1.1. Contributions

• An overview of Blockchain technologies, including components features, and characteristics of Blockchain, and secure application methods
• A focuses on how blockchain can be integrated into the Internet of Things (IoT) infrastructure, discussing recent methods and examples.
• A summary of blockchain application into IoT, implementation methods, and the requirements for integration.
• A discussion on Blockchain security, and how it can protect IoT from cyber-attacks, with suggestions and protective measures.
• An exploration of the most suitable methods for IoT- Blockchain integration, architectural challenges, and issues, along with an overview of consensus protocols and algorithms.
• A taxonomy detailing the security considerations and constraints involved in the IoT and blockchain integration process aiming at establishing a secure authentication framework.

1.2. Organization of the Paper

1.3. Related work

2.1. Methodology for Information Retrieval and Source Identification

2.2. Criteria for Incorporation and Exclusion

3.1. Components of Blockchain

3.2. Blockchain Technologies

• Data Layer: This layer consists of transactions based on a hash function, block, Merkle tree, and digital signature [33, 55]. The data block is divided into two parts as transaction records, organized in a Merkle tree, a binary tree structure that summarizes and securely checks content within a large data set[33.50]. Figure 6 illustrates the structure of the Merkle tree.
  Those Merkle trees are generated by hashing nodes in a kind of for-loop function until one hash is left, called the root hash. Lastly, another component of the data layer is the digital signature which is authenticated digital content that guarantees the integrity of the transaction [33, 47, 48, 50]. It is sometimes complicated to understand the architecture of IoT with Blockchain, Figure 7 visualizes the architecture of IoT and Blockchain in different layers and what are their characteristics in each layer, whereas the bottom of the figure shows the IoT and Blockchain layer and their function in the network.
• Application Layer: known for smart contrast, dApps, and chain code, this layer includes the presentation layer(scripts, user interface, APIs) and the execution layer ( smart contracts, chain code). Each transaction in the chain is run from the execution layer, which follows instructions given by the presentation layer [48, 57]. See Figure 7.
• Consensus Layer: serves a crucial function in maintaining the reliability of the Blockchain platform [50]. Consensus is a set of rules enforced by this layer itself, which each participant must follow to ensure that generation is done smoothly, making the transactions/block valid [48, 57, 58]. There are different kinds of consensus ways to guarantee Blockchain consistency, including probabilistic and deterministic methodologies. This layer ensures the reliability of the blockchain platform. Consensus rules enforced by this layer guarantee smoother transaction and block generation and validation[48, 50, 57]. Consensus methods include probabilistic and deterministic methodologies[48,50,57].
• Network Layer: Establishes communication between nodes, also known as a peer-to-peer network(P2P)[59]. This layer ensures all nodes are connected to propagate blocks through the network and synchronize the valid state of the blockchain. Figure 8 illustrates the P2P network architecture using six nodes.

Figure 6. Merkle Tree Structure.

Figure 6. Merkle Tree Structure.

Figure 7. The architecture of IoT with Blockchain.

Figure 7. The architecture of IoT with Blockchain.

Figure 8. Architecture of Peer-to-Peer Network using six nodes.

Figure 8. Architecture of Peer-to-Peer Network using six nodes.

Table 1. Types of Nodes in Blockchain Network.

3.3. Blockchain Characteristics and Features

• Decentralization: This is among the features that characterize Blockchain because it is a decentralized and distributed environment done through P2P communication between nodes as shown in Figure 10 [59, 60]. Decentralization utilizes all users’ processing power, decreasing latency and deleting the single-point failure[39, 46, 47, 51, 59]. The network participants can access all data records without being controlled by a central authority [33, 55]. Furthermore, Blockchain utilizes P2p communication, decreasing latency and eliminating single-point failure.
• Immutability: As shown in Figure 10, a key feature of Blockchain is its ability to maintain transaction integrity through immutable ledgers [51]. Unlike a central authority where data integrity is controlled and maintained by a single entity, Blockchain employs collision-free hash functions to link each block to the previous one, ensuring the integrity of each block’s content [33, 47]. Another aspect of immutability is that the blocks in the ledger cannot be altered unless all users agree to the change [39, 46, 47, 51]. Additionally, Blockchain maintains data integrity via collision-free hash functions, making blocks unmodifiable without user approval.
• Identity: The ownership of an Internet of Things device may shift throughout its life cycle, necessitating an identity management system that is both efficient and secure. There are a lot of characteristics associated with the Internet of Things devices, such as the manufacturer, GPS coordinates, serial number, and kind, all of which require an administration that is trustworthy and secure [91]. In every phase of the life of Internet of Things devices, Blockchain has the potential to be a viable solution that may help reduce the issues that have been discussed above. Through the use of a decentralized and distributed ledger, blockchain can offer approved and trustworthy identity management of linked Internet of Things devices, together with information on their intricate qualities and relationships. At every stage of the Internet of Things device’s life cycle, beginning with the manufacturer, the supplier, and the customer, it can monitor the item [90]. Overall, blockchain provides secure identity management for IoT devices through their lifecycle.
• Non-repudiation: the process of validating is done by using the private keys to utilize the signature of the transactions, which helps in confirming other participants with the equivalent public key. Therefore, each transaction that is signed cannot be refused by the transaction originator [33, 46, 47]. See Table 2 for more details about the decentralized and centralized system and their features. Furthermore, blockchain uses private key for transaction validation, confirming participant authenticity.
• Transparency: all the data encapsulated in the block can be viewed by all participants in the Blockchain, which means every user can access and interact with the Blockchain network while all users have equal rights [46, 47, 51, 56]. Overall, blockchain allows all participants to view data encapsulated in blocks.
• Traceability: all the transactions saved in the Blockchain are attached with the timestamp, which is recorded when the transaction is executed. This allows each user to quickly verify and trace the origins of historical data items after analyzing the Blockchain data with current timestamps. This enables users to trace back to the original transaction [47, 51]. Furthermore, blockchain records timestamps for each transaction, enabling users to trace data origins.
• Pseudonymity: each transaction in Blockchain uses a certain level of privacy by making Blockchain addresses anonymous[47]. Blockchain information can help identify scams and illegal transactions that may appear[33, 47]. However, Blockchain can only provide a certain level of privacy since Blockchain addresses are traceable. Overall, blockchain maintains privacy by making blockchain addresses anonymous.
• Anonymity: All the nodes in the Blockchain interact with the network using the public key that it uses to address the node on the entire network but do not know the real identities[55]. One point is that Blockchain cannot provide proper confidentiality protection because of critical limitations [47, 55]. Overall, blockchain uses public keys for network interaction without revealing real identities.
• Security: Another advantage of Blockchain technology is its provision of better protection for existing solutions [55, 56]. Using the public key infrastructure, Blockchain offers a secure environment from any attacks. The consensus component provides a trusted method that improves Blockchain security[26, 46, 47]. Overall, blockchain provides better protection through public key infrastructure and consensus mechanisms.

Table 2. A comparison between Centralized Decentralized System.

4.1. Applications of Blockchain in IoT

• Smart Manufacturing
  The manufacturing industry is changing a lot and is moving from automated manufacturing to what is called smart manufacturing [25, 26]. Data plays one of the most significant roles in this transmission. There is a vast amount of data generated from different phases of manufacturing, known as the product life cycle and supply, retail, distribution, raw materials, designing, and the sale service[25, 26, 33]. This tells us that all of these cycles provide a substantial amount of data, making data aggregation and analytics more difficult. That is why Blockchain can address the interoperability problem by connecting IoT systems using a P2P network which allows data sharing in all industrial sectors[26, 33, 48]. Blockchain also helps improve security in smart manufacturing since most IoT systems have been centralized. Nodes in Blockchain can install and import the hashes with the help of smart contracts, which are implemented by design. Automated manufacturing is often integrated with decentralized Blockchain, which provides security and confidentiality compared to the centralized system [26, 33, 47, 48]. Overall, blockchain improves security and data sharing in smart manufacturing through decentralized systems and smart contracts.
• Supply Chain
  The supply chain can be defined as a set of activities, components, and resources that must deliver a product or service to the customers [26, 33]. The final product sent across countries using different manufacturers also consists of forged components. Going through the various manufacturing processes makes those products risky[33, 48]. In this case, deploying an anti-fraud technology in the supply chain costs a lot, but the Blockchain is more affordable and can solve this problem. Blockchain and IoT affect supply chain goals such as speed, quality, costs, risk reduction, flexibility, and more [25, 26, 33, 47, 48, 67]. The trading contains a lawful contract [47] between those who buy and those who sell, with a thriving trade achievement as the product. That situation can be found in the appropriate transport technique for goods, where in this case there is a third party to check the trading procedure [47, 48]. In some cases, there will be an argument that trading the regulatory entity will solve their problems [33, 47, 48]. Overall, blockchain improves traceability and reduces fraud in the supply chain by recording each step of the product lifecycle on the blockchain.
• Food Industry
  When discussing Blockchain’s role in the food sector, we must consider the traceability of food products in ensuring food protection [47, 48]. Having the current IoT is challenging to provide food traceability within the food supply chain. Numerous providers can stipulate different food-producing organizations. Therefore, a need exists to digitize materials from sources in the manufacturing sector. Blockchain technology helps ensure the food’s origin and tracking ability [26, 33, 47]. Different sources claim that there is a need for Blockchain to establish a supply chain from farming to food manufacturers, which guarantees the food supply-chain data tractability[25, 47]. Blockchain, combined with food supply chain calls, allows consumers to track the total amount of food manufacturing procedures. The Colombian natural coffee industry is also using Blockchain technology. Also, there is encouragement by the Electronic Product Code to use IoT tags and Blockchain, which can stop information interference and confidentiality revelation [25, 26, 33, 47, 67]. Overall, blockchain ensures food traceability and safety.
• Healthcare
  Healthcare is rapidly evolving with the integration of digital technologies, which promise to revolutionize clinical data management and improve both outcomes and processes effectively [47, 48]. As one of the most pressing socioeconomic issues due to population growth [26, 68, 69], the healthcare sector faces significant challenges, particularly with limited hospital resources. One key advancement is the use of wearable devices, which enhances remote healthcare services both in clinics and at home [26, 33, 47]. For example, patients can now remain at home while using wearable devices that monitor metrics like heart rate and blood pressure [26, 33, 48]. These devices enable doctors and nurses to access healthcare data anytime and anywhere via the network. However, this data also raises concerns regarding security and privacy [26, 47, 48]. Blockchain technology offers a potential solution by ensuring the privacy and security of healthcare data stored on cloud servers and managing this data effectively [68, 69]. Medical sensor data can be collected and transmitted automatically to the system through smart contracts, enabling real-time patient monitoring [26, 33, 47, 48, 67]. By implementing these measures, Blockchain can address privacy and security issues in healthcare data management.
• Internet of Vehicles
  Internet of Vehicles integrates vehicle-to-vehicle networks, vehicle-to-infrastructure, vehicle-to-roadside networks, and vehicle-to-pedestrian networks [33, 48]. The automotive sector is leading technically superior branches by scaling from electric, hybrid, and self-driving smart cars in the Industrial Internet of Things in combination with IoT-linked cars[26, 33, 47]. Securing message transmission and execution are some of the challenges that come with the decentralization, heterogeneity, and trustworthiness in the Internet of Vehicles. If Blockchain and the Internet of Vehicles are integrated is going to solve those challenges[25, 26, 47]. Blockchain technology conserves the energy and information interactions between electric and hybrid electric vehicles using a smart grid. Unmanned Aerial vehicles today have a communication network that lacks a wireless communication network that is used to deliver product items and acquire real-time traffic flow data. Integrating Blockchain technology with the UAV network will help the confidence in UAVs[25, 26, 33, 47]. Much development on an autonomous platform based on the Ethereum Blockchain is used to provide trust management on UAVs. They are a developer working on developing a blockchain-based system to serve privacy and security for UAV data[33, 67, 70]. In general, blockchain integration will secure communication and data exchange in vehicular networks.
• Smart Grid
  Renewable Energy resources, the role of which is the energy consumers from the pure ones to prosumers who can generate energy and not just consume [33, 47]. The energy traded between consumers and prosumers is called P2P energy trading or between peers [26, 47]. With regard to energy trading, the challenge is to ensure and trust that energy trading between two trading parties in the environment has been distributed. Blockchain plays an essential role in ensuring P2P energy trading[47, 68]. One of the suggestions is to develop secure trading energy based on the blockchain consortium. This can reduce the trading cost because we do not go through a central broker via the distributed consensus of Blockchain[26, 33, 47, 67, 68]. They also received more suggestions on how to use blockchain to protect confidentiality in energy trading using decentralized smart grid systems. In general, blockchain secures P2P energy trading and protects confidentiality in decentralized smart grids.

4.2. AI in Blockchain

• Convergence of AI and Blockchain;
  The convergence of AI and blockchain [81] represents a powerful synergy that has the potential to drive significant advances in technology and various industries.
  • AI can improve several aspects of blockchain technology, such as security, supply chain, data storage, authenticity validation, healthcare, and financial services. ;
  • AI protocols can be integrated into blockchain in several formats:
    Smart contracts on blockchain platforms can be used to automate and execute AI-related tasks to manage AI training, data transactions, and payments. Table 3 presents protocols and platforms for smart contracts in AI applications.

Table 3. Protocols and Platforms for Smart Contracts in AI Applications.

Table 4. Blockchain-based AI Data Market Places for AI Applications.

4.3. Blockchain Tokenization for IoT-Enabled Smart Assets (IoT)

4.4. Non-Fungible Tokens (NFTs)

4.5. NFTs in Security and Authentication

6.1. Blockchain Integration with the IoT

Figure 15. Internet of Things infrastructures.

Figure 15. Internet of Things infrastructures.

6.2. Challenges in a Blockchain–IoT integration

Figure 16. Integration issues of IoT and Blockchain.

Figure 16. Integration issues of IoT and Blockchain.

Figure 17. Taxonomy of Consensus Mechanism.

Figure 17. Taxonomy of Consensus Mechanism.

7.1. Blockchain-Based IoT Cyberattacks

• Sybil Attack: This attack involves a malicious party controlling a blockchain network by owning multiple malicious nodes[45, 56, 58, 74]. Sybil attacks manipulate transactions or flood the network with bad transactions, causing problems. PoW makes Sybil attacks expensive to execute, requiring significant computational resources[45, 74, 77, 78]. Some blockchain networks use consensus protocols alongside PoW to prevent Sybil attacks. Furthermore, to launch this attack, the attacker should use many computational resources to produce a block. Sybil’s attack, same as PoW, is expensive to launch. The attacker must have many native cryptocurrency coins to add a new block to the Blockchain[56, 74]. When an attacker uses this kind of attack, they can compromise the entire network by manipulating a large number of virtual nodes in the network; that is the reason that they used several Blockchain which use consensus protocol in conjunction with PoW to avoid Sybil attacks[56, 58, 74, 78], in this case, PoW is conducted after every 100 blocks [34, 42, 45, 74]. Some solutions are offered to prevent Sybil’s attack[45]. An IoT trust model is proposed for each user permission Blockchain with smart contracts to evaluate the trustworthiness of IoT device identities. Sybil attacks occur due to confusion some nodes can experience after a hard fork, which often happens within insecure cryptocurrency-based protocols. One of the cryptocurrencies still vulnerable to Sybil attacks is Ethereum [42, 56, 74, 77], which still has weak restrictions on the node generation process [34, 56, 78].
• Eclipse Attack: In an eclipse attack, an attacker isolates victim nodes from the normal blockchain network by stealing routing tables and adding fake nodes as neighbors[34, 42, 56, 77]. Eclipse attacks can lead to various issues such as route fraud, storage squeeze, and denial of service. This attack is closely associated with Sybil attacks, often requiring multiple malicious nodes. Research on eclipse attacks has shown their severe impact on network topology and resource-sharing efficiency. Furthermore, when the eclipse attacks the victim node, most of its external route paths are controlled by the attacker nodes [74, 77]. When the nodes are attacked, they can be dangerous because they can take various actions, such as route fraud, storage squeeze, denial of service, and others. So based on the discussion, an eclipse attack stands as one of the most severe threats to the blockchain network [45, 74, 77]. The eclipse attack is closely associated with the Sybil attack, where multiple malicious nodes are forged. It usually needs more Sybil nodes to mount an attack[34, 42, 56, 78]. Two main studies have been done to study eclipse attacks. One introduces the eclipse attack in Bitcoins’ P2P networks. The eclipse attack destroys the topology of the networks, which reduces the number of nodes and the efficiency of resource sharing [34, 45, 74, 78]. A consequence is that the attacker hijacks all Blockchain network requests, and most of the replies they receive are falsified, and normal sharing or downloading cannot be performed[34].Figure 18.
• DDoS Attack: Distributed denial of service (DDoS) attacks are a major threat to blockchain networks, closely connected to IoT devices[34, 45, 77, 78]. In a DDoS attack, the attackers use a client/server model to combine multiple computers, amplifying the power of denial-of-service attacks. This multiplication of resources target multiple nodes simultaneously, using blockchain as a DDoS attack engine[34, 42, 56]. This causes multiplying the power of denial-of-service attacks. This happens because many concurrent online nodes (millions of them) hold many resources in storage and bandwidth, and a Blockchain node needs to keep a copy of the whole network. A Blockchain network, in this case, is used as a DDoS attack engine[34, 56, 77, 78]. DDoS attacks can be divided into active and passive attacks. Active DDoS attack works in this way.
  The attacker actively sends a large amount of false information to the network node because the subsequent visits to this information will be forwarded to the victim to achieve the effect of a DDoS attack. Passive DDoS attacks push a based mechanism in the Blockchain network protocol [34, 42, 74]. From this, there will be a lot of information within a short period, which is not easy to record and analyze. This allows IP checks to be avoided by using fake source addresses, making it difficult to track and locate attack sources [34, 42, 45, 77, 78]. Blockchain-based passive DDoS attacks passively wait for the queries from other nodes where it modifies the Blockchain client/server software, and from there, it returns an incorrect response to achieve an attack effect. The target of this attack is to deploy multiple attack nodes, which include target hosts in one response message [34, 56, 78]. This attack deploys the pull mechanism in the Blockchain network protocol.
  DDoS attacks can be facilitated by Sybil attacks and eclipse attacks [34, 45, 56, 69]. The Sybil attack’s goal is for each physical node to generate many different identities on the Blockchain network. On the other hand, DDoS attacks on a single node send many false messages to the blockchain networks or provide incorrect responses [34]. Another way to launch DDoS attacks is by using only one computer to exploit an intelligent contract repeatedly, thus congesting the network with megabytes of bytecode [42, 74, 77, 78].
• Spoofing Attack: Another meaningful attack is a spoofing attack. In a Sybil attack, the attacker attempts to create false or virtual identities, and a spoofing attack attempts to spoof the identity of a legitimate user and use those privileges to exploit [45, 58, 74, 78]. Using this identity in spoofing attack can pretend to be a legitimate user in an IoT device by using a false identity, such as using the legitimate user’s IP address or MAC address[42, 45, 77, 78]. Using this way, the attacker can gain unauthorized access to the IoT network and open doors to exploit other attacks that can seriously risk the network[45]. Table 6 shows the types of attacks and what layer those attacks have as targets.
• A replay Attack: A replay attack is an attack where a valid transmission is maliciously repeated [34, 42, 45]. A blockchain reply attack occurs when the blockchain is hard fork and a transaction on one chain is replayed on another since both transactions are valid [34, 42, 78]. Because there are two chains, their addresses are the same as the algorithm used to generate the private key, where the transaction information is the same. This results in a transaction on one of the chains that are likely to be perfectly legal on the other[34, 45, 56, 77, 78]. Like this message, a replay attack occurs because verification of the message does not certify the correctness of the message’s sending time; any messages can be selectively captured and replayed later without alteration by the attacker [45, 74, 77]. Message replay attacks are often combined with message removal attacks [45].
• Off-Off Attack: In this attack, a malicious node behaves well and poorly. This behavior attacks before the trust system becomes aware of it[42, 45, 58, 77]. On-off attacks are known as selective attacks because malicious nodes can attack the multiservice IoT architecture by acting according to the type of service they provide to other nodes in the network [34, 42, 45, 56, 78]. On and off, attackers behave differently with different neighbors to obtain contradictory trust opinions for the same node. This attack is hard to detect because it uses traditional trust management schemes. To classify a node’s behaviors requires prior trust knowledge and time. Also, not all malicious devices do misbehave [45, 77, 78].

7.2. Countermeasures

• Countermeasures against Physical Attacks: When it comes to physical attacks, a mutual authentication protocol has been proposed based on PUF (Physically Unclonable function) for small devices, which exploits the inbuilt variability of an integrated circuit [42, 44, 56, 74, 77]. Authentication is carried out using a challenge-response mechanism whose output primarily depends on the device’s physical microstructure. In this case, forging the PUF [42] to clone the same structure is impossible because it eliminates attacks like tampering and malicious code injection [42, 56, 74]. In physical attacks, a heterogeneous architecture is also proposed, which is used on a customizable and trustable device mote, which can benefit energy and performance[42]. This architecture is used with reconfigurable computing with an IEEE 802.15.3 radio transceiver and hardcore micro-controller unit [34, 42, 44, 74, 78] which is host Contiki-OS [42, 56]. Another proposed solution is named REATO [42, 44, 56] which deals with different kinds of DoS attacks in IoT devices. It is proposed that a cross-domain and flexible middleware is named NetwOrked Smart object (NOS) and tailored REATO to it [42]. The solution is said to be based on an HTTP connection request to NOS, and validation, the encrypted information is sent back [34, 42, 56].
• Countermeasures against Network Attacks: On the other hand, in-network attacks (replay and Sybil attack), there are proposed solutions to defending from the attacks[34, 42, 56]. One of the solutions offered is detecting and isolating the nodes launching Sybil attacks. It presented a trust-aware RPL routing protocol which is named SecTrust-RPL the way how it works is that it uses a mechanism based on trust to fulfill the goal [42, 56, 78]. The framework is attached to ContikiRPL. The purpose is to serve as a trust engine for making routing decisions and malicious node detection [42, 44, 56, 74, 77]. There is proposed another model of protection, which is the “signcryption technique,” which is based on Identity Based Cryptography (IBC) that can satisfy confidentiality, integrity, and authenticity[42]. This technique combines encryption and signature and deletes the need to access a trusted third party to fulfill the authentication process [34, 42, 74]. This method is mainly proposed to be applied in a replay attack. Also, in-network attacks prevent the idea of a defensive framework against network DoS and DDoS attacks, mostly related to message flooding[42, 56, 77, 78]. This is done using a DDoS server from a third party, and the algorithm for the server consists of two parts. A part analyzes the incoming traffic to decide the suspicion of danger level [42, 44, 77, 78]. This explores the suspicious activity in DoS or DDoS. Also, has been proposed SD-IoT framework which uses an algorithm to detect and mitigate DDoS attacks using cosine similarity of vectors[34, 42, 44, 56]. This works because a threshold value is obtained using the cosine similarity of the vectors of the packet-in-message rate at the SD-IoT boundary [42]. When a DDoS attack is found, a threshold value is used, and the attacker is found out and blocked at the source.
• Countermeasures against Software Attacks: When it comes to software attacks, a framework is developed to integrate three security aspects to protect trojan hardware from being affected on IoT devices [34, 42, 44]. The first is vendor diversity, which enables trusted communication between untrusted nodes. Secondly, message encryption prevents unauthorized parties from accessing contacts. Lastly, mutual auditing is allowed to allow authorized nodes to verify the encryption status and content of a message [42, 56, 77, 78]. Also, another way to prevent hardware trojans is proposed, and that way is by using high-level synthesis (HLS). Security improves the hardware produced by HLS, which is an indirect way to prevent the injection of hardware Trojans into the network [34, 42, 44].

7.3. Blockchain Privacy and Issue

Table 7. Security .

7.4. IoT Trust Issues and Their Solution by Blockchain

• Data privacy - Due to the wide-ranging integration of services and networks, the data stored on a device is susceptible to attacks by compromising nodes within connected IoT systems [33, 51]. Additionally, attackers can access the data without the owner’s authorization.
• Data integrity - Within a centralized client-server architecture, an attacker might exploit unauthorized network access to alter the original data or information before forwarding it. For example, when Alice sends data to Bob, Watson, an intermediary, could intercept and modify these data before passing them on [45, 54, 78].
• Third-party - Information gathered in a centralized setup is held and managed by an external centralized organization, which might abuse this information or share it with others [45, 58].
• Reliable data source - Within an IoT context, identifying the source of data generated by various devices is challenging because the information is stored across the entire network and can be modified by any user [34, 56, 78].
• Access management - One of the primary challenges in IoT networks is access management. Determining which node is authorized to access and execute various functions across the entire IoT network can be complex [45].
• Single points of failure - The ongoing expansion of centralized networks for the IoT infrastructure can reveal single points of failure. If a central authority stores and verifies all data of the network [58, 60, 78], the entire network becomes vulnerable if the main point fails or experiences downtime.
• Scalability - The Internet of Things interlinks numerous sensors and diverse devices for data exchange and various applications over the Internet[65, 79]. It poses a challenge to the system’s architecture and its ability to grow swiftly to scale. Figure 19 illustrates the dimensions of scalability which are divided into horizontal and vertical scalability, depicting the dimensions of blockchain scalability.

7.5. Blockchain Solutions for IoT

• Data integrity - Blockchain operates as a peer-to-peer network where each node maintains a duplicate set of records. When a transaction is initiated, the originating node uses its private key to sign the transaction and then sends it to other nodes for validation. All miner nodes take part in the validation process to find a nonce [34, 42, 65]. The first node to discover the nonce gains the right to validate the transaction and receives a reward. It will then broadcast the validated transaction to all other nodes in the network. Once the transaction is added to the Blockchain, it is immutable and cannot be altered, rolled back, or deleted [33, 51].
• Data privacy - A consortium Blockchain ensures data privacy within a Blockchain network. As depicted in Figure 3, all nodes intended for a specific purpose are grouped to create a private network or Sidechain. Each Sidechain handles its respective IoT data [34, 78]. Nodes belonging to one Sidechain do not participate in the validation process of other Sidechains. To access data on the consortium Blockchain network, a requestor node must first register and join the relevant Sidechain network, then submit an access request. Consortium Blockchain includes access control mechanisms to prevent unauthorized access [34, 42, 65, 78].
• Addressing space - Blockchain utilizes a 160-bit address, while IPv6 employs a 128-bit address scheme. Consequently, Blockchain offers 4.3 billion more addresses than IPv6, thereby enhancing the addressing capacity compared to the IPv6 addressing scheme [51].
• Trusted accountability - Each operation record is required to be logged into the Blockchain network. This process assigns an identity to every operation, making each one traceable. If any abnormal behavior is identified, it is reported back to the origin for further investigation [56, 58].
• Fault tolerance - Decentralized devices are less likely to fail accidentally because they rely on many separate components. Blockchain is a point-to-point decentralizing network. Every device has the same record copy in it, which is why a single node’s failure has not affected the network [45, 56]. So, the Blockchain prevents a single point of failure.
• Trusted data origin - To track data in a Blockchain network, a unique ID is assigned to each IoT device [56].
• Removing third-party risks - Blockchain technology empowers devices to execute operations without relying on an intermediary or third party, thereby enabling thus making them risk-free from a third party [51, 58, 78].
• Access Control: The smart contract is one of the most effective features of Ethereum, first proposed by Nick Szabo in 1994 [44]. These smart contract programs for Blockchain are designed to establish access rights and various policies. For instance, a rule might be set so that when the meter hits 135 KW, devices automatically switch to energy-saving mode [33, 51, 65]. The Internet of Things (IoT) is a rapidly advancing technology due to the growth of high-speed networks and smart devices. However, IoT devices are particularly vulnerable to attacks and lack the ability to defend themselves. This paper explores the various characteristics of Blockchain networks, such as Proof of Work (POW), decentralization, persistence, and network scalability [45, 58]. It also examines the challenges faced by IoT devices, including data integrity, access control, and privacy, while presenting Blockchain-based solutions proposed in the literature.

7.6. Consensus Protocols/Algorithms in Blockchain

• The Proof-of-Work (PoW) is computed as a mathematical problem. The nonexistence of PoW today would make it impossible to talk about Blockchain[52, 62]. PoW has been considered to be hard to computationally heavy and expensive in energy consumption[34, 52]. PoW is often considered challenging because it would not be easy to do if you want to obtain a value that can be Bitcoin or another cryptocurrency because of performed work. One of the main goals of Proof of Work is to avoid spamming attacks. PoW should be an asymmetric task, which is hard to solve but would be easy to verify. A miner needs much more time to find the nonce that solves the hash problem, where other miners in the network can easily verify and validate the solution [45, 52, 64]. One of the main problems of PoW [52, 62, 69, 70] is that multiple miners working on a standard objective lead to tremendous wastage of computing power and electricity. With a high requirement in computing power, mining has advantages if done in pools by defeating decentralization for its sound. PoW-based consensus is vulnerable when a user takes control of 51 percent of processing power in the network, as illustrated in Figure 21, [39, 45, 62].
• Proof-of-Stake (PoS) – the main concept of proof of stake is the concept itself “stake,” where the nodes participating in this consensus process lock into account a specific number of coins. The idea is to ensure that every node will act respecting the protocol rules and will not deviate from them [34, 45, 56]. As a result, users with larger stakes have a stronger incentive to protect the system’s reliability, as they risk losing more if the stake is compromised. Thus, there is less chance for a node to become malicious. It has been said that PoS protocols have a low performance. For that reason, a few other variations from Algorand are proposed on top of the Byzantine agreement [56, 62, 69]. See Figure 22 for how those transactions work.
• Algorand solves the decentralization, scalability, and security by attaching a cryptographic proof where each new block checks the eligibility of the block proposed to be chosen, which is directly proportional to its stake. Based on security and performance, Algorand can tolerate malicious behavior and provide high scalability [34, 56, 64, 70]. Also, as part of PoS, there is Delegated PoS, which does require voting to reach a consensus [52, 62, 69]. The responsibility of network management is given to delegates who are not incentivized. The duties they have include fee schedules, block intervals, as well transaction sizes. There can also be changes, and those changes can be adopted based on the network’s voting. Proof of Authority (POA) is the successor of PoS, where the reputation of the validator acts as a stake [52, 64]. When it comes to reputation, it is hard to regain it when lost; that’s why there will be a better choice for “stake. PoA networks have high throughput but are centrally controlled by the validators [34, 52, 56, 61]. Figure 23 shows the hybrid network using Proof of work and Proof of stake and how that works.
• Proof-of-Activity is introduced as an alternative to Bitcoin mining, combining elements of both proof-of-work and proof-of-stake to achieve consensus. Its primary aim is to reward stakeholders who are actively involved in the network [39, 61, 69, 70]. Additionally, Proof-of-Activity is used to ensure distributed consensus by finding proof-of-work against an empty block, with no transactions included. From there, a group of validators is selected to vote on the validity of the mined block header [39, 62].
• Proof-of-Elapsed-Time (PoET) is designed to eliminate the need for the computational power required by PoW-based consensus protocols. Implemented in the Sawtooth platform, this protocol addresses the Byzantine agreement problem by using a lottery-like approach to ensure fairness, investment, and verification during leader election [39, 45, 56, 70]. In PoET, peers wait for a random amount of time to elapse, and the peer that finishes waiting first is selected as the leader to create a new block [56, 62, 64]. This process takes place in a secure memory area, such as Intel’s SGX, which is commonly used in the Sawtooth platform [56, 61].
• Proof-of-Capacity (PoC) is an alternative to PoW that leverages miners’ hard drive space instead of computational power to solve cryptographic challenges. In PoC, miners engage in a process called "plotting," where they store various solutions to problems in advance, removing the need to solve them on the spot. This method is more energy-efficient compared to PoW [45, 52, 56, 70]. However, there is a possibility that multiple users could collude to combine their storage power within a centralized network.
• The Kafka protocol utilizes a shared subscribe messaging pattern capable of transferring large volumes of log data with minimal latency. It involves producers, topics, consumers, and brokers. Producers publish recorded information as a stream of messages, which are segments of partitioned files [56, 61, 64, 69]. These messages are stored by brokers in the latest segment file, and subscribed consumers can read them by requesting access from the brokers [56, 62]. Kafka uses a Crash Fault Tolerant (CFT) consensus protocol, which can handle up to 50 percent of network failures, and is primarily implemented in fabric systems [45, 70].
• The Practical Byzantine Fault Tolerance (PBFT) protocol is a widely recognized protocol that has gained increased attention with the rise of Blockchain technology. PBFT operates under the assumption that less than 33 percent of network nodes behave maliciously. Consensus is achieved through three phases: pre-prepare, prepare, and commit [56, 61, 69]. In this process, each node acts as a validating replica that votes to elect a primary node, or leader, which initiates the three-phase consensus after receiving a request from clients [34, 39, 56, 70], starting by multicasting a pre-prepared message. Despite its effectiveness, PBFT lacks scalability, supporting only a limited number of nodes and requiring the transfer of numerous messages to reach consensus [56]. The Ripple Protocol Consensus Algorithm (RPCA) was designed to ensure security and stability within a cryptocurrency-based network for remittance transfers, without the need to implement smart contracts [62, 64]. Each node in RPCA maintains a unique node list, which is a set of trusted validators involved in the consensus process [34, 45, 56]. Nodes listen to these trusted validators, and if consensus is not achieved on a set of transactions, the node’s proposals are adjusted according to suggestions from their trusted validators [56, 61, 69, 70]. Transactions that receive more than 80 percent positive votes are processed, while others are either discarded or placed in a candidate pool for future ledger-inclusion [56, 64].
• The Stellar Consensus Protocol (SCP) is designed as a Federated Byzantine Agreement (FBA). Unlike other protocols, SCP does not require consensus from a threshold of all nodes in the network; instead, each node forms a subset of trusted nodes within the network [45, 52]. Nodes make decisions based on the consensus of their trusted circle, and any misbehaving nodes are excluded from both the trusted circle and the decision-making process. In Hyperledger Fabric, where PBFT is used, network nodes validate each transaction, and a leader is elected to propose the transaction sequence. In the Delegated Byzantine Fault Tolerant (DBFT) protocol, specific nodes are chosen to reach consensus on the next block to be added to the Blockchain [52, 61, 69, 70]. These nodes, known as bookkeepers, continuously vote and are selected through a registration process in the network [34, 52]. The table below provides a detailed summary of various consensus algorithms.

7.7. Taxonomy of Security Research in IoT