HFBTP: A Blockchain Performance Dataset

This dataset is obtained from our blockchain performance test using Hyperledger Caliper on Hyperledger Fabric 2.3.
There are eight fields in it.
Set Transaction Arrival Rate: the transaction arrival rate we set (from [10-200], step length 5)
Actual Transaction Arrival Rate: the transaction arrival rate in real-world situations
Block Size: the block size we set (from [10-800], step length 5)
Orderers: the number of Orderer nodes on Hyperledger Fabric blockchain (from [3, 5, 7, 9])
Throughput: Throughput (TPS)
Avg Latency: average latency (seconds)
Min Latency: minimum latency (seconds)
Max Latency: maximum latency (seconds)

For a fixed transaction delivery rate, we set different block sizes to test its blockchain performance. For each data, we conducted 1000 experiments and finally took the average value to get the least error and the most accurate performance test results.

Our dataset will be helpful for researchers to find the relationship (pattern) between block size and blockchain performance in Hyperledger Fabric. Obviously, most blockchain users do not change the block size (they tend to use the default block size) and they are not aware of the impact of different block sizes on the blockchain performance.

With this dataset, researchers can use methods such as deep learning to uncover the relationship between block size and blockchain performance to find the optimal block size (other parameters that may affect blockchain performance) in different scenarios.

Note that if this dataset is helpful for your research, we are grateful and hope you can cite these papers as follows.

[1] J. Wang, C. Zhu, C. Miao, R. Zhu, X. Zhang, Y. Tang, H. Huang, and C. Gao, "BPR: Blockchain-Enabled Efficient and Secure Parking Reservation Framework With Block Size Dynamic Adjustment Method," IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 3, pp. 3555-3570, March 2023, doi: 10.1109/TITS.2022.3222960.

[2] J. Wang, Y. Wang, X. Zhang, Z. Jin, C. Zhu, L. Li, R. Zhu, and S. Lv, "LearningChain: A Highly Scalable and Applicable Learning-Based Blockchain Performance Optimization Framework," IEEE Transactions on Network and Service Management, vol. 21, no. 2, pp. 1817 - 1831, April 2024, doi: 10.1109/TNSM.2023.3347789.

This will also motivate our future work, and we will continue to contribute more valuable datasets.