IEEE Consumer Electronics Magazine - January/February 2024 - 14

Special Section on Security and Privacy-Aware Emerging Computing
The rest of this article is organized as follows.
In the " RelatedWork " section, we discourse about
related works in the literature, and present a comparison
involving them and the proposal in this
article. The " Trust Modeling " section presents the
trust components of our model and its verification.
Some open research issues are discussed in
the " Open Research Challenges " section, and the
" Concluding Remarks and Future Work " section
concludes this article.
RELATED WORK
The literature presents many proposals to
model a trust metric applied to IoT. Wang and Vassileva15
proposed a Bayesian network-based trust
model and amethod for building reputation based
on recommendations in peer-to-peer networks.
They show that systems that communicate
according to trust and reputation outperforms
systems that do not consider such aspects in
terms of the percentage of successful interactions.
Fortino et al.16 developed a framework based on
software agents able to exploit social attitudes of
IoT devices. They use blockchain (BC) to compute
the reputation of devices, which can use network
services according to their reputation.
Tang et al.17 used the passport analogy to
propose a decentralized trust framework for
cross-platform collaborations using BC technology.
They use a combination of smart contracts,
with each interaction among devices signed by
the participants and recorded on the BC. In our
approach, we focus on recording the devices'
identities in BC to build devices' reputations (initial
trust). Chen et al.18 focused on wireless sensor
network applications and present a trust
model using weighted averages considering the
following three types of trust: direct, regional,
and historical data trust. Through simulations,
authors show that their approach is aware of
energy consumption with a significant reduction
of resource usage and a higher detection rate in
comparison with other approaches.
Hongjun et al.19 used information theory to
build trust among devices. They represent the
relationships with a directional graph and compute
the entropy of the capability of a device to
perform an action. This way, they can detect
malicious devices in the network. We also adopt
the information theory, but with a different
14
perspective focusing on the network level
instead of the application level.
Khan et al.20 proposed a trust-based approach
for managing the reputation of every device of an
IoT network based on routing protocol for low
power and lossy networks (RPL). The results show
the ability to detect and isolate malicious nodes
from the network, resulting in better network resilience,
as well as less number of misbehaving devices
(bad devices) identified in the network after
every RPL round, and less number of paths including
a bad device. Caminha et al.21 introduced a
smart trust management method based on
machine learning, which automatically assesses
the IoT trust by evaluating service provider attributes.
They also use an elastic slide window feature
that helps us to differentiate broken or malfunctioning
devices amongmisbehaving devices.
Bernal Bernabe et al.22 used fuzzy logic to provide
an end-to-end security solution through a
lightweight authorization mechanism and a novel
trust model that has been specially devised for IoT
environments. Their approach considers four
dimensions, i.e., quality of service, reputation,
security aspects, and social relationships, to compute
trust values about IoT devices. Zhou et al.23
considered an identity-based encryption (IBE)
implementation together with a BC implementation.
Authors split the devices in the chain to complete
user authentication and private key
protection. The results show the failure probability
is stabilized with the number of cycles during
which a device operates. Wei et al.24 employed a
BC approach to address the trust management
issue in the context of social IoT. They compute
the trustworthiness value of devices according to
their service evaluations stored in BC. Different
from our approach, they do not consider network
aspects that also impact trust values. Puthal et al.25
proposed a decentralized security framework
based on BC that could be applied to IoT systems,
among several other consumer applications,which
provides a newway of managing trust.
Table 1 shows a comparison between previously
discussed works and our approach in
terms of the following aspects:
1) which technique is used;
2) whether a dataset is considered and, if so,
what kind of dataset;
IEEE Consumer Electronics Magazine
IEEE Consumer Electronics Magazine - January/February 2024

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - January/February 2024
