Decentralized Storage and Computing Models for Web3 Platforms

Authors

  • Jonas Ivanov Assistant Professor, School of Data Science, Nordic Technical University, Stockholm, Sweden Author
  • Laura Costa Senior Lecturer, School of Data Science, Baltic AI Research University, Tallinn, Estonia Author
  • Lukas Costa Postdoctoral Researcher, School of Data Science, Mediterranean Institute of Technology, Rome, Italy Author

Keywords:

decentralised storage, IPFS, Filecoin, Arweave, decentralised computing, Akash Network, Web3 infrastructure, data availability

Abstract

A blockchain stores state, not data. An NFT token URI points to an image; the image itself lives somewhere else. A DAO proposal is a hash on-chain; the proposal text is hosted on a server. A DeFi protocol's frontend is a React bundle on Vercel -- one takedown notice away from disappearing. Web3 applications need decentralised storage and computing infrastructure that matches the decentralisation of the blockchain layer, yet most rely on centralised cloud services for everything except the smart contracts. We present the Decentralised Infrastructure Assessment Framework (DIAF), evaluating six storage platforms -- IPFS, Filecoin, Arweave, Ceramic Network, EthStorage, and Swarm -- and four decentralised computing platforms -- Akash Network, Golem, Fluence, and Internet Computer -- on real Web3 workloads including NFT metadata hosting, frontend deployment, off-chain data indexing, and decentralised ML inference. Our Infrastructure Quality Score (IQFS) measures data availability, retrieval latency, cost efficiency, censorship resistance, and developer experience. Arweave achieves the highest storage IQFS (0.896) through permanent one-time-payment storage. Akash achieves the highest compute IQFS (0.872) through Kubernetes-compatible decentralised deployment.

Author Biographies

  • Jonas Ivanov, Assistant Professor, School of Data Science, Nordic Technical University, Stockholm, Sweden

    Assistant Professor, School of Data Science, Nordic Technical University, Stockholm, Sweden

  • Laura Costa, Senior Lecturer, School of Data Science, Baltic AI Research University, Tallinn, Estonia

    Senior Lecturer, School of Data Science, Baltic AI Research University, Tallinn, Estonia

  • Lukas Costa, Postdoctoral Researcher, School of Data Science, Mediterranean Institute of Technology, Rome, Italy

    Postdoctoral Researcher, School of Data Science, Mediterranean Institute of Technology, Rome, Italy

Downloads

Published

2025-09-26

How to Cite

Decentralized Storage and Computing Models for Web3 Platforms. (2025). Blockchain, Web3 & Digital Trust Journal P-ISSN 3117-597X and E-ISSN 3117-5988, 2(3), 11-18. https://galaxiauniverse.com/index.php/BWDTJ/article/view/454