CTA: “In these threads, we attempt to further the discussion of a key problem in this category and evolve our understanding of the domain space where research work has not yet answered the specific problem or question being considered. These posts are living documents, and it is our hope that the community will continue to contribute to their structure and content.”
Key Problem / Topic Area
Ethereum’s transition to Proof-of-Stake (PoS) introduces several changes to its security model, especially as it relates to Maximum Extractable Value (MEV), but these dynamics are poorly understood.
Specific Question or Problem Statement
What intrinsic changes did The Merge introduce to Ethereum?
Approach / Methodology
The author provides technical background on the key changes introduced by The Merge, evaluates its potential implications to security, as well as the game-theoretic principles underpinning the new system.
Proof-of-Work (PoW): a mechanism introduced by Bitcoin to solve the coordination problem that distributed, permissionless networks face. This problem has been formalized as the Byzantine General’s Problem (BGP) and Bitcoin was the first network to solve it in a decentralized setting. PoW involves allocating computational resources to effectively “brute-force” the solution to a puzzle. Successful parties receive a reward for solving the puzzle and are granted permission to add a new block to the blockchain. Entities involved in this activity are called miners.
Proof-of-Stake (PoS): a mechanism that borrows the key coordination dynamics of PoW, but instead of brute-forcing the creation of new blocks, it coordinates block production on the basis of bonded participants. These are users that have locked up capital, or stake, and that synchronize on block production. If they misbehave, their stake is destroyed, or slashed. As such, relative to PoW, PoS is substantially less energy intensive.
The Merge: Ethereum was launched as a Proof-of-Work (PoW) network and was home to the world’s second-largest mining ecosystem. Some estimates suggest that this activity on Ethereum consumed 0.05% of the world’s total electricity consumption. In order to change that dynamic, the Ethereum team has been working on a Proof-of-Stake implementation for nearly 8 years, even predating its mainnet launch in 2015. That implementation went live in December of 2020 as a separate network called The Beacon Chain, which ran in parallel to Ethereum’s legacy PoW. This was done as a way to test PoS functionality before the full shift to PoS. The Merge is the moment where these 2 networks converge and Ethereum becomes a PoS system.
Source: Coin Metrics
Slot: in PoS Ethereum, the concept of “blocks” in a blockchain changes. In turn, that alters how the system measures time. Instead of blocks, time is measured in increments called slots. These can be defined as opportunities to create a block. At every slot, a stake is chosen to propose a block. This staker is called the Proposer and that entity has full control over the content of that block. Slots vary in time, but they are more predictable than blocks given the increased coordination in PoS. Roughly speaking, each slot takes an average of 15 seconds.
Attester Committee: once a block has been proposed by the Proposer, the network needs to agree (reach consensus) on whether this was a valid block. This is performed by another group of stakers called Attesters, whose role is to vote or attest to the validity of the block that was just proposed. This is done via a consensus mechanism called LMD-Ghost. This mechanism is similar to how blocks are agreed upon in PoW systems, but instead of computational power (hashrate) the attesters look at voting power (how many votes has the proposer’s block achieved).
Epoch: the Attester Committee uses LMD-Ghost to decide which blocks are valid, but there is another level of consensus required in order for a set of blocks to be deemed final. Simply relying on the Attester Committee would not offer strong guarantees that the blocks on the blockchain will persist, a property called finality. In order to provide stronger finality guarantees, Ethereum’s PoS uses another protocol to reach consensus on a wider set of blocks: Casper-FFG. This is done in set intervals of 32 slots which are called Epochs.
LMD-Ghost + Casper-FFG: Combined, these two protocols are the backbone of Ethereum and they intend to provide strong finality guarantees to network participants. Nevertheless, this system is more complex than PoW and stakers need to have a level of technical expertise, in addition to a balance of 32ETH, in order for them to be able to participate in this process.
Liquid Staking: Given the aforementioned challenges and complexities associated with participating in Ethereum’s consensus process, many users have preferred to delegate their stake to third parties. These third parties are exchanges and staking services that stake on behalf of users. Fundamentally, staking involves posting a bond, or capital, which assures that there are financial consequences if the staker misbehaves. However, the advent of Liquid Staking has changed that dynamic. These are ERC20 tokens that reflect ETH that is staked, which effectively “unlocks” staked assets. The most popular liquid staking provider is a project called Lido, which now controls over 33% of ETH that is staked.
Source: Coin Metrics
Maximum Extractable Value (MEV): the dynamic depicted above has interesting implications with regard to Maximum Extractable Value (MEV), which is the economic benefit extracted by block producers (miners in PoW, staker in PoS) from ordering transactions in a block. Often, MEV serves an added revenue overlay that can drastically improve total revenues of block producers, especially in times when there are a lot of arbitrage opportunities on-chain.
Safety Considerations Post-Merge: the centralization seen in Ethereum staking closely mirrors the previous centralization seen in Ethereum mining. The top 5 stakers contribute to 65% of staked assets, much like the top 5 mining pools contributed roughly to 65% of the hashrate prior to the Merge. However, the attack threshold in Proof-of-Stake is lower than in Proof-of-Work. In PoS, that threshold is 33% and a single entity controlling 33% of staking is able to break consensus, whereas in PoW that threshold is 50% in order for consensus to break via a so-called 51% attack.
Conclusions / Key Takeaways
The primary goal of the Merge was to transition to PoS and the ultimate result of that is a considerable reduction in electricity consumption. However, it can be argued that under the existing set of stakers the system is less secure. Applications and users must treat it accordingly and potentially wait longer (2 epochs) before considering large transactions final.
CTA: Future Work / RFP
Much remains to be evaluated in this new system. In order to counter the centralization seen in the set of stakers in the network, it might be required to separate stakers from those that are composing blocks: a new system called Proposer Block Separation (PBS). This might be the only way that the attack vector introduced by centralization can be addressed.