Ethereum's Rollup is centralized: Exploring Decentralization of Sorters

1. Key Points

• Transaction ordering has become an increasingly serious issue in the Layer 2 space. The main role of L2 rollups is to provide a secure place for cheap transactions, offering users an execution layer, and then submitting transaction data to the upper layer L1.

• The sorter is an entity that organizes transactions into groups. It receives users' unordered transactions, processes them off-chain into groups, generates compressed ordered transaction batches, and then sends them to the parent L1.

• Rollups actually do not require sequencers; this is just a design choice to provide a better experience for users. They can also use the base layer for sequencing, but the efficiency of the Ethereum base layer is relatively low and expensive. Currently, each major L2 rollup project has found that running centralized sequencers is more convenient, cheaper, and user-friendly.

• Due to the sorter controlling the order of transactions, it has the authority to review user transactions ( although a complete review is unlikely ). The sorter can also extract the maximum extractable value ( MEV ), which may cause economic losses to users. Validity is also a concern; if the only centralized sorter goes down, the entire rollup will be affected.

• The solution is a shared, decentralized sequencer. The shared sequencer provides decentralized services for rollups. In addition to addressing issues of censorship, MEV extraction, and validity, it also introduces cross-rollup functionality, opening up new possibilities. Espresso, Astria, and Radius are developing innovative shared sequencer solutions, each with unique features in their respective architectures.

2. Introduction

As the Ethereum L2 rollup ecosystem becomes more popular, the often-overlooked aspect of sorters is becoming increasingly important. Sorters are responsible for transaction ordering and can provide a better user experience, lower fees, and faster confirmations. However, the problem is that currently, all major Ethereum L2 companies find it most convenient, user-friendly, and cost-effective to run their own centralized sorter. Considering the power that sorters have in transaction review, MEV extraction, and single points of failure, this may be seen as an undesirable outcome and not in line with the spirit of cryptocurrency.

Although most companies regard decentralized sorters as part of their roadmap, there is still no real consensus on how to achieve this. Notably, Arbitrum and Optimism have launched solutions since the second half of 2021, but there has been little progress in decentralized sorters.

This report will delve into the role of sorters and the current state of the Ethereum rollup space, discussing projects that are researching solutions, namely the decentralized shared sorting network. We will detail the uniqueness of these projects and their solutions, and reflect on what this means for the future development of the Ethereum L2 rollup field.

3. What is a sorter?

Blockchain is a distributed ledger composed of time-stamped transaction data sorted by blocks. Initially, this transaction data was unordered and unorganized. After sorting, it can be organized into blocks and executed, creating a new state of the blockchain. For L1 blockchains like Ethereum, transaction sorting occurs at the Ethereum base layer itself.

In the most popular scalability solution of Ethereum, the L2 rollup layer, transaction ordering has become an increasingly serious issue. L2 rollups provide users with an execution layer and then submit transaction data to the upper layer L1. A single batch of transactions submitted to L1 typically contains hundreds or thousands of compressed L2 transactions, thereby reducing the cost of sending data to L1.

In the L2 rollup world, the sequencer is the entity that sorts transactions into batches. It receives unordered transactions from users, processes them off-chain into batches, and generates compressed ordered transaction batches. These transactions can be placed into blocks and sent to the parent L1. Batching transactions can also be used on the data availability (DA) layer (, typically on Ethereum ). The sequencer also provides users with near-instant "soft confirmation" receipts. The "hard confirmation" is received after the transaction has been sent to the L1 layer.

Why do Rollups need to use sequencers, and why is it a problem?

The fundamental goal of the sorter is to improve user experience. Using a sorter for L2 transactions is similar to using a "fast lane", which means lower fees and faster transaction confirmations. The sorter can batch hundreds or thousands of L2 transactions into a single L1 transaction, thus saving gas fees. In addition, the soft confirmations provided by the sorter allow rollup transactions to provide users with quick block confirmations. This combination helps to enhance the user experience when using L2 rollups.

It is important to remember that rollups do not need a sequencer; it is simply a design choice made for a better user experience. For example, rollups can also use Ethereum L1 for sequencing. However, the underlying layer of Ethereum is likely to be relatively inefficient and expensive, especially considering the high volume of L2 transactions. The transaction throughput of rollups will be limited by the data sequencing rate of Ethereum L1. Users will also experience the same confirmation delays as transactions on Ethereum. This means that, so far, every major L2 project has found that running a centralized sequencer is more convenient, cheaper, and user-friendly. Although L2 users can submit transactions directly to L1 to bypass the sequencer, they must pay gas fees to L1, and transactions may take longer to finalize. This largely undermines the purpose of using L2 rollups to execute transactions.

Given that the sequencer controls the order of transactions, theoretically it has the right not to include user transactions (. However, users can submit transactions directly to L1 ). The sequencer can also extract MEV from the transaction pool, which may cause economic losses to the user group. If there is only one sequencer, the risk of centralization is even greater. In such cases, validity can become a problem; if the only sequencer fails, the entire rollup will be affected. A multi-sequencer setup can reduce this risk.

With this setup, the sorter can be viewed as a semi-trusted party for the user. Although the sorter cannot prevent users from using L2, it can delay user transactions, causing users to pay additional gas fees and extract value from user transactions.

The correlation of MEV

MEV refers to the value obtained from block production, exceeding the first-order mining ( or staking ) block rewards and gas fees. It is extracted by manipulating transactions within the block, specifically through inclusion, exclusion, and altering the transaction order. Common forms include front-running and sandwich attacks.

Given the role of sorters in L2 rollup, they can understand all user transactions off-chain. Moreover, since these sorters are often operated by the projects themselves or affiliated teams, many users are concerned about not being able to see potential MEV extraction. Even without these concerns, as projects operate their own centralized sorters, the trustworthiness and degree of decentralization of these protocols will also be called into question.

Status of the Sorter Market

Currently, all major Ethereum L2 versions rely on centralized orderers. As more and more Ethereum transactions move to L2 solutions, despite the fact that the Ethereum validator set itself is decentralized, a large number of transactions (, namely those on L2, will be influenced by centralized forces in the form of a unique orderer.

As expected, most of these companies have already included the decentralization of their respective sequencers as part of their roadmap. While this is a positive sign indicating that decentralization is part of the L2 vision, we should note that Arbitrum and Optimism have launched solutions since the end of 2021 and have not made significant progress in terms of decentralized sequencers.

Most top companies seem to allocate resources to improve core products and features rather than focusing on decentralization. This is not entirely a criticism, as in a highly competitive environment, focusing on decentralization before having a competitive product is not in any company's best interest, which is somewhat understandable. However, as internet companies mature, this perspective is changing, and the discussion is rapidly shifting towards sorting decentralization and increasing credibility.

) Other issues

It is worth emphasizing that there is some discussion about the level of risk associated with relying on centralized sorters.

As mentioned above, since the sequencer controls the order of transactions, they can exclude user transactions and extract MEV. However, the sequencer ultimately cannot completely exclude users from rollup transactions. Users can bypass the sequencer and submit transactions directly to L1 as long as they are willing to pay the increased gas costs. Although misbehaving sequencers may cause transaction delays and users to incur additional fees, they cannot completely censor transactions. Prior to this, no major L2 company was extremely focused on decentralizing its sequencer, which is likely one of the reasons. Nevertheless, the reordering of transactions by sequencers to extract MEV remains a problem, especially in private mempools like the OP Mainnet.

Perhaps the bigger issue is effectiveness. Given that the main rollup programs are running a single centralized sequencer, if these sequencers encounter problems, the entire rollup program could be adversely affected. While users can still complete transactions by directly accessing L1, this is not a particularly sustainable method, and it is unlikely to work for most transactions. Remember, the whole point of using L2 rollups is to save on transaction costs. Given that one of the fundamental ideas behind cryptocurrency is to prevent reliance on a single centralized provider, just like in traditional finance, the centralization of sequencers is clearly an important issue that needs to be addressed, and it is one of the key unlocks that shared sequencers will bring to the L2 rollup market.

![Binance Research Institute interprets Decentralization Sorter: The cornerstone of Layer 2 security, expanding the use cases for native tokens]###https://img-cdn.gateio.im/webp-social/moments-825829026dfe6eac7918343e0340ba6a.webp(

4. Solution: Decentralized Shared Sorter

) Overview

The new solution to the above problems is a decentralized shared sorter. Although the solutions of different projects vary, the fundamental idea of replacing a single centralized sorter is the same. Here, "shared" refers to multiple different rollups being able to use the same network, with transactions from multiple rollups being aggregated in a memory pool before sorting, which helps reduce the possibility of MEV extraction and censorship (. "Decentralization" refers to the concept of leader rotation, meaning that it is not always a single actor who sorts all transactions, but a leader selected from a group of decentralized actors. This helps prevent censorship and provides validity guarantees.

This is very similar to how various different L1s operate using leader rotation mechanisms. In fact, building a decentralized sorting layer is similar to building a decentralized L1, which requires the construction of a validator set. Different projects have taken different approaches to meet this requirement.

The shared sequencer is designed to alleviate the MEV extraction problem, provide censorship resistance, and enhance the effectiveness guarantees of rollups, that is, addressing the issues faced by centralized sequencers. Additionally, there are two more points worth noting:

• Decentralization as a Service: The shared sequencer solution aims to provide sequencing decentralization services for any number of rollups. All these rollups will then benefit from the censorship resistance and efficacy that a decentralized network can offer, without having to establish that network themselves. Given that this could be a very costly and time-consuming process, this is a major selling point of the shared sequencer network. Keep in mind that no company has decentralized its sequencer yet, and most of them have sufficient funds to do so, which means this is not a trivial issue. If companies like Astria or Espresso can provide out-of-the-box sequencing decentralization services, then rollup companies can continue to focus on differentiation and optimizing performance, thereby better serving different users.

• Cross-rollup composability: Since these shared sequencer solutions are designed to handle transaction ordering for multiple rollups, they can provide unique interoperability guarantees that are currently unavailable. For example, users should be able to specify that transactions on Rollup 1 can only be included in a block if different transactions on Rollup 2 are also included in the same block.