Optimized Fusion Chain with Web3 Storage

Tahiya Tabassum Lisa

and Chaiyachet Saivichit

Wireless Network and Future Internet Research Unit,

Department of Electrical Engineering, Faculty of Engineering,

Chulalongkorn University, Bangkok 10330, Thailand

Keywords: Fusion Chain, IoT, Web3 Storage, Practical Byzantine Fault Tolerance Consensus Algorithm, Raft Consensus

Algorithm.

Abstract: The Internet of Things (IoT) has become increasingly popular over the past two decades. It has a significant

impact on many aspects of daily life, including smart cities, intelligent transportation, manufacturing, and

several other industries. Processing and networking abilities of the Internet of Things are condemned/

censorious in today’s highly technological environment. These systems are also reasonably priced when used

on embedded platforms and consume little power (Ismail and Materwala, 2019). Nearly all IoT devices have

bounded storage and random access memory with 8-bit or 16-bit microcontrollers (Bosamia and Patel, 2020).

Emerging technologies, such as blockchains, can be altered to accommodate IoT networks’ and devices’

limitations and requirements. Nowadays, blockchain has become the most secure medium of data

transmission. The main goal of this paper consists of the usage of web3 storage in the fusion chain platform

along with the fusion chain’s lightweight block structure that can be used by IoT devices resulting in minimal

CPU power consumption. Moreover, the development of two consensus algorithms for the performance

evaluation of the fusion chain blockchain to balance the computational power with the Internet of Things is

proposed.

INTRODUCTION

Blockchain’s underlying technology has the potential

to change how we consume information completely.

As the Blockchain grows, the operating node requires

significant storage, processing power, and data

transparency. The ability to establish trust in remote

settings without the need for authority is a

technological advancement that has the potential to

change several industries, including the Internet of

Things. Resources are typically limited in IoT

devices. The devices require much processing power

and scalability to handle planned and specified

computations. These devices’ high power

consumption is a significant drawback and

performance constraint.

We believe that Blockchain will be one of the

following disruptive technologies that IoT will use to

surpass its current limitations, following in the foot-

steps of other pioneering technologies like big data

https://orcid.org/0009-0004-5795-0514

https://orcid.org/0000-0002-6308-9191

and quantum computing. In contrast to Hyperledger

Fabric and Ethereum, the light block structure has

been suggested for the fusion chain, as seen from the

literature on block structures of different blockchain

systems (Na and Park, 2021) (Thakkar et al., 2018).

This paper aims to propose a blockchain topology

with web3 storage. To test the performance of CPU

power consumption, both the PBFT and RAFT

consensus algorithm implementation is included. (Na

and Park, 2021) (Yao et al., 2021).

The rest of the paper is organized as follows:

Section 2 provides an overview of blockchain

technology and Internet of Things (IoT) challenges.

The related work is presented in Section 3. Section 4

describes about the system architecture and

architecture flow. Section 5 includes the frontend and

backend flow. Section 6 describes about the achieved

results. Conclusion and Future Work is mentioned in

Section 7.

Lisa, T. and Saivichit, C.

Optimized Fusion Chain with Web3 Storage.

DOI: 10.5220/0012147800003541

In Proceedings of the 12th International Conference on Data Science, Technology and Applications (DATA 2023), pages 613-620

ISBN: 978-989-758-664-4; ISSN: 2184-285X

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

613

BLOCKCHAIN OVERVIEW

AND IoT CHALLENGES

2.1

Overview of Blockchain

The roots of blockchain technology were established

in the late 1980s and early 1990s (Lamport, 2019). To

prove that nothing had been changed in the collection

of signed copies, a signed data chain was employed

in 1991 as an electronic ledger for digitally signing

documents (Narayanan et al., 2016). Bitcoin is only

one of many blockchain-based inventions now under

development. Before Bitcoin, other, less well-known

electronic payment systems existed. Because no

single user could control the virtual currency and

there was no single point of failure, the distribution of

a blockchain-enabled Bitcoin increased its utility. As

a result, direct user-to-user transactions without the

involvement of third parties were made possible. The

Bitcoin Blockchain allows users to maintain their

anonymity, but even if users’ identities are concealed,

all transactions and account IDs are still viewable to

the general public. In the absence of authorized

intermediaries, the four fundamental characteristics

of blockchain technology outlined below foster the

necessary confidence within a blockchain network:

Ledger:

The system keeps a complete record of all

transactions in an append-only ledger.

Secure:

The ability to rely on blockchain technology to

safely store and verify data is ensured by sophisti

cated

cryptography.

Distributed:

Participants on the blockchain network are

transparent to each other because the ledger is

distributed among them.

Decentralized: A decentralized ledger grows the

number of nodes and makes a blockchain network

resistant to malicious attacks.

2.2

Categorization of Blockchain

Depending on the permission system used to

determine who is allowed to maintain a blockchain

network, several categories may be identified. Its four

classifications are a public, private, consortium, and

hybrid blockchain topology.

Publicly controlled blockchains are open,

unconstrained, and available to all users. The majority

of bitcoin trading and mining now takes place on

public blockchains. Blockchains that have

permissions owned by a single entity are called

private blockchains. They are only partially

decentralized because the general public can only use

them to a limited extent.

A consortium blockchain is a permissioned

blockchain maintained by multiple organizations

rather than a single organization.

Hybrid blockchains are ledgers controlled by one

organization but accessible to public blockchain

inspection, which is required to verify some

transactions.

2.3

Pillars of Blockchain

Blockchain technology is built upon four pillars:

Cryptography: Cryptography is a vital part of

blockchain security. A blockchain encrypts data

before transferring it to a destination using

cryptography, which utilizes symmetric and

asymmetric keys and hash functions (Bosamia and

Patel, 2020).

Distributed Ledger: It is widely available, shared,

and synchronized by agreement across many

locations. The distributed ledger is decentralized and

keeps track of every contract and transaction between

various parties and sites (Bosamia and Patel, 2020).

Smart Contracts: Smart contracts are blockchain-

stored computer programs that only launch when

specific requirements are satisfied. They are used to

automate contract execution and guarantee that

everyone is notified of the conclusion as quickly as

feasible (Salimitari et al., 2020).

Consensus Algorithms: The blockchain network’s

nodes reach a consensus on the distributed ledger’s

current state using a consensus algorithm. The

primary motivation of the publishing node is

unquestionably financial gain, not a concern for the

welfare of other publishing nodes or the network.

2.4

Classification of Blockchain

Consensus Algorithms

Depending on the working mechanism, blockchain

consensus algorithms are classified into two

categories: Proof-Based and Voting-Based, which is

shown in Figure 1. Further, the Voting-Based

algorithms are divided into two classes: Crash Fault

Tolerant and Byzantine Fault Tolerant. Among these

two, PBFT (Practical Byzantine Fault Tolerance

Consensus Algorithm) and RAFT (It is not quite an

acronym but named after Reliable, Replicated,

Redundant, And Fault-Tolerant) are proven suitable

for the Internet of Things. Since all IoT networks

desire high throughput, low latency, and minimal

computational overhead, the PBFT and RAFT are

preferred among most consensus algorithms. They

can operate normally even when more than one-third

of all nodes engage in malicious behavior.

DATA 2023 - 12th International Conference on Data Science, Technology and Applications

614

Fast throughput and low latency are all benefits of

PBFT that make it a good choice for IoT networks. In

RAFT, where as much functionality as possible is

concentrated on the leader, and leader election is

necessary as part of the consensus process (Salimitari

et al., 2020). PBFT may work adequately in the

presence of up to one-third of all nodes engaged in

malicious activity. As far as we know, RAFT is the

only consensus-based log replication method that

supports fewer message types.

Figure 1: Classification of Blockchain Consensus

Algorithms.

2.5

An Overview of IoT Architecture

An IoT architecture is a four-step procedure whereby

data is collected by sensors and sent to the corporate

data center for processing, analysis, and storage. The

first step of the procedure entails using sensors and

actuators to keep an eye on and control a physical

process or phenomenon (Na and Park, 2021). The

second step consists of internet gateways and data-

gathering tools. Pre-processing reduces the digital

data entering the data center or cloud in the third

stage. The final step involves a thorough analysis in

the data center or cloud, where a decision is made

after carefully analyzing the data for business

requirements (Yao et al., 2021).

2.6

IoT Device Classification

IoT devices fall into one of two categories:

Non-Constrained Devices: These devices have a

direct Internet connection capability (Ahmed and

Shilpi, 2018). An example is an Internet of Things

device connected to an automobile through a Mobile

Network Operator. Another example is a wearable

device that sends a patient’s health data to a distant

server.

Constrained Devices: Constrained devices are

those that have limited memory and processing speed.

Typically, they cannot access the Internet due to an

external barrier (Ahmed and Shilpi, 2018).

2.7

Challenges to Adopt Blockchain in

IoT

The following difficulties exist when utilizing

Blockchain on IoT devices:

Computation: The cost of blockchain activity is too

high for small-scale IoT devices. The demand that a

full node in the Blockchain validates and searches

every block and transaction may be a significant load

for IoT devices with limited resources. IoT devices

are incompatible with consensus techniques like

PoW. As a result, IoT devices cannot supply

sufficient CPU power for PoW operations.

Storage: The entire Bitcoin Blockchain is about 150

gigabytes, and the Ethereum Blockchain is almost

400 gigabytes (Vujičić et al., 2018). So, IoT devices

can not afford the extensive storage that Blockchain

requires. In the past nine years, Bitcoin has received

about 5105 additional blocks.

Communication: Blockchain, a peer-to-peer

network continuously exchanging data, must

maintain consistent records, including the most recent

transactions and blocks. Wireless communication

methods, often used to link IoT devices, are less

reliable than traditional connections in typical

Blockchain applications because of shadowing,

fading, and interference. Blockchain requires much

more processing power than wireless technologies do.

Energy: A single battery charge can power some IoT

devices for extended periods. However, blockchain-

related activities necessitate a lot of communication

and energy-guzzling computers. As a result, IoT

devices cannot support Blockchain (Lee et al., 2007).

IoT Device Mobility and Partition: Blockchain

performance may suffer due to IoT device mobility.

Device mobility in a wireless network with

infrastructure support may increase the amount of

signaling and control messages sent. On the other

hand, network partitioning separates wireless ad hoc

networks into units when mobile nodes move along

different paths.

Latency and Capacity: Excessive latency is used in

decentralized Blockchain networks to guarantee

consistency. The latency that Blockchain frequently

tolerates is a concern for many IoT applications.

RELATED WORK

3.1

Fusion Chain

A key objective of FUSION is to develop a platform

level public chain in the crypto-finance era that

connects all values, conducts thorough financial

Optimized Fusion Chain with Web3 Storage

615

operations, interacts with a wide range of

communities and tokens, and links centralized and

decentralized organizations to expedite the Internet of

Values’ commercialization. Using distributed nodes,

FUSION will build a public chain that can map other

blockchain tokens to it and support cross-chain smart

contracts by storing the private keys of many tokens

on it. As well as improving the scalability and

interoperability of the current Internet of Values, it

will establish itself as a comprehensive crypto

financial platform:

• By supplying their private keys to the distributed

nodes of FUSION, tokens are interoperably mapped

to the FUSION public.

• Putting these off-chain assets and data under the

control of centralized institutions enhances

scalability. As a result, blockchain-based smart

contracts can interact with data and physical assets.

• The performance of FUSION needs to be

improved to handle a large number of tokens by fully

utilizing distributed nodes to perform distributed

parallel computing.

Fusion Chain’s purpose is to leverage blockchain

technology in order to provide an IoT-compatible

platform by providing a lightweight block structure,

InterPlanetary File System(IPFS), and IoT-friendly

PBFT consensus algorithm. However, the IPFS of the

fusion chain consumes a lot of bandwidth and offers

limited storage.

3.1.1

Block Structure Affects Storage

Overhead and Power Consumption

In the blockchain, a block contains some data fields.

The number of data fields inside data fields varies

depending on the blockchain platform design. In

comparison with other blockchain platforms, such as

Hyperledger and Ethereum, the fusion chain offers

the lightest block structure. The fewer data fields a

block will have in it, the calculation performed to

reach the consensus will become easier. This is how

it will reduce the storage overhead and thus gets less

CPU computational power. Figure 2 shows how the

block structure affects the power consumption.

3.2

Web3 Storage

Web3 storage is a decentralized platform of services

and APIs that comes with blockchain technology. It

is built on top of IPFS and Filecoin. It stores data in a

distributed manner on a network of multiple nodes. It

uses identity protocols like IPFS and Filecoin to store

data. Moreover, these protocols provide information

Figure 2: How the Block Structure Affects the Power

Consumption.

about the storage destination of the data and the

recovery process of it using its unique Content ID. In

order to interact with its service, it provides some web

user interface, client library, and URL.

It is trustworthy because all is in cryptography.

Content addressing and Filecoin proofs play a

prominent role here. It computes CIDs (Content

Identifiers) locally before uploading the data and

retrieving the data CIDs are validated locally. The

data is safe in Filecoin; the user can check the Filecoin

blockchain at any point to confirm. Users can pull the

data down via IPFS once it is on the network. Thus it

provides redundancy by storing the data on several

IPFS nodes. Filecoin webstorage token I’d creates

instance for storing the data. Then loT device data is

sent to Filecoin, and Content ID (CID) from Filecoin

is stored in blockchain. Figure 3 shows filecoin

system.

ARCHITECTURE

This section describes the system architecture. The

system architecture and the architecture flow

diagrams are depicted in Figure 4 and Figure 5,

respectively.

DATA 2023 - 12th International Conference on Data Science, Technology and Applications

616

Figure 3: Filecoin System:Global Services.

Figure 4: System Architecture.

Filecoin web storage is used as the storage system,

and the token ID is generated wherein we create an

instance for storing the data. So, we have a connection

among the Filecoin, IPFS, and that instance within

our node. When the new devices are added, a

structured json is created, and these json data will be

passed to the web storage that generates transactions.

Then the loT device data is sent to filecoin, and

Content ID (CID) from filecoin is stored in

blockchain. The IPFS storage will return CID, that’s

nothing but a hash value based on the file properties.

On storing the data, IPFS will return the hash to

retrieve the data from CID. These act as the API to

get data from filecoin from the web3 storage.

PBFT and RAFT Consensus algorithms are

applied to eliminate the malicious faulty nodes

preventing unauthorized nodes from validating bad

transactions in the blockchain network, thus

enhancing security. Figure 6 shows IPFS and Filecoin

Instance connection along with PBFT and RAFT.

Figure 5: Architecture Flow Diagram.

IMPLEMENTATION

This section contains information about the

implementation, which includes the frontend flow

and backend flow.

The fusion chain API has been re-written in the

implemented system, which is imported the

opensource fusion chain API (package.json in the

code) and re-written in the node.js backend by

modifying the json file to improvise the PBFT and

RAFT by setting it as the source package denoting as

iotblockchain-pbft-raft. It acts as the backend API

URL as /infusion. The code is written in such a way

by remapping the integrity and configuring the

dependency file with SHA-512 algorithm that

reduces computational power by enhancing the

transaction speed to get it stored in the blockchain.

The screenshots of the code implementation and the

json data are shown in Figure 7.

The process flow starts from the blockchain

network, which has individual nodes that represent

the IoT devices. When the IoT application is

executed, the backend API provides data to the

frontend through a port of peer-to-peer connection.

The nodes exhibit peer-to-peer connection

communicating with each other, which is collectively

within a blockchain network. It denotes home

automation control data connecting each other that

has to be stored in blockchain, which creates

transactions. Thus, performing the transactions on

blockchain also displays the number of nodes

connected in the frontend through the web socket.

Once the transaction has taken place, PBFT pre-

prepares announcing to all connected nodes that a

new transaction has been created to commit and

validate the signature from where it has been received

by verifying the transaction so that the successful

transactions are ready to get added in the blockchain.

Once all the nodes are validated by the PBFT

Consensus method, all the successful transactions

will be added to the blockchain eliminating the

malicious nodes.

Optimized Fusion Chain with Web3 Storage

617

Figure 6: IPFS and Filecoin Instance connection.

Figure 7: Configuring the Dependency File with SHA-512

Algorithm.

After which, the IoT device data will be stored in

the filecoin network passing the CID, which again

stores in blockchain upon retrieval keeping all the

transactions within the blockchain. The timed-out

nodes are handled by the RAFT algorithm through

electing the leader node by the follower node

instructions based on the vote to rejoin the network.

The handshake timeout is a time period in which a

node should establish a connection with another node.

In the peer-to-peer network handshake timeout issue

will happen when the establishment fails within this

period.

This implementation code can be found online

RESULTS

This section contains the results part. The output is

presented through the dashboard, and some messages

about the number of peers availability have been

displayed.

Our dashboard shows six sections: Devices, Met-

rics, Peers, Transactions, Blocks, and Comparision in

Figure 8.

https://github.com/61714100/Optimized-FusionChain-

with-Web3-Storage

In the Devices section, we can add a new device to

the home automation system, and then it will be added

to the dashboard and metrics. Here, we can also

choose the device power level and device name. The

addition of a new device is presented in Figure 9.

In the Metrics section, we can see the exact

information about the time and date of adding the

device. Moreover, it shows the turn-off time and usage

of the device as a graph. All these pieces of

information are shown in Figure 10 shows those

information.

The Peers section displays the average CPU usage,

current CPU usage, and current RAM usage for the

running node. Figure 11 shows those information. The

Transaction section shows the list of transactions.

When a new device will be added, it will have a new

transaction added. In Figure 12, we can observe the

blocks list, and then previously added and verified

transactions will be executed in the transaction

displaying the list of transactions.

The Blocks section shows the chain of blocks

added to the blockchain. When a transaction is made,

and the transaction reaches its threshold, a new block

will be added to the chain. The list of blocks will also

be displayed then. In Figure 13, we can observe it.

In Figure 16, we can see the transaction and block

has been added to the console.

Say, for example, it might have four transactions

and four blocks with two peers mostly, it’s using 100

MB min and even 1GB max till 1.4 percent that

displays metrics transaction comparison between

devices. This is the result example, the output that we

get when the code is executed.

The CPU and memory consumption is reduced to

1 percent and 3 percent, respectively. The CPU

utilization is reduced due to the implementation of a

lightweight block structure of fusion chain supporting

IoT devices through an optimized filecoin system. The

storage size is reduced due to the IPFS implementation

that decentralizes content storage through CID to store

data in the form of the hash value to the blocks in the

blockchain network.

DATA 2023 - 12th International Conference on Data Science, Technology and Applications

618

Figure 8: Dashboard.

The datasets aren’t a specific approach, but latency

time for loading data to IPFS, processor information

such as time taken, and storage utility can be used as

data to compare the system performance. The current

system utilizes minimal transactions and blocks

creation using 100MB min to 1GB max.

Figure 9: Adding New Device.

Figure 10: Device Usage Information.

We have executed the code in the M1 processor

Macbook Pro which has a RAM (Random Access

Memory) of 8 GB. Only two nodes could be run at a

time in this local machine. Moreover, it does not show

fluctuating percentage differences respective of the

devices as the loading data is in Kilobytes. In order to

add more nodes to the system and work with big-size

data, it will need a high-processor machine with

extended RAM. Moreover, this application is

executable both in Windows and Linux operating

systems.

Figure 11: Average CPU usage, Current CPU usage, and

Current RAM usage for Node.

Figure 12: Transaction Added in User Interface or Front

end.

CONCLUSION AND FUTURE

WORK

The design and implementation of the core idea of a

fusion chain blockchain with better storage is

implemented in this study. This work integrated web3

storage as improved storage from earlier work to

decrease power consumption and improve storage

quality (Na and Park, 2021). As for IoT devices, a

home automation system has been used to test the

performance of the work using the PBFT and RAFT

consensus algorithms. The proposed system execution

Optimized Fusion Chain with Web3 Storage

619

results in reduced CPU and memory consumption to 1

percent and 3 percent respectively, having the size of

the blockchain reduced that uses 100MB min to max

1GB for loading data with minimal transactions and

blocks.

Figure 13: Block Added in User Interface or Front end.

Figure 14: CPU Consumption Comparison: Fusion chain

using PBFT using 13% of CPU with four nodes. In the new

solution, code was optimized using the new concurrency

model CPU usage was reduced to 1%.

This project indicates a clear path toward the

implementation of the web3 storage, PBFT, and

RAFT inside the fusion chain platform’s internal

system architecture. Moreover, connecting real time

IOT devices like raspberry pi to merge with the

blockchain platform in order to analyze the

performance with real life implementation.

Figure 15: Memory Consumption Comparison: Fusion

chain using PBFT using 5 MB of memory with four nodes.

In the new solution code, it used 3 MB as we included

RAFT it required new memory to process.

Figure 16: New Transaction and Block Added in Console.

REFERENCES

Ahmed, I. and Shilpi, M. A. (2018). Blockchain technology

a literature survey. Int. Res. J. Eng. Technol.,

5(10):1490–1493.

Bosamia, M. and Patel, D. (2020). Comparisons of

blockchain based consensus algorithms for security

aspects. Int. J. Emerg. Technol., 11(3):427–434.

Ismail, L. and Materwala, H. (2019). A review of

blockchain architecture and consensus protocols: Use

cases, challenges, and solutions. Symmetry,

11(10):1198.

Lamport, L. (2019). The part-time parliament. In

Concurrency: the Works of Leslie Lamport, pages 277–

317. Lee, J.-S., Su, Y.-W., and Shen, C.-C. (2007). A

comparative study of wireless protocols: Bluetooth,

uwb, zig-bee, and wi-fi. In IECON 2007-33rd Annual

Conference of the IEEE Industrial Electronics Society,

pages 46–51. Ieee.

Na, D. and Park, S. (2021). Fusion chain: a decentralized

lightweight blockchain for iot security and privacy.

Electronics, 10(4):391.

Narayanan, A., Bonneau, J., Felten, E., Miller, A., and

Goldfeder, S. (2016). Bitcoin and cryptocurrency

technologies: a comprehensive introduction. Princeton

University Press.

Salimitari, M., Chatterjee, M., and Fallah, Y. P. (2020). A

survey on consensus methods in blockchain for

resource-constrained iot networks. Internet of Things,

11:100212.

Thakkar, P., Nathan, S., and Viswanathan, B. (2018).

Performance benchmarking and optimizing

hyperledger fabric blockchain platform. In 2018 IEEE

26th international symposium on modeling, analysis,

and simulation of computer and telecommunication

systems (MASCOTS), pages 264–276. IEEE.

Vujičić, D., Jagodić, D., and Ranđić, S. (2018). Blockchain

technology, bitcoin, and ethereum: A brief overview. In

2018 17th international symposium infoteh-jahorina

(infoteh), pages 1–6. IEEE.

Yao, W., Ye, J., Murimi, R., and Wang, G. (2021). A survey

on consortium blockchain consensus mechanisms.

arXiv preprint arXiv:2102.12058.

DATA 2023 - 12th International Conference on Data Science, Technology and Applications

620