The road to efficient and stateless clients on Ethereum

At the core of the Nimbus team’s development philosophy is a focus on building lightweight, high-performance Ethereum clients to further client diversity and network decentralisation.

Achieving widespread decentralisation and the ability to trustlessly and reliably verify chain data on lightweight, end-user devices is aligned with the long-term vision for Ethereum’s security and resilience.

Nimbus aims to build configurable, optimised Ethereum Validator clients capable of running on a broad range of hardware alongside Light Clients that can run on lightweight devices such as a mobile phone while still being able to trustlessly and fully verify the blockchain without relying on centralized service providers.

Among the biggest remaining challenges Ethereum faces in furthering its decentralisation is the requirement for nodes to store and update the blockchain’s historical and state data, which requires a significant amount of disk space. 

Statelessness, a situation where a full node does not necessarily need to store the state of the network, is envisioned as the endgame for Ethereum. The first steps to improving the efficiency of Ethereum clients are already being taken with EIP-4444 and The Purge, and the development of stateless clients is further explored through proposed optimisations that align with The Verge pillar of the Ethereum roadmap.

Let’s explore the steps to achieving lightweight and stateless clients on Ethereum, starting with a closer look at Nimbus’s client implementations.

Building lightweight clients for a high-throughput Ethereum

While Nimbus is focused on delivering the lightweight and performant Ethereum clients needed to realise the benefits of these proposed changes, the team is also committed to improving the scalability, throughput, and efficiency of the Ethereum layer-1 network.

Scaling the Ethereum network is a primary and immediate concern across the ecosystem, and Nimbus actively contributes to this effort through the development of its clients while upholding the principles of decentralisation that drive the vision of Ethereum one day running on portable devices.

Nimbus is developing and improving lightweight, high-performance execution-layer and consensus-layer clients with low resource requirements, making it easier to run an Ethereum node.

These efficient designs also benefit those with more powerful systems by freeing up resources for other tasks while maintaining reward performance and providing extra capacity during network stress.

To improve decentralisation for end-user devices such as the dGEN1 Ethereum Phone, Nimbus has developed the Verified Web3 Proxy, which verifies responses from RPC endpoints against the blockchain state, providing enhanced security and protecting against malicious or fraudulent behaviour.

Nimbus has also developed a client implementation for the Portal network, codenamed “Fluffy”, which allows node operators to offload historical blockchain data to and query data from a distributed storage network designed specifically for Ethereum.

The first step to scaling Ethereum: EIP-4444 and the Portal network

As mentioned earlier, one of the core obstacles hindering the ability to run Ethereum on lightweight hardware is the amount of storage space required to validate the blockchain state.

Ethereum Improvement Proposal 4444 (EIP-4444) aims to help solve the problem of data bloat by allowing Ethereum validators to prune historical data, moving this history to a trustless, secure, and distributed data storage network.

The implementation plan for EIP-4444 has been laid out, beginning with pre-merge data expiry being allowed from May 1, 2025. As node operators prune this data, it will need to be stored in a decentralised manner for nodes and users to reference in future. 

This is where the Portal network comes in.

The Portal network is an Ethereum Foundation initiative to bring trustless light client functionality to the layer-1 network. Its first and primary use case is to support the implementation of EIP-4444 by providing access to a decentralised archive of historical blockchain data.

Building and maintaining this network is critical to Ethereum’s future scalability and decentralisation. There are currently several active Portal clients storing historical Ethereum data, including the Nimbus Portal client, and you can track Portal network health and data availability here.

Note that as node operators begin to prune pre-merge data following the implementation of EIP-4444, many are helping to maintain a community-managed repository of era files as a fallback for the Portal network.

The widespread adoption of the Portal network and the continued pruning of data are crucial first steps toward scaling Ethereum, alongside the work on enabling stateless clients.

To play your part in helping Ethereum scale and remain decentralised, get started with Fluffy, the Nimbus Portal client.

The Verge: Unified binary trees and weak statelessness

Offloading historical data to a distributed network is an important first step in making Ethereum more accessible and decentralised, but the question still remains of how to address the need for node operators to store the blockchain’s state and to build stateless client implementations for Ethereum.

The Verge, a pillar of the Ethereum roadmap, proposes solving this by migrating Ethereum’s data structure from Merkle Patricia trees to a more advanced architecture to greatly optimise blockchain data storage and enable much smaller proof sizes.

Initially, The Verge proposed the adoption of Verkle trees to achieve this optimisation, but more recent solutions have shifted to adopting approaches such as unified binary trees instead for a number of important reasons:

• Verkle trees rely on cryptography that isn’t secure against quantum attacks, while the new proposal uses only hash functions, which remain safe in a post-quantum world.
• Advances in proving systems now allow for the generation of efficient state proofs without the need for Verkle trees' compact proof sizes.
• The unified binary tree design is likely the long-term solution, avoiding the need for future replacements of the tree architecture to ensure quantum resistance.

Read more about the proposal to switch Ethereum’s state tree to a unified binary tree in draft EIP-7864.

Over time, the scope of The Verge as a pillar of Ethereum’s roadmap has expanded beyond migrating the state tree architecture of the blockchain to encompassing the easy and cost-efficient verification of blocks in a broader capacity.

In one of his blog posts on the possible futures of Ethereum, Vitalik Buterin breaks down two key goals of The Verge. The first is to require no more than a few GB of storage of fully verifying nodes and staking nodes, and the second, longer-term goal is to be able to fully verify the Ethereum blockchain on a smartwatch.

These goals may be realised by a solution akin to “weak statelessness”, where the need for state storage in all nodes on the network is not entirely removed. Instead, specific nodes are responsible for state storage, while other nodes use optimised proofs to verify blocks without needing to store the full state.

To achieve this type of statelessness, the work done by validators will need to be separated across multiple nodes, with some proposing and others building blocks. This will help to further decentralise the validation process and improve the scalability of the network by enabling danksharding upgrades.

Read more about proposer-builder separation.

In this configuration, block proposers will still need to store the full state of the blockchain, even as other nodes use proofs to independently validate transactions without needing to store state data at all.

Even the short-term goal of The Verge, to run a fully verifying node with only a few GB of storage space, would drastically improve decentralisation and the resilience of the Ethereum network by allowing Ethereum clients to run on a much wider range of hardware.

Research is also ongoing into the possibility of implementing “strong statelessness” on Ethereum, where no nodes are responsible for storing full state data, and the state is aggregated across witness proofs leveraged by users on the distributed network.

The future of stateless clients on Ethereum

This vision of a stateless Ethereum is central to Nimbus’ long-term goals. We are working towards a future where Ethereum can run seamlessly on consumer devices – smartphones, tablets, and even smaller hardware – without users needing to be aware of the complexity underneath. 

Stateless clients will enable this, making Ethereum more accessible and decentralised than ever before, and Nimbus will be at the forefront of building and propagating lightweight and accessible Ethereum clients for both validators and end-users.Achieving statelessness may have a positive secondary effect on throughput, but Nimbus is also heavily focused on scaling Ethereum and improving transaction throughput at the layer-1 level. We ensure that while we aim towards the vision of stateless clients, we continue to push for network performance improvements in line with the principles of decentralisation and network resilience.

Today, our focus remains with the wider community on improving Ethereum’s scalability and throughput. In the long run, our goal is clear: to build the clients that allow Ethereum to run on everyday devices, bringing it closer to users while maintaining its integrity as a decentralised network.

If you want to help scale Ethereum and improve its decentralisation, you can start by downloading the Nimbus Portal client to play your part in offloading historical blockchain data from node operators.

Visit the Fluffy Quick-Start Guide for comprehensive documentation and step-by-step instructions on setting up your Nimbus Portal client. For more technical details, check out the GitHub repository.

Join the Nimbus community on Discord to get involved and provide feedback on the Portal client, and follow Nimbus on Twitter for the latest news and updates.