Bankless: A Review of Centralization Pain Points in the Ethereum Ecosystem and Their Solutions

Original Title: Where Ethereum is Still Too Centralized

Author: Bankless

Compiled by: Mary Liu, BitpushNews

The Ethereum ecosystem is more centralized than it needs to be.

We often praise the accessibility of staking, but rarely discuss the large amount of Ether staked in Lido. We support decentralized currency, yet still heavily rely on centralized stablecoin issuers. We criticize L1s for compromising on decentralization principles, yet we have similar issues in our own scaling solutions.

However, there is hope, as developers often have solutions in centralized areas. This article will explore the centralization pain points of Ethereum in staking, stablecoins, and rollups, as well as the measures being taken to address these issues.

Staking

Pain Point: Ethereum's Security is Centralized Among a Few Platforms

Lido's liquid staking derivative (LSD) "stETH" dominates DeFi; this token has the highest liquidity among all staking derivatives, standing out as a blue-chip liquid staking asset, with Lido controlling 17.6% of Ethereum's TVL. One in three Ethereum validators (30.6%) belongs to this protocol, and its staking derivatives hold a 74.9% absolute share in the LSD market.

While many decentralized protocols (like Rocket Pool) allow anyone to be a node operator, Lido's current V1 version only allows whitelisted operators to run the protocol's validators, with pre-approved partners staking large amounts of ETH, which poses a centralization "risk."

As of Q4 2022, Lido had 29 active Ethereum node operators.

Solution: PoS and Withdrawals

Lido is an easy target for comparison, but it should be acknowledged that Ethereum's transition to proof of stake consensus (the Merge) has contributed to the network's decentralization.

Lido may control 30.6% of the validator set, yet the largest Bitcoin mining pool controls 31.7% of the hash rate. The top 12 BTC mining entities collectively control 96.1% of the network's hash power, while the 12 largest stakers in the network hold 70.3% of Ethereum.

PoS systems serve small-scale participants, such as individual validators, better than PoW systems, allowing more participants to earn rewards for securing the network. Just like the previous Merge hard fork, Ethereum's latest network upgrade, Shapella, is a key step towards decentralized network security.

Shapella supports withdrawals, providing liquidity for staked ETH for the first time since the Beacon Chain launched in December 2020, offering options for stakers seeking liquidity: independent staking and staking pool solutions are now effective competitors to LSD for those wanting to maintain liquidity.

At the same time, next-generation staking protocols are trying to provide more reasons for Lido users to withdraw than ever before. New staking methods (Stakewise V3 and ether.fi) are creating a competitive market where node operators compete for capital by offering superior staking yields, while LSD-Fi protocols like unshETH are attempting to decentralize staking through token incentives.

Stablecoins

Pain Point: Transitioning Reliance on Centralized Stablecoins

After Circle announced it had $3.3 billion trapped in the collapsed Silicon Valley Bank, its stablecoin faced indiscriminate sell-off pressure in the market.

USDC fell to a historic low of $0.82 on centralized exchanges, with liquidity evaporating from on-chain stable swap pools. Maker's DAI and Frax's FRAX, both heavily collateralized by USDC, inevitably became victims, trading at significant discounts during the sell-off.

Fortunately, the market recognized the danger and chose to diversify away from USDC, with the stablecoin's market cap dropping by 30.1% since the March banking crisis!

Unfortunately, there are still many lessons to be learned: we did not vow to abandon centralized stablecoins but simply denied USDC, turning to Tether's opaque-backed stablecoin USDT instead. USDT's market cap has risen by 13.6% since the Circle incident, approaching the all-time high set before the Terra-Luna collapse.

Solution: Building Next-Generation Stablecoins

The extremely limited reserve transparency of centralized platforms is merely a stopgap; the crypto market should turn to truly decentralized stablecoin alternatives, with non-seizure/recovery and fully transparent on-chain stablecoin substitutes becoming the trend.

Although Reflexer’s RAI has long been praised for having these advantages, its reliance on ETH as the sole collateral is limited in scale, and the ongoing negative redemption rate hinders adoption, as no one wants to hold a "stablecoin" that depreciates against the dollar!

HAI is a RAI fork set to deploy on Optimism soon, aiming to address its predecessor's shortcomings by accepting yield-generating stETH as collateral, hoping to turn the redemption rate positive.

Meanwhile, Berachain's HONEY stablecoin will be collateralized by various assets staked to the chain's unique liquidity proof consensus layer, helping HONEY achieve UST-like scalability without the risks of endogenous collateral.

Rollups

Pain Point: Rollups are Centralized and Lack Necessary Security

Sequencers sort L2 transactions and build local chains, functioning similarly to validators on Ethereum but requiring higher trust assumptions.

While the single sequencer model used by most rollups is efficient and provides low gas fees and fast transaction times, it is controlled by a single entity, representing a threat to network vitality and raising questions about centralization.

For a long time, Alt-L1 chains, aside from Bitcoin, have been criticized for sacrificing uptime and decentralization in pursuit of low fees and high TPS. Currently, Ethereum's scaling experiments yield similar results, undermining the benefits of transacting on a trustless network.

A quick glance at the "risk" classifications of the top ten Ethereum scaling solutions on L2BEAT reveals a significant gap between the current state of L2 and our ideal vision.

Optimism and forks like Metis Andromeda lack fault-proofing, while leading rollup solution Arbitrum only allows whitelisted participants, meaning malicious sequencers could include false transactions without permissionless state validation.

Additionally, many rollup contracts are upgradeable, a feature that allows owners to change the code (presumably to fix bugs). Unfortunately, updating battle-tested code resets the Lindy clock and may introduce errors, compromising the integrity of rollups. The $200 million Nomad bridge attack in August 2022 proved the pitfalls of upgradeable contracts.

Solution: Implement Decentralized Sequencers and Continue Building

A vague commitment to decentralizing the sequencer set can be seen on every rollup's roadmap. The difficult part is how to achieve it! For example, Optimism lists "decentralized sequencers" as a goal but notes that "the precise mechanism for sequencer rotation has not yet been finalized."

While possible L2 teams are privately researching internal solutions to achieve sequencer decentralization, various exciting external solutions are emerging to address this pain point.

One potential solution gaining attention is "based" sequencing, as explained by Ethereum researcher Justin Drake in a forum post on March 10. In this model, L1 block proposers collaborate with seekers and builders to include the next rollup block as part of an L1 block without permission.

Based rollups inherit Ethereum's liveness and security guarantees, but their throughput is limited by L1's transaction ordering and data availability. Additionally, transaction confirmations are constrained by a twelve-second block confirmation time, eliminating the rapid pre-submission enjoyed by L2 traders like Arbitrum.

Espresso, Astria, and Radius are some innovative protocols aiming to provide middleware sequencer solutions, allowing rollups to access a decentralized sequencer network while avoiding the L1 transaction processing bottleneck based on sequencing.

In addition to decentralized sequencer sets, shared sequencer layers also promise native cross-rollup atomicity and bridging options for rollups entering the network, similar to Optimism's OP stack "superchain" vision. However, regardless of their construction, any Optimism or zero-knowledge rollups can access these shared sequencers. Furthermore, these solutions can quickly leverage EigenLayer's restaking to enhance capital efficiency and improve sequencer performance.

Unfortunately, regarding permissionless proof systems and the restrictive nature of rollup contract upgrades, more time and further development are the only remedies to improve the risk situation of rollups. The lack of effective fraud proofs and upgradeable contracts reflects the reality that achieving scalability on Ethereum today requires compromises.

Vitalik proposed a three-phase taxonomy for classifying rollups based on the number of required "training wheels." Phase one rollups, the second classification, have proof-of-work systems, while phase two "no training wheels" rollups eliminate the upgradeability of smart contracts.

Most rollups are still in zero phase. However, significant progress is being made, with Arbitrum recently reaching phase one for EVM-compatible rollups, and Vitalik predicting that all rollups will enter phase two within a year.

Conclusion

Without these new developments, Ethereum might not be struggling with these centralization issues. Crypto is at the forefront of permissionless financial innovation, and we are still on the path to solving these problems. Don't let one or two strange trust assumptions hinder your progress; let them serve as catalysts for transformative innovation. Overcoming them not only enhances security but also provides users with new (and potentially lucrative) opportunities.

The future is bright, and the crypto frontier is constantly evolving. Despite the shortcomings, we will continue to build.