An Efficient Blockchain-Based IoT System …

173

framework of IoT devices. Dorri et al. [8] describe how blockchain design may

be changed to improve the infrastructure’s ability to support IoT devices, partic-

ularly in terms of transaction speed. Some research focuses on the integration of

blockchain and machine learning in addition to blockchain in conjunction with IoT

[2]. Until now, the focus has been on linking blockchain with the two-revolutionary

technology,whichisIoTormachinelearning,ratherthanimplementingallthreetech-

nologies at the same time. The actual potential of these new, developing technologies,

however, will solitary be realized if they are merged. Kumar Singh et al. [9] propose

a blockchain-based framework for IoT and machine learning. Unlike Kumar Singh

et al. [9], their study presents a non-procedural review of all invention’s advantages

and in the way they accompaniment one another.

The widespread use of IoT devices is allowing different areas of our everyday

life to be automated [10, 11]. The computerization of houses and cities, denoted

as smart homes and smart cities, is one of the IoT’s success stories [11]. Securing

the transmitted and deposited data in the home system from malevolent activities

wishing to cause havoc in someone’s house is a big barrier to attaining the true

vision of smart homes [12]. Although there are several competing technologies aimed

at protecting data in smart homes from attacks, blockchain has developed as the

utmost capable technology for both safeguarding the home network counter to data

management attacks and offering a safe platform for the entire gadgets in the network

to interact with one another during computation in the cloud (cloud computing)

[13–15]. Considering the primary agreement protocols—a method by which the

entire transactions are confirmed by the complete nodes—the data in a blockchain is

immutable [16]. Administration attacks on conveyed or deposited data are therefore

unlikely to succeed with a single hacked node, and a majority of nodes must be

hacked [17]. Different consensus protocols are discussed, as well as their application

to IoT networks [12, 18].

Without any examination, Gupta et al. proposed the idea of adding ML approaches

to blockchain’s consensus procedure as an impending study project [19]. The advan-

tages of combining blockchain and artificial intelligence (AI) have been discussed

by Dinh et al. [20]. They claim that a blockchain regulated by an ML algorithm may

identify assaults and trigger appropriate protection measures or isolate the compro-

mised component. This concept known as AI-enabled blockchain has been success-

fully designed and implemented (AIBC). Dey [21] presented a utility function for

detecting anomalies [21] that is comparable to the function used in [17]. Then, he

believes, this value may be fed into a supervised ML system to predict the likelihood

of an attack and prohibit the consensus protocol’s blockchain confirmation of that

transaction. He does not, however, propose a strategy or implementation for creating

a useful consensus protocol. They use Hyperledger fabric to create a 3-layer archi-

tecture to evaluate the validity of our 2-phase consensus protocol for IoT systems.

The application layer, which houses various IoT devices, is the initial layer. The edge

blockchain layer and the principal blockchain layer are the second and third levels,

respectively, and contain various AIBC components.