Scalable Blockchain Architectures for High-Throughput Decentralized Applications

Authors

  • Matteo Hansen Assistant Professor, Department of Artificial Intelligence, Swiss Institute of Machine Intelligence, Zurich, Switzerland Author
  • Daniel Rossi Research Scientist, Department of Machine Learning, Central European Tech University, Vienna, Austria Author

Keywords:

blockchain scalability, Layer-2 rollups, zk-rollups, sharding, modular blockchain, DAG consensus, DeFi throughput, decentralisation

Abstract

Scalable blockchain architectures are essential for enabling high-throughput decentralised applications (dApps) that require transaction processing rates exceeding the capabilities of first-generation blockchain platforms. Ethereum mainnet processes approximately 15-30 transactions per second (TPS), while enterprise and DeFi applications demand 10,000-100,000 TPS with sub-second finality. This paper proposes the Scalable Blockchain Architecture Framework (SBAF), a systematic evaluation of five scaling architectures -- Layer-2 rollups (optimistic and zk-rollups), sharding, parallel execution chains, modular blockchain stacks, and DAG-based consensus -- across four dApp performance categories: DeFi trading platforms, supply chain provenance systems, decentralised social networks, and NFT marketplaces. SBAF introduces the Scalability-Decentralisation Quality Score (SDQS) integrating throughput, finality latency, decentralisation preservation, and security guarantee strength. Key results: zk-rollups achieve the highest SDQS (0.912) balancing throughput (4,000+ TPS) with strong security inheritance from Layer-1; modular stacks achieve the highest raw throughput (50,000+ TPS) but lower decentralisation; sharding achieves the best native Layer-1 scaling (1,000+ TPS) without off-chain trust assumptions. The framework provides architecture selection guidance for dApp developers and blockchain protocol designers.

Author Biographies

  • Matteo Hansen, Assistant Professor, Department of Artificial Intelligence, Swiss Institute of Machine Intelligence, Zurich, Switzerland

    Assistant Professor, Department of Artificial Intelligence, Swiss Institute of Machine Intelligence, Zurich, Switzerland

  • Daniel Rossi, Research Scientist, Department of Machine Learning, Central European Tech University, Vienna, Austria

    Research Scientist, Department of Machine Learning, Central European Tech University, Vienna, Austria

Downloads

Published

2024-03-28

How to Cite

Scalable Blockchain Architectures for High-Throughput Decentralized Applications. (2024). Blockchain, Web3 & Digital Trust Journal P-ISSN 3117-597X and E-ISSN 3117-5988, 1(1), 1-9. https://galaxiauniverse.com/index.php/BWDTJ/article/view/436