Important Ethereum Capacity Expansion Solution: Optimistic Rollup Status Report (Part 2)

Author: Daniel Goldman (free software engineer, technical consultant and writer)

Source: Medium

This article is the third part of "Ethereum's Important Expansion Plan: Optimistic Rollup Status Report".

Block Production

Most projects use an open, license-free block production model; that is, either party has the right to issue bonds and extend the ORU chain by proposing new blocks.

The only exception is IDEX 2.0, because IDEX has unique permissions for the production block. In this case, if the user can withdraw the funds without obtaining the operator's permission, this can maintain the untrustworthy property (that is, withdraw the funds by publishing a transaction directly on the parent chain). Without this option, users will be helpless when encountering a malicious operator, because the operator will review all withdrawal attempts by the user, in fact, locking the user's funds on the rollup chain indefinitely. Similarly, in a licensing system, it is not difficult for operators to gradually slow down the production of blocks or even stop production completely; in this case, users need to withdraw funds.

IDEX 2.0 does provide users with an exit option initiated by the main chain. The more complicated point is that if there is fraud, these exit options must be verified / possibly restored, so a guarantee must be attached to prevent malicious destruction. However, considering that these exit options are supported, IDEX's licensing model itself cannot guarantee that the system is in a regulated state (other license-free chains also plan to use the main chain to initiate exit options in special cases).

In addition, most operations include some insights into splitting block production into two steps: separate inputs for transaction inclusion and transaction sequencing. These models address issues related to reviewing resistance and reducing early trading opportunities.

In the Arbitrum model, any party can add transactions to the queue (they call them "inbox"); block producers are forced to fetch transaction items from the end of the queue and place them in the next block. Only invalid transactions are covered, but no state transitions are made. Regardless of whether an invalid transaction or a valid transaction fails, it can be judged as a punitive criminal act by anti-counterfeiting certification, thereby eliminating simple examination matters .

Others, such as Optimism * and WCL, plan to use the combustion proof auction model to sort transactions; after a set of transaction items are included, the transaction rights are owned by the party willing to burn the most Ether equity. The purpose of this is to reduce costs and open competition to minimize the profit margins of parasitic first-mover opportunities and ideally reduce overall risk. Please refer to "Miner Extractable Value Auctions" for details of how similar schemes work.

The Fuel model has a block production cycle. During this cycle, there is a time window where one party (that is, the company / team) can obtain permission to sort. In this model, the team has an advantage in gaining value and monetizing the application, and does not have to limit block production to what is allowed.

Finally, Celer's Rollup is unique because block production itself includes its own independent consensus mechanism. Prior to the development of rollup, Celer had planned to issue an Ethereum sidechain proof of stake ( State Guardian Network ) to serve as a watchtower for their national channel network. Participation in block production is not allowed and Celer's tokens need to be supported. Sidechains will use BFT consensus-such as Tendermint or similar. Celer plans to use this sidechain infrastructure to propose and release ORU blocks .

Trust hypothesis

Most projects strictly adhere to the trust requirements of the layer2 protocol described at the beginning of the article.

IDEX 2.0's reliance on data availability challenges makes this protocol a special case and is not well defined by ORU. When a block producer proposes a block, they initially only release the Merkle root and implicitly assume that they will share the relevant block data with the off-chain validator. For example: if the block data is never shared with the validator, the validator can issue a data availability challenge, forcing the block proposer to publish the data as calling data (as was done in the original typical ORU structure). If the result proves that the block is valid, the validator will be punished for invalid dispute challenges and cut some of their rights.

The problem at hand is that the differences outlined above are essentially arguing which of the two parties actually owns the data; according to the speaker / listener fault equivalence , it is not objective to resolve such a dispute. possible. Therefore, given concerns about reputation and the motivation to continue participating and earning fees in the exchange ecosystem, IDEX needs to be prevented from forcibly maliciously sabotaging and ultimately punishing validators. Regardless of whether it is sufficient to prevent malicious behavior in practice, these assumptions are a security guarantee for ORU or Ethereum itself, so this challenge and punishment for data availability can be used as an additional trust assumption.

The only other (indisputable) project that requires additional trust assumptions is ZK-ORU, which requires a trusted setup to create its snark line. Projects like ZCash and Aztec deal with this issue through an elaborate ceremony . As long as at least one of the n participants is acting honestly, it has cryptographic security; ANON plans to do something with ZK- ORU moves similarly.

Ultimately, some projects, including Fuel, Optimism, and Nutberry, have discussed providing users with trusted fast payment options; a user, if they choose, can consider including it in the block after completing the payment. If the words are not true, this fraud will not be proved to the parent chain, but it can be proved to any third party. The important point is that this option is not mandatory and there are obvious differences in the user interface; users who are completely distrusted by this can still choose to wait for block confirmation. Similarly, Celer plans to provide users with a series of security / trust parameters. Users can require all transactions to be announced immediately in the rollup block, or the side chain that can accept the parameterized trust window becomes smaller, And mission criticality will also decrease.

If and / or when payment commitments cannot be fulfilled, instant confirmation of trust can be made by allowing block producers to significantly reduce collateral; several teams are considering launching implementation and putting this feature in place.


All projects plan to release their work as free open source software.

ANON is planning to strictly implement the implementation as a public welfare achievement. Currently, there is no plan to directly obtain value. PinkieBell (the only anonymous developer supporting Nutberry) also has no plans to monetize it, but is currently considering developer fees and charging a small portion of the fees from the block producers to pay developers engaged in the project.

Five of these projects: Fuel Labs, Optimism, Interstate, WCL, and IDEX are planning to monetize their work in three ways: project implementation cases, monetizing transaction fees, and providing users with liquidity to withdraw funds quickly.

Although block production in Fuel, Optimism, and Interstate is license-free, if they (ie, the company) are the main block producers, everything is fine. They plan to make money by charging fees because they have an advantage as a priority aggregator (discussed previously).

At least one project has expressed uncertainty about the company's legal status as a payment processor and is currently awaiting further clarification on the issue.

Offchain Labs does not plan to enumerate Artibrum Rollup instances. Instead, it plans to use its own work to provide a project planning and support for the project, as well as some stronger economic and security guarantees to achieve profit.

IDEX's economic plan is the most detailed; monetization comes from a cryptoeconomic setup that includes charging exchange fees and the inherent value of native tokens. Both the block producer (IDEX) and the validator have been compensated, making IDEX the only project that attempts to directly motivate verification. IDEX and Celer are the only projects that plan to merge application tokens. Several other teams scoffed at the concept of tokens inherent in the Layer 2 system; others were more open to possible economic token models, but felt that more research was needed and that they were not specific as far as the project was concerned Actionable plan.

User experience / verification

Most projects plan to support rollup on-chain transactions through the metamask plugin. In addition to publishing their own front-end interfaces, the infrastructure tools of the metamask plugin are basically ready.

If not all users need to maintain uptime and verify all blocks, all projects will run; that is, they expect two types of users: ordinary end users and advanced users (ie, validators). The essence of all ORUs is that all users are safe as long as at least one honest party verifies and issues a security certificate at any given time. The verified economic incentive method is to directly issue rewards (only in the case of IDEX), and have the opportunity to produce blocks and collect fees, and have the opportunity to issue anti-counterfeiting certificates and get rewards. Of course, the 22nd principle of anti-counterfeiting proof is that if the system is designed properly, it is still hoped that fraud will not occur as much as possible (if it has actually happened).

Another direct incentive for verification is to trust the validator, which can quickly ensure that the transactions that everyone cares about have been completed; that is, you don't have to wait for the exit window to appear. The expected result is that the parties benefiting from it will involve exchanges, wallet providers, liquidity providers with quick exit options, users who can accept large significant transfers, high-end users / enthusiasts, and project creators. One team suggested that miners can take the verification rollup chain as their responsibility and ask some interesting questions about the interaction of incentives between the first and second layers of data used.

Open topics on research / discussion

BLS signature

Easily became a hot topic at the moment; with BLS signatures, block producers can obtain a signature from each rollup block and replace them with a single aggregate signature, making the data requirements on the chain closer to ZK- rollups. One problem that is more troublesome is that if verified, verifying the BLS signature consumes more gas than a typical ECDSA signature. Barry Whitehat said: He has identified a feasible structure to prepare for the ORU focus area, but has not yet reached a consensus on the best signature aggregation scheme.


It can be imagined that if rollup technology is used as a specification on Ethereum, what will the future look like when most of the chain data exists in the form of "roll-up"; that is, it is available in practical terms, but not directly for implementation; This change may have significant economic and technical impact on Ethereum client implementation. One solution is to prepare for such a future. In fact, it is to introduce a special field of transaction data, temporarily called "post-data", which is only used for availability and cannot touch the state module or EVM. With this, customers can process and store data more optimally, and ideally, the price of gas will eventually decrease. The EIP has a general overview of this plan, but it is currently in its infancy.

Bond size and challenge stage parameters

Ideal parameters for bonds required for block production and time windows

Since most teams have not yet determined the final number, it is still in the public discussion stage until the block is completed (so it is possible to withdraw to the main chain). If the challenge period is longer, it will better ensure that users have more time to detect fraud and better pressure the miners who are trying to review the security certificate. For some projects, the best is to be in the security period for a few days to a week, because they will move closer to security, and users will usually withdraw more quickly. The withdrawal method is either atomic conversion to the chain or tokenization. Not just use the exit option as a bond ( already discussed in Plasma's structure). According to Ed Felten's analysis , a shorter dispute window like 3 hours is sufficient.

In the layer 2 structure, if the compensation is invalid, the bond can be used as a necessary collateral to compensate the user, but the bond in the ORU is different. In addition to suppressing gas waste, it is only to suppress invalid block production (maybe It also serves as a reward for successfully demonstrating fraud.) The recommended bond demand ranges from 32 ETH to 1 ETH, which is comparable to the ETH 2.0 equity demand, but some people suggest that it should be equivalent to the economic value on the rollup chain (off-chain laboratory plan "1 Eth or on-chain total 2% of value, whichever is greater ")


Compared to Plasma and the channel structure, one of the shortcomings of rollups is that it cannot get the privacy benefits of data being completely dropped; the only item mentioned here that can provide privacy benefits is ZK-ORU. Other projects are discussing creating privacy-focused rollups in two ways: structurally implementing other well-known privacy technologies (ie, Tornado Cash-esque mixers or Aztrec-esque privacy layers) or creating hybrid rollup / federal data models, such that Users can rely on a large amount of available data, thereby reducing the number of publications on the chain.

Block production model

In addition to the complete EVM / application-specific segmentation, one of the main disputes between the teams seems to be which model of license-free block production is acceptable, and what results the different models will produce, that is:

  • Does the single-priority aggregator model only stimulate more destructive early transactions when early transactions decrease.
  • Whether the rich side will maliciously destroy and effectively review the pure proof of stake-block production.
  • How can the market value be fully reflected and displayed in the combustion certification auction.

Some research and modeling may provide some inspiration for these issues, but we may not be able to draw any conclusions until we see them being tested in the field.

Optimistic Game Semantics

The Optimism team's research on the theoretical nature of the layer 2 structure finally led to the concept of Optimistic Game Semantics , trying to explain the different layer 2 structure under the unified framework of fraud / controversy. Designed to create shared contract logic and infrastructure between different types of applications in layer 2 (including not only rollups, but channels and plasma, etc.). Optimism and the Cryptoeconomics Lab (currently focusing primarily on Plasma structures, but may develop rollup operations in the future) have both tried to develop tools within such a large unified framework.

in conclusion

The team using the Optimistic Rollup pattern came to a conclusion from different perspectives: exploring the theoretical design space of different structures of layer 2 and finding ways to effectively run the execution engine in EVM, trying to create a more scalable version of existing Ethereum applications Wait, converging to ORU to get enough scalability benefits with minimal practical complexity is the best way to move forward.

Only by seeing these issues entering the mainnet and seeing the ecosystem choosing to use these tools can we answer these open questions; that is, whether the project will apply the infrastructure to a broader rollup solution, or design Something more in line with practical use, plus some more specific benchmarks on gas costs and verification burdens to give exact numbers in transactions. But there are already several projects on testnet, and no secret research and usability challenges have been found. One thing is certain, we will soon see its practical application. Scalable autonomous smart contracts are emerging.

* Correction: A previous version of this article showed that the block production model planned by Optimism allows them to be prioritized as locators; their MEV auction agreement does not actually give anyone privileges. Thanks Ben Jones .

Thank you to all the teams that participated in the interview, and thank you for your helpful discussions.

As of this writing, all the contents of this report have been released.