) is
required to provide the appropriate protection. Most WAFs have their own issues because of
their rule-based nature. They usually incorporate advanced techniques like deep learning for
detecting web attacks (Shahid et al., 2022).
where, N, P, D, L, A are the Network factor, Proto-
col factor, Network design, Device Level, and Attack
attributes, respectively. Note that, different weights
can be assigned to the different factors. However, for
simplicity reasons, in this paper, we assume that these
factors impact the asset equally.
2.3 Risk Likelihood Evaluation (S)
Apart from the direct threat to assets as a result of any
cybersecurity issue, the likelihood or the probability
of the vulnerability getting exploited and the threat
getting materialized matters a lot. Therefore, to
evaluate the likelihood of security risk to happen,
we identify three parameters, namely device attack
history (H), Device layer security (L), and Device
criticality (C), which are explained in the following.
Device Attack History (H): This parameter is also
important as it reflects the history of attacks that have
taken place on a particular device. For instance, an
IoT device in a healthcare facility might have been
attacked multiple times over a certain span of time.
A device that has been a victim of frequent attacks in
history is highly likely to get affected once again.
Device Layer Security (L): Since IoT is a multi-
layered architecture, where devices (network assets)
operate at multiple layers, therefore the likelihood of
attack varies significantly. For instance, resources at
the perception layer are far less prone to attacks and
eventual risks as compared to network resources at
the network or the application layer. Similarly, the
business layer is relatively safer, compared to the
network layer. Hence, there is a need to relate the
likelihood of an attack and eventual risk with the IoT
layer as well.
Device Criticality (C): Moreover, the more critical
the device, the more targeted it can be, and so its
compromise can be more fatal as compared to other
devices.
IoT devices in a healthcare environment, such
as a pacemaker and an insulin pump, are way more
critical, compared to blood pressure monitors and
blood sugar checkers, because they directly affect
human lives. Thus, their compromise would be
detrimental to the entire organization’s reputation.
Based on the aforelisted likelihood parameters, we
propose to evaluate the risk likelihood for a given as-
set a, as shown in Equation 3.
S
a
= (H + L +C)/3 (3)
where, H, L, and C represent the device attack history,
device layer security, and device criticality, respec-
tively. Similarly, different weights can be assigned to
different parameters. For simplicity reasons, in this
paper, we assume that these factors impact the asset
equally.
2.4 Attack Coverage (Number of IoT
Devices)
In most real-world situations, risk impact and risk
likelihood are not sufficient to evaluate the overall risk
to an asset and an organization. This becomes more
obvious when talking about IoT networks, where the
number of connected devices is much higher than
those in orthodox networks. Moreover, keeping in
view the network attacks we discussed earlier (e.g.,
SECRYPT 2023 - 20th International Conference on Security and Cryptography
670
DDoS attack), the number of devices plays a crucial
role in the magnitude of the attack.
2.4.1 Threshold (T )
We use the threshold to distinguish what the proper
parameters are when evaluating the security risk. The
threshold can be determined by system managers, se-
curity experts, or business owners. This can pro-
vide more flexibility when having many critical de-
vice profiles with very few device proportions in each
profile, which is the case in almost all IoT systems
(e.g., healthcare). In such a scenario, when the num-
ber of devices falls below the threshold, and the pro-
portion of devices in each profile is used as a parame-
ter to evaluate risk, the risk will not be accurate even
if these devices are critical (heart monitor or insulin
pomp) and under a potential attack. Thus, we further
elaborate on the importance of the number of devices
as a parameter and threshold with the help of the fol-
lowing scenarios.
1. Scenario 1: Let us suppose a particular attack has
taken place on any IoT device(s) that might not be
critical enough. Moreover, the likelihood of the
attack was also not very high. Under normal cir-
cumstances, such attacks will be ignored and their
corresponding risks will not be catered for. But,
what if the number of devices compromised as a
result of such an attack is very high? In this sce-
nario, such an attack cannot be ignored. Indeed,
because the number of infected/compromised de-
vices is so high, the corresponding risk needs to
be taken seriously.
2. Scenario 2: Suppose that a single but crucial IoT
asset in the organizational IoT network has been
threatened/attacked through a mechanism whose
likelihood is also very high. In this case, if the as-
set is isolated and not further spreading the attack
by acting as a pivot point, then the overall magni-
tude of risk should not be as inflated as it appears
to be when it is calculated through Equation 1.
Therefore, it is clear that the number of devices could
accentuate the need to come up with yet another pa-
rameter, which is also important in contributing to
the overall risk. We call this parameter Attack Cov-
erage, as it conveys the number of affected/infected
devices. Besides, it is important to normalize the ra-
tio of devices by assigning them some weights so that
the overall risk is uniformly evaluated across all IoT
devices in the organizational network. The new risk
formula after introducing the attack coverage (num-
ber of devices) is presented in Equation 4 and Equa-
tion 5, whereas Table 3 shows how device weights are
assigned, with respect to their ratios.
R
a
= W
a
× S
a
× w
a
(4)
where, w
a
represents the weight of the asset a.
Thus, the Risk to the asset a becomes now as
shown Equation 5:
R
a
=
(N + P +D + L + A)/5
×
(H + L +C)/3
× w
a
(5)
Table 3: Device Weight Allocation through Ratio.
Device Ratio Assigned Weights
60% 1
49% to 50% 0.7
20% to 39% 0.5
below 20% 0.2
Therefore, for each asset, the risk is evaluated and
a score, between 0 and 1, is returned. The latter is then
used to classify the risk into one of these categories:
very low, low, medium, high, or very high, based on
their criticality criteria presented in Table 4.
Table 4: Risk Ranges and Criticality.
Risk Values Criticality
0.8 - 1.0 Very high
0.6 - 0.8 High
0.3 - 0.6 Medium
0.2 - 0.3 Low
0 – 0.2 Very low
3 CONCLUSION AND FUTURE
WORK
The rapid growth of IoT creates new security risks for
organizations. In order to evaluate these risks, orga-
nizations need to develop a Risk Assessment Frame-
work for IoT as organizations have a dire need to
focus their efforts on creating a comprehensive risk
assessment framework for IoT. Such a framework
should identify, assess, and respond to the potential
risks of IoT, including the risk impact and its likeli-
hood. This could include selecting an appropriate se-
curity architecture, conducting an assessment of the
system, deploying protective controls, and regularly
testing the security of the system.
In addition, organizations should also consider the
physical security of IoT systems, as well as their abil-
ity to respond to any incidents. Organizations should
also consider the specific needs of each IoT system,
A Comprehensive Risk Assessment Framework for IoT-Enabled Healthcare Environment
671
such as application-level security, authentication, au-
thorization, and encryption.
Our risk assessment framework takes into account
the key aspects of risk impact, likelihood impact, and
the number of devices as well, which are the key
contribution of the framework. The number of de-
vices has a direct impact on the overall risk evalua-
tion, as they are directly linked with the device thresh-
old. Moreover, the proposed framework shows that
the risk faced by an IoT device changes if the device
threshold is modified.
In the near future, we are planning to validate the
effectiveness and usability of the proposed framework
on a simulated IoT-enabled healthcare system. We
will then expand our testing to other scenarios, and
ultimately test it with IoT datasets captured from real-
world test beds.
REFERENCES
Bahizad, S. (2020). Risks of increase in the iot devices. In
2020 7th IEEE international conference on cyber se-
curity and cloud computing (CSCloud)/2020 6th IEEE
international conference on edge computing and scal-
able cloud (EdgeCom), pages 178–181. IEEE.
Chen, B., Yang, Z., Zhang, Y., Chen, Y., and Zhao, J.
(2020). Risk assessment of cyber attacks on power
grids considering the characteristics of attack behav-
iors. IEEE Access, 8:148331–148344.
Datta, S. K. (2020). Draft-a cybersecurity framework for iot
platforms. In 2020 Zooming Innovation in Consumer
Technologies Conference (ZINC), pages 77–81. IEEE.
Jasour, A., Huang, X., Wang, A., and Williams, B. C.
(2022). Fast nonlinear risk assessment for au-
tonomous vehicles using learned conditional proba-
bilistic models of agent futures. Autonomous Robots,
46(1):269–282.
Koohang, A., Sargent, C. S., Nord, J. H., and Paliszkiewicz,
J. (2022). Internet of things (iot): From awareness to
continued use. International Journal of Information
Management, 62:102442.
Matheu, S. N., Hernandez-Ramos, J. L., Skarmeta, A. F.,
and Baldini, G. (2020). A survey of cybersecurity cer-
tification for the internet of things. ACM Computing
Surveys (CSUR), 53(6):1–36.
Matsuda, W., Fujimoto, M., Hashimoto, Y., and Mitsunaga,
T. (2021). Cyber security risks of technical com-
ponents in industry 4.0. In 2021 IEEE Interna-
tional Conference on Omni-Layer Intelligent Systems
(COINS), pages 1–7. IEEE.
McIlwraith, A. (2021). Information security and employee
behaviour: how to reduce risk through employee edu-
cation, training and awareness. Routledge.
Nurse, J. R., Creese, S., and De Roure, D. (2017). Secu-
rity risk assessment in internet of things systems. IT
professional, 19(5):20–26.
Oser, P., Feger, S., Wo
´
zniak, P. W., Karolus, J., Spagnuelo,
D., Gupta, A., L
¨
uders, S., Schmidt, A., and Kargl, F.
(2020). Safer: Development and evaluation of an iot
device risk assessment framework in a multinational
organization. Proceedings of the ACM on Interac-
tive, Mobile, Wearable and Ubiquitous Technologies,
4(3):1–22.
Radanliev, P., De Roure, D., Page, K., Nurse, J. R., Man-
tilla Montalvo, R., Santos, O., Maddox, L., and Bur-
nap, P. (2020). Cyber risk at the edge: current and fu-
ture trends on cyber risk analytics and artificial intelli-
gence in the industrial internet of things and industry
4.0 supply chains. Cybersecurity, 3(1):1–21.
Radanliev, P., De Roure, D. C., Nicolescu, R., Huth, M.,
Montalvo, R. M., Cannady, S., and Burnap, P. (2018).
Future developments in cyber risk assessment for the
internet of things. Computers in industry, 102:14–22.
Raghuvanshi, A., Singh, U. K., and Joshi, C. (2022). A re-
view of various security and privacy innovations for
iot applications in healthcare. Advanced Healthcare
Systems: Empowering Physicians with IoT-Enabled
Technologies, pages 43–58.
Razdan, S. and Sharma, S. (2022). Internet of medical
things (iomt): overview, emerging technologies, and
case studies. IETE Technical Review, 39(4):775–788.
Roy, P. P. (2020). A high-level comparison between the
nist cyber security framework and the iso 27001 in-
formation security standard. In 2020 National Confer-
ence on Emerging Trends on Sustainable Technology
and Engineering Applications (NCETSTEA), pages 1–
3. IEEE.
Samad, J., Reed, K., and Loke, S. W. (2018). A risk aware
development and deployment methodology for cloud
enabled internet-of-things. In 2018 IEEE 4th World
Forum on Internet of Things (WF-IoT), pages 433–
438. IEEE.
Shahid, W. B., Aslam, B., Abbas, H., Khalid, S. B., and
Afzal, H. (2022). An enhanced deep learning based
framework for web attacks detection, mitigation and
attacker profiling. Journal of Network and Computer
Applications, 198:103270.
Wang, B., Li, M., Jin, X., and Guo, C. (2020). A reliable
iot edge computing trust management mechanism for
smart cities. IEEE Access, 8:46373–46399.
Zhao, Y., Cheng, G., Duan, Y., Gu, Z., Zhou, Y., and Tang,
L. (2021). Secure iot edge: Threat situation aware-
ness based on network traffic. Computer Networks,
201:108525.
SECRYPT 2023 - 20th International Conference on Security and Cryptography
672