DVT Research and Standardization effort from the Nimbus team

Building upon our existing roadmap for adding secret shared validator functionality in Nimbus, the Nimbus team is proposing to conduct research into potential improvements to the existing Ethereum DVT protocols with the goal of making the upcoming LIDO DVT offerings maximally robust and easy to use.

Goals

Our research will focus on identifying and proposing missing standards that aim to:

  1. Deliver the best possible validator performance.
  2. Improve the inter-client interoperability.
  3. Facilitate the creation of more accessible and functional client management interfaces.

Deliveries

More specifically, we expect to complete the following deliveries:

  • Standardized formats for distributed keystores and keystore shares.
  • Keymanager API extensions for managing distributed keystores and keystore shares.
  • EIPs featuring precise cryptographic specifications and test vectors for the following processes:
    • Distributed keystore generation
    • BLS threshold signing
    • Keystore shares rotation
  • LibP2P protocols and Beacon API extensions for coordinating the execution of the validator duties in a variety of trust models.
  • Guidelines for the CLI interface of the clients.

Use cases

We believe that any research into DVT should strive to be as neutral as possible and should take into consideration the multitude of benefits and use cases that the technology could bring to:

  • Solo stakers and professional operators seeking to improve the resilience of their setups.
  • Professional operators seeking to gain protection from rogue employees.
  • New staking protocols aiming to distribute the responsibility of operating an Ethereum validator among a group of weakly trusted users.

What differentiates this proposal?

Before standardization, we will explore the problem space through pilot implementations in the Nimbus client that will aim to validate that the proposed standards can achieve the desired performance and robustness. We acknowledge that many of these problems have been addressed in prior art such as the SSV network, Obol, Diva and SafeStake, but none of these project offers a similar breadth of standardization. We intend the leverage the successful practices of these projects, but we will also try to innovate in few key areas:

  • Setups for solo stakers and professional operators can leverage the higher level of trust between the nodes of the DVT cluster by taking advantage of more efficient coordination algorithms, derived from Raft.
  • We believe that our deeper understanding of the factors (including corner cases) that affect validator performance will allow us to design more robust coordination protocols that are better integrated into the existing BN/VC architecture, and are easier to scale to a large number of validators.

Requested budget

To support this effort, we are requesting 50,000 DAI from the LIDO community, with 50% unlocked upon the completion of the listed deliveries. This would constitute one third of our projected budget and we are deeply grateful to the Ethereum Foundation and Ether.fi for their expressed support and encouragement for this initiative.

9 Likes

This sounds like a solid proposal for DVT Research. Whats the timeline for delivery?

2 Likes

We expect the initial pilot implementations in Nimbus and the first draft of the proposed protocols to be available within 4 months. Our team will continue improving upon the designs and will aim to achieve interoperability with other implementation teams by the end of 2024.

I was advised to specify the Ethereum address where the requested funds can be provided. The Nimbus team maintains a multi-sig wallet intended for grants and donations at 0xDeb4A0e8d9a8dB30a9f53AF2dCc9Eb27060c6557.

3 Likes

Great initiative, will help in any way we can

3 Likes

Thanks for this proposal! I’m excited to see more research being done in the DVT space from additional teams.

A few questions for you:

In your eyes, what does a successful outcome of the research look like at completion? Is the goal for all of the existing DVT providers to change their protocols to follow the same standard?

Assuming completion of the research, how much of the suggested changes do you expect to be in the realm of CL clients vs. DVT providers?

In terms of improving client-interoperability, is this deliverable specific to Nimbus, or will you be looking at other clients as well?

Thanks!

4 Likes

We would welcome any DVT staking protocol adopting our standards, but we’ll also continue to cooperate as much as possible with other competing standards. As an example, Nimbus fully supports the Beacon API extensions proposed by the Obol team and we are actively helping the Diva Staking team develop the best possible Web3Signer interface.

For us, a successful outcome is defined mostly in terms of achieved robustness and ease-of-use of the final solution. As outlined above, we’ll also be addressing some use cases that feel neglected at the moment.

While we can make commitments only on behalf of our own team to implement the proposed standards within Nimbus, we’ll try to get all other implementation teams involved in the standardisation process and we’ll try to make everyone’s job as easy as possible by providing comprehensive specifications, reference implementations and reusable building blocks. We expect that a minimum level of support in other clients will be possible through a modification limited to the validator client, but one of the main goals of standardisation for us is to promote implementation diversity in the DVT space (it has all the same benefits as the general Ethereum client diversity).

4 Likes

Thanks for the proposal Zahary and Nimbus team! Given the Lido DAO’s previous commitment to supporting all a wide array of DVT research, I think it makes sense to continue in this manner and support via LEGO.

One thing that I would personally love to se see is closer alignment between client teams and the DVT infrastructure “flavors” in alignment on an updated and more robust spec, so that we don’t just end up with rivalrous interpretations and different sets of requirements/asks for clients to keep up with.

4 Likes

Hey @zahary thank you for this proposal and I’m glad to say that the Council decided to support it!
The first leg was transferred.

1 Like

That’s amazing news! We express our deepest gratitude to everyone who made this possible :heart:

2 Likes