Web and AI Influences on Distributed

Consensus Protocols in Cloud Computing:

A Review of Challenges and Opportunities
Wahab Kh. Arabo1, Lozan M. Abdulrahman2, Nasiba Mahdi Abdulkareem3, Baraa
Wasfi salim4
1
Information Technology Department, Technical College of Informatics-Akre, Akre
University of Applied Sciences, [Link]@[Link]
2
Information Technology Department, Duhok Technical College, Duhok Polytechnic
University, Duhok, Iraq, [Link]@[Link]
3
Information Technology Department, Duhok Technical College, Duhok Polytechnic
University, Duhok, Iraq, [Link]@[Link]
4
Information Technology Department, Duhok Technical College, Duhok Polytechnic
University, Duhok, Iraq, [Link]@[Link]

Abstract
In the rapidly evolving domain of distributed systems, consensus protocols play a crucial
role in maintaining reliability and coordination among distributed components. This
paper delves into the significance of these protocols in distributed systems, highlighting
their indispensability for ensuring consistency and fault tolerance. The primary aim of
this review is to meticulously examine recent studies in this area, focusing on their
contributions, methodologies, and recommendations for future research. This
comprehensive analysis not only elucidates current trends and practices in consensus
protocols but also serves as a valuable reference for researchers and practitioners
interested in this vital aspect of distributed computing.
Keywords: Web Technology, AI, Distributed Systems, Cloud Computing, Consensus
Protocols, Challemges, Opportunities.

1. Introduction
In the ever-evolving realm of cloud computing, the orchestration of distributed systems
is not just a technical challenge but a fundamental necessity for operational success [1,
2]. The efficiency, scalability, and reliability of cloud services are deeply intertwined with
the effectiveness of this orchestration[3]. Central to this orchestration are distributed
consensus protocols, which serve as the backbone for coordinating activities and
decisions across various nodes in a cloud environment [4, 5]. These protocols are not
merely technical constructs; they represent the critical mechanisms by which distributed
systems achieve a unified state of operation. Their role extends to managing data
consistency, ensuring that every part of the distributed system reflects the same
understanding of shared data, which is paramount for maintaining the integrity and
reliability of these complex networks [6, 7].
The importance of distributed consensus protocols becomes increasingly pronounced as
cloud-based applications become more integral to business and societal functions[8].
These protocols are foundational in enabling cloud systems to operate seamlessly, even
in the face of hardware failures, network issues, or other disruptions. They provide the
structure through which distributed nodes can reach agreement, a process vital for tasks
ranging from updating databases to coordinating microservices[9]. The robustness of
these protocols directly influences the resilience and stability of cloud services, making
their design and implementation a subject of critical significance. As such, they are key
to upholding the overall integrity and performance of distributed computing networks,
which are the backbone of modern digital infrastructures[10, 11].
The evolution of cloud computing, marked by its expanding scale and complexity, has
brought about a significant transformation in the operational landscape[12]. This
transformation has profound implications for distributed consensus protocols. Traditional
protocols, which were conceived and designed in an era where computing environments
were relatively static and predictable, are increasingly found to be inadequate in
addressing the needs of contemporary, dynamic cloud environments[13].These modern
environments are characterized by their vast scale, heterogeneity, and the need for rapid
adaptability. As a result, there is a growing need to re-evaluate and adapt these traditional
protocols, or develop new ones, to meet the demands of today's cloud computing
landscapes[14, 15]. This shift in the cloud computing paradigm presents not only
challenges but also a plethora of opportunities for innovation and optimization in the field
of distributed consensus protocols. The challenges are indeed profound, encompassing
issues like scalability, fault tolerance, and the need for higher performance metrics in an
ever-growing cloud infrastructure[16]. However, these challenges also serve as catalysts
for innovation, driving research and development towards more efficient, resilient, and
adaptable consensus mechanisms[17]. The opportunity lies in harnessing the latest
technological advancements, such as artificial intelligence, machine learning, and
advanced networking, to enhance the capabilities of these protocols[18]. Thus, the current
era of cloud computing is not just testing the limits of existing consensus protocols but
also setting the stage for the next wave of breakthroughs in distributed systems
technology[19-21]. This review paper aims to delve into the intricate world of distributed
consensus protocols within the cloud computing paradigm[22]. It seeks to explore the
challenges these protocols face in contemporary cloud architectures, ranging from
scalability issues and fault tolerance to security concerns and performance
bottlenecks[23]. Concurrently, it aims to uncover the opportunities that these challenges
present for researchers and practitioners alike[24].
Through a comprehensive analysis of existing literature and a critical examination of
various protocols, this paper aspires to provide a holistic understanding of the current
state of distributed consensus protocols in cloud computing, while also projecting future
directions and potential areas for advancement[25, 26]. This paper's objective is to
thoroughly investigate the consensus protocols within the realm of distributed systems. It
focuses on the basic principles, the latest advancements in the field, and the potential
challenges and opportunities that may arise. The organization of the paper is as follows:
Section 2 presents the foundational theories. Section 3 and 4 encompasses a review and
analysis of existing literature, providing an understanding of both progress and obstacles
in the field. Sections 5 provide some extracted statistical information, while section 6 give
some common recommendations. In conclusion, Section 7 presents a summary of the
main discoveries.

2. Background Theory
Essentially, a distributed consensus protocol is a technique in computer science that
ensures multiple distributed processes or systems reach a unified agreement on a specific
data value[27]. These protocols are essential in scenarios where it's crucial to ensure that
all participants in a distributed system agree on a common state, despite the presence of
faults or varying network conditions[28]. In cloud computing, where data and processes
are spread across multiple locations, achieving consensus reliably and efficiently is a
cornerstone for maintaining the system's integrity and performance [29]. Consensus
protocols must address two primary challenges: guaranteeing that all correctly
functioning nodes concur on the same value (consensus or agreement), and ensuring that
once a value is agreed upon, it cannot be changed (finality). They must also be resilient
to various types of failures, including node failures (where a part of the system goes
down) and network failures (where communication between parts of the system is
unreliable)[30].

2.1. Proprieties of Distributed Systems

Security and vitality are critical attributes offered by every distributed system[31]. The
accuracy of a distributed system can be characterized by these attributes, implying that
for a distributed system to be deemed correct, it must adhere to the requirements of both
safety and liveness. Similarly, there are two types of failures in a distributed system, Crash
Faults and Byzantine Faults[32]. A crash fault is a common scenario where nodes
suddenly cease functioning without warning, like a node crash. Byzantine faults,
conversely, involve a situation where a node starts exhibiting abnormal behaviour, such
as malfunctioning modes, leading to the data transmitted by the node being either
contradictory or entirely nonsensical[33, 34]. Dealing with crash faults is relatively
straightforward. Byzantine faults, however, pose a greater challenge since they involve
no restrictions or assumptions about the type of behaviour a node may exhibit. In
distributed systems, the capability to persistently operate without disruption despite the
failure of one or more nodes is crucial. This characteristic is referred to as fault
tolerance[35, 36].
Furthermore, consensus is one of the main fundamental aspects of distributed systems,
ensuring reliable and coordinated decision-making across different nodes. It takes various
forms to meet diverse requirements. Leader election involves selecting a guiding
node[37], while mutual exclusion prevents simultaneous access to a critical section,
maintaining system integrity[38]. Distributed transactions depend on consensus for
committing or aborting actions, crucial for database consistency. In database ordering,
consensus determines the sequence of transactions[39], a principle also applied in block
ordering for blockchain systems[40]. Finally, reliable multicast ensures every node
receive the same updates in same sequence, essential for system consistency[41] . These
forms illustrate the versatility and necessity of consensus in distributed environments[42].

2.2. Distributed Consensus Protocols

Paxos and Raft are two of the common classical distributed consensus protocols, Paxos:
Developed by Leslie Lamport in the late 1980s, Paxos represent one of the earliest and
most influential consensus algorithms. It was designed to be a part of distributed systems
where individual components might fail or act unreliably[43]. Paxos ensures that a
consensus is reached even in such failure-prone environments. The key idea of Paxos is
that a set of proposers suggest values, acceptors agree on these values, and a set of learners
are then informed of the consensus. While Paxos is known for its correctness, it has been
critiqued for its complexity and difficulty in understanding and implementation[44, 45].
On the other hand, Raft: Raft emerged as an alternative to Paxos, aiming to provide a
more understandable approach to consensus. Developed by Diego Ongaro and John
Ousterhout in 2013, Raft breaks down the consensus process into smaller subproblems,
making it more straightforward to comprehend and implement. Raft organizes the system
into a cluster with one leader and several followers. The leader is in charge of log
replication management; the followers then replicate leader’s log. When leader fail to
operate, another one is elected. This protocol is known for its simplicity and ease of use
compared to Paxos[46].
ZAB (ZooKeeper Atomic Broadcast) protocol have an important role in the management
of distributed systems, particularly in Apache ZooKeeper. It's designed for environments
where write operations are less frequent than reads. ZAB operates on a primary-backup
model and is focused on maintaining strong consistency and ordered transactions[47].
The protocol involves a leader that handles all write requests and broadcasts these updates
to follower nodes[48]. This architecture ensures data consistency and reliability, even in
cases of leader failures, making it an effective choice for systems that require high
availability [49]. PBFT (Practical Byzantine Fault Tolerance) emerged as a
groundbreaking solution to the Byzantine Generals Problem, offering a practical
approach to achieving consensus in asynchronous systems with potentially malicious
nodes. It has the capacity to manage scenarios where up to 30% the total nodes are
malicious. PBFT is notable for its robustness in providing security and fault tolerance.
However, its requirement for a high number of message exchanges can impact scalability,
making it more suitable for environments where security takes precedence over
performance[50]. RBFT, or Redundant Byzantine Fault Tolerance, builds upon the
foundations of PBFT to enhance scalability. It operates by running multiple instances of
the consensus protocol simultaneously and monitoring their performance[51]. This design
allows for a backup instance to take over if the primary instance becomes inefficient,
maintaining the system's performance without sacrificing Byzantine fault tolerance.
RBFT represents an evolution in consensus protocols, addressing scalability while
retaining the robustness required in Byzantine environments[5, 52].
BFT-SMaRt stands as a cutting-edge library in the field of Byzantine fault-tolerant state
machine replication. It intends to offer a scalable, high-performance, and user-friendly
platform for implementing various consensus mechanisms. BFT-SMaRt's modular design
allows for ease of configuration and deployment, catering to a wide range of distributed
systems. It is especially relevant in scenarios requiring high throughput and low latency
in Byzantine environments[53]. Proof of Work (PoW), notably utilized in Bitcoin and
other cryptocurrencies, represents a fundamental shift in consensus mechanisms,
particularly in the blockchain domain. PoW requires participants to perform complex
computational tasks, thereby securing the network against fraudulent activities. While it
offers robust security, the main downside of PoW is its significant energy consumption,
posing scalability and environmental concerns [54]. Proof of Stake (PoS) emerged as a
more energy-efficient substitute for Proof of Work (PoW) in blockchain systems. In PoS,
the probability of creating a block and receiving the associated rewards is proportional to
a participant's ownership stake in the network [55]. This mechanism decreases the energy
consumption and incentivizes participants to maintain network stability[56]. PoS is
gaining traction in newer blockchain systems that prioritize sustainability and
efficiency[57].
2.3. Web and AI Influences on Cloud Computing
The foundation of modern cloud computing rests upon the ability to maintain data
integrity and system functionality across a vast network of geographically dispersed
servers. This critical task falls to distributed consensus protocols, ensuring all nodes
within the system agree on the current state of data, even in the face of potential hurdles
like failures, network delays, or malicious actors. With the explosive growth of web-scale
applications and the integration of Artificial Intelligence (AI) into cloud services, both
the demands and opportunities for distributed consensus protocols are undergoing a
significant transformation[58]. This paper delves into the web's influence on these
protocols, explores the impact of AI integration, and delves into the future trajectory of
this critical technology in the ever-evolving cloud computing landscape. The phenomenal
rise of web-based services has dramatically increased data volume, user traffic, and the
complexity of cloud deployments[59]. These trends have several key implications for
distributed consensus protocols:
• Scalability Imperative: Traditional consensus protocols often falter when faced
with a growing number of participating nodes. Web-scale applications necessitate
protocols that can "scale horizontally," meaning they can efficiently handle the
addition of more nodes to the system without compromising performance.
• Minimizing Latency: Real-time web experiences demand low latency
communication between distributed servers. Consensus protocols need to be
optimized for speed while still ensuring data consistency. A slow consensus process
can significantly impact user experience and application responsiveness.
• Always-On Availability: The modern web thrives on near-constant uptime.
Consensus protocols should prioritize fault tolerance, allowing the system to
continue functioning even if some nodes fail. This ensures uninterrupted service
delivery and minimizes potential revenue losses for web-based businesses.
These web-scale considerations have spurred the development of novel distributed
consensus protocols specifically tailored for cloud environments. Some prominent
examples include:
• Paxos Family: This family of protocols is renowned for its robust theoretical
foundation and exceptional fault tolerance capabilities. Variants like Raft offer
practical implementations with good scalability and performance characteristics,
making them well-suited for web-scale deployments.
• Zab Protocol: Designed for high availability, Zab prioritizes replicating data across
geographically dispersed data centers. While prioritizing consistency over raw speed,
it makes Zab ideal for mission-critical applications where data integrity is paramount.
• Flake Protocol: This lightweight protocol prioritizes low latency and high
throughput, making it ideal for real-time web applications where speed is crucial. It
achieves faster consensus decisions at the expense of some consistency guarantees,
which might not be a significant trade-off for applications that prioritize
responsiveness over absolute data consistency[60].
The selection of the most suitable protocol hinges on the specific needs of the web
application. Factors like desired consistency level, latency requirements, and fault
tolerance needs all play a crucial role in the decision-making process. The integration of
AI into cloud services has further diversified the landscape of distributed consensus
protocols. Here's a closer look at how AI is influencing this critical domain:
• Machine Learning for Failure Prediction: AI algorithms can analyze system logs
and historical data to predict potential node failures proactively. This allows
 protocols to take preventative measures to maintain consistency and minimize
downtime by gracefully replacing failing nodes before complete failure occurs.
• AI-powered Byzantine Fault Tolerance (BFT): Traditional BFT protocols can be
computationally expensive, hindering their scalability. AI can be used to develop
more efficient BFT algorithms that can effectively handle malicious actors
attempting to disrupt the system. These AI-powered BFT protocols can provide
robust security without sacrificing performance.
• Self-healing Distributed Systems: AI can be used to automate tasks like node
replacement and resource allocation within a distributed system. This can lead to
more dynamic and adaptable consensus protocols that can self-heal in response to
changing conditions, such as unexpected surges in traffic or node failures. AI can
continuously monitor the system and take corrective actions to ensure optimal
performance and resource utilization[61].
However, AI integration also introduces new challenges that need to be addressed:
• Explainability and Transparency: Decisions made by AI algorithms within
consensus protocols need to be transparent and explainable. This is crucial for
ensuring trust and security in the system. Without clear explanations for AI-driven
decisions, it becomes difficult to identify potential biases or vulnerabilities within the
algorithms themselves.
• Security Vulnerabilities: AI models themselves can be vulnerable to attacks.
Malicious actors could potentially exploit these vulnerabilities to compromise the
consensus process. Robust security measures need to be implemented to safeguard
AI models and prevent them from being manipulated to disrupt the system or
introduce inconsistencies[62].
The interplay between web-scale demands and AI integration will continue to shape the
future of distributed consensus protocols in cloud computing. Here are some key trends
to watch for:
• Heterogeneous Consensus Protocols: Different parts of a cloud system might have
varying needs regarding consistency, latency, and fault tolerance.

2.4. Web Attacks Influences on Distributed Systems

The ever-expanding realm of cloud computing relies heavily on distributed consensus
protocols to maintain data integrity and system functionality across geographically
dispersed servers. These protocols ensure all nodes (individual computing units) within
the system agree on the current state of the data, even when faced with challenges like
node failures, network delays, or malicious actors[63]. With the relentless growth of web-
based applications and the integration of Artificial Intelligence (AI) into cloud services,
the landscape of distributed consensus protocols is constantly evolving, not only in terms
of functionality but also in terms of vulnerabilities. This paper explores how web attacks
influence these protocols, the potential consequences, and strategies for mitigating these
threats[64].

The explosive growth of web-based services has unfortunately coincided with a rise in
sophisticated web attacks. These attacks target various aspects of cloud computing
infrastructure, with distributed consensus protocols emerging as a prime target due to
their critical role in maintaining data consistency and system availability[65]. Here's a
breakdown of some common web attacks that pose a significant threat:
• Sybil Attacks: These attacks attempt to disrupt consensus by creating a large number
of fake or compromised nodes within the system. This can overwhelm the protocol and
potentially lead to inconsistencies in the data or even complete system failure. Sybil
attacks are particularly concerning for protocols that rely on voting mechanisms, as the
attacker can manipulate the voting process by controlling a significant portion of the
seemingly legitimate nodes[66].
• Man-in-the-Middle (MitM) Attacks: In a MitM attack, the attacker intercepts
communication between nodes within the consensus protocol. This allows them to
eavesdrop on sensitive data, modify messages in transit, or even inject malicious
messages into the communication flow. MitM attacks can compromise the integrity of
the consensus process and lead to inconsistencies in the data.
• Denial-of-Service (DoS) Attacks: These attacks aim to overwhelm the system with a
flood of traffic, rendering it unavailable to legitimate users. DoS attacks can target
individual nodes or the entire protocol itself. In the context of distributed consensus
protocols, DoS attacks can disrupt communication and prevent nodes from reaching
consensus, hindering system functionality[67].
• Zero-Day Attacks: These are previously unknown vulnerabilities that exploit
weaknesses in the software or hardware components of the consensus protocol. Zero-
day attacks are particularly dangerous because there are no existing patches or security
measures in place to mitigate them. They can have a devastating impact on system
security and data integrity[68].

These are just a few examples of the various web attacks that can significantly influence
the security and functionality of distributed consensus protocols. As attackers develop
new and more sophisticated techniques, it's crucial for cloud service providers and
protocol designers to stay ahead of the curve and implement robust security measures[69].

A successful web attack on a distributed consensus protocol can have severe

consequences for cloud-based services. Here are some potential ramifications:

• Data Inconsistency: Attacks can compromise data integrity, leading to inconsistencies

between different nodes in the system. This can have a cascading effect, causing errors
and disruptions in various cloud services that rely on accurate data.
• System Downtime: DoS attacks and other disruptions can lead to system outages,
preventing users from accessing their data or applications hosted on the cloud platform.
This can result in significant financial losses for businesses and inconvenience for users.
• Loss of Trust: Security breaches can damage the reputation of cloud service providers
and erode user trust. Businesses that rely heavily on cloud services may become hesitant
to adopt them due to security concerns.
• Reputational Damage: A successful web attack on a high-profile cloud service can
have a significant reputational impact on the provider. This can lead to negative
publicity and a loss of market share[70].

These consequences highlight the critical importance of robust security measures for
distributed consensus protocols. Cloud service providers and protocol developers must
prioritize security throughout the design, development, and deployment processes[71].

Combating web attacks and ensuring the security of distributed consensus protocols
requires a multi-pronged approach. Here are some key strategies:
• Secure Communication Channels: Implementing secure communication channels
using encryption protocols like Transport Layer Security (TLS) is essential. This
helps to protect data in transit from eavesdropping and modification by attackers.
• Authentication and Authorization: Strong authentication and authorization
mechanisms are crucial for preventing unauthorized access to the consensus
protocol[71]. This includes techniques like digital certificates and role-based access
control (RBAC).
• Regular Security Audits: Conducting regular security audits of the consensus
protocol and the surrounding infrastructure helps to identify and.

3. Literature Review
In [72], a new consensus protocol for linear agents within a dynamic
communication network is introduced by Rezaei and Khanmirza, challenging traditional
paradigms in the field. Distinctively, this protocol allowed agents to update their states to
either the maximum or minimum of their neighbours, or maintain their current state,
without the constraint of positive lower bounds on communication link weights typically
required in conventional consensus protocols. This approach demonstrates a significant
advancement in facilitating consensus under more generalized and varying conditions.
The paper's innovative contribution lies in its introduction of varied operating modes
(Tracker, Following, Cross) for each agent, enhancing the convergence rate and reducing
computational demands. These improvements are validated through numerical examples,
including a consensus problem and a formation problem, showcasing the protocol's
practical applicability and robustness, especially in scenarios with time-varying and
negative communication link weights. The paper concludes by suggesting future
extensions of this work to systems with communication delays, indicating a potential area
for further exploration in the realm of distributed control systems and network
communication.
Li et al. [73] addressed a critical aspect of multiagent systems (MASs), particularly
focusing on leaderless configurations issue as a consensus problem under Denial of
Service (DoS) attacks. Central to their methodology is the construction of a state
predictor, which was ingeniously utilized to recreate state information that is lost during
DoS attacks activation. This approach led to the design of a strong, irregular consensus
protocol, capitalizing on predictor states to counter the open-loop issues produced by
these attacks. A significant contribution of this study is the derivation of sufficient
conditions under which agents can obtain consensus inside a specific region, highlighting
the improved anti-disturbance and anti-attack system properties. This study stands out for
its practical relevance, as it moves beyond theoretical formulation to validate the proposed
protocol through numerical simulations. Moreover, the integration of fractional power
into the protocol not only accelerates convergence but also amplifies the system's
resistance to disturbances. The paper concludes with an emphasis on the need for further
exploration into the consensus challenges faced by complex MASs in scenarios laden
with DoS attacks, pointing towards a significant area for future research in the field of
network security and MASs.
Afzaal et al. [74] delve into enhancing the security and trustworthiness of
crowdsourcing systems, particularly addressing the susceptibility to various attacks such
as sybil, double-spending, and eclipse attacks. This research is pivotal as it integrates
blockchain technology into crowdsourcing, unveiling the Trust and Transactions Chain
(TTC) consensus protocol. The protocol is distinct in its approach to selecting a trusted
leader and validators, based on a comprehensive trust model that considers factors like
deposit ratio, block generation and validation rate, and waiting rate. The TTC protocol
aims to ensure the correctness of critical crowdsourcing systems, which is vital given the
severe consequences of their failure. A standout feature of this work is the application of
formal methods, specifically model checking, for the verification of the TTC protocol.
This method, utilizing Communicating Sequential Programs and Linear Temporal Logic,
presents a more rigorous and mathematical approach compared to traditional simulations.
The Process Analysis Toolkit is employed to verify the model, focusing on fraud-
resilience and the system's ability to withstand DoS attacks. The study concludes with
insights into future enhancements, including improving the efficiency of the model
checking technique, addressing scalability issues, and incorporating probabilistic aspects
for ensuring the correctness of critical applications. This research represents a significant
step in ensuring the security and integrity of crowdsourcing systems, leveraging
blockchain technology's decentralized and tamperproof nature to create a robust, fraud-
resilient consensus protocol.
Zhang et al. [75] addressed the complex challenges in unmanned aerial vehicle
(UAV) networks, such as dynamic changes, high latency, and scalability issues, through
the innovative integration of blockchain technology. They proposed a unique framework
combining network tomography and an adjustable reputation-based Raft-Poa layered
consensus protocol, aimed at enhancing network performance and security. This approach
includes a novel topology prediction method and a trust management mechanism using
group signatures, significantly improving network communication, reducing latency, and
ensuring node security. The effectiveness of this protocol is demonstrated through
simulations, showing substantial improvements over traditional methods. This research
marks a significant step forward in UAV network optimization, pointing towards future
enhancements in reputation management and system refinement for practical
applications.
Wen et al. [76] proposed a unique consensus protocol designed for complex MASs
characterized by high-order nonlinearity. This protocol stands out by incorporating the
Lipschitz constant, a mathematical condition that restricts nonlinearity, and leveraging
distinct communication topology structures. Its innovative approach involves the
integration of the Laplace matrix eigenvalues with the system's Lipschitz constant,
effectively creating a flexible and adaptable framework. The protocol's design is further
enhanced by a cleverly crafted Lyapunov function, which splits the system's convergence
conditions into two segments linked by an unknown quantity 'k', introducing a degree of
looseness and independence in the system's components. This nuanced approach not only
provides robustness against non-ideal connection weights but also paves the way for
future enhancements in addressing practical challenges like communication delay and
topology switching, making it a significant contribution to the field of nonlinear, high-
order MASs.
Luo et al. [77]innovated in the realm of MASs by introducing a novel consensus
protocol using enhanced Zeroing Neural Network (ZNN) models. This study is a
significant advancement in the field, particularly addressing the challenges of noise
tolerance and predefined-time stability in cooperative control. Two new noise-tolerant
ZNN models are proposed, equipped with advanced activation functions - the noise-
tolerant sign-bi-power (NTSBP) and noise-tolerant sign-exp-power (NTSEP). These
models exhibit greater robustness compared to traditional ZNN models, a claim
substantiated by detailed mathematical analysis and numerical experiments. A key aspect
of this research is the estimation of the upper bounds of the settling-time function using
an innovative method based on improper integral, enhancing the precision in measuring
system stability. The study demonstrates that these new models maintain robustness and
stability even under dynamic bounded vanishing and non-vanishing noise conditions,
marking a notable improvement over existing models. Practical applications, such as
formation control, are also explored, further validating the effectiveness of the proposed
models. This research not only provides a more robust and stable framework for
consensus in MASs but also lays the groundwork for future advancements in networked
control systems.
Cacace et al. [78] presented an innovative distributed protocol designed to achieve
consensus in discrete-time networks of scalar agents. This protocol is particularly notable
for its ability to assign an arbitrarily fast convergence rate to consensus, regardless of the
network's size. This aspect marks a significant advancement over more standard
consensus protocols. The protocol, evolving from a centralized consensus model
previously proposed in the literature, leverages a multi-rate forward computation across
various consensus steps, allowing for a distributed implementation. The effectiveness and
improvements of this new protocol are validated through several simulations,
demonstrating its robust performance. Looking forward, the authors plan to extend this
protocol to manage multi-consensus challenges in heterogeneous networks, including
addressing the presence of delays. Additionally, future work includes exploring adaptive
control strategies and enabling each agent in the network to set different values for the
weighting parameter. This research opens new possibilities for more efficient and flexible
consensus mechanisms in discrete-time systems, potentially impacting various
applications in networked systems and control theory.
Ma et al. [79] introduced two novel reduced-order protocols for linear MASs,
focusing on adaptive control using only relative output information. The edge-based
protocol adjusts network edges using information from two connected agents, while the
node-based protocol utilizes data from all neighbouring agents. These fully distributed
protocols reduce communication burdens and enhance network security, addressing key
challenges in MASs. The paper confirms the protocols' effectiveness and their
applicability in scenarios where existing methods fall short, marking a significant
advancement in the field of MASs consensus and network security.
Zhou et al. [80] addressed the finite-time consensus problem in high-order linear
multi-agent networks using a smooth state-feedback method. They introduced a new gain-
scheduled approach with two types of parameters: a state-independent parameter based
on network connectivity and a node-based adaptive parameter for fully distributed
consensus. These parameters, rooted in finite-time stability theory, ensure the agents
reach consensus within a specific time frame, as illustrated through a practical satellite
formation simulation. Despite its efficacy, the protocol poses challenges related to infinite
gain and potential unbounded input as convergence time approaches, a concern that the
authors aimed to address in future work, particularly under input saturation constraints.
This research marks a significant advancement in the field of MASs, offering a more
precise and adaptable consensus protocol for complex networked systems.
Zhang and Tang [81]addressed the consensus problem in discrete-time MASs
(MASs) under stochastic communication noises. The paper introduces a novel Kalman-
filter-based two-time-scale protocol, alternating between estimation and control phases,
which is a significant deviation from classical Kalman-filter-based consensus algorithms.
This protocol is innovative in that it does not require agents to receive their neighbours’
control input information, thus relaxing the conditions on the communication
environment. The Kalman-filter-based estimator used in this protocol is shown to be
asymptotically unbiased even without information about the noises and initial states. The
study delves into the stability and convergence properties of the proposed protocol,
analysing the closed-loop solution's mean square convergence. The effectiveness of this
approach is demonstrated through numerical examples. The research also highlighted a
trade-off between the accuracy of the consensus value and the convergence rate,
influenced by the holding time in the protocol. While the paper focuses primarily on
MASs with scalar state values, it acknowledges limitations and proposes future
directions, including minimizing the variance of the final consensus value and extending
the protocol to vector state cases and higher-dimensional MAS models. This study
provides a new perspective and methodology for addressing stochastic consensus
problems in MASs, paving the way for broader applications in complex networked
environments.
In [82] Zhang et al. introduced a new consensus protocol designed (named EB-
BFT) to enhance blockchain throughput and reduce transaction latency. EB-BFT
dynamically adjusts the number of consensus nodes and transaction batch sizes based on
transaction volume and physical resources. It improves efficiency by identifying and
suspending low-performance nodes, focusing on high-performing nodes for consensus.
The protocol also scales transaction batches post-validation for optimal throughput.
Experiments demonstrate its ability to handle up to 20000 transactions per second with a
maximum 15-second latency, rivalling traditional centralized systems. However, the
protocol's reliance on a master node ordering method, poses challenges if a single
unverified transaction affects the entire batch, highlighting the need for further refinement
in future research. This innovation in blockchain consensus mechanisms marks a
significant step toward more adaptable and efficient blockchain networks.
In the study [83], Xu et al. introduced innovative adaptive dynamic event-triggered
(ADET) strategies for consensus control in MASs with linear dynamics. These strategies,
unique for their online triggering parameters and dynamic thresholds, facilitate fully
distributed, observer-based consensus without requiring global information. The paper
showcased two ADET protocols for both synchronous and asynchronous event-triggered
communication, effectively addressing Zeno behaviour and ensuring system
performance. Validated through a numerical example, this approach marks a significant
advancement in distributed control, with future extensions proposed for more complex
systems and network structures.
Wu et al. [84] explored the enhancement of cybersecurity in Cyber-Physical
Systems (CPS) using an improved blockchain technology approach. Addressing the
increasing risks of data leakage and system performance issues due to malicious activities,
the study presented advancements in the Practical Byzantine Fault Tolerance (PBFT)
protocol, developing a new variant named Reinforced Practical Byzantine Fault
Tolerance (RPBFT). RPBFT modifies PBFT by broadcasting transaction data network-
wide, selecting master nodes based on the longest chain principle, and implementing a
blacklist mechanism for malicious nodes. It also introduced a two-stage process by
eliminating the commit phase, improving system throughput and reducing consensus
communication time. The study's simulations showed that RPBFT significantly enhances
system efficiency while maintaining security. Future research directions include adapting
RPBFT to handle massive data in fintech and AI, optimizing it for different blockchain
types, and integrating technologies for enhanced privacy protection. This research
contributes to fortifying the security and efficiency of blockchain applications in CPS,
addressing critical challenges in the field.
Ru et al. [85] presented a new approach to enhance the security of (MASs) against
Denial-of-Service (DoS) attacks. The study developed a sliding mode observer to
accurately estimate the state, actuator, and sensor faults of each agent in the system.
Building on this, a coding-decoding communication protocol is proposed, significantly
strengthening the security of communications between agents. This protocol is a key
innovation, providing robust protection against cyber threats. Furthermore, the paper
designed a leader-following consensus controller based on the sliding mode observer and
the secure communication protocol. This controller is engineered to ensure not only
consensus among the agents but also input-to-state stability of the MASs under DoS
attacks. The efficacy of this comprehensive security and control approach was validated
through a simulation example, demonstrating its practical application. Future research
directions suggested include developing more comprehensive communication protocols
for diverse practical needs and investigating secure communication strategies from the
leader to followers in MASs. This research contributes significantly to the field of cyber-
physical system security, offering a sophisticated solution to maintaining system integrity
and functionality under cyber attacks.
Huang et al. [86] introduced WRBFT, a significant improvement to the Practical
Byzantine Fault Tolerance (PBFT) protocol, addressing its limitations in primary node
selection, communication overhead, and dynamic node participation. WRBFT innovates
by combining node consensus workload with a verifiable random function (VRF) for
selecting a reliable primary node leading the consensus process. This method optimizes
the selection of consensus participants based on their workload, enhancing the efficiency
and reliability of the consensus mechanism. Compared to PBFT, WRBFT demonstrates
higher throughput, lower consensus latency, and improved algorithmic efficiency in
simulations. The algorithm's focus on reliable node selection not only resists malicious
attacks but also reduces consensus failure risks. Future work aims to refine the calculation
of node workload value for increased algorithm dynamics and adaptability, with a
potential integration of group signatures to accelerate consensus achievement. This
research represents a notable advancement in Byzantine fault tolerance protocols,
offering a more dynamic and efficient approach to achieving consensus in distributed
networks.
Menegatti et al. [87] proposed a groundbreaking Federated Learning (FL) algorithm
that decentralizes the model-averaging process. This algorithm, leveraging a Multi-Hop
consensus protocol, allows federation clients to share models efficiently using
information from only their m-step neighbours, significantly reducing communication
overhead. It achieves nearly the same performance as centralized FL with a slight
performance drop of about 1%, ensuring faster convergence rates. This approach,
validated through experiments with various communication topologies and datasets like
MNIST and MedMNIST v2, is particularly beneficial for large federations and time-
critical applications. The study paves the way for future enhancements in FL, particularly
in delay-critical and large-scale scenarios such as IoT networks, signifying a major shift
towards more scalable and efficient decentralized learning frameworks.
Zhao et al. [88] introduced a new consensus protocol, Proof of Evolutionary Model
(PoEM), it was to cater to the specific needs of Blockchain as a Service (BaaS) in IoT
environments. Addressing the limitations of conventional power-intensive consensus
protocols in power-constrained BaaS-enabled IoT systems, PoEM stands out as a
lightweight, model-based evolutionary approach. It utilized machine learning to
iteratively train models for achieving consensus, thereby enhancing efficiency and
enabling participation of low-performance IoT devices. PoEM also incorporated a
mechanism to manage the dynamic joining and exiting of nodes, considering the ever-
changing nature of IoT environments. Through extensive analytical and experimental
evaluations, the protocol demonstrates improved consensus efficiency and high-level
security, making it highly suitable for dynamic BaaS-based IoT systems. The research
indicates future directions towards optimizing the cold start process of PoEM and
integrating technologies like edge computing and federated learning for decentralized
model generation, aiming to further enhance efficiency and reduce complexity. This study
represents a significant advancement in developing efficient and secure consensus
mechanisms tailored for BaaS in the IoT domain.
Lippi et al.[89] addressed a new problem in MASs: distributive tracking the
minimum infimum or maximum supremum of time-varying signals in finite-time. This
study introduced an adaptive protocol specifically designed for undirected connected
network topologies. Uniquely, the protocol operates without any assumptions regarding
network size or bounds on the derivatives of exogenous signals. Each agent in the network
independently accesses a local time-varying exogenous signal, and through the protocol,
they collectively track the dynamic extremum of these signals. The paper's strength lied
in its adaptive approach, which is validated through numerical simulations in a precision
farming scenario. This methodology demonstrates the protocol’s ability to efficiently
achieve consensus on extremum values among agents. Future research directions include
extending this protocol to directed communication graphs and enhancing its robustness
to perturbations or attacks. This work contributes significantly to the field of MASs,
offering a scalable and adaptable solution for dynamic consensus problems in various
network configurations.
Sun et al. [90] presented the consensus problem in (MASs) impacted by input and
communication delays. The study introduced a predictor-based state feedback protocol
designed to compensate for these delays and achieve consensus in linear MASs. The new
approach involves equating the overall MASs to a feedback interconnection system,
comprising a linear time-invariant system and a time-delay operator, leveraging the
characteristics of the Laplacian matrix. The maximum delay permissible under this
protocol is then assessed using the small gain theorem (SGT). The research is validated
through two numerical examples, demonstrating the effectiveness of the consensus
condition derived. Looking ahead, the paper suggests future research directions, including
the development of observer-based consensus protocols for MASs with heterogeneous
input and communication delays. This work makes a significant contribution to the field
of MASs, offering an innovative approach to overcoming the challenges posed by delays
in achieving system-wide consensus.
Zhang et al. [91] addressed the consensus challenges in discrete-time linear MASs
within wireless sensor networks, focusing on mitigating issues caused by channel fadings.
They introduce a novel control protocol, DDTETS, tailored for the stochastic Rice fading
model. DDTETS utilizes local state information and a dynamic auxiliary system,
significantly improving system performance. The protocol prolongs interevent times
between communications, reducing data packages, mitigating channel fading effects, and
decreasing information transmissions among agents. Numerical simulations validate its
effectiveness in maintaining asymptotic consensus under challenging network conditions.
This research marks a significant advancement in networked control systems, offering a
resilient and efficient framework for consensus in wireless communication-constrained
environments
Zhang et al.[92] delved into the realm of multi-agent differential graphical games
for high-order systems. they introduced a modified cost function for each agent and crafts
a fully distributed control protocol. This protocol is notable for its ability to ensure
leaderless consensus and optimization in terms of Nash equilibrium, even with
communication topologies that only contain a directed spanning tree. The protocol's
design is particularly significant in that it is independent of global information about
network topologies, enabling fully distributed management. The efficacy of this protocol
is demonstrated through a numerical example, which confirms its capability to achieve
both consensus and Nash equilibrium in the system. This study makes a substantial
contribution to the field of MASs, offering a new perspective on leveraging graphical
games for distributed consensus and optimization in complex networked systems.
Abdella et al. [93] introduced a new hierarchical consensus protocol to enhance
blockchain-based peer-to-peer energy trading systems (called HiCoOB). It was designed
to manage both local and global transactions concurrently, significantly improving
system throughput and reducing latency. It employed separate leaders for local and global
consensus, thereby reducing the load on individual nodes and incorporating a
synchronization method to prevent double-spending issues. The protocol's effectiveness
was demonstrated by its substantial increase in throughput and reductions in both local
transaction waiting times and leader node overhead. This approach marks a significant
advancement in blockchain applications for energy trading, proposing a more efficient
and scalable solution. Future work aims to extend support to mobile energy participants,
broadening HiCoOB's applicability.
Wang et al. [94] presented a new approach to the consensus tracking problem in
MASs with Lipchitz nonlinear dynamics. They introduced an integrated adaptive control
approach, distinctively using a single adaptive coupling gain matrix to simultaneously
adjust the feedback control gain and coupling weights. This integration offered a more
flexible and convenient solution compared to traditional adaptive consensus protocols,
which typically require common predetermined feedback control gains. The effectiveness
of this approach is theoretically validated using Lyapunov stability theory, with a
Lyapunov function partially constructed from the trace-norm of the adaptive coupling
gain matrix. The protocol is further extended to handle robust consensus tracking with a
leader having nonzero inputs and external disturbances. It employs the σ modification
technique for bounded disturbances to ensure ultimate boundedness of the consensus
error, and an integrated H∞ consensus protocol for L2-type disturbances to guarantee
desirable disturbance rejection performance. The proposed approach's efficacy is
demonstrated through a numerical example involving forced pendulums. This study
represents a significant advancement in adaptive control for MASs, providing a more
versatile and analytically tractable solution for consensus tracking under nonlinear
dynamics and external perturbations
Voleti et al. [95] proposed a new set of consensus protocols designed to synchronize
the arrival of multiple agents or vehicles at a destination simultaneously. Their study
introduced three distinct time-to-go protocols: max-time, min-time, and achievable-time,
each utilizing a backstepping-inspired speed modulation approach suitable for systems
modelled as double integrators. These protocols were developed to ensure same-time
position consensus, with a particular focus on systems with arbitrary connected graph
communication topologies. To facilitate consensus in diverse network structures, a
tokenized algorithm is presented, particularly leveraging the achievable-time protocol.
This research stands out for its comprehensive approach to consensus in dynamic
systems, addressing situations where speed constraints might impede consensus under the
min-time protocol and offering a robust alternative with the achievable-time protocol.
The effectiveness and versatility of these protocols are demonstrated through numerical
simulations, showcasing their potential in scenarios such as secure rendezvous or
coordinated movement of multiple pursuers. Looking ahead, the authors suggest
extensions of this framework to include event-triggered consensus evaluations and
applications in planar multiple pursuer-target engagement problems, emphasizing its
broad applicability and relevance to complex coordination challenges in MASs.
In [96], Zou et al. focused on developing finite-time consensus (FTC) protocols for
second-order MASs characterized by non-identical nonlinear dynamics and operating
under directed networks. They introduced two innovative protocols tailored for leader-
follower and leaderless scenarios, respectively, utilizing a reference trajectory-based
method. A key feature of these protocols is the coupling between each agent's state
dynamics and the reference trajectory, with a crucial constraint that reference trajectory
information is not transmitted through communication channels. This approach allowed
relative state errors among agents to converge to the origin within a finite time, despite
the presence of unknown parameters in the agents' dynamics. The protocols did not
require nonlinear terms in agents' dynamics to be bounded by known constants or
functions, a requirement that limits many existing consensus schemes based on sliding
mode control. Sufficient conditions for achieving finite-time consensus are established
through graph theory, the Lyapunov functional method, and finite-time stability theory.
The effectiveness of the proposed protocols is validated through simulation examples,
demonstrating their capability in handling unknown parameters and achieving complete
FTC in MASs with complex dynamics. This study presents a significant advancement in
consensus protocols for MASs, offering more flexible and robust solutions for systems
with diverse and unpredictable dynamics
Li et al. [97] explored a new approach to achieving consensus in (MASs) subjected
to denial-of-service (DoS) attacks. They introduced an innovative G-noise
communication protocol, diverging from traditional noise-based protocols driven by
Brownian motion. This protocol is characterized within the sublinear expectation
framework, thereby addressing the challenges posed by the unique properties of G-noise,
which traditional theories find infeasible. The research establishes a new theoretical
framework, using Lyapunov methods, stochastic analysis, and graph theory, to develop a
sufficient criterion for ensuring consensus. This criterion accounts for various factors,
including DoS attack duration, noise intensity, and G-Brownian motion parameters. The
protocol's effectiveness is underscored by its ability to handle probability ambiguity and
nonlinear dynamics, which are common in practical scenarios like islanded microgrid
systems. The study not only highlighted the positive impact of G-noise on consensus
achievement but also opened avenues for future research, particularly in adapting agent
dynamics and exploring G-noise-based event-triggered and sampled-data communication
in both continuous and discrete-time MASs. This paper presents a significant
advancement in consensus protocols for MASs, offering new perspectives and solutions
in environments with cybersecurity threats.
Qyinloye et al. [98] proposed a new tool, SIM-P (Simplified Consensus Protocol
Simulator), to address the challenge of experimental verification in newly proposed
blockchain consensus protocols. SIM-P is designed as an agent-based stochastic
simulator that employs the sequential Monte Carlo method to model block publisher
selection, representing nodes as agents involved in independent trials of a binomial
experiment. This tool provides a base model simulating Proof of Work (PoW) for
benchmarking, which can be adapted to other protocols. The flexibility of SIM-P is
demonstrated through the simulation of two additional models: Proof of Reputation-X
(PoRX) and Proof of Contribution (PoC), both lacking experimental verification in their
original designs. SIM-P efficiently generates key performance metrics such as
throughput, resistance against the 51% attack, and energy consumption. The accuracy of
SIM-P is validated by comparing the simulated results of PoW with theoretical estimates
and historical Bitcoin data. This research showcases how slight modifications in the base
model enable the simulation of diverse protocols, illustrating the advantages of PoRX and
PoC over traditional PoW in terms of difficulty scaling mechanisms. SIM-P emerges as
a valuable tool for analyzing various consensus scenarios and optimizing performance
trade-offs in blockchain network design.

4. Literature Analysis
Through analysing the studies in previous section, the collection of papers on
consensus protocols covers a wide range of topics and applications, highlighting the
significance of achieving agreement in distributed systems. These papers contribute to
advancing the understanding and development of consensus algorithms in various
domains, including communication networks, cyber-physical systems, blockchain
technology... etc. Table 1 provides a consolidated overview of recent advancements in
the fields of consensus protocols, blockchain technology, and multi-agent systems. Each
entry in the table summarizes a key research paper or article, highlighting its main
objectives, significant contributions, methodologies employed, and potential avenues for
future research. This compilation serves as a valuable resource for scholars and
practitioners alike, offering a quick reference to the current state of research in these
rapidly evolving domains. It not only encapsulates the essence of each study but also
suggests directions for upcoming investigations, thereby fostering further innovation and
exploration in these fields.

Table 1: Summary of the recent studies.

Ref Main Objective/Problem Key Contributions/ Methodology Future Work Suggestions

Addressed Findings /Approach
[72] Consensus in time-variant New consensus Development of a Extend findings to systems
2023 communication topology protocol allowing unified consensus with communication delays
convergence protocol with
without lower various operating
bound weights modes (Tracker,
Following, Cross)
for agents.
[73] Consensus under DoS Robust consensus Designing a non- Further research on
2023 attacks protocol mitigating smooth consensus consensus challenges in
effects of DoS protocol DoS attack scenarios
attacks incorporating state
predictors.
[74] Fraud-resilience in Trust and Using CSP and LTL Addressing scalability
2023 crowdsourcing consensus Transactions Chain for TTC Protocol issues and incorporating
protocol (TTC) protocol with Verification in advanced model checking
formal verification Crowdsourcing techniques.
 Security via Process Considering probabilistic
Analysis Toolkit aspects in critical
applications of the system
[75] Data transmission Integration of Network Explore additional
2024 reliability in UAV blockchain for tomography-based challenges in UAV
networks improved reliability framework network optimization
Proposal of a
reputation-based
Raft-Poa layered
consensus protocol
to improve
scalability and
efficiency
[76] Developing a consensus Utilization of the Employing Expanding the protocol to
2022 protocol for high-order Laplace matrix mathematical and address communication
nonlinear multi-agent eigenvalues and the theoretical delays and switching
systems Lipschitz constant approaches to create topology challenges
in the protocol's a flexible and
design. adaptable consensus
Development of a protocol, with a
unique Lyapunov focus on the system's
function that divides stability and
the system's convergence
convergence
conditions into two
interconnected parts
[77] Developing a robust and Introduction of two Utilization of Exploration of further
2023 predefined-time novel NTZNN zeroing neural applications and
consensus protocol for models with networks with new enhancements of NTZNN
MASs using (NTZNN) enhanced (NTSBP) activation functions models in networked
models and (NTSEP) and mathematical control systems
activation functions techniques to
enhance the
robustness and
stability of
consensus protocols
in MASs
[78] Achieving consensus in Generalization of a based on a Extension of the protocol to
2023 discrete-time networks of centralized distributed handle multi-consensus in
scalar agents with an consensus protocol implementation with heterogeneous networks,
arbitrarily assignable for distributed multi-rate forward including in scenarios with
convergence rate implementation computation, delays
extending a
previously
established
centralized
consensus model
[79] Developing fully Development of two Utilization of Proven the consensus of
2023 distributed adaptive reduced-order adaptive reduced- MASs under general
control for linear multi- protocols: an edge- order consensus conditions
agent systems using only based protocol and a protocols relying
pure relative output node-based protocol solely on relative
information output information
among neighboring
agents.
[80] Developing a finite-time Achievement of Utilization of a gain- Addressing the finite-time
2023 consensus protocol for finite-time scheduled approach consensus problem under
high-order linear multi- consensus with and smooth state- input saturation constraints
agent networks specific feedback for to manage potential
convergence times, designing control unbounded inputs.
based on finite-time protocols
stability theory
[81] Addressing the discrete- Development of a Utilization of a Investigating methods to
2023 time multi-agent two-time-scale Kalman-filter-based minimize the variance of
consensus problem in the protocol alternating estimator in a two- the consensus value
presence of stochastic between estimation time-scale protocol
communication noises and control, based for discrete-time
on Kalman filtering stochastic (MASs).
[82] Improving blockchain Development of Utilization of a BFT- Refinement of the master
2023 throughput and reducing EB-BFT, an elastic based consensus node ordering method to
transaction latency batched consensus mechanism with a efficiently identify
protocol that focus on elasticity malicious nodes or lost
dynamically scales and adaptability transactions
consensus nodes
and transaction
batch sizes.
[83] Developing fully Introduction of two Employing ADET Extension to leader-
2022 distributed and observer- novel adaptive strategies for follower consensus issues.
based consensus control dynamic event- synchronous and Application to more
for multi-agent systems triggered (ADET) asynchronous event- general graphs such as
with general linear strategies for both triggered directed or time-varying
dynamics one-to-all and one- communication, graphs
to-one event- focusing on
triggered observer-based
communication consensus protocols
that are adaptable
and efficient
[84] Enhancing the security Development of the Utilizing blockchain Adapting RPBFT for
2023 and efficiency of Cyber Reinforced Practical technology to handling massive data in
Physical Systems (CPS) Byzantine Fault enhance the PBFT fintech and AI
Tolerance (RPBFT) protocol for better
protocol, an security and
improved version of efficiency in CPS
PBFT
[85] Developing a secure Creation of a sliding Leveraging sliding Exploration of more
2023 consensus control method mode observer to mode control theory comprehensive
for MASs subjected to estimate the state, and a novel coding- communication protocols
DoS attacks actuator, and sensor decoding for various practical
faults of each agent communication scenarios
protocol to enhance
security and stability
in MASs under
cyber threats
[86] Enhancing the Practical introduction of the Utilizing workload- Improving the calculation
2022 Byzantine Fault workload-based based criteria and of node workload values
Tolerance (PBFT) randomization randomization for increased dynamics and
protocol Byzantine fault techniques for adaptability
tolerance protocol selecting consensus
and primary nodes
[87] Developing a Utilization of a Leveraging a Multi- Development of a
2023 decentralized Federated Multi-Hop allowing Hop grounded in decentralized consensus
Learning (FL) algorithm federation clients to dynamical system protocol for delay-critical
using a Multi-Hop share the same consensus theory applications and massive
consensus protocol average model and Lyapunov scenarios,
using only stability arguments
information from to facilitate
their m-step decentralized FL
neighbors
[88] Developing a lightweight, Proposal of the Utilizing a machine Exploration of
2023 efficient consensus Proof of learning-based decentralized model
protocol for Blockchain Evolutionary approach for generation incorporating
Model, a model- iterative training and edge computing and
based evolutionary consensus federated learning
consensus protocol achievement,
leveraging machine tailored for dynamic
learning
 BaaS-based IoT
systems
[89] Developing an adaptive Each agent in the Utilizing an adaptive Extending the protocol to
2023 protocol for MASs to network approach in the directed communication
distributivity track the independently protocol design for graphs
minimum infimum or accesses a local finite-time dynamic
maximum supremum of a time-varying consensus in multi-
set of time-varying exogenous signal. agent systems,
signals in finite-time suitable for various
undirected network
topologies
[91] Addressing the consensus Development of a Derivation of Addressing consensus in
2023 problem in MASs predictor-based consensus conditions MASs with heterogeneous
impacted by input and state feedback that account for the input and communication
communication delays protocol to maximum delays
compensate for eigenvalues of the
input and Laplacian matrix
communication
delays in linear
MASs
[92] Investigating the Obtaining sufficient Utilization of local Focusing on the finite-time
2023 consensus problem in conditions to ensure state information and consensus for discrete-time
discrete-time linear MASs asymptotic construction of a multiagent systems with
with channel fadings in consensus under the dynamic auxiliary quantization over wireless
wireless sensor networks new protocol system to address sensor networks
the consensus
problem
[93] Developing a distributed Design of a fully Utilizing graphical Evaluate protocol
2022 consensus protocol for distributed control games and a performance with
high-order MASs through protocol that modified cost increasing agent numbers
optimization methods in ensures leaderless function to design a and complex network
graphical games consensus and consensus protocol topologies
optimization in
Nash equilibrium
[94] Enhancing the proposing a Employing a Expansion of the protocol
2023 performance and hierarchical hierarchical to support mobile energy
scalability of blockchain- consensus protocol approach to consumers/producers
based peer-to-peer energy that enables blockchain
trading systems concurrent consensus, with
processing of local distinct processes for
and global local and global
transactions transactions, and a
synchronization
method to ensure
consistency
[95] Developing an integrated Introduction of an Applying an Explore the application of
2023 adaptive control approach integrated adaptive integrated adaptive integrated adaptive
for consensus tracking law employing a control method and protocols in more complex
MASs single adaptive Lyapunov stability dynamic systems and real-
coupling gain theory to address world scenarios.
matrix for adjusting consensus tracking
feedback control in nonlinear dynamic
gain and coupling environments
weights
[96] Developing time-to-go Introduction of Implementing time- Extending the framework
2023 consensus protocols to three consensus to-go protocols to include an event-
achieve same-time protocols: max- within a consensus triggered consensus
position consensus of time, min-time, and framework for evaluation mechanism
multiple agents or achievable-time, systems modeled as
vehicles to a destination utilizing a double integrators,
backstepping- and verifying their
inspired speed effectiveness .
modulation
approach
 [97] Addressing the finite-time Developing two Utilizing reference Investigate the application
2023 consensus (FTC) in reference trajectory- trajectory-based and scalability of the finite-
second-order multiagent based finite-time methods, graph time consensus protocols in
systems with non- consensus protocols theory, Lyapunov heterogeneous MASs
identical nonlinear for both leader- functional methods,
dynamics follower and and finite-time
leaderless MASs stability theory
[98] Solving the quasi-sure Introduction of a G- Utilizing Lyapunov Exploring modifications to
2023 consensus problem in noise method, stochastic accommodate agent
MASs under DoS attacks communication analysis theory, and dynamics and noise-
protocol that graph theory to induced consensus in
incorporates time- establish consensus continuous and discrete-
varying and time MASs.
nonlinear
characteristics,
specifically
designed for MASs
under DoS attacks
[99] Development of a Creation of a base Employing agent- Potential use of SIM-P with
2023 simplified consensus model simulating based modeling and optimization algorithms to
protocol simulator for Proof of Work for the sequential Monte find optimal performance
accurate simulation of benchmarking and Carlo method in trade-offs in various
blockchain consensus adaptability to other SIM-P to simulate consensus scenarios
protocols protocols and analyze various
blockchain
consensus protocols

In the field of consensus protocols for multi-agent systems, a series of innovative

studies have emerged, each contributing uniquely to the field. [72] provide a framework
for understanding continuous-time consensus mechanisms, setting a foundational base for
subsequent studies. Building upon this, [76] and [77] delve into more complex
scenarios, introducing considerations for nonlinear dynamics and robustness against
noise, respectively. [79] marks a significant advancement in linear multi-agent systems,
proposing a novel approach to achieve consensus with reduced-order protocols. Further,
[91] explores the challenges posed by channel fadings, offering a dynamic event-based
solution. The integration of differential graphical games into the consensus process, as
presented in [92] adds a strategic layer to multi-agent interactions. Similarly, [94] and
[96] contribute to the field by addressing the complexities of consensus tracking and non-
identical dynamics, respectively. Collectively, these papers not only advance our
understanding of consensus protocols within multi-agent systems but also highlight the
diverse range of challenges and solutions inherent in this rapidly evolving field.
On the other hand, the exploration of consensus protocols within the realm of
blockchain and distributed ledger technologies has led to significant advancements, as
evidenced by a collection of pivotal studies. [75] introduces a novel layered approach
that combines the Raft and Proof of Authority consensus mechanisms, tailored
specifically for UAV networks, thereby enhancing both security and efficiency. In [82]
propose an innovative elastic batched Byzantine Fault Tolerance protocol, which not only
improves scalability but also ensures robust fault tolerance in blockchain systems. [93]
presents a hierarchical and optimistic approach to blockchain consensus, specifically
designed to facilitate efficient and secure peer-to-peer energy trading. Lastly, [88] offers
a groundbreaking solution for integrating blockchain services into the Internet of Things
(IoT), through a lightweight, model-based evolutionary protocol. These papers
collectively underscore the dynamic interplay between blockchain technology and
consensus protocols, highlighting innovative approaches that address scalability, security,
and integration challenges in diverse application domains. Their contributions not only
pave the way for more resilient and efficient blockchain systems but also open new
avenues for the application of distributed ledger technology in various sectors.
Furthermore, the integration of security and fault tolerance mechanisms in
consensus protocols is a critical area of research, as evidenced by a series of insightful
papers. [73] tackles the challenge of designing consensus protocols resilient to Denial-
of-Service attacks, a crucial concern in multi-agent system security. Similarly, [84]
presents an enhanced approach to Byzantine fault tolerance, specifically tailored for the
unique demands of cyber-physical systems, thereby fortifying them against a range of
security threats. [85] introduces a novel coding-decoding strategy, further bolstering
multi-agent systems against DoS attacks. In [86] a unique approach is explored that
employs workload randomization to improve the robustness of Byzantine fault tolerance
in consensus processes. Finally, [97] investigates the application of G-Noise
communication protocols to maintain consensus reliability even in the presence of DoS
attacks. These papers collectively highlight the increasing sophistication of security and
fault tolerance strategies in consensus protocols, emphasizing the importance of robust
and resilient design in an era where cyber threats are increasingly prevalent and diverse.
Their contributions are instrumental in advancing the field towards more secure and
dependable multi-agent and distributed systems.
The advancement of time and event-based consensus protocols represents a
significant stride in the domain of multi-agent systems, as highlighted by a series of
innovative research papers. [80] introduces a finite-time consensus protocol, offering a
novel gain-scheduled method for high-order linear systems, thereby enabling more
efficient consensus achievement. [95]presents an intriguing leaderless approach,
focusing on synchronizing the timing of actions across agents, a critical aspect in
numerous applications. [83] explore the realm of adaptive event-triggered schemes,
which significantly reduce communication demands while maintaining consensus
effectiveness. [81] provides a unique perspective by employing a two-time-scale
approach, adeptly handling the stochastic nature of consensus in discrete-time systems.
These papers collectively emphasize the growing importance of temporal precision and
communication efficiency in consensus protocols. By focusing on finite-time, event-
triggered, and stochastic methods, they contribute to a deeper understanding of dynamic
consensus processes and pave the way for more responsive and resource-efficient multi-
agent system operations.
The applications and adaptations of consensus protocols in various specialized
contexts are also been address in many studies, such as [74] delves into the critical aspect
of security in crowdsourcing environments, proposing a protocol that is resilient to
fraudulent activities. While [78] explores the implementation of consensus protocols in
discrete-time systems, a fundamental step towards broader applicability in various digital
and computational platforms. [87] addresses the integrating consensus protocols into the
burgeoning field of decentralized federated learning, addressing the challenges of multi-
hop communication in such networks. [89] presents an adaptive approach to achieving
consensus in dynamic environments, particularly where rapid convergence to extreme
values is required. Lastly, [98] introduces a simulation tool that aids in the evaluation and
analysis of consensus protocols, particularly in the context of reputation and contribution-
based systems. These papers collectively highlight the versatility and adaptability of
consensus protocols, demonstrating their potential in addressing specific challenges and
requirements in varied application domains. Their contributions are not only pivotal in
advancing the theoretical underpinnings of consensus mechanisms but also instrumental
in their practical implementation across diverse technological landscapes.
Based on the summary table ( Table 1), future research in the area of distributed
consensus protocols is abundant and diverse. There are clear opportunities to extend the
current understanding of consensus mechanisms in various contexts, such as improving
security in multi-agent systems, enhancing the efficiency and scalability of blockchain
protocols, and integrating these protocols in emerging technologies like the Internet of
Things (IoT). Additionally, there is room to explore novel approaches in dealing with
noise and delay in communication networks, as well as in the application of consensus
protocols in complex systems like cyber-physical systems and decentralized learning
networks. The table suggests that future work can also focus on optimizing consensus
protocols for specific application domains and addressing the unique challenges they
present.

5. Extracted Statistics
Based on the detailed review of the two previous sections, we extract some
statistical data that can be useful for having a comprehensive understanding of the trends
and patterns observed, which will be discussed in this section.
In terms of common research areas (see Figure 1 ), the data reveals significant
attention in several key fields: First, New Consensus Protocols, with 20 papers like [78],
and others, focus on distributed network agreement protocols. Then, Multi-Agent
Systems are covered in 17 papers, including [74] highlighting interactions among
autonomous agents. Blockchain Technology is the focus of 7 papers such as
[75]exploring its application in various sectors. Networked Control Systems are discussed
in 14 papers like [76] ,emphasizing integrated management of automated systems. Lastly,
Cybersecurity features in 9 papers, including [81], addressing the protection of data in
MASs and blockchain contexts. This comprehensive overview not only sheds light on the
current priorities and emerging topics but also indicates the interdisciplinary nature of
these research areas, with some papers contributing to multiple categories.

Common Study Areas

25

20

15

10

0
New Consensus Multi-Agent Blockchain Networked Security
Protocols Systems Technology Control Systems

Figure 1. Common Study Areas.

 Furthermore, our analysis sheds light on some common challenges and issues faced
in the field (see Figure 2). A significant concern is Consensus under Denial of Service
(DoS) Attacks, with 3 papers, such as [81] addressing this issue. This underscores the
vulnerability of consensus mechanisms in the face of cyber-attacks. Synchronization in
Multi-Agent Systems (MASs) is another area of focus, with 2 papers, numbered [90] and
[95] ,exploring this complex issue. The challenge of High-Order Nonlinearity is dealt
with in 1 paper, numbered [76] , pointing to the intricacies involved in handling
sophisticated algorithmic behaviours. The topic of Scalability in Blockchain Systems is
discussed in 2 papers, [82] and [93], reflecting ongoing efforts to enhance the efficiency
and capacity of blockchain networks. Lastly, the issue of Cybersecurity in Cyber-Physical
Systems is addressed in 3 papers, including [85] highlighting the critical need for secure
and robust systems in an increasingly interconnected digital environment. These
challenges, varying in their nature and impact, illustrate the dynamic hurdles that
researchers and practitioners must navigate to advance the field.

Common Challenges

Consensus under Denial of Service

(DoS) Attacks
Multi-Agent Systems

High-Order Nonlinearity

Blockchain Systems

Cybersecurity

Figure 2. common challenges.

6. Recommendations
The rapid evolution of cloud computing and the increasing importance of
distributed consensus protocols present numerous opportunities for further research and
advancements. In this section, we outline several recommendations that can guide
researchers and practitioners in exploring new directions and addressing existing
challenges in the field.
One area of focus for future research is the scalability and performance optimization
of distributed consensus protocols. As cloud computing environments continue to grow
in scale and complexity, it becomes crucial to develop protocols that can handle large
numbers of nodes and transactions efficiently. Researchers should explore techniques
such as parallelization, load balancing, and optimization algorithms to enhance the
scalability and performance of consensus protocols.
Ensuring fault tolerance and resilience is another critical aspect that requires
attention. As cloud systems are vulnerable to various types of failures, including node
failures and network disruptions, it is essential to design consensus protocols that can
tolerate and recover from such failures. Future research should focus on developing
mechanisms for fault detection, fault recovery, and adaptive fault tolerance to enhance
the robustness of distributed consensus protocols.
With the increasing reliance on cloud computing for sensitive applications and data,
security and privacy considerations become paramount. Researchers should explore
techniques for securing consensus protocols against malicious attacks, such as Byzantine
faults, and ensuring the confidentiality and integrity of data during the consensus process.
Additionally, privacy-enhancing mechanisms, such as secure multiparty computation and
zero-knowledge proofs, can be investigated to protect sensitive information during
consensus.
The integration of distributed consensus protocols with emerging technologies can
unlock new possibilities and capabilities. Researchers should explore the synergy
between consensus protocols and technologies such as artificial intelligence, machine
learning, and blockchain. Leveraging these technologies can enhance the efficiency,
adaptability, and trustworthiness of consensus protocols, enabling novel applications and
use cases in cloud computing.
As distributed consensus protocols continue to evolve, it becomes crucial to
establish standards and promote interoperability among different protocols and systems.
Researchers and practitioners should collaborate to develop common frameworks,
protocols, and APIs that facilitate seamless integration and interoperability between
various consensus mechanisms. This standardization effort can enhance the adoption and
compatibility of distributed consensus protocols in diverse cloud computing
environments.
By addressing these recommendations, researchers and practitioners can contribute
to the advancement of distributed consensus protocols, enabling more efficient, reliable,
and secure cloud computing systems.

7. Conclusion
In this paper, we've comprehensively reviewed the advancements in consensus
protocols across various domains, including multi-agent systems, blockchain
technologies, and their applications in enhancing security and fault tolerance. Our
exploration reveals significant innovations in continuous-time consensus mechanisms,
robustness against noise, and reduced-order protocols. We've seen novel approaches in
blockchain consensus for UAV networks, IoT, and peer-to-peer energy trading, alongside
advancements in Byzantine fault tolerance for cyber-physical systems. Time and event-
based consensus protocols have shown promise in achieving efficiency and temporal
precision. The adaptability of these protocols in crowdsourcing, federated learning, and
other dynamic environments underscores their versatility. Collectively, these studies not
only broaden our understanding of consensus protocols but also demonstrate their
evolving role in addressing complex challenges in distributed systems.
The diversity of research in the field of consensus protocols underscores the vast
array of challenges in distributed systems, each spawning unique opportunities for further
exploration. This variety reflects the complexity of the field and indicates a rich landscape
of research prospects. The range of topics, from the intricacies of multi-agent systems to
the robustness required in blockchain technologies, not only highlights current challenges
but also opens doors for innovative solutions and advancements in future studies. This
breadth of research signifies the dynamic and ever-evolving nature of consensus protocol
studies, promising continued advancements and breakthroughs.

References
1. Hassan, R.J., et al., State of art survey for iot effects on smart city technology: challenges,
opportunities, and solutions. Asian Journal of Research in Computer Science, 2021. 22:
p. 32-48.
2. Yahia, H.S., et al., Comprehensive survey for cloud computing based nature-inspired
algorithms optimization scheduling. Asian Journal of Research in Computer Science,
2021. 8(2): p. 1-16.
3. Abdulrahman, L.M., et al., A state of art for smart gateways issues and modification. Asian
Journal of Research in Computer Science, 2021. 7(4): p. 1-13.
4. Al-Zebari, A. and A. Sengur. Performance comparison of machine learning techniques on
diabetes disease detection. in 2019 1st international informatics and software
engineering conference (UBMYK). 2019. IEEE.
5. Zeebaree, A., et al., Designing an ontology of E-learning system for duhok polytechnic
university using protégé OWL tool. J Adv Res Dyn Control Syst Vol, 2019. 11: p. 24-37.
6. AL-Zebari, A., et al., ELMS–DPU ontology visualization with Protégé VOWL and Web
VOWL. Journal of Advanced Research in Dynamic and Control Systems, 2019. 11: p.
478-85.
7. Adel, A.-Z., S. Zebari, and K. Jacksi, Football Ontology Construction using Oriented
Programming. Journal of Applied Science and Technology Trends, 2020. 1(1): p. 24-30.
8. Omar, N., A. Al-zebari, and A. Sengur. Deep learning approach to predict forest fires using
meteorological measurements. in 2021 2nd international informatics and software
engineering conference (IISEC). 2021. IEEE.
9. Zebari, G.M., D.A. Zebari, and A. Al-zebari, Fundamentals of 5G cellular networks: A
review. Journal of Information Technology and Informatics, 2021. 1(1): p. 1-5.
10. Selamat, S.A.A.-Z.A., Electronic learning management system based on semantic web
technology: A review. Int. J. Adv. Electron. Comput. Sci, 2017. 4(3): p. 1-6.
11. Turkoglu, M., et al., Deep rhythm and long short term memory-based drowsiness
detection. Biomedical Signal Processing and Control, 2021. 65: p. 102364.
12. Zebari, D.A., D.A. Ibrahim, and A. Al-Zebari. Suspicious region segmentation using deep
features in breast cancer mammogram images. in 2022 International Conference on
Computer Science and Software Engineering (CSASE). 2022. IEEE.
13. Al-zebari, A., Peer to Peer File Sharing System. 2017: LAP LAMBERT Academic
Publishing.
14. Alsandi, N.S.A., et al., A Multi-Stream Scrambling and DNA Encoding Method Based Image
Encryption. Computer Systems Science & Engineering, 2023. 47(2).
15. Omar, N., A. Al-zebari, and A. Sengur. Improving the clustering performance of the k-
means algorithm for non-linear clusters. in 2022 4th International Conference on
Advanced Science and Engineering (ICOASE). 2022. IEEE.
16. Barwary, M.J., K. Jacksi, and A. Al-zebari, An Intelligent and Advance Kurdish Information
Retrieval Approach with Ontologies: A Critical Analysis. International Journal of
Intelligent Systems and Applications in Engineering, 2023. 11(11s): p. 189-199.
17. Abdulqadir, H.R., et al., A study of moving from cloud computing to fog computing.
Qubahan Academic Journal, 2021. 1(2): p. 60-70.
18. Al-Zebari, A., Ensemble convolutional neural networks and transformer-based
segmentation methods for achieving accurate sclera segmentation in eye images. Signal,
Image and Video Processing, 2024. 18(2): p. 1879-1891.
19. Barwary, M.J., K. Jacksi, and A. Al-Zebari, Constructing a Multilingual E-Learning
Ontology through Web Crawling and Scraping. International Journal of Communication
Networks and Information Security, 2023. 15(3): p. 137-153.
20. Ahmed, S.H., et al., Parameter Tuned Deep Learning Based Traffic Critical Prediction
Model on Remote Sensing Imaging. Computers, Materials & Continua, 2023. 75(2).
21. Haji, S.H., et al., Document Clustering in the Age of Big Data: Incorporating Semantic
Information for Improved Results. Journal of Applied Science and Technology Trends,
2023. 4(01): p. 34-53.
22. Ramadhan, M. and A. Al-zebari, Proposed A Web-Based Intelligent System to Manage the
Blood Bank in Zakho District. QALAAI ZANIST JOURNAL, 2023. 8(5): p. 1247-1263.
23. Omar, N., A. Al-Zebari, and A. Sengur. Fourier Domain Kernel Density Estimation-based
Approach for Hail Sound classification. in 2022 3rd International Informatics and
Software Engineering Conference (IISEC). 2022. IEEE.
24. Mustafa, Q., F. Naglaa, and A. Adel, Predation Capacity and Some Biological Parameters
of Chrysoperla Carnea (Stephens) On Aphis Craccivora Koch under Laboratory
Conditions.
25. Al-Zebari, A., N. Omar, and A. Sengur. Vision Transformers-based Hand Gesture
Classification. in 2022 3rd International Informatics and Software Engineering
Conference (IISEC). 2022. IEEE.
26. Idrus, Z., et al. Swarm Plot: Data Redistribution in Non-Random Technique. in 2021 6th
IEEE International Conference on Recent Advances and Innovations in Engineering
(ICRAIE). 2021. IEEE.
27. Abdulrahman, L.M., S.R. Zeebaree, and N. Omar, State of Art Survey for Designing and
Implementing Regional Tourism Web based Systems. Academic Journal of Nawroz
University, 2022. 11(3): p. 100-112.
28. Omer, M.A., et al., A Survey on Cloud Security: Concepts, Types, Limitations, and
Challenges. Journal of Applied Science and Technology Trends, 2022. 3(02): p. 101-111.
29. Abdulrahman, L.M., et al., Web phishing detection using web crawling, cloud
infrastructure and deep learning framework. Journal of Applied Science and Technology
Trends, 2023. 4(01): p. 54-71.
30. Malallah, H.S., et al., Performance analysis of enterprise cloud computing: a review.
Journal of Applied Science and Technology Trends, 2023. 4(01): p. 01-12.
31. Abdulrahman, L.M., A.M. Abdulazeez, and D.A. Hasan, COVID-19 world vaccine adverse
reactions based on machine learning clustering algorithm. Qubahan Academic Journal,
2021. 1(2): p. 134-140.
32. Yazdeen, A.A., et al., Internet of Things Impact on Web Technology and Enterprise
Systems. Journal of Applied Science and Technology Trends, 2023. 4(01): p. 19-33.
33. Mehyadin, A.E., et al., Distributed Fundamentals based Conducting the Web Crawling
Approaches and Types (Focused, Incremental, Distributed, Parallel, Hidden Web, Form
Focused and Breadth First) Crawlers. Journal of Smart Internet of Things, 2023.
2022(1): p. 10-32.
34. Ismael, H.R., et al., Web Technology Grounded Effects of Task Scheduling in Distributed
and Cloud Systems. Journal of Smart Internet of Things, 2023. 2022(1): p. 196-218.
35. Armya, R.E., et al., Web-based Efficiency of Distributed Systems and IoT on Functionality
of Smart City Applications. Journal of Smart Internet of Things, 2023. 2023(2): p. 142-
161.
36. Abdullah, R.M., et al., Modular Platforms based on Clouded Web Technology and
Distributed Deep Learning Systems. Journal of Smart Internet of Things, 2023. 2023(2):
p. 154-173.
37. Amanoul, S.V., et al., Orchestrating Distributed Computing and Web Technology with
Semantic Web and Big Data. Journal of Smart Internet of Things, 2023. 2023(2): p. 174-
192.
38. Shamsaldin, N., The Influence of SIMD Architecture on the Efficiency of Parallel
Processing Approaches: From Web and AI Perspective. Journal of Information
Technology and Informatics, 2024. 3(1).
39. Sadeeq, M.M., et al., IoT and Cloud computing issues, challenges and opportunities: A
review. Qubahan Academic Journal, 2021. 1(2): p. 1-7.
40. ABDULKAREEM, N.M. and S.R. ZEEBAREE, OPTIMIZATION OF LOAD BALANCING
ALGORITHMS TO DEAL WITH DDOS ATTACKS USING WHALE OPTIMIZATION
ALGORITHM. Journal of Duhok University, 2022. 25(2): p. 65-85.
41. Rashid, Z.N., S.R. Zeebaree, and A. Shengul. Design and analysis of proposed remote
controlling distributed parallel computing system over the cloud. in 2019 International
Conference on Advanced Science and Engineering (ICOASE). 2019. IEEE.
42. Haji, L.M., et al., Dynamic resource allocation for distributed systems and cloud
computing. TEST Engineering & Management, 2020. 83(May/June 2020): p. 22417-
22426.
43. Zeebaree, S.R., et al., Characteristics and analysis of hadoop distributed systems.
Technology Reports of Kansai University, 2020. 62(4): p. 1555-1564.
44. Ageed, Z.S., et al., A state of art survey for intelligent energy monitoring systems. Asian
Journal of Research in Computer Science, 2021. 8(1): p. 46-61.
45. Ibrahim, R., S. Zeebaree, and K. Jacksi, Survey on semantic similarity based on document
clustering. Adv. sci. technol. eng. syst. j, 2019. 4(5): p. 115-122.
46. Zebari, D.A., et al. Enhance the mammogram images for both segmentation and feature
extraction using wavelet transform. in 2019 International Conference on Advanced
Science and Engineering (ICOASE). 2019. IEEE.
47. Saeed, J. and S. Zeebaree, Skin lesion classification based on deep convolutional neural
networks architectures. Journal of Applied Science and Technology Trends, 2021. 2(01):
p. 41-51.
48. Omer, M.A., et al., Efficiency of malware detection in android system: A survey. Asian
Journal of Research in Computer Science, 2021. 7(4): p. 59-69.
49. Zebari, S. and N.O. Yaseen, Effects of parallel processing implementation on balanced
load-division depending on distributed memory systems. J. Univ. Anbar Pure Sci, 2011.
5(3): p. 50-56.
50. Zebari, D.A., et al. Multi-level of DNA encryption technique based on DNA arithmetic and
biological operations. in 2018 International Conference on Advanced Science and
Engineering (ICOASE). 2018. IEEE.
51. Jacksi, K., S.R. Zeebaree, and N. Dimililer, Lod explorer: Presenting the web of data. Int. J.
Adv. Comput. Sci. Appl. IJACSA, 2018. 9(1): p. 1-7.
52. Jacksi, K., N. Dimililer, and S. Zeebaree, State of the art exploration systems for linked
data: a review. Int. J. Adv. Comput. Sci. Appl. IJACSA, 2016. 7(11): p. 155-164.
53. Jacksi, K., N. Dimililer, and S.R. Zeebaree, A survey of exploratory search systems based
on LOD resources. 2015.
54. Abdullah, P.Y., et al., HRM system using cloud computing for Small and Medium
Enterprises (SMEs). Technology Reports of Kansai University, 2020. 62(04): p. 04.
55. Zeebaree, S., et al., Security approaches for integrated enterprise systems performance: A
Review. Int. J. Sci. Technol. Res, 2019. 8(12): p. 2485-2489.
56. Salim, N.O., et al. Study for food recognition system using deep learning. in Journal of
Physics: Conference Series. 2021. IOP Publishing.
57. Salih, A., et al. A survey on the role of artificial intelligence, machine learning and deep
learning for cybersecurity attack detection. in 2021 7th International Engineering
Conference “Research & Innovation amid Global Pandemic"(IEC). 2021. IEEE.
58. Jacksi, K., et al. Clustering documents based on semantic similarity using HAC and K-mean
algorithms. in 2020 International Conference on Advanced Science and Engineering
(ICOASE). 2020. IEEE.
59. Dino, H., et al., Facial expression recognition based on hybrid feature extraction
techniques with different classifiers. TEST Engineering & Management, 2020. 83: p.
22319-22329.
60. Zeebaree, S.R., et al. Design and simulation of high-speed parallel/sequential simplified
DES code breaking based on FPGA. in 2019 International Conference on Advanced Science
and Engineering (ICOASE). 2019. IEEE.
61. Zeebaree, S., DES encryption and decryption algorithm implementation based on FPGA.
Indones. J. Electr. Eng. Comput. Sci, 2020. 18(2): p. 774-781.
62. Hasan, D.A., et al. Machine learning-based diabetic retinopathy early detection and
classification systems-a survey. in 2021 1st Babylon International Conference on
Information Technology and Science (BICITS). 2021. IEEE.
63. Hasan, D.A., et al., The impact of test case generation methods on the software
performance: A review. International Journal of Science and Business, 2021. 5(6): p. 33-
44.
64. Zebari, I.M., S.R. Zeebaree, and H.M. Yasin. Real time video streaming from multi-source
using client-server for video distribution. in 2019 4th Scientific International Conference
Najaf (SICN). 2019. IEEE.
65. Malallah, H., et al., A comprehensive study of kernel (issues and concepts) in different
operating systems. Asian Journal of Research in Computer Science, 2021. 8(3): p. 16-31.
66. Mohammed, S.M., K. Jacksi, and S. Zeebaree, A state-of-the-art survey on semantic
similarity for document clustering using GloVe and density-based algorithms. Indonesian
Journal of Electrical Engineering and Computer Science, 2021. 22(1): p. 552-562.
67. Abdulrazaq, M.B., et al. An analytical appraisal for supervised classifiers’ performance on
facial expression recognition based on relief-F feature selection. in Journal of Physics:
Conference Series. 2021. IOP Publishing.
68. Mahmood, M.R., et al., Classification techniques’ performance evaluation for facial
expression recognition. Indonesian Journal of Electrical Engineering and Computer
Science, 2021. 21(2): p. 176-1184.
69. Khalid, Z.M. and S.R. Zeebaree, Big data analysis for data visualization: A review.
International Journal of Science and Business, 2021. 5(2): p. 64-75.
70. Shukur, H., et al., Cache coherence protocols in distributed systems. Journal of Applied
Science and Technology Trends, 2020. 1(2): p. 92-97.
71. Ibrahim, B.R., et al. Embedded system for eye blink detection using machine learning
technique. in 2021 1st Babylon International Conference on Information Technology and
Science (BICITS). 2021. IEEE.
72. Rezaei, V. and E. Khanmirza, Continuous-time min-max consensus protocol: A unified
approach. Mathematics and Computers in Simulation, 2024. 218: p. 292-310.
73. Li, S., et al., Design of non-smooth consensus protocol for multi-agent systems under DoS
attacks. Applied Mathematics and Computation, 2024. 463: p. 128360.
74. Afzaal, H., M. Imran, and M.U. Janjua, Formal verification of fraud-resilience in a
crowdsourcing consensus protocol. Computers & Security, 2023. 131: p. 103290.
75. Zhang, Y., W. Wang, and F. Shi, Reputation-based Raft-Poa layered consensus protocol
converging UAV network. Computer Networks, 2024. 240: p. 110170.
76. Wen, H., et al., A consensus protocol considering Lipschitz constant and communication
topology condition of high-order nonlinear multi-agent systems. Communications in
Nonlinear Science and Numerical Simulation, 2022. 111: p. 106499.
77. Luo, J., et al., A new class of robust and predefined-time consensus protocol based on noise-
tolerant ZNN models. Applied Soft Computing, 2023. 145: p. 110550.
78. Cacace, F., et al., A new distributed protocol for consensus of discrete-time systems.
European Journal of Control, 2023. 74: p. 100833.
79. Ma, Y., X. Li, and S. Li, Consensus of Linear Multi-agent Systems with Pure Relative Output
Through Fully Distributed Reduced-Order Protocols. IFAC-PapersOnLine, 2023. 56(2): p.
10162-10167.
80. Zhou, Y., Y. Liu, and Y. Zhao, Design Finite-time Consensus Protocol for High-order Linear
Multi-agent Systems: A Gain Scheduled Approach⋆. IFAC-PapersOnLine, 2023. 56(2): p.
8123-8128.
81. Zhang, W. and H. Tang, Discrete-time stochastic consensus: A Kalman-filter-based two-
time-scale protocol. Systems & Control Letters, 2023. 174: p. 105487.
82. Zhang, B., et al., EB-BFT: An elastic batched BFT consensus protocol in blockchain. Future
Generation Computer Systems, 2023. 139: p. 267-279.
83. Xu, W., et al., Fully distributed observer-based consensus protocol: Adaptive dynamic
event-triggered schemes. Automatica, 2022. 139: p. 110188.
84. Wu, Y., L. Wu, and H. Cai, Reinforced practical Byzantine fault tolerance consensus
protocol for cyber physical systems. Computer Communications, 2023. 203: p. 238-247.
85. Ru, T., H. Lin, and C. Cai, Secured consensus control for multi-agent systems under DoS
attacks: Coding–decoding communication protocol. Nonlinear Analysis: Hybrid Systems,
2024. 52: p. 101445.
86. Huang, B., et al., Workload-based randomization byzantine fault tolerance consensus
protocol. High-Confidence Computing, 2022. 2(3): p. 100070.
87. Menegatti, D., et al., A Discrete-Time Multi-Hop Consensus Protocol for Decentralized
Federated Learning. IEEE Access, 2023.
88. Zhao, Y., et al., A lightweight model-based evolutionary consensus protocol in blockchain
as a service for IoT. IEEE Transactions on Services Computing, 2023.
89. Lippi, M., et al., An adaptive distributed protocol for finite-time infimum or supremum
dynamic consensus. IEEE Control Systems Letters, 2022. 7: p. 401-406.
90. Sun, H.-Y., et al., Consensus of mass with input and communication delays by predictor-
based protocol. IEEE Transactions on Cybernetics, 2022.
91. Zhang, H., Y. Fan, and J. Qiu, Consensus protocol for discrete-time linear multiagent
systems with channel fadings: A dynamic event-based approach. IEEE Transactions on
Control of Network Systems, 2022. 10(1): p. 345-354.
92. Zhang, S., et al., Distributed Consensus Protocol for Multi-Agent Differential Graphical
Games. IEEE Transactions on Circuits and Systems II: Express Briefs, 2023.
93. Abdella, J., et al., Hicoob: Hierarchical concurrent optimistic blockchain consensus
protocol for peer-to-peer energy trading systems. IEEE Transactions on Smart Grid,
2022.
94. Wang, Q., Y. Lv, and Q. Wang, Integrated Adaptive Protocols for Consensus Tracking With
Lipchitz Nonlinear Dynamics. IEEE Transactions on Control of Network Systems, 2022.
95. Voleti, R.S., D. Bhattacharjee, and K. Subbarao, Leaderless Time-To-Go Protocols for
Same-Time Position Consensus. IEEE Transactions on Systems, Man, and Cybernetics:
Systems, 2023.
96. Zou, W., et al., Novel reference trajectory-based finite-time consensus protocols for
multiagent systems with non-identical nonlinear dynamics. IEEE Transactions on
Network Science and Engineering, 2022. 10(2): p. 1107-1118.
97. Li, S., J. Zhao, and X. Ding, Quasi sure consensus of multi-agent systems: G-noise
communication protocols under dos attacks. IEEE Transactions on Control of Network
Systems, 2022.
98. Oyinloye, D.P., et al., SIM-P—A Simplified Consensus Protocol Simulator: Applications to
Proof of Reputation-X and Proof of Contribution. IEEE Internet of Things Journal, 2022.
10(6): p. 5083-5094.