In October 2020, Ethereum co-founder Vitalik Buterin introduced the “Ethereum Roadmap Centered on Rollups.” Currently, we live in an ecosystem dominated by multiple Rollups. While Ethereum’s Rollup roadmap has brought significant advantages such as improving user experience, it has also triggered a range of new challenges. For instance, developers often face fragmentation and decision-making dilemmas when dealing with a diversified Rollup ecosystem. Moreover, centralized sequencers have led to issues like transaction censorship and Miner Extractable Value (MEV) exploitation, which cause some users to face unfair costs or operational experiences.

Current mainstream solutions include Shared Sequencers and Based Rollups to solve the issues posed by centralized sequencers. Shared Sequencers offer unified sequencing services for multiple Layer 2 chains, but as third parties, they carry trust and incentive mechanism problems. Based Rollups, on the other hand, rely on Ethereum L1 proposers for sequencing, which enhances decentralization. However, the transaction confirmation time depends on Ethereum L1’s block time (around 12 seconds), which cannot achieve fast finality.

The Historical Background of Preconfirmations

Preconfirmations, or “Preconfs,” can be understood as a form of “reliable preview before final confirmation.” A simple analogy is when you make a reservation at a restaurant: the restaurant confirms your reservation, but you haven’t arrived yet. This is a form of preconfirmation. You’ve received a trustworthy commitment that ensures a spot for you at the specified time, which saves you the time you would otherwise spend waiting in line.

Preconfirmations are not a fresh concept. In the early days of Bitcoin’s development, the Bitcoin community introduced the concept of “0conf.” This referred to treating a transaction as valid before the Bitcoin network fully confirmed it. This reduced the waiting time for final transaction confirmation. In 2023, Uri Klarman extended this idea to Ethereum by introducing the concept of “Chained Preconfirmations.” The core idea is that future preconfirmers can not only pre-confirm the current transaction but also inherit the preconfirmations made by all previous preconfirmers. Primev then further @muratlite/primev-v01#What-is-Primev">explored this concept, proposing a novel bidding mechanism called “Preconfirm (Pre-conf)” which allows a block builder alliance to commit to including a transaction in future blocks within a specific time window, so that users are allowed to participate in preconfirmation bidding.

(Primev is developing mev-commit, a decentralized platform specifically designed to coordinate MEV transactions. By using real-time cryptographic commitments and efficient bidding mechanisms, it helps users and providers ensure the reliability and speed of transactions. It is applicable to complex scenarios that require rapid confirmation and execution, such as preconfirmations and other high-frequency trading strategies.)

Justin Drake further advanced this concept by proposing the idea of “Based preconfirmations.” Currently, many Layer 2 systems use centralized sequencers, which offer a significant advantage in providing preconfirmations. This confirmation indicates that the sequencer has committed to including the user’s transaction in a future block. However, this confirmation is not final; users may still encounter transaction reordering or cancellation. The “Based Preconfirmation” mechanism proposed by Justin Drake aims to provide immediate preconfirmation while ensuring that L1 completes decentralized sequencing (Based Rollup).

The Proposal and Optimization of Based Preconfirmations Mechanism

The concept of Based Rollups was first introduced by Vitalik Buterin in his 2021 article “An Incomplete Guide to Rollups”. In this article, Vitalik introduced the idea of “Total Anarchy,” which describes an entirely decentralized environment where anyone could submit transaction batches at any time, without any restrictions.

In March 2023, Justin Drake further clarified the concept of Based Rollups in his article “Based rollups—superpowers from L1 sequencing”. When the sequencing of a Rollup is driven by L1, it is referred to as a Based Rollup or an L1-sequenced Rollup. In this architecture, the generation and sequencing of Rollup blocks are carried out by proposers on L1, who can collaborate with searchers and builders to directly include Rollup blocks in L1 blocks. Notably, by default, Based Rollup blocks are constructed by L2 builders, so Based Rollups do not add to the burden of L1 validators.

For further details on Based Rollups, you can check out the article What is Based Rollup, Inheriting Ethereum’s Vitality?

By transferring the responsibility of transaction sequencing to L1, Based Rollups inherit the security and liveness of L1 while optimizing the performance of L2. This mechanism brings several advantages, such as reducing infrastructure and operational costs for L2. However, challenges remain, such as the fact that most of the MEV (Miner Extractable Value) generated by Based Rollups flows to L1. This, however, is not a major issue since centralized Rollups primarily generate revenue from L2 transaction fees and MEV opportunities, with the former being the main revenue source and the latter requiring significant infrastructure and maintenance costs. Based Rollups preserve L2 transaction fees as their primary revenue stream while outsourcing the role of proposers to L1 builders, thus reducing infrastructure and operational costs. Another challenge is that Based Rollups cannot achieve rapid transaction finality. Since Based Rollup transaction confirmation times depend entirely on Ethereum L1 block times (currently about 12 seconds), this contrasts sharply with the quick soft confirmations provided by centralized sequencers.

To address the issue of transaction confirmation speed in Based Rollups, Justin Drake introduced the concept of Based preconfirmations in November 2023. It uses Slash execution to provide preconfirmation for Based Rollups. In the initial proposal, Justin Drake outlined two key prerequisites: first, L1 proposers must opt into additional penalty conditions to become preconfirmers (a re-staking mechanism); second, L1 proposers must be able to enforce the inclusion of transactions on-chain (inclusion lists).

Subsequently, Justin Drake optimized the proposal, which allows Rollups to use Ethereum L1 for shared sequencing and preconfirmation without the need for a hard fork. In the new proposal, some validators voluntarily become sequencers, while validators who do not wish to sequence transactions become “includers.” Includers can include transactions but do not sequence them. Sequencers can reorder transactions from includers and even insert additional transactions. Users can submit transactions in two ways: first, by sending the transaction to a backup mempool and paying a small fee for includers to include it; second, by communicating with the next sequencer, paying a higher preconfirmation fee to request preconfirmation. If the sequencer fails to settle the transaction within the designated slots, includers can still process these transactions.

Compared to the initial proposal, the optimized solution directly leverages Ethereum L1’s economic security, which allows L1 proposers to voluntarily become sequencers without additional penalty measures. This reduces the dependence on extra infrastructure.

From Concept to Practice: The Application of Based Preconfirmations

Taiko: The Pioneer in Developing and Promoting Based Rollups

Since the second half of 2023, Taiko has been evolving towards becoming a Based Rollup. As a Based Rollup, Taiko relies on a decentralized proposer system where proposers run specific clients (taiko-geth) and stay synchronized with the L2 mempool. When a proposer detects a batch of transactions that can form a profitable block, they submit the pre-compiled block to Ethereum L1.

In the transaction processing flow, user-submitted transactions on Taiko first enter the L2 mempool. L2 searchers seek profitable transactions within the mempool and package them into L2 transaction batches. The L1 searchers, as L2 block proposers, then sort these L2 transaction batches into L2 blocks, which are subsequently included in their L1 transaction packages and integrated into an L1 block. Finally, L1 validators verify and submit these blocks to the Ethereum main chain to ensure transaction finality.

To improve efficiency, Taiko introduces a preconfirmation mechanism. Proposers can publish preconfirmation information to other participants in the network before officially submitting a block. This notifies them which transactions will be included in the upcoming block. Additionally, proposers can regularly publish small preconfirmation batches in sequence to block builders, who can then choose to release these preconfirmed transactions in smaller batches to L1, rather than a single large data block, thus reducing data publication costs.

Moreover, in a decentralized proposer system, multiple proposers may attempt to submit blocks containing the same transactions simultaneously. In such cases, only one block will be accepted by L1, and the others will be reverted, and the unsuccessful proposers will lose their block fees. To mitigate this potential conflict, Taiko introduces a leader election mechanism. At any given time, only one proposer is elected as the leader, granting them exclusive rights to finalize the block. This ensures that the block created by the leader is added to the blockchain, while blocks created by other proposers are discarded, thus preventing multiple conflicting blocks and avoiding fee losses for failed proposers.

Puffer UniFi: Reshaping the Decentralized Application Chain Ecosystem

Puffer Finance has also introduced a preconfirmation-based solution, Puffer UniFi. As a Based Rollup, the transaction sequencing on UniFi is outsourced to Ethereum L1, while the preconfirmation mechanism in Puffer UniFi is implemented through Puffer’s native restaking validators.

In terms of process, transactions submitted by users are first handled by Puffer validators, who are registered as “Native Restaking” nodes on Ethereum. The validators provide preconfirmation commitments to users within approximately 100 milliseconds, which quickly informs them that their transactions have been received and will be included in future blocks. To ensure that validators adhere to their preconfirmation commitments, Puffer has also deployed the UniFi AVS mechanism, which imposes additional penalty conditions on validators. After providing preconfirmation, Puffer validators package these transactions with others and submit the blocks to Ethereum L1. Finally, the Puffer Sequencer Contract, part of the Puffer UniFi smart contract, accepts batch transactions and ensures that the transaction state has been confirmed and cannot be reverted.

According to the latest roadmap, Puffer’s UniFi testnet is set to launch in September 2024, and the UniFi AVS mechanism is expected to go live in Q4 2024, alongside the UniFi mainnet.

The vision of Puffer UniFi is to address the current fragmentation of liquidity across chains. In the future, applications built on Puffer UniFi will be considered as independent application chains. These application chains rely on the L1 sequencing and preconfirmation mechanisms provided by UniFi, which reduce development costs and enable seamless interoperability with other L1-sequenced Rollups or application chains. For instance, consider user Alice, who wants to use tokens from Rollup A as collateral to borrow assets on Rollup B. Since both Rollup A and Rollup B rely on the same Ethereum L1 validators for transaction sequencing, the validator can handle Alice’s transactions on both Rollups within the same Ethereum block. This allows Alice to operate across different Rollups without the need for complex cross-chain bridges.

What Else is Worth Watching?

In addition to the two projects mentioned above, there are several other noteworthy developments related to preconfirmations. For example, Primev has launched the mev-commit platform, which can provide preconfirmation services for any Based Rollup. Primev has also proposed a solution combining mev-commit and inclusion lists (IL) for blob preconfirmations. Meanwhile, Espresso has @EspressoSystems/bft-and-proposer-promised-preconfirmations">introduced the Proposer-promised preconfirmation model, allowing each Rollup to customize the sequencing order and penalty conditions of proposer-promised preconfirmations. Spire has launched the Preconfirmation Registry, a system where proposers (such as large operators and independent stakers) can post collateral in ETH. Chainbound has also unveiled Bolt, a protocol that allows Ethereum proposers to make commitments about the contents of their blocks.

Additionally, to simplify the user experience and better coordinate preconfirmation requests, researchers are exploring the introduction of preconfirmation gateways to abstract the complexity of preconfirmations from users. Through gateways, proposers can delegate their preconfirmation rights, and the gateway handles more complex tasks, such as communicating with users and maintaining the uptime of full nodes. Notable preconfirmation gateway projects include Aestus, Titan, and Ultra Sound.

Facing the current liquidity fragmentation within the Rollup ecosystem, Based Rollups are widely seen as an effective potential solution. For instance, Adam Cochran, a partner at Cinneamhain Ventures, tweeted that Based Rollups could address Ethereum Layer 2’s economic challenges. Preconfirmation plays a crucial role in enhancing the user experience of Based Rollups. As several major protocols continue to make positive progress, we will continue to monitor the latest developments in this area.

Disclaimer：

1. This article is reprinted from [ChainFeeds Research], All copyrights belong to the original author [Linda Bell]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.