UL-blockDAG : Unsupervised Learning based Consensus Protocol for Blockchain

B, Swaroopa Reddy and Sharma, G V V (2020) UL-blockDAG : Unsupervised Learning based Consensus Protocol for Blockchain. In: 40th IEEE International Conference on Distributed Computing Systems, ICDCS 2020, 29 November - 1 December 2020, Singapore.

[img] Text
Proceedings_International_Conference.pdf - Published Version
Available under License Creative Commons Attribution.

Download (923kB)

Abstract

In this paper, we propose a consensus protocol by considering the ledger as Directed Acyclic Graph (DAG) called blockDAG instead of chain of blocks. We propose a two-step strategy for making the system robust to double-spend attacks. The first step is the graph clustering algorithm based on spectral graph theory for separating the blocks created by the non-cooperating miners (attacker) in the blockchain network followed by the second step-the ordering algorithm based on the topological ordering of the blockDAG using the references included in block header. The first step is an unsupervised learning classification of the vertices of a graph into two classes. The simulation results show that the proposed clustering Algorithm based consensus protocol counter-attack the attacker’s double-spending strategy by eliminating the attacker blocks created during attacking phase from the confirmed list of the blocks. In bitcoin’s longest chain rule protocol, the ledger takes the chain of blocks and it operates with the overestimation of the network’s end-to-end propagation delay which results in a low transaction throughput. Bitcoin protocol guarantees the security through longest chain rule but it suffers from the limited transaction scalability. The proposed consensus protocol works better for higher block creation rates in turn improves the transaction throughput without compromising the security of the blocks from double-spending attack. © 2020 IEEE

[error in script]
IITH Creators:
IITH CreatorsORCiD
Sharma, G V Vhttps://orcid.org/0000-0001-7691-1706
Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: BlockDAG; Directed Acyclic Graph; Graph Clustering; Transaction Throughput; Unsupervised Learning
Subjects: Electrical Engineering
Divisions: Department of Electrical Engineering
Depositing User: . LibTrainee 2021
Date Deposited: 24 Nov 2022 11:10
Last Modified: 24 Nov 2022 11:10
URI: http://raiith.iith.ac.in/id/eprint/11413
Publisher URL: http://doi.org/10.1109/ICDCS47774.2020.00159
Related URLs:

Actions (login required)

View Item View Item
Statistics for RAIITH ePrint 11413 Statistics for this ePrint Item