Centralized Clouds or Distributed Ledgers: Evaluating Tradeoffs in Storage Architectures


  • PRATHAM SANTAPUR Department of Finance, School of Business, George Mason University, Fairfax, VA
  • Jiasun Li Department of Finance, School of Business, George Mason University, Fairfax, VA




Cloud storage services like Amazon S3 have become increasingly popular over the past decade, providing customers with flexible and scalable on-demand storage. However, concerns around vendor lock-in, lack of full control, and cost at scale have led to the emergence of decentralized storage models like Filecoin and IPFS. This research compares the performance and economic tradeoffs between centralized cloud and decentralized blockchain-based storage to evaluate the benefits and limitations of each approach. The analysis implements pricing functions based on cloud provider SLAs to account for factors like usage costs, service credits, and uptime guarantees. It also considers decentralized storage qualities like permissionless access, verifiable audits, and transparent incentives structures facilitated by blockchain. Experiments on AWS show centralized clouds currently outperform on latency while decentralized networks offer stronger resilience and fault tolerance. Ultimately, both models have advantages – clouds excel at convenience and speed while decentralization promotes openness and control. The findings provide storage customers a framework to assess which approach best fits their needs based on access patterns, scale, and cost requirements. This research aims to give unbiased technical guidance on adopting the optimal storage architecture. 





School of Business: Department of Finance