Using Blockchain Technologies to

Solve Trust Issues for the

Internet of Things (IoT)

Submitted on March 17, in partial fulfilment of the conditions of the award

of the certificate course
in
Cyber Security

By
Gideon Nuertey

Supervisor Mr. Ampofo

Date: 17/03/2023

The due acknowledgment must always be made of the use of any materials contained in or
derived from this thesis”.
Acknowledgments

Whilst studying at IPMC, I have received an excellent amount of support from every member
of staff I have encountered at the school.

I would like to thank all members of staff for helping me extend my knowledge throughout
my time at the Institution. I would like to especially thank my supervisor, Mr .Ampofo for his
amazing continuous support and guidance throughout the course.

Abstract
As devices deployed in the Internet of Things (IoT) continue to increase their connectivity
and intelligence, so too will our demand for them to autonomously shape networks, exchange
data, and perform specific behaviours on our be halves.
Traditional privacy and security methods are highly inapplicable when looking to secure the
IoT, due to not only the technology's decentralized topology but also the resource constraints
posed on the majority of smart devices.
Blockchain has risen to prominence as one of the most significant disruptive technologies of
this century. Originally developed to expedite the bitcoin cryptocurrency, the technology has
now been suggested as an answer to the IoT's privacy and security vulnerabilities (Dorri et al,
2016).

While there is exceptional potential, blockchain technologies must grow and evolve
considerably to meet the numerous unique requirements of the Internet of Things (IoT).
Therefore, the primary aim of the ensuing research project will look to carry out an extensive
analysis of the major trust issues that currently plague smart devices in the Internet of Things.
This analysis will be undertaken in an attempt to understand whether blockchain
technologies, as well as other methods, can be used as a suitable approach to address the
identified issues. By evaluating the strengths and limitations of both blockchain technologies
and other approaches, future work will be recommended.

Page 1
Table of Contents

Introduction .......................................................................... 4
Background ........................................................................ 5
Research Project Chapter Breakdown ............................. 7
Aims and Objectives ............................................................. 8
Aims and Research Questions ........................................... 8
Objectives of the Study ...................................................... 9
Literature Review ............................................................... 11
Trust Issues and Trust-Related Attacks of the Internet of
Things (IoT) ...................................................................... 11
Approaches to Address the Internet of Things (IoT)
Device Trust Issues .......................................................... 13
Using Blockchain Technologies to Address the Internet
of Things (IoT) Device Trust Issues ................................ 15
Methodology and Approach .............................................. 17
Results and Findings .......................................................... 18
Comparative Analysis ...................................................... 19
Proposed Design Guideline for New Blockchain
Approach .......................................................................... 22
Analysis and Conclusion .................................................... 22
Limitations and Impact of Research .............................. 23
Summary of Contributions ............................................. 23
Future Work..................................................................... 24
References ........................................................................... 25

List of Tables

Table 1: Blockchain Approaches Compared

Page 2
Fig 1.0 on the infusion of Blockchain in IOT in our daily lives

Page 3
Introduction

In recent years, the Internet of Things (IoT) has experienced remarkable growth in both
research and industry. Many have even labelled the rise as the "fourth industrial revolution",
with Gartner estimating that over 25 billion connected devices will be in use across the globe
in the next 18 months (Gartner, 2019). The detailed knowledge gained by such smart devices
will continue to improve consumer efficiencies and deliver progressive services to a wide
range of industries including pervasive healthcare and smart home services. However, with
all the promise the IoT hopes to deliver, smart devices are known to still suffer from
consistent trust issues. In the Internet of Things (IoT) availability, privacy/security, and
integrity have been defined as the main elements of trust (Numminen, 2016).
Smart devices use a plethora of networking technologies such as ZigBee IEEE802.15.4, Z-
Wave, Wi-Fi 802.11 etc. in the IoT, to communicate with other devices and offer a unique,
ubiquitous service to the consumer. Such devices continually exchange data, creating a
heterogeneous network of devices. However, the problem lies in the integrity of the
information collected. Although smart devices provide an improved level of convenience, it
would be wrong to suggest they are in constant use by the consumer. The fact that such
devices are left disregarded for large periods of time, whilst communicating using wireless
protocols can lead to them becoming highly susceptible to physical attacks.

As a result of this, data risks falling into the hands of malicious participants. When the
devices deployed in the Internet of Things plan and carry out tasks based on our actions, there
has to be a solid reassurance that the information they utilize to make such decisions is
trustworthy. However, assuring the data is trustworthy can prove to be extremely difficult,
especially when a central authority coordinates device configuration, data collection/cleaning
and data dissemination (Bowman & Morhardt, 2018). Nonetheless, distributed networks are
unable to utilise a central authority. This is since smart devices act on the instructions of
various owners whom all have specific motives are produced by various manufacturers, often
execute various operating systems and interact with different protocols. As a result of these
factors, conventional methods to affirm and authenticate user identity and integrity
continually fail. The challenge, therefore, is to devise a framework that establishes trust
between communicating devices when exchanging information with each other.

Page 4
Background
Trust is viewed as a level of confidence. It is considered at two levels; whether a smart device
trusts the data/information it is receiving or whether a human truly trusts the smart devices
that it uses. This research project will look to focus on the latter; however, both are equally as
important. Furthermore, the research project will look to analyse the many explicit trust
issues that plague IoT ecosystems, when devices look to exchange data. Ultimately once
identified, the research will it look to focus on the most common trust issue, establishing a
common root of trust that can be accepted by all parties within an IoT environment.
Moreover, blockchain technologies will then be analysed to understand whether the
technologies properties can be used as an adequate solution to address the trust issue
explained above.

The best way to introduce this trust issue is through the use of a scenario. A service provider
proposes a service, coupled with its terms of use. For example, consider a little lunchtime
restaurant who looks to provide consistent Wi-Fi service to their customers for a small
supplementary fee. Presently large-scale operators and restaurant chains possess bilateral
agreements that cater to this, however, there is no such clarification that fits little restaurants.
Therefore, a solution that grants anyone the ability to utilise these services easily by
implementing a public obligation for achieving the terms of use as specified by the service
provider would be welcomed. Once the suitable practices have been met, it is not expected of
the service provider to automatically obtain an obligation from any consumer, but it should
first assess the overall risk obtaining that obligation. Generally, a considerable portion of
possible consumers may not possess the correct levels of reputation to access the service.

Blockchain which is primarily based on the distributed paradigm was first introduced in
2008. However, in recent years the emerging technology has received outstanding praise and
success as it has been touted to take the technological world by storm. In short, blockchain
technologies are viewed as digital ledgers that remain at the centre of decentralised
environments and in return, such ledgers are tasked with keeping track of all changes to the
system by taking a record of each transaction. Blockchain gained its notoriety with the
introduction of bitcoin. Bitcoin is a distributed cryptocurrency payment method designed
solely as a peer-to-peer version of electronic cash as well as possessing the ability to provide
secure direct payments from the payer to payee.

In a more detailed manner, blockchain technologies can be viewed as an organised and back-
linked list of blocks responsible for transporting transactions which conceal exchanged data
between two or more people. Each singular block contains a collection of transactions and is

Page 5
linked to the preceding block in the blockchain, ultimately establishing a sequential order of
blocks that in return builds a sequential order of transactions. Furthermore, peers work to
construct and preserve the blockchain via various consensus algorithms, hence accomplishing
a state machine replication system, which is the basis of any blockchain technology.

The promising uses of blockchain technology go way beyond Bitcoin. Blockchain has the
following characteristics:
- Decentralised Control: A decentralised framework, where there is no principal entity
that dictates instructions.
- Data Transparency and Auditability: An entire copy of every single transaction
ever completed within the system is securely stored in the blockchain and can be
publicly viewed by all accepted peers.
- Distribute Information: All nodes attached to the network maintain a complete copy
of the blockchain to avert having a centralised authority confidentially keeping all that
data.
- Decentralised Consensus: Instead of a central authority validating transactions, they
are validated by all of the nodes attached to the network. Moreover, this stops with the
paradigm of centralised consensus.
- Secure: The blockchain itself is known to be completely tamper-proof and so,
therefore, cannot be exploited by malicious users.

Above are just some of the leading benefits that blockchain technologies have to offer. The
protected, decentralised and independent nature of the blockchain makes it an optimal
component in becoming an essential element of future IoT approaches.

Page 6
Research Project Chapter Breakdown
The ensuing dissertation is structured as followed:

- Introduction: This section will be used to introduce the IoT, how it works and what
makes the technology susceptible. Furthermore, a brief description of blockchain and
its key components will be given.

- Aims and Objectives: The dissertation looks to answer a few main research
questions as well as a plethora of aims and objectives, those will be stated in this
section.

- Literature Review: The literature review will be split into three sections. Trust issues
and trust-related attacks in the IoT, which will look to establish the main trust issues
that plague devices within the IoT. Additionally, the review will then go on to identify
and critically analyse existing blockchain approaches into addressing such trust
issues. Finally, a comparison of other approaches separate from the use of blockchain
will be given.

- Methodology and Approach: This section will be primarily used to describe and
analyse the two different literature review approaches taken.

- Results and Findings: This section will be used to compare and analyse the existing
blockchain approaches proposed to address the specified trust issue, to assess to what
extent has the trust issues been resolved. Identifying the benefits and limitations, as
well as the technical differences of each approach will help propose guideline for a
new or improved blockchain approach.

- Analysis and Conclusions: Conclusions will be drawn on the limitations and impact
of the research as well as suggestions for future work.

Page 7
Aims and Objectives

Aims and Research Questions

As mentioned previously in the Introduction, due to the unpredicted development of
connected devices in the IoT era, the importance of evaluating trustworthiness among IoT
devices has grown. Therefore, the ensuing dissertation looks to achieve three main aims.

- Firstly, the research project looks to understand the major trust issues that smart
devices face whilst interacting in the IoT and the attacks that can occur due to a lack
of trust. Although a thorough analysis of existing trust issues will take place, the
dissertation will primarily focus on the problem of establishing a common root of
trust within an IoT ecosystem.
- Additionally, once such issues have been established the work will evaluate the
approach of using blockchain technologies as well as other approaches to address
these trust issues.
- Ultimately, comparing the strengths and limitations of existing blockchain approaches
to promoting trustworthiness, will allow the work to propose a guideline for a new or
improved blockchain solution that can be viewed as an alternative to establishing a
shared/central root of trust.

As a result of the above, the following primary research questions have been constructed in
an attempt to meet the aims of this work:

"What are the main trust issues smart devices face whilst exchanging data on the
Internet of Things (IoT)?"

"Can blockchain technologies be used to address these issues?"

The subsequent sub-research questions have been constructed to obtain a deeper

understanding of the primary research questions.

"What attacks can occur due to a lack of trust between devices?"

"Should particular trust issues take priority over others?"

"What are the core components of blockchain technologies?"

"What specific blockchain components are utilised to achieve trust between devices?"

Page 8
"What are the existing approaches to building trust using blockchain technologies?"

"Apart from using blockchain technologies, are there other approaches can be used to
build trust?"

Objectives of the Study

To sufficiently answer the above research questions, a further six objectives have been
constructed to help the dissertation achieve its three main aims.

1. Carry out a thorough analysis of existing IoT trust issues.

To truly understand the problems that devices face when looking to building trust in an IoT
environment, it is important to carry out a thorough analysis of existing trust issues.
Furthermore, identifying such trust issues will allow the researcher to gain an initial
understanding of why blockchain technologies have been viewed as an adequate solution. In
order to truly achieve this objective, numerous secondary resources will be used. These
include; published research articles, webpages and textbooks.

2. Investigate modern blockchain technologies.

Blockchain technologies were introduced over a decade ago, however, to the average
consumer it is still viewed as a relatively new concept. It is extremely important that a
comprehensive understanding of how blockchain technologies work is gained. Each
component of the blockchain architecture will be carefully scrutinised. In return, a decision
can be made on how effective emerging technology can be when overcoming IoT trust issues.
Again this objective will be achieved through the use of research journals, textbooks and
manufacturer information.

3. Identify and establish the existing blockchain technology approaches used to

address identified trust issues.

The main research question argues whether blockchain technologies can be used as an
alternative to address the identified trust issues. Therefore, this objective will focus on the
various ways in which blockchain technologies can be integrated into an IoT ecosystem in an
attempt to provide trust between devices. A literature review will be conducted, primarily
using published research journals to dissect existing blockchain frameworks proposed by
researchers.
High importance should be placed on completely understanding why these researchers
believe blockchain to be an adequate solution and exactly why they have used a specific
component of the blockchain architecture to address the issues.

4. Evaluate the benefits and limitations of integrating blockchain technologies into

the IoT to solve trust issues.

Page 9
Although blockchain provides plenty of promise, a critical analysis should be undertaken that
identifies both the benefits and limitations of integrating blockchain into the IoT. Critically
evaluating the approaches proposed by these researchers through the use of a literature
review, allows to truly recognise why blockchain offers such promise. Furthermore, using
blockchain may solve the problem of building trust between devices but it may result in other
issues arising. Therefore, this possibility should be explored. Ultimately, if this objective is
completed the approach of using blockchain technologies to address the trust issues can be
sufficiently evaluated.

5. Compare other approaches to address trust issues with blockchain.

The use of blockchain for building trust in the IoT is not the only method of doing so. Other
approaches should be researched to see if they offer better promise than blockchain.
Furthermore, there is a possibility for aspects of these approaches to be combined with
blockchain technologies so the potential of this should be explored. In order to truly achieve
this objective, numerous secondary resources will be used. These include; published research
articles, webpages and textbooks.

6. Research new or improved blockchain approaches to propose a new guideline

for building trust in the IoT.

It can be argued that this is the most important objective, but it will prove the hardest to
complete. The information gained in objective four will form the basis for this objective.
Identifying the strengths and weaknesses of existing IoT blockchain trust mechanisms will
not only provide the relevant knowledge on the topic but it also exposes the main gaps in
research. Therefore, the proposed new or improved blockchain solution will look to not only
improve on the limitations found in existing works but to build on the strengths found as
well.

The aims and objectives have been made to be specific and measurable, ultimately making
them achievable. Adequately achieving the six stated objectives will assist in reaching the
ultimate goal of the ensuing dissertation. It should be noted; further research will be
conducted outside of these objectives but they provide a firm basis to work against.

Page 10
Literature Review

Trust Issues and Trust-Related Attacks of the Internet of Things (IoT)

As the Internet of Things (IoT) continues to grow, one of the greatest hurdles for this
disruptive technology is to bridge the currently fragmented trust domains. It is no secret that
the IoT continues to experience exponential growth, however, for the technology to realise its
true potential, the billions of devices currently connected to the Internet must be able to
consistently interact with each other.
Therefore, this section looks to analyse existing works on trust issues mentioned in the
introduction as well as introducing other issues.

Nabil et al (2018) understand that building secure trust relationships between devices in the
IoT must be viewed as a fundamental requirement, however, they appreciate implementing
adequate trust management in a ubiquitous environment can represent a true threat.
The work of Di Pietro et al (2018) appreciates that more and more devices are becoming
completely connected, nonetheless, the overall goal of the IoT is still far from being
adequately accomplished as smart devices still fail to communicate with one another. It is
important to note that there are several reasons why such devices are not able to correspond
with each other, however, the major reason identified is the absence of trust between devices,
which is crucial when building secure communication in the IoT. Di Pietro et al clearly
understands that conventional trust frameworks typically used to equip the Internet are not
compatible with the fit the scale and diversity of the IoT ecosystem, as there is no common
root of trust (centralized authentication system) that is accepted by all parties involved. As a
result of this, we are exposed to numerous domains, in which device manufacturers
implement a root of trust that primarily allows smart devices located within each domain to
interact securely. The work of Di Pietro has categorised these domains as ‘Islands of Trust',
as in each domain the trust is solely administered and monitored by an entity completely
separate from the entities providing the trust for other domains. Many companies have tried
to design their industry-standard consortiums in an attempt to address this issue by agreeing
on a common root of trust, most notably the Open Connectivity Foundation (OCF).
Nonetheless, OCF, as well as numerous other companies, fail to cover the full IoT landscape.
Therefore, the present IoT landscape consists of separate manufacturers and platforms that
are only able to communicate safely if a common root of trust is mutually agreed upon
(through the use of a consortium) or if direct mutual trust is approved through a bilateral
agreement.

Ning et al (2013) go on to explain that IoT devices perform their tasks at three separate
levels. Firstly, a physical level, (aka perception level) is solely accountable for sensing the

Page 11
environment around the smart device. To carry out such sensing responsibilities a range of
technologies are used such as; Radio Frequency Identification (RFID), Near Field
Communications (NFC), infrared as well as many others. Furthermore, the next level is
known as the network level. This layer is used to generate network communication through
the use of interfaces and routers. Finally, the application level is primarily responsible for
implementing "context-aware intelligent services in a pervasive manner" (Yan et al, 2014).
With that all being said, for trust to be achieved, it must be generated vertically throughout
the system not just at each level.

This is a view shared by Fernandez-Gago et al (2017). The authors believe that two major
aspects hinder devices building trust with each other; interoperability and dynamicity. Within
an IoT environment, heterogeneity is viewed as a very important component. Smart devices
possess contrasting storage, computational and communication capabilities as well as using
various protocols to interact and collaborate. Nonetheless, different trust management
protocols can be utilised in IoT devices. As a result of this, sufficient solutions should be
provided to assist trust management systems to interact, interoperate and work cohesively in
a persistent manner despite the differences explained above. Furthermore, IoT environments
are known to be extremely dynamic. Smart devices are constantly being connected and
disconnected from a network, therefore changing the state of that network ultimately
disrupting communication. In a situation like this, the trust management system must have
the ability to update their settings, so devices can adapt to any significant changes in an IoT
environment.

Apart from the problems explained above, Guo et al (2017) believe that another major issue
is that a lack of trust between devices in an IoT environment can lead to several different
attacks. In particular, a compromised node possesses the ability to carry out eavesdropping,
forgery, denial-of-service, man-in-the-middle attacks as well as many others. However, some
malicious attacks have been specifically designed to only arise when a particular node is
trying to determine the trust value of another node in an attempt to disturb the trust system.

Trust is widely viewed as a notion that an entity is truthful and therefore will not pose a threat
towards other entities. However, it should be noted that this notion is biased and based on
past experiences. Nevertheless, reputation is generally perceived as an entity's behaviour
based on the levels of trust other entities have established. Reputation systems are known to
compile and distribute feedback on how an entity has behaved in the past, therefore
reputation is viewed as a much more objective notion than trust (Resnick et al, 2000).
Consequently, the primary objective of a Trust and Reputation System (TRS) is to ensure that
any action undertaken by an entity within the system must reflect their known reputational
values, and thus their reputation cannot be mimicked by any unauthorised parties. Typically,
a TRS consists of entities (end devices), observers, disseminators and reputation servers. End
devices tend to only use reputation, instead of processing them directly. Observers are
assigned to TRS's components which generate and handle the trust leveraged by entities (end
devices). Observers also possess the ability to build trust data based on historical information
of these end devices and in some cases, they can even cooperate among them. Data traded
between observers or sent to other end devices consistently rely upon safe and trusted
communications, carried by the disseminators. (Fraga et al, 2012).
Be that as it may, despite TRS effectiveness when bridging trust between secure domains,
they are susceptible to a variety of attacks. As explained above, a TRS depends on the
successful collaboration among distributed entities. With that being said, cooperation

Page 12
amongst these entities can be easily damaged, resulting in the following trust-related attacks,
identified by Guo et al (2017):
- Bad-Mouthing Attacks (BMA): This specific attack occurs when a malicious entity
exploits the reputation of a trusted entity by providing bad recommendations.

- Ballot Stuffing Attacks (BSA): Also known as good-mouthing attacks, BSA is

similar to bad-mouthing attacks. However, where it differs is the fact malicious
devices group together to trigger the attack.

- Opportunistic Service Attacks (OSA): In this attack, the malicious device attempts
to keep the reputation high by providing a good service. This is intending to misguide
the TRS by carrying out both bad/good mouthing attacks.

- On-Off Attacks (OOA): Has proven to be one of the hardest threats to detect. In this
attack, the compromised device switches between providing a good and bad service.
The aim of this is to keep the reputation of the malicious node high so it can comprise
the network by implementing a good recommendation for other malicious devices or
bad recommendations for previously trusted devices.

Furthermore, Sun et al (2008) have identified a further two more attacks that are known to
disturb trust systems:
- Selective Behaviour Attack: This attack consists of a compromised device having
the ability to be viewed as a trusted device by some of its neighbours whilst being
perceived to be malicious to other devices. Therefore, the average recommendation
with stay positive whilst still being able to conflict damage to particular devices.

- Sybil Attack: In terms of authentication and access control breaches, a device can
generate, imitate or mimic different devices within the network. As a result of this, the
malicious node can exploit the recommendations, promoting itself as a trusted device.
Furthermore, a Sybil attack gives a compromised device the ability to rescind its bad
reputation by generating a completely new identity.

Approaches to Address the Internet of Things (IoT) Device Trust Issues

The IoT offers a plethora of benefits to consumers, however, the technology's security
continues to pose a major problem when implemented in real-world scenarios. As explained
previously, in the IoT cyber attackers have the ability to hijack communication between
sensors or precisely control a sensor. Consequently, many approaches for building secure
trust relationships between devices in the IoT have been studied in the literature regarding
IoT trust issues.

In the last few years, many researchers have looked to focus on the importance of combining
traffic filtration and sampling when analysing trustworthiness between IoT devices. In
particularly, Meng (2018) who recognises that building a secure trust map for IoT nodes is
imperative. The author concludes that trust exhibits the level of confidence in which a device
will agree to another device. Nonetheless, as the number of IoT continues to rise, it is
becoming increasingly harder to detect malicious attacks. Meng (2018) proposes
implementing traffic-based intrusion detection mechanisms as a means to analyse the
trustworthiness of devices. Furthermore, he states that trust computation that relies on traffic

Page 13
status is a highly capable mechanism for Intrusion Detection Systems (IDSs). Typically,
traffic-based trust computation contributes to an abundance of benefits for IoT environments.
Some of which include:
- Simple Implementation: Network Intrusion Detection Systems (NIDSs) manage and
verify all traffic coming in and out of the network, therefore building a trust map
between numerous IoT nodes based on collected traffic data will prove to be easy.

- Detection Accuracy: As mentioned by Fernandez-Gago et al (2017 IoT nodes can

dynamically join or leave the network, resulting in making detection harder.
Nonetheless, the majority of IoT nodes have to interact with other nodes, therefore the
traffic generated from such nodes can be compiled across various nodes.
Consequently, the collection of such mutual data can lead to high detection accuracy.

- Error Tolerance: Due to the false rates that Intrusion Detection Systems possess,
some devices may prompt false alerts. Furthermore, trust computation via traffic
managing for an extended time can establish error tolerance that decreases the effect
of mistakes made during the examination.

In conventional network environments, the process of using traffic-based trust computation to

detect insider attacks has proven to be an effective technique. Be that as it may, network sizes
continue to develop at an expeditious pace. The amount of traffic produced vastly eclipses the
maximal processing capability of any detection system.
In return, overhead traffic can lead to an intrusion detection system dropping numerous
packets without them being thoroughly examined. Additionally, in the era of IoT, such a
scenario would render the use of traffic based trust computation inadequate when analysing
trustworthiness between IoT devices.
To combat this problem, Meng (2018) proposes the use of traffic filtration and sampling
when looking to decrease the amount of packets that will be managed by an IDS.

Traffic Filtration: The primary objective of a traffic filtration mechanism is to diminish the
target packets managed by an Intrusion Detection System. This method can be used in an
initial stage to enhance the efficiency of trust computation. Traffic filtration has been
categorised into two types;
- Blacklist-based – The system creates a predefined ‘blacklist' which is based on a
specific set of security rules, therefore if the source IP address of an incoming packet
equals an item in the ‘blacklist' that packet is blocked.
- Whitelist-based – Whitelisted packets can gain direct access to the target network.

Traffic Sampling: Traffic sampling can be utilised to manage the remaining packets,
ultimately being viewed as the second stage to increase the performance of trust computation.
Sampling is mainly used to plan information about a specific characteristic of the parent
population at a minor cost than a full examination.

Traffic filtration and sampling propose great potential for assisting IoT devices to securely
communicate with each other. Nonetheless, various limitations exist due to the IoT's dynamic
nature. In regards to traffic filtration, organisations are generally able to determine both a
whitelist and blacklist. However, in an interconnected world like the IoT, it is challenging to
build a full list due to the vast amount of IoT devices. As a result of this, implementing traffic
filtration effectively in an IoT environment can prove to be difficult. In terms of traffic
sampling, it is important to explore the accuracy and workload of both packet/flow-based

Page 14
sampling. The analysis clearly shows that flow-based sampling can produce better accuracy,
but the issue resides on how to manage its required resources.
In conclusion, Meng (2018) introduces an encouraging solution to enhance the performance
of trust computation in an ever-growing IoT world via suitable traffic sampling and filtration
techniques. With that being said, the focal point of the authors' future research should focus
on examining how to lessen false rates of detection, as well as finding the correct balance
between performance, sampling and filtration.

In other related works, Chen et al (2008), proposes a trust management mechanism that
utilises a watchdog scheme to oversee the node's behaviour and calculate trust ratings.
Whereas, Kwok et al (2013) design a trust management framework primarily based on a
Bayesian model through analysing the status of packets and establishing harmful nodes in
terms of a threshold.

Using Blockchain Technologies to Address the Internet of Things (IoT)

Device Trust Issues
The main issues that IoT devices face whilst trying to build trust and communicate with each
other have been identified. Moreover, having carried out a thorough critical analysis of
existing approaches to address such issues, it is clear they all offer potential promise.
However, another main motivation for this research project is to evaluate whether blockchain
technologies can also be viewed as an effective solution to these issues.
In recent years, many works have introduced original frameworks that can be used to solve
security issues that devices face when communicating within an IoT ecosystem. Therefore,
this section will be used to critically analyse these existing works to evaluate the
effectiveness of using blockchain technologies to address trust issues and mitigate trust-
related attacks.

As mentioned in the work of Di Pietro et al (2018), due to the full autonomy of the IoT
devices must possess the ability to recognise and authenticate one another to assure the
integrity of the data exchanged with each other. The work of Hammi et al (2018) recognises
that it is virtually impossible to implement a competent centralized authentication system.
Therefore, in order to combat this problem, the authors introduce an original decentralised
system named ‘bubbles of trust'. Hammi et al (2018) utilise the numerous security advantages
provided by blockchain technologies to generate protected virtual zones, (also known as
bubbles), where devices can identify and communicate with each other in a secure manner.
Ultimately resulting in devices establishing trust with each other. The approach examined in
this ensuing journal article, not only ensures that devices within an IoT environment undergo
a vigorous identification and authentication procedure but the framework also safeguards the
integrity and availability of data exchanged.
The authors should be praised for testing and applying their approach to a vast amount of IoT
use case scenarios. Testing the framework in a vast amount of scenarios gives a clear picture
of its adaptability which is very important. Furthermore, the proposed approach is tested
against some trust-related attacks, with the framework successfully mitigating such attacks.
Although this approach offers promise, it must address three fundamental issues to be
implemented efficiently. Firstly, the approach has not been adapted to real-time applications
as it relies on a public blockchain. However, this problem can be dealt with if a private
blockchain is used. Secondly, for the approach to work, it needs an initialization phase. As
mentioned the analysed framework relies on the use of a public blockchain, which involves

Page 15
costs that solely depend on the specific cryptocurrency used by the blockchain system.
Consequently, the rapid evolution of cryptocurrency rates may hinder the successful
implementation of this approach

Within the majority of IoT ecosystems, trust is presumed implicitly. However, this implicit
assumption of trust is where the problem lies, as this trust can be abused by malicious
attackers. Previously stated above, Sun et al (2008) have identified Sybil attacks as one of the
major attacks that can occur when the adequate levels of trust have not been met. The work
of Asiri & Miri (2018) recommend an IoT trust framework that primarily utilises
permissioned blockchains that use smart contracts to assess the trustworthiness of IoT nodes
by recording and substantiating IoT nodes' identities to restrict Sybil attacks. A strength of
this proposed framework is its ability to adapt to the ever-changing needs of IoT, as it takes
full advantage of decentralized processing. Additionally, this allows for the proposed
approach to be scaled to fit any kind of IoT ecosystem, without the hindrance of being
constrained to a certain context. Although the proposed framework is viewed as an efficient
solution, the approach does suggest improvements for future work. Within this proposed trust
model, one certificate authority (CA) is responsible for managing Member Service Providers
(MSP) responsibilities (e.g. aiding registrations of new IoT devices). Furthermore, the
handling of these cryptographic keys can lead to the depletion of network resources, as the
original CA is the only authority capable of performing such a task. In order to improve this
framework, the authors should add secondary CA's that inherit that authority. Distributing
key management does not only greatly decrease network overheads, but it is also sure to
improve performance as well.

Similarly, Huh et al (2017) also propose an approach utilising blockchain smart contracts.
The framework depends on implementing a specific smart contract for every object that
specifies the objects its actions. Although the approach offers promise, it is still in an infancy
state. The authors fail to apply any use case scenarios to their solution, which could result in
unforeseen issues. Additionally, full anonymity of the used objects is proposed. This weakens
the approach as any user can utilise the system, even harmful ones.

Agrawal et al (2018) share similar thoughts to Nabil et al (2018) as they both recognise the
absence of trust between communicating devices as well as a single point of failure as being
the major blocks for the IoT realizing its full potential. In an attempt to combat this problem,
the ensuing journal article proposes a blockchain-based IoT security model, where trust is
initiated via the immutable and decentralized nature of blockchain technologies. Each user
interaction within an IoT ecosystem is securely saved in the blockchain as a transaction, with
a sequence of these transactions representing the user’s IoT trail. Furthermore, a unique
digital crypto token is needed to prove the user interaction is genuine. This particular token is
then used as an access control tool to prohibit any unapproved access to the system. It should
be noted; such tokens are pre-created using a prediction configuration solely based on users’
IoT trail stored in the blockchain. After a careful analysis, it is clear to see the proposed
approach greatly improves the overall security of the IoT environment without any user
interference as tokens are pre-created using numerous prediction models. However, to
achieve higher levels of accuracy for prediction, bigger datasets should be used in the Long
Short-Term Memory (LTSM) model.
The proposed solution has been touted as the first of its kind, therefore the use of larger, more
varied datasets in the testing phase would lead to better, detailed results.

Page 16
Methodology and Approach
This section has been used to provide a theoretical structure for how the stated aims and
objectives will be met. Furthermore, a suitable rationale for why the specific approach was
chosen will be provided.

To achieve the research project aims of identifying IoT device trust issues as well as
evaluating the approach of using blockchain technologies to address these trust issues, a
comprehensive literature review has been conducted. The work is primarily based on
secondary resources from existing research articles, journals and publications. It should be
noted literature reviews act as a form of research that not only reviews relevant literature, but
it also critiques and synthesizes the chosen works eventually offering new frameworks,
methodologies and perspectives. When developing the literature review, both systematic and
critical analysis strategies were undertaken. (Nakano & Muniz J, 2018)

Systematic literature reviews mainly intend to uncover suitable answers to the research
questions developed. Scholarly articles/journals relevant to the topic are simply gathered and
presented in an orderly manner. (Bhattacharya, 2018) To establish what other specific trust
issues plague IoT devices when looking to build trust with one another, as well as what
attacks can occur due to a lack of trust, this type of review was used. The reason is that the
chosen research topic is very broad, therefore a lot of previous work has been undertaken by
previous researchers, ultimately generating a lot of useful data that can be used. There are
many benefits to using the systematic review technique. Firstly, you can use literature from
an extensive range of sources. An aspect of this methodology that should be praised, is that it
gives the researcher the ability to use a plethora of keywords and phrases to find relevant
literature. The keywords used for this review included: IoT, IoT trust issue, trust IoT, trust
reputation, islands of trust, smart devices, radio frequency identification, root of trust, IoT
trust domain, heterogeneity, near field communications, trust attack, IoT network,
interoperability, ballot stuffing, bad-mouthing, sybil, dynamicity. Although helpful, a
keyword search can limit creativity and intuition. Many well known digital libraries such as
IEEE Xplore, ACM Digital Library, Research Gate, Science Direct etc. were used during this
review process. Although these libraries offer an expansive range of literature, however, it
also overlooks important "grey literature" e.g. reports. Furthermore, systematic reviews aid
the process of synthesis, as well as increasing the transparency and replicability of the
review. Although several trust issues have been examined, a scenario has been given in the
introduction of this dissertation, clearly stating what explicit trust issue this project will be
focusing on.

On the other hand, the primary function of a critical review is to produce a critical assessment
of the research subject proposed by comparing numerous scholarly articles that have already
been attempted in the topic. Therefore, when looking to assess the suitability of using
blockchain and other approaches to addressing IoT device trust issues, a critical literature
review methodology was used. The main reasons why this type of literature review was

Page 17
chosen over common structures such as systematic reviews and meta-analysis reviews is
because this aim requires a deeper analysis to be understood. Furthermore, the paper selection
is intuitive. It is important that as the researcher, one possesses full control on what is
included, discussed and summarised in the review. Ultimately given this full autonomy, a
precise picture of the chosen field can be given. Furthermore, when the topic becomes much
more focused, e.g. assessing the suitability of blockchain technologies, implementing a
critical review provides the researcher with the necessary information that will help the
reader understand if blockchain can truly be established as a suitable solution to solving
stated trust issues. The blockchain architecture can be thoroughly assessed and analysed to
see how it is specifically used to solve trust issues.

Utilising a literature review has proven to be the most suitable methodology when looking to
achieve the aims and objectives of this research project. Constructing a literature review
allows for a theoretical background to be established. To completely understand how
blockchain technologies can be used to build trust in the IoT, it is important to not only
dissect the blockchain architecture but a solid understanding of what the IoT is and how it
operates should be gained. Furthermore, literature reviews provide a sense of depth that
cannot be achieved through other secondary resources. The methodology chosen allows for
this, as it provides a foundation of knowledge on the topic. Furthermore, a comprehensive
review identifies inconsistences and gaps in research. It is common knowledge that the IoT
and blockchain will continue to shape how consumers use technology in the future. However,
constructing a literature review has shown how the two emerging technologies can be used
together to complement each other. Systematic and critical reviews also allow the researcher
to identify areas of prior scholarship. This is very important as it prevents duplication and in
return gives the credit that the researchers deserve. The final reason why this methodology
was chosen is due to the fact it identifies the need for additional research, ultimately
justifying the research undertaken in this dissertation. Blockchain is viewed as an adequate
solution, but carrying out a literature review allows for a deeper understanding of why it
offers such promise.

A revision of the existing blockchain frameworks and approaches to building trust in an IoT
environment has led to the proposition of a guideline for an improved blockchain solution
that utilises technical aspects of some commonly used blockchain frameworks identified in
the literature review above. Furthermore, the blockchain solution proposed looks to solve the
most common IoT trust issues also identified above. However, more detail will be given in
the following section.

Results and Findings

After carrying out the literature view, it is clear the issue of establishing a common root of
trust is one of the main factors stopping the IoT achieving its grand vision of devices being
able to securely communicate with each other. This section will be used to compare existing
works who propose the use of specific components of the blockchain technologies for
identifying a common root of trust or as an alternative to looking to establish a common root
of trust within an IoT ecosystem, that can be accepted by all communicating parties.
Ultimately this comparison will allow for
the proposition of a guideline for a new blockchain approach that improves existing
frameworks.

Page 18
 As explained previously, one of the most difficult challenges that the IoT must face is the
task of bridging the currently fragile trust domains. The conventional way of solving this
issue, typically utilised in the PKI model, is to establish a common root of trust. Nonetheless,
this approach continually fails due to the exceptionally heterogeneous nature of the IoT,
which is made up of billions of devices, varying in available resources and belonging to
individual administrative domains. In return, we are exposed to various ‘islands of trust',
which are domains where trust is provided and regulated by individuals separate from entities
providing trust for other domains.

Comparative Analysis

Research has shown that there is not much work exploring the use of core blockchain
components in designing frameworks that promote roots of trust and improve overall security
between trusted and untrusted devices as well as using blockchain as a secure communication
channel.
The table below will be used to compare existing approaches, which aim to use specific
blockchain properties for the purposes explained above. It is important to carry out this initial
comparative analysis, as the reader can truly understand to what extent these approaches
resolve the issue of establishing a root of trust within the IoT. Furthermore, exploring the
advantages and limitations of each framework, as well as analysing their different approaches
of technically addressing the trust issue will help for the proposition of a guideline for a new
blockchain framework that builds on the limitations identified.

Proposed Blockchain Benefits Limitations

Work approach used to
address the trust
issue
Di Pietro et Chain - Obligation - Service providers and - Privacy concerns.
al (2018) Chain consumers do not have to enact
a contractual agreement.
- Use of self-enforced and
tamper-resistant reputations to
further assist obligation chain.

Jesus Smart Contracts - Removes the need for verifying - Performance whilst
(2018) both parties before exchanging implemented in the
information. blockchain.
- Provides a centralised method - Smart contracts are open
of storing unmodifiable data. and data related to SC
- Provides verifiable means to always available.
handle functions such as the
generation of keys and storage
of hashes.
Novo Use blockchain as - Mobility - Immutability of storage
(2018) immutable and - Accessibility can create serious risks.
verifiable storage. - Concurrency
- Lightweight
- Scalability
- Transparency

Page 19
Yu et al Consensus Protocols - Cooperation - Data Privacy
(2018) Smart Contracts. - Robustness - Smart contracts are open
and data related to SC
always available.
Table 1: Blockchain Approaches Compared

Instead of relying on a central root of trust, the work of Di Pietro et al (2018) proposes the
use of blockchain technologies in an attempt to bridge the trust between different secure
domains by creating a distributed trust mechanism. The researchers argue that terms and
obligations can be used by service providers and service consumers to implement trusted
communication between untrusted devices. Firstly, a new blockchain is introduced named
‘Obligation Chain', which is attached to another blockchain. This is then viewed as a new
platform for a distributed credit-like system (this differs to bitcoin, which is known as a cash-
like system). Moreover, this brand-new credit-like system consists of a built-in, tamper-proof
reputation tool that gives peers the ability to choose whether or not to accept an obligation
solely based on the credit history of a consumer. It should be noted, an entity that regularly
doesn't fulfil its obligations is very prone to have its newly minted obligations approved.
In short, the framework designed to address the stated trust issue consists of a three-way
handshaking protocol made up of a setup, a spending, and a fulfilling phase. During the
initial setup phase, service consumer back ends initiate and sign obligations solely based on
the terms presented by the service provider back ends. Moreover, such obligations are then
delivered to service consumer end devices, so that they can be used whilst communicating
with service provider end devices. The spending phase of this proposed protocol consists of
end devices moving from both service providers and consumers communicating with one
another and exchanging a personalized version of terms. The obligations generated during the
setup phase are then delivered to the service provider’s back end if a match is detected.
Additionally, if the service provider back-end agrees to welcome the obligations based on the
local service consumers' reputation score, the service consumer obtains access to the service.
As a result of this action, its obligations are broadcasted across the network to make it aware
that obligations should now be admitted into the obligation blockchain. Finally, in the
fulfilment phase service consumers create a connection with both the obligation chain and
bitcoin by paying for them. It should be noted the final step of this protocol causes the local
reputation scores stored within the service providers to be updated, which in return boosts the
probability of the service consumer's new obligations being approved in the future.

One of the major benefits of this proposed solution and the use of an obligation chain is the
fact that both service providers and consumers do not necessarily have to enact a contractual
agreement. Service providers declare the feasible business models as part of their terms,
whilst at the same time, service consumers have the ability to decide which business model
best suit their needs. Ultimately the service providers then have the choice of whether to
accept the obligations or not. Additionally, another crucial benefit of the proposed framework
is the installation of self-enforced and tamper-resistant reputations. An assumption is made
that service consumers and service providers are unaware of each other in advance, therefore
the only way for them to assess each other's reliability is through the reputations stored
within the proposed obligation chain. Although the process of having transactions openly
recorded poses many benefits, it does raise some privacy concerns.

Page 20
In comparison, Jesus (2018) proposes the use of blockchain smart contracts as a means of
enhancing or creating roots of trust, instead of completely diminishing the idea. Where this
proposed framework improves on the approach explained previously is its use of three
representative scenarios in the lifecycle of a device to clarify how smart contracts can
increase trust. The proposed strategy aims to utilise blockchains immutable and auditing
properties to achieve secure verification and general management of IoT devices. Initially,
smart contracts are used to implement a simulation of hardware cryptographic services
comparable to remote and virtual trusted platform modules. Furthermore, another aspect of
this approach is its use of a hidden blockchain that implements a protected message bus that
allows both smart devices and the cloud to exchange sensitive data such as public keys.
These ideas were further developed by Machado and Frohlich (2018), who proposes an
integrity/verification approach which not only uses blockchain technologies such as smart
contracts but introduces various consensus algorithms. The framework has been specifically
designed for constrained devices and real-time smart applications which operate in IoT
environments. Additionally, Novo (2018) propose a specific architecture that performs tasks
such as access control as well as the verification of IoT updates by implementing a
blockchain framework where nodes completely integrate instead of typically using it as a
service. Where this blockchain approach differs from those analysed above is the fact that
they have opted against the use of smart contracts. Instead, they utilise blockchain as
immutable and verifiable storage in their custom blockchain framework.

From the comparison above, it is evident the use of smart contracts has been touted as
another a possible answer to the issue of establishing roots of trust within an IoT ecosystem.
Such technology proposes a significant hope for building shared trust in all data generated
and shared by smart devices. Blockchain technologies are compiled of an immutable log of
events which present an opportunity to establish the origin of all information exchanged,
recorded and implement strict policies for accessing the information and provides a platform
for such information to be acted on individually through the use of smart contracts.
A factor that diminishes the string of trust between communicating devices, that is often
overlooked, is the treatment of the smart devices from development in factory till activation.
Commissioning a device requires a trusted process which will be primarily based on
uncertain assumptions. For instance, on arrival the device must be in a trusted state, then
provisioned, and decisively authenticated at least once before being mobilized for production.
Furthermore, and central to this research project, it is a necessity for smart devices to trust
whatever information they are accepting from their cloud counterparts. E.g. acquiring
firmware updates, or the smart device acting as an actuator, accepting commands.
Jesus (2018) recognises the use of smart contracts as a competent way of acting as a root of
trust and thus mitigating such weaknesses described above.

As mentioned, establishing a root of trust is not technically viable for a completely

decentralized and distributed IoT environment that contains a large amount of distributed
devices and consumers. Yu et al (2018) recognize the privacy weaknesses of smart contracts
and consider the introduction of consensus protocols to complement the use of smart
contracts. For example, in any collaborative system, entities such as banks and government
agencies manage the trust problem. Within a decentralized network, blockchain can be
viewed as one of these centralized institutions, distributing trust where necessary. In other
words, specific blockchain components can be utilized to establish a root of trust.

Page 21
It is no secret; smart contracts are viewed as contemporary but the intrigue remains as they
continue to create potential for the IoT to grow. It should be noted; smart contracts do possess
a few unique characteristics that in return generate countless opportunities for innovative
secure applications. These include their transparent nature, they are executed in a
deterministic manner, meaning their execution output cannot be influenced by any entity.
Moreover, they implement a means for both user/device authentication as well as token
transfer, with all communication with a smart contract recorded in the blockchain. To the
contrary, smart contracts are not entirely faultless as they come with their privacy risks and
limitations. Smart contracts are entirely immutable. Once they have been deployed, they
cannot be altered, meaning if there is a defect in the contract's code, the defect will exist
forever. What’s more is their inability to preserve user privacy, a concern that was also
identified in Di Pietro’s (2018) Obligation Chain approach. A fundamental characteristic of
blockchain-based systems is transparency. Everybody possesses the ability to gain access to
all of the records stored in the blockchain. As smart contracts form a part of the blockchain,
not only can everybody view the source code of the smart contract, they are also able to
execute it offline. Decisively smart contracts are unable to generate or store sensitive data,
example. private keys/protected records. They are also unable to execute operations that
require sensitive information, example. generation of a digital signature. Finally, smarts
cannot create secret data, create a secret key. Additionally, especially when it comes to the
IoT, further scrutiny should be taken into account. ‘Things’ may not always possess the
correct computational power to communicate with a blockchain, and in many cases they do
not even possess the computational power to perform as a ‘light node’ (a node that simply
sends and receives transactions to the blockchain, but is unable to store the blockchain as
well as implement any kind of verification). In order to combat this, a gateway-based
approach is recommended. This gateway should be responsible for not only storing the
‘Things’ sensitive data but should also act on behalf of the ‘Things’. All in all, I believe smart
contracts can be viewed as the answer when it comes to establishing a root of trust, however
many factors must be considered before this tool can be deemed a success.

Proposed Design Guideline for New Blockchain Approach

The careful comparative analysis was undertaken above,

Analysis and Conclusion

In conclusion, blockchain technologies have the potential to address many of the trust
issues associated with the Internet of Things (IoT), by providing a more secure, transparent,
and efficient way to manage data and transactions. By eliminating the need for a centralized
authority, blockchain can reduce the risk of data breaches, hacks, and other types of attacks,
which can increase trust among stakeholders in the IoT ecosystem.
However, it is important to note that blockchain is not a panacea for all the trust issues
associated with IoT. Blockchain technology is still in its early stages, and several challenges
need to be addressed, such as scalability, interoperability, and energy efficiency.
Additionally, blockchain-based solutions may not always be the most appropriate or cost-

Page 22
effective solution for every use case, and it is important to carefully evaluate the specific
application before implementing this technology.
Despite these challenges, the potential benefits of using blockchain technologies in IoT are
significant, and ongoing research and development will likely lead to more efficient,
interoperable, and energy-efficient blockchain protocols, as well as new use cases and
governance frameworks that can enable responsible use of this technology in IoT.
In summary, the use of blockchain technologies to solve trust issues for the Internet of Things
is a promising area of research, and with careful evaluation and development, it has the
potential to significantly enhance the security, transparency, and efficiency of IoT
ecosystems.
Limitations and Impact of Research

• Sample size: If your research involves collecting data from a specific group of people
or a limited number of participants, the sample size may be too small to generalize the
findings to a larger population.
• Data collection methods: The method used to collect data may have limitations, such
as biases in self-reported information or lack of accuracy in data collected from
sensors or other devices.
• Time constraints: Depending on the scope of your research, time constraints may limit
the depth of analysis or the amount of data that can be collected.
• Lack of resources: Limited access to funding, equipment, or other resources may
impact the scope or quality of your research.

Summary of Contributions

The potential contributions of using blockchain technologies to solve trust issues for the
Internet of Things (IoT) can be summarized as follows:

a. Improved Security and Privacy: By using blockchain, IoT devices can potentially
improve security and privacy by eliminating the need for a centralized authority and
reducing the risk of data breaches or hacks. This can increase trust among users and
stakeholders in the IoT ecosystem.

b. enhanced Transparency: Blockchain technologies can provide a more transparent and

tamper-proof system for tracking data and transactions within the IoT ecosystem,
which can be especially valuable for supply chain and logistics applications. This can
help to reduce fraud and increase accountability.

c. Increased Efficiency: By using blockchain, IoT devices can potentially reduce the
need for intermediaries and streamline transaction processing, which can lead to cost
savings and improved efficiency. This can be particularly valuable for applications
that require a high volume of transactions, such as micro-payments or peer-to-peer
transactions.

d. Potential to Impact Regulation and Governance: As blockchain-based solutions for

IoT become more prevalent, they may have implications for regulatory and
governance frameworks, which will need to adapt to accommodate this new

Page 23
 technology. This can potentially lead to changes in how data privacy and security are
regulated, as well as how IoT devices are governed and managed.

e. Overall, using blockchain technologies to solve trust issues for the Internet of Things
has the potential to make the IoT ecosystem more secure, transparent, and efficient. It
also has the potential to impact regulatory and governance frameworks and could lead
to changes in how IoT devices are managed and governed. However, it is important to
carefully evaluate the specific use case and determine whether blockchain is the
appropriate solution and to consider the potential costs and benefits associated with
implementing this technology.

Future Work

Scalability: One of the main challenges of using blockchain in IoT is scalability. Future
research will focus on developing more efficient blockchain protocols that can handle the
large volume of data generated by IoT devices.

Interoperability: To increase the adoption of blockchain in IoT, there needs to be

interoperability between different blockchain networks and IoT devices. Future research will
focus on developing standards and protocols that can enable cross-platform compatibility.

Energy efficiency: Blockchain can be energy-intensive, which is a challenge for IoT devices
that have limited battery life or processing power. Future research will focus on developing
more energy-efficient blockchain protocols that can be implemented on low-power devices.

New use cases: There is potential for blockchain to be used in a wide range of IoT
applications, including smart homes, smart cities, and industrial IoT. Future research will
focus on identifying new use cases and developing blockchain-based solutions that can
address the specific trust issues associated with these applications.

Governance and regulation: As blockchain-based solutions for IoT become more prevalent,
there will be a need to develop new governance and regulatory frameworks to ensure the
security and privacy of IoT data. Future research will focus on developing new governance
and regulatory frameworks that can ensure the responsible use of blockchain technology in
IoT.

Overall, there is a great deal of potential for using blockchain technologies to solve trust
issues for the Internet of Things. With ongoing research and development, we can expect to
see more efficient, interoperable, and energy-efficient blockchain protocols, as well as new
use cases and governance frameworks that can enable the responsible use of this technology
in IoT.

Page 24
References

1. www.ieeexplore.ieee.org/document/7467408/

2. www.scirp.org/(czeh2tfqw2orz553k1w0r45))/reference Papers.aspx

3. www.ncbi.nlm.nih.gov/pmc/articles/PMC7308857

4. arxiv.org/ftp/arxiv/papers/1704/1704.00073.pdf

5. www.scirp.org/(S(i43dyn45teexjx455qlt3d2q))/reference/ReferencesPapers.aspx

6. www.hindawi.com/journals/scn/2021/7142048

7. www.sciencedirect.com/science/article/pii

8. ipmc bookmannual provided.

Page 25