research paper on blockchain in cyber security

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 11 January 2024

Blockchain security enhancement: an approach towards hybrid consensus algorithms and machine learning techniques

• K. Venkatesan 1 , 2 &
• Syarifah Bahiyah Rahayu 1   nAff2  

Scientific Reports volume  14 , Article number:  1149 ( 2024 ) Cite this article

2040 Accesses

Metrics details

• Engineering
• Mathematics and computing

In this paper, we propose hybrid consensus algorithms that combine machine learning (ML) techniques to address the challenges and vulnerabilities in blockchain networks. Consensus Protocols make ensuring agreement among the applicants in the distributed systems difficult. However, existing mechanisms are more vulnerable to cyber-attacks. Previous studies extensively explore the influence of cyber attacks and highlight the necessity for effective preventive measures. This research presents the integration of ML techniques with the proposed hybrid consensus algorithms and advantages over predicting cyber-attacks, anomaly detection, and feature extraction. Our hybrid approaches leverage and optimize the proposed consensus protocols' security, trust, and robustness. However, this research also explores the various ML techniques with hybrid consensus algorithms, such as Delegated Proof of Stake Work (DPoSW), Proof of Stake and Work (PoSW), Proof of CASBFT (PoCASBFT), Delegated Byzantine Proof of Stake (DBPoS) for security enhancement and intelligent decision making in consensus protocols. Here, we also demonstrate the effectiveness of the proposed methodology within the decentralized networks using the ProximaX blockchain platform. This study shows that the proposed research framework is an energy-efficient mechanism that maintains security and adapts to dynamic conditions. It also integrates privacy-enhancing features, robust consensus mechanisms, and ML approaches to detect and prevent security threats. Furthermore, the practical implementation of these ML-based hybrid consensus models faces significant challenges, such as scalability, latency, throughput, resource requirements, and potential adversarial attacks. These challenges must be addressed to ensure the successful implementation of the blockchain network for real-world scenarios.

Similar content being viewed by others

Highly accurate protein structure prediction with AlphaFold

John Jumper, Richard Evans, … Demis Hassabis

High-threshold and low-overhead fault-tolerant quantum memory

Sergey Bravyi, Andrew W. Cross, … Theodore J. Yoder

High-throughput prediction of protein conformational distributions with subsampled AlphaFold2

Gabriel Monteiro da Silva, Jennifer Y. Cui, … Brenda M. Rubenstein

Introduction

The Consensus protocols employed in the blockchain network provide high security and more efficient operations. Hybrid consensus algorithms are developed by combining the key elements of various consensus algorithms. This might be useful to prevent double-spending and 51% of attacks. Combining Proof of Work (PoW) and Delegated Proof of Stake (DPoS) improves computation performance and enhances high security 1 . DPoS is used for block validation, and PoW is used for block creation, making it more difficult for an attacker to control the network. The combination of Proof of Stake (PoS) and Proof of Work (PoW) results in better security performance along with the network's decentralization 2 . PoS is used for block validation, and PoW is used for block creation, increasing the network's security and decentralization 3 . Integrating the DPoS and Practical Byzantine Fault Tolerance (PBFT) provides higher security, scalability, and efficiency. Here, DPoS is used for block creation, and PBFT is used for block validation, providing better security and scalability 4 .

Hybridization of the Casper and PBFT consensus algorithms can provide a higher level of security against 51% of attacks in blockchain technology. Casper Algorithm adopts PoS for block validation 5 . Here, validators are selected based on their network stake size. PBFT works differently compared to other algorithms. It confides a group of validators to grasp the following block consensus. Hybridization of PoS and PBFT leverages high-level security and rapid consensus time. This makes it very difficult for attackers to access the network and perform cyber-attacks. In addition, consensus hybridization balances scalability and decentralization, which is convenient for microgrid networks 6 . For any algorithm, achieving a complete security audit and system update, which confirms the efficacy, is essential. Compared to other individual consensus algorithms, these hybridization algorithms enhance security by preventing 51% of attacks 7 . However, analyzing the trade-offs before implementing them in blockchain networks is mandatory. Figure  1 shows the basic block diagram using the blockchain network and ML techniques.

Basic block diagram using blockchain and ML techniques.

Many researchers propose a new consensus protocol that performs better security and scalability and reduces the probability of cyber-attacks on the blockchain network. In 8 , the author proposed Algorand, which achieves transaction finality and high scalability 9 introduces consensus algorithms, which enhance throughput, scalability, and security. The authors Zhu 10 introduced the GHOST protocol, the modified PoW consensus protocol, which achieves the best security and high throughput transaction. Meanwhile 11 , A. Kiayias et al. propose ouroboros, the modified PoS protocol that enhances scalability and security. Lashkari et al. 12 eliminate 51% of attacks using Bitcoin-NG, improving throughput and scalability. These works accentuate the necessity of creating new consensus protocols that perform highly scalable security and have the potential to handle vast volumes of transactions.

The Hybrid consensus algorithms with Machine Learning (ML) techniques gained significant approaches, which led to high performance, improved scalability, and enhanced security in blockchain networks. Several researchers, such as Zhang 13 , Andoni 14 , Wang 15 , and Yang 16 , conduct various simulations and experiments and discusses hybrid protocols, which evaluate their efficacy and perform better results and security than the existing consensus mechanisms. Additionally, Mahmood et al. 17 propose a comprehensive review of consensus protocols that enlighten the potential of ML approaches for enhancing performance and security. The author highlights the hybrid consensus mechanism and ML integration to achieve optimal block validation and improve network performance.

Over the years, the performance of the hybrid consensus protocol combined with Particle Swarm Optimization (PSO) has been studied. In 18 , Zhu et al. propose a PoS-based consensus model with PSO that achieves high scalability, better performance, and enhanced security. Razali et al. 19 Introduce a new consensus model, which uses PSO for block validation optimization. Ali et al. 20 Work on hybrid consensus protocol with PSO approaches, which results in high security and scalability and maintains high performance. Ullah et al. 21 propose ML-based PSO for faster block validation and better throughput. Kumar et al. 22 Provide a comparative study of existing consensus protocol and highlight the potential of the hybrid consensus protocol with PSO optimization techniques. It also demonstrates the strength of hybridization and achieves better performance and scalability, enlightening future research requirements in blockchain networks.

Unsupervised, supervised, and rule-based ML approaches 23 are vital in responding to and detecting microgrid attacks in blockchain networks. ML techniques may also be used to solve communication and behavior-based attacks. ML-based hybrid consensus algorithms improve security and other performance factors and provide a scope for active research 24 . Specifically, new consensus protocols with ML approaches detect and prevent the significant threats of attacks in blockchain networks. These papers propose simulation models and demonstrate experiments that can perform better security than any other consensus mechanisms 25 . Additionally, these papers provide a comprehensive overview, realize the potential of hybridization, and highlight the importance of further research to develop more effective and efficient methods to secure blockchain mechanisms 26 .

However, the researchers propose a diverse solution, develop a creative approach, and offer a state-of-the-art framework to overcome cybersecurity issues. In 27 , the author proposes Fed-Inforce-Fusion, which is a reinforcement learning-based fusion model for the Internet of Medical Things (IoMT) networks. This method incorporates federated and reinforcement learning to improve accuracy and detection. Meanwhile, the "PC-IDS" framework uses hybrid machine learning approaches to identify harmful behaviors and secure privacy in cyber-physical power networks 28 . However, deep autoencoder IDS performs well and accurately in real-time intrusion detection in IIoT networks 29 .

Additionally, the multi-stage AV framework combines the state-of-the-art framework and deep learning techniques, outperforming existing systems in understanding and identifying cyber risks within autonomous vehicles 30 . Therefore, the relevance of these studies lies in their exploration of hybrid consensus algorithms and integration of machine learning approaches, offering valuable insights into blockchain and cyber security attacks. Despite their contributions, several limitations were prevalent in the studies reviewed. These limitations underscore the need for further research to address these gaps and refine our understanding of integrating machine-learning approaches in hybridizing consensus algorithms.

Open challenges and motivation

Consensus algorithms face several open challenges that need to be addressed for the widespread adoption of blockchain technology in real-world applications. One of the primary challenges is scalability , as consensus mechanisms must efficiently handle many transactions per second without compromising security and decentralization. Another critical challenge is energy efficiency , particularly in Proof of Work (PoW), where the high energy consumption is unsustainable and costly in the long run. Developing energy-efficient consensus algorithms or improving existing ones is vital for practical applications. Ensuring fast transaction confirmation times is another challenge to meeting the demands of real-time applications. Long confirmation times can hinder the usability of blockchain technology, making it necessary to optimize latency and transaction confirmation. Security is an ongoing concern, and consensus mechanisms must resist attacks such as double-spending, Sybil, and 51% attacks. Enhancing security measures is crucial for building trust and widespread adoption.

Governance and compliance are significant challenges, as consensus algorithms must align with legal and regulatory requirements while maintaining decentralization. Finding the right balance between compliance and decentralization is crucial for the finance, healthcare, and supply chain management industries. Interoperability between blockchain networks and consensus algorithms is essential for collaboration and communication across systems. Developing standards and protocols for seamless integration and data exchange is challenging. Privacy and confidentiality are paramount, and consensus algorithms must incorporate robust techniques to protect sensitive data while maintaining transparency and audibility. Consensus mechanisms should adapt to dynamic networks , where nodes can join or leave anytime and handle challenges such as node churn, network partitions, and malicious nodes. Improving the user experience is vital for widespread adoption, requiring consensus algorithms to minimize transaction fees, reduce latency, and provide a seamless and intuitive interface.

Addressing the environmental impact of consensus algorithms, particularly energy-intensive ones like PoW, is a pressing challenge. Developing sustainable and eco-friendly consensus algorithms is essential to align blockchain technology with global sustainability goals. Overcoming these challenges will require continuous research, innovation, and collaboration between industry, academia, and regulatory bodies. By addressing these open challenges, consensus algorithms can pave the way for the widespread adoption of blockchain technology in various real-world applications.

Research gap identification

The analysis of existing literature reveals a distinct gap in integrating hybrid consensus algorithms with machine learning approaches. While previous studies have laid essential groundwork, there remains unexplored territory in understanding the concepts of hybridization and difficulties in integrating machine learning approaches to enhance the security of the blockchain network. This research strives to fill this gap by defining the research objectives, thereby advancing our understanding of improving security in blockchain networks and protecting the network from malicious attacks.

Research objectives

The main goal of this research paper is to implement a hybrid consensus mechanism with ML techniques, which enhances the security of the consensus mechanisms and avoids cyber attacks. The contribution of this research is listed below.

To identify and understand the vulnerabilities in existing mechanisms. Here, analyzing and evaluating the security shortcomings enables a comprehensive understanding of the vectors and potential threats.

To reframe the hybrid consensus Algorithms. A comprehensive analysis of consensus algorithm hybridization and its necessity in cyber security is discussed here.

To develop the ML framework for extraction of the features and anomaly detection. The critical aspects of the ML framework that effectively performs feature extraction, anomalies, and malicious activity detection within the blockchain network must be performed. This framework will leverage advanced ML algorithms to analyze network behavior, identify suspicious patterns, and distinguish normal activities from potential attacks.

To integrate the ML framework with consensus mechanisms. The developed ML framework will be integrated into the consensus algorithms discussed in this work to enhance the security of consensus mechanisms. This integration will enable real-time monitoring and proactive defense mechanisms against attacks, thereby ensuring the integrity and stability of blockchain networks.

To evaluate the effectiveness of the hybrid approach. The proposed hybrid ML approach can be thoroughly evaluated. The evaluation would be focused on security enhancements achieved through the proposed solution, which leads the system to an adoptive selection of consensus algorithms and intelligent decision-making.

By addressing these challenges, this research paper aims to contribute to blockchain security by proposing novel consensus algorithms and hybrid ML approaches that enhance the overall security of blockchain networks and mitigate the risk of cyber-attacks.

The contribution of the proposed hybrid consensus algorithms is listed below.

This proposed hybridization combines the strength of different consensus mechanisms, mitigates the vulnerabilities, and enhances scalability.

The hybrid consensus algorithm uses energy-efficient mechanisms to reduce environmental impact without compromising security.

This model can adapt to dynamic conditions and confirms robustness.

This proposed method can integrate privacy-enhancing features and protect sensitive information.

Additionally, hybrid models can improve trust by providing robust and resilient consensus mechanisms.

The contributions of the ML techniques are listed below:

ML can greatly enhance threat detection capabilities and improve the trustworthiness of the network by providing a multilayer defense.

It improves the adaptability to emerging threats and acts as an intelligent layer to optimize consensus mechanisms based on network conditions and improve scalability challenges.

ML techniques can adjust the consensus mechanism based on energy availability and consumption patterns.

It also learns continuously from the network behavior and adjusts security measures to meet network dynamics.

Cryptographic techniques and advanced privacy-preserving algorithms can further enhance the confidentiality of transactions and user data.

Decentralized systems can detect and respond to security threats using machine learning to identify patterns and anomalies, reducing the risk of successful attacks.

Resource allocation can be optimized by dynamically assessing threat levels and adjusting security measures accordingly. This ensures efficient allocation of resources to areas with the highest risk.

The rest of the paper is organized as follows: Section " Background study " establishes the preliminaries for the challenges and vulnerabilities in existing consensus mechanisms. A detailed discussion of the previous research on 51% of attacks and their impact is also presented in this section. ML Techniques for Security Enhancement are discussed in Section " Machine learning techniques for security enhancement ". Section " Research methodology and materials " presents the research methodology and materials and its advantages and optimizations achieved through the proposed approach. The experimental implementations and results are discussed in Section " Security enhancements achieved through the proposed solution ". Section " Open issues and challenges of the hybrid consensus approach " presents the proposed hybrid consensus approaches' open challenges and future scope. The conclusions and future work of the paper are discussed in Section " Conclusion ".

Background study

Challenges and vulnerabilities in existing consensus mechanisms.

The progress in blockchain technology has played a crucial role in addressing challenges related to decentralization, security, and consensus formation using various consensus mechanisms. However, these mechanisms have their own set of challenges and vulnerabilities. This section will delve into the concerns and susceptibilities of existing consensus mechanisms.

Proof of work (PoW)

PoW is a consensus algorithm that exhibits several challenges and vulnerabilities. Firstly, PoW requires substantial computational power and energy consumption to solve complex cryptographic problems, resulting in high energy consumption that is environmentally unsustainable 31 . This energy-intensive nature raises concerns about the ecological impact of blockchain networks utilizing PoW. Secondly, scalability becomes a concern as the network grows, as the PoW algorithm needs help to maintain efficient consensus and transaction validation in the face of increasing participant numbers 32 . The computational requirements become increasingly demanding, potentially limiting the scalability potential of PoW-based blockchains.

Additionally, the concentration of mining power in a few large mining pools introduces concerns regarding mining centralization and the potential for collusion 33 . This centralization raises questions about the democratic nature of the blockchain network and the potential for malicious activities by a concentrated mining power 34 . The priority toward long-term sustainability and PoW-based blockchain adoptions are required to address the challenges.

Proof of stake (PoS)

The Proof-of-Stake (PoS) consensus protocols face vulnerabilities and challenges that include the possibility of wealth concentration and must be addressed within the network 35 . In PoS, validators are chosen based on the number of coins they have staked, meaning that those with more enormous stakes are more likely to be selected as validators. This wealth concentration can undermine the network's decentralization and compromise its security 6 . Another issue is the nothing at stake problem, where validators can vote on multiple forks simultaneously without consequences. Validators can create forks and change transaction history, which may lead to double-spending attacks 36 . To mitigate these challenges, solutions such as implementing penalties for malicious behavior, implementing robust governance mechanisms, and ensuring widespread participation can help maintain the security and integrity of the PoS network. Additionally, ongoing research and development efforts are needed to address these vulnerabilities and improve the overall robustness of the PoS consensus mechanism 37 . PoS has energy efficiency advantages, but its vulnerabilities must be acknowledged to ensure blockchain network security.

Delegated proof of stake (DPoS)

The DPoS algorithm is widely adopted for its effectiveness and scalability; however, it poses significant challenges and vulnerabilities. A primary concern is the potential for centralization, as DPoS relies on a small group of trusted delegates responsible for block production and validation 38 . This concentration of power introduces the risk of influential entities engaging in vote buying or collusion, compromising decentralization and fairness. Another area for improvement is low voter participation, leading to a lack of representation and potential governance problems 39 . In order to overcome these challenges, measures should be implemented to promote decentralization and encourage greater voter engagement. This includes preventing manipulation through mechanisms that safeguard the integrity of the election process 40 . Educating token holders about the importance of voting and its impact on network governance can increase participation. Introducing mechanisms for delegate rotation or limiting their terms can prevent long-term centralization 41 . Through continuous research and protocol improvements, DPoS can balance efficiency and decentralization, offering a more inclusive consensus mechanism for blockchain networks by enhancing transparency and promoting active participation 42 .

Practical byzantine fault tolerance (PBFT)

PBFT offers the advantage of withstanding Byzantine faults but also presents challenges and assumptions that must be addressed. Scalability is a limitation of PBFT, as the latency increases with more nodes due to the communication required for consensus 43 . This makes PBFT more suitable for smaller networks or consortium blockchains. The assumption of PBFT regarding the number of faulty nodes is critical, as it assumes that at most one-third of nodes are faulty. When a higher proportion of nodes become malicious or inaccurate, PBFT's ability to maintain consensus can be compromised 44 . However, researchers have dedicated their efforts to strengthening the scalability and resilience of PBFT in order to overcome these challenges 45 . Optimizations like parallelization and batching have been proposed to reduce communication overhead and latency in more extensive networks. Fault-tolerant algorithms and Byzantine fault detection techniques aim to handle situations where the assumed threshold of faulty nodes is exceeded 46 . Hybrid consensus models combining PBFT with Proof of Stake or Proof of Work have also been explored to balance scalability and fault tolerance. Researchers have proposed various methods to overcome these challenges and improve the scalability, durability, and usability of PBFT in various blockchain networks 47 .

Proof of authority (PoA)

PoA relies on a predetermined set of approved validators responsible for validating and adding new blocks to the blockchain. While PoA offers certain benefits, it also poses notable limitations and vulnerabilities. A key concern is the potential for centralization 48 . However, carefully selecting validators and proactive measures to prevent infiltration can ensure continued decentralization and strengthen network security, leading to a robust and resilient system. The concentration of authority in the hands of a small group of validators contradicts the decentralization goal that blockchain technology aims to achieve 49 . Another challenge in PoA is the need for incentives for validators. Unlike Proof of Work and Proof of Stake, PoA does not provide incentives or rewards to validators, as block creation authority is based on reputation or identity rather than a commitment of resources. The absence of incentives will lead to lower participation and highly compromise the network's security 50 . Validators may become less vigilant or refrain from active participation in block validation, leading to a less secure and reliable network. Certain modifications have been proposed for PoA to address these challenges. For instance, implementing a reputation-based system or penalties for misbehavior can mitigate the risk of collusion among validators 51 .

Additionally, offering incentives in the form of transaction fees or token awards can promote active participation and ensure the stability and security of the network. The PoA system offers faster transactions and uses less energy. However, there are concerns about security and trust because it relies on a set of predetermined validators. Additionally, when implementing PoA, it is necessary to consider the balance between centralization and decentralization 50 .

Casper presents a more secure and reliable consensus mechanism, addressing challenges traditional proof-of-stake (PoS) algorithms face. However, a significant challenge in Casper lies in parameter selection. Careful consideration is required to balance security, liveness, and fault tolerance 52 . Improper parameter values can lead to vulnerabilities and compromise the protocol's effectiveness. Conservative parameters may hinder efficient block finalization, while permissive parameters can increase the risk of malicious behavior 53 . Thorough analysis and understanding of network dynamics and trade-offs are necessary to determine appropriate parameter values for Casper. Another critical aspect of Casper is using slashing conditions to penalize validators for malicious behavior 54 . However, defining and enforcing slashing conditions without introducing false positives or negatives is a complex challenge. False positives penalize honest validators wrongly, while false negatives allow malicious validators to escape penalties. Striking the right balance is crucial to prevent unfair penalization and ensure appropriate punishment 55 . Designing robust and accurate slashing conditions requires careful consideration and analysis to minimize false positives and negatives.

Researchers are enhancing Casper to address system challenges through better parameter selection and slashing conditions. This method involves rigorous empirical analysis, simulation studies, and formal verification techniques 56 . It also enables us to confidently determine the most influential parameter choices and optimize the settings for feasible results. Extensive testing and experimentation assess the impact and behavior of slashing conditions in real-world scenarios 57 . Future advancements may involve automated or adaptive parameter selection mechanisms that dynamically adjust based on network characteristics. Similarly, improvements in slashing conditions can be achieved through ML or incorporating external reputation systems. Continued research and development efforts will lead to more robust and practical implementations of Casper.

Previous research on cyber 51% of attacks and their impact

Previous research on 51% of attacks has explored the potential risks and consequences of these attacks in blockchain networks. By gaining control over 50% of the network's mining power, an attacker can manipulate the blockchain's transactions, potentially leading to double spending or other fraudulent activities . It is necessary to assimilate the importance of 51% of attacks for establishing adequate security and authenticity maintenance of the blockchain network.

Attack vector identification

This is a challenging research problem related to 51% of attacks. Several studies and experiments have been conducted to understand the various attack vectors. By analyzing these vectors, researchers can enhance security and improve countermeasures. Mining centralization is one of the prominent attacks. Studies are performed to examine the mining power concentration 58 . This study includes investigating the distribution of mining resources, examining the incentives for mining centralization, and identifying the potential risks associated with such centralization. Network partitioning usually occurs when we split the blockchain network into multiple subnets. These results from various technical issues, intentional attacks, and other network disruptions, which blockchain networks need to address 59 . However, researchers have analyzed the influence of network partitioning on consensus mechanisms and their vulnerabilities. After examining these problems, researchers aim to mitigate the issues connected with network partitioning and ensure blockchain integrity. Rent attacks acquire more computational power and control the hash rate's prominence, enabling attackers to manipulate the transaction. Researchers can examine the influence of this attack on other consensus mechanisms, which emphasize the importance of prevention protocols to avoid the presence 20 . However, researchers can also investigate the chance of group mining to launch the attack. When mining pools or entities collude with each other, they combine the computational power and control the hash rate, as discussed in the rental attack, and manipulate the transaction in the blockchain network 60 . This results in the importance of protection mechanisms to avoid and reduce collusion between the mining entities. Furthermore, researchers also analyzed these attacks to detect and mitigate the related vulnerabilities.

Double spending and transaction reversal

These attacks aim to achieve double-spending and transaction reversal, exploiting vulnerabilities in the transaction verification process. By controlling most of the network's mining power through a 51% attack, an attacker can manipulate the blockchain's transaction history, allowing them to spend the same coins multiple times 61 . Previous research has extensively explored the economic incentives and feasibility of double-spending attacks, considering factors such as attack cost, potential gains, and impact on the network's reputation 62 . Evaluating the economic viability of these attacks helps understand their motivations and enables countermeasures to be developed. Proposed countermeasures include increasing the number of confirmations required for transaction finality, implementing mechanisms for detecting suspicious transactions, and enhancing consensus algorithm security 63 . Advancements in blockchain technology have introduced additional measures to mitigate double-spending risks, such as faster block confirmation times and additional security layers like two-factor authentication and multi-signature transactions 64 . Research on the economic incentives and potential impact of double-spending attacks has led to the developing enhanced security protocols, promoting trust and reliability in cryptocurrency transactions.

Blockchain security and trust

Blockchain security and trust are critical considerations in designing and operating blockchain networks. The occurrence of 51% of attacks represents a significant threat to the security and trustworthiness of these networks. Extensive research has been conducted to assess the impact of such attacks on transaction integrity and overall reliability 65 . Successful attacks will compromise the accuracy and reliability of blockchain technology. Controlling network-mining power will lead attackers to manipulate the transaction history. This will cause double-spending attacks and confirms the earlier transactions 66 . This action will destroy the confidence of the user in the blockchain network. This will make users stumble on network transactions and cause fraudulent or fearful activities. However, the blockchain network always depends on user participation and adopting successful transactions, which causes economic repercussions 67 .

Earlier studies discussed the significant consequences of these attacks and recommended the importance of security measures to protect 68 . In addition, researchers are also focusing on mitigating and detecting 51% of attacks, which improves consensus mechanisms, decentralization governance, and enhancing network resilience 69 . Addressing the security and trust issues requires multi-faceted approaches. This approach involves technical solutions, regulatory measures, governance frameworks, and other industrial standards 70 . Research collaboration with policymakers, industrialists, and stakeholders ensures effective practices and security measures. Educating users and stakeholders will establish trust, comprehensive adoption, and usage.

Countermeasures and prevention

Hybrid consensus protocol combines PoW, PoS, and other mechanisms to improve security and avoid these attacks. These models will leverage the strengths of the approaches and mitigate the weaknesses. For instance, Hybrid PoW and PoS will improve security and reduce the mining centralization problem, which addresses the nothing at stake in PoS 71 . Likewise, hybrid PBFT leads to an increase in fault tolerance and scalability. These models can improve security and avoid 51% of attacks 72 .

Improving mining decentralization is also another preventive measure that has been proposed by researchers 73 . Promoting a distributed network of miners may lower the mining concentration power, which makes it more difficult for single or group entities to control the computational resources 74 . This technique can mitigate the problem of 51% of attacks through power distribution among various participants and confirms diverse and resilient networks. Introducing penalties for adverse behavior is vital to discouraging and avoiding 51% of attacks 75 . Consensus algorithms introduce various mechanisms that reduce the probability of 51% of attacks. However, improving network monitoring and malicious detections are essential to identify potential attack patterns, which trigger timely responses and avoid 51% of attacks. In machine learning, anomaly detection approaches are used to analyze the data and help detect suspicious activities that show the potential 51% of attacks 76 . Active network monitoring leads to the detection of attacks and ensures the security and integrity of the blockchain.

Impact on decentralization and consensus

The 51% of attacks can cause severe implications in the decentralization of the blockchain network. Hence, it should be considered as an immediate concern. Extensive research provides an understanding of the consequences of attacks on network governance, concentration power, and decision-making processes 77 . Decentralization shows that any single or group of entities does not control the network, immutability, resistance, and fostering transparency. However, a perfect 51% of attacks can allow an attacker to control the network and threaten decentralization. This power concentration contradicts decentralization, which introduces vulnerabilities and compromises the blockchain network's trust and integrity 78 . The existing consensus algorithms show the consensus reaching and validating the transactions . When a 51% attack occurs, the attacker can manipulate the consensus process, potentially invalidating transactions or reversing confirmed blocks. This disrupts the integrity of the consensus mechanism and raises concerns about the validity of the entire blockchain 79 . Research on the impact of 51% of attacks emphasizes the criticality of maintaining a decentralized network structure and robust consensus mechanisms. Efforts are directed toward developing countermeasures that promote decentralization and enhance the resilience of consensus protocols against such attacks 80 . Hybrid consensus models, for example, aim to combine multiple consensus algorithms to mitigate the vulnerabilities of individual approaches and achieve a more balanced and secure network. Safeguarding decentralization and consensus mechanisms also involves addressing factors such as governance and decision-making processes. Research explores ways to ensure fair and democratic governance structures where decision-making power is distributed among network participants 81 . Decentralized governance models prevent authority concentration with on-chain voting and transparent protocols.

In the real world, 51% of attacks serve as valuable resources for understanding the methodologies, impact, and responses associated with such attacks. Previous research has analyzed notable incidents, including attacks on Bitcoin Gold, Verge, and Ethereum Classic, to gain insights into the nature and consequences of these attacks 82 . By examining case studies, researchers can delve into attackers' specific techniques to gain majority control over the network's mining power. This analysis helps identify vulnerabilities within the consensus mechanisms and highlight areas where improvements are needed.

In recent years, many case studies have provided insights about the impacts of these attacks in the blockchain community, corresponding responses, and steps to reduce their effects 83 . Understanding the successful 51% of attacks can help to assess the economic losses, potential damage, and disintegration of user trust. This can be crucial for analyzing network participants to represent vulnerabilities, strengthen network security, evolve countermeasures, and provide active security strategies. Previous studies on 51% of attacks can provide the fundamentals to improve security, governance framework, developing mechanisms, monitoring, and detecting systems 84 . By gaining knowledge through these case studies, we can identify vulnerabilities, patterns, and best practices to develop more resilient blockchain networks. It is necessary to understand that the process of 51% of attacks can change frequently, leading to new attack vectors and the emergence of other techniques. Therefore, ongoing research and collaboration are essential to anticipate threats and address vulnerabilities associated with these attacks. The insights from case studies and real-world examples highlight the importance of continuous research efforts and interdisciplinary collaboration among researchers, developers, policymakers, and industry stakeholders. All these factors are combined to enhance security, trustworthiness, and resilience in the face of 51% attacks and other emerging threats to blockchain networks.

Consensus algorithms have several limitations that must be addressed to implement them effectively in real-world applications. Scalability is a significant concern, as many consensus algorithms need help to handle high transaction volumes and large network sizes. Finding efficient solutions to scale while maintaining security and decentralization is crucial for accommodating the demands of real-world applications. Energy efficiency is another limitation, especially in consensus algorithms like Proof of Work (PoW) that consume substantial energy. This not only makes them environmentally unfriendly but also economically unsustainable. Developing energy-efficient consensus algorithms is essential to reduce the carbon footprint associated with blockchain technology and ensure long-term viability.

Specific consensus algorithms introduce centralization risks, such as Proof of Stake (PoS) and Delegated Proof of Stake (DPoS). These algorithms can concentrate control in the hands of validators with more significant stakes or selected delegates, compromising the decentralization and trust that blockchain technology aims to provide. Balancing decentralization and stakeholder influence is crucial to maintaining a robust and inclusive network. Trust assumptions in consensus algorithms pose another limitation. Many algorithms rely on pre-approved validators or trusted authorities, which may not align with blockchain technology's decentralized and trustless nature. This limitation restricts the applicability of consensus algorithms in specific real-world applications that require higher levels of trust and security without relying on centralized entities.

Privacy and confidentiality are challenging within consensus algorithms prioritizing transparency and immutability. Striking a balance between data privacy and transparency is a complex task that consensus algorithms must address to protect sensitive information while maintaining the auditability and transparency required by various applications. Addressing these limitations in consensus algorithms will require continuous research, innovation, and collaboration between stakeholders in the blockchain community. By overcoming these challenges, consensus algorithms can unlock the full potential of blockchain technology in a wide range of real-world applications while ensuring scalability, energy efficiency, decentralization, trust, and privacy.

Machine learning techniques for security enhancement

Overview of machine learning algorithms applicable to blockchain security.

Machine learning algorithms offer a range of techniques that can be applied to enhance blockchain security. These algorithms leverage artificial intelligence and data analysis to detect anomalies, identify patterns, and make predictions, thereby strengthening the resilience of blockchain networks against potential security threats. This section discusses an overview of machine learning algorithms applicable to blockchain security.

Supervised learning algorithms

Supervised learning algorithms, such as Support Vector Machines (SVM) and Random Forests (RF), play a vital role in blockchain security by enabling classification tasks and bolstering the detection and prevention of fraudulent or malicious activities 85 . Support Vector Machines (SVM) stand out as a widely utilized supervised learning algorithm in the context of blockchain security. SVMs excel in binary classification tasks, where transactions must be categorized as legitimate or malicious. By creating a hyperplane that separates the two classes in a high-dimensional feature space, SVMs strive to find the optimal hyperplane that maximizes the separation of data points while minimizing classification errors. SVMs boast a robust theoretical foundation and are particularly effective in scenarios where the data is not linearly separable. Kernel functions manage the high-handling feature spaces more efficiently and achieve linear and non-linear classification tasks .

RF is an ensemble learning algorithm that utilizes multiple decision trees and performs predictions 86 . Each decision tree in this algorithm is trained on the data subset and uses random feature selections. The final predictions can be made by combining the individual tree predictions. RF has the potential to handle high-dimensional data and its robustness. This technique is effective in handling tasks that involve regression and classification. In blockchain networks, RF is highly effective in determining whether the transactions are legitimate or fraudulent through investigating features and patterns related to their security. RF can also detect malicious nodes by analyzing their interactions and behaviors.

Both SVM and RF offer distinct advantages in terms of performance and interpretability. SVMs are acclaimed for their adeptness in handling complex data and finding optimal decision boundaries, while RF excels in managing large and diverse datasets 24 . Both algorithms can deliver accurate and reliable results within blockchain security applications. It is important to note that the efficacy of these supervised learning algorithms relies on the quality and representativeness of the training data. Labeled datasets containing examples of legitimate and malicious transactions or nodes are crucial for effectively training the models 38 . Furthermore, feature engineering plays a critical role in extracting meaningful features from blockchain data, thereby enhancing the performance of these algorithms.

Unsupervised learning algorithms

Unsupervised learning algorithms such as clustering and anomaly detection will aid blockchain security by identifying threats and patterns without labeled training data. Clustering algorithms group similar transactions or network entities in blockchain security. Algorithms such as k-means or DBSCAN partition the data into clusters, each representing a group of similar data points 42 . Clustering algorithms can differentiate between regular and unusual behavior by examining the patterns in each cluster. This helps to identify security threats or suspicious activities in the blockchain network. This capability is precious when specific attacks or anomalies are unknown in advance, as clustering algorithms can unveil unknown patterns or groups within the data.

On the other hand, anomaly detection algorithms focus on identifying outliers or anomalies that significantly deviate from the expected behavior 87 . Algorithms such as Isolation Forest or One-Class SVM construct models of the expected behavior within the data and flag any instances that fall outside this norm. In blockchain security, anomaly detection algorithms can help identify unusual or suspicious transactions, network nodes, or activities that may indicate potential security threats, fraud, or network intrusions 85 . Identifying these irregularities will lead to rapid implementation of security measures, minimize the risk, and protect the blockchain network's dependability. Unsupervised learning can also provide valuable insights into the character and structure of blockchain data. This makes it feasible to detect the potential security risks and irregularities that may not have been previously labeled or identified. These algorithms can provide a comprehensive and proactive approach that enhances the capabilities of blockchain security. An unsupervised learning algorithm requires expert parametric tuning to perform precise and significant outcomes 84 . In cluster algorithm selection, several clusters or thresholds for anomaly detection can significantly influence the algorithm's efficiency. However, data preprocessing and feature engineering are essential in unsupervised learning that prepares the data, which relies on inherent structure and data distribution.

Deep learning algorithms

Deep learning algorithms such as CNNs and RNNs significantly influence blockchain security 88 . CNNs are highly effective in analyzing structured data like transaction graphs. Meanwhile, RNNs excel in analyzing sequential data like transaction histories and brilliant contract execution. These algorithms can effectively identify malicious and fraudulent activities by detecting anomalies and patterns. It can provide advantages such as autonomous learning, accurate anomaly detection, and handling large datasets.

Moreover, deep learning algorithms can continuously improve their performance, allowing them to adapt to new attack patterns and evolving security threats 89 . Nevertheless, it is crucial to consider that deep learning algorithms require substantial amounts of labeled training data and significant computational resources to train and deploy effectively. Model interpretability and explainability can be challenging with deep learning models, given their complex nature and operation as black boxes. Ensuring the privacy and security of sensitive blockchain data during the training process is also a critical consideration 90 . By integrating deep learning algorithms, such as CNNs and RNNs, into blockchain security frameworks, researchers can enhance threat detection and strengthen the trustworthiness and resilience of blockchain networks.

Reinforcement learning

Reinforcement Learning (RL) is a valuable branch of machine learning that enables the training of intelligent agents to make sequential decisions to maximize cumulative rewards 91 . RL algorithms like Q-learning or DQNs can help create intelligent agents to secure blockchain with optimal security policies. Blockchain agents boost security by verifying blocks, selecting consensus methods, and preventing attacks. Through interactions with the blockchain network, RL agents observe the current state, take actions, and receive rewards or penalties based on the outcomes of their activities, thereby learning from their experiences during training 92 . RL agents explore the environment during training and learn through trial and error. The goal is to obtain a policy that increases rewards over time by enhancing security measures and reducing potential threats to blockchain security. RL empowers blockchain networks with adaptive and intelligent decision-making capabilities. RL agents can learn to identify attack patterns, anticipate vulnerabilities, and respond to emerging threats in real-time 93 . This adaptability is particularly valuable in blockchain security's dynamic and evolving landscape, where new attack vectors and vulnerabilities continually occur.

Furthermore, RL algorithms can be combined with other techniques like supervised or unsupervised learning to enhance the learning process. Pre-training RL agents with previous data or labeled examples improves their learning speed and performance in new situations. However, deploying RL algorithms in blockchain security presents challenges 94 . One critical challenge involves defining an appropriate reward structure that accurately reflects the security objectives of the blockchain network. RL agents must balance exploring new security measures and relying on established ones by finding a balance between exploring new actions and exploiting known strategies.

Bayesian networks

This algorithm uses probabilistic reasoning, which helps the system model and analyze the elemental correlations between the variables and blockchain security. It shows uncertain and incomplete information with directed acyclic graphs 95 . The calculations are performed for the event probability based on their observed evidence by quantifying conditional variable dependencies and probabilities. The Bayesian network provides valuable information, such as a blockchain network's security risks and vulnerabilities, and enables stakeholders to act appropriately 96 . By incorporating additional communication and handling missing or incomplete data with probabilistic reasoning, this algorithm is flexible and quickly adapted to changing the real-world scenario.

Generative adversarial networks (GANs)

In a blockchain network, GANs are a vital tool to enhance security in distributed networks. It has two components, a generator, and a discriminator, that create artificial datasets that replicate the real-world data exactly. GANs provide several benefits, such as creating diverse, realistic data for attacking scenarios 97 . This approach can simulate the attackers' actions and improve the network's defense system. However, this model can also enhance the anomaly detection system, improving blockchain security . In order to train the GANs, enormous computation resources and extensive training data sets are required 98 . Furthermore, it requires high-quality and diverse datasets, which influence the reliability of the synthetic data and improve the performance. Even though it has limitations, implementing GANs in blockchain provides real-world opportunities for developing attacking scenarios, testing blockchain security, and performing an effective anomaly detection system.

Privacy-preserving techniques (PPTs)

This technique has two security components, namely Differential Privacy (DP) and Secure Multi-Party Computation (SMPC). These components are necessary for maintaining confidentiality and anonymity in the distributed blockchain network 99 . DP can enable control over the noise, data perturbation, and query response, which avoids data point identification. However, SMPC works on multiple parts that collaborate with ML tasks and computations and improve privacy. Combining these approaches, PPTs can ensure the safety of the applicant's data and compliance with data protection regulations. Furthermore, it can also enable the detection of anomalies and prevent malicious activities 100 . While selecting the appropriate methods, we should consider the requirements such as privacy level, data type, analysis or prediction task, and computational overhead.

Feature extraction and anomaly detection for 51% attack prevention

As a researcher, it is imperative to highlight the significance of feature extraction and anomaly detection in preventing 51% of attacks in blockchain networks. These techniques are pivotal in identifying and flagging abnormal patterns or behaviors that may signify potential attacks, allowing prompt intervention to mitigate associated risks. By delving into feature extraction and anomaly detection details, we can gain deeper insights into their crucial role in preventing 51% of attacks in blockchain networks.

Feature extraction

Analyzing blockchain data requires feature extraction, which is crucial in identifying and extracting meaningful information that captures the essential characteristics of transactions, network behavior, and participant activities 101 . Analysts obtain meaningful insights such as the blockchain network's security, performance, and efficiency, enabling information-based decision-making. Transaction-based features provide insights into individual transactions, encompassing transaction size, frequency, time stamp, input–output ratios, graph properties, and volumes 36 . Furthermore, these functions expedite the anomalies, detecting the malicious transactions and recognizing patterns associated with the malicious behavior.

The behavior, connectivity patterns, centrality, and consensus participation of blockchain are analyzed through network-based features. This analysis helps evaluate the network structure, identifies the significant nodes that influence it, and identifies any abnormal network activity 98 . Participant-based features center on individual participants within the blockchain network, encompassing reputation scores, stake sizes, and consensus participation history. These features assess the trustworthiness of participants, detect potential malicious actors, and evaluate the reliability of network contributors 102 . Smart contract-based features involve the analysis of smart contract code and properties, allowing vulnerability assessments, identification of attack vectors, and monitoring for suspicious behaviors during contract execution 59 .

Feature extraction is the foundation for subsequent analysis and modeling tasks in blockchain analysis. ML algorithms acquire the extracted features and perform statistical analyses. Researchers can gain insights that help to identify the pattern and make formal decisions based on these inputs. The feature selection always depends on the data under examination in blockchain and meets the goal, which helps the researchers better understand the blockchain network.

Anomaly detection

Detecting anomalies is critical, which secures the blockchain and maintains integrity. The process involves the identification of deviations that lead to potential attacks, abnormalities, and fraudulent activities. Irregularities can be identified using graph-based methods, behavioral analysis, ML, and statistical techniques 103 . Statistical approaches utilize statistical techniques to identify abnormalities in transaction patterns, network behavior, or consensus parameters. Clustering techniques group similar transactions or network entities to identify outliers or unusual patterns 98 . Outlier detection methods focus on identifying data points that significantly deviate from the expected distribution. Time-series analysis detects anomalies in temporal patterns or trends, enabling the identification of abnormal behaviors over time.

Machine learning approaches are impressive for anomaly detection in blockchain networks. Unsupervised learning algorithms automatically identify anomalies based on patterns and statistical deviations from normal behavior. These algorithms learn from previous data, detect subtle changes, and adapt to evolving attack patterns. Supervised learning algorithms can also be utilized with labeled data to classify instances as normal or anomalous based on known patterns 23 . Graph-based approaches leverage the structure of the blockchain network to identify anomalies. Graph analysis techniques detect changes in node centrality measures, unexpected or suspicious connections, or alterations in community structures. Analyzing the network graph can detect malicious activity like changes in consensus participation or new connections 36 .

The behavioral analysis focuses on establishing normal behavior profiles based on previous data. Analyzing transaction patterns, network interactions, or consensus participation can establish a baseline of expected behavior 104 . Deviations from these profiles, such as sudden changes in transaction volumes, irregular consensus participation, or unusual network interactions, are flagged as anomalies. Threshold-based approaches involve setting thresholds for specific features or behaviors and monitoring deviations beyond those thresholds. For example, exceeding a certain threshold in transaction volume or encountering a predetermined limit of consensus failures may indicate an anomaly 105 . These approaches are straightforward to implement and provide an early warning system for potential anomalies. By utilizing these approaches, researchers can enhance the security and robustness of blockchain networks by promptly detecting and mitigating anomalies.

Real-time monitoring and alerting

Real-time monitoring and alerting are integral components of a robust anomaly detection system in blockchain networks. The ability to detect anomalies in real-time and promptly respond to potential cyber-attacks is necessary for upholding network security and integrity 90 . Real-time monitoring ensures that any abnormal behavior or suspicious activities are swiftly identified and addressed, thereby minimizing the potential effect of attacks. Continuous analysis of blockchain data through ongoing anomaly detection methods is essential for real-time monitoring. As discussed previously, machine-learning models can be trained to detect anomalies in real-time and deployed to analyze incoming data streams for deviations from normal behavior or expected patterns 84 . The system uses machine-learning algorithms to identify and alert users of potential attacks or malicious activities as soon as they are detected. Automatic alerts are challenging to identify quickly and immediately respond to anomalies. This information can be sent through email, SMS, or mentioned on a dashboard. Integrating the ML model with the defined framework process can lead to real-time monitoring. Developing well-defined responding mechanisms involving active protocols isolating the compromised nodes and enabling additional security measures is necessitated 106 .

Dynamic adaptation

Dynamic techniques lead the blockchain network to adopt 51% of attacks that detect and respond to emerging threats. The continuous learning process in ML models can analyze the attacking pattern, detect new threats, prevent potential attacks, and improve accuracy 107 . This never-ending learning makes the system practical for identifying potential security risks. Feedback loops are critical for effective functioning. It can provide valuable information about false positives and negatives, improving anomaly detection accuracy, refining thresholds, and enhancing overall performance 100 . Dynamic adaption made the adjustments by consensus parameters or security measures and mitigated the potential attacks based on anomaly detection. Network configuration can be modified, limiting the action of malicious transactions or nodes and preventing possible attacks. Feature extraction is a critical factor in enabling dynamic adaptation and minimizing the effects of the attacks 108 . The system can identify the characteristics by extracting valuable information from blockchain transactions. This is more valuable to detect anomalies and identify deviations from the patterns. This technique will be updated based on new attacks and includes the latest indicators and behaviors related to 51% of attacks.

Machine learning-based consensus decision-making

ML-based consensus hybridization is a promising method that enhances blockchain networks' security and trust. ML Techniques such as data analysis and pattern recognition can improve the network mechanisms. Intelligent decision-making is performed by leveraging ML algorithms. By investigating the previous data, ML models can collect the essential information that predicts the effects of various consensus performance parameters. This leads to optimizing the consensus parameters such as block size, difficulty level, and adaption over the change in network and improving scalability. In addition, ML-based consensus decision-making commits to blockchain security by implementing the identification, detection, and mitigation of suspicious attacks. ML models monitor the consensus process, detect anomalies, perform more proactive measures, and serve the network's integrity. Furthermore, this method also attains the potential to recast the consensus approach toward data availability, interpretability, privacy, and governance. This section discusses the different factors and their influence on the ML-based consensus decision-making process.

Consensus parameter optimization

Consensus parametric optimization is required to develop a high-performance, secure, and scalable blockchain network. ML algorithms optimize these parameters using the previous data and real-time network conditions. These models can investigate the previous data, which helps to identify the patterns and trends of the network performances through block size, difficult adjustment, time, and validation rules. ML model can identify the different parameters affecting the network's performance and optimize values that improve the efficiency of blockchain operations. In real-time scenarios, monitoring networking conditions allows ML models to adopt the dynamic consensus parameters, which examines the network metrics and provides necessary modification over the consensus mechanisms. These models provide valid suggestions to increase the block size and decrease block time, enhancing scalability in higher transactions. However, when the network faces any security threats, the proposed model learns from experience and recommends changing the difficult level or other validation rules and adapting consensus parameters to improve security. The proposed model can analyze the previous data and understand how the parametric setting affects the network performance and behavior. This learning model will refine the proposed system, which will result in improving the accuracy and optimizing the performance. Through ML, optimizing consensus parameters can improve network performance and security while also improving blockchain systems' scalability and adaptability. ML models can adjust the consensus parameters as the network changes to fit the new conditions, workloads, and threat levels. Adaptability allows the blockchain network to maintain its efficiency, security, and resilience even when faced with dynamic and changing environments.

Adaptive consensus selection

Blockchain networks need a suitable consensus algorithm that adapts to their dynamic conditions and needs. By analyzing past data and network metrics, ML is vital in choosing the best consensus algorithm. ML models can analyze performance metrics of various consensus mechanisms, such as transaction throughput and confirmation latency. These analyses can identify patterns and trends that highlight the strengths and weaknesses of each algorithm in specific network environments. Real-time monitoring allows the machine learning models to make adaptive decisions regarding consensus algorithm selection. The models constantly analyze network load, security needs, and node capabilities to select the best algorithm for the current situation. For instance, during heavy traffic, the system may choose a consensus algorithm that increases transaction speed. However, if security is the top priority, a Byzantine fault-tolerant algorithm may be selected instead. The ML models learn from previous data and adapt to changing conditions, improving their decision-making accuracy. Adaptive consensus selection brings numerous benefits to blockchain networks. Our system carefully selects the best consensus algorithm for every situation to maximize network efficiency, ensuring top-notch performance and resource utilization. Another important aspect is that it boosts network security by adapting the consensus algorithm according to the current security needs and potential threats. The network can protect itself against attacks and preserve the integrity of the blockchain system by adjusting to the situation.

Intelligent block validation

Ensuring the safety and accuracy of a blockchain network relies heavily on intelligent block validation. ML models can play a crucial role in this process. ML algorithms can help to make informed decisions about block validation by analyzing transaction data, network behavior, and consensus rules. By analyzing validated blocks from the past, these algorithms can detect patterns that signify valid transactions. This helps them develop a thorough understanding of a legitimate transaction. ML models categorize the valid and invalid blocks, improving block validation. This procedure can reduce manual inspection and detect malicious activities, improving efficiency in data transactions and network behavior. ML models can correctly identify potential threats, which leads to learning and evolving from the newly arrived data. This introduces effective adaption, changes the network conditions, and elevates themselves from attacking strategies. Furthermore, intelligent block validation improves security and confirms the valid blocks added to the blockchain.

Fraud detection and prevention

Blockchain networks can utilize the ML model and sophisticated algorithms, which analyze transaction patterns and network behavior and detect and prevent malicious activities effectively. The ML model analyzes previous data and detects unusual patterns and behaviors identified as fraudulent activity precisely and accurately . This makes the ML tools more powerful to maintain security and integrity. The proposed system uses supervised learning approaches and trains on labeled data to accurately classify whether the transactions are valid. This method also considers parameters such as transaction amount, time stamps, network interaction, and user behavior for exact predictions. Unsupervised learning methods can analyze the transaction pattern and network behavior, identify deviations from the normal ones, and finally detect anomalies in the blockchain data. This method detects fraud and unknown practices more effectively and identifies the double-spending or Sybil attacks as fraud attempts .

Furthermore, Natural language processing (NLP) techniques are used to analyze the textual data, feedback, and forum discussions to detect potential vulnerabilities. Keeping ML algorithms up-to-date and learning continuously from the new data will always recognize the patterns and prevent fraudulent activities. Earlier detection can reduce potential damage and enable high security. Hybridization of ML and Blockchain develops the network into more trustworthy and resilient.

Predictive analytics for consensus optimization

Predictive analytics are essential and enhance consensus mechanisms. ML models can analyze previous data and patterns, forecast network conditions, and optimize consensus processes. This model goes through different situations and predicts the network behavior by examining network latency, resource usage, and transaction throughput. This leads to the proposed model's ability to perform proactive, intelligent decision-making, which optimizes the consensus mechanisms. Predictive analysis can offer intelligent parametric optimization and adaptive consensus selection, which is anticipated to mitigate security threats. This allows the proposed system design to perform efficient and adaptable operations of the blockchain network under dynamic workloads and varying environments. Predictive analytics is useful for detecting and preventing fraud. Machine learning models can identify patterns of fraudulent activity, which can then trigger alerts and preventive measures for participants in the network. This proactive approach helps maintain the blockchain network's integrity and trustworthiness.

Research methodology and materials

This section will discuss the proposed architecture's steps, components, and modules. The proposed architecture is developed using the ProximaX blockchain infrastructure platform that combines blockchain with the distributed service layers. This integrates blockchain networks with decentralized storage, database, streaming, and enhanced smart contract services to create an all-in-one user-friendly platform. ProximaX is designed to achieve high scalability, and throughput provides low latency. This platform is available in private, public, and consortium configurations and accommodates additional services without compromising the performance. This unique platform is built on reliable technologies that can be used in all industry segments. Researchers and practitioners can easily design and develop an application on a secure with high availability at a low cost.

The proposed architecture combines ML approaches with consensus to achieve agreement in a blockchain network. Integrating ML algorithms with consensus mechanisms can significantly enhance decision-making in distributed systems. These hybrid algorithms aim to address the limitations of consensus protocols by utilizing ML models. The algorithm gathers important data, extracts meaningful features, and trains ML models with data from the system model. These models are then incorporated into the consensus protocol to optimize decision-making, improve security through anomaly detection, and enable adaptive learning. Figure  2 provides a visual representation of how ML models are integrated with hybrid consensus algorithms in a ProximaX blockchain network to enhance efficiency, scalability, and fault tolerance while ensuring the integrity of the consensus process through the prediction model.

Proposed system using blockchain and machine learning layer.

Initially, the module is developed for collecting and extracting the necessary information from the ProximaX blockchain network. Next, the feature extraction module processes the data to extract meaningful features that capture pertinent information for consensus. The ML training module then uses various algorithms to train models based on the extracted features. The anomaly detection module analyzes incoming data using the trained ML models to identify abnormal behaviors or attacks. If any anomalies or attacks are detected, the consensus decision-making module evaluates them, assesses their impact, and determines appropriate actions to maintain consensus integrity. Finally, the consensus enforcement module ensures that the decisions made by the consensus decision-making module are enforced within the network. This iterative process involves continuous feedback, where data is collected, features are extracted, models are trained, anomalies and attacks are detected, consensus decisions are made, and enforcement is carried out. This enables the consensus architecture to adapt to changing network conditions, identify anomalies or attacks, and maintain consensus integrity based on intelligent ML decision-making. Figure  3 shows the design flow of the proposed work. The detailed discussion over the methodology involved in this research is discussed below through step-by-step analysis.

Design flow diagram for the proposed work.

Review and identify the attack scenarios : The first step is reviewing existing consensus mechanisms and analyzing their limitations in required applications. Additionally, it identifies the different attacks that can occur in an application using blockchain. This information can be used to create a set of labeled data that can be used to train the machine learning algorithms.

Choose a consensus algorithm : The second step is to choose an appropriate consensus algorithm that meets the specific requirements of the blockchain-based system. Here, the ProximaX-based blockchain environment is considered. Hybrid consensus algorithms, which combine elements of different consensus algorithms, can provide robustness and security to the blockchain system. This can help prevent cyber-attacks, such as 51% of attacks and double-spending attacks. This algorithm will ensure the integrity and reliability of the blockchain. Some of the hybridizations of consensus algorithms discussed in this research work are listed below.

Delegated proof of stake work (DPoSW) : DPoS validates blocks in the blockchain while PoW creates them, making it harder for attackers to manipulate the network. It uses a limited number of elected validators, enabling faster block confirmation times and higher transaction throughput than PoW. However, DPoS comes with the risk of centralization and relies on trust in elected representatives, which can compromise decentralization. These factors should be considered when evaluating consensus algorithms.

Proof of stake and work (PoSW) : Blockchain uses PoS and PoW to enhance security and efficiency. PoW increases security against attacks, while PoS allows for better energy efficiency and scalability. However, PoS can lead to power concentration, requiring mechanisms to address the "nothing at stake" problem. Balancing decentralization and efficiency involves trade-offs. PoS creates governance complexity and requires careful management for transparency and inclusivity. Long-term security is a consideration as reliance on PoS increases and PoW decreases.

Proof of CASBFT : Casper-PBFT is a hybrid consensus algorithm that combines Proof of Stake (PoS) and Practical Byzantine Fault Tolerance (PBFT) to improve network security and transaction speed. It offers strong consistency, rapid transaction finality, and scalability advantages. However, careful selection and governance of validators are necessary to avoid centralization risks, and the initial stake distribution may lead to power imbalances. Proper design, testing, and maintenance are necessary due to increased complexity. Adoption should be based on network requirements and associated trade-off management.

Delegated byzantine proof of stake (DBPoS) : A secure and scalable system is achieved through a DPoS-PBFT hybrid algorithm. DPoS elects delegates for faster block confirmation and increased transaction throughput, while PBFT enhances resilience against failures. However, DPoS's small delegate set may lead to centralization and collusion risks, requiring proper governance and transparency. Careful design and monitoring are necessary to balance decentralization and efficiency. DPoS may affect decentralization compared to PBFT.

Choose machine-learning algorithms : Select ML algorithms, such as supervised, unsupervised, or rule-based learning, suitable for detecting and responding to attacks.

Steps involved in ML : ML approaches can be used to help prevent attacks on blockchain-based applications by:

Anomaly detection : ML algorithms can identify and flag unusual network behavior, allowing network participants to quickly detect and respond to potential attacks.

Prediction modeling : Predictive models can be trained to identify the likelihood of an attack based on previous data, allowing network participants to take preventative measures proactively.

Clustering and classification : Clustering and classification algorithms can be used to identify and categorize different attacks, making them easier to understand.

Network traffic analysis : Machine-learning algorithms can analyze network traffic and identify patterns that indicate potential attacks, allowing network participants to respond quickly.

Blockchain data analysis : Machine-learning algorithms can be used to analyze the data stored on the blockchain, such as transaction history and network activity, to identify potential attacks.

Fraud detection : ML algorithms can detect fraudulent transactions, such as double-spending or fake transactions, and prevent them from being added to the blockchain.

Risk assessment : ML algorithms can be used to assess the risk posed by different nodes on the network and prioritize security measures based on the risk level.

Reinforcement learning : ML algorithms can learn from network interactions and optimize the security measures to respond to potential attacks.

Data collection : Collect data from the required system to train the machine learning algorithms. This data should include normal and abnormal behavior patterns.

Train machine-learning models : Collect and label data from the system to train machine-learning algorithms, such as supervised, unsupervised, or rule-based learning, to detect and respond to cyber-attacks. This will help the algorithms identify abnormal behavior patterns that indicate an attack.

Integrate the consensus algorithm and machine learning models : Integrate the consensus algorithm and the machine learning models into the blockchain-based system such that the machine learning algorithms can identify and trigger a response mechanism through the consensus algorithm.

Implementing ML in the blockchain : Responsible machine learning (ML) involves developing, deploying, and using ML models that are ethical and accountable. This includes fairness, transparency, interpretability, and privacy. To achieve this, avoid biases and discrimination in data collection and model training, document decisions and assumptions, interpret predictions, and implement privacy safeguards. The ML life cycle includes formulation, acquisition, development, testing, deployment, and ongoing monitoring and maintenance. Responsible ML ensures the trustworthy and ethical use of ML technologies.

Performance metrics : To evaluate a machine-learning model, use metrics like Confusion Matrix, Accuracy, Precision, F1 Score, R-squared, ROC Curve, Area under ROC Curve, and Goodness of Fit. Analyzing these metrics helps identify areas for improvement and determine if the model suits the production environment.

Monitor the system : Continuously monitor the proposed system application and update the machine learning algorithms and consensus algorithm to adapt to evolving attack patterns.

Respond to an attack : If an attack is detected, the machine learning algorithms will trigger a response mechanism that is determined by the consensus algorithm. The response mechanism may include limiting the attacker's access, rolling back the blockchain, or triggering a secure emergency shutdown.

Test and evaluate : Test the hybrid system in a controlled environment to assess its effectiveness in detecting and responding to cyber-attacks.

Deploy : Once the system has been tested and evaluated, deploy the hybrid consensus algorithm and machine learning approach in the real-world environment.

Monitor and update : Continuously monitor the system for performance and security and update the ML and consensus algorithms to ensure their effectiveness against evolving attack patterns. However, implementing a hybrid consensus algorithm with an ML approach requires choosing an appropriate consensus algorithm, training ML models, integrating the consensus algorithm and ML models, testing and evaluating the system, and deploying the system.

Experimental results and discussion

Figure  4 shows the experimental diagram for this research work. In an IoT environment, obtaining vital information by deploying sensors that can detect various parameters from real-world scenarios is difficult. Establishing a robust system to collect this data in real-time or at regular intervals is paramount. Additionally, it is necessary to ensure the structured and balanced data and keep track of the time it was collected. Advanced techniques can be used to obtain valuable insights from the data. When training machine learning models, selecting and combining the most relevant features with labeled data is essential to building a training dataset.

Experimental test setup with intelligence analysis.

Furthermore, correct labeling of the target variable for machine learning is necessary. When setting up a Blockchain network, it is essential to choose a proposed consensus algorithm that fits the network's needs, such as Delegated Proof of Stake Work (DPoSW), Proof of Stake and Work (PoSW), Proof of CASBFT (PoCASBFT) Delegated Byzantine Proof of Stake (DBPoS). This algorithm determines how nodes agree on transactions and adds new blocks to the blockchain. Collecting data from the network is required to detect potential threats. Information like transactions, block details, timestamps, and network states help us stay keen and secure. Relevant features related to transaction volume, block size, transaction fees, or other parameters should be extracted to identify potential attacks. Feature extraction techniques can analyze transaction patterns, mining activities, block propagation delays, or network connectivity measures. Anomaly detection algorithms, such as statistical methods, machine learning techniques, or graph-based analysis, can be used to identify abnormal patterns or behaviors. If an anomaly is detected, it should be labeled as a specific type of attack, such as double-spending, selfish mining, Sybil attacks, or 51% attacks. Ensuring equal representation of typical and attack instances in the training dataset is crucial to boosting the model's performance. This approach is necessary to prevent bias and establish the proposed model to achieve the user's requirement.

After preparing the model, it can be implemented into an IoT environment to analyze future data or make predictions in real-time. In order to make decisions fast and accurately, it is necessary to integrate the model coherently with the current IoT infrastructure . Meanwhile, deploying the trained model in the blockchain network can help to detect and monitor potential attacks in real time. Integrating the model into the blockchain infrastructure enables proactive defense mechanisms against malicious activities. After preparing the model, it can be implemented into an IoT environment to analyze future data or make predictions in real-time. Integrating the model with the existing IoT infrastructure is necessitated to ensure efficient and effective decision-making. To ensure strong security measures, continuously monitor and improve the model by gathering new data and adapting to emerging attack patterns.

Improving consensus algorithms' performance depends on factors like block confirmation time, transaction throughput, energy efficiency, latency, scalability, and fault tolerance. In this work, we can enhance the algorithm's performance by fine-tuning its parameters using optimization algorithms, grid search, or random search techniques. They can then evaluate the optimal parameter values using metrics that reflect the desired performance factors. By comparing different parameter configurations, we can choose the best values that maximize the desired performance factors. By implementing this methodology, we anticipate a more promising and enhanced efficacy within blockchain technology. In blockchain networks, various changes occur, such as network size fluctuations, workload, latency, or adversarial activities. An adaptive consensus algorithm that can adjust its behavior as needed is required to achieve optimal performance. Our work involves developing rule-based mechanisms that dictate the circumstances in which the algorithm should adjust. These guidelines can be determined by various factors such as network parameters, performance metrics, or system-level thresholds. By adhering to these conditions, the algorithm can adapt to changing network conditions by modifying its mode or parameters. Machine learning or reinforcement learning techniques can be leveraged for intelligent adaptive selection. By training models on previous data and network conditions, algorithms can autonomously adapt based on learned patterns or reinforcement learning rewards and penalties. It is necessary to consider the preferences and needs of the users during the adaptive selection procedure. Consensus algorithms are crafted by integrating user-specified criteria that align with the unique requirements of the blockchain application, including but not limited to scalability, security, and energy efficiency.

Understanding the needs of stakeholders, identifying the priorities, and considering specific applications are essential for developing an adoptable ML-based blockchain network. Integrating Consensus Algorithms with ML approaches attains the potential to optimize the network performance and effectively respond to the user's requirements. Stakeholders may possess transaction throughput, privacy, decentralization, energy efficiency, or consensus speed-based priorities. Analyzing these requirements will introduce intelligent selection, which selects hybrid consensus algorithms such as Delegated Proof of Stake Work (DPoSW), Proof of Stake and Work (PoSW), Proof of CASBFT (PoCASBFT), and Delegated Byzantine Proof of Stake (DBPoS). The proposed algorithms have their strengths and trade-offs in security, scalability, efficiency, and decentralization. However, these performance factors are determined by finding the required algorithms based on the user's requirements.

Optimization of the consensus algorithm requires the adjustment of the network parameters based on the user requirements. This can be achieved through several optimization techniques. Here, an ML-based blockchain network has been developed to monitor the network's performance and allows the system to make an adaptive decision based on the feedback from the real-world scenario. Continuous monitoring, evaluation, model refinements with periodic assessments, experimentation based on stakeholders' responses, and user feedback are highly required to implement successful models. Valuable insights are gained by performing the investigation using ProximaX, and results are expected to overcome the different types of attacks in the proposed work. The expected results are listed below.

Real-time attack detection and response,

Maintenance of blockchain integrity,

The smooth operation of a proposed application,

Adaptability to evolving threats,

Improved security,

Minimized attack damage,

Efficient defense mechanism,

Increased trust in the system, adoption, and better performance.

The prime objective of this research work is to identify and respond to the attacks in real-time scenarios immediately. This can be achieved by implementing the ML model to determine the paranormal behavioral patterns that act as a threat. This improves the security and resilience of the blockchain network and minimizes the damage. Maintaining the integrity of the blockchain network is another important objective. Consensus algorithms can validate the transaction and ensure the reliability and stability of the Blockchain network. By combining these Consensus algorithms with a hybrid approach, the proposed system attains higher transactional data integrity and lowers unauthorized activities. ML model and Hybrid consensus mechanisms are combined to ensure convenient operations. These results in highly comprehensive defense mechanisms. This approach can reduce the downtime risk and provide effective operations. However, this approach is adaptable to an evolving new threat and maintains reliable and efficient operations in a specific application-oriented system. The proposed ML models can be kept updated and customized to encounter emerging attacks, which performs the guaranteed defense mechanism for real-world environments.

This approach can proactively identify, detect, respond to, and mitigate the threats. However, it can also reduce the system vulnerability and damages caused by the threats. Expeditious measures are performed to protect the stability and functionality of the network, ensuring better system operations. The proposed hybrid approaches can efficiently protect the blockchain-based system against attacks. The proposed method adapts and updates continuously, which performs fine-tuning to understand emerging threats. This also enhances trust and encourages adaptability in intelligent grid sectors. These features can increase the stakeholder's confidence and improve system performance, resulting in stability, security, and efficient defense against attacks.

Security enhancements achieved through the proposed solution

By combining ML techniques with consensus protocols, the blockchain network performs anomaly detection, adaptive decision-making, and detection of malicious activities. These techniques can detect and mitigate various cyber-attacks by continuous network monitoring and analyzing real-time datasets. However, this hybrid approach provides robust security enhancements that enhance decision-making. This section discusses the various security enhancements achieved through the proposed methodology.

Increased attack resistance : Incorporating ML techniques and hybrid consensus protocols in blockchain networks increases the attack's resistance. This identifies and prevents attacks such as 51%of attacks, Sybil, and double spending attacks. ML models can detect potential attacks by collecting transaction details, network connectivity, and participant behavior. However, the proposed method ensures network security and integrity. This ML model analyzes the network behavior and transaction patterns using previous data obtained from the network, which helps to detect malicious activities and perform precautionary measures. This can prevent almost half of the attacks, detect malicious mining behaviors, and immediately send alert information to prevent attacks. These proactive techniques will minimize the negative impacts of the blockchain network and maintain high security, which is appreciable for the stakeholders and network participants.

Furthermore, the ML model can prevent Sybil and double spending attacks on blockchain networks. These models can use the previous network information and detect irregularities, flags, and fraudulent activities. Combining ML algorithms with a hybrid consensus process will improve the network's ability, promoting confidence in the blockchain network and protecting the stakeholder's assets and transaction details.

Dynamic threat detection : ML algorithms play a crucial role in detecting new threats and identifying the attacking patterns in the blockchain network. The proposed ML model can identify the unusual behaviors and patterns that represent potential attacks by analyzing the datasets from the existing networks. In addition, continuous learning in ML techniques leads to the immediate adaption of new approaches used by malicious attackers, which develop the network into highly secure. These algorithms can detect attacks, including DDoS, Network intrusions, data breaches, and malware propagation. ML model can perform dynamic threat detection, which enhances blockchain security and promotes trust among the network participants and stakeholders.

Anomaly detection and prevention : Effective anomaly detection and introducing preventive measures lead to maintaining the high security and integrity of blockchain networks. ML models are irreplaceable tools that detect mischievous transaction behavior, network connectivity, and other participant activities. It can perform proactive detection and prevent suspicious activities. Analyzing transaction records, network logs, participant interactions, and other system parameters leads the ML model to learn the standard patterns and behaviors exhibited within the network. However, these models establish the expected baseline and detect the deviations from these patterns that indicate fraudulent activities. ML techniques such as clustering, outlier detection, time-series analysis, and graph-based techniques are used to identify potential attacks. Integrating these techniques with blockchain networks can enhance proactive identification, which detects malicious activities in real time.

Furthermore, anomaly detection can deter potential attackers and discourage fraudulent activities . This can improve the confidence level of the stakeholders and participants and promote the adoption of blockchain technology. The proposed ML model can integrate the network decision-making process. For instance, detecting anomalies leads the network to trigger additional verification automatically and employs priority-based security over efficiency. This adaptive response makes the system resilient, secure, and trustworthy to face potential threats.

Adaptive decision-making : ML models facilitate the blockchain networks based on previous data in real time and adapt the decision-making process. This systematic approach can improve performance and optimize security by investigating network congestion and resource availability. We can analyze the metrics and adjust the consensus parameters model with ML approaches. This leads to block time and size variations that improve the transaction throughput and network performance in congested networks. Similarly, for detecting malicious behavior, the proposed model suggests actions such as improving validation regulations and tweaking the consensus protocols. These models can be trained based on previous data, which performs the adaptive choices to enhance security. It can identify patterns or trends that affect network performance and security by analyzing experiences. This knowledge allows the models to make informed decisions and adjust the consensus parameters accordingly. Adaptive decision-making is essential for blockchain networks because it enables quick response to changes by adjusting real-time consensus parameters. It also enhances security by detecting and responding to emerging threats using ML models. This approach optimizes resource allocation, improving efficiency, scalability, and network participation. Continuously adapting decision-making strategies ensures the network remains responsive to emerging challenges and continually improves performance, security, and resilience.

Robust network monitoring : The safety and stability of blockchain networks heavily rely on dependable monitoring. Hybrid consensus protocols use ML to continuously monitor and analyze the network, detecting security threats, attacks, and irregular behaviors. ML algorithms are particularly effective at identifying patterns and anomalies in large amounts of data, such as traffic and transaction data. Monitoring the blockchain network in real-world scenarios provides earlier warnings and other appropriate trigger responses. Robust monitoring can detect critical behavior, potential attacks such as DDoS, Sybil, tampering with the data transactions records, and other security breaches. This also uncovers the security vulnerability and pattern with exploitable loopholes, mitigates the attacks, and maintains the network integrity. ML models address these issues proactively and improve the system's ability to fight security threats and enhance overall network security. Furthermore, these models can identify recurring patterns and adjust their response strategies to improve outcomes in strengthening high adaptability. By using federated learning and homomorphic encryption, consensus mechanisms can detect anomalies and security breaches, improving the privacy and security of the entire blockchain network.

Scalability, decentralization, and energy efficiency considerations

The hybridization of consensus algorithms and ML models addresses challenges such as scalability, decentralization, and energy efficiency in distributed systems. This approach can disseminate the workload across various machines, perform effective data processing, and execute parallel algorithms. The decentralization process can be achieved using P2P networks, which improves reliability and prevents failures. To ensure data availability and integrity, redundancy and replication techniques like sharding or erasure coding can be used, even in node failures. Energy efficiency has become a critical concern as the demand for sustainable computing solutions grows. Consensus algorithms and ML models should be hybridized while considering energy efficiency to minimize environmental impact and operating costs. Use specialized hardware or cloud services to save energy when performing resource-intensive tasks like machine learning or complex consensus computations. These specialized infrastructures are designed to maximize computational efficiency, reducing energy requirements. Developing consensus algorithms with energy efficiency in mind can also minimize computational and communication overhead. PoS consumes less energy than PoW algorithms, relying on stakeholder voting rather than resource-intensive mining. ML models can be compressed or quantized using techniques like pruning, quantization, and knowledge distillation to reduce their size and energy consumption without sacrificing accuracy. Hybridizing consensus algorithms and machine learning models can address distributed systems' scalability, decentralization, and energy efficiency challenges. Considering these factors, we can develop sustainable solutions that leverage the best of both worlds.

Advantages and optimizations achieved through the proposed approach

Hybrid consensus algorithms and ML approaches are widely used to overcome various attacks in blockchain technology. The advantages and optimizations achieved through the proposed approaches are listed below. The key insights and SWOT analysis of the proposed research work are shown in Fig.  5

SWOT analysis of the hybrid approach.

Improved security

Advanced security is necessary for safe and reliable blockchain network systems. Practical ML techniques with hybrid consensus create a secured network with high integrity and reliability, preventing potential attacks and negative impacts. This model can establish a trust-based framework that ensures the network nodes perform valid transactions and avoid unauthorized attackers. Furthermore, some attacks require other measures that represent abnormal behavior. ML techniques can investigate substantial data volumes, identify patterns, and detect malicious attacks in real time. Here, continuous monitoring of energy transactions, communication patterns, and other system parameters is performed by integrating ML techniques. The system can utilize the previous data and effectively identify the unusual activity of ongoing attacks. Combining consensus and ML algorithms can create a robust defense mechanism and prevent cyber-attacks. Integrating ML techniques leads the system to detect and prevent the attack before any damage. These techniques can adapt and improvise through continuous learning from new data and evolving attack techniques. This adaptability can improve the system's ability to detect threats and respond effectively to unknown attacks, strengthening security.

Real-time detection and response

Real-time detection over frequent responses is necessary to ensure the security and stability of the proposed systems. ML techniques are generally trained on the previous data and establish the baseline to understand normal behavior, energy consumption, communication patterns, system metrics, and other relevant data. By identifying the anomalies, ML techniques can provide the response mechanisms to isolate the affected nodes, reroute the energy transactions, and activate the alerts and backup systems for further investigation. For example, The proposed system with microgrid application can provide services to hospitals, emergency services, and remote community sectors, requiring real-time detection and responses and ensuring uninterrupted operations and reliability. ML techniques can identify the attacks in time, respond quickly to process the data and ensure the decision is based on learned patterns. This system can ensure security, detect real-time threats, and reduce the vulnerability window that causes minimum damage by utilizing ML techniques.

Evolving security

Evolving threats can be avoided using hybrid consensus and ML algorithms. Consensus algorithms can secure the blockchain, and ML techniques analyze various data sources and detect new attacks, adapting defense systems to ensure security. Data analysis tools can monitor parameters such as network traffic, transaction patterns, user behavior, and other system logs. These systems can analyze new data, improve their ability to detect suspicious activities and perform immediate alerts for further investigation. ML algorithms can perform collaborative acts and rely on information sharing for evolving defense mechanisms. Integrating consensus and ML algorithms can strengthen the synergistic effect and be updated regularly through ongoing research, collaboration, and information sharing within the blockchain community. This system can stay connected to the latest attack techniques, which helps to mitigate emerging threats and ensures long-term security.

Efficient defense mechanism

An effective defense system can be developed using consensus and ML algorithms that detect and respond automatically. In microgrid applications, consensus algorithms can provide authorized transactions that ensure integrity and reliability. ML algorithms can analyze the data, identify patterns, learn from previous attacks, and predict the detection. These ML models will monitor the microgrid system, identify network behavior, and analyze the transaction patterns. This leads to the proposed algorithm detecting potential attacks, identifying the deviations, evolving the triggering process, and mitigating their impacts on microgrid security. Combining hybrid consensus algorithms and ML techniques creates an effective defense mechanism that performs an immediate detection and response to any attacks on the microgrid. Through continuous learning, the ML model develops into an automation process that reduces the administrator's workload, enhances the efficiency of the defense mechanism, and enables real-time detection and response to attacks. These processes can be optimized through iterative learning approaches in ML techniques. Through the application of these methodologies, the proposed system has the potential to enhance the resilience and efficiency of microgrid systems . These algorithms can automate processes and continuously learn, which reduces the workload on system administrators and promotes a proactive approach to security. Ultimately, this helps to defend against attacks while ensuring the uninterrupted operation of the system.

Increased trust in the system

Trust and security are crucial in microgrid systems, and blockchain-based microgrid systems offer a solution that goes beyond. Combining hybrid consensus algorithms and ML approaches can enhance trust in the system's design and improve its adaptability. Using hybrid consensus algorithms and machine learning techniques creates a robust defense system to protect microgrids against potential attacks. This way, it establishes trust by verifying transaction validity and preventing malicious activities through consensus among network participants while also allowing for continuous learning from new data and emerging attack patterns through ML algorithms. Adaptability is of utmost importance in maintaining the effectiveness of defense mechanisms in microgrid systems, especially against constantly evolving and sophisticated strategies. Through ML algorithms, this hybridization can detect anomalies, predict real-time attack patterns, and adjust the defense mechanism accordingly. As a result, this instills increased trust in the microgrid system, enhances its resilience, and reduces its vulnerability to evolving threats.

Open issues and challenges of the hybrid consensus approach

Open issues of the proposed research approach.

Integrating ML, deep learning, and RL with blockchain protocols can improve security, performance, and decision-making capabilities. However, it also presents open issues and challenges that researchers and practitioners must consider carefully. In order to discover the benefits of these technologies in blockchain networks, it is essential to understand and address these challenges. This research analyzes the challenges of integrating intelligent learning algorithms with consensus protocols in blockchain networks. However, it also explores the factors researchers and practitioners should consider to ensure blockchain networks' effective and efficient operation. This analysis identifies critical limitations such as machine-learning algorithms' computational complexity and resource requirements, the need for labeled training data in decentralized and pseudonymous blockchain networks, and adaptive learning approaches in dynamic blockchain networks. Investigations are performed to understand the significance of security and privacy in integrating ML techniques, which includes the vulnerability of models, malicious attacks, and other challenges that ensure interpretability in decision-making processes. However, it is observed that valuable insights into the difficulties of Implementing ML-based consensus algorithms in blockchain. Our findings can also provide future research and development efforts to address the practical concerns and overcome these limitations.

Incorporating hybrid consensus algorithms with ML models will be challenging and complex. To achieve this, we need a strong understanding of consensus protocol, ML algorithms, and sufficient computational resources for training and executing the models. Coordinating the computational resources with consensus protocol and integrating the ML approaches into the hybridization are significant challenges in real-world scenarios.

Data availability and quality

Accurate training of ML models requires diverse and high-quality data. Nevertheless, this can be challenging in blockchain networks, where data privacy and confidentiality are crucial. Furthermore, obtaining labeled data for supervised learning requires specialized expertise and extensive manual work that can be time-consuming and expensive.

Model robustness and generalization

For developing successful hybrid consensus algorithms, the ML model must possess resilience and the capability to perform effective generalization in untested data. Inadequate adaptation to new attack patterns or overfitting can compromise the effectiveness of the models, leading to the identification of false positives or false negatives, which can ultimately weaken the entire system.

Interpretability and explainability

Understanding the reasoning behind ML models can be difficult, as they often function like mysterious black boxes. However, when it comes to consensus protocols, transparency and accountability are crucial, which means that methods must be employed to clarify the choices made by these models. This task can be incredibly daunting when dealing with complex models such as deep learning architectures.

Malicious attacks

It is challenging to be aware of potential threats to the machine learning models, such as false training data, harmful adversarial examples, and weaknesses in the learning process. These attacks can negatively affect the reliability and effectiveness of the hybrid consensus approach. Therefore, it is essential to implement robust defenses to minimize their impact.

Computational overhead

Incorporating ML models into the consensus process could increase the amount of computational work needed. Developing and implementing intricate machine learning models may demand substantial computational resources, which could affect the scalability and effectiveness of the blockchain network.

Ethical considerations

It is imperative to consider ethical implications such as privacy, fairness, and bias when utilizing ML models in consensus protocols. We must take precautions to prevent infringement on user privacy or discriminatory behavior towards certain participants due to biased training data or decision-making processes. Addressing these issues will lead to more effective and equitable use of ML models.

Regulatory and legal considerations

Applying ML models in consensus protocols may raise regulatory and legal challenges, especially in industries with strict data protection and compliance requirements. Compliance with data privacy regulations and ensuring the lawful use of data becomes crucial when integrating ML into consensus mechanisms. Addressing these limitations and challenges requires careful consideration of the specific use case, continuous research and development, and collaboration between domain experts in consensus protocols, machine learning, and cybersecurity. By addressing these challenges, the hybrid consensus approach with ML models can unlock its full potential in enhancing blockchain networks' security, scalability, and adaptability.

Future research directions

The future scope of this research leads to adaptive hybrid models, which focus on developing new consensus mechanisms and ML techniques. These combinations can dynamically adapt to networking situations. It also includes switching between other ML models or consensus mechanisms and responding to the security threats that evolve based on the scenario.

Privacy-preserving ML techniques are essential for advancing research. These hybrid models effectively integrate homomorphic encryption, federated learning, and differential privacy. This ensures that the user's sensitive information remains secure and confidential while performing the learning process.

Future research should focus on self-learning systems, which have the potential to hybrid consensus mechanisms. These systems can adapt autonomously to new threats and optimize the system parameters per the network conditions' real-time feedback. However, energy efficiency is the paramount factor in the blockchain network. Research efforts should focus on developing effective energy consumption-based consensus mechanisms and ML models.

The security and the consensus mechanism's performance can be ensured by establishing security standards and benchmarks for the evaluation process. This facilitates comparing various approaches and contributes to developing new effective models. Cross-collaboration across experts in consensus algorithms, machine learning, cryptography, and cyber security leads to novel methods and insights to overcome real-world problems.

In this research, we have thoroughly analyzed the hybrid consensus algorithm with ML techniques. The challenges and vulnerabilities that exist in a proposed system are concluded using a ProximaX-based decentralized network platform. The findings from this research paper have emphasized the need for effective preventive measures to combat the detrimental impact of cyber-attacks. The proposed research uses hybridized consensus approaches to enhance network security, scalability, and resilience. The proposed ML-based hybrid consensus algorithm mitigates the challenges and vulnerabilities in decentralized public networks. The ML model implemented in this research elevates threat detection, optimizes consensus mechanisms, and ensures the confidentiality of transactions and user data. The decentralized nature of ML-driven security performs proactive attack identification, feature extraction, and anomaly detection, enhancing the consensus protocol's security and reducing cyber-attacks. The proposed model understands the consensus mechanism and retrieves the real-time data and network state, which improves network resilience and decision-making accuracy in the dynamic field of cyber security.

Furthermore, this paper also discusses the implementation challenges of the proposed consensus approaches and their adaptability to a real-world scenario. Further investigation and refinements are required to solve the complexity of the ML model with scalability, resource requirements, computational overhead, and susceptibility to achieve the effectiveness of security and trustworthiness. The future scope of this research leads to the development of an adaptive hybrid model that utilizes novel consensus mechanisms and ML techniques. Privacy-preserving ML techniques and generative learning can autonomously adapt to new threats and optimize the system parameters based on the real-world environment.

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Oyinloye, D. P., Sen Teh, J., Jamil, N. & Alawida, M. Blockchain consensus: An overview of alternative protocols. Symmetry 13 (8), 1363 (2021).

Article   ADS   Google Scholar  

Wang, B., Li, Z. & Li, H. Hybrid consensus algorithm based on modified proof-of-probability and DPoS. Futur. Internet 12 (8), 1–16 (2020).

Article   Google Scholar  

Shafay, M. et al. Blockchain for deep learning: Review and open challenges. Cluster Comput. 26 (1), 197–221 (2023).

Article   PubMed   Google Scholar  

Khobragade, P. & Turuk, A. K. Blockchain consensus algorithms: A survey. Lect. Notes Netw. Syst. 595 , 198–210 (2023).

Bachani, V. & Bhattacharjya, A. Preferential delegated proof of stake (PDPoS)—modified DPoS with two layers towards scalability and higher TPS. Symmetry 15 (1), 4 (2023).

Wu, Y., Song, P. & Wang, F. Hybrid consensus algorithm optimization: A mathematical method based on POS and PBFT and its application in blockchain. Math. Probl. Eng. https://doi.org/10.1155/2020/7270624 (2020).

Sakhnini, J., Karimipour, H. & Dehghantanha, A. smart grid cyber attacks detection using supervised learning and heuristic feature selection. in Proceeding of 2019 7th International Conference on Smart Energy Grid Engineering SEGE 2019 , pp. 108–112, (2019).

Mololoth, V. K., Saguna, S. & Åhlund, C. Blockchain and machine learning for future smart grids: A review. Energies 16 (1), 528 (2023).

Ortega-fernandez, I. & Liberati, F. Smart grid using reinforcement learning. 1–15 (2023).

Sahani, N., Zhu, R., Cho, J. H. & Liu, C. C. Machine learning-based intrusion detection for smart grid computing: A survey. ACM Trans. Cyber-Phys. Syst. 7 (2), 1–31 (2023).

Kiayias, A., Russell, A., David, B. & Oliynykov, R. Ouroboros: A provably secure proof-of-stake blockchain protocol , LNCS. 10401, (2017).

Lashkari, B. & Musilek, P. A comprehensive review of blockchain consensus mechanisms. IEEE Access 9 , 43620–43652 (2021).

Xiao, Y., Zhang, N., Lou, W. & Hou, Y. T. A survey of distributed consensus protocols for blockchain networks. IEEE Commun. Surv. Tutorials 22 (2), 1432–1465 (2020).

Andoni, M. et al. Blockchain technology in the energy sector: A systematic review of challenges and opportunities. Renew. Sustain. Energy Rev. 100 , 143–174 (2019).

Wang, W. et al. A survey on consensus mechanisms and mining strategy management in blockchain networks. IEEE Access 7 , 22328–22370 (2019).

Yang, F. et al. Delegated proof of stake with downgrade: A secure and efficient blockchain consensus algorithm with downgrade mechanism. IEEE Access 7 , 118541–118555 (2019).

Shah, A. A., Malik, H. A. M., Muhammad, A. H., Alourani, A. & Butt, Z. A. Deep learning ensemble 2D CNN approach towards the detection of lung cancer. Sci. Rep. 13 (1), 1–15 (2023).

Wang, X. et al. A long single-span dispersion-decreasing-like fiber transmission system. Opt. Laser Technol. 116 , 338–344 (2019).

Razali, N. F., Isa, I. S., Sulaiman, S. N., Noor, N. K. & Osman, M. K. CNN-Wavelet scattering textural feature fusion for classifying breast tissue in mammograms. Biomed. Signal Process. Control 83 , 104683 (2023).

Yazdinejad, A., Parizi, R. M., Dehghantanha, A. & Choo, K. K. R. P4-to-blockchain: A secure blockchain-enabled packet parser for software defined networking. Comput. Secur. 88 , 101629 (2020).

Alam, T., Ullah, A. & Benaida, M. Deep reinforcement learning approach for computation offloading in blockchain-enabled communications systems. J. Ambient Intell. Humaniz. Comput. 14 (8), 9959–9972 (2023).

Sanwar Hosen, A. S. M. et al. Blockchain-based transaction validation protocol for a secure distributed IoT network. IEEE Access 8 , 117266–117277 (2020).

Michalski, R., Dziubaltowska, D. & MacEk, P. Revealing the character of nodes in a blockchain with supervised learning. IEEE Access 8 , 109639–109647 (2020).

Nasir, M. U., Khan, S., Mehmood, S., Khan, M. A., Zubair, M. & Hwang, S. O. Empowered with blockchain technology. (2022).

Zeadally, S. & Tsikerdekis, M. Securing internet of things (IoT) with machine learning. Int. J. Commun. Syst. 33 (1), 1–16 (2020).

Joshi, K. et al. Machine-learning techniques for predicting phishing attacks in blockchain networks: A comparative study. Algorithms 16 (8), 366 (2023).

Khan, I. A. et al. Fed-inforce-fusion: A federated reinforcement-based fusion model for security and privacy protection of IoMT networks against cyber-attacks. Inf. Fusion 101 , 102002 (2023).

Khan, I. A. et al. A privacy-conserving framework based intrusion detection method for detecting and recognizing malicious behaviours in cyber-physical power networks. Appl. Intell. 51 (10), 7306–7321 (2021).

Khan, I. A. et al. Enhancing IIoT networks protection: A robust security model for attack detection in internet industrial control systems. Ad Hoc Netw. 134 , 102930 (2022).

Khan, I. A. et al. An enhanced multi-stage deep learning framework for detecting malicious activities from autonomous vehicles. IEEE Trans. Intell. Transp. Syst. 23 (12), 25469–25478 (2022).

Sapra, N., Shaikh, I. & Dash, A. Impact of proof of work (PoW)-based blockchain applications on the environment: A systematic review and research agenda. J. Risk Financ. Manag. 16 (4), 218 (2023).

Schinckus, C. Proof-of-work based blockchain technology and anthropocene: An undermined situation?. Renew. Sustain. Energy Rev. 152 , 111682 (2021).

Saad, M., Njilla, L., Kamhoua, C., Kim, J., Nyang, D. & Mohaisen, A. Mempool optimization for defending against DDoS attacks in PoW-based blockchain systems. in ICBC 2019–IEEE International Conference on Blockchain and Cryptocurrency , pp. 285–292 (2019).

Li, W., Cao, M., Wang, Y., Tang, C. & Lin, F. Mining pool game model and nash equilibrium analysis for PoW-based blockchain networks. IEEE Access 8 , 101049–101060 (2020).

Lepore, C. et al. A survey on blockchain consensus with a performance comparison of PoW, PoS and pure PoS. Mathematics 8 (10), 1–26 (2020).

Cao, B. et al. Performance analysis and comparison of PoW, PoS and DAG based blockchains. Digit. Commun. Networks 6 (4), 480–485 (2020).

Liu, D., Alahmadi, A., Ni, J., Lin, X. & Shen, X. Anonymous reputation system for IIoT-enabled retail marketing atop PoS blockchain. IEEE Trans. Ind. Inform. 15 (6), 3527–3537 (2019).

Xu, G., Liu, Y. & Khan, P. W. Improvement of the DPoS consensus mechanism in blockchain based on vague sets. IEEE Trans. Ind. Inform. 16 (6), 4252–4259 (2020).

Zhou, T., Li, X. & Zhao, H. DLattice: A permission-less blockchain based on DPoS-BA-DAG consensus for data tokenization. IEEE Access 7 , 39273–39287 (2019).

Liu, J., Xie, M., Chen, S., Ma, C. & Gong, Q. An improved DPoS consensus mechanism in blockchain based on PLTS for the smart autonomous multi-robot system. Inf. Sci. 575 , 528–541 (2021).

Article   MathSciNet   Google Scholar  

Chen, S., Xie, M., Liu, J. & Zhang, Y. Improvement of the DPoS consensus mechanism in blockchain based on PLTS. in Proceeding of–2021 7th IEEE Intl Conference on Big Data Security on Cloud (BigDataSecurity), IEEE Intl Conference on High Performance and Smart Computing,(HPSC) and IEEE Intl Conference on Intelligent Data and Security (IDS) , pp. 32–37 (2021).

Wang, L., Xu, P., Su, W., Li, Y. & Chen, X. Research on Improvement of blockchain DPOS consensus mechanism based on HK clustering. in Proceeding–2021 China Autom. Congr. CAC 2021 , pp. 1167–1172 (2021).

Li, W. et al. A scalable multi-layer PBFT consensus for blockchain. IEEE Trans. Parallel Distrib. Syst. 32 (5), 1146–1160 (2021).

Chiu, W. Y. & Meng, W. EdgeTC–a PBFT blockchain-based ETC scheme for smart cities. Peer-to-Peer Netw. Appl. 14 (5), 2874–2886 (2021).

Liu, J., Feng, W., Zhang, Y. & He, F. Improvement of PBFT algorithm based on CART. Electronics 12 (6), 1460 (2023).

Liu, S., Zhang, R., Liu, C. & Shi, D. P-PBFT: An improved blockchain algorithm to support large-scale pharmaceutical traceability. Comput. Biol. Med. 154 , 106590 (2023).

Wu, Y., Wu, L. & Cai, H. Reinforced practical byzantine fault tolerance consensus protocol for cyber physical systems. Comput. Commun. 203 , 238–247 (2023).

Hu, Y. et al. A practical heartbeat-based defense scheme against cloning Attacks in PoA blockchain. Comput. Stand. Interfaces 83 , 103656 (2023).

Ometov, A. et al. An overview on blockchain for smartphones: state-of-the-art, consensus, implementation, challenges and future trends. IEEE Access 8 , 103994–104015 (2020).

Sasikumar, A. et al. Blockchain-based trust mechanism for digital twin empowered industrial internet of things. Futur. Gener. Comput. Syst. 141 , 16–27 (2023).

Shi, L., Wang, T., Li, J., Zhang, S. & Guo, S. Pooling is not favorable: Decentralize mining power of PoW blockchain using age-of-work. IEEE Trans. Cloud Comput. 11 (3), 2756–2769 (2022).

Google Scholar  

Bandara, E. et al. Casper: A blockchain-based system for efficient and secure customer credential verification. J. Bank. Financ. Technol. 6 (1), 43–62 (2022).

Zhu, S., Cai, Z., Hu, H., Li, Y. & Li, W. zkCrowd: A hybrid blockchain-based crowdsourcing platform. IEEE Trans. Ind. Inform. 16 (6), 4196–4205 (2020).

Buterin, V., Reijsbergen, D., Leonardos, S. & Piliouras, G. Incentives in Ethereum’s hybrid Casper protocol. in ICBC 2019–IEEE International Conference of Blockchain Cryptocurrency , pp. 236–244 (2019).

Sriman, B., Ganesh Kumar, S. & Shamili, P. Blockchain technology: consensus protocol proof of work and proof of stake. Adv. Intell. Syst. Comput. 1172 , 395–406 (2021).

Chen, Y. & Liu, F. Research on improvement of DPoS consensus mechanism in collaborative governance of network public opinion. Peer-to-Peer Netw. Appl. 15 (4), 1849–1861 (2022).

Hasanova, H., Baek, U. J., Shin, M. G., Cho, K. & Kim, M. S. A survey on blockchain cybersecurity vulnerabilities and possible countermeasures. Int. J. Netw. Manag. 29 (2), 1–36 (2019).

Schlatt, V., Guggenberger, T., Schmid, J. & Urbach, N. Attacking the trust machine: Developing an information systems research agenda for blockchain cybersecurity. Int. J. Inf. Manag. 68 , 102470 (2023).

Chaganti, R., Bhushan, B. & Ravi, V. A survey on blockchain solutions in DDoS attacks mitigation: Techniques, open challenges and future directions. Comput. Commun. 197 , 96–112 (2023).

Bhardwaj, A. et al. Penetration testing framework for smart contract Blockchain. Peer-to-Peer Netw. Appl. 14 (5), 2635–2650 (2021).

Liao, K. & Katz, J. Incentivizing Blockchain Forks via Whale Transactions. In Financial cryptography and data security. FC 2017. Lecture notes in computer science Vol. 10323 (eds Brenner, M. et al. ) 264–279 (Springer, 2017).

Iqbal, M. & Matulevicius, R. Exploring sybil and double-spending risks in blockchain systems. IEEE Access 9 , 76153–76177 (2021).

Xu, C. et al. A lightweight and attack-proof bidirectional blockchain paradigm for internet of things. IEEE Internet Things J. 9 (6), 4371–4384 (2022).

Nicolas, K., Wang, Y., Giakos, G. C., Wei, B. & Shen, H. Blockchain system defensive overview for double-spend and selfish mining attacks: A systematic approach. IEEE Access 9 , 3838–3857 (2021).

Li, B. et al. LBS meets blockchain: An efficient method with security preserving trust in SAGIN. IEEE Internet Things J. 9 (8), 5932–5942 (2022).

Nofer, M., Gomber, P., Hinz, O. & Schiereck, D. Blockchain. Bus. Inf. Syst. Eng. 59 (3), 183–187 (2017).

Shin, D. D. H. Blockchain: The emerging technology of digital trust. Telemat. Inform. 45 , 101278 (2019).

Kumar, R. & Sharma, R. Leveraging blockchain for ensuring trust in IoT: A survey. J. King Saud Univ.-Comput. Inf. Sci. 34 (10), 8599–8622 (2022).

Anjum, A., Sporny, M. & Sill, A. Blockchain standards for compliance and trust. IEEE Cloud Comput. 4 (4), 84–90 (2017).

Ismail, L. & Materwala, H. A review of blockchain architecture and consensus protocols: Use cases, challenges, and solutions. Symmetry 11 (10), 1198 (2019).

Khatri, N., Shrestha, R. & Nam, S. Y. Security issues with in-vehicle networks, and enhanced countermeasures based on blockchain. Electronics 10 (8), 1–33 (2021).

Pourrahmani, H., Yavarinasab, A., Monazzah, A. M. H. & Van Herle, J. A review of the security vulnerabilities and countermeasures in the Internet of Things solutions: A bright future for the blockchain. Internet Things 23 , 100888 (2023).

Xia, J., Li, H. & He, Z. The effect of blockchain technology on supply chain collaboration: A case study of Lenovo. Systems 11 (6), 299 (2023).

Connolly, L. Y. & Wall, D. S. The rise of crypto-ransomware in a changing cybercrime landscape: Taxonomising countermeasures. Comput. Secur. 87 , 101568 (2019).

Nguyen, C. T. et al. Proof-of-stake consensus mechanisms for future blockchain networks: Fundamentals, applications and opportunities. IEEE Access 7 , 85727–85745 (2019).

Yusoff, J., Mohamad, Z. & Anuar, M. A review: Consensus algorithms on blockchain. J. Comput. Commun. 10 (09), 37–50 (2022).

Fu, X., Wang, H. & Shi, P. A survey of blockchain consensus algorithms: mechanism, design and applications. Sci. China Inf. Sci. 64 (2), 1–15 (2021).

Bamakan, S. M. H., Motavali, A. & Babaei Bondarti, A. A survey of blockchain consensus algorithms performance evaluation criteria. Expert Syst. Appl. 154 , 113385 (2020).

Foti, M., Mavromatis, C. & Vavalis, M. Decentralized blockchain-based consensus for optimal power flow solutions. Appl. Energy 283 , 116100 (2021).

Shahsavari, Y., Zhang, K. & Talhi, C. Toward quantifying decentralization of blockchain networks with relay nodes. Front. Blockchain 5 , 1–11 (2022).

Yadav, A. K. et al. A comparative study on consensus mechanism with security threats and future scopes: Blockchain. Comput. Commun. 201 , 102–115 (2023).

Liu, Y., Ke, J., Xu, Q., Jiang, H. & Wang, H. Decentralization is vulnerable under the gap game. IEEE Access 7 , 90999–91008 (2019).

Kim, H., Kim, S. H., Hwang, J. Y. & Seo, C. Efficient privacy-preserving machine learning for blockchain network. IEEE Access 7 , 136481–136495 (2019).

Farooq, M. S. et al. Blockchain-based smart home networks security empowered with fused machine learning. Sensors 22 (12), 1–13 (2022).

Miglani, A. & Kumar, N. Blockchain management and machine learning adaptation for IoT environment in 5G and beyond networks: A systematic review. Comput. Commun. 178 , 37–63 (2021).

Ali, J., Khan, R., Ahmad, N. & Maqsood, I. Random forests and decision trees. IJCSI Int. J. Comput. Sci. Issues 9 (5), 272–278 (2012).

Ren, Y. S., Ma, C. Q., Kong, X. L., Baltas, K. & Zureigat, Q. Past, present, and future of the application of machine learning in cryptocurrency research. Res. Int. Bus. Financ. 63 , 101799 (2022).

Ferrag, M. A. & Maglaras, L. DeepCoin: A novel deep learning and blockchain-based energy exchange framework for smart grids. IEEE Trans. Eng. Manag. 67 (4), 1285–1297 (2020).

Amirzadeh, R., Nazari, A. & Thiruvady, D. Applying artificial intelligence in cryptocurrency markets: A survey. Algorithms 15 (11), 428 (2022).

Keshk, M., Turnbull, B., Moustafa, N., Vatsalan, D. & Choo, K. K. R. A privacy-preserving-framework-based blockchain and deep learning for protecting smart power networks. IEEE Trans. Ind. Inform. 16 (8), 5110–5118 (2020).

Jameel, F. et al. Reinforcement learning in blockchain-enabled IIoT networks: A survey of recent advances and open challenges. Sustainability 12 (12), 1–22 (2020).

Zhang, F., Wang, H., Zhou, L., Xu, D. & Liu, L. A blockchain-based security and trust mechanism for AI-enabled IIoT systems. Futur. Gener. Comput. Syst. 146 , 78–85 (2023).

Dai, Y. et al. Blockchain and deep reinforcement learning empowered intelligent 5G beyond. IEEE Netw. 33 (3), 10–17 (2019).

Qiu, C., Ren, X., Cao, Y. & Mai, T. Deep reinforcement learning empowered adaptivity for future blockchain networks. IEEE Open J. Comput. Soc. 2 , 99–105 (2020).

Mohammed, Z. K. et al. Bitcoin network-based anonymity and privacy model for metaverse implementation in Industry 5.0 using linear Diophantine fuzzy sets. Ann. Oper. Res. https://doi.org/10.1007/s10479-023-05421-3 (2023).

Jang, H. & Lee, J. An empirical study on modeling and prediction of bitcoin prices with bayesian neural networks based on blockchain information. IEEE Access 6 , 5427–5437 (2017).

Raja, L. & Periasamy, P. S. A trusted distributed routing scheme for wireless sensor networks using block chain and jelly fish search optimizer based deep generative adversarial neural network (deep-GANN) technique. Wirel. Pers. Commun. 126 (2), 1101–1128 (2022).

Elsayed, R., Hamada, R., Hammoudeh, M., Abdalla, M. & Elsaid, S. A. A Hierarchical deep learning-based intrusion detection architecture for clustered internet of things. J. Sens. Actuator Netw. 12 (1), 3 (2023).

Dwivedi, A. D., Srivastava, G., Dhar, S. & Singh, R. A decentralized privacy-preserving healthcare blockchain for IoT. Sensors 19 (2), 1–17 (2019).

Viswanadham, Y. V. R. S. & Jayavel, K. A framework for data privacy preserving in supply chain management using hybrid meta-heuristic algorithm with ethereum blockchain technology. Electronics 12 (6), 1404 (2023).

Ogundokun, R. O., Misra, S., Maskeliunas, R. & Damasevicius, R. A review on federated learning and machine learning approaches: categorization, application areas, and blockchain technology. Information 13 (5), 263 (2022).

Mekdad, Y. et al. A survey on security and privacy issues of UAVs. Comput. Networks 224 , 362–367 (2023).

Frimpong, S. A. et al. RecGuard: An efficient privacy preservation blockchain-based system for online social network users. Blockchain Res. Appl. 4 (1), 100111 (2023).

Qi, J. & Guan, Y. Practical Byzantine fault tolerance consensus based on comprehensive reputation. Peer-to-Peer Netw. Appl. 16 (1), 420–430 (2023).

Sajana, P., Sindhu, M. & Sethumadhavan, M. On blockchain applications: Hyperledger fabric and ethereum. Int. J. Pure Appl. Math. 118 (18), 2965–2969 (2018).

Nakandala, S. et al . A tensor compiler for unified machine learning prediction serving. in Proceeding of 14th USENIX Symposium on Operating Systems Design and Implementation, OSDI 2020 , pp. 899–917 (2020).

Alsheikh, M. A., Lin, S., Niyato, D. & Tan, H. P. Machine learning in wireless sensor networks: Algorithms, strategies, and applications. IEEE Commun. Surv. Tutorials 16 (4), 1996–2018 (2014).

Mao, Q., Hu, F. & Hao, Q. Deep learning for intelligent wireless networks: A comprehensive survey. IEEE Commun. Surv. Tutorials 20 (4), 2595–2621 (2018).

Download references

Acknowledgements

“This research work was supported and funded by the National Defence University Malaysia (NDUM)—UPNM/2023/GPPP/ICT/1 and UPNM/2022/GPJP/ICT/3”.

This research received funding from the corresponding research grants: UPNM/2023/GPPP/ICT/1 and UPNM/2022/GPJP/ICT/3.

Author information

Syarifah Bahiyah Rahayu

Present address: Department of Science Defense, Faculty of Defense Science and Technology, National Defence University Malaysia (UPNM), Kuala Lumpur, Malaysia

Authors and Affiliations

Cyber Security & Digital Industrial Revolution Centre, National Defence University Malaysia (UPNM), Kuala Lumpur, Malaysia

K. Venkatesan & Syarifah Bahiyah Rahayu

Department of Science Defense, Faculty of Defense Science and Technology, National Defence University Malaysia (UPNM), Kuala Lumpur, Malaysia

K. Venkatesan

You can also search for this author in PubMed   Google Scholar

Contributions

Dr.K.V: coordinating the work, investigation, methodology, writing original draft, Dr.S.B.R: methodology, project administration, writing, review and editing, All authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Syarifah Bahiyah Rahayu .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Venkatesan, K., Rahayu, S.B. Blockchain security enhancement: an approach towards hybrid consensus algorithms and machine learning techniques. Sci Rep 14 , 1149 (2024). https://doi.org/10.1038/s41598-024-51578-7

Download citation

Received : 05 October 2023

Accepted : 07 January 2024

Published : 11 January 2024

DOI : https://doi.org/10.1038/s41598-024-51578-7

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Blockchain as a Cyber Defense: Opportunities, Applications, and Challenges

Ieee account.

• Change Username/Password
• Update Address

Purchase Details

• Payment Options
• Order History
• View Purchased Documents

Profile Information

• Communications Preferences
• Profession and Education
• Technical Interests
• US & Canada: +1 800 678 4333
• Worldwide: +1 732 981 0060
• Contact & Support
• About IEEE Xplore
• Accessibility
• Terms of Use
• Nondiscrimination Policy
• Privacy & Opting Out of Cookies

A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. © Copyright 2024 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• Perspect Health Inf Manag
• v.20(3); Summer-Fall 2023
• PMC10701638

Blockchain Revolutionizing Healthcare Industry: A Systematic Review of Blockchain Technology Benefits and Threats

Fatma m. abdelsalam.

Belk College of Business at the University of North Carolina Charlotte

Blockchain technology has been gaining significant traction in the healthcare industry in the past few years. The value proposition of using blockchain technology is to augment interoperability among healthcare organizations. However, the disruptive technology comes with costly drawbacks. The aim of this paper is to explore the benefits and threats of blockchain technology as a disruptive innovation in the healthcare sector. Current blockchain applications were reviewed through studies conducted to identify uses and potential challenges of blockchain technology based on its current implementations. This literature review highlights gaps in research and the need for further blockchain studies, particularly in the healthcare domain.

Introduction

One of the challenges encountered by the healthcare industry is the inability to safely manage and retrieve personal health information (PHI) in a timely manner. Effective management and retrieval of patient data would enable healthcare providers to capture a holistic picture of a patient's health, improve patient-physician interaction, and achieve better use of healthcare-related data 1 . Interoperability has enormous potential to transform the health sector through the development of affordable cures and cutting-edge treatments for numerous diseases but depends upon smooth, effective data exchange, and distribution across all the well-known network participants and health professionals 2 . Privacy and security threats are common challenges faced by the healthcare industry. The rise in cybersecurity attacks and security breaches of healthcare records has stimulated the pressing need of healthcare organizations to invest in advancing security technologies 3 . As a disruptive innovation, blockchain technology is paving the way for new potential of solving serious data privacy, security, and integrity issues in healthcare and facilitating the paradigm shift of patient-centric interoperability, while enabling decentralization and transparency of stored information 4 . The global pandemic has revealed a lack of interoperability in the current healthcare system and the need for accurate clinical data that can be widely distributed to healthcare providers in an efficient and secure manner 5 .

Blockchain is seen as a key breakthrough that will likely have a considerable influence on a myriad of different industries such as healthcare, supply chain management, and business. A peer-to-peer network called blockchain was initially proposed by Satoshi in 2008 and then commercialized in 2009 when Bitcoin emerged as its first use case 46 . Kassab et al reported that in 2016, “healthcoin” was developed by Diego Espinosa and Nick Gogerty as the first platform based on blockchain to manage and reward Type-2 diabetes prevention 39 . Users submit their biomarkers into the blockchain. If the biomarker is an improvement, the system rewards the patient with digital tokens: healthcoin that can be applied toward government tax breaks and/or discounts on multiple fitness brands 6 . Future technology may open the door to significant opportunities, ranging from research and economics to interactions between patients and physicians 7 . Blockchain technology conflates complexity, novelty, and diversity, which has posed challenges in gauging the value proposition of incorporating the technology 47 . Due to its complexity, blockchain may be used for managing business processes or as a workflow system 8 .

Several research studies have been conducted on the benefits and challenges of blockchain technology in the healthcare industry. However, some of the potential applications have not yet deployed 9 . The objective of this literature review is to explore the research studies that have been conducted on applications of blockchain technology as a disruptive innovation in healthcare industry 10 , addressing current and potential uses, benefits, and threats of the technology based on the historical research studies. Several researchers suggested studying the outcomes of leveraging blockchain technology in the context of improving security of health records, meeting social determinant of health needs, and improving health outcomes 11,12, 3, 4 . Based on this context, the previously available scholarship on blockchain were analyzed through a systematic review as an assessment tool. The findings convey key insights on the current state of research investigation on blockchain, including benefits and implications as a disruptive innovation in healthcare industry 13 . The study also highlights the gaps in research and the need for further blockchain research in healthcare domain.

This paper was framed to guide future researchers and decision-makers on the current knowledge of benefits, drawbacks, and gaps in blockchain research landscape. The findings were conveyed to proactively identify key challenges pertaining to blockchain adoption and application in the healthcare domain to support improvement opportunities and tackle challenges at their initial stages. This paper was framed to explore the theoretical lens of disruptive innovation theory and innovation diffusion theory. The study was organized to begin with a background of blockchain technology, then explore its key uses and potential benefits within a healthcare context based on the research studies and addressing possible threats discussed by literature from an organizational, social, and technological level. Finally, this review provides recommendations to guide future research, bridge the gaps identified in literature, and further examine the prototypes implemented in the healthcare sector.

Literature Review

Blockchain is considered a relatively recent invention that first appeared in 2008 and provided the technical foundation for the birth of the cryptocurrency known as “bitcoin.” In general, blockchain may be thought of as a method of network organization that combines distributed ledgers and databases. In this design, records are updated or maintained by a certain authority but are dispersed over all computers connected to the network so that no one node has the power to change the data that is being stored. For the handling of sensitive data, such as health information or financial transactions, this specific component might be useful 14 . The healthcare industry, one of the biggest in the world, frequently must deal with a complicated network of interrelated stakeholders that are subject to a variety of rules and have their patient data dispersed across numerous databases. Blockchain technology can help healthcare professionals in this difficult situation address the present inefficiencies in the sector 15 .

Healthcare data management systems encounter issues including data transparency, traceability, immutability, audit, data provenance, flexible access, trust, privacy, and security. By overcoming these obstacles and bringing about significant advances, blockchain technology can completely transform healthcare data administration, blockchain establishes confidence in health data by enabling the tracking of changes from their source to their present form. Current projects and recent case studies show how useful blockchain technology is for a range of healthcare applications. However, there are issues that need to be resolved for blockchain to be successfully adopted in the healthcare industry. Overcoming these difficulties and further investigating the possibilities of blockchain in healthcare data management should be the main goals of future research 8 .

Several review articles on blockchain technology's use in industries including banking, the internet of things (IoT), the energy sector, government, and privacy and security are now available in the open literature. A broad thorough critical assessment of the most recent research on blockchain-based healthcare applications is not addressed, despite a few review papers discussing the uses of blockchain technology in healthcare. For instance, most of the studies give a brief overview of blockchain-based healthcare applications. Despite being the first to provide a high-level overview of new blockchain-based healthcare applications, the study largely focuses on the practical applications and advantages of this technology 16 .

Blockchain technology can change the topology of a healthcare network such that data are added in a decentralized fashion. Blockchain improves data security, confidentiality, and interoperability while allowing patients to integrate themselves into an ecosystem 17 . Bibliometric analyses of blockchain technology in the healthcare sector are few. In this regard, there is a growing body of literature examining and debating the potential and current applications of blockchain in healthcare. To our knowledge, however, none of these studies examine the potential environmental and health effects of this industry's potential use of blockchain. This lack of attention should be addressed because, in theory, any technical improvements to the healthcare sector should be made in a way that does not hurt either the environment or people's health. This study addresses blockchain technology and healthcare studies to bridge the gap. It also discusses potential directions for future research with the right depth and breadth in pertinent areas.

Disruptive innovation theory has analyzed and addressed growth driven by innovation 18 . The theory was originally initiated by Clayton Christensen et al. in 1995 and has pervaded the clinical healthcare dialect over the past years. Increased adoption of blockchain technology in the healthcare domain will lead to a disruptive shift in the foundation of the healthcare system 13 . Despite the growing use of the concept in literature, there are gaps in comprehending disruptive innovations in a healthcare context as there is no objective definition in healthcare literature 19 . In addition, there is no published literature that compares perceived healthcare disruptive innovations. Therefore, key innovations in the sector remain in silos, which limits our ability to identify disruption.

Innovation diffusion theory states that characteristics of innovation affect how organizations gather knowledge, which consequently affects the decision to adopt or reject the innovation. These characteristics are: (1) relative advantage; (2) compatibility; (3) complexity; (4) trialability; and (5) observability 20 . Haleem and Hartley 3, 20 have noted that lack of blockchain understanding is a barrier to technology diffusion. Given the relatively early stage of blockchain development, most healthcare organizations often rely on consultants when adopting modern technology 2 . Additional barriers to diffusion success are switching costs and the network effect 10 .

Methodology

Systematic reviews are an effective way of evaluating and interpreting research relevant to a particular research question, topic area, or phenomenon of interest based on previous research outcomes 21 . Systematic reviews are common in the medical field and healthcare domain. Nonetheless, there are many research studies addressing blockchain technology applications in healthcare 4, 13, 22, 23 . For example, 24 conducted a systematic review of the adoption of blockchain platforms in healthcare and how they improved the industry outcomes.

To compile data and insights on blockchain in healthcare research, meta-analysis was conducted and identified studies were included in the review using a list of relevant terms through the search of several electronic databases including PubMed, MEDLINE, Scopus, EBSCO, and IEEE Xplore, and other databases for research including ScienceDirect, and Google Scholar. By choosing the mentioned databases, the intention was to focus on peer-reviewed articles that have been published in healthcare journals. The database was searched to determine whether a publication contained at least one of the keywords or search terms in the title, abstract, or keywords. In total, 1,830 articles were identified. The Boolean operator was utilized with a combination of “AND/OR” of search terms. The following search string was used: blockchain AND (healthcare OR medical) AND (challenge, threat, OR benefit OR uses OR 1 application). Following this process, 37 articles were determined to be relevant to the study. Subsequently, a backward reference-list checking was conducted to identify other relevant literature 5 . As a result, 10 more articles were identified. In total, 47 articles were identified to be relevant to this literature review.

To narrow down the literature selection process to the relevant articles, all publications that are fully available in English language and published between 2016 and 2022 were included. Duplicate articles, book chapters, and papers that discussed blockchain from a technical and engineering perspective were excluded. Based on figure ​ figure1 1 , 33 articles were identified in the final population for analysis as relevant literature. EndNote software was utilized for duplicate removal and final screening. To ensure reliability, the search process was comprehensively documented to identify studies, assess relevance, and synthesize the structure of the paper. The goal was to find research articles focused on blockchain applications, benefits, and threats in healthcare domain. This literature will answer the following research questions: How has blockchain been defined in literature? What are the potential blockchain applications in healthcare domain? What are the blockchain benefits in healthcare literature? What are the possible threats of blockchain technology in the healthcare industry? For the purposes of the review, blockchain research was categorized into three categories: 1) Applications in healthcare industry, 2) Benefits of blockchain, and 3) Threats of the technology.

PRISMA for Identification and Inclusion Process of Systematic Review

Most of the scholars describe blockchain using their properties 13, 48 defined blockchain as a decentralized transparent ledger with transaction records. Blockchain technology is characterized as “an open, distributed network that may record transactions between two individuals rapidly and in a verified and conspicuous way.” Blockchain is described by several authors as a digitized decentralized ledger to allow recordkeeping of all peer-to-peer transactions without the need for a centralized authority 19 . Blockchain was also described as “a distributed ledger system, which maintains all transactions synced across users” 25 . Researchers highlighted that information that has already been used in a transaction cannot be altered or deleted, and users can openly and transparently audit any transactions. The technology protects data from manipulation and alteration. The studies addressed that blockchain offers tremendous efficiency and affordable solutions in the healthcare industry. The essential technology characteristics include decentralization, traceability, immutability, and provenance 26 .

Since 2016, the demand for blockchain technology has increased globally, and several large technology firms, such as IBM, Intel, and Microsoft, are heavily invested in blockchain technology development. The World Economic Forum estimates that, by 2025, 10 percent of the global gross domestic product will be stored on blockchain technology 27 . The marketplace for blockchain technology was estimated to be worth around $339.5 million globally in 2017, and it is expected to increase to $2.3 billion by 2021. By 2030, blockchain is anticipated to provide $3.1 trillion in economic value. According to International Data Corporation (IDC), worldwide spending on blockchain will increase from about $1.5-$2.9 billion in 2018- 2019 and rise?? to $11.7 billion in 2022 10 . For the anticipated period of 2017–2022, the anticipated annual compound growth rate is 73.2 percent 49 . The US healthcare industry is the world's largest and absorbs more than $1.7 trillion per year 28 . Today, the average annual cost of healthcare per person in America is $10,739, which is more than residents of any other country 28 . Abdel-Basset 5 , noted that blockchain technology can be used to manage pandemics by considering different data sources, which can be statistically analyzed to extract essential features and patterns for healthcare professionals and the government.

Although understanding blockchain technology might be challenging, the fundamental ideas behind it are rather straightforward. Blockchain is a database of a group of data that is electronically stored on a computer network 50 . In an examination of academic literature where blockchain applications have been applied to diverse topics, it can transform the traditional industry with its features, which include decentralization, anonymity, persistency, and auditability 29 . The studies reviewed have covered several instances of blockchain technology being used in healthcare, as well as the issues and potential fixes. The design decisions and compromises made by the researchers were addressed in the many situations where this innovative technology was used 11 . The Office of the National Coordinator for Health Information Technology (ONC) has described several features critical to the development of an interoperable health system, which are addressed by blockchain 9 . The research studies have covered a wide range of settings for using this technology, including blockchain-based applications across many different sectors 11 . Then the researchers describe some aspects of blockchain technology for medical record management, insurance claim process, biomedical research, and health data ledgers 30 . There is a consensus among researchers that, with blockchain technology, patient data will be truly owned and controlled by the right owner of the data, which is the patient 31 . The healthcare industry is a suitable candidate for the use of blockchain technology since it may address critical concerns including computerized claim verification and global health management 23 . With the assistance of this technology, patients may maintain their personal information and choose with whom it will be shared, overcoming the present problems with data ownership and exchange. Despite the general belief that the benefits of adopting blockchain-based technology may be exaggerated, a new study suggests that enterprises will still make significant investments in this area in the future 2 .

It might be argued by researchers that this system has not yet lived up to expectations, a reality that may be explained by the widespread deployment of blockchain, particularly in relation to governmental restrictions and other difficulties 19 . Another key barrier to the widespread adoption of blockchain is that both general and specialized users such as patients or doctors are unaware of how it operates, its technological aspects, or its benefits for processing data 1 . The researchers proposed that it could take some time for this technology to build all anticipated and expected stages of transformational change in business, mostly due to implementation obstacles in the manner of organizational and social issues such as security concerns and governance 8,19,24 . This may also be made worse by widespread misunderstandings about how blockchain technology is used in government policies and regulations. By removing these obstacles, recent research aims to assist blockchain clear changes and expedite its spread 32 .

The papers reviewed described many blockchain uses and potential issues often at the conceptual level. However, empirical studies are limited as blockchain research remains early-stage and immature, particularly in healthcare 5,11,27 . Blockchain technology is a prominent example of disruptive innovation. However, poor identification can lead to poor understanding of the technical features and potential of an innovation and the possible barriers to adoption and ways to overcome them 33 .

Healthcare Industry Challenges

Some of the numerous concerns hospitals and other healthcare organizations deal with daily include patient data access, medication storage logs, and medical records. Patient care, information security, and privacy must all be balanced in the healthcare sector. Major challenges the healthcare sector faces include putting the patient first, privacy and access, accuracy of medical data, pricing, management of supply chains and prescription records. Even if the conventional technique of storing data through a centralized database can be damaging, as indicated in research, it can also be susceptible to hacking or even a single failure point 13 .

The fact that all the servers temporarily go offline while the changes are being made to the databases used to store medical data is another problem with a traditional database. Given that healthcare is a 24/7 industry, this little gap might prove to be quite deadly 23 . Another concern with medical records is the cost associated with transferring records among different entities. The lack of availability of test results can be dangerous in terms of delayed treatment. Also, sending data via email is considered a security risk. A system integrating patient consent and access to authorized individuals would improve efficiency and save on financial costs 9 . Blockchain technology is being promoted as the “solution” to issues in a variety of healthcare issues 34 . By doing a thorough literature review and responding to the research questions posed in the research, this study attempts to discover blockchain technology capabilities in the healthcare sector. The potential of blockchain technology has extended to the healthcare sector, enabling a change in the way the present system and its utilization of technology currently operate.

The study seeks to emphasize the potential paths for blockchain research in healthcare, as well as to emphasize the possible uses of the platform. According to literature, blockchain technology is currently being researched in the field of healthcare, where it is mostly employed for network access, data exchange, and record management 23 . Additionally, it demonstrates that many studies lack implementation or prototype information. The authors of literature reviewed reached the conclusion that blockchain application-based research is expanding and growing at an exponential rate 5 . The research has also demonstrated that the exponential growth of blockchain technology initiatives in the healthcare industry are projected to have a major influence. A systematic study method was conducted, employing a well-planned monitoring strategy to look for pertinent papers. Several studies have put out various scenarios for the application of blockchain in healthcare systems. The assessment also identifies benefits as well as shortcomings and potential future research topics. To further comprehend, define, and assess the usefulness of blockchain in healthcare, additional study is still required 9 .

Main Features of Blockchain Technology

The four key characteristics of blockchain were identified by research studies and serve as the foundation upon that it has expanded. Technology's four distinguishing characteristics are: decentralization, immutability, transparency, and provenance 8 . Healthcare systems have used centralized systems up to the advent of blockchain to fulfill data exchange requirements. A centralized institution is employed to hold all the information in a central network, and only that entity and the user may communicate with each other. Even though centralized systems have indeed been in use for a long time, there are certain restrictions associated with this kind of network. Since the data is kept in a single main place and by a single organization, this turns into a red flag for would-be cybercriminals or hackers and even represents a lone source of potential failure 36 , 37 .

Blockchain offers a decentralized network as an alternative option to a centralized one, removing the necessity for a single centralized power to rule over the network 22 , 23 , discussed the idea of immutability, which states that once data or information has been generated it should not be changed. When a blockchain record has been created, it cannot be changed once it has joined the network 9 . This is a crucial aspect of the blockchain that may be used to stop a lot of unethical or questionable behavior in any sector 41 . Blockchain transparency is a term that is frequently misunderstood. With the use of sophisticated encryption, a person's identity is concealed and just their upgradable is shown 8 . The provenance feature of the blockchain implies that any additions to the blockchain are visible to all the patient's network members 39 .

Blockchain Applications in Healthcare

Blockchain is a relatively emerging and developing technology that offers creative uses in the healthcare industry. The development of affordable cures and cutting-edge treatments for numerous diseases depends on smooth, effective data exchange and distribution across all the well-known network participants and health professionals. In the upcoming years, this will hasten the expansion of the healthcare sector. The studies reviewed highlighted that Ethereum and Hyperledger fabric seem to be the most used platforms/frameworks in this domain 12 . The studies unveiled blockchain technology prospects in the supply chain highlighting the benefits for the healthcare business. This is among the primary areas that the digital revolution enhances and innovates since it immediately affects living quality. Blockchain technology is also growing in popularity in the healthcare industry. It presents several significant and spectacular opportunities, ranging from research and economics to interactions between patients and physicians 7 . The most significant research explored and organized according to several use cases in this domain, include electronic health records (EHRs), remote monitoring of patients, pharmaceutical distribution network, and healthcare insurance claims 8,10,24 .

1. Electronic Health Records

The administration of health data, which might be enhanced by the capacity to integrate disparate systems and enhance the precision of EHRs, should be given priority in the effort to change healthcare. While the phrases electronic patient records (EPRs) and electronic health records (EHRs) are sometimes used interchangeably, they have different meanings. EMRs, or electronic medical records, are a more recent name for the paper charts kept by clinicians in their offices. The medical and treatment histories of patients in a single practice are recorded in an EMR. EHRs, on the other hand, put a greater emphasis on a patient's overall health, going beyond the usual clinical data gathered at the doctor's office and taking a more comprehensive approach to a patient's care.

According to the studies reviewed, blockchain helps manage EHRs. To handle authorization and data exchange across healthcare entities, Ekblaw et al. described MedRec, an EHR-related solution that suggests a decentralized method. The MedRec platform provides patients with information and understanding about who may access their medical records. FHIR Chain (Fast Health Interoperability Records and Blockchain) is another program that incorporates EHRs 36 . It is a medical record management-focused, blockchain-based platform for exchanging clinical data that is developed using bitcoin, and patients can get solutions from FHIR Chain. Nonetheless, Xia et al. introduced Medshare, an ethereum program for systems that experience a lack of communication for information sharing among cloud computing owing to the negative risks towards disclosing the content of personal data information. When exchanging medical data in cloud archives, Medshare offers data monitoring, and governance among large data organizations. MedBlock and BlockHIE are two further EMR apps built on the blockchain. MedBlock offers a method for searching records.

The suggested method keeps track of the addresses of health records blocks that are organized by health professionals. Each patient assessment has a link to the relevant blockchain record. Jiang et al. proposal for BlockHIE presents a blockchain-based healthcare system 34 . BlockHIE blends off-chain retention, in which data is kept in database systems of external institutions, with on-chain validation to continue taking advantage of current databases. Another blockchain-based healthcare platform addressed in the literature is called Ancile, which employs smart contracts to ensure data security, confidentiality, access management, and EMR compatibility 45 .

2. Remote Patient Monitoring

Remote patient surveillance refers to the gathering of medical data using smart phones, wireless body sensor sensors devices, and Internet of Things (IoT) devices to be able to monitor various patients’ conditions 30 . Blockchain technology is crucial for the storage, exchange, and retrieval of remotely gathered health data. It offers a solution in this setting where information is sent from mobile devices to a blockchain-based application on Hyperledger 2, 23 . By providing real-time patient monitoring applications, ethereum platform contracts may allow automated interventions in a safe setting 51 , 12 . Other literature suggested ways highlight the enormous potential of the IoT in various fields, particularly how it is being widely utilized in e-health. Io Health, a data-flow architecture that integrates the IoT with blockchain and may be used for accessing, storing, and managing e-health data, is a suggestion made in this area 36 .

3. Pharmaceutical Supply Chain

The pharmaceutical sector is another recognized use case for blockchain as patients may suffer severe effects if they get fake or subpar medicine. According to a study by the World Health Organization (WHO), over 100,000 people die in Africa due to improper dosing from counterfeited drugs obtained from untrusted vendors 4 and research has determined that blockchain technology has the power to solve this issue. Drug counterfeit has also been tackled by the researchers, who suggest a safe, irreversible, and verifiable supply chain for pharmaceuticals built on blockchain-based technology to prevent it 19 , 34 . In relation to drug regulating issues, drug standardization difficulties were addressed. Researchers?? have drawn attention to the challenges in identifying fake medications and suggested a blockchain-based approach to do so. Even though the suggested approach is only implemented in a small number of articles, several intriguing studies have addressed problems with the pharmaceutical supply chain 4 .

4. Health Insurance Claims

One area of healthcare that can profit from blockchain's absoluteness, openness, and traceability of stored data on it is healthcare insurance claims. Blockchain technology has promising solutions to handle health insurance claims. However, there are few prototypes and applications of these systems 9 . MIStore, a cryptocurrency health coverage system that offers the medical coverage data that is well-secured and maintained, was located 34 , 35 .

Benefits of Blockchain in Healthcare Sector

The blockchain technology enables medical professionals to embrace the notion of a public database that can be used to develop shareable, customized healthcare plans for their patients. As a result, this may readily assist in the facilitation and creation of personalized health plans that classify the patients based on their shared genetic data, age, and gender. Researchers have identified and divided blockchain benefits into individual benefits, organization-related benefits, and government benefits. Since users may only establish their identities once in the blockchain network, and the recorded identification traits are encrypted and kept in every blockchain server, users will not need to re-register their identities for accessibility in the foreseeable future.

Additionally, several researchers have highlighted the benefits of blockchain technology and how they addressed existing challenges in healthcare applications 12,19,42 . For example, ChengYing et al., 2018, explored the benefits of blockchain to link patients’ EHRs across different healthcare services.

Patient-level Benefits

The literature on blockchain technology offers proof that the technology can get around some of the problems with the current healthcare system. The advantages of blockchain technology allow for efficient maintenance and interchange of health records. The decentralization of patient information creates a single point of truth for connectivity and efficiency 2 . Data reconciliation among all parties engaged in the transaction is made unnecessary by leveraging blockchain, which improves cost effectiveness 10 . Only authorized people are granted access to sensitive and important patient data and protected health records, and a lifelong and continuous health status record may be created using blockchain technology 38 .

Patient data in the current healthcare information systems is frequently corrupted, prone to data breaches, or at elevated risk of failing. Data security is hence the main advantage of blockchain technology. According to a survey on the present status of EHRs with a sample size of 8,774, almost 40 percent of physicians view connectivity and EHR design as the main causes of their dissatisfaction 32 . It is challenging to transfer, retrieve, and analyze data due to the restricted data exchange and absence of compatibility among healthcare storage solutions. Berryhill et al. 43 noted that better compatibility is made possible by blockchain technology.

Organization-level Benefits

In terms of organizational advantages, blockchains have the capacity to offer safe patient data sharing across healthcare organizations. The group of authorized healthcare organizations taking part in the private network would be able share and access the information stored in the blockchain in a safe and trustworthy fashion 3 . Other studies emphasized the need of using blockchains to streamline the management of clinical trials because the study involves extremely sensitive patient-related data 27 .

Government Benefits

Blockchain technology has enabled the government to offer new public healthcare designs, assist in addressing fraud and waste, reduce the cost and sophistication of different health activities, and identify misuse and fraud activities 31 . It is thought that establishing a public blockchain will save costs, speed up learning, reduce risk, boost technology acceptance, and have an impact on regulations 28 . Another advantage of blockchain applications is successful care surveillance, especially for extremely ill patients since this technology can help physicians perform appropriate medical treatments. To do this, patients’ wearable technology, including smart watches, cell phones, and smart glasses, must be linked to the public blockchain of the healthcare provider 4 . In this section of the literature, the blockchain benefits that are most explored and addressed by previous studies were highlighted.

1. Securing Patient Data

Protecting patient information is one of the most important aspects of the healthcare industry. Falsifying patient records might contribute to difficulty for hospitals and physicians to diagnose and treat their patient's illness or issue. According to research studies, more than 176 million medical data records were compromised between 2009 and 2017. The data was hacked by cybercriminals, who then exploited it unethically 35 . Health data may be gathered using blockchain without having to move it all to a single place or centralized database. In the current EHR system, healthcare professionals hold the records, while patients have the right to access their own health records. Improved security and data integrity are made possible by the dissemination of health records and the data integrity of the data 13 . Data integrity is essential to healthcare since the current healthcare system has problems providing patients with accurate or sufficient information. Blockchain reduces the likelihood that unauthorized users would be able to extract health information 29 .

2. Medical Drugs Supply Chain Management

Medications or pharmaceuticals are created in laboratories and pharmaceutical firms all over the world. According to each country's needs, these medications are further distributed across the world. What happens if the medications are tampered with while being transported across the nation? As a result, the importers and exporters must have access to a transparent, tamper-proof healthcare supply chain. Blockchain minimizes this issue because of its transparency, decentralization, and tamper-proof properties 3 . Each carrying point for the medicine will be added to the blockchain after a distributed ledger has been established, making the whole transportation process visible 37 .

3. Single Longitudinal Patient Records

Every medical chart will be added to the blockchain ledger since it is made up of a chain of blocks called a blockchain. Examining the pre-compiled records would allow healthcare providers to have a broader picture of patients’ medical conditions. Additionally, it will assist in mastering patient indices, streamlining data meticulously, and avoiding expensive errors 29 .

4. Supply chain optimization

Authenticating the origin of medical supplies to assure the legitimacy of medications is a problem facing the healthcare industry. Supplies may be tracked from manufacture to every step of the supply chain with the use of blockchain technology. This makes it possible to acquire items transparently and visibly. This may assist businesses in implementing artificial intelligence (AI) and improving demand forecasting and supply optimization, while also boosting consumer confidence 44 .

5. Drug Traceability

The most trustworthy, dependable, and safe way to trace every medicine back to its source is via blockchain. There will be a hash value associated with every data block including drug-related information. By using this hash code, the data is protected against manipulation. All parties with permission to see the blockchain can see the events. By scanning the QR code and pulling up all the essential details, such as the manufacturer's information, the legitimacy of the acquired drugs will be seamlessly verified 44 .

6. Updated medical supply chain management

Blockchain is ideally suited for organizing and tracking the flow of medicine supply because of its security, dependability, and decentralized storage. Technology improves patient safety through building a reliable supplier network. In a single unchangeable record that's also securely held, blockchain unifies all the operations including manufacturing, packaging, marketing, shipping, and warehousing information. Blockchains adopt GS1 (open global standard for tracking healthcare products) 27 .

7. Improved electronic health record systems

Systems for keeping track of patient's health information digitally are known as electronic health records (EHRs). By connecting EHRs and distributing property of the records across all stakeholders, blockchain overcomes issues with availability, compatibility, and verification 19 .

8. Improved recruitment for clinical trials and Research

A cryptocurrency blockchain that replicates the hiring process has been developed by researchers to safeguard study participants’ anonymity while enabling access to study results for all academics 4 . Data integrity and provenance are critical characteristics in clinical trials. Blockchain network can transparently show the data from the origin to the final clinical report 27 . Technology allows researchers to access vast amounts of unprocessed data that might lead to important medical advancements without jeopardizing patient confidentiality 38 .

Threats of Blockchain Technology in Healthcare

Blockchain technology has a myriad of benefits, however, there are also considerable risks associated with the technology. Risks in this research were divided into three categories: organizational; societal; and technological threats. Scaling problems, authorization and security problems, and excessive power and energy usage were all recognized by researchers as the common three technical dangers 32 . The most important technical risk to blockchain advanced technologies is scalability. Since there is no limitation on the number of people who join the network, the scaling issue has evolved into a major worry for blockchain-based applications. Additionally, issues occur when utilizing wearable technology to track blockchain networks since the amount of data provided by these sensors grows exponentially 40 . Researchers have claimed that private permissioned blockchain deployment brings the most benefits for health care applications, however, it is usually combined with security risks 30 . Private permissioned blockchains are most prone to a 51 percent attack 37 . Additionally, blockchain is vulnerable to cyber-attacks in which the attackers can seize control of the network. If the attackers disrupt or even reverse transactions that have been validated inside the network, a disaster may result. Additionally, this evaluation identified high energy use as a hazard since it pertains to the usage of public blockchains and is a mining method that causes a lot of energy consumption. This issue got worse when more people joined the public blockchain and more payments were being processed every second.

The absence of legal authority-issued blockchain technology rules was another major societal danger highlighted. Meanwhile, interoperability problems, shortage of technical expertise for integrating pharmacological suppliers, setup expenses, and transaction costs were the main sources of organizational risks. Interoperability was seen as one of the main obstacles to blockchain technology acceptance in the healthcare industry due to lack of trust among healthcare organizations and a shortage of information technology (IT) personnel qualified to use blockchain technology. Employing blockchain technology without the necessary technical knowledge and capacity might have fatal results 8 . The included research revealed eight challenges to blockchain technology, which were categorized as organizational, societal, or technical/technological concerns. Studies discovered two different forms of social dangers, three distinct types of organizational threats, and three distinct types of technological threats. The following section provides more information on the risks explored by researchers 5 .

1. Technical or technological threats

The scalability problem with blockchain technology was due to the network's constrained processing capacity for transactions. Additionally, according to two studies, the exchange between trading volume and the amount of processing power needed to handle those transactions is the major limitation of scalability. Authorization and security were issues and constraints associated with blockchain technology. According to several studies, distributed ledger technology is vulnerable to assaults. Other research studies identified significant issues, particularly with blockchain networks, including high consumption of energy and sluggish processing speed brought by a significant increase in network users 31,39,40 .

2. Social threats

According to research studies, the societal acceptability of blockchain technology was a key obstacle to implementation. Scholars revealed that it is challenging for the legal authorities to grant access due to the decentralization of medical data and the withdrawal of a trusted third-party emphasizing privacy as a valid concern. Literature reviews also emphasized the absence of governance norms and standards as a barrier to blockchain adoption in the healthcare industry 30 .

3. Organizational threats

According to research studies, compatibility is one of the main problems with blockchain adoption in the healthcare sector from an organizational standpoint. Studies described interoperability issues as lack of confidence among parties and absence of transparent standards, which make it difficult for healthcare organizations to communicate full patient data. The upkeep of an interconnected supply chain for pharmaceuticals for the networks that lack the necessary technical knowledge to manage the system was another issue noted by research. In addition, the initial cost of installation is rather significant for blockchain, even though it can save costs in the long term 46 .

Some solutions have been proposed to address the highlighted challenges. For example, as a countermeasure to the challenge of scalability, given the large volume of clinical data involved, the trend is to store the actual healthcare data on the cloud and store only the pointers of the data on blockchain, along with their fingerprints 22 . A considerable number of papers were found on the implementation of blockchain-based EMR applications in which different strategies were considered to tackle these challenges. Yet, some publications propose different workarounds to improve the security and privacy challenges of blockchain 11,23,42 .

Blockchain as an Opportunity to Approach Medicine in a Novel Way

Blockchain is a potential solution for health data security because of its eternity, autonomy, and total openness 36 . Patients’ identity and medical information will continue to be retained in confidence using blockchain if the system is secure. By eliminating inefficient instrumentation, this ground-breaking solution will simplify the challenging billing procedure 40 . Blockchain technology may usher in a new framework for the exchange of health data by improving the efficiency, dependability, and security of EHRs as a decentralized ledger that stores important transactional data 11 . By allowing the safe transfer of patient medical records, controlling the medication supply chain, and enabling the regular and accurate of patient records, ledger technology assists healthcare scientists in deciphering genetic code. Medical files protection, diverse genomes management, electronic information management, interoperability, digitized tracking, and issue outbreak are a few of the outstanding and technologically derived aspects used to create and implement blockchain technology 3 .

Chen et al. 2019 23 noted that blockchain-based digital structures would ensure that unauthorized changes to the logistical data are avoided. They foster confidence and inhibit those who are interested in obtaining drugs from handling information, funds, and medicine in an unauthorized manner. The use of technology can significantly enhance patients’ conditions while keeping costs low. In multi-level authentication, it removes all hurdles and difficulties. Patients, physicians, and other healthcare professionals may all quickly and securely exchange the same information because of the technology's decentralized nature. Medical entities are constantly experimenting, researching, and learning about blockchain technology particularly for health records solutions. By adopting medications, enhancing payment alternatives, and decentralizing patient health history information, technology has established itself as an indispensable innovation in healthcare. The medical industry is heavily dependent on blockchain in addition to advanced technologies like machine learning and AI. There are several legitimate ways that blockchains are transforming the healthcare sector. A single blockchain system stores all the data, protecting it from loss and change. Leveraging this approach, physicians may simply get all the information required to make an accurate diagnosis and suggestions. A substantial organization with blockchain database that is encrypted may get protected from hazards and attacks from the outside world. Such rescue, assaults, and other issues, including computer malfunction or hardware breakdown, will have minimal impact on healthcare organizations appropriately deploying a blockchain network 10 .

The research studies highlighted the technology's potential to fundamentally transform the current segmentation in which patients sign fresh consents for every consultation, clinical procedure, and medical test 23 , 43 . It has the potential to become a crucial component of healthcare consent management that promotes information sharing. A blockchain-based supply chain system ensures security, reliability, and promptness of pharmaceuticals delivery. The presence of this technology solves issues that cannot be addressed by current conventional methods 32 . Reliability, protection, and data interchange among many systems are necessary for great healthcare 42 .

The research has been describing blockchain technology as a disruptive innovation.

However, blockchain research is an emerging field in healthcare, which indicates that it is mostly used for data sharing, health records, and access control along with other areas such as supply chain management or drug prescription management. Some scholars addressed other applications including the interchange of clinical testing dataset and the potential for uncovering advantages for test subjects. Technology has the potential to become a crucial component of healthcare consent management that promotes information sharing. However, much potential for blockchain is still unexploited.

A blockchain-based supply chain system ensures the security, reliability, and promptness of the delivery of pharmaceuticals. It enables the manufacturer to keep the correct formulation mixture in accordance with medical standards. Medical devices can charge for patient information, confirm that the designated patient is receiving the therapy, and communicate operational data with authorities in an anonymized manner 5 .

Recent years have seen notable advancements in medical research and enhanced medical treatments. Reliability, protection, and data interchange among many systems are necessary for a great healthcare system. Research proposed to use blockchain for building a personal health record system to bridge the gap between patient and organization 34 . Blockchain has the potential to support health records and transfer the ownership of the medical records to the patients. The use of blockchain technology in the healthcare industry is exciting. It is recommended that challenges encountered in implementing blockchain solutions should be explored in these applications. Furthermore, none of the reviewed studies described how the blockchain application was compliant with healthcare regulations, which is another area that needs to be more explored on an extended level. Also, blockchain is prone to cyber-attacks along with interoperability issues and lack of technical skills for integrating systems. In addition, high energy consumption was highlighted in this review as a threat since it relates to public blockchain use, which consumes a great amount of energy.

Limitation and Future Direction

The studies in this review describe many blockchain potential uses, benefits, and issues, often at the conceptual level. Despite the growing use of the concept in literature, there are gaps in comprehending it on empirical and theoretical levels due to the limited number of studies. However, the current and proposed studies are growing exponentially. Disruptive innovation is a term that has diffused into the healthcare industry, but there is widespread ambiguity in the use of the term 19 . Data driven studies on outcomes of specific blockchain solutions in the healthcare industry are highly recommended to pave the way for future applications. Like any emerging technology, it will introduce innovation, benefits, and risks into society. Future research is suggested to include blockchain's instrumental role in population health management and how to mitigate risks associated with utilizing the technology. Expanding healthcare research from the administrative and strategic perspectives of blockchain adoption and its economic impact on healthcare organizations will fill some gaps in the research landscape.

There is currently extremely limited research on certain applications and prototypes of blockchain solutions that would open unlimited opportunities for future research to delve into. There is also further research needed to expand on the value of blockchain uses in healthcare through developing proof of concepts to deepen researchers’ understanding of the technology in relation to healthcare system strategic needs. Future research is recommended around blockchain scalability and risk of specific blockchain cybersecurity attacks that can halt the entire system and jeopardize users’ information. Frizzo-Baker 10 discussed the argument that only 20 percent of the barriers to blockchain adoption and success are technological, while the other 80 percent are related to organizational practices. Conducting research on organizational strategies and practices in the adoption and implementation of innovative technologies in healthcare was proposed.

The purpose of this systematic review was to examine the current state and research topics of blockchain technology in healthcare, along with the applications and key benefits and challenges associated with this technology. The findings show that in the past few years blockchain has gained traction to be implemented in the healthcare sector with a potential to improve the authenticity and transparency of healthcare data, while highlighting the major challenges uncovered in this review. Blockchain's decentralization, immutability, and transparency features have enabled better management of patient health records and supply chain management. However, many healthcare organizations remain hesitant to adopt blockchain technology due to threats such as security, interoperability issues, and lack of technical skills related to blockchain technology.

The studies reviewed suggest that we are still at the beginning of the road toward the full utilization of blockchain technology in the healthcare sector. It was proposed that research be conducted on each of digital platforms discussed in the literature to identify use cases of blockchain technology and to assess its feasibility. However, doubts remain regarding the value of blockchain technology in relation to the technical experiences of users. The goal is to empower patients with the ownership of their medical data accessing and sharing. The proper utilization of blockchain can increase interoperability while maintaining privacy and security of data. Increased interoperability would be beneficial for health outcomes. However, more research still needs to be conducted to better understand and evaluate the utility of blockchain technology in healthcare. Furthermore, this paper contributes to the research on blockchain technology by highlighting current studies and identifying potential research gaps that could positively impact the industry if properly addressed.

• Reference Manager
• Simple TEXT file

People also looked at

Original research article, a novel approach toward cyberbullying with intelligent recommendations using deep learning based blockchain solution.

• 1 Faculty of Computing and Information Technology, King Abdulaziz University, Department of Information Technology, Rabigh, Saudi Arabia
• 2 Faculty of Computing and Information Technology, King Abdulaziz University, Department of Information Systems, Rabigh, Saudi Arabia

Integrating healthcare into traffic accident prevention through predictive modeling holds immense potential. Decentralized Defense presents a transformative vision for combating cyberbullying, prioritizing user privacy, fostering a safer online environment, and offering valuable insights for both healthcare and predictive modeling applications. As cyberbullying proliferates in social media, a pressing need exists for a robust and innovative solution that ensures user safety in the cyberspace. This paper aims toward introducing the approach of merging Blockchain and Federated Learning (FL), to create a decentralized AI solutions for cyberbullying. It has also used Alloy Language for formal modeling of social connections using specific declarations that are defined by the novel algorithm in the paper on two different datasets on Cyberbullying and are available online. The proposed novel method uses DBN to run established relation tests amongst the features in two phases, the first is LSTM to run tests to develop established features for the DBN layer and second is that these are run on various blocks of information of the blockchain. The performance of our proposed research is compared with the previous research and are evaluated using several metrics on creating the standard benchmarks for real world applications.

1 Introduction

Within the dynamic sphere of social media, the persistent issue of cyberbullying demands inventive and robust solutions to ensure user safety and cultivate a secure digital environment. Recent insights from the “Cyberbullying Statistics, Facts, and Trends (2023) with Charts” ( 1 ) underscore concerning statistics, revealing that over 61% of teens on social media have encountered online bullying related to their appearance, while 41% of adults have personally confronted harassment on social media. A thorough examination of cyberbullying rates among adolescents further underscores the gravity of the issue, with a study in England revealing an incidence of 17.9%, and research in Saudi Arabia reporting a prevalence of 20.97% ( 2 ). Despite recognized correlations between socio-economic factors, environmental influences, mental health, and cyberbullying tendencies, there remains an unexplored dimension—the creation of an online self-sufficient system to address cyberbullying and offer necessary guidance to identified victims and bullies.

As our digital interconnectedness expands, so too does the urgency to confront the challenges posed by malicious online behaviors. This paper proposes a novel approach to combat cyberbullying by integrating findings from cyberbullying statistics with innovative solutions. Our approach involves the fusion of two cutting-edge technologies: Blockchain and Federated Learning (FL) ( 3 ). Blockchain, known for its decentralized nature and transaction integrity, serves as the foundation of our solution, while Federated Learning facilitates collaborative machine learning without compromising individual data privacy. Alloy Language is utilized for the formal modeling of social connections, with specific declarations defined by our novel algorithm shaping the foundation of our proposed methodology. The incorporation of Long Short-Term Memory (LSTM) and Deep Belief Networks (DBN) into our system architecture enables established relational checks as well as feature detection within the DBN layer. Recognizing the importance of user accessibility, we augment our approach with an eXplainable Artificial Intelligence (XAI) layer, which sits atop our integration of Deep Learning and Blockchain technologies, making the solution more understandable to users in real-world circumstances. In the dynamic scenario of online interactions, natural language processing with AI capabilities emerges as an important aspect in the study of Cyberbullying, this plays an important role in developing useful features textual data. With the growth in usage of social media communication and utilization of day to day activities, prevalence of NLP with AI capabilities to study and analyze human interactions, innate sentiments, and discourse patterns has become increasingly relevant. The availability of vast amounts of data and the development of NLP and AI capabilities are the main drivers which cause the surge in the field of Sentiment Analysis, Tone detection etc. ( 4 ). The same is also used in fields such as information retrieval, topic modeling, sentiment analysis, and more. Cyberbullying has developed as a major issue in today’s socially connected generation, with reference to the purposeful and repetitive use of digital communication by miscreants to harass, intimidate, or hurt individuals. Cyberbullying includes a wide range of damaging activities such as spreading rumors, publishing sexual or slanderous content, sending abusive communications, and participating in online hate speech. Individuals’ mental health, social interactions, and overall well-being are all negatively impacted by cyberbullying ( 5 ).

The design is kept such that the proposed solution can be deployed using existing packaging and MLOps processes. The work explored in this document aims to contribute to the existing studies on detection and prevention of cyberbullying by proposing a novel approach and make online spaces safer. It combines three powerful technologies: federated learning, blockchain, and deep learning with natural language processing (NLP). Federated learning protects user privacy by training the cyberbullying detection model on individual devices without sharing the data itself. Blockchain ensures the security and tamper-proof nature of the training process. Deep learning and NLP enable the model to accurately identify cyberbullying content.

Through this Blackbox model powered by federated learning and NLP techniques, we develop a model that works primarily on two factors – Preservation of Social Media User Privacy and increasing the accuracy of Cyberbullying detection. The work done in this paper works in line with objective of creating safer online spaces by detecting cyberbullying and hence giving a boost to the mental health of individuals in the digital era. Our study follows a well-defined federated training sequence of various blocks, that has been developed to implement both user privacy and high-speed block chain based deep learning methods, toward cyberbullying detection.

In this paper, we have made the following contributions:

• To propose a novel framework using Blockchain and Federated Learning based Cybersecurity Solution (BFL-CS) to handle cyberbullying in social media space.

• To develop novel algorithms which works as a Hybrid Block Chain & Federated Learning model for the prevention Cyber bullying solution.

• To evaluate the proposed method with other deep learning-based methods, by using a dual layer deep learning architecture using LSTM and DBN techniques.

• To assess the effectiveness of the work using metrics and visualization tools.

The paper has been organized as follows: Section 1 discusses the Introduction and contributions made, Section 2 highlights the previous researches done in the field. Section 3 mentions the detailed proposed framework and methodology. Section 4 presents the evaluation and discussion of the results and last section concludes with some future directions.

2 Literature review

Muniyal el al. ( 3 ) introduced Federated Learning [FL] as a procedure to secure sensitive user data across the process pipeline. The authors emphasize more toward the possibility of a security breach on a Cyberbully detection and prevention system when the same is based on a Central Server. In addition to this, the performance parameters of the proposed solution is shown only on a IID (Independent and Identically Distributed) dataset only. The solution developed is named as “FedBully,” which used NLP techniques such as sentence-embedding based classifier, Sentence-BERT (Bidirectional Encoder Representations from Transformers) to detect cyberbullying, incorporating the training procedure from federated learning. Iwendi et al. ( 6 ) proposes a pure Deep Learning based solution for detection of Cyberbullying in Social Media. Advanced techniques like Bidirectional Long Short-Term Memory (BLSTM), Gated Recurrent Units (GRU), Long Short-Term Memory (LSTM), and Recurrent Neural Network (RNN) are used in ensemble to generate a higher accuracy – AOC (Area Under the Curve) for the proposed solution. In addition to that, the solution also does a significant amount of text cleaning and tokenization efforts. The paper also explores a comparative analysis of various other deep learning methods and provides a qualitative result of each method with respective accuracies and process performances. Samee et al. ( 2 ) showed detection of cyberbullying with federated learning. The work improved the identification of cyberbullying cases by offering a richer knowledge of the emotional context within communications by developing eight novel emotional elements retrieved from textual tweets. The use of privacy-preserving federated learning enabled collaborative cyberbullying detection, maintaining data privacy while encouraging collaboration across varied groups for a more scalable and successful method. Furthermore, similar to Iwendi et al. ( 2 ) where the analysis done in the paper used a client selection strategy for overall model ensemble preparation which was purely based on statistical performance of the model, the output was desired to be more accurate. The paper showed that the BERT model used in Gohal et al. ( 2 ) outperforms other traditional models such as CNN, DNN, and LSTM, that too with such low number of epochs, i.e., 200.

2.1 Research gap

Based on the literature review, we see that in previous research works on cyber-bullying detection and mitigation, a drawback that we constantly notice is the centralization of sensitive user data compared to social media for deep learning model training, highlighting a major privacy concern ( 12 ). This disadvantage may also make the adoption of such systems problematic when applied to real-world applications, as consumers will be hesitant to provide data with systems that take no precautions to safeguard their data ( 13 ). Furthermore, we show that traditional approaches frequently struggle to perform effectively due to a lack of different user behavior data and linguistic patterns. In our research, we effectively solve the above mentioned issues by combining federated learning with a secure block chain-based backend and alloy data modeling techniques. Federated learning uses a decentralized strategy to ensure that user data is handled and stored ensuring user privacy. Furthermore, the basic working of primary deep learning methods provides us with opportunity of continuous model tweaking, which, combined with other data security measures helps us in achieving our goal without giving away the third-party data security ( 7 ). Our paper uses features of federated learning to handle these shortcomings of earlier methodologies, resulting in a ground-breaking approach to cyberbully identification that maintains the highest level of user information privacy and data security.

2.2 Comparative study of systems proposed in earlier works

3 Proposed design methodology

This paper envisages novel method named Blockchain based Federated Learning based Cybersecurity Solution (BFL-CS) methodology to handle cyberbullying in social media space and its prevention ( 14 ). In the approach defined in this study, a Federated Learning methodology is employed with methods such as a modified LSTM in tandem traditional DBN to improve on the statistical parameters of the model and the privacy security of the model. The LSTM has traditional parameters such as batch size, timesteps and input feature vectors. It is to be noted that the DBN model is used as per its usual implementation without any modifications.

The proposed methodology works on two layers of memory:

1. A short-term memory (LSTM) that helps in generating blocks and federated learning nodes.

2. A long term memory (DBN) that keeps the learning from federated learning nodes and propagates it across the model during future epochs.

In this way, the model achieves faster run time due to actively forgetting information that does not value the model in the long run. And also generates highly accurate results from its long memory model implementation.

In a classical Ensemble implementation, the accuracy of two or models is combined to get a unified result. However, in our model, we have two DL models working together on the same data but at different stages to generate a result.

The architecture given below shows the complete data flow and working of the proposed design ( Figure 1 ).

Figure 1 . Overall structure of the BFL-CS model.

The framework model is listed and explained in the following steps.

3.1 Data warehousing

In our system architecture, the data is mainly collected from Social media platforms using Web Scrapping APIs. This scrapping is running on a preset scheduler to collect information at regular intervals of time and new data is added to the existing information set ( 15 ). In our model, data is stored in PostgreSQL. Currently the solution is hosted locally, however, as the complexity and size of data increases, we plan on scaling the solution toward AWS S3 with 3 AZs.

3.2 Data pre-processing

At this stage the data is made ready for ingesting into the model for obtaining desired performance. In the signal, it clears unnecessary effects, prevents issues, and improves accuracy. In this stage dataset namely the “BFL-CS dataset” and operations such as data cleaning, normalization and development of data stream is done.

3.3 Data cleaning and normalization

All the blank value fields and social media comments which clean word stems are not established are deleted from the database to prevent any kind of influence on the model due to high level outliers. Also, in order to eliminate the influences presented in the dissimilar scale features is executed in this process which reduces the model’s run time.

In many cases in the data science space, data scientist use the method of min-max normalization process. However, this method has its own problems, since this is rather a feature scaling method – this normalization significantly lowers the biasness of the model. While a lot of cases see biasness as a vice, in our case the biasness of the model actually points us toward the habitual bullies ( 16 ). Therefore, in our model we apply a rather lesser known normalization process which creates a correlation between the dataset and the standard deviation of the dataset.

3.4 Data stream for real time data publication to base database

This step involves a sophisticated integration of advanced data streaming and storage methodologies, as this step is very crucial in sensing repeated offenders and sensing their patterns. The various concepts incorporated in the model are as follows:

Event-Driven Architecture is a process that enables real-time processing by triggering and responding to events as they occur via web hooks, making it instrumental in capturing and handling data streams in real time. Kafka facilitates the building of real-time data pipelines and streaming applications. The process of collecting and importing real-time data streams into the base database for immediate storage and analysis. Utilizing messaging protocols (such as AMQP and XMPP) that minimize the time it takes for data to travel from source to destination, ensuring low-latency data delivery.

3.5 API Integration

Representational State Transfer APIs follow a set of architectural principles for designing networked applications, providing a standardized way for systems to communicate ( 17 ). Webhooks enable real-time communication between systems by triggering events in one system based on actions or updates in another, enhancing the responsiveness of API integrations. OAuth is a protocol for secure API authorization, allowing applications to access resources on behalf of a user with limited permissions. A centralized entry point that manages and optimizes API requests, ensuring scalability, security, and efficient data flow between systems. The entire design is parametric in nature without any hardcoded values. These parameters will be controlled by API driven microservices.

3.6 Data broadcaster to blockchain

At this stage a data broadcaster is developed which pushed the information to the blockchain, marrying the real-time dissemination of information with the immutable, decentralized characteristics of blockchain technology.

Key Components and Technical Processes involved at this stage. The deployment of a specialized protocol, such as DBP (hybrid ICMP & POP3), facilitates the secure and efficient real-time broadcasting of diverse data types onto a blockchain network. Decentralized Ledger Technology ensures a decentralized and distributed ledger, eliminating single points of failure and fortifying data availability across a network of nodes. The integration of a sophisticated execution engine ensures the seamless automation and enforcement of predefined rules embedded within smart contracts associated with the broadcasted data. The utilization of cryptographic hash functions, which is SHA-512 (specialized for our application), safeguards the immutability of data on the blockchain, rendering each block impervious to unauthorized modifications. The consensus algorithm, like Proof of Work (PoW) or Proof of Stake (PoS), orchestrates the agreement among network nodes, validating transactions and solidifying the security of the data broadcasting process. Blockchain’s inherent transparency provides an audit trail that allows participants to scrutinize the origin, journey, and modifications (if any) made to the broadcasted data, fostering accountability and trust. The comprehensive security architecture ensures the resilience of the data during transmission and storage, encompassing encryption, public-key infrastructure (PKI), and other robust security measures.

3.7 Blockchain administration system

This system tracks that individual changes are meticulously recorded within blocks, contributing to a transparent and tamper-resistant ledger with time & pseudo random number based identification module. The system allows for individual data entries to be added to the blockchain, with each piece of information forming a block in the distributed ledger. This decentralization eliminates the need for a central authority, enhancing transparency and reducing the risk of single points of failure ( 18 ). The heart of blockchain’s power lies in its unchangeability. Information in a block, once added, is cryptographically secured, making it virtually impossible to modify or erase. This feature guarantees the integrity of the recorded data throughout its entire existence. Every block in the blockchain is timestamped, providing an accurate record of when each data addition occurred. This temporal dimension adds another layer of transparency and traceability to the administration system. Smart contracts, self-executing contracts with predefined rules, can be incorporated to automate specific administrative functions. This enhances efficiency and reduces the need for manual intervention in routine processes. The administration of the blockchain is distributed across network nodes, eliminating the need for a centralized administrator. This decentralized governance model aligns with the principles of autonomy and inclusivity.

3.8 Deep learning engine

The deep learning engine that we are using in our architecture has two methods built in it. We first run classifications using LSTM and then we run another classification using Deep Belief Networks which then throws out the result.

Long Short-Term Memory (LSTM) is modified process of recurrent neural network (RNN) architecture designed to address the diminishing gradient situations in usual RNNs, enabling more effective modeling of sequential data. The key innovation of LSTMs lies in their memory cells, which allow them to capture and store information over long sequences.

Mathematically, as per theory, the following is to be noted in terms of LSTM model:

The base model contains of three units—the input unit i p , forget unit f , and output unit o p .

In addition to that, data state is stored in – cell state c s .

The input unit handles the process flow of new information into the cell,

The forget unit controls the retention of existing information,

and the output unit handles the knowledge to be output from the cell.

In addition to LSTMs, the model proposed in the paper also used Deep Belief Networks (DBNs) in tandem.

Here, α is a weighting parameter that determines the influence of each component on the final output. This hybrid approach aims to exploit the complementary strengths of LSTM and DBN, providing a more robust and expressive model for tasks such as sequence generation, where capturing both short-term and long-term dependencies is crucial. The choice of α allows for flexible adjustment of the contribution of each component, enabling fine-tuning based on specific task requirements and data characteristics.

3.9 Mathematical model

From the above mathematical model, we define a base algorithm on directions of which the entire architecture is built, the algorithm is as follows:

ALGORITHM 1 : Deep learning engine of BFL-CS.

3.10 Federated learning node

In the context of the proposed model on combating cyberbullying through a decentralized defense system, Federated Learning (FL) emerges as a main technology backbone of the solution. By distributing the model training process across individual devices, FL ensures that sensitive user data, integral to understanding and mitigating cyberbullying, remains localized. The use of Federated learning is used to handle separate learning activities across the data. This step has actually made the system faster by running complex algorithms across small scale datasets with limited features.

This decentralized approach mitigates privacy concerns associated with centralization, a critical consideration in the realm of cyberbullying detection. Moreover, FL’s iterative model refinement, conducted collaboratively while preserving individual data, holds significant promise in enhancing the system’s understanding of evolving cyberbullying patterns. The incorporation of FL in the proposed system aligns with the broader goal of empowering users and institutions to actively contribute to the development of robust cyberbullying detection models, fostering a collective defense against online harassment while respecting individual privacy. The Algorithms 1 , 2 for the complete model is given below:

ALGORITHM 2 : Complete BFL-CS Model.

This code implements a secure federated learning system for training a combined LSTM and DBN model. In each round, clients are chosen to participate. They receive a global model, train their local versions on their own data, and calculate updates. To protect privacy, these updates can be masked with noise or securely combined before being sent back to a central server. The server aggregates the updates and improves the global model. Finally, for tamper-proof tracking, each improved model is recorded on a blockchain ledger. This process repeats for multiple rounds, resulting in a collaboratively trained model without ever sharing the raw data from individual clients.

3.11 Result node, feedback loop: to deep learning engine and corrective data loop: to social media

In the proposed system, the culmination of federated learning, LSTM, DBN, data collection, preprocessing, and blockchain management converges at the result node ( 21 ).

This node serves as the repository for the outcomes of the intricate processes conducted during each communication round. Subsequently, these results are broadcasted into the system feedback loop, initiating a sequence of actions for system parameter optimization. The system feedback loop strategically utilizes the obtained results to refine global model parameters, enhancing the overall effectiveness of the cyberbullying detection system. Simultaneously, the results are channeled into the social media loop, triggering actions against systemic bullies. This dynamic loop interfaces with social media platforms to deploy measures aimed at curtailing cyberbullying activities. The feedback-driven optimization process and decisive actions against online aggressors collectively contribute to the robustness and adaptability of the decentralized defense system, fostering a safer and more secure online environment.

3.12 Alloy modeling

In this paper, Alloy language helps in formalizing and modeling the intricate social connections within the context of cyberbullying detection ( 22 ). Alloy, a declarative modeling language, provides a robust framework for expressing and analyzing complex relationships between entities in a system. Specifically, we employ Alloy language to create formalized declarations and constraints that define the features and dynamics of social interactions within the cyber realm ( 23 , 24 ). We construct a formal model that captures the essential features and constraints relevant to cyberbullying scenarios. This model helps in shaping the foundation of our proposed methodology, influencing the design of our novel algorithm. Alloy’s ability to articulate intricate relationships and constraints enhances the precision of our modeling efforts, contributing to the overall effectiveness of the decentralized defense system against cyberbullying.

4 Experimental results and discussions

The working of the BFL-CS method for detection and prevention of Cyberbullying in social media is tested with the Federated Deep Learning Processes which employ various methods.

The method is tested against various measures such as Recall, F1, Accuracy etc., and the results are compared with existing methods such as Vanilla RNN (v-RNN), Deep Reinforcement Learning (DRL), Residual Networks (ResNet) and Capsule Networks (CapNets). It is to be noted that the design is specifically made for English language analysis, it is seen that with appropriate data training, the results on various regional languages also show same results as shown by Pawar et al. ( 25 ) and Haider et al. ( 26 ).

4.1 Experimental setup

In this paper, the proposed methodology is implemented using Python and R. Pre-built packages are used for the implementation ( 27 ).

4.2 Programming setup parameters

In this research, Mean Squared Error (MSE) serves as the error function, while the RELU activation function is employed.

The rate of learning is set to 0.001, with a bundle size of 300 and a dropout rate of 0.1. To enhance the performance of the BFL-CS method, a Gradient-based target optimizer is applied, as illustrated in Eqs. 12– 14 , for hyperparameter optimization in this study ( 28 ). Another important aspect is that the data is purely textual in nature ( 29 ).

4.3 Dataset description

In the paper, we have utilized the dataset of Cyberbullying which is available on Kaggle by Sahane et al. ( 30 ) & KLEJ ( Kompleksowa Lista Ewaluacji Językowych ) ( 31 ) to implement the BFL-CS method for detection of Cyberbullying.

4.4 Evaluation measures

The performance of the proposed method for Cyber bullying is evaluated through evaluation statistics such as Recall, Accuracy, Specificity, F1-score, etc. ( 32 ). The performance evaluation of these metrics is based on the mathematical expressions mentioned below.

Accuracy: This is the measure that measures the efficacy of the model with respect to correct classification of data-points on Cyberbullying scope.

Precision: This is the measure that shows the overall consistency of the model and shows how many instances does the model provide accurate classifications ( 12 ).

Recall: This measure shows the number of positive values that are measured on a random basis from the total number of positive classifications feedback ( 13 ).

F1-score: This is a derived value which is the mixture of Recall and Precision – basically the Harmonic mean of both these functions ( 33 ).

Specificity: This is again a very simple measure which sort of is the opposite of precision. This is the total negative hits of the model out of the total negative values ( 34 ).

4.5 Performance analysis

The statistical performance evaluation of the proposed model for detection and prevention of Cyberbullying in social media is tested with the Federated Deep Learning Processes which employ various methods.

The BFL-CS method is evaluated with various evaluation measures against existing methods such as Vanilla RNN (v-RNN), Deep Reinforcement Learning (DRL), Residual Networks (ResNet) and Capsule Networks (CapNets) ( 27 , 35 ). From Figures 2 – 6 , the performance of various methods as mentioned above are compared with respect to the BFL-CS. It is pertinent to note that the results are with respect to the overall accuracy of detection ( 36 ).

Figure 2 . Comparison of accuracy.

Figure 3 . Comparison of precision.

Figure 4 . Comparison of recall measure.

Figure 5 . Comparison of F1-score.

Figure 6 . Graphical representation of specificity analysis.

The plot shown in Figures 2 , 3 illustrates the accuracy and precision of the BFLCS method in comparison to other known models. Impressively, BFLCS method has managed a remarkable accuracy of 98.92%. In contrast, established methods like v-RNN, DRL, ResNet, and CapNet demonstrated lower accuracies, recording values of 93.21, 96.43, 95.38, and 97.20%, respectively. Furthermore, examining precision, the BFL-CS method excels with a notable precision score of 97.91%.

We see that the system shows that it has achieved a high recall of 97.61% while the existing methods show much less recalls.

Figure 5 represents the graphical analysis to illustrate the F1-score of the BFL-CS method and the existing methods and again the superiority of the proposed solution.

The proposed methodology achieved high specificity of 97.55% while the existing methods obtained low specificity of 96.37, 95.61, 94.53, and 94.16%, respectively ( Figures 7 , 8 ).

Figure 7 . AUC-ROC plots.

Figure 8 . Comparison of run time.

In above plot shows the area under curve of the proposed methodology. The proposed solution method showed a higher ROC of 0.9812 while the existing models such as vRNN, DRL, ResNet, and CapNet obtained a low AUC-ROC of 0.9691, 0.9592, 0.9494 and 0.9576, respectively.

We have tabulated the comparison of various statistical parameters of the proposed solution and the existing models such as vRNN, DRL, ResNet and CapNet. The details of our analysis are given ( Table 1 ).

Table 1 . Tabulation of statistical performance measure of various laid down processes against the proposed methodology.

BFL-CS achieved the highest accuracy (98.92%) and AUC-ROC (0.9812), indicating that it correctly classified the most data points and has the best ability to distinguish between positive and negative classes. However, it also has the second highest computational time (16 s).

v-RNN, DRL, and ResNet all have similar performance in terms of accuracy (around 95–96%) and computational time (around 20 s). They also have good precision, recall, and F1-score, which means they are good at identifying both positive and negative cases correctly. CapNet has a slightly lower accuracy (97.2%) and AUC-ROC (0.9526) compared to the other methods, but it has the highest computational time (33 s). This suggests that CapNet may be less efficient than the other methods, even though it has a good overall performance. In addition to comparison of BFL-CS with respect to other Deep Learning models, we also compared the accuracy of other implemented solutions ( Table 2 ).

Table 2 . Comparison of technique, dataset & accuracy of previous work done on the subject.

The table suggests that models using sentence encoders (SBERT, DAN) perform well on publicly available data (Kaggle, Youtube, Twitter) and achieve high accuracy (above 97%). The model using a multi-technique approach (SVM, CNN, XGBoost) shows competitive performance on a specific dataset (VISR) ( 37 , 38 ). The BFL-CS model, which combines blockchain, federated learning, and deep learning (LSTM & DBN), achieves the highest accuracy but the data source is not specified.

5 Conclusion

The study done on the paper is a novel approach named Blockchain & Federated Learning based Cybersecurity Solution (BFL-CS) Algorithm for detection and prevention of Cyberbullying in social media. In this study, LSTM-DBN in-tandem is utilized along with block chain based federated learning. We see from our design that a major roadblock of the proposed methodology is the usage of multiple technologies in the model, therefore making it very complex for implementation, particularly in implementation of Federated Learning where two complex deep learning methods are already running, while FL is being carried out across the blocks of a real time updated ledger. This level of interconnectedness with various cutting edge technologies will required significant computational resources and strong network data transfer capabilities, however, we have tried to solve this problem by keeping only one epoch of Block-chain updation post training of data, when we increase the frequency of block updations, this approach may prove to very computationally expensive, as each updation will need a hashing process and consensus building. In the future, we should explore in making the blockchain and vanilla federated learning processes more effective. At this point of time, we have high efficacy with respect to the Deep Learning engine, however, this only contributes to only a fraction of what this approach is all about. However, handling the federated learning layer is very crucial when the size of data increases. While there has been attempts in the past at making this process more efficient, however, all of these have created compromises in the security part of it. Therefore, the future scope of work will play out in this direction. In the future scope of work, we try in developing an in-line module in one of the social networks to do a real time reporting and correction of cyberbullying online.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

AA: Conceptualization, Formal analysis, Funding acquisition, Methodology, Project administration, Writing – original draft. AM: Conceptualization, Formal analysis, Data curation, Validation, Visualization, Writing – review & editing.

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This research work was funded by Institutional Fund Projects under grant no. (IFPIP: 55-865-1442). Therefore, authors gratefully acknowledge technical and financial support from the Ministry of Education and King Abdulaziz university, DSR, Jeddah, Saudi Arabia.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

1. Djuraskovic, O. . Cyberbullying statistics, facts, and trends (2023) with charts. FirstSiteGuide. (2023). Available at: https://firstsiteguide.com/cyberbullying-stats/

Google Scholar

2. Gohal, G, Alqassim, A, Eltyeb, E, Rayyani, A, Hakami, B, Al Faqih, A, et al. Prevalence and related risks of cyberbullying and its effects on adolescent. BMC Psychiatry . (2023) 23:39. doi: 10.1186/s12888-023-04542-0

PubMed Abstract | Crossref Full Text | Google Scholar

3. Shetty, NP, Muniyal, B, Priyanshu, A, and Das, VR. FedBully: a cross-device federated approach for privacy enabled cyber bullying detection using sentence encoders. J Cyber Sec Mobil . (2023) 12:465–96. doi: 10.13052/jcsm2245-1439.1242

Crossref Full Text | Google Scholar

4. Chakraborty, K, Bhatia, S, Bhattacharyya, S, Platos, J, Bag, R, and Hassanien, AE. Sentiment analysis of COVID-19 tweets by deep learning classifiers—a study to show how popularity is affecting accuracy in social media. Appl Soft Comput . (2020) 97:106754. doi: 10.1016/j.asoc.2020.106754

5. Yosep, I, Hikmat, R, and Mardhiyah, A. Preventing cyberbullying and reducing its negative impact on students using E-parenting: a scoping review. Sustain For . (2023) 15:1752. doi: 10.3390/su15031752

6. Iwendi, C, Srivastava, G, Khan, S, and Maddikunta, PKR. Cyberbullying detection solutions based on deep learning architectures. Multimedia Systems . (2023) 29:1839–52. doi: 10.1007/s00530-020-00701-5

7. Sebastiani, F . Machine learning in automated text categorization. ACM Comput Surv . (2002) 34:1–47. doi: 10.1145/505282.505283

8. Fati, SM, Muneer, A, Alwadain, A, and Balogun, AO. Cyberbullying detection on twitter using deep learning-based attention mechanisms and continuous Bag of words feature extraction. Mathematics . (2023) 11:3567. doi: 10.3390/math11163567

9. Bruwaene, DV, Huang, Q, and Inkpen, D. A multi-platform dataset for detecting cyberbullying in social me-dia. Lang Resour Eval . (2020) 54:1–24. doi: 10.1007/s10579-020-09488-3

10. Bozyigit, A., Utku, S., and Nasiboğlu, E.. Cyberbullying detection by using artificial neural network models. 2019 4th International Conference on Computer Science and Engineering (UBMK), Samsun, Turkey. (2019).

11. Samee, NA, Khan, U, Khan, S, Jamjoom, MM, Sharif, M, and Kim, DH. Safeguarding online spaces: a powerful fusion of federated learning, word embeddings, and emotional features for cyberbullying detection. IEEE Access . (2023) 11:124524–41. doi: 10.1109/ACCESS.2023.3329347

12. Zheng, W, Lu, S, Cai, Z, Wang, R, Wang, L, and Yin, L. PAL-BERT: an improved question answering model. Comput Model Eng Sci . (2024) 139:2729–45. doi: 10.32604/cmes.2023.046692

13. Liu, X, Zhou, G, Kong, M, Yin, Z, Li, X, Yin, L, et al. Developing multi-labelled corpus of twitter short texts: a semi-automatic method. Systems . (2023) 11:390. doi: 10.3390/systems11080390

14. Liu, Z, Kong, X, Liu, S, and Yang, Z. Effects of computer-based mind mapping on students' reflection, cognitive presence, and learning outcomes in an online course. Distance Educ . (2023) 44:544–62. doi: 10.1080/01587919.2023.2226615

15. Xu, JM, Burchfiel, B, Zhu, X, and Bellmore, A. An examination of regret in bullying tweets. In Proceedings of the 2013 conference of the North American chapter of the association for computational linguistics: human language technologies (2013) pp. 697–702.

16. Dadvar, M, Trieschnigg, D, Ordelman, R, and de Jong, F. Improving cyberbullying detection with user context In: P Serdyukov, P Braslavski, SO Kuznetsov, J Kamps, S Rüger, and E Agichtein, et al., editors. ECIR 2013. LNCS , vol. 7814. Heidelberg: Springer (2013). 693–6.

17. Foong, Y. J., and Oussalah, M., "Cyberbullying system detection and analysis," 2017 European intelligence and security in-formatics conference (EISIC), Athens, Greece. (2017), pp. 40–46.

18. Poeter, D . Study: a quarter of parents say their child involved in cyberbullying. (2011). Available at: https://www.pcmag.com/article2/0,2817,2388540,00.asp

19. Salawu, S, He, Y, and Lumsden, J. Approaches to automated detection of cyberbullying: a survey. IEEE Trans Affect Comput . (2020) 11:3–24. doi: 10.1109/TAFFC.2017.2761757

20. Rosa, H, Pereira, N, Ribeiro, R, Ferreira, PC, Carvalho, JP, Oliveira, S, et al. Automatic cyberbullying detection: a systematic review. Comput Hum Behav . (2019) 93:333–45. doi: 10.1016/j.chb.2018.12.021

21. Nadali, S., Murad, M. A. A., Sharef, N. M., Mustapha, A., and Shojaee, S.. A review of cyberbullying detection: an overview. Proceedings of the 2013 13th international conference on intellient systems design and applications. Salangor, Malaysia. (2013), pp. 325–330.

22. Kim, S, Razi, A, Stringhini, G, Wisniewski, PJ, and De Choudhury, M. A human-centered systematic literature review of cyberbullying detection algorithms. Proc. ACM Hum. Comput. Interact. (2021) 5:325. doi: 10.1145/3476066

23. Potha, N., and Maragoudakis, M., Cyberbullying detection using time series modeling. Proceedings of the 2014 IEEE international conference on data mining workshop, Shenzhen, China. (2014), pp. 373–382.

24. Perera, A, and Fernando, P. Accurate cyberbullying detection and prevention on social media. Proc Comput Sci . (2021) 181:605–11. doi: 10.1016/j.procs.2021.01.207

25. Pawar, R., and Raje, R. R.. Multilingual cyberbullying detection system. Proceedings of the 2019 IEEE international conference on electro in-formation technology (EIT), Brookings, SD, USA. (2019), pp. 40–44.

26. Haidar, B., Chamoun, M., and Serhrouchni, A.. Multilingual cyberbullying detection system: detecting cyberbullying in Arabic content. Proceedings of the 2017 1st cyber security in networking conference (CSNet), Rio de Janeiro, Brazil. (2017), pp. 1–8.

27. Kargutkar, S. M., and Chitre, V.. A study of cyberbullying detection using machine learning techniques. Proceedings of the 2020 Fourth In-ternational Conference on Computing Methodologies and Communication (ICCMC), Erode, India. (2020), pp. 734–739.

28. Dinakar, K, Reichart, R, and Lieberman, H. Modeling the detection of textual cyberbullying In: Proceedings of the International AAAI Conference on Web and Social Media (2011). Vol. 5, pp. 11–17.

29. Bhatia, S, Sharma, M, Bhatia, KK, and Das, P. Opinion target extraction with sentiment analysis. Int J Comput . (2018) 17:136–42. doi: 10.47839/ijc.17.3.1033

30. Cyberbullying Dataset . (2020). Available at: https://www.kaggle.com/datasets/saurabhshahane/cyberbullying-dataset

31. KLEJ . The KLEJ benchmark (Kompleksowa Lista Ewaluacji Językowych) is a set of nine evaluation tasks for the Polish language under-standing . (2020)

32. Basheer, S, Bhatia, S, and Sakri, SB. Computational modeling of dementia prediction using deep neural network: analysis on OASIS dataset. IEEE Access . (2021) 9:42449–62. doi: 10.1109/ACCESS.2021.3066213

33. Nahar, V, Al-Maskari, S, Li, X, and Pang, C. Semi-supervised learning for cyberbullying detection in social networks In: H Wang and MA Sharaf, editors. Databases theory and applications. ADC 2014. Lecture notes in computer science . Cham: Springer (2014)

34. Liu, X, Wang, S, Lu, S, Yin, Z, Li, X, Yin, L, et al. Adapting feature selection algorithms for the classification of Chinese texts. Systems . (2023) 11:483. doi: 10.3390/systems11090483

35. Yin, D, Xue, Z, Hong, L, Davisoni, BD, Kontostathis, A, and Edwards, L. Detection of harassment on web 2.0. In Proceedings of the Content Analysis in the WEB, 2(0) (2009) p. 1–7.

36. Yang, J, Yang, K, Xiao, Z, Jiang, H, Xu, S, and Dustdar, S. Improving commute experience for private car users via blockchain-enabled multitask learning. IEEE Internet Things J . (2023) 10:21656–69. doi: 10.1109/JIOT.2023.3317639

37. Shen, J, Sheng, H, Wang, S, Cong, R, Yang, D, and Zhang, Y. Blockchain-based distributed multiagent reinforcement learning for collaborative multiobject tracking framework. IEEE Trans Comput . (2024) 73:778–88. doi: 10.1109/TC.2023.3343102

38. Rahmani, MKI, Shuaib, M, Alam, S, Siddiqui, ST, Ahmad, S, Bhatia, S, et al. Blockchain-based trust management framework for cloud computing-based internet of medical things (IoMT): a systematic review. Comput Intell Neurosci . (2022) 2022:9766844. doi: 10.1155/2022/9766844

Keywords: public health, prediction, health monitoring, blockchain, cyberbullying, federated learning, decision making

Citation: Alabdali AM and Mashat A (2024) A novel approach toward cyberbullying with intelligent recommendations using deep learning based blockchain solution. Front. Med . 11:1379211. doi: 10.3389/fmed.2024.1379211

Received: 30 January 2024; Accepted: 15 March 2024; Published: 02 April 2024.

Reviewed by:

Copyright © 2024 Alabdali and Mashat. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Arwa Mashat, [email protected]

This article is part of the Research Topic

Cluster-based Intelligent Recommendation System for Hybrid Healthcare Units

China Cyber Security Annual Conference

CNCERT 2018: Cyber Security pp 55–72 Cite as

An Overview of Blockchain Security Analysis

• Hai Wang 14 , 15 ,
• Yong Wang 16 ,
• Zigang Cao 14 , 15 ,
• Zhen Li 14 , 15 &
• Gang Xiong 14 , 15  
• Conference paper
• Open Access
• First Online: 20 February 2019

24k Accesses

29 Citations

1 Altmetric

Part of the book series: Communications in Computer and Information Science ((CCIS,volume 970))

The blockchain, with its own characteristics, has received much attention at the beginning of its birth and been applied in many fields. At the same time, however, its security issues are exposed constantly and cyber attacks have caused significant losses in it. At present, there is little concern and research in the field of network security of the blockchain. This paper introduces the applications of blockchain in various fields, systematically analyzes the security of each layer of the blockchain and possible cyber attacks, expounds the challenges brought by the blockchain to network supervision, and summarizes research progress in the protection technology. This paper is a review of the current security of the blockchain and will effectively help the development and improvement of security technologies of the blockchain.

• Network security
• Cyber attacks
• Network supervision

You have full access to this open access chapter,  Download conference paper PDF

1 Background

1.1 origin and development of the blockchain.

The first blockchain was conceptualized by a person (or group of people) known as Satoshi Nakamoto in 2008 [ 1 ]. It was implemented the following year by Nakamoto as a core component of the cryptocurrency bitcoin, where it serves as the public ledger for all transactions on the network.

Comparing to the rapid development of blockchain technology, relevant norms and standards on it are still incomplete. The first descriptive document on the blockchain is the “Bitcoin: A Peer-to-Peer Electronic Cash System” written by Nakamoto, in which blocks and chains are described as a data structure recording the historical data of the bitcoin transaction accounts. “A timestamp server works by taking a hash of a block of items to be timestamped and widely publishing the hash, such as in a newspaper or Usenet post. The timestamp proves that the data must have existed at the time, obviously, in order to get into the hash. Each timestamp includes the previous timestamp in its hash, forming a chain, with each additional timestamp reinforcing the ones before it (Fig.  1 ).” The blockchain is also called the Internet of value [ 2 ], which is a distributed ledger database for a peer-to-peer network.

The structure of blockchain.

As a rule, most innovations do not appear out of nowhere, nor does the blockchain. The blockchain is actually a natural result of that the ledger technology developed into distributed scenarios. Ledger technology has evolved from single entry bookkeeping, double-entry bookkeeping, digital bookkeeping to distributed bookkeeping. The blockchain structure (Fig.  1 ) naturally solves the problem of multiparty trust in distributed bookkeeping [ 3 ].

Due to its decentralization, tamper-resistance, safety and reliability, the block-chain technology has received extensive attention since its birth. After nearly 10 years developing, the blockchain technology has experienced the period of v1.0-bitcoin, v2.0-Ethernet and v3.0-EOS. Not only has the technology itself been greatly expanded and developed, but it has also been applied in many fields.

1.2 Blockchain Classification

According to the way users participate, blockchains can be classified into Public Blockchain, Consortium Blockchain and Private Blockchain, and also can be classified into main chains and side chains based on the relationship of chains. In addition, several blockchains can form a network. The chains in the network are interconnected in order to generate the Interchain [ 4 ].

Public Blockchain: a consensus blockchain that everyone can get an access to. He or she in the blockchain topology can send transactions and validated. Everyone can compete for billing rights. These blockchains are generally considered to be “completely decentralized”, typical use like the bitcoin blockchain, in which the information is completely disclosing.

Private Blockchain: a blockchain in which the permission to write remain in one organization. The permission to read can be public or limited to some extent. Within a company, there are additional options, such as database management, audit, and so on. In most cases, public access is not necessary.

Consortium Blockchain: in between Public Chain and Private Chain, it refers to the blockchain whose consensus process is controlled by pre-selected nodes. For example, there is a system of 15 financial institutions, each of which manages one node, and at least 10 of which must confirm each block to be recognized as valid and added to the chain. The right to read the blockchain can be open to the public, or limited by participants, or “hybrid”. Such chains can be called “partially decentralized”.

1.3 Paper Organization

At present, the blockchain has received much attention for its own characteristics, and has been applied in many fields including finance. However, there is little concern and research on its network security. Therefore, this paper introduces the birth, development and application of blockchain technology in detail, comprehensively searches and investigates various documents targeted on the security needs of blockchains, and systematically analyzes the security threats and defense technologies of blockchains.

The Sect.  2 of this paper introduces applications of the blockchain in different fields. The Sect.  3 focuses on the security threats in different layers of blockchains and summarizes common attacks. The Sect.  4 summarizes the research progress of blockchain security protection technologies. At the end of this paper, we summarize the work of the full paper.

2 Blockchain Applications

The large-scale digital currency system represented by the Bitcoin network runs autonomously for a long time, through which it supports the global real-time reliable transactions that are difficult to achieve in the traditional financial system. This has caused infinite imagination for the potential applications of the blockchain. If the business value network based on the blockchain gets real in the future, all transactions will be completed efficiently and reliably, and all signed contracts can strictly follow the agreement. This will greatly reduce the cost of running the entire business system, while sharply improving the efficiency of social communication and collaboration. In this sense, the blockchain might trigger another industrial revolution as the Internet did.

In fact, to find the right application scenario, we should proceed from the characteristics of the blockchain itself. In addition, you need to consider the reasonable boundaries of the blockchain solution. For example, blockchain applications for mass consumers need to be open, transparent, and auditable, which can be deployed on a borderless public chain or on a blockchain that is commonly maintained by multicenter nodes.

The application of blockchain in the financial services is the most concerned currently, and many banks and financial institutions around the world are the main promoters. At present, the processing after global securities trading is very complicated. The cost of liquidation is about 5–10 billion dollars. The post-trade analysis, reconciliation and processing costs exceed 20 billion dollars. According to a report by the European Central Bank [ 5 ], the blockchain, as a distributed ledger technology, can make a good deal with the cost of reconciliation and simplify the transaction process. Relative to the original transaction process, the ownership of the securities can be changed in near real time.

Blockchain can be used for ownership and copyright management and tracking. It includes transactions of valuables such as cars, houses and artworks, as well as including digital publications and digital resources that can be tagged. For example, Factom tried to use blockchain to revolutionize data management and logging in business societies and government departments. Similarly, in response to the problem of food fraud, IBM, Wal-Mart and Tsinghua University jointly announced at the end of 2016 that blockchain will be used to build a transparent and traceable cross-border food supply chain [ 6 ]. This new supply chain will improve the traceability and logistics of food and create a safer global food market.

While enjoying the convenience of cloud storage, we will inevitably mention privacy concerns. This concern comes from two aspects. One is that the storage center may be attacked by hackers, causing their own data outflow, and the second is that the company wants to get more profits to abuse the privacy of users. Blockchain solves these problems perfectly. At present, there are many distributed cloud storage projects, such as Sia, Storj, MadeSafe, and IPFS in foreign countries, and FIGTOO and GNX in China. InterPlanetary File System (IPFS) is a global, peer-to-peer distributed file system, which aims to supplement (or even replace) Hypertext Transfer Protocol (HTTP), seeks to connect all computing devices with the same file system. Replacing domain-based addresses with content-based addresses to get a faster, safer, more robust, and more durable web [ 7 ].

The relationship between FIGTOO and IPFS: IPFS is a peer-to-peer hypermedia protocol and a distributed web and FIGTOO is developed on the basis of its open source. It is a branch of IPFS, which is equivalent to bitcoin and Ethereum in the blockchain. The infrastructures are all based on the blockchain. FIGTOO creates a shared trading market for free storage space and shares global storage resources through the shared economy model. It uses red chain technology to store files in slices, builds decentralized cloud storage and becomes the infrastructure of global red chain distributed file storage [ 8 ].

User Generated Content (UGC) is one of the important aspect of blockchain application. In the era of information explosion, how to quickly find the most important content from the overloaded information has become a core issue of the Internet. UGC Network is the world’s first content value forecasting platform, a fair and value-driven content-incentive network with the mission of creating a content-driven blockchain value community that differentiates truly valuable content and achieves a reasonable return [ 9 ]. It committed to solving problems such as excellent content discovery and pricing on the UGC platform, unreasonable distribution of benefits, and centralized content storage.

Other UGC applications include YOYOW (You Own Your Own Word) - a blockchain-based UGC platform that all processes rely on interest-based implementation. It solves the problems in current content platform like lacking of high-quality content incentives, community pollution (piracy and Advertising) serious [ 10 ]. BiHu - a token investor vertical community. In the BiHu, the user’s contribution will be rewarded with the token (KEY) representing the BiHu and its surrounding ecological use rights [ 11 ].

Due to its decentralization, eliminating trust, tamper-resistance, safety and reliability characteristics, the blockchain technology has been used in lots of fields including financial services, credit and ownership management, trade management, cloud storage, user-generated content, copyright protection, advertising and games. In these cases, blockchain either solves the problems of multiparty trust in the transaction, or reduces the costs and risks of traditional industries.

3 Blockchain Security Analysis

3.1 security situation.

With the blockchain technology has been widely used, various types of attacks have emerged. Such as from the more and more digital currencies have been stolen to the exchanges have been attacked and other events. According to the statistics of the BCSEC on the blockchain attack events, about 2.1 billion dollars of economic losses due to blockchain security incidents in 2018 [ 12 ]. These are only a part of the currently exposed, and as the value of blockchain increases, the number of attacks will continue to increase (Fig.  2 ).

Economic losses caused by blockchain security incidents (ten thousand dollars).

Blockchain technology itself is still in the initial stage of rapid development, and its security is far behind the needs of development. The risks may come from attacks by external entities or internal participants. The popularity of blockchain makes new demands on security and privacy protection on data storage, transmission and applications, and puts forward new challenges to existing security solutions, authentication mechanisms, data protection, privacy protection and Information regulation.

With the current recurrence of a series of digital currency theft, hacking of exchanges, and theft of user accounts, it is urgent to establish one or more collaborative security solutions to improve the security performance of the blockchain system.

3.2 Security Analysis of Each Layer of Blockchain

The current blockchain structure can be roughly divided into application layer, smart contract layer, incentive layer, consensus layer, network layer and data layer from top to bottom. The security analysis of each layer will be performed separately below.

Application Layer. Application layer security mainly covers the security issues of centralized nodes such as the exchanges which involve digital currency transactions and manage large amounts of funds. These nodes are at any point of failure of the entire blockchain network, and the attack yield is high and the cost is low, which is the preferred target of the attackers [ 13 ].

Unauthorized Access to An Exchange Server. Exchanges often deposit large amounts of money and are easily targeted. Once the exchange server authority is obtained and the key information is modified, the attacker can steal the funds key, tamper with the transaction amount or leak sensitive information, causing economic and reputational devastating blows to the exchange.

For example, the Youbit (formerly Yapizon) stolen event. On April 22, 2017, 4 hot wallets of Youbit were stolen, lost 3,816 BTC, with a total value of about $5,300,000, accounting for 36% of the exchange’s funds. On December 19, 2017, Youbit announced that it was attacked again, lost approximately 17% of its assets, and at the same time announced the exchange closed and entered the bankruptcy process [ 14 ].

Exchange DDoS. Due to the high demand for network bandwidth in the trading platform, once a DDoS attack occurs, it is very serious for the platform and the entire industry. If the trading platform is attacked by DDoS, not only will itself suffer losses, but the transaction volume of the blockchain currency will also be greatly reduced, which will indirectly affect the rise and fall of the blockchain currency [ 15 ].

According to the report of global DDoS threat landscape Q3 2017 by Incapsula [ 16 ], although its industry scale is still relatively small, Bitcoin has become one of the top 10 industries which are most vulnerable to DDoS attacks. This reflects to a certain extent that the entire blockchain industry is facing serious DDoS security challenges. For example, from November 2017 to December 2017 Bitfinex announced that it had suffered the DDoS attack for three times, and all the services of the exchange had been shut down for a long time [ 17 ]. The attacker creates pressure on the server by creating a large number of empty accounts, causing related services and APIs to go offline for hours.

Employees Host Security. On June 20, 2011, the large Bitcoin exchange Mt.Gox was attacked. Its server was not compromised, but the attacker gained access to a computer used by an auditor of Mt.Gox, and got a read-only database file, resulting in about 60000 users’ username, email address, and encrypted password [ 18 ] to be leaked. After obtaining this sensitive information, the attacker cracked the password of one of the large accounts, issued a large sales message through this account, and sold 400,000 BTC [ 19 ] under it, trying to transfer funds through the legal transaction process. Fortunately, because the exchange protection measures are effective, it limits the maximum value of $1,000 BTC per account per day, so it does not cause much damage to this account. However, a large number of BTC sale requests caused the exchange BTC price to drop to 1 cent, resulting in an impact of approximately $8,750,000 in assets.

Malicious Program Infection. Once a malicious program is implanted into the exchange system, it is likely to cause a large amount of sensitive information leakage, including key and wallet files. The key is everything, and the leakage of sensitive information often means losing control of all assets. The exchange Mt.Gox was attacked in 2014. The key file of Mt.Gox was stored locally in clear text, and the key file wallet.dat leaked due to Trojan infection, resulting in a large amount of asset loss and eventually, Mt.Gox went bankruptcy [ 20 ]. It is worth noting that in this attack, the attacker used two years to gradually transfer assets in order to avoid the community recovering the loss through hard forks. The emergence of this type of APT attack means that monitoring of the threat of attack in the blockchain industry cannot rely solely on short-term anomaly transaction monitoring.

Initial Coin Offering. Tampering Attack: When ICO raises funds, it usually hangs the receiving address on the project official website, and then the investor will transfer money to this address for the corresponding token. Hackers can tamper with the collection address through attacks such as domain hijacking, web vulnerabilities, or social engineering.

Phishing attack: The attacker uses social engineering and other means to impersonate the official, allowing the user to transfer money to the attacker’s wallet address. For example, an attacker can use an approximate domain name and highly phishing website to defraud investors or use email to disseminate fake information, such as ICO project’s payment address change notice, etc. or disseminate phishing information on social software and media to defraud investors.

Mining Machine System. The cyber security awareness of mining device manufacturers is uneven, and because of its closed source characteristics, the security of its code cannot be checked by the public. Once a cyber security issue occurs, the result is fatal. And whether the device manufacturer will intersperse the back door for remote control of the device, or steal the mining output, is still remain to be discussed.

0day: Most mining system is a general-purpose system. Once a mining system is found to have a 0 day vulnerability, the security barriers of the system will be broken in an instant. The attacker can use the vulnerability to obtain the modify permission and then tamper with reward receiving address and then hijack the user’s reward.

Weak password attack: At present, the mining system in the market is based on the B/S architecture. Access to the mining system is usually through the web or other means. If the weak password is used, it will be vulnerable to intrusion.

Mining Pool. By June 2018, the top five Bitcoin mining pools in the world are BTC.com, AntPool, SlushPool, BTC.TOP and F2Pool. About 60% of the world’s hash power is in the hands of Chinese miners [ 21 ].

Hash power forgery attack: The mining pool will test the actual hash power of the current miner through a certain proof of work test algorithm. The hacker can falsely report the hash power by finding the vulnerability of the algorithm, and then obtain the excessive reward that doesn’t match the actual contribution.

Selfish mining attack: A malicious mining pool decides not to release the block it finds, and thus creates a fork. When the private fork is longer than the public chain, the malicious mining pool issues the private fork. Because the fork is the longest chain in the current network, it will be recognized as a legal chain by honest miners, so the original public chain and the honest data it contains will be discarded. The results of the study indicate that the malicious mining pools will yield more benefits normally by using selfish mining strategies. But such attacks usually require huge hash power as a support.

Centralization: The existence of the mining pool violates the principle of decentralization of the blockchain. Theoretically, if it can control at least 51% of the hash power of entire network, it will be able to monopolize the mining right, billing right and distribution right, which will affect the ecological security of the blockchain, so that the credit system of the cryptocurrency will cease to exist and the cryptocurrency system will be completely destroyed.

Possible Methods. It is impossible for any one party to respond to various attacks at the application layer. The application developers should ensure that the softwares don’t contain discovered vulnerabilities and are thoroughly tested. As the central node, such as a trading platform, real-time monitoring of system health and some protected methods (e.g. data encryption storage, etc.) are required to ensure that the system is not subject to internal and external attacks. All employees should be systematically trained before they are employed to avoid becoming an attack portal. As a user, you should be able to keep your own account and key properly, distinguish between true and false information and be cautious in trading to avoid phishing attacks.

Smart Contract Layer. A smart contract is more than just a computer program that can be executed automatically. It is a system participant. It responds to the received message, it can receive and store value, and it can send out information and value [ 22 ]. For the security risks of smart contracts, the following attacks are summarized.

Reentrancy Attack. The essence of reentrancy attack is to hijack the contract control flow and destroy the atomicity of the transaction, which can be understood as a logical race condition problem. For example, The DAO was attacked, and the attacker used the vulnerability in the contract to launch a reentrancy attack and gained 60 million dollars. In order to recover this part of the funds, the Ethereum community decided to perform a hard fork, roll back all the transaction records since the start of the attack and fix the contract vulnerabilities in the new branch. The vulnerability is described below. Here is a simplified version of The DAO contract:

Participants call the donate function to donate their own Ether to a contract address, the donation information is stored in the credit array, and the recipient contract calls The DAO’s withdraw function to receive funds. Before actually sending the transaction, The DAO checks if there is enough donation in the credit array, and after the transaction is over, the transaction amount is reduced from credit.

The attacker first constructs a malicious contract Mallory, as follows:

After Mallory deployed, the attacker calls The DAO’s donate function to donate a bit of Ether to the Mallory contract. After triggering Mallory’s fallback function (unnamed function), there are many trigger methods, such as transfer money to Mallory. The fallback function will call The DAO’s withdraw function and extract all the funds that belong to it. It seems to be no problem so far. However, after msg.sender.call.value(amount)() in the withdraw is executed, Mallory’s fallback function is automatically called after the transfer is completed due to the transfer operation feature, so the withdraw function is called again. Because credit is not updated at this time, so you can still withdraw money normally, then you fall into a recursive loop, and each time you can extract a part of Ether in the DAO to the Mallory contract.

This loop will continue until one of three conditions occurs, gas is exhausted, the call stack is full, and The DAO balance is insufficient. An exception is thrown when one of the above conditions occurs. Due to the characteristics of the Solidity exception handling, all previous transactions are valid. Theoretically, repeating this operation can extract all the Ether of The DAO’s to Mallory.

Unauthorized Access Attack. Most of this attack due to failure to make explicit function visibility, or fails to do sufficient permission checks, which can cause an attacker to access or modify a function or variable that should not be accessed.

For example, a multi-signature contract vulnerability in the Parity wallet was exploited by an attacker to steal a total of 153,037 Ether in three times. Then Parity official blog and Twitter released security alert [ 23 ] and updated the new version of the library contract. The bug comes from the Multi-Sig library file enhanced-wallet.sol written by Parity’s founder Gavin Wood. The attacker exploited the bug to reset the wallet owner, took over the wallet and stolen all the funds. This is essentially a breach of authority in the contract.

Solidity Development Security. Possible bugs when writing smart contracts include:

Race condition: The biggest risk of calling an external function is that the calling behavior may cause the control flow to be hijacked and accidentally modify the contract data. This type of bug has many specific forms, such as reentrant and cross-function race conditions.

Transaction-Ordering Dependence: A attacker can construct his own transaction based on the order information contained in the pending transactions, and try to get his transaction to be written into the block before others.

Integer overflow and underflow: When programming, you should think about whether integer overflows can occur, how the state of uint variables will be transferred, and who has the authority to modify those variables.

Denial of Service Attack Based on Exception Rollback: For example, a crowdfunding contract gives a refund to a participant. The contract may need to traverse an array to process a refund for a group of users. The simple idea is that every refund is successful, otherwise the program should be rolled back. The consequence of this practice is that one of the malicious users forced the refund to fail and all users were unable to receive the refund. It is recommended to use a pull payment mechanism, which separates the refund operation into an independent function, which is called by the refund recipient to pull the refund.

Possible Methods. Once a smart contract is deployed in a distributed, decentralized network, it is difficult to change. It prevents data manipulation and establishes a trust mechanism based on the encryption algorithm. On the other hand, when the blockchain is facing a security attack, it lacks an effective correction mechanism and is difficult to reverse. Therefore, before the development of smart contracts, it is necessary to guard against the vulnerabilities that have already occurred. It should conduct sufficient security tests before issued. Professionals perform code optimizations in a timely manner, conduct regular code audits, and monitor abnormal behavior of deployed contracts to reduce losses.

Incentive Layer. The purpose of the incentive layer is to provide certain incentives to encourage nodes to participate in the security verification of the blockchain. The security of the blockchain depends on the participation of many nodes. For example, the security of the Bitcoin blockchain is based on the great hash power that many nodes participate in the proof of work which makes it impossible for an attacker to provide a higher amount of computation. The verification process of a node usually consumes computing resources and electric power. In order to encourage node participation, the blockchain usually rewards participants in the form of virtual currency. Bitcoin, Litecoin, and Ether are all products of this mechanism.

Blockchain projects need to adapt to the market to automatically adjust the rewards, rather than simply reducing them. In the blockchain project reward mechanism, when the node’s working cost is close to or greater than the income, they often choose not to work for this blockchain, which can easily lead to centralization problems.

Consensus Layer. The consensus mechanism gives the blockchain the soul to differentiate it from other P2P technologies. Commonly used consensus mechanisms are Proof of Work (PoW), Proof of Stake (PoS), and Delegated Proof of Stake (DPoS). The possible attacks include Bribe Attack, Long-Rang Attack, Accumulation Attack, Precomputing Attack and Sybil Attack. Table  1 shows the application scope of the attacks for the consensus mechanisms.

At present, the existing consensus mechanisms are not perfect, and it is necessary to explore a more secure and faster consensus mechanism while increasing the difficulty of existing attacks.

Network Layer. The information transmission of the blockchain mainly depends on the peer-to-peer network. The P2P network relies on nearby nodes for information transmission in which it must expose each other’s IP. If there is an attacker in the network, it is very easy to bring security threats to other nodes. The node of the public blockchain network may be an ordinary home PC, a cloud server, etc., and its security must be uneven. There must be a node with poor security, and attacking it will directly threaten the other nodes. The main attacks are as follows.

Eclipse attack: The node is kept in an isolated network by hoarding and occupying the victim’s slots. This type of attack is designed to block the latest blockchain information from entering the eclipse node, thereby isolating the nodes [ 24 ].

BGP hijacking: At present, the security researchers have proved the conceptual feasibility of the attack. From November 5, 2015, to November 15, 2016, through the analysis and statistics of the node network, most of the bitcoin nodes are currently hosted in a few specific Internet Service Providers (ISP), while 60% of Bitcoin connections are in these ISPs. Therefore, these ISPs can see 60% of Bitcoin traffic, and can also control the traffic of the current Bitcoin network. The researchers verified that at least two attacks are conceptual feasible through the hijacking scenario, and given validation code [ 25 ].

The security defense for the network layer can be mainly improved from two aspects: P2P network security and network authentication mechanism. In the transmission process of the network, a reliable encryption algorithm is used for transmission to prevent malicious attackers from stealing or hijacking the node network. Strengthen the validity, rationality and security of data transmission in network. Client nodes should do the necessary verification for important operations and information.

Block Data. Malicious information attack: Write malicious information, such as virus signatures, politically sensitive topics, etc. in the blockchain. With the data undelete feature of the blockchain, information is difficult to delete after it is written in the blockchain. If malicious information appears in the blockchain, it will be subject to many problems.

A team of researchers at the RWTH Aachen University and the Goethe University Frankfurt in Germany pointed out that among the 1,600 documents added to the Bitcoin blockchain, 59 files contained links to illegal children’s pictures, politically sensitive content or privacy violations [ 26 ]. Currently, only a few Bitcoin blockchain transactions contain other data. In the Bitcoin blockchain, about 1.4% of the 251 million transactions contain other data, that is, only a few of these transactions contain illegal or undesirable content [ 26 ]. Still, even such small amounts of illegal or inappropriate content can put participants at risk.

Signature and Encryption Method. Cryptography is the key to ensure the security and tamper resistance of blockchain, and blockchain technology relies heavily on the research results of cryptography, which provides a key guarantee for the information integrity, authentication and non-repudiation of the blockchain.

As a mainstay of the blockchain, the encryption technology is particularly important. For example, the MD5 and SHA1 hash algorithms popular in previous years but have been proved to be insufficiently secure. At present, the SHA256 algorithm is widely used in bitcoin. So far, this algorithm is still safe, but with the development of new technology and research, it may not be safe in the future. Therefore, when designing blockchain applications, it is important to carefully choose the encryption method. Current mainstream signature methods include aggregate signature, group signature, ring signature, blind signature, proxy signature, interactive incontestable signature (IIS), blinded verifiable encrypted signature (BVES), and so on.

Attacks on cryptographic algorithms, especially the hash functions, include brute-force attack, collision attack, length expansion attack, back door attack and quantum attack.

3.3 Network Supervision of Blockchain

While blockchain brings technological innovation, it also brings huge challenges for network supervision. The traditional supervision mode is mostly centralized management. How to use the blockchain technology and the current legal system to supervise the application of the blockchain is one of the problems that the government and the industry pay attention to.

In order to overcome the problems of blockchain in network supervision, it is necessary to cross the underlying technology and think about how to combine the specific cases of technology application with supervision. At present, by classifying application cases, they can be divided into three categories, “Recycling Box”, “Dark Box” and “Sandbox” [ 27 ]. The application cases in each category bring many challenges for the legal, supervision and decision-making departments. The three categories are fully analyzed below.

3.4 “Recycling Box”

“Recycling box” are those cases that attempt to solve industry pain points through blockchain solutions in a better, faster, and cheaper way. Their goals are not illegal, and the motivation is simple. In the process of the application launched, the network supervision authorities can implement supervision only by making minor modifications to the current supervision framework.

The most typical example is the interbank settlement system developed by Ripple. The payment solution uses a single distributed ledger to connect the world’s major financial institutions and cross-bank transactions that occur between each other can be done in real time. Compared with the traditional method, it not only saves a lot of time, improves efficiency, but also saves a service fee [ 27 ].

3.5 “Dark Box”

“Dark box”, its source is similar to “dark net”. Cases belonging to this category, without exception, all contradict the current law. Such cases are numerous, for example, the online drug market, the arms market or other illegal goods market, human trafficking networks, terrorist financing and communication networks, money laundering and tax evasion can all be classified as such. These illegal services have existed in the dark network for a long time. Nowadays, because of the application of blockchain technology, some of them are like discovering the New World. It’s easy to identify the “dark box”, but it can be difficult to try to stop them [ 27 ].

The reason why the “dark box” is difficult to be stopped is that in recent years, the digital currency has become an important tool for money laundering, illegal transactions, and escaping foreign exchange control due to its anonymity and decentralization. Digital currency does not require a credit card and bank account information. Criminals can avoid the supervision agencies and cannot trace the source and destination of funds through traditional capital transaction records, which makes traditional supervision methods malfunction.

3.6 “Sandbox”

The “sandbox” is one of the most exciting and headaches for legislators in these three categories, and many of the most disruptive and public interest cases fall into this category. The term “sandbox” was taken from a recent initiative by the Financial Conduct Authority (FCA) called “Regulatory Sandbox”. Application cases belonging to this category have very valuable business objectives, but the current situation is that due to the various characteristics of the distributed ledger technology, most of these cases cannot meet the existing supervision requirements. Their common feature is what the business pursued is legal, but it may cause various risks, so the government will not let it go and will have appropriate supervision.

The typical case is peer-to-peer(P2P) funding. It is necessary to mention the venture capital fund The DAO based on the blockchain. Although The DAO’s ICO is no different from ordinary venture capital, their goals are all to invest in a startup. It seems to have nothing to do with illegality. However, the way The DAO works is not normal at all, which is one of the reasons why it will be incompatible with the existing legal system.

The DAO has no physical existence, no legal status in any jurisdiction, no leadership, management, or even employees. All operations are automatically done by the blockchain in a decentralized manner. It is not responsible to anyone except those anonymous donors. TechCrunch commented on such organizations as “completely transparent”, “shareholders have full control”, and “unparalleled flexibility and self-governance”.

At present, the skills possessed by most of the regulators are highly specialized, and they are only suitable for a certain place. The applications of blockchain are mostly global, and the coverage area is very wide. This also explains why the FCA’s proposed regulatory sandbox program has suffered a cold spot as soon as it was launched, and many blockchain startups have expressed no interest in it.

4 The Current Status of Blockchain Security Protection

Blockchain technology is currently in the early stage of development. There are many security issues from the underlying technology to the upper application. The third chapter has analyzed the vulnerabilities of each layer of the blockchain and the possible attacks. At present, when studying blockchain security, most of the scholars mainly focus on integrity, privacy protection and scalability [ 4 ].

Defenses against these attacks have been given in some papers. In the blockchain integrity protection aspect, for example, for selfish mining attacks, Eya [ 28 ] and Heilman [ 29 ] both proposed defensive measures. The existence of Proof of Work mechanism and the large number of honest miners make the blockchain integrity protected.

Although the blockchain provides anonymization, it is not completely anonymous. The attacker can still perform certain mapping by analyzing network traffic and transaction information. In the literature [ 30 , 31 , 32 ], scholars analyzed and advanced a hybrid mechanism. It’s main idea is that the user sends some bitcoin from an address and puts the bitcoin into another address in such a way that it is difficult to find the correspondence between the input and output addresses of the same user. At present, there are two main types of methods for blockchain privacy protection: One is to add an anonymous protection mechanism to an existing blockchain through a technology such as “secure transmission”. Another possible approach is to create a new blockchain that is incompatible with the Bitcoin system, such as Zerocash, which provides anonymity by using new primitives in its block [ 33 ]. In fact, some more forward-looking technologies have been studied to obtain a better anonymity guarantee, such as Coin join solutions, software that provides anonymous functionality (e.g. Mimble wimble) and next-generation encryption technology represented by attribute-based encryption.

Cryptography is the cornerstone of blockchain technology. Once the hash function or encryption algorithm is no longer secure, the security of the blockchain will no longer exist. The hash function SHA256 and the encryption algorithm elliptic curve cryptography used for the blockchain are still safe, but with the development of new technologies (e.g. quantum computing), its security remains to be discussed. Therefore, we should pay attention to new research results in a timely manner and actively seek more secure algorithms.

Blockchain technology currently has many security problems, but any innovative technology needs a process of continuous problem solving from birth to maturity, so as the blockchain. What’s more, features of the blockchain like eliminating the center, eliminating trust, and tamper-resistance, can solve problems exist in many industries.

5 Conclusion

As an emerging technology, the inherent data security and effective privacy protection make the blockchain industry be used more and more widely. However, it is worth noting that with the expansion of its application, more and more new types of security threats are emerging targeted on the blockchain. The way to strengthen the security protection of the blockchain needs further research indeed.

The second chapter of this paper introduces the application scenarios of blockchain technology in different fields and analyzes the corresponding projects. The third chapter focuses on the security analysis of the technology and application of each layer of the blockchain, and summarizes the vulnerabilities and possible attacks. The fourth chapter summarizes the current status of blockchain security protection, it shows that more research is needed on the security aspect.

According to a large number of papers have been researched, most users and researchers of the blockchain pay more attention to the application of blockchains and technology itself, but less attention and researches to security. We think blockchain anonymity research and upper-level security, especially smart contract layer and application layer security requires continuous attention and research. I hope that the work of this paper can alert the practitioner “network security of the blockchain is still waiting for deeper research”.

Nakamoto, S.: Bitcoin: a peer-to-peer electronic cash system (2008)

Google Scholar  

Zhao, G.: Blockchain: the cornerstone of the value Internet. Publishing House of Electronics Industry, Beijing (2016)

Yang, B., Chen, C.: Blockchain Principle, Design and Application. China Machine Press, Beijing (2017)

Fang, W., Zhang, W., Pan, T., et al.: Cyber security in blockchain: threats and countermeasures. J. Cyber Secur. 3 (2), 87–104 (2018)

Distributed ledger technologies in securities post-trading. https://www.ecb.europa.eu/pub/pdf/scpops/ecbop172.en.pdf . Accessed 4 July 2018

IBM News. https://www.ibm.com/news/cn/zh/2016/10/19/D468881I72849Y25.html . Accessed 4 July 2018

Benet, J.: IPFS - Content Addressed, Versioned, P2P File System. https://github.com/ipfs/papers/raw/master/ipfs-cap2pfs/ipfs-p2p-file-system.pdf . Accessed 4 July 2018

RedChain White Paper. https://cdn.thiwoo.com/RedChain/reeed_white.pdf . Accessed 4 July 2018

U Network: A Decentralized Protocol for Publishing and Valuing Online Content. https://u.network/U_whitepaper_en.pdf . Accessed 4 July 2018

YOYOW White Paper. https://yoyow.org/files/white-paper3.pdf . Accessed 4 July 2018

BIHU White Paper. https://home.bihu.com/whitePaper.pdf . Accessed 4 July 2018

BCSEC Security Trend Analysis. https://bcsec.org/analyse . Accessed 4 July 2018

CHAITIN TECH, ConsenSys.: Blockchain Security Guide. https://chaitin.cn/cn/download/blockchain_security_guide_20180507.pdf . Accessed 4 July 2018

Youbit Files for Bankruptcy After Second Hack This Year. https://www.ccn.com/south-korean-exchange-youbit-declares-bankruptcy-after-second-hack-this-year . Accessed 4 July 2018

Blockchain Security v1. https://bcsec.org/report . Accessed 4 July 2018

GLOBAL DDOS THREAT LANDSCAPE Q3 2017. https://www.incapsula.com/ddos-report/ddos-report-q3-2017.html . Accessed 4 July 2018

Bitfinex Attacked Statement. https://twitter.com/bitfinex/status/940593291208331264 . Accessed 4 July 2018

MtGox Account Database Leaked. https://news.ycombinator.com/item?id=2671612 . Accessed 4 July 2018

LulzSec Rogue Suspected of Bitcoin Hack. https://www.theguardian.com/technology/2011/jun/22/lulzsec-rogue-suspected-of-bitcoin-hack . Accessed 4 July 2018

Bitcoin Trading Platform Mt.Gox Filed for Bankruptcy Protection. http://www.bbc.com/zhongwen/simp/business/2014/02/140228_bitcoin . Accessed 4 July 2018

Pool Distribution. https://btc.com/stats/pool?pool_mode=month . Accessed 4 July 2018

Smart Contract Wiki. https://github.com/EthFans/wiki/wiki/%E6%99%BA%E8%83%BD%E5%90%88%E7%BA%A6 . Accessed 4 July 2018

Parity Security Alert. https://paritytech.io/security-alert . Accessed 4 July 2018

Heilman, E., Kendler, A., Zohar, A., et al.: Eclipse attacks on Bitcoin’s peer-to-peer network. In: Usenix Conference on Security Symposium (2015)

BGP Hijack-btc. https://github.com/nsg-ethz/hijack-btc . Accessed 4 July 2018

Matzutt, R., Hiller, J., Henze, M., et al.: A quantitative analysis of the impact of arbitrary blockchain content on bitcoin. In: 22nd International Conference on Financial Cryptography and Data Security. Springer, Curaçao (2018)

Depth Long Text Interpretation of Blockchain and Supervision: “recycling boxes”, “black boxes” and “sandboxes”. https://www.pintu360.com/a49882.html?s=87&o=1 . Accessed 4 July 2018

Eyal, I., Sirer, E.G.: Majority is not enough: bitcoin mining is vulnerable. Commun. ACM 61 (7), 95–102 (2018)

Heilman, E.: One weird trick to stop selfish miners: fresh bitcoins, a solution for the honest miner (poster abstract). In: Böhme, R., Brenner, M., Moore, T., Smith, M. (eds.) FC 2014. LNCS, vol. 8438, pp. 161–162. Springer, Heidelberg (2014). https://doi.org/10.1007/978-3-662-44774-1_12

Chapter   Google Scholar  

Valenta, L., Rowan, B.: Blindcoin: blinded, accountable mixes for bitcoin. In: Brenner, M., Christin, N., Johnson, B., Rohloff, K. (eds.) FC 2015. LNCS, vol. 8976, pp. 112–126. Springer, Heidelberg (2015). https://doi.org/10.1007/978-3-662-48051-9_9

Bissias, G., Ozisik, A.P., Levine, B.N., et al.: Sybil-resistant mixing for bitcoin. In: Proceedings of the 13th Workshop on Privacy in the Electronic Society. ACM (2015)

Meiklejohn, S., Orlandi, C.: Privacy-enhancing overlays in bitcoin. In: Brenner, M., Christin, N., Johnson, B., Rohloff, K. (eds.) FC 2015. LNCS, vol. 8976, pp. 127–141. Springer, Heidelberg (2015). https://doi.org/10.1007/978-3-662-48051-9_10

Sasson, E.B., Chiesa, A., Garman, C., et al.: Zerocash: decentralized anonymous payments from bitcoin. In: Security and Privacy, pp. 459–474. IEEE (2014)

Download references

Author information

Authors and affiliations.

Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China

Hai Wang, Zigang Cao, Zhen Li & Gang Xiong

University of Chinese Academy of Sciences, Beijing, China

National Computer Network Emergency Response Technical Team/Coordination Center, Beijing, China

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Gang Xiong .

Editor information

Editors and affiliations.

CNCERT, Beijing, China

Xiaochun Yun

Peking University, Beijing, China

Weiping Wen

Beihang University, Beijing, China

Hanbing Yan

Rights and permissions

Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.

Reprints and permissions

Copyright information

© 2019 The Author(s)

About this paper

Cite this paper.

Wang, H., Wang, Y., Cao, Z., Li, Z., Xiong, G. (2019). An Overview of Blockchain Security Analysis. In: Yun, X., et al. Cyber Security. CNCERT 2018. Communications in Computer and Information Science, vol 970. Springer, Singapore. https://doi.org/10.1007/978-981-13-6621-5_5

Download citation

DOI : https://doi.org/10.1007/978-981-13-6621-5_5

Published : 20 February 2019

Publisher Name : Springer, Singapore

Print ISBN : 978-981-13-6620-8

Online ISBN : 978-981-13-6621-5

eBook Packages : Computer Science Computer Science (R0)

Share this paper

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

The Economic Times daily newspaper is available online now.

Eu drops sovereignty requirements in cybersecurity certification scheme: document.

The European Union has struggled to agree to a cybersecurity certification scheme (EUCS) to vouch for the cybersecurity of cloud services and help governments and companies in the bloc to select a secure and trusted vendor for their business.

Elevate Your Tech Prowess with High-Value Skill Courses

Discover the stories of your interest.

Cyber-safety

Read More News on

At Suzlon, CEO Chalasani is powering profitability. But is the turnaround sustainable?

12th pass: Meet Vishesh Khurana, the ‘college dropout’ who will lead Tribe Capital’s India foray

Prime Decoder: Understanding India’s proposed move from minimum wage to living wage

Entire INR25 crore poll bonds bought by Dangi’s Authum ahead of its Reliance-ADAG deals went to BJP

Stock Radar: GSFC shows signs of trend reversal after falling 30% from highs

Stark divergence in projected returns of private and PSU banks in next one year: Will analysts again be caught on the wrong foot?

Find this comment offensive?

Choose your reason below and click on the Report button. This will alert our moderators to take action

Reason for reporting:

Your Reason has been Reported to the admin.

To post this comment you must

Log In/Connect with:

Fill in your details:

Will be displayed

Will not be displayed

Share this Comment:

Uh-oh this is an exclusive story available for selected readers only..

Worry not. You’re just a step away.

Prime Account Detected!

It seems like you're already an ETPrime member with

Login using your ET Prime credentials to enjoy all member benefits

Log out of your current logged-in account and log in again using your ET Prime credentials to enjoy all member benefits.

To read full story, subscribe to ET Prime

₹34 per week

Billed annually at ₹2499 ₹1749

Super Saver Sale - Flat 30% Off

On ET Prime Membership

Unlock this story and enjoy all members-only benefits.

Offer Exclusively For You

Save up to Rs. 700/-

ON ET PRIME MEMBERSHIP

Get 1 Year Free

With 1 and 2-Year ET prime membership

Get Flat 40% Off

Then ₹ 1749 for 1 year

ET Prime at ₹ 49 for 1 month

Stay Ahead in the New Financial Year

Get flat 20% off on ETPrime

90 Days Prime access worth Rs999 unlocked for you

Exclusive Economic Times Stories, Editorials & Expert opinion across 20+ sectors

Stock analysis. Market Research. Industry Trends on 4000+ Stocks

​Get 1 Year Complimentary Subscription of TOI+ worth Rs.799/-​

Stories you might be interested in