Towards Stateless Clients in Ethereum: Benchmarking Verkle Trees and Binary Merkle Trees with SNARKs

2025-04-18

AI Summary

Ethereum faces growing decentralization risks due to state bloat, which raises hardware requirements for validators. This work systematically benchmarks two stateless client paradigms: Verkle trees—based on vector commitments—and binary Merkle trees augmented with zk-SNARKs—assessing their practicality in terms of proof overhead, verification efficiency, and state size reduction. Experimental results show that Verkle trees achieve a superior trade-off between proof generation/verification latency (~seconds) and proof size (~1 MB), whereas zk-SNARK–based solutions offer constant-time verification but suffer from prohibitively slow proof generation. To our knowledge, this is the first empirical, head-to-head evaluation of these two cryptographic constructions in the context of Ethereum statelessness. We open-source our implementations, providing concrete performance data and design insights to inform the community’s path toward stateless Ethereum.

Technology Category

• Search and Optimization: Distributed Search
• Data Mining & Knowledge Management: Data Compression
• Machine Learning: Hardware-aware ML

Application Category

• Security and Privacy: Blockchains and distributed ledgers
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• Economics, Online Markets and Human Computation: Economic aspects of blockchain and cryptocurrencies

Abstract

Ethereum, the leading platform for decentralized applications, faces challenges in maintaining decentralization due to the significant hardware requirements for validators to store Ethereum's entire state. To address this, the concept of stateless clients is under exploration, enabling validators to verify transactions using cryptographic witnesses rather than the full state. This paper compares two approaches currently being discussed for achieving statelessness: Verkle trees utilizing vector commitments and binary Merkle trees combined with SNARKs. Benchmarks are performed to evaluate proving time, witness size, and verification time. The results reveal that the Verkle tree implementation used for benchmarking offers proving and verification times on the order of seconds and proof sizes on the order of one MB. The SNARK-based Merkle trees exhibit slow proof generation times, while offering constant and fast verification time. Overall, the results indicate for Verkle trees to provide a more practical solution for Ethereum's stateless future, but both methods offer valuable insights into reducing the state burden on Ethereum nodes. We make the code used for benchmarking available on GitHub.

Problem

• Reducing Ethereum's state storage burden on validators
• Comparing Verkle trees and SNARK-based Merkle trees
• Evaluating performance of stateless client solutions

Innovation

• Verkle trees using vector commitments
• Binary Merkle trees with SNARKs
• Benchmarking proof and witness sizes