Journal of information and
communication convergence engineering
(J. Inf. Commun. Converg. Eng.)

The exponential growth of the Internet of Things (IoT) has resulted in the networking of diverse devices, enabling the connectivity of billions of devices and consequently generating an unparalleled volume of data [1]. These data encompass personal privacy, industrial control, smart-home systems, healthcare, and other related domains. The conventional approach to centralized data storage and management is inadequate for addressing the increasing need for data processing, primarily because of issues such as data silos, significant delays, vulnerability to system failures, and security hazards. Consequently, the development of a novel data-management model that is both more efficient and safer has emerged as a prominent area of current research [2].

The decentralization, transparency, and immutability of blockchain technology make it a universally recognized optimal solution for addressing the issues of managing IoT data. An InterPlanetary File System (IPFS) is a decentralized file system created to construct a more efficient, secure, and accessible Internet. IPFS enhances the effectiveness and dependability of data storage by utilizing content addressing to store files and distribute data across numerous nodes. The integration of blockchain and IPFS allows for effective decentralized data management [3]. Sultana et al. [4] presented a conceptual framework that utilized IPFS and blockchain to address certificate fraud in the academic domain and guarantee data authenticity. Razzaq et al. [5] presented a decentralized system for exchanging IoT data. This framework deploys contracts on the Ethernet blockchain and combines them with IPFS to ensure data security and privacy. Sangeeta et al. [6] developed a decentralized application (DApp) that utilizes IPFS and Ethernet blockchain technology to store CCTV and black-box videos securely, ensuring data transparency and integrity.

This paper presents an innovative methodology for managing IoT data by utilizing federated blockchain and IPFS technologies. This approach seeks to tackle the issue of managing the vast volume of data produced by IoT devices while ensuring data security, integrity, and efficient management. The primary goal of the proposed approach is to establish a datamanagement framework for IoT systems that is both secure and efficient while also being capable of scaling as needed. Utilizing IPFS, a distributed and decentralized data-storage system, leads to reduced storage expenses and enhanced access velocities.

The proposed architecture incorporates robust encryption algorithms such as AES-256 to encrypt data during storage and transport. The model facilitates the incorporation of various IoT devices and communication protocols, rendering it a versatile solution for diverse applications including smart homes, healthcare, and industrial IoT.

A. Status of Blockchain Processing Massive IoT Data

Significant advancements have been achieved in applying blockchain technology to enhance the security and privacy protection of IoT data. The tamperability and encryption properties of blockchain ensure excellent data protection during transmission and storage. For certain applications, such as smart-home systems, data produced by devices are encrypted and stored using blockchain technology to guarantee user privacy [7]. The combination of blockchain and distributed storage systems, such as IPFS, offers decentralized options for storing data. This method not only mitigates the potential for a singular point of failure, but also enhances the efficiency of data retrieval. Environmental monitoring data in smart city applications, such as air and water quality monitoring, are saved using IPFS [8].

Blockchain technology enhances the transparency and traceability of IoT data by recording all data actions in the chain. This enables users and regulators to track the origin of data easily and access operational records whenever needed. Food transport and storage data collected by sensors in the food supply chain are documented using blockchain technologies. This process guarantees food traceability and improves food safety [9].

Blockchain smart contract technology enables the automation of IoT data processing. Smart contracts autonomously execute processes according to predetermined circumstances, thereby significantly enhancing the efficiency and precision of a system. Smart contracts in energy management systems can autonomously modify power supply plans by analyzing energy consumption data and enhancing energy utilization efficiency [10].

Although blockchain holds significant promise for IoT data management, it faces numerous challenges [11]. First, the processing capacity and scalability of the blockchain must be enhanced to satisfy the requirements for handling large amounts of data in IoT. Furthermore, in the context of IoT, safeguarding data privacy is of utmost significance, necessitating additional investigation and enhancement of privacy protection methods. Furthermore, technical obstacles regarding distributed storage and data retrieval efficiency, network bandwidth and latency, and energy consumption management must be overcome.

B. Previous Works of Blockchain Applications in the Internet of Things

The integration of blockchain with IoT technology will enhance the advancement of IoT applications and offer more secure and effective data-management solutions for smart homes, smart cities, healthcare, and other domains.

Na et al. [12] developed and deployed Fusion Chain, a decentralized and efficient blockchain system specifically designed to facilitate the operation of IoT devices. Fusion Chain utilizes IPFS to address the problem of blockchain storage limitations and implements the Practical Byzantine Fault Tolerance (PBFT) consensus mechanism to minimize computational resource demands. Fusion Chain employs a Public Key Infrastructure (PKI) for data encryption to guarantee data privacy. Jiang et al. [13] introduced a cross-chain system designed to incorporate different blockchains into IoT data management. An interactive decentralized access paradigm was provided for IoT data-management scenarios, in which several blockchain platforms offer their own advantages. In this approach, the federation chain serves as the control site, whereas other blockchain platforms operate based on specific IoT scenarios. Maftei et al. [14] presented a blockchain-based data storage system designed exclusively for IoT devices.

The system utilized a dual-blockchain structure consisting of a lightweight blockchain (referred to as a local blockchain) and public blockchain. Public blockchains securely and permanently store the entire IoT data stream transferred over the entire Wireless Sensor Network (WSN) architecture. Gong et al. [15] introduced a Secure and Dynamic Access Control Scheme (SDACS) that is blockchain-based. This scheme was designed to improve the security, reliability, and transparency of IoT data sharing. The scheme ensures data privacy and security during the sharing process by integrating a hyperledger fabric, IPFS, and Attribute-Based Access Control (ABAC). This enables a decentralized and fine-grained access control.

A. Overview

The exponential increase in data volume has resulted from the advancement of IoT technology. The issue of effectively managing and storing this vast quantity of data has emerged as a pressing problem. Presently, the IoT data-management paradigm comprises five primary layers: blockchain, user–client, communication, data-storage, and IoT devices. Each layer performs a distinct and critical function in the entire system to guarantee the integrity and security of the data from collection to storage to access.

Various sensors and microcontrol units, including temperature, humidity, and pulse rate sensors, are included in the IoT device layer. Environmental and physiological data are collected using these devices to provide fundamental data support for upper-layer applications. Wireless communication technologies, including Wi-Fi, LoRa, and ZigBee, are typically employed in the communication layer to transmit data from IoT devices to the data-storage layer or peripheral computing nodes. The data are stored and managed by the data-storage layer, which employs either a centralized database or distributed storage system. Data transactions and operation logs are recorded using the untamperable and decentralized features of blockchain technology in IoT data administration. Users can access and administer IoT data through mobile devices or web applications to facilitate real-time monitoring and analysis. Fig. 1 illustrates the data-management model based on IPFS and blockchain.

Figure 1. Overview of IoT data-management models.

Blockchain- and IPFS-based data-management models offer novel solutions for the storage, transmission, and security of IoT data. We introduce a unified data pre-processing and encryption module in the IoT device layer to guarantee that the data are encrypted and processed prior to transmission, thereby enhancing data security. We utilize edge computing technology to assign a portion of the data-processing duties to the edge nodes, thereby reducing the delay in data transmission and bandwidth consumption. The IPFS technology is integrated to enhance the reliability and access efficiency of data storage and attain decentralized data storage. The blockchain's processing capacity and scalability is enhanced by optimizing its consensus mechanism. Simultaneously, smart contracts are created to facilitate the automated administration and operation of data. The front-end application and back-end data access interface is optimized to improve the performance and user experience of data access.

B. Model Design

This section provides a comprehensive explanation of blockchain and IPFS-based IoT massive data management (BIMDM) models. The objective of this model is to enhance the security, reliability, and efficiency of IoT data management by utilizing distributed storage and decentralized management technologies. Sensor nodes, ZigBee communication nodes, perimeter device or gateway nodes, IPFS nodes, blockchain nodes, and user clients are the six critical components of the BIMDM model. The node model architecture diagram is shown in Fig. 2.

Figure 2. Node model architecture diagram.

1) Sensor Node

The IoT data-management process is initiated by the sensor nodes that acquire and process data. Data is collected in realtime using sensor devices. Data, such as body temperature, blood oxygen, pulse, acceleration, sound, and gas concentration can be collected using sensors. These data are used for subsequent processing. Data are typically processed by sensor nodes before transmission to the upper stratum. This encompasses a reduction in the volume of data transmitted and enhancement of its quality through filtering, formatting, and compression. The data can be encrypted using sensor nodes to ensure security. Encryption safeguards the integrity of data and privacy of users during transmission. Using the wireless module, the sensor node transmits data to the gateway node or edge device.

2) ZigBee Communication Node

In the IoT model, ZigBee terminals wirelessly transmit data. They establish a connection between the gateway nodes and sensors. These sensors provide data to the ZigBee node. This encompasses encrypted pre-processed data from the MCU and unprocessed sensor data. The data are transmitted from the ZigBee node to the edge or gateway node through the ZigBee network. The ZigBee network is overseen by the ZigBee communication node. This encompasses the management of the network, addition and removal of nodes, and maintenance of routes. To ensure data security, ZigBee nodes can encrypt and authenticate data.

3) Edge Devices or Gateway Nodes

Edge devices are responsible for data aggregation, pre-processing, encryption, and uploading to the IoT. Sensors provide data to the gateway node or peripheral device. Data are transmitted to the edge device via various wireless communication methods to facilitate centralized management and processing. To enhance the quality and alleviate tension, the sensor data undergo pre- processing. Encryption ensures data security during transmission and storage. The encrypted data are uploaded to IPFS for storage, and the hash value is documented in the blockchain. To enhance responsiveness, edge devices can process data faster than clouds. Using edge computing, data can be analyzed, events can be identified, and local decisions can be made in real-time, thereby minimizing data usage and delays.

4) IPFS Node

IPFS nodes are crucial for storing data in the IoT. The IPFS nodes ensure data security and facilitate data retrieval. Data are stored by IPFS nodes on edge devices. Utilizing a distinct hash value for each file, the data are stored in the IPFS network. By utilizing the hash value of the file, IPFS nodes facilitate the rapid discovery of data. This method is more efficient than the conventional location-based approaches. Data sharing is facilitated by IPFS nodes. Data can be exchanged between nodes, and users can access public data via the IPFS network. This facilitates the exchange and retrieval of data. Different versions of the same data can be stored and retrieved using IPFS nodes. This is crucial when data are frequently updated. IPFS nodes ensure data security through encryption and access control.

5) Blockchain Node

Blockchain nodes are crucial for managed IoT data. A blockchain node records the data hash from an IPFS node. Therefore, the data cannot be modified. The hash value will also be altered, which is visible on the blockchain. To verify the accuracy of the data, users can examine the hash value on the blockchain using the blockchain nodes. Data are managed by blockchain nodes through smart contracts. Smart contracts execute predetermined rules and logic to automate data administration. Blockchain nodes deliver data for auditing and traceability.

6) User Client

Users and IoT devices communicate with the client. Users can access, manage, and control data and IoT devices. The client enables users to view data from IoT devices in real-time. Users can comprehend and analyze data using charts and dashboards. The client enables the users to search for and observe data within a specific period. The data can then be filtered, sorted, and exported. The client enables users to administer and control the IoT devices. This includes the addition, deletion, configuration, and monitoring of the devices. The client allows users to send commands to modify the device in real-time. The solutions required in various scenarios are diverse and include desktop, mobile, and web applications.

C. Model Process

Modern medical information systems require the precise entry, updating, and secure sharing of patient data. BIMDM is an approach intended to optimize this process and improve the quality and efficacy of healthcare services. Furthermore, BIMDM is a comprehensive system intended to manage vast quantities of data produced by IoT devices in a secure and efficient manner. The BIMDM model is ultimately divided into two components: the process of data acquisition and uploading to the chain for transaction storage and user access. The BIMDM system completes the entire IoT data administration process, including collection, pre-processing, transmission, encryption, storage, authentication, and user access. Users can easily access and manage IoT data, real-time monitoring, data queries, device control, and data verification through the user client, which enhances the overall performance of the IoT system and user interface.

Environmental and health data are collected using sensors. Microcontrollers, such as ESP32 or STM32, are connected to environmental or health sensors. The MCU performs pre-processing operations on the sensor data, including data cleansing, filtering, and compression, to enhance data quality and alleviate the transmission burden. Wireless communication technologies, including Wi-Fi and ZigBee, are employed to transmit pre-processed data to gateways or peripheral devices. To guarantee data security, an edge device receives the transmitted data for additional processing and subsequent encryption. To guarantee high availability and reliability of the data, the encrypted data were submitted to the IPFS for distributed storage.

A. Experimental Method

The MCU was connected to the sensors (DHT22, MAX9814, MQ-2, and MAX30102), which generated a program to collect the sensor data and ensure consistent data collection. Edge devices configured the MQTT agent and local database; deployed the hyperledger fabric blockchain network, which includes sorting nodes, peer-to-peer nodes, and smart contracts (chain code); and installed and configure the IPFS nodes. Raspberry Pi 4 and Intel NUC were set up as the edge devices. Servers must be configured to operate the IPFS nodes and hyperledger fabric. Data collection sensors (for temperature, humidity, vibration, gas concentration, and heart rate) and a microcontroller for data pre-processing were used (filtering and compression). The data were encrypted with AES-256 using the edge device and then uploaded to IPFS. The storage time of the data was recorded, and the hash value of the content was recorded in the hyperledger fabric blockchain.

B. Performance Evaluation

1) Throughput testing

The processing capacity of the proposed BIMDM model was assessed under high concurrency by local testing, which simulated 0-1000 IoT device nodes and recorded the processing capacity of the system under varying numbers of nodes. This evaluation was conducted to assess data-processing and storage rates. Throughput, which is a critical metric of data-processing capability, is the number of transactions that can be effectively processed in a given amount of time. Fig. 3 illustrates the throughput simulated for multiple device nodes. Our proposed model achieved a maximal value of 23893 TPS at 1000 nodes in the throughput comparison test, as illustrated in Fig. 3. This implies that they are more capable of processing in high-concurrency environments. The system proposed by Maftei et al. achieved a maximum value of 13434 TPS at 1000 nodes, which is a satisfactory performance and appropriate for medium-scale IoT systems. In contrast, the system proposed by Na et al. exhibited a substantial decrease in performance after 800 nodes, suggesting the necessity for additional optimization at an ultralarge-scale number of nodes.

Figure 3. Throughput test comparison diagram.

2) Latency Test

To assess the transaction rate in the proposed BIMDM model, the latency is defined as the total time necessary for the data to complete the following steps: acquisition, pre-processing, transmission, encryption, storage, and recording in the blockchain. To minimize the volume of data that must be stored in the blockchain, data storage is integrated with IPFS. The data are stored exclusively in the IPFS network, and the blockchain records only the HASH values. The latency was assessed as the sum of the total time required for data to be transmitted from the sensors to the blockchain and IPFS. Fig. 4 shows the latency. Fig. 4 shows that the proposed model obtained a maximum value of 998 ms under a high load with 1000 nodes, resulting in the lowest delay time, and performed well. The system proposed by Na et al. is suitable for applications with medium-delay requirements because it obtains a maximum value of 1900 ms under a high load with 1000 nodes. In contrast, the system proposed by Maftei et al. experienced a substantial increase in latency after the addition of more than 700 nodes, suggesting the need for additional optimization at the mega-node level.

Figure 4. Latency test comparison diagram.

3) Scalability Testing

The scalability of the proposed BIMDM model was assessed by analyzing the throughput and latency of the system in relation to the volume of data. The collected data were converted into a consistent JSON format and chunked at a fixed size of 100 KB. Fig. 5 illustrates the processing efficacy of the model as the amount of data increases. The scalability test in Fig. 5 demonstrates that the model maintains a high TPS performance with modest data sizes (100 KB to 1 MB) as the data size increases. This is achieved through the optimization of the IPFS storage and hyperledger fabric upload performance, as well as the efficient pre-processing and encryption of edge devices. This model exhibited excellent scalability and a high TPS at a data volume of 10 MB. The utilization of high-performance edge devices and low-latency high-bandwidth networks guarantees efficient data processing and storage.

Figure 5. Data size and model scalability test diagram.

(a) Throughput vs. data size

(b) Latency versus data size

In this paper, we propose BIMDM for the IoT. The proposed model is designed to address the issues of low data storage efficiency, high data transmission latency, and insufficient security and scalability in current IoT data-management systems by utilizing the efficient distributed storage of IPFS and immutability of blockchain technology. To verify the efficacy of the model, we conducted comprehensive experimental evaluations that encompassed data acquisition, pre-processing, encryption, transmission, storage, retrieval, and validation. We evaluated the scalability and availability of the system by simulating 0-1000 IoT device nodes and testing the throughput and latency of the model for varying data volumes and nodes.

The model will be further optimized in future research to investigate and implement more intelligent load-balancing algorithms, thereby improving the system performance in high-concurrency and high-load environments. To improve data security, more sophisticated data-privacy protection mechanisms, including homomorphic encryption and zero-knowledge proofs, are required during data transmission and storage. To alleviate the burden on the central node and enhance the processing capacity and response speed of the entire system, a greater number of computing duties are allocated to the edge devices.