Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

International Conference on Computational Intelligence, Data Analytics and Information

Sciences (ICCDI-2023)

Performance Benchmarking and Analysis of Blockchain Platforms

Anita Thakura, Virender Rangab, Ritu Agarwalc


Department of Information Technology
Delhi Technological University, Delhi, 110042, INDIA
a
anitathakur_2k21phdit504@dtu.ac.in
b
drvirender.ranga@gmail.com
c
ritu.jeea@gmail.com

Abstract: The popularity of Blockchain among researchers, developers, and tech-savvy is reaching the sky. Due to the Blockchain's growing attractiveness to
the new crowd, anyone familiar with cryptocurrencies must be aware of Blockchain. Blockchain technology provides trust in an untrusted environment of
computation and technology. Blockchain has the potential to store, manage and share information in a decentralized manner. This decentralized blockchain
property makes this technology robust and immutable to external malicious activities. In this paper, we have conducted a study to measure and assess the
performance of various blockchain platforms named Ethereum (private Deployment), Hyperledger Besu Ethereum Client, and Hyperledger Fabric. The
performance analysis, such as throughput, resource utilization, and latency, are also calculated. The results show that the Hyperledger Fabric performed very
well compared to Ethereum and Hyperledger Besu in all the performance measures.
Keywords: Hyperledger Fabric, Hyperledger Caliper, Ethereum, Hyperledger Besu, Chaincode, Smart Contract, Blockchain.

Abbreviation
A HLF: Hyperledger Fabric
1. Introduction B DLT: Distributed Ledger Technology
C MSP: Membership Service Provider
Blockchain technology has become extremely popular with the advent of
D SPoT: Single point of truth
cryptocurrencies and is continuously expanding exponentially. Blockchain
E SPoF: Single point of failure
is a continually updated and maintained list of transactions and an
F PoET: Proof-of-Elapsed Time
immutable decentralized ledger. Blockchain is recognized as the missing
G VSCC: Validation System Chaincode
trust layer for the internet. When a block is joined to the chain of the
H MVCC: Multiversion Concurrency Control
Blockchain, it is nearly impossible to delete or change the block since the
I IBFT: Istanbul Byzantine Fault Tolerance
attacker has to change the previous blocks too. Satoshi Nakamoto [1]
J OS: Operating System
introduced the P2P electronic cash system and popularized this DLT among
K Avg.: Average
society. Later, V. Buterin [2] introduced Ethereum, and the reach of
L TPS: Transaction per second
blockchain technology expanded to include an extensive and vast range of
applications. Buterin presented a paper on next-generation smart contracts public blockchain, and Multichain, Corda, and Hyperledger Fabric are
and decentralized application platforms referring to the limitation of the examples of private blockchains. Private blockchain systems additionally
bitcoin scripting language, such as the lack of Turing completeness. ensure that the Blockchain doesn't branch through the usage of specific
Ethereum supports all kinds of computation, including loops, and supports protocols. Hyperledger Fabric is a private blockchain that has emerged as a
the state of the transaction. Ethereum provides the functionality of a smart popular solution for many business problems and applications. Hyperledger
contract, a self-executing contract that permits trusted transactions between Fabric offers a wide range of use cases, flexibility, and ease of
parties without a central authority. A blockchain network can be private or implementation, making it popular among developers and researchers. The
public. The blockchain concept, which serves as the foundational core data public blockchain network has significantly more security, performance,
structure of public networks such as Ethereum [3], and Bitcoin [1], allows scalability, and transparency issues than the private network. Business
any participants to join the network, and that causes the problems such as solutions require low latency and high throughput blockchain networks to
high transaction fees, high resource consumption, low throughput, and rapidly handle the transaction in less time with minimal computational
many more. An essential characteristic of such systems is the anonymity of resources. With the advent of blockchain technology, people have been
the participants in public blockchain networks. A permissioned network is introduced to various public, private, and consortium blockchains. However,
ideal for businesses that need authenticated participants, and every performance concerns persist; therefore, offering a practical yet effective
transaction is verified by the authorized gatekeeper rather than an unknown solution is essential. The contribution of this study is to:
participant. To address the growing need for business-level blockchain • Offer an empirical analysis of the performance of blockchain
applications in many industries and businesses, developers presented a systems.
novel framework in response to this new archetype. This new private and • Evaluate the key metrics such as throughput, latency, and other
permissioned Blockchain only allows identifiable or authorized participants factors and analyze the significant variation in performance
to join the network, providing better information sharing, security, and • The performance of Hyperledger version v2.x, Ethereum, and
increased performance. A private blockchain is faster than a Besu are analyzed in this study, which also aids in understanding
the circumstances in which one platform facilitates higher
performance than the other platforms.
1

Electronic copy available at: https://ssrn.com/abstract=4385643


International Conference on Computational Intelligence, Data Analytics and Information
Sciences (ICCDI-2023)
The structure of the paper is as demonstrated: Section 2 provides preliminary Hyperledger Fabric design is very flexible and modular, allowing different
knowledge about the Hyperledger Fabric and the transaction flow in industries to use cases the blockchain potential with adaptability, efficiency,
Hyperledger Fabric. Section 3 illustrates the existing done on the HLF or any and innovation [8]. HLF employs container-based technology to enable
other platform relevant to measuring the performance of Hyperledger chaincode (smart contracts), which are written in the languages of JAVA,
Blockchain or any other blockchain platform. Section 4 describes the GO, and Node.js. The Fabric environment consists of types of entities such
methodology and the experimental setup to conduct the study. The as client, peer node, or endorsement, ordering services for ordering the
experimental results obtained from work are discussed in section 5, and section transaction and creating the block and MSP node for membership and
6 provides the conclusion of the study. identity management, as shown in fig.1. MSP verifies the client credentials
and provides the no SPoT and no SPoF [9].

2. Preliminary Knowledge

2.1. Ethereum

Ethereum, first proposed by V. Buterin in 2013 [2], is a distributed


computing platform based on a public blockchain having a Turing-
Complete scripting language. Smart contracts and cryptographic rules in
Ethereum enable anyone to create their own ownership and state transition
function rules. Smart contracts in Ethereum execute on Ethereum Virtual
Machine (EVM). In Ethereum, Smart contracts are written in HLL (high-
level language) and compiled into EVM code. To develop smart contracts,
java-script-like language, i.e., Solidity, is the most prominent and widely Fig. 1 - Overview of Hyperledger Fabric [9].
adopted language. Ethereum, the programmable Blockchain, was
The MSP exerts certificate authorities, essentially plug-and-play interfaces
launched in 2015 [5] and permitted the participants to create new
that verify or revoke the user certificate based on verified identity. In
applications. Unlike Bitcoin, Ethereum allows the participant to build
Hyperledger Fabric, the flow of transactions utilizes the simulate-order-
their own rules of ownership or operation of any complexity. Ethereum is
validate and commit model. Figure 2 illustrates the flow of transactions
a public distributed computing platform that enables the developer to
involving the endorsement/proposal, ordering, and validation phases.
download and set up the network as a private network with only
authorized nodes allowed to join the network.

2.2. Hyperledger Besu

Coded in Java programming language and Developed under the license of


Apache 2.0, Hyperledger Besu is an open-source Ethereum client.
Hyperledger Besu runs on both the private permissioned and public
Ethereum networks. Additionally, it can be tested on test networks like
Sepolia and Görli. A private network is not connected to the Ethereum
Testnet or Mainnet. Hyperledger Besu assists in the creation of enterprise
applications that need private network processing that is secure and highly
transactional. Hyperledger Besu supports the enterprise features of
permissions and privacy. The consensus protocols used in Hyperledger
Besu are Clique, IBFT 2.0, Proof-of-Authority, and proof-of-work
(Ethash). Besu uses PoA when participants are recognizable to one
Fig. 2 - Flow of transaction in Hyperledger Fabric [10].
another and there is a high level of trust between them [6][7].
The client submits a transactional proposal during the endorsement phase
2.3. Hyperledger Fabric specified by the endorsement policy to use the chaincode function, which
contains the write set, read set, and metadata [11]. The client uses its
With several significant advantages over other blockchain platforms and credentials to sign the proposal of the transaction and passes it to one or
distributed ledgers, HLF is an open-source private DLT platform several endorsing peers at the same time. After satisfying the endorsement
developed for usage in business contexts. The development of private policy of the transaction, peers send the endorsement to the client. The
blockchain networks provides a variety of capabilities in terms of client then prepares the transaction with the payload and transmits to the
customization and potential use cases. The implementation of ordering service the endorsement it has received from the endorsing peers.
Hyperledger Fabric provides distinctive properties suited for developing The ordering service enqueues and gathers the transactions received from
an extensive range of business and enterprise applications. The various channels by various clients during the ordering phase. The ordering

Electronic copy available at: https://ssrn.com/abstract=4385643


International Conference on Computational Intelligence, Data Analytics and Information
Sciences (ICCDI-2023)
service generates a block of pending transactions for each channel after a focusing on the evaluation of latency, throughput, and error rate. The results
few seconds (block timeout), maintains the order by timestamp, signs it are compared with the study by Nasir et al. [13] analyzed that v2.0
with its identity, and broadcasts it to the peers on that channel using the outperformed v1.0 and v.06. Thakkar et al. [12] examined the performance
gossip message protocol. Then comes the validation phase, where the bottleneck of HLF v1.0 and enhanced those hotspots by employing
committing peer and the endorsing peer receive the block, verify the parallelizing and aggressive caching for endorsement policy verification.

Table 1 - Literature on performance evaluation and analysis of blockchain platform

Performance

consumption
Throughput

Success rate
Blockchain

Scalability

Resource
Latency
Authors

Year

Pongnumkul et al. [17] 2017 HLF, Ethereum Yes No Yes Yes No No

Nasir et al. [13] 2018 HLF Yes Yes Yes Yes No No

Thakkar et al [12] 2018 HLF Yes No Yes Yes No No

Baliga et al. [19] 2018 Quorum Yes No Yes Yes No Yes

Kuzlu et al. [15] 2019 HLF Yes Yes Yes Yes No No

Monrat et al. [18] 2020 Ethereum, HLF, Corda, Quorum Yes Yes Yes Yes No No

Dabbagh et al. [16] 2020 HLF, Ethereum Yes No Yes Yes Yes Yes

Melo et al. [9] 2022 HLF Yes No Yes Yes No Yes

Khan et al. [11] 2022 HLF Yes Yes Yes Yes Yes No

Woznica et al. [20] 2022 HLF, Iroha, Sawtooth Yes Yes Yes Yes No No

Our study 2022 HLF, Ethereum, Hyperledger Besu Yes No Yes Yes Yes Partially yes

orderer's signature on the block, evaluate against the VSCC validation, They conducted experiments that introduced the optimization, such as
and then MVCC validation [12]. parallel VSCC validation, MSP cache and read/write during MVCC and
commit phase. Kuzlu et al. [15] analyzed the performance of HLF v1.4 in
terms of latency, throughput, and scalability by varying transaction counts,
speed, and types with the help of the Hyperledger caliper benchmarking
3. Related Work tool. The experimental results show that query types of transactions can
handle 200 transactions per second without any substantial network latency.
This section and Table 1 depict the performance evaluation performed on Dabbagh et al . [16] the performance of HLF and Ethereum using Fabric
the Hyperledger Fabric and other blockchain platforms, which assists versions v1.4.x, v1.3.0, v1.2.0, v1.1.0, and Ethereum 1.2.1. on latency,
developers, researchers, and readers in determining the compatible and throughput, success rate, and resource consumption. The results from the
suitable Blockchain based on industry requirements. experiments show that the Hyperledger Fabric outperforms Ethereum in
terms of throughput and latency, while Ethereum has an exceedingly high
3.1. Performance studies on Hyperledger Fabric: resource consumption and transaction success rate. Pongnumkul et al. [17]
investigated the execution time, throughput, and latency of Hyperledger
Nasir et al. [13] investigated the Hyperledger Fabric platform's Fabric v0.6 and the private Ethereum blockchain in varying workloads of
performance. They developed their own smart contract (chaincode) for up to 10000 transactions. The paper analyzed that the HLF achieved high
transferring virtual money. The experiment was carried out on throughput and low latency in the given workload.
Hyperledger Fabrics v1.0 and v0.6. The performance metrics such as
latency, throughput, and execution time are evaluated by variation in the 3.2. Performance studies on other Blockchain Platforms:
workload up to 10000 Tx. Results. The platform's scalability is then
calculated by making the variation in the number of nodes up to 20, and Monrat et al. [18] evaluated the scalability and performance of private
the results are compared. The experimental results show that v1.0 blockchain platforms, incorporating Quorum, Corda, Ethereum, and
performed significantly well as compared to v0.6. Dreyer et al. [14] Hyperledger Fabric. The platforms are evaluated and examined in terms of
conducted a study analyzing the performance of Fabric version 2.0, throughput and network latency by making variations in the number of

Electronic copy available at: https://ssrn.com/abstract=4385643


International Conference on Computational Intelligence, Data Analytics and Information
Sciences (ICCDI-2023)
transactions and nodes. The results from the experiment illustrated that Table 2 – Infrastructure setup and system specification
the performance of the HLF was more impressive than other private Component Description
platforms. Baliga et al. [19] measured the latency and throughput of the
Quorum blockchain using various consensus algorithms, including RAFT
CPU AMD Ryzen 5 5500U with Radeon Graphics
and IBFT, and workloads. In terms of throughput, the Raft consensus
RAM 16 GB
outperforms the IBFT. Due to the increased workload on the system, the
quorum provided lower throughput at an input transaction rate of 600 SSD 512 GB
Tx/sec. Woznica et al. [20] performed the scalability and performance Operating System Ubuntu v22.04 LTS
evaluation on the permissioned Blockchain incorporating the Hyperledger
Blockchain Infrastructure Hyperledger Fabric v2.2.9, v2.4.7, Ethereum
Fabric, Hyperledger Iroha with YAC consensus, and Hyperledger (private Deployment), Hyperledger Besu (
Sawtooth with the PoET algorithm. The performance is assessed based on Ethereum client)

the avg. transaction latency, transaction failure rate, and network Docker Engine v20.10.21

throughput and scalability are assessed by varying the network size. The Docker Compose v1.28.5
results show that Iroha's minimal latency is independent of transaction
Benchmarking tool Hyperledger Caliper v0.4.2, Latest
sending rate or network size.
Programming Language Go language, JavaScript

IDE VSCode

4. Methodology
The methodology adopted in this study is employed to measure the HLF permits the developer to use any consensus protocol, including SOLO,
performance metrics and analyze the performance of these different KAFKA, and RAFT. The databases supported by the HLF are CouchDB and
blockchain platforms. The result obtained from this study can assist the Level DB. The CouchDB is used here in the deployment model.
practitioner, developer, and learner in understanding bottlenecks and
adopting a suitable platform for business application development. 4.3. Benchmarking Tool and Test Environment:
To benchmark Blockchain's performance, the Hyperledger Caliper tool [21]
4.1. Key metrics definitions:
is incorporated in this study. The Hyperledger Caliper represents the
Key metrics must be measured and analyzed to study the blockchain multiple-client thread; one client (load-generating) submits the transaction to
platform's effectiveness. There are four primary vital metrics: Success Rate: the blockchain network on the participant's behalf, and another (observing
the success rate in the experiment depicts the count of successful client) checks for the status of the submitted transaction. Caliper provides the
transactions out of all the transactions sent. Latency: The benchmarking user with a predefined solution to test the Blockchain, such as Hyperledger
tool evaluates the max. Latency, min. Latency and avg. Latency. The Besu harnessing the Ethereum adapter, Ethereum, Hyperledger Fabric, and
latency is the time elapsed between the initialization of the request and the FISCO BCOS.
time the response is received. Throughput: Throughput is measured in
transactions per second, representing the number of transactions
successfully performed in a second. The high throughput with low latency
is an admirable feature in a blockchain. Resource Consumption: The
resource consumption depicts CPU usage, memory usage, and Network IO
in the Blockchain.

4.2. Blockchain platforms and infrastructure setup:


The study is conducted on Hyperledger Fabric v2.2 and v2.4, Ethereum
(private Deployment), and Hyperledger Besu Ethereum client latest version
described in table 2. Experiments are conducted on the personal computer
with the system specification, processor AMD Ryzen 5 5500U with Radeon
Graphics, 16GB RAM, 512GB SSD, and running ubuntu operating system
22.04 LTS. Benchmarking and testing blockchain performance are
implemented using the Hyperledger caliper [21] with version v0.4.2 and
Latest. In order to evaluate the performance of the Hyperledger Fabric
platform, the entire HLF deployment was installed as well as run on Linux-
Ubuntu OS. The installed binaries of the Fabric can be checked on the
docker container. The HLF component comprising peers, certificate
Fig. 3- A Hyperledger Fabric chaincode code snippet of function
authority, and ordering service is initiated as a docker container
Open( ) in GO Language [21]

Electronic copy available at: https://ssrn.com/abstract=4385643


International Conference on Computational Intelligence, Data Analytics and Information
Sciences (ICCDI-2023)
However, the "query" type of transaction in Ethereum and Besu handled the
1000 transactions without delay. The results also show that the for the open
and transfer function, the Hyperledger Besu Ethereum adapter worked well
compared to the Ethereum. In contrast, the latency for the query function for
HLF v2.2 and V2.4 is 0.01%

Fig. 4- A Ethereum smart contract code snippet of a contract simple


for money transfer in Solidity language [21]
The smart contract in the Hyperledger Fabric is known as the chaincode Fig. 5- Success rate comparison between the blockchain
and is popularly written in the GO programming language; similarly, the platforms.
Ethereum Smart contracts are written in the Solidity language. Smart
contract imparted by the caliper-benchmarking tool is a simple money
transfer code that includes the functions open( ) fig.3, query ( ), and transfer(
) where the open ( ) in chaincode is employed to open the account of the
participant; similarly, the query ( ) queries the node, and the transfer ( ) for
transferring the money from one account to another fig.4. and before
evaluating the Blockchain's performance, the smart contract is deployed in
the blockchain platform.

5. Experimental Results
The study's results incorporate the tests of all the parameters defined in the
performance metrics. The test results are analyzed as follows:

Case 1: Success Rate: Functions addressed in the chaincode, such as open,


query, and transfer, have 1000 transactions in our study. The success rate is Fig. 6- Latency comparison of the blockchain platforms.
calculated for all the blockchain platforms over 1000Tx. The success rate
shows how many transactions are performed successfully. The success rate
calculated here is:
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑢𝑐𝑐𝑒𝑠𝑠𝑓𝑢𝑙 𝑇𝑟𝑎𝑛𝑠𝑎𝑐𝑡𝑖𝑜𝑛𝑠
𝑆𝑢𝑐𝑐𝑒𝑠𝑠 𝑅𝑎𝑡𝑒 =
𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑡𝑟𝑎𝑛𝑠𝑎𝑐𝑡𝑖𝑜𝑛

Figure 5 shows that the Hyperledger Fabric version v2.2 has fewer
successful transactions than the other platform. The issue that appeared
during the execution of the transfer function was related to the MVCC
read_conflict. According to Chacko et al. [22], the MVCC read conflicts
occur when there is an increased time between the endorsement and
validation of a transaction. Ethereum and Besu have a 100% success rate
and outperform the HLF. Figure 5 also illustrates that the "query" type
of transactions has a 100% success rate in all the blockchain platforms.

Case 2: Average Latency: Latency is the difference between the


transaction's completion time and deployment time. Figure 6 shows that the
Hyperledger Fabric v2.2 and v2.4 handle all the transactions relatively well. Fig. 7- Throughput comparison between the Blockchain platforms.

Electronic copy available at: https://ssrn.com/abstract=4385643


International Conference on Computational Intelligence, Data Analytics and Information
Sciences (ICCDI-2023)
Case 3: Throughput: Figure 7 represents that the throughput for the "query" suitable for enterprise-level applications.
type transaction is almost equal in all the Blockchain, whereas throughput
for "open" type transactions and the "transfer" type transaction are quite
unsatisfactory. It has also been illustrated from the study that the
6. Conclusion
Hyperledger Fabric version v2.4 has increased throughput in comparison to
the other platforms. It has been observed from the study that the
Hyperledger Fabric has outperformed the other Blockchain by a significant Blockchain is a DLT that offers immutable and verifiable ledgers secured
margin. by the cryptographic algorithm and various consensus mechanisms.
Blockchain technology furnishes users with trust in the untrusted
environment, which helps achieve Blockchain's distinctive features, such as
security, privacy, scalability, and performance. Use cases of Blockchain
technology are not restricted to cryptocurrency but are widely adopted in
eHealth, eGovernment, eVoting, drug tracing, and many more industry
applications. Various blockchains, such as permissionless, permissioned,
and consortium, are endorsed by the developer to build projects and
business applications. Selecting the application's appropriate blockchain
platform, which offers low latency, high throughput, and low resource
consumption, is essential, making our study relevant to the user, developer,
or any tech-savvy interested in blockchain technology. This study presents
a performance analysis of the Hyperledger Fabric, Ethereum (private
Deployment), and the Hyperledger Besu Ethereum adaptor. The study
evaluates the blockchain platforms against the performance metrics, i.e.,
throughput, latency, success rate, and resource consumption. The study's
Fig. 8- Latency of the transaction in varying batch size. observations are as follows: the success rate of Ethereum and Besu is high
compared to the Hyperledger Fabric when transferring money from one
account to another. The average latency of the Hyperledger Fabric is
significantly less, which makes it faster. Hyperledger Fabric provides a high
throughput, making the HLF more efficient and the resource consumption
of the HLF is also significantly less compared to other blockchains. The
study has yet to explore the reasons for transaction failure, evaluation of
fault tolerance, and scalability. Varying the workload, the network size, the
number of organizations, and the number of nodes per organization to
evaluate the performance and scalability can be done in the future.

REFERENCES

[1] Nakamoto, S. (2008). A peer-to-peer electronic cash system. Online:


https://bitcoin. org/bitcoin.
[2] Buterin, V. (2013) Ethereum: A Next-Generation Smart Contract and
Decentralized Application Platform.Online:
https://ethereum.org/en/whitepaper/.
Fig. 9- Throughput of the transaction in varying batch size. [3] Dannen, C. Introducing Ethereum and Solidity: foundations of blockchains
programming for beginners.
[4] Benji, M, & Sindhu, M. (2018). A Study on the Corda and Ripple Blockchain
In Hyperledger Fabric, the batch size and the batch time are key factors in
Platforms, Advances in Intelligent Systems and Computing.
achieving high throughput. Batch Size specifies the number of
[5] What is Ethereum? Accessed 2022. [Online]. Available
transactions the orderer collects before cutting a block. In the resultant https://ethereum.org/en/what-is-ethereum/.
study, at the batch size with 100 message count, the latency and [6] Hyperledger Besu. Accessed 2022. [Online]. Available
throughput show linear/slower growth. https://besu.Hyperledger.org/en/stable.
[7] Hyperledger Besu. Accessed 2022. [Online]. Available
Case 4: Resource Consumption: The resource consumption incorporates https://www.blockchaincouncil.org/blockchain/Hyperledger-besu-
anexhaustive-guide/
CPU usage, memory usage, and Network IO. The average memory [MB]
[8] Hyperledger Fabric. Accessed 2022. [Online]. Available
usage of the Ethereum blockchain for open, query, and transfer functions https://www.Hyperledger.org/use/Fabric.
is 247, 602, and 363, and the average CPU% for the Ethereum for the [9] Melo, C., Oliveira, F., Dantas, J., Araujo, J., Pereira, P., Maciel, R., & Maciel,
same function as above is 1.86, 0.36, and 1.46, respectively. The obtained P. (2022). Performance and availability evaluation of the blockchain platform
result of the Ethereum and Besu is compared with the resource Hyperledger Fabric. The Journal of Supercomputing, 1-23.
consumption studies of existing studies. The study observed that the [10] Transaction Flow Hyperledger Fabric 2022. [Online], Available
https://Hyperledger-Fabric.readthedocs.io/en/release-2.5/txflo.
Hyperledger Fabric outperforms all the other blockchain platforms and is
6

Electronic copy available at: https://ssrn.com/abstract=4385643


International Conference on Computational Intelligence, Data Analytics and Information
Sciences (ICCDI-2023)
[11] Khan, D., Jung, L. T., Hashmani, M. A., & Cheong, M. K. (2022). Empirical In 2017 26th International Conference on Computer Communication and
Performance Analysis of Hyperledger LTS for Small and Medium Networks (ICCCN) (pp. 1-6). IEEE.
Enterprises. Sensors, 22(3), 915. [18] Monrat, A. A., Schelén, O., & Andersson, K. (2020, December). Performance
[12] Thakkar, P., Nathan, S., & Viswanathan, B. (2018, September). Performance evaluation of permissioned blockchain platforms. In 2020 IEEE Asia-Pacific
benchmarking and optimizing Hyperledger Fabric blockchain platform. Conference on Computer Science and Data Engineering (CSDE) (pp. 1-8).
In 2018 IEEE 26th international symposium on modeling, analysis, and IEEE.
simulation of computer and telecommunication systems (MASCOTS) (pp. [19] Baliga, A., Subhod, I., Kamat, P., & Chatterjee, S. (2018). Performance
264-276). IEEE. evaluation of the quorum blockchain platform. arXiv preprint
[13] Nasir, Q., Qasse, I. A., Abu Talib, M., & Nassif, A. B. (2018). Performance arXiv:1809.03421.
analysis of Hyperledger Fabric platforms. Security and Communication [20] Woznica, A., & Kedziora, M. (2022). Performance and scalability evaluation of
Networks, 2018. a permissioned Blockchain based on the Hyperledger Fabric, Sawtooth, and
[14] Dreyer, J., Fischer, M., & Tönjes, R. (2020, November). Performance analysis Iroha. Computer Science and Information Systems, (00), 2-2.
of Hyperledger Fabric 2.0 blockchain platform. In Proceedings of the [21] Hyperledger Caliper Benchmarking tool. [Online] 2022. Available
Workshop on Cloud Continuum Services for Smart IoT Systems (pp. 32-38). https://Hyperledger.github.io/caliper/.
[15] Kuzlu, M., Pipattanasomporn, M., Gurses, L., & Rahman, S. (2019, July). [22] Chacko, J. A., Mayer, R., & Jacobsen, H. A. (2021, June). Why do my blockchain
Performance analysis of a Hyperledger Fabric blockchain framework: transactions fail? A study of Hyperledger Fabric. In Proceedings of the 2021
throughput, latency, and scalability. In 2019 IEEE international conference International Conference on Management of Data (pp. 221-234).
on Blockchain (Blockchain) (pp. 536-540). IEEE.
[16] Dabbagh, M., Kakavand, M., Tahir, M., & Amphawan, A. (2020, September).
Performance analysis of blockchain platforms: Empirical evaluation of
Hyperledger Fabric and Ethereum. In 2020 IEEE 2nd International
Conference on Artificial Intelligence in Engineering and Technology
(IICAIET) (pp. 1-6). IEEE.
[17] Pongnumkul, S., Siripanpornchana, C., & Thajchayapong, S. (2017, July).
Performance analysis of private blockchain platforms in varying workloads.

Electronic copy available at: https://ssrn.com/abstract=4385643

You might also like