2023年09月27日14:09:03 科技 1365

本文內容來自於網路，若與實際情況不相符或存在侵權行為，請聯繫刪除。本文僅在今日頭條首發，請勿搬運。

Introduction

With the rapid advancement of hardware, software, and communication technologies, the Internet of Things (IoT) has garnered increasing attention in both academia and industry. IoT applications span across various aspects of people's lives, revolutionizing sectors such as smart healthcare, transportation, agriculture, and entertainment.

One critical challenge lies in handling the vast amounts of data generated by IoT applications, given that IoT devices often lack sufficient processing and storage resources. This is where blockchain, an encrypted distributed computing storage system, steps in. Its purpose is to create real-time, tamper-proof records.

By merging IoT and blockchain, a theoretically verifiable, secure, and permanent method for recording data processed by IoT smart devices is established.

Related Studies

Existing frameworks integrating IoT systems with cloud computing can be broadly categorized into two types: those focusing on specific applications and those providing a general Platform as a Service (PaaS) model. Rahmani et al. have developed a prototype-based framework for supporting IoT healthcare systems, emphasizing local storage and data processing. In this framework, the cloud serves as the backend system for data analysis and decision-making. Azimi et al. have used Raspberry Pi and Intel Edison development boards as edge nodes in IoT, employing authentication to safeguard data privacy. However, in this framework, data is stored in the cloud.

Services within this framework are divided into two types: core operations for managing the framework, including resource management and security assurance, and application-specific requirements. The framework's security is guaranteed by authentication and access mechanisms.

Hybrid Blockchain Structure for IoT Wireless Sensor Networks

This article introduces a novel hybrid architecture for monitoring and managing IoT systems, as illustrated in Figure 1.

Traditional smart home systems monitor all activities through wireless sensor networks and send the collected data to the blockchain via wireless sensor networks for storage in the cloud. This enables the utilization of all advantages offered by blockchain. In the new architecture proposed in this article for IoT monitoring, smart devices are equipped with sensors and Raspberry Pi for controlling these sensors. Each smart node consists of different transceivers capable of capturing data and incorporating it into the Wireless Sensor Network (WSN). The sensors include an accelerometer model LIS3DH, which accurately measures acceleration within different ranges (0 and ±16g) and automatically adjusts to usage conditions while maintaining constant precision.

For temperature control, a low-power sensor TC1047A is used to collect temperature data. Through this process, data is transformed from IoT devices to the blockchain in the Edge Computing layer during the ETL (Extract, Transform, Load) process, rather than directly in the blockchain. This significantly reduces the energy consumption of the main chain in the cloud when converting data and constructing blocks, enabling the smart controller to perform the same tasks with lower energy consumption. Simultaneously, a side chain executed on Raspberry Pi 4B is created using smart contracts, storing all the data collected by WSN nodes.

Smart contracts verify the identity of smart devices, and once a WSN node is identified as a member, the smart contract executes an event to send data to the side chain. Once the smart contract is verified (i.e., data is fully inserted into the side chain), the side chain is inserted into the main chain. The cloud-based blockchain main chain is constructed from subchains built by smart devices and merged into one. To build the blockchain, this novel architecture uses side chains (i.e., small-scale blockchain for preprocessing data) to create an intermediate layer (i.e., Edge Computing layer) between IoT devices and the blockchain.

These edge chains optimize transactions sent to the blockchain by performing data processing at the network edge close to the data source. Executing side chains with WSN data reduces communication between WSN nodes (i.e., sensors) and the blockchain. This approach requires leveraging Edge Computing resources that cannot maintain a continuous connection to the blockchain network, reducing the waste of computing resources in the IoT layer.

Edge Computing pushes applications, data, and computing capabilities away from the blockchain. As a computing paradigm, Edge Computing is also referred to as grid processing, peer-to-peer computing, autonomic and distributed computing.

The advantages of Edge Computing relative to current computing lie in (1) significantly smaller amounts of data needing to be transmitted over the network, (2) elimination (or reduction) of centralized computing, and (3) enhanced virtualization capabilities for improved scalability.

Empirical Study on the Performance of Blockchain Framework Design

To assess the performance of the entire framework, an integrated computing environment consisting of multiple cloud and edge nodes was prepared. Table 1 describes the complete configurations of servers and related running components. The environment used to build nodes for the blockchain was uniformly a Docker environment, resulting in the construction of seven server nodes. One node was allocated for customer-related control operations, and the remaining nodes were assigned internal private IP addresses for internal data transmission and authentication. Of the six nodes, three were remote servers serving as data storage for the blockchain main chain. Two Raspberry Pi 4B devices were used as edge devices for collecting data from various sensors, constructing subchains, and executing smart contracts. After the smart contract execution was completed, the subchain data was uploaded to the cloud and inserted into the cloud's main chain, reducing the communication frequency between edge devices and the main chain. Additionally, a microcontroller was used as the control center between edge devices.

Figure 2 represents the average Docker sizes of different components in the framework under compressed and uncompressed conditions. The compressed Docker image sizes were obtained based on the average sizes stored in Docker hub for multiple architectures, while the uncompressed Docker image sizes were obtained based on extracted Docker images from instances. The compressed Docker image sizes indicate that the framework is lightweight and can be downloaded on different platforms, ranging from a few megabytes to a maximum of 100 megabytes.

Furthermore, the uncompressed Docker image sizes demonstrate that the framework components do not occupy excessive storage space. Figure 3 displays the average RAM sizes of different components in the hybrid architecture during runtime. From the figure, it can be observed that the proposed framework occupies very little memory, making it suitable for deployment on edge devices. The memory usage of different components ranges from 25 MB to 45 MB.

However, for existing edge devices, individual components can still be deployed on different edge devices with low resource requirements. Therefore, from an overall performance analysis, the entire framework does not excessively consume physical resources. It also better utilizes the existing resources of edge nodes, allowing various components to run reasonably through a distributed structure. Figure 4 demonstrates the average startup times of different components in the hybrid architecture. This includes the time required to start the containers until they are in a fully available state for processing incoming requests.

It is found from Figure 4 that this framework only takes a few seconds to complete startup and enter the processing state. The startup time of the system is related to its existing hardware resources. If faster startup speed is needed, edge devices with better computing performance can be employed, significantly enhancing overall response speed. Moreover, in the deployment phase, it greatly extends many important scenarios, widening the overall application prospects. Figure 5 describes the task quantities generated under different experimental settings in the hybrid architecture. It is observed that when deployed only at the edge, more tasks can be generated compared to tasks generated in the cloud or in a combination of cloud and edge.

This occurs because the edge is closer to the data source, allowing rapid delivery of results from the previous task while generating the next task. The communication latency is lower in this scenario. However, if computation is only performed at the edge, the data processing speed may be too slow, negatively impacting the overall data processing

speed. On the other hand, the cloud offers higher data processing speed.

From Figure 5, it is evident that combining edge and cloud computing generates a higher quantity of tasks while maintaining a balanced data processing speed. This hybrid architecture presented in this paper exhibits a well-rounded performance.

Conclusion

This paper introduces a novel hybrid blockchain structure based on IoT wireless sensor networks. This structure maximizes the computational capabilities of edge nodes, distributing a portion of the cloud's computational burden to the remaining edge nodes. Subchains are constructed on edge nodes, and once smart contracts are completed, the subchains are inserted into the cloud's main chain. This effectively reduces communication pressure between edge and cloud, resulting in a system with enhanced scalability.

The study also assessed the resource consumption of various components in the framework and their system response times. The experimental results demonstrate that the proposed framework aligns with the practical resource constraints of edge devices. The entire system exhibits high task processing capabilities and data processing speed, significantly improving blockchain processing throughput and efficiency. In practical applications, it can cater to a wide range of scenarios, offering significant practical value.

In conclusion, the hybrid architecture presented in this paper showcases a promising approach to harnessing the potential of both IoT and blockchain technologies. By effectively utilizing edge computing capabilities and optimizing data flow between edge and cloud, the framework offers a scalable and efficient solution for handling IoT-generated data.

References

[1] HUH S., CHO S., KIM S. Managing IoT devices using blockchain platform [C]// Proceedings of 19th International Conference on Advanced Communication Technology (ICACT). Pyeongchang, Korea (South): IEEE, 2017: 464-467.

[2] WALKER M. A., DUBEY A., LASZKA A., et al. PlaT-IBART: a platform for transactive IoT blockchain applications with repeatable testing [C]// Proceedings of the 4th Workshop on Middleware and Applications for the Internet of Things. New York, NY, USA: ACM, 2017: 17-22.

[3] CROSBY M., PATTANAYAK P., VERMA S., et al. Blockchain technology: beyond bitcoin [J]. Applied innovation, 2016, 2: 6-10.

[4] PAN J., YANG Z. Cybersecurity challenges and opportunities in the new edge computing+IoT world [C]// Proceedings of the 2018 ACM International Workshop on Security in Software Defined Networks & Network Function Virtualization. Tempe, AZ, USA: ACM, 2018: 29-32.

[5] RAHMANI A. M., GIA T. N., NEGASH B. Exploiting smart e-Health gateways at the edge of healthcare Internet-of-Things: a fog computing approach [J]. Future generation computer systems, 2018, 78 (P2): 641-658.

[6] AZIMI I., ANZANPOUR A. M., RAHMANI A. M. HiCH: hierarchical fog-assisted computing architecture for healthcare IoT [J]. ACM transactions on embedded computing systems, 2017, 16 (5s): 174.

以上內容資料均來源於網路，本文作者無意針對，影射任何現實國家，政體，組織，種族，個人。相關數據，理論考證於網路資料，以上內容並不代表本文作者贊同文章中的律法，規則，觀點，行為以及對相關資料的真實性負責。本文作者就以上或相關所產生的任何問題任何概不負責,亦不承擔任何直接與間接的法律責任。