Decentralized Exploration of Rollup Sorters: Espresso, SUAVE, and Polygon

100y
2023-04-12 10:39:25
Collection
Implementing decentralized sorting for Rollup is crucial for addressing censorship issues and improving MEV extraction efficiency.

Author: 100y

Compiled by: 0x11, Foresight News

  • The adoption of Ethereum Rollup networks surged in 2022, but there are still areas for improvement.
  • The sequencer is responsible for ordering user-submitted transactions. Currently, almost all Rollups rely on a single sequencer. Given that Rollups benefit from Ethereum's security through fraud or validity proofs, centralized sequencing does not pose significant issues. Nevertheless, decentralized sequencing is ultimately a better solution, as centralized sequencers may pose risks in terms of MEV extraction or censorship.
  • This article will introduce various decentralized sequencer projects, including: Espresso Sequencer, SUAVE, Polygon PoE, EigenLayer, and Cosmos ICS.

Introduction

The trend of TVL dominance among the 7 major networks outside of Ethereum, Source: Messari

2022 was a remarkable year for Rollups, with some Rollup networks achieving significant success amid the overall downturn in the cryptocurrency market. In 2020 and 2021, L1s shone brightly, and many were talking about the substantial gains from investing in SOLUNAVAX (Solana, Luna, Avalanche) and FOAN (Fantom, Harmony One, ATOM, NEAR). On the other hand, in 2022, due to macroeconomic impacts, frequent hacks, and the collapse of Terra, the adoption rate of L1s rapidly declined, which in turn drew public attention to Rollup networks.

Ethereum Rollups store state (e.g., wallet balances) and transaction data summaries (state roots) on the Ethereum network, relying on Ethereum's security when executing transactions off-chain, thus achieving scalability. So, can we fully trust and use Rollup networks like Optimism or Arbitrum? Unfortunately, this technology is still in its early stages, and current Rollup networks still have some risk factors.

Risk analysis of Rollup networks, Source: l2beat

The table above is a risk analysis from the l2beat website, showing that there are quite a few risk factors currently present in Rollups. Although the risks mentioned in the table are not the focus of this article, it is still worth taking a moment to briefly review their implications:

  • State Validation: State validation refers to how to prove the validity of L2 execution. For networks using zero-knowledge proofs, zk-SNARKs are labeled as ZK proofs (SN), while zk-STARKs are labeled as ZK proofs (ST). Many Optimistic Rollups seem to have risks in state validation: some projects like Optimism, Metis Andromeda, and Boba Network are developing fraud proof mechanisms, while others like Arbitrum One or Arbitrum Nova do have fraud proof schemes, but they only allow whitelisted participants to submit fraud proofs.
  • Data Availability: This concerns where the transaction data of the L2 network is stored. It is crucial that transaction data remains intact so that users can prove validity or withdraw tokens from the L2 network. A network can only be considered a Rollup if transaction data is stored on Ethereum; otherwise, it cannot be classified as such.
  • Upgradability: Whether the Rollup smart contracts on the Ethereum L1 network are upgradable. If marked as "yes" in the table, centralized participants can upgrade them at any time; if marked as "N days delay," such networks will enforce a delay in the upgrade time.
  • Sequencer Failure: This relates to the sequencer responsible for ordering transactions in the L2 network going offline or censoring transactions. "Using L1 for transactions" and "forced exit to L1" are two secure methods; the former allows users to submit transactions directly through L1, while the latter allows users to force exit to L1.
  • Validator Failure: This relates to the block producers of the L2 network going offline or deciding to censor. "Propose Blocks" and "Escape hatch (MP)" are secure; the former allows anyone to become a validator after a period, while the latter allows users to safely withdraw collateral by submitting state roots.

So far, we have reviewed various risk factors in Ethereum L2 networks. However, there is one risk that remains unaddressed: centralized sequencing. The sequencer is responsible for receiving and ordering transactions submitted by L2 users, and currently, centralized sequencing is the norm for almost all L2 networks. Even the two largest Rollup networks, Optimism and Arbitrum, operate a single sequencer to determine the order of all transactions. Does centralized sequencing pose significant risks? Compared to the risk factors listed above, the answer is no.

In fact, the risks listed in the table can directly harm users' funds, but centralized sequencing does not. Whether the sequencing is concentrated in a single entity or distributed across multiple entities, the validity of L2 network transactions is proven on the L1 network through fraud proofs or validity proofs. Nevertheless, if a single sequencer behaves improperly, users of the L2 network may suffer indirect losses. The ability to unilaterally decide the order of transactions may foster malicious MEV extraction, such as sandwich attacks causing financial losses to users, or even censor transactions (this varies by project, but due to the mechanism that forces transactions to be included in L2 blocks through L1, it is difficult for malicious sequencers to censor and completely block the addition of transactions. However, if a malicious sequencer censors transactions in urgent situations that require timely execution, such as liquidations in lending protocols, it may lead to delays in L2 transactions).

Although most L2 projects ultimately plan to decentralize sequencers, this requires the introduction of a new transaction ordering consensus mechanism. Incorporating consensus mechanisms for multiple sequencers will inevitably compromise scalability, which contradicts the original intention of Rollups. However, decentralized sequencing based on the spirit of blockchain does have advantages and is the ultimate direction to pursue. This article explores recent attempts and efforts made for decentralized sequencers.

Attempts at Decentralized Sequencers

Espresso Systems' Espresso Sequencer

Espresso Systems is a provider of blockchain scalability and privacy solutions, raising $32 million in funding from Electric Capital and Sequoia Capital in May 2022. Espresso Systems is also developing other solutions: HyperPlonk, a zkEVM proof system; jellyfish, an open-source cryptographic library; CAPE, a smart contract application that provides selective privacy for assets in existing EVM chains; and VERI-ZEXE, a system that provides privacy for smart contracts. This article will focus on the decentralized sequencer Espresso Sequencer developed by Espresso Systems.

Espresso Sequencer is a solution that helps improve the current use of centralized sequencers in zk and Optimistic Rollups. As a layer for decentralized sequencing and data availability for Rollups, it can serve as a middleware layer between L1 and L2. Transactions are processed by L2, but their finality ultimately depends on Ethereum's L1, as the network's consensus algorithm prioritizes liveness over security. Espresso Sequencer addresses the issue of Rollup sequencing and finality needing to rely on L1 by introducing the HotShot consensus algorithm based on HotStuff BFT. Based on this algorithm, multiple sequencers can quickly process L2 transactions under optimistic conditions.

Recently, HotShot released a testnet called Americano, achieving a throughput of 10,000 to 20,000 ERC-20 transactions in a network of 100 nodes configured with 4GB of memory and 2 CPU cores. The high performance is attributed to the HotShot algorithm leveraging the centralized network structure under optimistic conditions. When issues arise in the network, the algorithm switches to another mechanism, prioritizing stability over performance under pessimistic conditions.

Espresso Sequencer roadmap | Source: Espresso Systems

Currently, the Americano testnet only adopts an optimistic response transaction processing method, with plans to introduce a pessimistic processing method in the future Doppio testnet.

Flashbots' SUAVE

Source: Flashbots

Flashbots is a research and development organization dedicated to the efficient extraction and decentralized distribution of MEV on the Ethereum network. It previously provided Flashbot Auction for PoW Ethereum and is now contributing MEV-Boost, a middleware that implements MEV decentralization for PoS Ethereum. However, MEV-Boost also has its limitations: the centralization of block builders.

While the centralization of validators or proposers poses a serious threat to network security, the centralization of block builders does not carry the same level of risk. Nevertheless, centralized builders are not an ideal situation, as EOF (external order flow, where users send transactions to specific block builders rather than the mempool, greatly enhancing the power of builders) may lead the network away from the goal of efficiently extracting and distributing MEV. Flashbots launched a solution called SUAVE (The Single Unifying Auction for Value Expression) on November 23, 2022.

Source: Flashbots

SUAVE is an EVM-compatible blockchain network, but its purpose is to handle transaction ordering and block building rather than support the deployment of various dApps. In simple terms, SUAVE serves as both a block builder network for transaction ordering and a mempool for storing pre-confirmed transactions from the L1 network. Additionally, SUAVE is compatible with various EVM-compatible networks, meaning it can be used not only for Ethereum but also for L2 networks like Optimism, Polygon, and Arbitrum.

Source: Flashbots

The initial purpose of SUAVE is to 1) use encrypted mempool to encrypt users' transactions to ensure privacy and mitigate EOF situations through optimal execution markets, thereby preventing the centralization of block builders, and 2) decentralize cross-chain MEV extraction by processing the ordering of multiple networks, thus preventing certain block builders from gaining excessive power. However, using SUAVE also brings additional benefits.

By serving as an ordering layer for multiple networks, SUAVE can decentralize the ordering for L2 networks that choose to use it. Furthermore, this will enhance the efficiency of the MEV extraction process, leading to a more sustainable network, which is an additional benefit. However, delegating the role of centralized entities to a decentralized layer like SUAVE has a potential downside, which could lead to performance degradation in aspects like latency. Since SUAVE has not yet been released, its actual performance remains to be seen.

Polygon's PoE

Polygon zkEVM, renamed from Polygon Hermez, plans to achieve the decentralization of sequencers in Rollup networks by introducing a consensus algorithm called Proof of Efficiency (PoE). The way PoE works is described as follows, splitting the existing sequencer role into sequencers and aggregators. Sequencers aggregate users' transactions into batches and submit them to smart contracts on Ethereum L1, while aggregators verify the batches of transactions published by the sequencers. For a detailed explanation of PoE and its predecessor, Proof of Donation (PoD), please refer to Section 5 of the article "Polygon, the Front Runner of Modular Blockchain Future."

How PoE works, Source: Polygon

EigenLayer

EigenLayer is the first protocol to introduce the concept of re-staking, where stakers attach additional slashing conditions to already staked ETH (see EigenLayer: Expanding Trust Networks through ETH Re-staking). The benefit of this approach is that other protocols can leverage the strong security established by Ethereum with its massive market cap. Additionally, by adding another slashing condition, EigenLayer can better achieve decentralization goals. EigenLayer can be used to decentralize the current Rollup sequencers, and institutions wanting to become decentralized sequencers can re-stake ETH through EigenLayer as an entity and can be slashed for malicious behavior. While it is possible to build decentralized sequencers using the native tokens of Rollup networks through PoS methods, EigenLayer offers better protection by using the relatively more secure ETH tokens. The introduction of EigenLayer is essentially adding ETH tokens and slashing conditions to the PoS process of Rollup sequencers. Therefore, before adopting EigenLayer, consideration must be given to how to decentralize sequencers. The protocol has not yet launched and is planned for release in 2023. The aforementioned proposals for decentralizing Rollup sequencers are still under discussion, and it remains to be seen how EigenLayer will perform in actual use after its release.

Cosmos Interchain Security

Source: Celestia

Nick White, COO of Celestia, one of the representative projects of modular blockchains, recently shared an interesting proposal regarding the decentralization of Rollup sequencers on the Celestia forum. He suggested utilizing Interchain Security from the Cosmos ecosystem. Interchain Security (ICS) uses the validators of the Cosmos Hub to provide shared security for consumer chains (specific application chains using ICS). In other words, the idea is to leverage the existing validators of Cosmos as Rollup sequencers through ICS, as launching new tokens for Rollups and building decentralized sequencers is challenging. This is fundamentally similar to the idea of using EigenLayer for decentralized sequencing.

Introducing this method would provide the Celestia network with a powerful tool to achieve "atomic cross-Rollup composability." The Celestia network acts as a data availability layer for multiple Rollup networks, which could lead to liquidity fragmentation and decreased composability. If multiple Rollup networks operate using the same set of sequencers established through ICS, this would allow for atomic transactions, as the same participants would generate batches of Rollups. dApps on different Rollups would be able to interact seamlessly, addressing the issues of liquidity dispersion and composability degradation on the Celestia network.

Suppose you intend to borrow Y tokens from Rollup B by providing X tokens from Rollup A as collateral. In that case, it would take a considerable amount of time to verify whether the X tokens from Rollup A have been appropriately locked as collateral and whether the Y tokens from Rollup B have been correctly borrowed. However, if Rollup A and Rollup B share the same decentralized sequencer set, their blocks would be generated by the same entities, significantly simplifying the verification process for the loan. This also means that the networks can interact with each other almost as if they were on the same network, greatly enhancing the user experience for cross-Rollup interactions.

Like EigenLayer, Celestia and Cosmos also plan to launch ICS in 2023. Therefore, the proposals discussed so far should be viewed as potential solutions for decentralized sequencing.

Conclusion

This article examined different approaches to the decentralization of Rollups: Polygon's PoE consensus algorithm directly implements decentralized sequencing at the Rollup network level, Espresso Systems' Espresso Sequencer and Flashbots' SUAVE address the issue through another layer, and EigenLayer and Cosmos' ICS leverage a set of network validators that have already built strong security. Given the critical importance of decentralizing sequencers, Alex Beckett from Celestia published an article on this topic in June 2022. I would like to conclude this article with a brief overview of Alex's solutions.

The first is "permissionless PoS with leader election." In this approach, anyone can participate in the network as a Rollup sequencer by staking tokens and determining the order of block production through a leader election process. As mentioned above, Espresso Sequencer, EigenLayer, and Cosmos ICS all adopt this approach.

The second is "permissionless PoS through MEV auctions." Instead of using a leader election process to determine which sequencer will create the next block, the sequencer that bids the highest during the auction process is responsible for generating the block. Flashbots' SUAVE or the previously discussed Polygon zkEVM's PoE consensus algorithm employs a similar approach.

Finally, "permissioned sequencer sets for fair ordering." While the two solutions discussed above are permissionless, allowing anyone to participate in the ordering process as long as certain conditions are met, solutions involving permissioned sequencers refer to creating a set of sequencers in a private manner with trusted entities. You can think of it as sequencers collecting transactions and fairly determining the order based on a first-come, first-served basis. However, this approach requires trust and cannot be considered a true decentralized solution for Rollup sequencing.

Although Rollup networks gained significant attention in 2022, there are still substantial risks associated with them, and decentralized sequencers are often considered a low-priority risk. Nevertheless, achieving decentralized sequencing is crucial for addressing censorship issues and improving the efficiency of MEV extraction. With the widespread adoption of Rollup networks in 2022, they are handling increasingly large amounts of funds, which raises the demand for enhanced network security and fairness. It is hoped that the decentralization of sequencers will improve in the future, allowing users to utilize safer and fairer Rollup networks.

References

https://www.espressosys.com/blog/decentralizing-Rollups-announcing-the-espresso-sequencer

https://www.espressosys.com/blog/espresso-hotshot-consensus-designed-for-Rollups

https://www.espressosys.com/blog/releasing-espresso-testnet-1-americano

https://writings.flashbots.net/the-future-of-mev-is-suave/

https://www.alexbeckett.xyz/decentralized-sequencers-where-do-we-go-next/

https://forum.celestia.org/t/decentralized-Rollup-sequencing-as-a-service-via-interchain-security/423

ChainCatcher reminds readers to view blockchain rationally, enhance risk awareness, and be cautious of various virtual token issuances and speculations. All content on this site is solely market information or related party opinions, and does not constitute any form of investment advice. If you find sensitive information in the content, please click "Report", and we will handle it promptly.
banner
ChainCatcher Building the Web3 world with innovators