Scalable and Secure Gradient Smart Dempster Shafer Reputation for Development of Smart Cities


 Over the last decade, blockchain has been considered an encouraging solution to secure distributed ledgers. Moreover, with the introduction of a pseudonymous payment method without a centralized database or authoritative person, blockchain has also evolved as the future generation for online payment system. However, with the eruption of a large scale database, scalability has also become a demanding issue. In addition to the obstacle mentioned above, challenges like security and scalability stop accelerated adjustments for the development of smart cities. Without directing this essential scalability and privacy issue, such an encouraging method may not help develop smart cities. This paper bestows a measure to analyze both scalability and security aspects of existing blockchain methods with applications of smart city networks. The proposed method is known as Gradient Smart Load Balancer and Blockchain Dempster Shafer Reputation (GSLB-BDSR). Gradient Smart Load Balancer is designed so that even though with the increase in the number of participating sensors, the load is said to balance by applying gradient function, therefore ensuring scalability. Next, to cover the security aspect, with the aid of scalable blocks in the blockchain network, a Blockchain Dempster Shafer Reputation model is proposed. Evaluation outcomes of proposed security solutions outperform conventional solutions.


Introduction
Tasks and smart home characters are continually progressing due to contemporary Information and Communication Technology (ICT) and the Internet of Things (IoT). With the constant growth globally and with the anticipated overall population exponentially increasing, smart cities have a tendency. According to the home's network arrangement, users explore numerous keep an eye on and manage themselves based on the user settings. However, this shift has generated a smart home circumstance. The formation creates significant security susceptibilities, as several devices in smart homes include centralized networks.
TP2SF framework consisted of three modules. They were the trustworthiness module, two-level privacy module, and intrusion detection module. With these three structure module, accuracy, detection rate and precision were found to be improved significantly.
But, few challenges are identified with enhanced IoT nodes, time is taken in file uploading, and block mining enhances. In the future, we improve this work by various load balancing criteria to enhance network performance. Gradient Smart Load Balancer model is designed to aid gradient function to ascertain load condition to address issues in this work. It organizes block gradient surface separately in two different layers. Therefore it ensures scalability via throughput and latency time even with the increase in participating IoT devices.
The contribution of gateways in smart homes is outstanding, but its centralized organization dispenses numerous security susceptibilities. A blockchain-based smart home gateway network was proposed in [2] via three layers to address these security susceptibilities, comprising of device, gateway, and cloud layers. To start with, initially, the blockchain framework was utilized in the gateway layer. The security response time and accuracy were found to be improved. Despite improvement observed inaccuracy, additional computational complexity was incurred due to the increasing number of IoT devices, therefore compromising security. In our work, a Blockchain Dempster Shafer Reputation model is proposed to address this issue with the aid of trust and reputation measured via the Dempster Shafer proposition, ensuring security via accuracy and precision.
The inspiration for work is to resolve the scalability and security of transactions in implementing blockchain computing-based solutions in the smart city. Blockchain with cloud computing environment is the next evolution in the smart city environment, and several countries identify services beyond existing scalability and security-based solutions. Smart cities generate numerous IoT data streams that assist in transforming prevailing standard computing-based results in significant results. Hence, a novel method for smart city permits numerous and frequent streams of human and environment collective contextual data to benefit scalable and secured computing-based applications. Contribution of work enumerated as: • This article offers related work performed to indicate requisite foundations of IoT, blockchain and smart city architecture in circumstances of IoT.
• GSLB-BDSR method shows various domains of smart city network, namely IoT Fridge activity, IoT Garage Door activity, IoT GPS tracker activity, IoT Modbus activity, IoT Motion Light activity, IoT Thermostat activity and IoT Weather activity to construct scalable and secured computing-based applications.
• Technologies that make it possible for the GSLB-BDSR method to work in a smart city.
• The opportunities and challenges which arise in executing the GSLB-BDSR method. We describe the need for scalability and security to deploy the smart city in a blockchain network, the concerns of precision, accuracy, throughput and latency time in addressing scalability and security concerns.
The residual structure of the article is ordered as below. Related works of blockchainbased security and scalability method for smart cities are offered in Section 2. GSLB-BDSRmethod for the deployment of smart cities in IoT is described in Section 3. Section 4 estimates throughput, accuracy precision, and latency time and scrutinizes the recital parameters of the GSLB-BDSR method with a detailed experimental setting for fair comparison among GSLB-BDSR and state-of-the-art methods. Lastly, Section 5 concludes the work.

Related Works
A comprehensive and systematic review in distributing loads among different nodes was proposed in [3]. The insurgence of the IoT is metamorphosing several notions, therefore making them smart. It has transfigured numerous areas of real-life-critical notions of this kind of revolution are Smart City. Though several cities are found to be transformed digitally, still there are found to be several hurdles making the system a more cumbersome process. In [4], areas, where blockchain is utilized are highlighted with the inclusion and advantages of utilizing blockchain in a smart city.
A blockchain-based smart home gateway network was proposed in [5] with the aid of the blockchain method provided measures for potential attacks on the gateway of smart homes. In addition to protecting transparency and ensuring security for each smart sensor activity, a novel secure wireless mechanism utilizing Blockchain technology was proposed [6]. A review of security and privacy concerning the development of the smart city was investigated in [7].
In [8], a smart city network architecture using the cognitive model, ensuring scalability and flexibility were designed.
Present-day evolutions in IoT has validated gathering, processing and different forms of data analysis with data about personal to create necessary knowledge, making more productive services [9] for stakeholders. However, additional security and privacy issue occur due to the tremendous scale of IoT networks. In [10], a novel privacy preservation blockchain called TrustChain eliminating the delay and ensuring privacy was provided. Yet another study on the convergence of blockchain and Artificial Intelligence was proposed in [11].
A blockchain-based security solution was introduced in [12] for Industry 4.0-based applications. A comprehensive review of blockchain involving industrial aspects was made in [13]. Another experimental study involving the impact of blockchain in the smart city was proposed in [14]. A conceptual framework that includes three dimensions, namely, human, technology and organization, was presented in [15].
Next-generation smart cities are countenance of the concurrence of these developments.
An enormous amount of data is produced by the mass crowd and IoT devices daily. These data have to be processed and acknowledged securely and cognitively. In [16], a blockchain-based infrastructure was designed using artificial intelligence, therefore supporting secured smart city services. Another fine-grained access control mechanism for smart healthcare using hash calculation was proposed in [17], therefore contributing to security.
A privacy-preserving strategy via Healthchain with Blockchain technology ensured security, privacy, scalability and integrity concerning smart healthcare data was discussed in [18]. Concepts of blockchain, smart city and file system utilized for smart cities were investigated [19]. Yet another blockchain-based loan system push-pulls mooring effects was proposed in [20].
So far, researchers estimated numerous use cases of blockchain. But, small-scale research is performed on the blockchain concept in smart cities. Several authors have simulated their graphs in esteem to scrutinize sensitivity, specificity structure. This work aims to present a smart city environment with blockchain and calculate graphs on various parameters in security and scalability via blockchain.

Gradient Smart Load Balancer and Blockchain Dempster Shafer Reputation
Blockchain is a distributed network, shares and stores data between sensors in IoT devices participating in the communications. As far IoT-driven smart cities are concerned, data created by sensors in IoT devices are stored in a distributed blockchain ledger. With this, data centralization and reliability is said to be ensured. However, with the increase in participating IoT nodes, both security and scalability remain a significant factor to be analyzed. Scalability is influenced by numerous factors, like, throughput (i.e., maximum throughput or the maximum rate of how many transactions can be confirmed by the network), latency time (i.e., how quickly transactions are confirmed).
The main contribution of this work is concentrating on how different scalability influence the security of proof-of-work blockchain and what improvements in regards to an increased number of transactions per second and latency they bring. In this work, a method called GSLB-BDSR for smart cities is proposed that provides security and scalability using blockchain for the development of smart cities. Here, only after the validation of each transaction between sensors in IoT devices execution is performed, hence contributing to both security and scalability to a greater extent. A blockchain network model is designed, followed by which the scalability and security models are elaborated in detail.

Network model
Blockchain applied for smart cities development is a critical part of data transmission authentication necessitating confidentiality among devices and sensors. Smart Cities possess a centralized network form; it has been applied in recent years as centralized to the distributed network via utilizing blockchain at the cloud layer. Smart city gateway presented based on designed blockchain with three layers. It consists of a device layer, gateway layer and cloud layer [2]. Figure 1 given below shows the network model used in our work.

Fig. 1 Overview of the network model
As shown in the above figure, the first layer, also referred to as a device layer ' ' comprises sensors ' = 1 , 2 , … . , ' and devices ' = 1 , 2 , … , ' and monitor data ' = 1 , 2 , … , ' in a smart city environment via numerous heterogeneous IoTs configured in a smart city. It is mathematically stated as given below.
The second layer utilized in our work refers to the gateway layer ' ' that stores and offers to users as required. It is mathematically expressed as given below.
Finally, the cloud layer registers ID ' ' for gateway ' ' and data processed by every gateway ' ' in the corresponding blockchain and formulated as given below.

Gradient Smart Load Balancer model
With enhanced IoT nodes for the development of smart cities, the time consumed in block mining moderately rises, compromising scalability. Gradient Smart Load Balancer (GSLB) is designed which enhance the performance of blockchain network to address this issue. Figure 2 shows the Proximity Administration and Block Gradient Surface block diagram used in Gradient Smart Load Balancer.

Fig. 2 Block diagram of Proximity Administration and Block Gradient Surface
As an example, the figure depicts a Proximity Administration and Block Gradient Surface system with a 4 x 3 rectangular configuration and assumes that sensors ' 6 ' and ' 11 ' are lightly loaded. As shown in the above figure, the GSLB uses a two-layered load balancing algorithm. The first layer let each block ascertain its loading condition. The time differing load state of a block may be light, average, or dense. Hence, if a block is light, more load is said to be given to the block in the blockchain network, on the other hand, if a block is dense, then some of the load has to be freed, or else, the blocks in the blockchain network is said to be average and left as it is. is the most considerable distance between any two blocks of ' ', and this is mathematically expressed as given below.
The gate of a sensor in the form of a block. ' ' then represents a binary function. ' '.
Here, the gate is open if the block in the blockchain network is lightly loaded, and on the other hand, it is closed. It is defined as given below.
The second layer of the Gradient Smart Load Balancer step is to organize a gradient ' is defined as ' ' and this is represented as given below.
From the above equations (7) and (8), the proximity of a light sensor is denoted as zero.
In contrast, its instantaneous adjacent sensor's proximity represents that these sensors in the respective blocks are one hop away from a light block. The proximity of the adjacents' adjacent is two, etc. The pseudo-code representation of Gradient Smart Load Balancer is given below.

12: End
As given in the Gradient Smart Load Balancer algorithm, the objective remains in improving the scalability with the increase in participating IoT devices. It is achieved by designing a two-layered Gradient Load Balancer for smart cities. In the first layer, load conditions are ascertained, and accordingly, in the second layer, block gradient surface for each sensor are organized on time, therefore ensuring scalability.

Blockchain Dempster Shafer Reputation model
In

Rating Blockchain Dempster Shafer Reputation
Scalable and secured smart city environment From the above equation (9), entry-level transaction. ' ' is derived based on the random number, ' ', public key ' ', timestamp, ' ' and the signature of the issuer, ' ' respectively. The entry-level transaction, ' ' is next validated in the blockchain network where the transaction signing is said to occur and is then forwarded to the blockchain network via the updated transaction as given below.
From the above equation (10) ' or malicious device, ' ' respectively. A new IoT device attains a more excellent reputation score, more significant than threshold trust value, ' ℎ ' of the blockchain network, ' > ℎ ' to start.
Therefore, an IoT device with a transaction must have sufficed reputation, ' > ℎ ' to resume as a candidate in the blockchain network. With the reputation score and trust aid, security is ensured through Dempster Shafer and hence referred to as the Blockchain Dempster Shafer Reputation model. In the Blockchain Dempster Shafer Reputation model, all cloud service providers independently issue voting transaction for the IoT devices with ' < ℎ ' and is mathematically expressed as given below. Dempster Shafer-based reputation is utilized that integrate evidence from different sources (i.e., cloud service providers). It arrives at a degree of belief (represented by belief function) takes every available evidence (obtained from all cloud service providers) to evaluate the trustworthy of IoT device. Then, the possible set of conclusions 'Θ' is given below.
From the above equation (12) From the above equation (13), '∅' refers to the empty set possessing '0' probability as one of the outcomes obtained from the cloud service provider has to be true. Each of the other outcomes in the possible subsets has a probability of either '0' or '1'. Then, belief in an IoT device, ' ' refers to the sum of the masses that are subsets of IoT device, ' ' respectively and is mathematically expressed as given below.
The pseudo-code representation of Blockchain Dempster Shafer-based Reputation for a secured blockchain with smart city networks is given below.  (9) 6: Validate entry-level transaction as in equation (10) and update accordingly 7: Issue voting transaction as in equation (11) 8: Evaluate possible set of conclusions as in equation (12) 9: Evaluate the set of all possible subsets of 'Θ' as in equation (13) 10: Measure belief as in equation (14) 11

Experimental results and Discussion
Experimental analysis of GSLB-BDSR is conducted in CloudSim and Java programming language with IoT-based datasets ToNIoT [21], [22]. The results of GSLB-BDSR are compared with [1] and [2] for in-depth analysis of latency time, throughput, security response time, accuracy. A prototype has been implemented to examine the viability and performance of the method. The configuration utilized for analyzing the simulation is an intel core i7-4790@3.60GHz processor, 4GB RAM.

Dataset description
The UNSW Canberra @ the Australian Defense Force Academy (ADFA). Besides, datasets were attained in a parallel processing manner for obtaining numerous regular and cyber-attack events.
It included raw datasets, processed datasets, train test datasets, description stats, and security events ground-truth data set.

Performance analysis of throughput, accuracy and precision
The first essential and significant metric of consideration for scalable and secured development of smart cities is accuracy. The higher the accuracy rate, the large numbers of IoT devices tasks are being accessed, and more significant events are noted from IoT networks.
Accuracy ' ' refers to the percentage ratio of correctly identified instances for sustainable smart cities by leveraging blockchain and reputation model to the total number of observations in the test case. It takes both a valid positive rate, ' ' and an actual negative rate, ' ', into account for measuring the accuracy. It is mathematically expressed as given below.
From the above equation (15), the accuracy rate ' ' is measured based on the actual positive rate ' ', actual negative rate ' ', false-positive rate ' ' and the false-negative rate ' '. It is calculated in percentage (%). Second parameter is precision. It refers to the percentage ratio of malicious activities detected to the total number of observations measured as an attack.
The precision rate is mathematically expressed as given below. It is calculated in percentage (%). Table 1 below shows the throughput, accuracy, and precision of three different methods: GSLB-BDSR, TP2SF [1], and Blockchain-based smart home gateway network [2]. From table 1, the performance of the GSLB-BDSR method using the ToN_IoT dataset illustrates improved throughput, accuracy and precision.   inferred that the accuracy is better using the GSLB-BDSR method than [1] and [2].
The reason behind the improvement was due to the application of the Gradient Smart Load Balancer algorithm. The gradient block surface was organized only after ascertaining the load put in the blockchain network by applying this algorithm. The scalability was improved, and the rate of transactions accurately addressed involving different IoT due to the binary function applied, therefore ensuring smooth task migrations between blocks in the blockchain network.
Due to this, the accuracy using the GSLB-BDSR method was better than 2% compared to [1] and 6% compared to [2].
Finally, we estimated precision to address the scalability aspect for different IoT features.
Precision illustrated the number of transactions confirmed based on actual positive and false positive by the cloud service provider via the blockchain network. Also, the precision using the GSLB-BDSR method was better than [1] and [2]. The improvement was due to the application of Proximity Administration and Block Gradient Surface in the GSLB-BDSR method. With this, both the breadth and the proximity factor were considered for the increasing number of participating devices or transactions. The precision was better using the GSLB-BDSR method by 3% compared to [1] and 4% compared to [2], respectively.

Performance analysis of latency time
Next, to ensure scalability using blockchain for the development of smart cities, latency is highly considered. It refers to the time consumed in confirming the transactions. It is mathematically expressed as given below.
From the above equation (18)    the latency time is also said to be increased. However, with '500' number of transactions to be performed in the blockchain network and the time consumed in transacting single blockchain being '0.145 ' using GSLB-BDSR, '0.185 ' using [1] and '0.205 ' using [2], the overall latency time was observed to be '72.5 ', '92.5 ' and '102.5 ' respectively.
The improvement using GSLB-BDSR is due to the application of the Blockchain Dempster Shafer-based Reputation algorithm. By applying this algorithm, scalability is said to be ensured. In other words, the transactions are quickly confirmed by the cloud provider in the blockchain network. To reduce the latency time, for each subset of IoT device with numerous transactions, validation of the entry-level transactions is first performed. Then the actual voting transactions are issued. The probable conclusion set and subsets were obtained based on the belief obtained for each transaction. Finally, based on the belief and reputation score, validity was made; this, in turn, reduced the latency time of the GSLB-BDSR method by each device is obtained. The security faster adaptations of 13% compared to [1] and 25% comparison to [2] respectively.

Conclusion
An efficient GSLB-BDSR method for building a secure and scalable method using blockchain to develop smart cities is proposed to improve throughput, accuracy, and precision with minimum latency time. In work, the scalability is addressed via throughput, latency time, whereas security is addressed utilizing precision and accuracy. The key objective of GSLB-BDSRmethod is to ensure throughput, precision, accuracy maximization and minimize latency time for a smart city environment. The objective of GSLB-BDSRmethod is attained with the application of Gradient Smart Load Balancer and Blockchain Dempster Shafer-based Reputation algorithm. First, a two-layer modelled was structured to ensure scalability where ascertaining load conditions and organization of each block for numerous transactions were made in a computationally efficient manner. By employing the Blockchain Dempster Shafer-based Reputation algorithm, reputation scores were obtained through Blockchain Dempster Shafer and rating was made accordingly for each transaction. Finally, for each belief factor were obtained and only upon reputation score evaluation, the IoT devices were accessed, ensuring security. The efficiency of the GSLB-BDSR method is estimated in terms of throughput, accuracy, precision and latency time. Simulation results ofGSLB-BDSRmethod present better performance with enhanced scalability and security for smart cities via blockchain than conventional works.

Conflict of Interests
On behalf of all authors, the corresponding author states that there is no conflict of interest.

Data Availability statement:
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.

Code Avaliability
Not Applicable.