Scholarlink

<p>Governments worldwide are increasingly leveraging blockchain technology to streamline tax collection, owing to its immutable, decentralized, and transparent nature. India’s Goods and Services Tax Network (GSTN) is plagued with numerous challenges, including system downtime, lack of transparency, data security concerns, and an increased risk of tax fraud. Blockchain technology can promptly tackle these issues by ensuring real-time tracking of transactions and enhancing transparency within the tax collection process. By utilizing this technology, the GSTN could substantially reduce the likelihood of tax fraud and increase overall tax collection efficiency. However, it may introduce performance overheads that collide with the GST system’s inherent need for high throughput and low latency. This paper presents a performance evaluation of three GST systems: a centralized GST system, a blockchain-based GST system with Raft consensus, and a blockchain-based GST system with Dynamic Priority Leader Election (DPBLE). Performance evaluation is carried out using Hyperledger Caliper for blockchain-based GST systems and Apache JMeter for centralized GST systems. Additionally, comprehensive functional validation of the gtscc chaincode deployed on the GST-Blockchain-DPBLE system was performed to ensure correctness across all key functionalities, including invoice lifecycle management, GSTIN validation, ITC management, fraud detection, and audit compliance. Experimental results reveal that while the centralized GST system (GST-CS) achieves the highest throughput and lowest latency, maintaining latency between 12 ms (200 TPS) and 201 ms (2000 TPS), it suffers from longer recovery times due to reliance on externally managed failover mechanisms. The DPBLE-based blockchain system (GST-Blockchain-DPBLE) outperforms the Raft-based blockchain system (GST-Blockchain-Raft) by 22% in throughput and approximates 72% of the performance of GST-CS. Moreover, it exhibits superior fault tolerance, achieving an MTTR of 1332 ms even under six-node failure by prioritizing reliable nodes and reducing leader election delays. These results advocate the viability of the GST-Blockchain-DPBLE system for GST systems requiring both high availability and trust without sacrificing performance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Performance Evaluation and Validation of a Blockchain-Based GST Network with Dynamic Priority-Based Leader Election

  • 期刊论文

  • 发表日期：

    2026年2月28日

查看全文
  • Palak Aar,
  • Jawahar Thakur

摘要

Governments worldwide are increasingly leveraging blockchain technology to streamline tax collection, owing to its immutable, decentralized, and transparent nature. India’s Goods and Services Tax Network (GSTN) is plagued with numerous challenges, including system downtime, lack of transparency, data security concerns, and an increased risk of tax fraud. Blockchain technology can promptly tackle these issues by ensuring real-time tracking of transactions and enhancing transparency within the tax collection process. By utilizing this technology, the GSTN could substantially reduce the likelihood of tax fraud and increase overall tax collection efficiency. However, it may introduce performance overheads that collide with the GST system’s inherent need for high throughput and low latency. This paper presents a performance evaluation of three GST systems: a centralized GST system, a blockchain-based GST system with Raft consensus, and a blockchain-based GST system with Dynamic Priority Leader Election (DPBLE). Performance evaluation is carried out using Hyperledger Caliper for blockchain-based GST systems and Apache JMeter for centralized GST systems. Additionally, comprehensive functional validation of the gtscc chaincode deployed on the GST-Blockchain-DPBLE system was performed to ensure correctness across all key functionalities, including invoice lifecycle management, GSTIN validation, ITC management, fraud detection, and audit compliance. Experimental results reveal that while the centralized GST system (GST-CS) achieves the highest throughput and lowest latency, maintaining latency between 12 ms (200 TPS) and 201 ms (2000 TPS), it suffers from longer recovery times due to reliance on externally managed failover mechanisms. The DPBLE-based blockchain system (GST-Blockchain-DPBLE) outperforms the Raft-based blockchain system (GST-Blockchain-Raft) by 22% in throughput and approximates 72% of the performance of GST-CS. Moreover, it exhibits superior fault tolerance, achieving an MTTR of 1332 ms even under six-node failure by prioritizing reliable nodes and reducing leader election delays. These results advocate the viability of the GST-Blockchain-DPBLE system for GST systems requiring both high availability and trust without sacrificing performance.