The Rollup -Taiko's Based Contestable Rollups with CEO Daniel Wang

In this episode of The Rollup, which took place on February 9, 2024, Robbie, Andy, and Dino discuss based contestable rollup design, ZK proofs, and prover incentives. Read our notes below to learn more.

Background

Robbie (Host) – Merkle Tree Maxi, Co-Founder of The Rollup
Andy (Host) – zkProof Maxi, Co-Founder of The Rollup
Daniel Wang (Guest) – Co-Founder and CEO of Taiko
Taiko – decentralized, Ethereum-equivalent (Type 1) ZK-EVM.

Introduction to Taiko and Rollups

Daniel says that Taiko has two major milestones: the mainnet launch with BCR (Base Contestable Rollup) design and a testnet based on BBR (Based Boosted Rollups). BCR is a base rollup that allows for scalability without sequencers. Ethereum validators become the block builders who decide the sequencing of transactions. Being a base rollup means there is no extraction of value from block building, allowing existing infrastructure to evaluate opportunities.
He adds that base rollups have permissionless block proposing or sequencing, allowing anyone to become a builder without needing permission. The vision of Taiko is to be decentralized and permissionless from day one to enable scaling Ethereum effectively.

Decentralizing Prover Networks and Layer-1 Transactions

Daniel says that decentralizing prover networks like Scroll can work with Taiko’s base rollup. Booster rollups involve cross-layer transactions between layer-1 and layer-2. These transactions may have storage rights to the layer-1 state. Performing a layer-1 transaction on a boosted rollup layer-2 can be as expensive as a pure layer-1 transaction, but typically there are fewer writes to the base layer. The focus is often on reading from the base layer, such as oracles and AMM pools.
He adds that decentralizing prover networks is beneficial, but base rollups primarily focus on decentralized block proposal and sequencing, not just proving. Different blocks proposed by various builders can vary in difficulty to prove, requiring solutions for hard-to-prove blocks. While discussing booster rollups was not intended, it’s mentioned that a decentralized approver network like Succinct could potentially prove on Taiko.
Daniel says that random numbers can be obtained from layer-1 for data verification purposes. Verifying the data with other later transactions that use the same data. Writing back to layer-1 is expensive, discouraging people from transacting on layer-1.
Daniel says that Taiko inherits the decentralized proposer-builder separation concept from Ethereum. Malicious participants are not allowed to join Taiko and extract values.

Benefits of Contestable Rollup

Daniel says that ZK proofs are theoretically trustworthy but require extensive testing due to complex implementation. Handling malicious actors who create blocks with difficult or impossible-to-prove transactions. Introducing a guardian actor to handle situations where ZK proofs cannot be provided or are invalid.
He adds that app developers may find Taiko interesting due to its simple design for ZK rollup. Reducing costs by not requiring proofs for every block. Allowing different proof systems to be used based on application requirements.
Daniel says that comparing Taiko’s contestable rollup model with hybrid rollups that offer optimistic fraud proofs followed by ZK proofs. Not all transactions need to be proven in Taiko’s model, providing flexibility compared to other models.

Understanding the Reasoning behind Contestable Rollup

Daniel says that contestable rollup is suitable for applications where there is not much financial value at stake or when fast execution is required. The main reason for adopting contestable rollup is cost, as it allows for massive adoption by reducing expenses. Different apps have different security needs, and ZK rollup may not be the best solution for all of them. For example, social media apps may only require Valum and not a full rollup.
He adds that the Taiko technology protocol offers a wider range of options for different layer-to-layer series. Initially, when launching layer two, validity proof will be ensured for each block. Eventually, it will transition into a pure ZK rollup with all blocks ZK proven.

Choosing the Right Solution for Different Apps

Daniel says that different apps have different security needs, and ZK rollup may not be suitable for all of them. Social media apps, for example, may only require Valum and do not necessarily need a full rollup. The Taiko technology protocol provides a wider range of options to cater to various layer-to-layer series.

Fraud Proof in Contest Rollup Design

Daniel says that in contestable rollup design, there is no fraud-proof mechanism. Only validity proof exists. Validity proof ensures that blocks have correct block hash or state roots, proving that they are correct. The focus is on proving something is right rather than identifying errors or proving something is wrong.
He adds that guardians are chosen at the smart contract level, primarily through a decentralized autonomous organization. The DAO, which will be established later, will have a security council responsible for making decisions regarding fixing issues, forking, or upgrading smart contracts based on community consensus. The security council consists of various actors from both Taiko and the community, such as token holders and delegates.
Daniel says that the security council decides who will become guardians. Each guardian runs their own full node independently to correct any errors on-chain. If the aggregated approval from guardians surpasses a certain threshold (e.g., 2/3), the previous proof is overridden by the guardian’s correction to ensure network integrity.
He adds that guardian improving should be rare due to the presence of Taiko token-based bond designs. Validity bonds are burned if a proof is proven incorrect, while contestation bonds are burned if someone contests a proof and is proven wrong. These mechanisms discourage spamming the network unless individuals want to burn money.
Daniel says that once ZK Roll Up becomes more mature and bug-free after rigorous testing with ZK VM, the need for guardians will diminish. He suggests that after a year or two, and possibly a few more years, guardians can be removed as they are considered training wheels. ZK can also be considered trustworthy in the future.

Choosing Guardians and Their Independence

Daniel says that the selection of guardians is initially done at the smart contract level through a DAO. The security council within the DAO decides who becomes a guardian based on community consensus. Each guardian runs their own full node independently without needing to reach a consensus with other guardians. Guardians correct errors on-chain individually, ensuring network integrity.
He adds that the security council, consisting of actors from both Taiko and the community, plays a crucial role in selecting guardians. Token holders, delegates, and other community members participate in decision-making regarding guardian selection.
Daniel says that guardians collectively maintain network integrity by correcting errors on-chain independently. They do not need to reach a consensus among themselves but rather focus on correcting any wrong information on-chain. If aggregated approval from guardians surpasses a certain threshold (e.g., 2/3), previous proofs are overridden to ensure network correctness.

Cheaper and More Realistic Design

Daniel says that the design is cheaper because most proofs will not be ZK initially. ZK proofs are only required when needed, reducing costs. Other rollup solutions may challenge the claim of being fully ZK, but this design focuses on being cost-effective.
He adds that after discussions with app developers, it was realized that not everything needs to be fully ZK. The focus shifted to providing a cheaper alternative while still maintaining trustworthiness. Developers are encouraged to deploy their apps on this platform and gradually transition towards full ZK rollup if desired.
Daniel says that with a contestable rollup, configuration changes can be made along the way without needing a hard fork or genesis reset. A single layer-1 transaction can change the chain from 100% optimistic to 100% ZK. The initial state doesn’t matter as long as ZK becomes more affordable and trustworthy over time.

Complexity vs. Benefits

Daniel says that contestable rollup is more complex in terms of smart contract design compared to straightforward designs. The code may be more intricate, but the benefits outweigh the complexity.
He adds that the contestable rollup design has been running successfully with minor issues that are being addressed. The ability to challenge proofs and change configurations adds security and flexibility to the system.

Security Considerations

Daniel says that the contestable rollup introduces a cool-down window for validity proofs, adding an extra layer of security. Valid proofs can be challenged within a specified time frame, allowing for higher-tier proofs to override them if necessary.
He adds that trust in guard improvers plays a crucial role in determining overall chain security. If guard improvers are trusted, they enhance security. However, if they act maliciously, it compromises trust in the entire chain. While there are potential risks associated with guard improvers, similar centralized systems also rely on trusted administrators.

Maturity and Future Development

Daniel says that ZK is both the future and currently immature in its current state. It is acknowledged that ZK cryptographers may have reservations about its current implementation.
He adds that similar to protocols setting caps on TVL, adopting middle grounds allows for secure growth and testing while prioritizing user safety. If users desire a different design without guard improvers, they can fork the chain and implement their preferred system. The goal is to make ZK mature enough to remove guard improvers in the future.
Daniel says that ensuring user asset security is a top priority for developers and token holders. If issues arise due to ZK proofs, it should not be blamed on the design but rather seen as an opportunity for improvement. The aim is to secure user assets using all available means.

The Challenges of Type 1 ZK EVM

Daniel says that implementing a Type 1 ZK EVM is easier to debug and fix for simple applications. However, for a programmable Type 1 ZK EVM, it becomes challenging to trust the proofs. The initial plan is to use Optimistic and SGX PSE zkEVM and then transition to risk-zero-based ZK proofs. The goal is to eliminate the complicated codebase of PSE ZKEVM and rely on the existing PSE codebase for performance improvements.
He adds that ultimately, the aim is to have a zkEVM with no custom circuits, using languages like Rust for easy writing and debugging.

Token Utility and Governance in Mainnet

Daniel says that initially, $TKO tokens will be used to facilitate network operations by Taiko’s Engineers. Tokens will be distributed to those who have helped on testnet. After stability is achieved, an airdrop will distribute tokens to participants of the testnet, Ethereum contributors, etc.
He adds that $TKO will become the utility token for proposers and approvers who can earn layer-2 transaction fees. Eventually, Taiko aims to transition into a decentralized organization governed by the community through proposals.

Selection Process for Guardians

Daniel says that initially, guardians are primarily Taiko’s Engineers who maintain their own machines and run proposer/improver software. Technical expertise is required as they need to respond to notifications and investigate issues related to proof or block overrides.
He adds that after the DAO is available, there are two options for selecting guardians: direct selection or community voting. Concerns exist regarding individuals offering services as guardians without sufficient knowledge and time commitment.

Prover Selection and Compensation

Daniel says that they have not yet decided on how provers will be selected. They propose offering compensation for hardware to incentivize individuals to become guard improvers. Transaction fees are also mentioned as a potential form of compensation.

Check out these important links

Show Information

Medium: Youtube
Show: The Rollup
Show Title: Taiko’s Based Contestable Rollups 101 with CEO Daniel Wang
Show Date: February 9, 2024