Shelby: Decentralized hot storage protocol competitive with AWS S3 performance (arxiv.org)

Shelby is a new decentralized storage protocol that claims Web2-grade throughput, latency, availability and cost-efficiency—competitive with AWS S3—for demanding, read-heavy applications like video streaming, large-scale analytics and AI training. The project targets a major shortfall in current Web3 storage: many decentralized systems are too slow, costly or unreliable for production workloads, so apps fall back to centralized providers. Shelby’s design aims to close that gap and enable truly decentralized, data-intensive services. Technically, Shelby borrows proven Web2 patterns while preserving decentralization: it separates control and data planes, runs RPC and storage nodes over a dedicated backbone, and uses erasure coding to cut replication overhead and minimize repair bandwidth. Reads are paid to economically reward high-performance providers, and the paper introduces a novel auditing protocol that provides strong cryptoeconomic guarantees without the heavy performance penalties typical of other proofs-of-storage schemes. The combination of paid reads, low-overhead redundancy, and efficient repair/audit mechanisms suggests a practical path to production-scale decentralized storage—though the reliance on a dedicated backbone and paid-read incentives highlights trade-offs in network topology and economics that will shape adoption.