In this section, the research in the direction of the IoT is subdivided into nine application areas: IoT security, smart fisheries and smart agriculture, smart toys and IoT games, smart fitness, smart city monitoring, smart transportation, smart grid, smart construction projects, and smart energy (energy saving direction).
4.1. IoT Security
In the current research, privacy, confidentiality, and integrity security of data received more attention than other security requirements. Since the large amount of data generated by IoT devices is mainly processed by centralized cloud services, it is difficult to guarantee the privacy and confidentiality of these data. Wang [
22] proposed a Hyperledger (Fabric 1.1.0) based the IoT data integrity verification scheme. In
Figure 1, it is shown that the IoT data are split into multiple fragments and automatic verification and processing of device metadata is achieved and records are stored through predefined smart contracts. The cloud service provider is only responsible for returning the validation results to the user. This reduces overhead and computational costs, but lacks the design to handle more complex data types. To solve the problem of transaction security between different cloud service providers, Yang [
23] proposed a federated cloud system based on Hyperledger (Fabric 1.0). In
Figure 2, it is shown that this system is designed to determine the trusted level mechanism through user credit value instead of centralized management, and chaincode is signed between different cloud service providers for secure transactions. To a certain extent, trust is ensured and the utilization of cloud computing resources is improved. In cloud services that store datasets, there is a risk of malicious tampering and a single point of failure of the dataset model of the data owner. Dib [
24] proposed a dataset utilization system based on Hyperledger (Fabric 1.1). In
Figure 3, it is shown that the cloud service stores only the data model encrypted by the data owner, and consumers pay for the service through Hyperledger when sharing the dataset. The transparency of the dataset being utilized and the security of the dataset are enhanced, but no regulatory policy is designed for high trust level users. In supply chain systems, where data security is the first concern, Cao [
25] proposed a traceability system (Sawtooth) for the steel industry. In
Figure 4, it is shown that the data of each link is stored through a smart contract, and the regulator obtains all the circulation data through the block. Consumers scan the RFID code to obtain the final traceability information.
The automatic handling of compromised devices can avoid dangerous behaviors in time; Rodriguez [
26] proposed a Hyperledger (Fabric)-based IoT device monitoring scheme. In
Figure 5, the source and target devices are shown to verify the transaction reliability through the endorsing node in Hyperledger, and the dangerous devices are automatically isolated by chaincode. The security of device data are ensured. To solve the problems of latency and efficiency, Kim [
27] proposed a lightweight scheme combining deep learning and Hyperledger (Fabric). In
Figure 6, it is shown that the system, based on the node behavior, latitude and longitude of the network nodes, etc., and clustering, can generate multiple clusters using the clustering K-means algorithm. The system generates the corresponding chain verifier (consisting of four nodes screened) to verify the communication legitimacy and store the transaction records. The security of the data is improved to some extent. The configuration data of the IoT devices is an important part of the IoT data, and once tampered with will directly affect the original task direction. Helebrandt [
28] proposed a Hyperledger (Composer)-based configuration file system for IoT devices. In
Figure 7, it is shown that on-chain and off-chain (storing large configuration files) storage is designed to encrypt the messages that modify the configuration and load the management ID, device ID, and timestamp into a new block. However, it lacks the supervision of more configuration information, such as power, CPU utilization, and disk space. The multi-level proxy approach helps to secure the transmission of the IoT data, so Mbarekp [
29] proposed a multi-level proxy-based IoT data protection system based on Hyperledger (Fabric 1.1.0). In
Figure 8, it is shown that the validity of the blocks is verified by the check of the three level agents, which ensures the security of the data.
Since most of the key management in Hyperledger is issued and managed by government nodes, there are still security problems such as key tampering and forgery, so Ribeiro [
30] proposed a distributed key management scheme (Fabric 1.4.0). In
Figure 9, it is shown that by signing a smart contract between the device and the connection server, the system establishes a temporary session key to safeguard the device privacy, and this scheme solves the security problem of the device key to some extent.
Several studies focused on security requirements such as authentication, authorization, and billing in IoT security. Hang [
31] proposed an IoT communication platform based on Hyperledger (Fabric 1.2). In
Figure 10, it is shown that the system uses smart contracts to achieve secure access between devices, and stores data in Hyperledger to improve the security of transactions. Due to its lightweight architecture, it provides feasibility for implementing large-scale IoT device communication. To improve the reliability of smart contracts, Liu [
32] proposed a data access control system (Fabric 1.4.3). In
Figure 11, multiple users are shown to jointly develop access control policies, and the system stores records and URLs for these data through Hyperledger. This scheme reduces the pressure on on-chain storage. To address the centralized root management in top-level domain authorization, Zhang [
33] proposed a distributed root management scheme based on Hyperledger (Fabric 1.4). In
Figure 12, it is shown that the transactions for a domain authorization are sent to multiple authorization nodes, and only the authorization nodes that respond within the time threshold are considered valid. The authorization messages are counted and processed automatically by a smart contract. To improve the efficiency of authentication, Chi [
34] proposed a data co-authentication scheme (Fabric). In
Figure 13, it is shown that the user’s identity information is split into labeled data and real data. The network is divided into multiple communities according to the K-medoids algorithm [
35], and the similarity between the labeled data and the community data of the nodes is measured using the cosine similarity algorithm [
36]. Users retrieve relevant information based on tags. The efficiency of identity related data retrieval and sharing is improved.
In Hyperledger, the centralized authorization and authentication of CAs may generate risks such as tampering and forgery. To solve the problem of centralized CA authorization, Siris [
37] proposed two decentralized authorization strategies based on Hyperledger (Fabric). In
Figure 14, it is shown that multiple organizations are authorized instead of unified authorization by CA nodes, and the authorized nodes for transactions at a certain moment are filtered according to the corresponding time of the nodes. The authorization efficiency is improved while ensuring the security of distributed authorization, but the first strategy requires higher computational cost. To solve the problem of centralized CA authentication, Kakei [
38] proposed a strategy for distributed CA authentication (Fabric). In
Figure 15, it is shown that the CA nodes in Texas are divided into meta-CA and CA. The cross-authentication between meta-CA and CA determines whether this CA node is a trusted party, and this scheme improves the reliability of CA nodes to a certain extent.
To provide a generalized Hyperledger-based authorization architecture, Pajooh [
39] proposed a multilayer blockchain model (Fabric) based on a cellular system. In
Figure 16, it is shown that the network is divided into three layers based on SI (swarm intelligence) and EC (evolutionary computation) algorithms. Multiple base stations are connected in Hyperledger to achieve distributed authorization and authentication of the IoT devices. The model reduces the network load, but does not actually build a testbed.
An IoT system developed based on Hyperledger should address the security requirements for service availability, and encryption of data is one of the effective ways to ensure that data are not attacked. Zhou [
40] proposed a fully homomorphic computing scheme (Fabric) for IoT data protection. In
Figure 17, it is shown that, by encrypting the session message using a homomorphic encryption algorithm, the system verifies that the message did not change through multiple servers. It effectively protects the IoT data from attacks with good performance. Hou [
41] proposed a scheme for edge computing to protect data. In
Figure 18, it is shown that the messages of the devices are obtained through LoRa gateway and the uplink messages are stored in Hyperledger, which reduces the possibility of the messages being attacked.
In
Table 1,
Table 2 and
Table 3, this paper presents a comprehensive comparison of the above schemes. The table compares their differences in six aspects: year, consensus algorithm, incentive mechanism, application domain, issue addressed, and performance evaluation. In the performance evaluation metrics, this paper presents some of the main experimental results of the schemes.
4.2. Smart Fisheries and Smart Agriculture
Advances in information technology contributed to the digital transformation of fisheries and agriculture. At the conceptual level, smart fisheries are similar to smart agriculture. Both offer, through the deep integration of big data, blockchain, artificial intelligence and other information technology, access to real-time data collection, quantitative decision making, intelligent control, accurate investment, yield prediction, and other personalized services [
42]. Smart fisheries focus on water quality monitoring to achieve analysis and regulation of water quality in large areas. In contrast, the main need of smart agriculture is to make intelligent decisions through real-time monitoring, and analysis to improve productivity and resource efficiency [
43]. Currently, Hyperledger is less used in smart fisheries and agriculture, and the problems solved are mainly focused on data tamper resistance and real-time data flow.
It is difficult to regulate the fishery accurately, and the data are not tampered, so Hang [
44] proposed a smart fish farming platform based on Hyperledger (Fabric 1.4.3). In
Figure 19, it is shown that the actual water level data are predicted by the water level sensor, and the error is eliminated by using the Kalman filter algorithm. The system calculates the actual required water level and duration for automatic regulation. This platform provides a safer development idea for smart fisheries, but lacks interaction with different fisheries.
In a smart farm system, there are issues regarding the real-time monitoring of crops and reliability of product data. Lee [
45] proposed a middleware for monitoring the food growth environment based on Hyperledger (Sawtooth). In
Figure 20, it is shown that the crop data collected by the sensors are up-linked, and Hyperledger performs 10 cycles of authentication of the monitored data. The POET (proof of elapsed time) consensus is proven to have practical applicability with faster processing efficiency.
4.3. Smart Toys and IoT Games
While both smart toys and IoT games are intelligent entertainment services, they achieve different goals. The users of smart toys consist primarily of children, and are enabled by the integration of IT technologies to make phone calls, educate children, browse websites, as well as provide location tracking and other services. The types of smart toys available in the global market include additional mechanical toys, sound/image recognition toys, screenless toys, lifestyle toys, educational and construction games, as well as health tracking/wearable toys [
46]. However, smart toys have the problem of the inability to exchange horizontal data. This is due to how difficult it is for heterogeneous APIs to accomplish data exchange between different systems [
47], and results in a large amount of redundant data (data not needed by the user) that cannot be used effectively. IoT games break away from the traditional meaning of image and video-based games, which are games powered mainly by IoT technologies to interact with real objects in the physical world to obtain rewards. As a result, IoT games are oriented towards decentralized objects, mainly including location-based perception games. However, such games lack a robust technology to guarantee the authenticity of the tasks and the privacy of the users from being violated. Hyperledger provides an effective solution to the above problem.
In the data sharing of smart toys, horizontal data security exchange is difficult. Yang [
47] proposed a toy data exchange model based on Hyperledger (Fabric 1.0). In
Figure 21, it is shown that the toy data are desensitized and then the supplier generates a unique identifier for the toy, and Hyperledger checks and stores the toy data in Couch DB to ensure storage security.
In the Hyperledger-based IoT gaming system, regarding real-time updates of game tasks, player privacy, and reliability of game task locations, Manzoor [
48] proposed a location-aware mobile hunting game (Fabric). In
Figure 22, it is shown that the hunting task submitted by the player is validated by the smart contract, and only the reward information is posted without showing the hunting details. Players’ rewards are secured through the wallet function that stores information about completed missions in Hyperledger. This enhances the transparency and security of rewards in location-based games, but the detection of IoT beacons is largely delayed and there is no guarantee that the location of the hunt is secure. Considering the situation that some players are unable to complete hunting tasks, Pittaras [
49] developed a location-based mobile game for the interconnection of Ethereum and Hyperledger (Fabric 1.4). In
Figure 23 (since the literature does not specify the design of Ethereum), only the design related to the Hyperledger is shown. Additionally shown in
Figure 23, the system developed an advertising function and used chaincodes to count the number of times players watch the ads (advertisements) and automatically issues rewards.
4.6. Smart Grid
A smart grid is an advanced digital bi-directional tidal power system that is self-healing, adaptive, resilient, and sustainable, with the ability to predict uncertainty [
53]. Smart grids have high requirements for reliable, sustainable power supply [
54], and secure two-way power transactions are an important factor in ensuring sustainable supply. Security includes reasonable privacy protection in addition to secure storage and traceability of transactions. Hyperledger-based research is focused on addressing power transactions, privacy protection, and energy consumption load.
In a smart grid system based on Hyperledger, a real-time scheduling strategy is an important part of power trading. Zhao [
55] developed a micro grid market model based on Hyperledger (Fabric 1.1). In
Figure 26, it is shown that multiple chaincodes are used for real-time dispatching of power resources, and transaction records are stored on Hyperledger. The transaction price and volume are determined according to the Bayesian Nash equilibrium theory of incomplete information static game, which effectively reduces the purchase cost of electricity users, but cannot guarantee the systematicity in handling a large number of transactions. Li [
56] proposed a two-way electricity trading system based on Hyperledger (Fabric 1.4.0). In
Figure 27, it is shown that a real-time scheduling policy is developed for EVs through an iterative two-tier optimization-based charging and discharging policy, and chaincodes are used for scheduling transactions and clearing. The structure of hierarchical power scheduling helps to improve the scalability of the system. Considering the stability of transactions, Li [
57] proposed a power scheduling scheme (Fabric 1.4.0). In
Figure 28, it is shown that the charging/discharging schedule for electric vehicles is developed based on an optimization model with an improved krill swarm algorithm, which minimizes the load variance of the grid and thus improves the security and stability of the electricity trading of electric vehicles. In power trading, a reasonable bidding strategy helps in power dispatching. Yu [
58] proposed a power trading model based on Hyperledger (Fabric). In
Figure 29, it is shown that the best bid strategy is provided to users by improving the Bayesian bidding algorithm, including the possible bid types, the best bid, and the probability distribution of the adversaries. A three-layer structure of user layer, agent layer, and Hyperledger layer is used to ensure that detailed transaction information is not accessed by agents and Hyperledger. To address the supply chain imbalance of users due to over scheduling, Lohachab [
59] discussed a novel framework for electrical energy transactions (Fabric 1.4.0, Fabric 1.4.1). Instead of centralized microgrid scheduling of electricity, real-time scheduling of the dispatching of the Hyperledger is used. Reward algorithms and scheduling algorithms are designed to encourage users to sell excess electricity, maintain the demand balance of electricity, and guarantee the energy level of each user between minimum and maximum demand. This solution improves the utilization of electricity to a certain. extent.
To ensure the stability of energy trading in different periods, Jamil [
60] proposed a smart power trading platform by combining machine learning and Hyperledger (Fabric 1.2). In
Figure 30, it is shown that customer information is collected based on the physical network, and machine learning is used to analyze data characteristics and predict short-term and long-term scheduling transactions. The network load is effectively ensured, but a single metric is predicted. To better alleviate network congestion during power system peaks, crowdsourcing the transaction is an effective solution. Sciume [
61] proposed an energy consumption load response scheme (Fabric). In
Figure 31, it is shown that the network load reduction transaction is crowdsourced to users by predicting the next day’s network load through a data hub. The actual load capacity of each user involved in reducing the power network load is evaluated, based on the baseline, using a smart contract, and the corresponding reward task is assigned to effectively solve the network congestion caused by peak loads in the power system.
In smart power trading, regarding user privacy protection, Wang [
62] proposed an electrical energy management system. In
Figure 32, it is shown that an authentication method combining entity mapping protocol and zero-knowledge proof is used to separate user information and ensure the privacy of users.
In
Table 5, this paper provides a comprehensive comparison of smart grid schemes. The table compares their differences in six aspects: year, consensus algorithm, incentive mechanism, application domain, issued addressed, and performance evaluation. In the performance evaluation metrics, this paper presents some of the main experimental results of the schemes.
4.7. Smart Transportation
Smart traffic is the development of big data-driven intelligent traffic management solutions that harness the potential for artificial intelligence to enable effective decision making [
63]. Most of this decision making refers to the effective avoidance, mitigation of traffic congestion, and traffic accidents [
64]. Making fast and accurate regulations in the face of highly mobile and dynamic traffic situations becomes an urgent challenge to be solved. The research of Hyperledger in the field of smart transportation covers several aspects, including automatic authentication, intersection regulation and monitoring, ETC (electronic toll collection), air–land integrated authentication, and connected vehicle data security.
In a Hyperledger-based vehicle system, regarding real-time authentication, Feng [
65] proposed an automatic authentication vehicle information system based on Hyperledger (Fabric, Composer 0.20.7). In
Figure 33, it is shown that the on-board unit is used as the unique identity of the vehicle, and the roadside unit and the on-board unit are used for real-time detection and automatic authentication by chaincode. Among them, the vehicle’s identity is encrypted during the authentication process, which improves the privacy of authentication. To address cross-domain identity authentication, Li [
66] proposed a vehicle location-aware system based on Hyperledger (Fabric 1.2, Ursa). In
Figure 34, it is shown that the I-SIG system is used to obtain the data of the vehicle, providing the optimal signal scheme for the intersection. Encryption of the vehicle information was achieved using the ZKPR (zero knowledge range proofs protocol), and finally verified the legitimacy of the vehicle identity through the intelligent gateway. It has better advantages in terms of transaction latency, throughput, and success rate. Due to the limited monitoring range of roadside units, some schemes are dedicated to combining air resources for monitoring. Luo [
67] proposed an air–land integrated vehicle cross-domain identity monitoring system (Indy). In
Figure 35, it is shown that, using USRP (universal software radio peripheral) technology to provide the identity of the vehicle, the vehicle identity is authenticated by the UAV (unmanned aerial vehicle), and the legitimacy of the UAV’s identity is ensured by using the cross-authentication of neighboring UAVs. The use of airborne nodes extends the monitoring range, but the latency of authentication is high.
To address the traffic safety problem at intersections, Buzachis [
68] proposed a system for monitoring vehicles at intersections based on Hyperledger (Fabric). In
Figure 36, the trajectory of the vehicle is simulated in real time by chaincode and the endorsement node is responsible for detecting the simulation results. The usability of the system is demonstrated by testing self-driving vehicles through intersections at 1–2 intersections, but there is no design for multiple intersections. To address the problem of real-time assistance in case of vehicle hazards, Mbarek [
69] proposed a multi-level endorsement vehicle communication system (Fabric). In
Figure 37, it is shown that the BF-DF-AF-IF (belief function–desire function–analysis function–intention function) model is used to refine the vehicle’s needs into specific repair action needs. Endorsement level mechanisms are designed (according to the score obtained by the exchange, chaincode automatically upgrades or downgrades the endorsement level), and each transaction is endorsed by a higher-level endorsement node to ensure the reliability of the transaction. An intelligent endorsement mechanism is realized to enhance the efficiency of endorsement, but the scoring mechanism is not complete. To address the authenticity of accident information in telematics, Xiao [
70] proposed a telematics fake news detection model (Fabric). In
Figure 38, a Bayesian algorithm is used to detect the probability of authenticity of telematics messages and stored in Hyperledger. Load balancing is achieved and its feasibility is demonstrated in terms of prior probability, transaction processing speed, and accuracy. To enable timely access to road conditions and avoid traffic accidents, Chen [
71] proposed an edge server-based vehicle area information auction scheme. In
Figure 39, an edge server is used to divide the area and issue a request task for information reporting in a certain area. The vehicles completing the task are identified by the road side unit (RSU) technology and the authenticity information is evaluated using the expectation maximization (EM) algorithm. Suitable for low-power devices, it ensures data quality and rewards.
Several studies are devoted to the problem of secure transactions in smart transportation. Gao [
72] proposed a V2G (vehicle-to-grid) payment model based on Hyperledger (Fabric 0.6). In
Figure 40, it is shown that privacy is ensured by the ability of the payers to create multiple accounts in the same transaction. Chiu [
73] proposed an ETC system based on Hyperledger (Fabric 2.2). In
Figure 41, it is shown that the vehicle is cross-authenticated with the ETC gate, which detects the legitimacy of the vehicle’s identity and stores the transaction records in Hyperledger. It has stability and high performance, but PBFT (practical Byzantine fault tolerance) consensus is not applicable to large networks. In the toll station system, to solve the problem of electronic identity, Viera [
74] proposed a 5G-based C-V2X (vehicle-to-everything) road tolling system (Fabric). In
Figure 42, it is shown that Indy’s portable identity technology is used to send identity information through smartphones instead of RSU to obtain identity information. Toll requests are processed and transaction records are stored via cell phones. This proposal demonstrates for the first time the feasibility of combining 5G with Hyperledger in a V2X system. Lee [
75] proposed a traffic system (Fabric) based on an auction mechanism and fog computing. In
Figure 43, it is shown that fog computing is used to allocate public transportation resources, and an auction mechanism is designed to select the highest bidder for the connected vehicle user. A rational allocation of public transportation resources is achieved, but the winner is selected in a single way. In addition, the neighboring RSU nodes are secure by default, which reduces the credibility of the endorsement results.
In
Table 6 and
Table 7, this paper provides a comprehensive comparison of smart transportation schemes. The table compares their differences in six aspects: year, consensus algorithm, incentive mechanism, application domain, issue addressed, and performance evaluation. In the performance evaluation metrics, this paper presents the main experimental results of the schemes.
4.8. Smart Construction Project
In this study, smart construction projects refer to the high integration of construction projects and cutting-edge IT technologies to achieve real-time updates in building modeling, transaction security, reduced delivery costs, and effective collaboration [
76,
77]. Most of the current research focused on solving the multiparty information exchange in construction-type projects.
To address the problem of information exchange in construction projects, Suliyanti [
78] proposed a system for multiple interested parties to exchange construction information (Fabric, composter). In
Figure 44, it is shown that the system developed a system for bidding construction projects and stores the whole cycle of construction completion in Hyperledger. A complete record and exchange of building information modeling (BIM) information is explored, but the solution is owner centric, resulting in an owner’s choice without an appropriate regulatory approach. To address the issue of financial allocation in construction projects, Elghaish [
79] proposed a building information modeling system (Fabric). In
Figure 45, it is shown that the system uses smart contracts to check the financial allocations of the construction team, and allocates the appropriate finances to each participant based on the net amount of total profit, cost savings and reimbursed costs. The scheme demonstrates the feasibility of applying Hyperledger to integrated project delivery (IPD) systems. To address the issue of the privacy of different construction project ledgers, Yang [
80] discussed a multi-channel design scheme (Fabric). In
Figure 46, it is shown that smart contracts enable communication between architects, suppliers, engineers, clients, building surveyors, and urban planners, as well as store information from each of them in different channels. This scheme identified advantages unique to Hyperledger-based construction project systems in terms of scalability, traceability, and auditability features, as well as challenges in terms of transaction processing efficiency, business changes, identity, cost, and security of smart contracts. To address the problem of incomplete information for construction projects, Sheng [
81] proposed a construction project information management system based on Hyperledger (Fabric 1.4). In
Figure 47, it is shown that the system checks the authenticity of construction information through the endorser of Hyperledger, and uses orderer to sort transactions, and queries the complete construction project information through the web. To a certain extent, it solves the problem of incompleteness and difficult traceability of construction project information.
In
Table 8, this paper provides a comprehensive comparison of the smart construction project’ schemes. The table compares their differences in six aspects: year, consensus algorithm, incentive mechanism, application domain, issue addressed, and performance evaluation. In the performance evaluation metrics, this paper presents the main experimental results of the schemes.
4.9. Smart Energy (Energy Saving Direction)
Smart energy trading aims to achieve autonomous energy regulation and efficient energy use by selling their surplus energy or buying the energy they need between consumers and businesses. The concept of energy in this section may be electrical energy or carbon emissions. Smart energy trading improves the utilization of energy and reduces manual errors and management costs. Hyperledger mainly solves the problems of secure transactions and transaction integrity verification.
To address the secure scheduling of energy emissions, Yuan [
82] proposed an energy emissions trading system based on Hyperledger (Fabric 1.1). In
Figure 48, it is shown that nodes allocate emissions through specific channels and store and review energy emission transactions using smart contracts. To ensure the legitimacy of the energy emission trading identities, Hu [
83] proposed a model of distributed energy trading (Fabric). In
Figure 49, it is shown that the identity of the company and the requested emissions are verified by the endorser, and the transaction information is stored on the chain. To improve the validation efficiency of energy emission transactions, Che [
84] proposed a scheme to jointly validate energy transactions (Fabric 1.1). In
Figure 50, it is shown that a certain number of transactions are packaged and verified by the matching unit, and then re-verified and stored by the peer point on the chain. To improve the efficiency of energy dispatching, Silva [
85] proposed an electric vehicle energy bidding system. In
Figure 51, it is shown that the bidding of electric energy is designed using Hyperledger (Fabric, composter) and connected to the controller of the local parking lot for electric energy scheduling. The system uses chaincode to complete electricity transactions, which has better advantages in terms of the integrity and transparency of transactions, but the buyer is close to centralized in the transactions, and there are obvious shortcomings in regulating the buyer.
In
Table 9, this paper presents a comprehensive comparison of smart energy schemes. The table compares their differences in six aspects: year, consensus algorithm, incentive mechanism, application domain, issue addressed, and performance evaluation. In the performance evaluation metrics, this paper presents the main experimental results of the schemes.