Trust Management Technique Using Blockchain in Smart Building ^†

Abstract

Security is a big challenge for developing and implementing IoT in smart building situations. In this context, our goal is to create a secure blockchain-based trust management system. To do so we take advantage of the security features that blockchain technology provides in terms of reliability, traceability, and data integrity. We design and implement a blockchain-based trust strategy that collects trust evidence, assigns each device a trust score, and securely stores and shares them with other devices in the network by integrating them into blockchain exchanges. According to the findings of our performance evaluation, our concept includes security features such as tamper-proofing and assault resistance, reliability, and easy implementation for IoT environments and applications.

1. Introduction

The IoT in smart buildings performs a very vital role in daily life. IoT in smart buildings deals with various fields such as smart houses, vehicles, games, and organizational equipment [1]. It is an emerging field that can be used with various devices. In IoT in smart buildings, a considerable number of devices are connected. Nowadays, smart cities and smart grids are included in IoT in smart buildings [2]. The connectivity of IoT in smart building devices belongs to wireless connectivity of the sensors and other smart devices. Every object is part of the IoT in the smart building network, supporting interoperability with the current system [3].

Smart buildings are becoming a possibility with the integration of Building Management Systems (BMS) with an underlying control and networking infrastructure largely comprised of smart devices such as alarms, cameras, RFIDs, miles, and sensors [4]. IoT encompasses mobile devices and networking networks (IoT in smart buildings). The Building Management System (BMS) is in charge of a number of critical building elements such as the ventilation system, power, light, protection, and flame systems. It is capable of communicating with IoT cameras used in smart buildings [5]. A few studies on merging machine learning with model-based control approaches for building management have recently been developed. ML (Machine Learning) is used to improve the computing competence of the model-based controller in particular. Panteli et al. [6] present research on Building Information Modelling (BIM) smart building applications. They present the applications of BIM in all phases of smart building life. They discuss the applications of BIM in the phases of planning, building, and post-construction. Siountri et al. [7] present research on the development of smart buildings through Blockchain, BIM, and IoT in smart buildings. This paper focuses on implementing emerging technologies like BIM, IoT in smart building, and Blockchain for the construction sector. Carli et al. [8] present a research article on IoT in smart building HVAC system analytical management using a Smart Buildings Model. This paper discusses an indoor thermal comfort and energy usage optimization design for the predictive regulation of HVAC systems. Dey et al. [9] proposed a remote fault approach based on a case study of a multi-level machine learning approach to detection in smart buildings. This article presents a unique automatic fault detection system (MLe-AFD) based on machine learning for remote HVAC fan coil unit (FCU) behavior analysis [10].

It is necessary to safeguard the system and the reliability and trustworthiness of the shared information. Our network’s goals include tamper proof, information reliability, and authentication [11]. The rest of the paper is organized as follows. Section 2 describes the proposed solution. Section 3 describes the result and implementation, and Section 4 concludes the study.

2. Proposed Solution

Our system is made up of many smart building floors, each of which contains a set of hardware facilities (such as computers, ordinary actuators, IoT nodes, and so on) as well as a verification manager in charge of making access control decisions, attempting to verify device characteristics, and producing access control tokens. Moreover, each device on the smart building floor is connected to a trust manager responsible for determining the level of trustworthiness and assessing and calculating a comprehensive trust value for each associated device. This entity is expected to be located in a more powerful network node linked directly to every node in the scenario of IoT nodes with stringent resource limits; otherwise, it is implemented within the device itself. Following that, a collection of particular devices (miners) is implemented to receive trust values, create components, and transmit them into the blockchain system. More information is available in the section below.

2.1. System Architecture Detail

In our proposed paradigm, there are three basic conceptual layers. These are illustrated in Figure 1.

2.1.1. IoT Layer

This layer contains IoT devices that gather and process data. This category includes measuring instruments, controllers, RFID readers, computers, robotic systems, and other equipment. Their main responsibilities are requesting and acquiring data, executing tasks, and other primary activities. These devices will perform additional activities relating to trust management, which is unique and relevant to our proposed scheme. They will be able to assess, communicate, acquire, and collect information connected to trust from and to the management system in this manner. They also communicate with industrial services through the Internet, enquiring about and monitoring their status and any associated information, based on their quantitative trust scores, assessed, processed, and kept by system management.

2.1.2. Trust Layer

This layer comprises advanced interconnected hardware that are in charge of securing the proposed scheme’s setup, operation, and reliability. It also gives data on trust and reputation score activities, which are safely saved and indexed in a decentralized environment, allowing them to be acted on later when requested or necessary. Every completed action is propagated into a blockchain network of consensus objects to ensure that the action in question is verified, audited, and validated. Finally, for general system trust data access, the layer is in charge of authenticating each node and managing their action keys.

• IoT devices trust manager:

This device enables the creation of a secure and dependable environment in which devices can communicate with one another, including commercial IoT services, without concern of compromised trust scores’ integrity and validity. As a result, individuals may make trustworthy decisions and get data based on the measured trust levels.

The following variables were used to characterize the relationship between trustor “tr” and trustee “ti” at a specific time “n.” T: trust(trti)t.

This relationship is given the value Tri(t), which indicates the trust value of any device ‘a’ for any other device ‘b’. This trust value ranges from −5 to +5, with −5 denoting full ignorance and +5 denoting perfect trust.

T_ab = tri1 × (T_ab)^{(t − 1)} + tri2 × (T_ab)^(Δt)

(1)

Δt was assigned the threshold values tri1 and tri2, where tri1 + tri2 = +5; −5 = tri1 = +5 and −5 = tri2 = +5. The entity’s behavior is always changing, although it frequently changes throughout time.

• Machine Learning:

After calculation of trust, we store all trust in an array. After storing the trust in an array, we use a machine learning approach for trust classification. We employed the ID3 method to identify the maximum trust among all devices through a decision tree.

We denote the array in letter ‘E’. Numbered lists can be added as follows:

entropy(E)=(−p)/(p + n) [log₂](p/(p + n)) − n/(p + n) [log₂] (n/(p + n))

(2)

• Miner:

This component is in charge of confirming the authenticity, authenticity, and security of trust records and transaction data. This will be transmitted to miners to check its authenticity before packaging it into a block, chained into the ledger once received. Multichain blockchain technology, a private blockchain protocol that regulates block access through a list of registered participants, is used in our architecture. Participants who have already registered have access to the reading or writing blocks of the database. The main motivation we had for choosing Multichain is because it fulfills the overwhelming bulk of our criteria. Multichain is a private blockchain with permissions. Streams, which function as an autonomous append-only collection of objects, ensure that shared data is kept private. Second, it stands out for its versatility, which allows for changes in authorization and delegation.

3. Result and Implementation

The experiments were done on a Linux operating system, Ubuntu 18.04, on an ACER Aspire 5349 machine with a 3rd generation Intel Core i5 processor and 6 Gb of RAM for the multichain network. The standard simulation tool “NS3” was used to develop the proposed system. It is a readily available discrete network simulator primarily developed for education and research. Using this tool, a smart building with IoT devices was modeled and experimented. We used Multichain to create the blockchain network. An IoT in a Smart Building environment was considered with various smart building devices, using quantities of 30, 50, 70, 90, and 100. We evaluated a 50 to 100-device Internet of Things (IoT) scenario. Nodes from the same community may have comparable interests or duties to accomplish. Each device was assigned a random number between 1 and 10, indicating which of the 10 conventional communities it belonged to within the smart building management system. It might also have been a part of one, two, or three communities simultaneously. In this last case, malicious devices represented for the compromised model had 40% of the entire number of wireless devices. To illustrate the effectiveness of our blockchain-based storage and sharing solution, we assessed the response times and the number of generated transactions and the processing power utilized by every individual involved in the Blockchain.

3.1. Resiliency against Attacks

In this section, we look at the short-term sustainability of our notion in the face of harmful attacks performed by a group of IoT devices. First, we examined how the overall trust value like a good node fluctuates when the cumulative quantity of bad nodes executing bad-mouthing attacks fluctuates.

The network’s total number of devices was limited to 60. In order to accomplish this, we set the amount of dishonest nodes in the example to 40%.

We examined the evolution of a malicious node’s trust value in Figure 2 as the fraction of total bad nodes executing ballot stuffing and on–off attacks increases. The malicious node alternated between behaving well and badly in packet delivery behavior in an on–off attack. Figure 3 represents that once a problematic node’s trust value goes below 0.4 and repeatedly acts correctly, it is punished. Its score is kept at 0.4 thanks to a function provided by blockchain technology.

3.2. Performance Evaluation

The reaction time was assessed while altering the size of the file containing trust information and trust ratings for every evaluated object in the network. The files came in different sizes, from one kilobyte to two megabytes. The average response of the blockchain system increase in lockstep with the size of the trust file, as seen in Figure 4.

4. Conclusions

The design and implementation of a safe trust management system based on a blockchain network to gather trust evidence and reliably storing and sharing it within and around the blockchain network are described in this article. Our system is practical, deployable, and suited for IoT contexts, according to our review, because it is decentralized, ensures security and resistance against a combination of attacks, and has a low overhead in addition to other properties. We wish to apply various benchmarking blockchain consensus techniques to optimize the overhead and performance of our approach.