❓FAQ
Taiyi
What issues are Taiyi solving?
Answer:
Speed: Reduces effective block time from 12 seconds to milliseconds, dramatically reducing confirmation latency
Cost Predictability: Provides deterministic gas costs, eliminating unpredictable spikes and improving transaction cost forecasting
What's the relationship between Taiyi and based sequencing?
Answer:
Taiyi as a L1 preconfirmation protocol paves the way for Based Sequencing/Preconfirmation, which will be an extension of the Taiyi's initial version in the near future. We designed Taiyi to be totally forward compatible with L2 based sequencing/preconfirmation in mind. One could imagineAn Analytical Framework for Proposer Delegation Mechanism for Based Rollups that the opt-in proposers would automatically be the based preconfirmer for Based Rollups.
How are preconfers selected in Taiyi?
Answer:
Preconfers will be selected with slot auction conducted before the start of the target block. In the future, we will have a customized allocation function to factor in metrics such as historical performance, censorship-resistance, etc.. See our analytical framework for designing the preconfer selection mechanism An Analytical Framework for Proposer Delegation Mechanism for Based Rollups.
How to make sure preconfirmations are honored in Taiyi?
Answer:
Through economic guarantees. Economic guarantees for preconfirmations are created through restaking and a fraud proof based mechanism. Preconfirmers and proposers (re)stake ETH, which can be slashed if they fail to fulfill their commitments.
Who are actually selling preconf?
Answer:
Preconfirmations are sold by network participants called preconfirmers. These could be block builders, proposers, specialized entities, or potentially any party with the ability to influence block construction.
What's the difference between Taiyi and its peers?
Answer:
Taiyi focuses on gas hedging and preconf assignability. Preconfirmation is essentially blockspace futures, where buyers long base fee volatiilty while seller(preconfer) short base fee volatility. Applications such like Telegram bots, aggregators, or self sequencing application are blockspace "dropshippers" that have constant demand for blockspace not only for inclusion/exeuction garantees, but also deterministic pricing so that end-users could be quotes prices that are demernistic pre-execution. Similar dynamic exists for Rollups. See Ed Felton's talk L2 Resource Pricing.
Additionally, Taiyi separates preconfirmation from transaction construction, where blockspace buyers do not need to specify transaction data when preconfirmation gets purchased. This not only addresses the issues created by state dependent transactions but also makes preconfermation freely transferable. In other words, there exists a secondary market for preconfirmations, much like secondary markets for futures.
What's the lookahead period for Taiyi's preconf?
Answer:
2 epochs, which translates to 64 slots and roughly 12.8 minutes.
Our protocol has a 2-epoch(64 slots, 12.8 minutes) lookahead for preconfirmation. This aligns with Ethereum's beacon chain design, which provides:
Minimum 1 epoch lookahead for validator committee assignments
Proposer assignments visible for the current and next epoch
At any given time, this mechanism allows visibility of:
Minimum 32 future proposers (1 full epoch)
Maximum 64 future proposers (current epoch remainder + full next epoch)
How to become a preconfirmer?
Answer:
To become a preconfirmer, one needs to stake tokens in the Luban network and run the necessary software to participate in the preconfirmation process.
How do preconfirmers price preconf?
Answer:
It's up to how preconfirmers implement their own algorithm. For us, see Preconfirmation Pricing & Backtest
What's the risk involved for preconfer?
Answer:
The main risk for preconfirmers is underpricing preconfirmations, which could result in losses if gas prices spike unexpectedly. There's also a risk of being slashed if they fail to fulfill their commitments.
How many opt-in proposers does Taiyi needs?
Answer:
Taiyi could become operational at least 1 validator has opted in. However, it's better to have at least 1 opt-in proposer in every lookahead(). To calculate how many proposers we need to opt into the protocol, this is the formula
P = 1 - (1 - \frac{v}{\text{total_validators}}\right)^{64}, where
represents the probability of having at least one opt-in validator within a span of two epochs.
represents number of validators that have opted in to participate in the preconfirmation system.
represents total number of validators in the Ethereum network.
Assuming equals 5%, the probability of having at least one validator opt-in within two epochs is approximately 96.25%
How to buy preconf on Taiyi?
Answer:
Users can buy preconfirmations through the preconfirmers. The process involves specifying the desired block for inclusion.
Which parts of Taiyi v0 are handled by relays and which parts are not? For example, preconf request pricing, preconfer selection, fair exchange guarantee, etc.
Answer:
Commit-Boost primarily focuses on proposer signature. Other critical aspects, such as pre-confirmation request pricing, pre-confirmer selection, and fair exchange guarantees, are not within the scope of Commit-Boost.
From a technical standpoint, our solution does not rely on Commit-Boost. Instead, we view it as a bootstrapping "accelerator" that can facilitate initial adoption, rather than a core dependency. Strategically, we are committed to maintaining independence and avoiding over-reliance on entities beyond our control. This approach ensures the robustness and long-term viability of our pre-confirmation solution.
What's the pricing model work for contentious state, i.e., a top-of-block swap in 20 blocks? Will a proposer earn less from MEV if he sold the blockspace via Taiyi vs. MEV-boost?
Answer:
The pricing model outlined in our documentation serves as the foundation of our comprehensive pricing strategy, not its entirety. For simple transactions like ETH transfers, which constitute a significant portion of network activity and priority fees paid (as shown in our dashboard), base fee prediction is a reliable pricing mechanism for preconfirmations.
For complex transactions involving MEV, we will develop a more robust market system to encourage price discovery, with the goal of approaching the theoretical optimum of a block. Here's the roadmap:
During early implementations, proposers and the relay will only allow preconfirmation on non-contentious state. They can configure constraints to control whether preconfirmation requests can touch specific smart contracts (e.g., Uniswap). Preconfirmers can also differentiate preconfirmation requests based on the sender, similar to how market makers differentiate between toxic and organic flow.
The preconfirmation market will mature rapidly, starting with an RFQ-type system and progressing towards an auction mechanism, which is currently being researched. With a competitive market, multiple searchers will vie to secure the top-of-block swap in 20 blocks, ensuring proposers receive competitive bids with a risk premium. Ultimately, an open market for top-of-block swaps will generate more value compared to the current vertical integration between searchers and builders, funneling more value to proposers than the status quo.
How does Taiyi work with "based rollups"?
Answer:
Taiyi V0's architecture inherently supports "based preconfirmation." A proposer can configure the preconfirmation constraints so that the preconfirmer gets the exclusive right to update the rollup contract in the future block, effectively becoming the "based sequencer" for the rollup, providing preconfirmation to L2 transitions.
Taiyi v1 will provide a more out-of-the-box solution for based rollups when the demand and requirements become clearer.
Taiyi is indifferent to blockspace, transaction or blob. Blobs are a separate part of the block and won't interfere with preconfirmed L1 transactions. Proposers/preconfirmers already preconfirming L1 transactions would likely start preconfirming transactions for based rollups. The pricing of blob preconfirmation will be similar to transaction preconfirmation, as the 4844 base fee generally mirrors 1559 base fee dynamics.
To hedge base fee volatility, rollups could preconfirm their blob inclusions in advance. With more opt-in proposers (roughly 5% of all validators), we can confidently say every epoch will have an opt-in proposer. In this scenario, rollups could purchase long-term inclusion guarantees weeks or even months in advance to manage costs.
Preconf
What is preconfirmation? How is it related to blockspace?
Answer:
Preconfirmation is a mechanism that allows users to secure their transaction's inclusion/execution in a future Ethereum block. It's directly related to blockspace as it essentially reserves a portion of a future block for a specific transaction, guaranteeing its inclusion/execution.
How is blockspace traded today?
Answer:
Currently, blockspace is primarily allocated through a priority gas and MEV auction mechanism. Users compete by setting priority gas, with higher-priced transactions more likely to be included sooner, or alternatively searchers compete by extracting MEV to be able to bid higher price bundle for inclusion.
Who wants preconf and why?
Answer:
Those who wants a predictable supply of blockspace, and a less volatile gas fee. Namely:
Self-sequencing dapps, like Sorella Labs and Semantic Layer. These dapps sequence transactions by themselves, and only submit a monolithic bundle to Ethereum. The higher value that bundle carries, the higher certainty of inclusion the dapps need. Therefore, they are incentivized to buy preconf for their future bundles. See our writing on the topic From Self-sequencing to Preconfirmation.
Trading bots, like Banana Gun. Trading bots can pre-buy future blockspace in anticipation of surging demand driven by, i.e., memecoin launch, by purchasing preconf. Regular cost incurred by limit orders can be smoothed by preconf too.
Intent solvers, smart wallets, bridges, aggregators. They are big transaction originators, and often work cross-domain. They are incentivized to buy preconf for deterministic transaction outcome.
Any issues created if preconf protocol is not well-designed?
Answer:
Potential for new forms of front-running if the ordering mechanism is flawed. Taiyi uses verifiable sequencing rule(see Preconf Sequencing Rules) to prevent preconfers front-running users' transactions
Poorly designed preconf pricing algorithm would either overcharge the user that renders the preconf request overly expensive, or alternatively undercharge the user the expose the preconfer to under-hedged base fee volatility.
EVM state dependent transactions would render transaction construction in advance less useful as state dependent transactions would rely on knowing the latest EVM state. For example, preconfirming a swap transaction 30 blocks in advance is risky as swapper wouldn't know what the pool state will be 30 blocks later. Knowing the latest pool state is critical for state dependent transaction. Taiyi solves this by separating the preconfirmation from transaction construction. User only needs to submit a transaction just-in-time but preconfirm inclusion/execution/ordering in advance.
How about execution auction(EA) and execution ticket(ET), etc - what's the relationship between preconf and these designs?
Answer:
EA and ET are generally orthogonal to preconfirmation as they are proposer selection mechanisms as opposed to how blocks are constructed. However, our preconfirmer selection design was inspired from the ideas of EA and ET. See An Analytical Framework for Proposer Delegation Mechanism for Based Rollups
What's the relationship between preconfirmation, inclusion list, multiple concurrent proposer, and partial block building?
Answer:
Preconfirmation, inclusion list, multiple concurrent proposer, and partial block building are all in essence different manifestations of the same idea - block co-creation. Preconfirmation creats a nice entry point for weaving in the design goal of these ideas, namely censorship resistence, into one design. This is exactly the future iteration of Taiyi. See our design proposal Consensus-Based Preconfirmation with Multiple Concurrent Proposers. Also see Taiyi's Roadmap.
How would the role of the current mev-boost relays change in the context of preconfirmation? Will it be a hurdle to preconf adoption?
Answer:
The role of current MEV-Boost relays will undoubtedly evolve in the context of pre-confirmation. However, this change is already underway, with leading relays such as Ultrasound, Titan, and Aestus actively collaborating with emerging initiatives like Commit-Boost. As a result, I anticipate a smooth transition rather than significant hurdles to pre-confirmation adoption.
Notably, pre-confirmation opens up a clear path to monetization for relay operators, addressing a critical challenge that previously hindered relay sustainability. This shift in dynamics is likely to incentivize pre-confirmation teams, including our own, to establish their own relays. The alignment of incentives between pre-confirmation and relay operation is poised to drive innovation and adoption in this space.
How competitive can preconf tips be when compared to just-in-time block building? Since predicting future MEV sounds like a very hard problem?
Answer:
We don't expect preconfirmation to directly compete with just-in-time (JIT) block building. Instead, they are complementary. The most contentious state (top-of-block) is likely to be reserved for JIT auctions in the foreseeable future. Preconfirmation will target those willing to pay for the extra value created by the expressiveness brought by preconfirmation.
As shown in the image below, only preconfirmation of execution, at the right end of the spectrum, will directly compete with JIT auctions. Luban doesn't plan to push for this.
Along the spectrum, expressiveness will grow, allowing preconfirmation buyers to specify more conditions (transaction position, relative order in a smart contract, etc.), creating more value compared to pure JIT building. This translates to higher premiums and preconfirmation tips. With preconfirmer sophistication and market maturity, we believe more transactions will become "easy to price" in a sense.
Why do self-sequenced dapps need inclusion guarantees from preconfirmation? Isn't a small priority fee enough for builders to include their bundles?
Answer:
There are two main reasons:
Tail risk: While self-sequenced dapps can generally work with a small priority fee, there are extreme cases where it doesn't suffice, such as sudden on-chain congestion or builder propagation issues. Since the bundle would carry little value for builders, it would be more easily dropped in these scenarios (similar to an L2 blob).
Interoperability: As more apps adopt self-sequencing, interoperability between them could be broken. A sophisticated user would need to ensure ALL auctions can land on-chain before bidding, which is challenging.
Moreover, not only AMMs want to self-sequence. Telegram bots can self-sequence too if they buy preconfirmation. Currently, they form private deals with builders without inclusion guarantees, under a vague profit model (as they don't know how much user gas to pay builders and how much to retain as profit). With preconfirmation, they can be certain of the cost and profit.
This applies to all transaction originators, like wallets and solvers. Preconfirmation turns order flow wholesale into real-time per-transaction bids.
Is multi-block MEV a concern if a preconfirmer can secure slots in consecutive future blocks?
Answer:
Multi-block MEV (MMEV) is indeed a concern. The way to address it is to increase the entropy between slots. There are many ways to do this, such as making preconfirmer selection purely random. We have shared an analytical framework on preconfirmer selection. Ultimately, the exact proposer selection mechanism will be decided through governance processes controlled by token holders. The largest token holders will be proposers, who have the most skin in the long-term health of Ethereum. This is reflected in the fact that Lido, in theory, could perform MMEV every epoch but refuses to do so to maintain network health.
Does the pre-slot deterministic nature of preconfirmer auctions lead to additional centralization?
Answer:
It highly depends on the selection mechanism. The way we design the allocation function will greatly influence participant behavior. For example, when an allocation is concave, it doesn't increase the chances of winning linearly simply because one party has more access to capital. Additionally, we will introduce other metrics, such as bids back to the rollup in a based rollup scenario, into the allocation function to encourage "good" behavior for the participants. Token holders get to decide what behaviors are considered "good."
How can preconfirmers deal with latency-related spam?
Answer:
There's no perfect solution to spam, but here are the solutions we are exploring:
Rate Limiting: Implement strict rate limits on preconfirmation requests per user/address.
Stake-Based Prioritization: Require users to stake tokens (e.g., BURR) to submit preconfirmation requests. Prioritize requests based on stake size, discouraging spam from low-stake accounts.
Reputation Systems (currently used by Flashbots and AA bundlers): Develop a reputation scoring system for users based on their historical behavior. Allocate resources preferentially to users with higher reputation scores.
Dynamic Pricing: Implement surge pricing during high-load periods. Increase the cost of preconfirmation requests as system load increases, making spam attacks economically unfeasible.
Proposer
How to become an opt-in proposer?
Answer:
Restake on a restaking protocol
Run Commit-Boost
Done!
What are the risks associated with running Taiyi as an opt-in proposer?
Answer:
Proposer has zero pricing risk as any base fee volatility will be absorbed by the preconfirmer. In other words, the worst case scenario for an opt-in proposer is to earn 0 additional profits. However, proposer will be slashed if Commit-Boost was configured to accept blocks from non-Commit-boost integrated relays.
Note: as of 23/Jul/2024, Titan builder's Helix is the only fully integrated relay, though other relays such like Ultrasound and Aestus have expressed interests.
How does a proposer delegate blockspace to a preconfer?
Answer:
Preconfers are like wholesalers of blockspace. They bid for the right to preconf on behalf of a proposer. The proposer sell his future blockspace, subject to restrictions like total gas limit, smart contracts interacted with, etc., to the highest bidder.
How much of a block can be preconfed via Taiyi? Does it have to be the whole block?
Answer:
No. The proposer can sell arbitrary size of future blockspace via preconf in Taiyi. The size will be configured inside of Commit-Boost
How can/do you think it's likely that proposers will take advantage of their agency with, e.g., computational overhead of simulating blocks?
Answer:
If the clear advantage of having agency outweighs the cost, proposers will be incentivized to do so. We expect some proposers to have the required sophistication, while others don't. In such cases, we provide an open access marketplace for proposer commitment delegations.
The utility function becomes: if the profits stemming from having agency outweigh the cost of delegation, proposers will be incentivized to do so. The cost of delegation generally comes from having to pay for it. However, if we design the marketplace so that the cost of delegation is contingent upon profit (see Contingent Fee), i.e., proposers only pay when delegatees make a profit, there's a clear incentive for proposers to exercise agency through delegation.
How does preconfirmation enhance proposer agency, and how would proposers exercise their agency?
Answer:
The simple answer is that proposers will exercise agency when they can make more money by doing so. Here are some examples:
Pre-PBS block building exemplifies how proposers initially leveraged their agency for profit.
The introduction of PBS itself demonstrates a form of agency, but with delegation. There's nothing inherently wrong with delegation, but it has to be designed carefully to encourage efficient resource allocation, which is something PBS lacks but preconfirmation shines at.
In our opinion, blockspace should be priced differently to reduce consumer surplus and overall deadweight loss, due to varying degrees of state contention under different market conditions. Preconfirmation of state/execution is a first attempt to address this problem. There are more elegant designs that could address this issue by extending preconfirmation, which we will explore in the future.
We believe proposer agency, through delegation or not, paves the way for the multidimensional fee market on Ethereum that Vitalik talked about without protocol-level updates. As such, proposers will earn more profit by differential pricing, which also facilitates a more efficient use of blockspace.
Bussiness
Will there be a token? What's the core token design?
Answer:
There may or may not be a token:). But let's say there will be. Let's give a name $BURR since our logo is a Burr Puzzle. Burr is a token that governs time in Ethereum blockspace. Some amount of $BURR needs to be burnt every time a preconfirmed (partial) block happens, by the preconfirmer. $BURR can be staked for fresh $BURR emission. $BURR can be delegated, and staked by the delegate. The initial distribution of $BURR will be conducted via an incentivized proposer onboarding program.
How does value accrue to $BURR?
Answer:
$BURR is the preconfirmation currency. Preconfirmers need $BURR to do preconfirmation, which creates demand on $BURR.
$BURR is a delegation token. Anyone can express their preferences on preconfirmers by delegating $BURR to them, which can be further staked and generate new $BURR to cover preconfirmation cost, giving competitive advantage to preconfirmers that are favored by the people.
$BURR is a governance token. Decisions regarding slashing, preconfirmation market, preconf and based sequencing rules, and anything that falls into the realm of Luban's governance, will be made by $BURR holders.
Proposers will receive the lion's share of liquid $BURR in the early stage of the protocol, making sure the blockspace producers get to choose how and who their blockspace is sold.
Will there be a fee?
Answer:
In the future. We will focus on go to market and growth in the early stage. All potential fees collected by the protocol will go to protocol treasury that is governed by $BURR holders.
At the mean time, there will be protocol-operated preconfirmer and based sequencer, which will bring positive cashflow to the protocol.
Last updated