By now, you’ve probably heard about Distributed Validator Technology (DVT) and its importance for the future of Ethereum and other proof of stake chains (If not, you can check out our DVT intro).

As a quick refresher, DVT is a technology primitive that allows a single Ethereum validator to be split into multiple nodes that complete duties as a cluster. This builds in a level of fault tolerance—as long as a certain percentage of the nodes are active, the cluster as a whole will be active. This improves the current state of Ethereum by:

Increasing the resiliency of validators by reducing validator downtime and protecting against key compromise and byzantine behavior
Reducing slashing risk of honest validators, particularly those that are currently running active-passive setups
Allowing for operators to validate as a community (vs. single entity), creating active-active redundancy across geographies
Improving decentralization in Ethereum by mitigating stake centralization risk

You may have realized it already, but these benefits impact every type of validator: institutional validators, liquid staking protocols, community validators, and solo validators. In this series, we will go deeper into how DVT helps each of these types of validators.

(Note: We will be using Obol’s DVT middleware design (aka Charon) as the standard for exploring how DVT works and the benefits provided for each type of validator. There could be other approaches to DVT that would potentially impact each of these validator types differently.)

We will start by the type that is the largest and growing the fastest today—liquid staking protocols, or LSPs.

What are Liquid Staking Protocols?

Liquid staking is a mechanism with which stakers provide ETH to a provider in exchange for a receipt token that is (or will be once withdrawals are enabled) redeemable for the staked token. What’s critical is that the receipt tokens are also immediately liquid and can be used in trading, borrowing, lending, providing liquidity, held as collateral, or any other defi activity. This allows stakers to earn rewards without sacrificing all of the liquidity and flexibility of the underlying token.

There are now many liquid staking providers, with the top four in Ethereum being Lido, Coinbase, Rocket Pool, and Stakewise in terms of share of staked ETH.

What risks do Liquid Staking Protocols face today?

By most measures, liquid staking has been a wild success since its inception a few years ago. Today, LSPs makeup more than a third of all staked ETH. It is likely that this number will continue to grow post-Merge as more investors look to earn passive income on their crypto without completely sacrificing liquidity.

However, as the percentage of stake in LSPs grows, there are increasing risks to the Ethereum network (and other proof of stake networks) that need to be addressed to ensure long-term resiliency and decentralization.

Operator risk

All LSPs rely on their operator set to stake the ETH given to them. The performance of the protocol also depends on the quality of the underlying operators. If operators go offline, the protocol will lose out on staking rewards. To manage this risk, LSPs, like Lido, heavily evaluate the performance of their operators, tracking metrics like uptime, inclusion delay, and more. They also try to distribute the stake evenly across their operator set to reduce the risk from any single operator. Rocketpool, on the other hand, distributes their stake across 1,400+ operators to minimize exposure from any single operator. Stakewise, in their V3 release, is taking a unique approach to reducing operator risk by providing greater inclusion, transparency, and choice in selecting the operator to entrust stake to.

Lido operator distribution

While all those measures help to reduce operator risk, operators still represent single-points-of-failure with potentially $100M+ of staked ETH each. If an operator goes offline, that represents a significant amount of rewards that could be lost, and worse, if an operator becomes malicious, that could cause a mass slashing event for the liquid staking pool, impacting the entire Ethereum network. Let’s discuss this type of risk further.

Slashing risk

Another risk that’s related to operator risk is centered around slashing. Slashing penalties are mechanisms built into PoS protocols, like Ethereum, to prevent validator misbehavior. Back in September, there was a mass slashing event where 20+ validators were penalized due to an incorrect setup.

RIP. 20 validators slashed over the last 3 hours due to attester equivocation. Can't find much info with the validator indices and deposit addresses. My guess is duplicated setups / forgot doppelganger detection pic.twitter.com/TJupDCle8j
— terence.eth (@terencechain) September 23, 2022

To reduce downtime, many professional operators also utilize active-passive environment setups (i.e. a backup environment that’s spun up if the active environment goes down). This can significantly increase the slashing risk as the backup environment also controls a full copy of the private key, and in the event both environments are active and signing with the same key (perhaps due to a config or code error), that can create a significant slashing event.

Slashing risks only grow as stake becomes more centralized, and centralization can occur from correlation in any component in the validation tech stack. That leads us to the next type of risk.

Correlation risk

Both operator and slashing risk can be exacerbated by higher levels of correlation. Correlation risk occurs when there are shared components in a large percentage of validators within a liquid staking protocol. These could be shared consensus or validator clients, network, geography, hosting service, and more. Each shared component then becomes points of failure that, when representing a large portion of the network, could greatly affect a liquid staking protocol.

Today, 80% of the Ethereum network uses just two consensus clients (Prysm and Lighthouse) and only a single execution client (Geth). That fact poses significant correlation risks to any liquid staking protocol.

Ethereum consensus client distribution

Ethereum execution client distribution

Stake centralization

The final risk that LSPs face is associated with having large amounts of stake pooled within a few entities. Stake centralization can occur at different levels, from the operators to the liquid staking protocol itself. Like with correlation risk, having large amounts of stake increases the exposure of each liquid staking protocol as well as the network as a whole.

Over the history of the Beacon Chain, you do see a trend towards some LSPs gaining a larger percentage of stake. Even though each of these entities are making significant efforts towards lowering correlation risk and decentralization within their protocols, a greater distribution of providers, especially in smaller and solo validators, can help to reduce stake centralization.

Now let’s discuss how DVT can help address these challenges, but before we do, let’s first clarify how DVT works with LSPs.

How does DVT work in Liquid Staking Protocols?

As a middleware component, Obol’s Charon acts as an additive layer in the technology stack of any liquid staking protocol. Instead of running full validator nodes, operators would form Distributed Validator (DV) clusters of 4, 7, or 10+ nodes within their own environments or as a shared environment with other operators in the same liquid staking protocol (something we are actively testing with LSPs today). Each node in a cluster could be running on different machines with different consensus, validator, and execution clients. These machines could also be hosted in the same or different geographical locations.

Graphic showing how Obol's DVT middleware, Charon, works in a DV cluster

Next, let’s discuss how implementing DVT can help LSPs.

How does Obol DVT Middleware help Liquid Staking Protocols?

DVT greatly minimizes many of the risks that LSPs face today, improving resiliency, scale, and decentralization.

DVT improves LSP operator effectiveness

DVT helps to improve operator effectiveness by reducing downtime and slashing risk. If an operator implements DVT within their own environment, they can deliver active-active redundancy that allows validators to stay running even if a certain percentage of their nodes go offline. They also no longer need to run active-passive setups to improve uptime, which greatly reduces the risk of slashing.

Better yet, DVT can allow multiple operators in the network to collaborate and form inter-operator clusters. This means that even if an entire operator fails, the shared distributed validators would still be able to stay active and attest.

DVT improves operator participation

By eliminating single-points-of-failure for each operator, DVT also allows LSPs to increase their operator pool without adding additional risk. Smaller, unproven operators can be grouped with larger validators to create clusters that have sufficient redundancy. This provides a path for smaller operators to build their reputation over time without lowering the effectiveness of the protocol as a whole. In the long-term, we envision a future where operator sets for LSPs can be completely trustless, allowing even at-home validators to permissionlessly enter distributed validator clusters with other at-home validators or professional node operators.

DVT reduces correlation risk by diversifying client configurations and geographies

Since each node within a DV cluster can run different client configurations and run in different geographical locations, this greatly reduces the correlation risk associated with any single client or location.

Due to the large percentage of stake managed by LSPs, improving the diversity of clients, geographies, and other correlating factors in their operator set can go a long way to improving diversity in the overall network.

DVT allows LSPs to decentralize stake

DVT is a technology that improves performance and reduces the barriers for smaller, solo validators to participate in LSPs, and creates a path to trustless staking that allows any operator to join an LSP while minimizing the risk to the protocol. Many LSPs including Lido and Stakewise have publicly announced plans to incorporate DVT as a means to decentralize stake in their protocols.

The future of liquid staking with DVT

We strongly believe that liquid staking will continue to play a large part of how the Ethereum network and other PoS chains are secured. However, that could potentially be a centralizing force in the Ethereum network that needs to be counteracted. By implementing DVT, LSPs can continue to provide liquidity for staked assets while ensuring resiliency and decentralization in the network.

We are actively working with many major LSPs today to implement DVT and believe that soon DVT will be a standard middleware technology for all LSPs.