The State Network: A Scalable State Sharing Protocol for Low-Storage Validator Nodes in Blockchain Networks

The perpetual growth of data stored on popular blockchains such as Ethereum leads to significant scalability challenges and substantial storage costs for full-node operators. Increasing costs may lead to fewer independently operated nodes, which poses risks to decentralization (and hence security), but also pushes decentralized app developers towards centrally hosted API services. This motivates lightweight participation models for validators that cannot store the full blockchain state.

We introduce the State Network , which allows low-storage validators to participate without storing the full state. The protocol treats the blockchain as both a replicated state machine and a distributed storage system. State is distributed across nodes and retrieved via a Kademlia-like routing protocol, thereby reducing validator storage requirements. Cryptographic proofs (e.g., Merkle proofs) allow nodes to ensure data correctness despite Byzantine nodes. While the protocol trades off data storage for increased network bandwidth, caching and Verkle proofs can further minimize increased bandwidth needs.

We evaluate the State Network using both a prototype and data-driven network simulations. A prototype in

demonstrates feasibility, achieving up to \(\sim\) 80% reduction in per-commit disk writes and maintaining stable \(\sim\) 3.6s block processing even when most state is fetched remotely. Simulations confirm scalability: under realistic availability assumptions (30% node unavailability and a \(10^{-9}\) loss target), a homogeneous configuration with prefix length \(p=5\) achieves about 97% aggregate storage reduction in a 1,000-node network. The additional bandwidth remains moderate (1.5–5 MB per block) and can be further reduced through caching and more space-efficient Verkle proofs. Block propagation remains within Ethereum's 12-second slot window.

Limitations and future work include improving data availability guarantees beyond the probabilistic model by mitigating issues arising from uneven prefix length selection and adversarial selective withholding and by introducing incentives, and adapting the protocol to alternative overlay network designs and the upcoming Verkle-based state structure.