Single Slot Finality: The Missing Piece in Ethereum’s Usability Puzzle?

It’s no secret that Ethereum has a usability problem: transaction fees can cost several dollars and take around fifteen minutes before they are considered confirmed and finalized. While the fee problem is being solved by danksharding and Layer 2s like Optimism and Arbitrum, the issue of finality remains. Fortunately, a solution that could ultimately allow Ethereum transactions to be finalized after a single block, around 15 seconds, is in the beginning phases of research - single slot finality. 

A blockchain is, at its most basic, a distributed digital ledger. What makes it such an innovation is the ability to achieve trustless consensus about the ledger's state in a cryptographically secure manner. For every block, about 10 minutes for Bitcoin and 12 seconds for Ethereum, a validator earns the opportunity to batch the transactions since the last block into a new block and publish it for the consideration of other validators. If validators reach a consensus about the block, it is considered a part of the chain, and the process continues. 

However, this process of block finalization is not instant. In Ethereum, two-thirds of validators must attest to a new block’s accuracy, and with over half a million validators and a low minimum validator computing requirement, it takes about 15 minutes for a block to reach finality. Before finalization, there is the potential that a block could not end up in the finalized chain, such as if an attacker created a different chain not including that block to extract lucrative MEV from transactions. This is why centralized exchanges require users to wait around 15 minutes for their Ethereum transaction to be confirmed.

Other blockchains and consensus mechanisms allow for instant finality but with the trade-off of either security or decentralization. For example, the Algorand blockchain guarantees finality as long as an attacker holds less than one-third of ALGO but also can produce empty blocks in the case of a potential validator disagreement. Tendermint, the consensus mechanism of Cosmos and any Cosmos SDK-based chains, also provides instant finality, but halts if more than one-third of validators go offline.

Fortunately, a solution is possible, called single slot finality. As the name suggests, SSF would allow blocks to be finalized within the same slot they were confirmed. While certainly a complex solution, its basis is made possible by a few concepts. First, since the design of Etherueum’s current consensus mechanism, there have been advances in cryptography that have proven the cryptographic signature mechanism used by Ethereum, called the BLS signature aggregation scheme, can be computed significantly faster. Moreover, since Ethereum’s validator set is so massive, trade-offs can be made between decentralization, scalability, and security that do not significantly hamper the protocol’s resilience to malicious attacks. One proposal is for supercommittees with random validators to be randomly chosen and vote on a block instead of the entire validator set. This would drastically increase the speed of voting but also reduce the requirement to attack Ethereum to two-thirds of the staked Ether in the supercommittee instead of two-thirds of the entire validator set. With a proposed size of 125,000, this would still require 2.6 million ETH, an amount higher than any singular address holds and worth approximately $52 billion.

Another potential change might be to change the validator set to create a new role for specialized signature aggregators, similar to the proposed changes to block building by separating the block proposer from the block builder. This would allow for improved signature aggregation, the main bottleneck, while ensuring network decentralization and a low minimum requirement for validators.

Several problems still need to be solved before SSF can be implemented in Ethereum. It will likely be the focus of the protocol only after Danksharding, proposer builder separation, and Verkle trees are integrated on the mainnet. Regardless, the problem of bringing Ethereum’s transaction finalization time down from 15 minutes to 15 seconds is critical to the long-term success of the protocol and would make the user experience similar to that of a centralized payment network like Visa or PayPal, bringing Ethereum even closer to its goal of becoming a seamless global settlement layer.

By Lincoln Murr

Related Articles