Original Title: Ethereum’s Rollups are Centralized: A Look Into Decentralized Sequencers
Source: Binance Research
1. Key Points
❖ Transaction sequencing has become an increasingly serious issue in the Layer 2 (L2) space. The main role of L2 rollups is to provide a secure place for cheap transactions. L2 rollups provide an execution layer for users and then submit their transaction data to the higher-level Layer 1 (L1), such as Ethereum’s Arbitrum, Optimism, zkSync, etc.
❖ Sequencers are entities with the authority to group these transactions into ordered batches. Sequencers receive unordered transactions from users, process them off-chain into batches, and generate compressed, ordered transactions. These transactions can then be included in blocks and sent to the parent L1.
- LD Capital Product Expansion and Mechanism Updates of Lybra V2
- Overview of Global Cryptocurrency Transaction Taxes Current Taxation Status and Future Outlook
- LianGuai Morning News | Analysis Bitcoin’s halving in 2024 could push prices to $400,000
❖ Rollups don’t actually require sequencers; this is just a design choice aimed at providing a better user experience with lower fees and faster transaction confirmation. For example, just as most rollups use the Ethereum base layer for data availability, they could also use the base layer for sequencing. However, the Ethereum base layer may be relatively inefficient and expensive. This means that, so far, every major L2 rollup project has found it more convenient, cheaper, and user-friendly to operate a centralized sequencer.
❖ Since sequencers control the ordering of transactions, they have the authority to censor user transactions (although full censorship is unlikely as users can submit transactions directly to L1). Sequencers can also extract maximum extractable value (MEV), which can cause economic losses to the user base. Additionally, validity can also be a major concern, i.e., if the sole centralized sequencer goes down, users will be unable to use that sequencer, and the entire rollup will be affected.
❖ The solution to this problem is shared, decentralized sequencers. Shared sequencers essentially provide decentralized services for rollups. In addition to addressing issues like censorship, MEV extraction, and validity, shared sequencers also introduce cross-rollup functionality, opening up various new possibilities. Espresso, Astria, and Radius are developing innovative shared sequencing solutions, each with unique features in their respective architectures. Espresso is attempting to leverage EigenLayer to bootstrap its network, while Astria maintains close ties with the modular data availability network Celestia. Radius, on the other hand, brings its unique encrypted mempool into the conversation.
As the Ethereum L2 rollup ecosystem continues to grow, one aspect that is often overlooked is sequencers. Sequencers are responsible for transaction ordering and using sequencers through rollups can provide a better user experience, lower fees, and faster transaction confirmation. However, the issue is that, so far, all major Ethereum L2 companies have found it most convenient, user-friendly, and cost-effective to run their own centralized sequencer. Considering the power that sequencers possess in terms of transaction censorship, MEV extraction, and creating single points of failure (i.e., validity issues), this may be seen as an undesirable outcome and not in line with the spirit of cryptocurrencies.
Although most cryptocurrency companies have solved the decentralization problem of their own sequencers and included it as part of their roadmaps, there is still no real consensus on how to achieve decentralization. It should also be noted that Arbitrum and Optimism have launched their own solutions since the second half of 2021, but they have not made substantial progress in decentralized sequencers.
In this report, we will carefully study the role of sequencers and the current situation of the Ethereum rollup space. Then, we delve into the projects that are researching solutions, namely decentralized shared ordering networks. We will provide detailed introductions to these projects and the uniqueness of their solutions. We also contemplate what this means for the future development of the Ethereum L2 rollup space.
3. What is a sequencer?
At a basic level, a blockchain is a distributed data ledger composed of transaction data with timestamps sorted into blocks. Initially, this transaction data is unordered and unorganized. After being sorted, it can be organized into blocks and executed to create a new state of the blockchain. For first-layer (“L1”) blockchains like Ethereum, this transaction ordering occurs within the Ethereum base layer itself.
In the most popular scalability solution for Ethereum, Layer-2 (“L2”) rollup layer, transaction ordering has become an increasingly serious problem. Remember, the main purpose of rollups is to provide users with a secure and cheap transaction environment. In simple terms, L2 rollups provide an execution layer for users, and then submit their transaction data to the parent L1, such as Ethereum’s Arbitrum, Optimism, zkSync, and so on. The batch of transactions submitted to L1 usually contains hundreds or thousands of compressed L2 transactions, reducing the cost of sending data to L1.
In the L2 rollup world, a sequencer is an entity with the authority to group transactions into batches. The sequencer receives unordered transactions from users, processes them off-chain into batches, and generates batches of compressed and ordered transactions. These transactions can then be put into blocks and sent to the parent L1. Batches of transactions can also be used on the data availability (“DA”) layer, usually Ethereum, for most current rollups. It also provides users with soft commitments, meaning that after receiving a user’s transaction, the sequencer provides an almost instant receipt as a “soft confirmation” (1). “Hard confirmation” is received after the transaction is sent to the L1 layer.
Figure 1: Application scope of sequencers?
Why do Rollups use sequencers, and why is it a problem?
Fundamentally, the sequencer has a very clear goal: to improve the user experience. Using a sequencer for L2 transactions is like using the “express lane,” which means lower fees and faster transaction confirmations. In fact, the sequencer can batch and compress hundreds or thousands (2) of L2 transactions into a single L1 transaction, saving gas fees. Additionally, the soft confirmations provided by the sequencer mean that rollup transactions can provide users with fast block confirmations. This combination helps improve the user experience of using L2 rollups.
It is important to remember that rollups do not require a sequencer; it is simply a design choice made for better user experience. For example, just like most rollups use Ethereum L1 to improve data availability, they can also use it for ordering. The Ethereum Foundation’s Justin Drake recently referred to these as “rollup-based” (3). However, Ethereum’s base layer is likely to be relatively inefficient and expensive, especially considering the large volume of L2 transactions. Essentially, the transaction throughput of rollups will be limited by the Ethereum L1 data ordering rate. Users will also experience the same transaction confirmation delays as trading on Ethereum. This means that so far, every major L2 scalability project has found it more convenient, cheaper, and easier for users to run a centralized sequencer. Although L2 users can submit transactions directly to L1 to bypass the sequencer, they must pay transaction gas fees to L1, and transactions may take longer to complete. This largely contradicts the original intention of using L2 rollups to execute transactions.
Figure 2: Sequencers can help aggregate multiple transactions into a single L1 transaction, making transaction costs on L2 several times lower than on Ethereum L1.
Given that the sequencer controls the ordering of transactions, in theory, it has the right to exclude user transactions from it (although if users have the ability and willingness to pay gas fees, they can also submit transactions directly to L1). The sequencer can also extract MEV (more details later) from the transaction batch, which can cause economic losses to the user group. If there is only one sequencer, as is the case with all major rollup transactions currently, the centralization risk is greater. In this case, validity may become an issue, meaning that if the only sequencer fails, the entire rollup will be affected. A multi-sequencer setting can reduce this risk.
In this setting, the sequencer can be seen as a semi-trusted party for users. While the sequencer cannot prevent users from using L2, it can delay their transactions, causing users to pay additional gas fees and extract value from their transactions.
Relevance of MEV
MEV is particularly important here. MEV refers to the value obtained from block production that goes beyond the first-order mining (or staking) block rewards and gas fees. It is the value extracted by manipulating transactions within a block, i.e., by including, excluding, and changing the order of transactions. For example, common forms of MEV extraction include frontrunning and sandwich attacks.
Given the role that sequencers play in L2 rollups, they have access to all user transactions off-chain. Additionally, since these sequencers are usually operated by the project itself or affiliated teams, such as the Optimism Foundation for OP Mainnet (4) and the Arbitrum Foundation for Arbitrum One and Nova (5), many users are concerned that they cannot see potential MEV extraction. Even without these concerns, the credibility and decentralization of these protocols are also impacted by running their own centralized sequencer. The trustworthiness and decentralization of these protocols will certainly be questioned.
The Current Situation of Sorter Market
At the time of writing this article, all major Ethereum L2 versions rely on centralized sorters. As more and more Ethereum transactions move to L2 solutions, although the validator set of Ethereum itself is decentralized, it seems that a large number of transactions (i.e., those on L2) will be influenced by centralized power in the form of a unique sorter.
Figure 3: All top Ethereum L2 rollups use proprietary centralized sorters
As expected, most of these companies have already addressed the decentralization issue of their respective sorters and have included it as part of their roadmap. While this is a positive sign that decentralization is part of the L2 vision, we should note that Arbitrum and Optimism have already launched their own solutions since the end of 2021, indicating that they have not made substantial progress in the decentralization of sorters.
Figure 4: All top rollups address the decentralization issue of sorters in their documents
Most top companies seem to allocate resources to improving their core products and features rather than focusing on decentralization. This is not entirely a criticism, as it is understandable to focus on decentralization after having competitive products in a fiercely competitive environment. However, as network companies mature, this view is changing, and discussions are quickly shifting towards sorter decentralization and improving credibility.
It is worth noting that there is some discussion about the level of risk posed by relying on centralized sorters.
As mentioned above, since sorters control the sorting of transactions, they can exclude user transactions and extract MEV. However, sorters cannot completely exclude users from rollup transactions. Users can bypass sorters and submit transactions directly to L1 (as long as they are willing to pay the increased gas cost). Although improper sorters may cause transaction delays and additional costs for users, they ultimately cannot fully censor them. This is likely one of the reasons why no major L2 company has been extremely focused on decentralizing its sorters. Nevertheless, sorter reordering transactions to extract MEV remains a problem, especially in private mempools like OP Mainnet(6).
Perhaps the bigger issue is real-time performance. Given that major rollup protocols are running unique centralized sorters, the entire rollup program will be adversely affected if these sorters encounter problems. While users can still complete transactions by accessing L1 directly, this is not a particularly sustainable method and is unlikely to work for most transactions. Remember that the whole point of using L2 rollups is to save transaction costs. Given that one of the fundamental ideas behind cryptocurrencies is to prevent reliance on a single centralized provider (as in traditional finance), the centralization of sorters is clearly an important issue that needs to be addressed and one of the key unlocks that shared sorters will bring to the L2 rollup market.
4. Solution: Decentralized Shared Sorter
The new solution to address the above problems is a decentralized shared sorter. While the specific solutions may vary between different projects, the fundamental idea is to replace the single centralized sequencer. The term “shared” refers to the ability for multiple different rollups to utilize the same network, meaning that transactions from multiple rollups are aggregated in a memory pool before being sorted (which helps reduce the possibility of MEV extraction and front-running). The term “decentralized” refers to the concept of leader rotation, where transactions are not always sorted by a single actor, but a leader is selected from a decentralized group of actors. This helps prevent censorship and provides guarantees of validity.
This operates in a similar way to various L1s that use leader rotation mechanisms. In fact, building a decentralized sorting layer is similar to building a decentralized L1, as it requires constructing a validator set. As we will see later in this section, different projects have taken different approaches to meet this requirement.
The shared sorter aims to mitigate the MEV extraction problem, provide censorship resistance, and improve the validity guarantees of rollups (as described above). Additionally, there are two noteworthy points:
- Decentralized as a service: The shared sorter solution aims to provide decentralized sorter services to any number of rollups. In turn, all these rollups can benefit from the censorship resistance and real-time capabilities provided by the decentralized network without having to build the network themselves. Given that this can be a very expensive and time-consuming process, this is a major selling point for the shared sorter network. Keep in mind that currently no company has decentralized its sorter, and most of them have enough funding (7)(8)(9) to do so, which means this is not a completely trivial problem. If companies like Astria or Espresso can provide ready-to-use decentralized sorter services, rollup companies can continue to focus on differentiation and performance optimization, better serving different users.
- Cross-rollup composability: Since these shared sorter solutions aim to handle transaction sorting for multiple rollups, they can provide unique interoperability guarantees that are currently not possible. For example, users should be able to specify that transactions from Rollup 1 can only be included in a block if different transactions from Rollup 2 are also included in the same block. By enabling this conditional transaction inclusion, the shared sorter can unlock new possibilities, including atomic cross-rollup arbitrage.
Many projects are researching shared sorting solutions. We will focus on a few and their strategies in the following sections.
Espresso Systems is a company dedicated to bringing Web3 tools into the mainstream, with a particular focus on L2 rollups and the Ethereum ecosystem. Prior to developing the shared sorter, they have been working on improving blockchain privacy and have developed the CAPE (10) application. They have also contributed to open-source developer tools through initiatives such as the Jellyfish (11) cryptographic library and Hyperplonk (12).
In November 2022, Espresso started sharing their work on the Espresso Sequencer.
The Espresso Sequencer is a decentralized shared ordering network designed for decentralized rollups, providing secure, high throughput, low latency transaction ordering and data availability.
Its design purpose is to handle decentralized ordering and data availability for rollups, acting as a middleware network between the rollup and the underlying L1.
The design of the Espresso Sequencer is agnostic to the virtual machine (“VM”), meaning it can be used for non-Ethereum virtual machines as well as zero-knowledge (“zk”) virtual machines and optimistic virtual machines.
- How does it work?
The core of the Sequencer is the consensus protocol HotShot. HotShot is based on the HotStuff (13) consensus protocol and combines the latest developments from various domains (14) such as pacemakers and verifiable information dissemination (“VID”).
HotShot is open and permissionless, decentralizing the power in the Sequencer network while providing high throughput and fast finality, as well as ensuring security and effectiveness. HotShot adopts a Proof-of-Stake (PoS) security model, and one of the key requirements posed by the Espresso team is to achieve strong performance without impacting the scale of the validator set. Specifically, HotShot should be able to scale to include all Ethereum validators (currently over 700,000 (15)) participating.
Espresso Systems aims to achieve Ethereum-level security for its Sequencer by utilizing the existing Ethereum validator set. There are two key reasons for this setup:
Security: The cost of launching a decentralized PoS consensus protocol is extremely high and requires significant energy consumption. Even then, acquiring a sufficient number of network participants can be a significant challenge. By using the same validators as Ethereum, the Sequencer can achieve a level of security, effectiveness, and decentralization that would be difficult to achieve on its own. The Espresso Sequencer can benefit from sharing the cryptographic economic security of the second-largest decentralized cryptocurrency, recognized as second only to Bitcoin.
Incentivized Consensus: Conceptually, it makes sense to involve Ethereum L1 validators in running the protocol that powers Ethereum L2 rollups. In practice, in a centralized Sequencer setup, almost all costs and MEV generated by the rollup could be captured by the Sequencer. If this value is not shared (or shared minimally) with L1 validators, there is a concern about the impact on the security of the rollup. For example, L1 validators could be bribed to fork the rollup and gain more profit than honestly managing the rollup contract. Decentralizing the Sequencer and collaborating with L1 validators ensures its security, reducing such concerns.
Espresso will seek to establish this collaboration by resetting contracts, particularly with EigenLayer. Through the revaluation by EigenLayer, users can stake their Ethereum and Ethereum liquidity staking tokens (“LST”) across multiple protocols, extending the economic security beyond Ethereum itself. By doing so, they can earn fees as rewards but also agree to additional reduction conditions. Re-staking is a subsidized way to enter the system as stakers do not need to deploy additional capital, only use their previously staked Ethereum. This lowers the capital cost of securing other protocols, meaning the Espresso Sequencer can tap into the capital base and decentralized validator set of Ethereum without launching its own validator set.
- Tiramisu Data Availability (16)
As mentioned earlier, most rollups rely on L1 blockchains (such as Ethereum) to provide data. However, this is not ideal because block space on L1 blockchains like Ethereum is scarce and expensive, resulting in high transaction fees for users – an undesirable outcome. Espresso Systems addresses this issue with its efficient Tiramisu data availability solution.
Like the classic Italian dessert, the Tiramisu solution has three novel layers. Together, they ensure data is made available to the parties that need it – in our case, the various rollups that order transactions.
The foundational layer of Tiramisu is called Savoiardi. It is an anti-bribery layer (similar to Ethereum’s Danksharding proposal) that provides the highest level of security. However, due to this feature, it is the least user-friendly layer among the three. To address this, Espresso adds two additional layers to its solution.
Mascarpone is the middle layer, ensuring efficient data recovery by electing a small data management committee.
Cocoa is appropriately named the “topping” of the entire system. Cocoa helps Tiramisu deliver “Web2-level performance” by providing a content delivery network for Tiramisu. This contributes to efficient data recovery and significantly speeds up data propagation. Given that this layer is inherently centralized (17), it is entirely optional, and Tiramisu can work perfectly without it. It helps accelerate data availability and can be easily modified or removed.
It is worth noting that Espresso Systems designed its protocol with flexibility and modularity in mind. Rollup devices using its sequencer can choose to use Tiramisu or any other data availability solution if they prefer.
Figure 5: Three layers of the Tiramisu data availability solution
- Notable Partners (18)
Since July, the Espresso Systems team has been continuously announcing partnerships. EigenLayer was the first company to announce such a partnership, given its significance in the Espresso Sequencer architecture, and it is worth closely monitoring its developments. EigenLayer itself launched its first mainnet on June 14.
Alongside the announcement of the Doppio testnet, Espresso also announced a partnership with Polygon zkEVM. This partnership represents the first end-to-end integration of Espresso’s sequencer with a full-fledged zk-rollup (a fork of Polygon zkEVM). The testnet allows users to submit transactions to the fork and then have them routed and ordered by nodes running Espresso’s HotShot protocol.
Espresso supports the integration of Injective’s sequencer into Cascade, enabled by the Injective chain with IBC (19) Cosmos SDK. Cascade is the first cross-chain Solana SVM rollup in the IBC ecosystem, allowing for the deployment of Solana contracts on Injective and the broader IBC ecosystem for the first time. Integration with Cascade’s testnet is expected to be completed by the end of 2023, with the mainnet expected to be completed in 2024.
AltLayer has also joined the Espresso Systems ecosystem. AltLayer is a rollup-as-a-service platform that allows developers to launch highly scalable rollups with support for multiple virtual machines. Through this collaboration, developers will be able to decide whether to use AltLayer’s solution and/or Espresso Sequencer to launch their rollups. The teams will also work together to develop other integrated products and see how their designs complement each other.
Espresso Systems is partnering with Caldera to deploy an optimistic rollup based on the OP Stack, which utilizes Espresso Sequencer and Tiramisu. Caldera enables developers to deploy custom rollups for their applications. After deploying this extension, future L2s built on Caldera will have the option to easily incorporate Espresso Sequencer and Tiramisu as plug-in components for their extensions.
Spires, a Layer 3 (L3) as a service company, has announced its integration with Espresso Sequencer and Tiramisu. Spires’ infrastructure allows developers to deploy their own L3 app chains on top of zkEVM L2. Spires will work with the Espresso team to integrate their solution into the Spire L3 framework. The test network is expected to be completed in 2024.
- Latest Updates
November 28, 2022: Americano is the first test network for Espresso Sequencer and HotShot. The initial post contains more technical details, but it should be noted that this is an internal test network and not open to the public.
Figure 6: Project roadmap released with Americano testnet and initial announcement
- July 20, 2023: Doppio is the second major milestone and testnet for HotShot and Espresso Sequencer. At the same time, Espresso Systems releases the whitepaper for the entire project. Doppio brings many efficiency improvements to HotShot, including Verifiable Information Decentralization (VID), a new view synchronization sub-protocol, and signature aggregation of quorum certificates (20). Doppio also implements the first two layers of Tiramisu, with the future testnets expected to include the third and final layer. Espresso Systems also releases the first end-to-end integration of their sequencer with a full-fledged zk-rollup, specifically a fork of Polygon zkEVM.
- August 4, 2023: Doppio testnet opens to the public. Documentation on how users can submit transactions to the zkEVM fork is also released. Performance benchmarks (21) and expected next steps are also published. Specifically, they announce the start of onboarding some rollups and rollup-as-a-service companies into their sequencer. They also announce their contribution to the OP Stack through the concept of Optimism Leader Election Work (following the recently accepted RFP (22)).
Astria is building a shared sequencer network and is one of the leading companies in phasing out centralized sequencers. At the same time, they are developing Astria EVM, which will be the first rollup supported by their shared sequencer network. The project will obtain fast and censorship-resistant transaction ordering from its network and will leverage Celestia for data availability. Celestia is a modular blockchain network and DA layer that Astria is familiar with. Founder Josh Bowen has previously worked on Celestia, and the project and its ecosystem are mentioned multiple times in Astria’s introductory blog.
Astria’s shared sequencer network allows multiple different rollups to share a single, permissionless, decentralized sequencer network. With this network, Astria provides an out-of-the-box solution that allows rollups to have censorship resistance, fast block confirmations, and atomic cross-rollup composition capabilities.
- How does it work?
Astria’s shared sequencer network itself is a middleware blockchain that achieves consensus on a set of ordered transactions using CometBFT (a fork of Tendermint Core). The design goal of this network is to accept transactions from multiple rollups, then order them into a block and write them to the data availability (DA) layer.
Rollups can immediately retrieve ordered blocks from Astria after creating a block, providing fast block confirmations to users through “soft commitment”. Alternatively, rollups can retrieve ordered blocks from the DA layer for “hard commitment”, as once written to the DA layer, transaction orderings are considered final. This provides users with the strictest finality, which can be particularly useful in high-value transactions, among other scenarios.
Figure 7: Astria’s shared sequencer network
- Astria EVM
As mentioned above, Astria EVM will be the first cryptocurrency powered by Astria’s shared sequencer network.
Currently, most rollup projects execute and order transactions themselves, using Ethereum as the data availability layer. Astria EVM will focus on execution while utilizing Astira’s shared sequencer for ordering and Celestia for the DA layer.
Figure 8: Three key layers focusing on the L2 process, showing how rollups tend to leverage their proprietary sequencer and Ethereum’s DA capabilities (we also show Ethereum L1 itself for comparison)
The goal of Astria’s EVM is to serve as a liquidity and bridge hub to help bootstrap the Celestia rollup ecosystem. This also means that the Astria team has a living test case to understand how rollups can best integrate with their shared sequencer network.
Astria’s future vision includes thousands of decentralized sovereign rollups, each tailored to unique use cases and applications.
Their shared sequencer network plays a key role in their vision, simplifying the rollup development process. Their solution means that rollup developers can focus on innovative use cases while being able to easily integrate with decentralized networks, providing them with fast, censorship-resistant transaction ordering and cross-rollup composability.
- Astria Development Cluster
On August 16th, Astria released its development cluster (24), which includes all the different components needed to launch rollups on Astria’s shared sorter network. The goal of this cluster is to make the development and testing of the Astria network, as well as integration with Astria, as simple as possible.
- The components include:
Astria Sorter: A block generating node used for transaction sorting. The development cluster relies on a single node. In the main network, a decentralized set of nodes will be used.
Data Availability Layer: The local Celestia network that provides hard terminations.
Rollup: Geth (25) rollup node used for task execution and state storage.
Composer: Retrieves pending transactions from the rollup’s mempool and submits them to Astria’s CometBFT mempool.
Conductor: Filters these program blocks for each rollup after receiving a single program block. These filtered blocks are then passed to the rollup for execution.
Relayer: Sends sequentially ordered data blocks to the conductor and data availability layer Celestia.
Recently, Astria announced the deployment of rollup technology on its development cluster, and we will pay attention to which companies decide to deploy rollup technology.
Figure 9: Different components of Astria’s development cluster
- Latest Updates
In April 2023, Astria announced a $5.5 million seed round investment (26).
As mentioned above, in August 2023, the team unveiled their development cluster.
The Astria team is also developing a Devnet to kickstart related work. It is expected to be completed in the coming weeks.
Their code is open source, and further documentation can be found on their official GitHub page.
Radius is building a trustless shared sorting layer that uses encryption technology to decentralize sorters, prevent censorship, and minimize harmful MEV. Their solution is blockchain-agnostic and can be used for various types of rollups.
- How does it work?
Radius uses an encrypted mempool to achieve its goals. Essentially, the content of each user transaction is encrypted upon submission. When the sorter sorts the transaction group, it cannot see the content of each transaction, preventing the sorter from extracting MEV or conducting audits.
Figure 10: Radius transaction flow
This ultimately means that Radius’ solution can address MEV and censorship issues with just one sorter. Since the transaction content is encrypted, even a single sorter cannot act maliciously. This means that there is no need to introduce a consensus mechanism, which could be advantageous in terms of speed and scalability. This is also the difference between Radius’ solution and Astria and Espresso solutions, both of which rely on consensus mechanisms to sort transactions.
Although the encrypted mempool on a single sorter solves two key issues of centralized sorters: MEV and censorship, it still has a single point of failure. In order to ensure real-time performance, Radius adopts a decentralized sorter network with multiple sorters running simultaneously. One of these sorters is selected to run as the sorting layer. There are various suggestions (27) on how to choose a single sorter, including secret election mechanisms and sorter group sharding.
- Practical Verifiable Delay Encryption (PVDE)
Radius adopts the zk-based encryption scheme PVDE (28) to create an encrypted mempool.
User transactions are temporarily encrypted based on time-locked puzzles. Then, the sorter sorts the encrypted transactions. The sorter needs to unlock the locked puzzles to obtain the decryption keys. This takes time and computational resources, and prevents the sorter from decrypting transactions too early (i.e., before transaction sorting).
To prevent attacks, users generate ZK proofs to prove the validity of their transactions and decryption keys. The sorter can verify these proofs before sorting, effectively preventing meaningless decryption (i.e., attacks) and resource waste.
- MEV Market
Radius also proposes an optimized block space design. They attempt to create an auction-based market (29) where traders submit bundled cross-rollup MEV transactions. The transaction with the highest bid will be included in a block by the sorter, helping maximize the cross-rollup arbitrage profit and creating a more efficient rollup market.
- Latest Updates
In June 2023, Radius announced the completion of a $1.7 million seed round financing.
While we have introduced some of the larger and more well-known projects in the shared sorting field, there are also other projects developing similar or closely related solutions.
NodeKit: The NodeKit team is building NodeKit SEQ, a decentralized shared sorter built into a custom L1 blockchain.
They are also building NodeKit Chain, which is an EVM-based rollup.
Their Twitter page also indicates that their solution will be launched on the Avalanche subnet (30).
AltLayer: AltLayer is a “rollup-as-a-service” platform that allows developers to launch highly scalable L2 rollups with support for multiple virtual machines.
Although “rollup-as-a-service” companies are a separate field and not within the scope of our report, AltLayer’s decentralized sorter network (31) is worth mentioning.
AltLayer’s shared sorter network is called Beacon Layer, which is a permissionless middleware blockchain. Nodes in the blockchain are called validators (similar to any PoS network).
When users want to create a rollup using AltLayer’s platform, they can specify the number of sorters required to run the rollup, the minimum staking amount required for each sorter, and the token set that the staked assets can be priced in. AltLayer recommends having at least five different sorters for each rollup project.
Once validators join the beacon chain and provide the minimum collateral, they can act as sorters in different rollups. The beacon chain selects validators to be sorters for various rollup projects based on their stake and some randomness. Like any PoS blockchain, there is a risk of collateral reduction if improper behavior occurs.
This process means that developers can relatively quickly deploy a cryptographic rollup using the infrastructure of AltLayer and then use the beacon chain to ensure its decentralization. If you agree with the idea of rollup-centric future, services like AltLayer are definitely worth paying close attention to.
Existing L2 rollups seem to have to make choices. On the one hand, they can maintain the status quo and continue to use a single centralized sorter. On the other hand, they can start integrating with third-party shared sorting networks or develop their own internal solutions.
1. Continue to use a single centralized sorter:
a. This is the simplest and possibly the most financially cautious approach. The monetization of sorters is an important source of revenue for all major scaling solutions (32), undoubtedly an important part of the business model. In fact, Base, a newly established L2 promotion company, recently confirmed its intention to monetize sorters during Coinbase’s Q2 earnings conference call (33).
b. Maintaining a centralized sorter raises issues such as censorship, MEV extraction, and single point of failure risks. In addition, it goes against the fundamental spirit of cryptocurrencies. Imagine a scenario where a key member of a major crypto organization mysteriously disappears or gets into serious trouble. If they run a centralized sorter, it is likely to affect their crypto rollup, daily operations, and user experience. If this happens, many other participants in the industry are likely to start seriously committing to decentralized sorters according to their roadmaps. This is a simple example that illustrates why the decentralization of sorters may be more important than it initially appears.
2. Integrate with third-party shared sorting networks:
a. With the continuous development of shared sorting networks like Espresso and Astria and the launch of mainnets, this will become a major choice for existing networks. In fact, considering the integration of Espresso with Polygon zkEVM fork, some major network companies seem to be actively exploring this solution.
b. Compared to the risks of managing centralized sorters or the effort and cost required to develop internal solutions, outsourcing sorting work to experts is a wise choice for many companies.
c. One of the most important factors to consider here is the interoperability of the sample libraries. This may be one of the most obvious advantages of running L2 on shared sorters compared to those running in their proprietary “silos”. As emphasized earlier in this report, running on shared sorters and the interoperability it brings can unleash various new possibilities, including cross-rollup arbitrage and conditional transaction inclusion.
3. Developing proprietary solutions:
a. Because this may be the most time-consuming and expensive option among the three, it will be interesting to see which companies choose this path.
b. So far, a key issue with large-scale cryptocurrencies we have seen is the accumulation of token value. Most top Ethereum L2 companies have used ETH as the token for gas fees, which prevents their native tokens from accumulating value. One possible solution is for rollup companies to develop internal sorting solutions backed by token holders; for example, users can stake their native rollup tokens and become sorters, charging fees for their services.
c. The drawback of this approach is its impact on interoperability. Compared to systems running their own proprietary sorting solutions, rollup systems running on shared sorters have better interoperability.
d. Optimism announced its “Law of Chains,” which is a recent development worth considering. The Law of Chains is a set of guiding principles for chains within the OP Stack superchain ecosystem. Its essence is to establish a framework for these chains to work in a more unified way. This may likely extend to shared sorting solutions for OP Stack-based chains, which could be a solution to the interoperability issue discussed above (at least for OP Stack chains).
As layer-two cryptocurrencies continue to emerge and grow in scale and transaction volume in the crypto world, issues surrounding centralization and interoperability will continue to ferment. This topic has received much attention in the past year, and we expect it to expand further as major cryptocurrency issuers approach their one-year and two-year anniversaries, and more cryptocurrency issuers enter the market.
We believe that at least some companies will choose to integrate with third-party sorter networks like Espresso and Astria, but we also see other companies choosing to develop their own internal solutions. Some larger companies, especially those that have launched native tokens, are likely to see value in developing their own solutions, allowing them to maximize profits and increase token utility. Whatever happens, this is a very important aspect we need to pay attention to, and we will be closely watching with interest.
Users want and prefer faster transaction confirmation and lower fees. While centralized sorters have been the solution of choice for major L2 companies so far, ideally, companies and users should be able to choose the best decentralized version of this technology. This is where companies like Espresso Systems, Astria, Radius, and others play a key role in the L2 story.
Two key driving factors here are decentralization and rollup interoperability. Decentralization is critical for many reasons. It is the philosophical foundation of cryptocurrencies, among other things. On a more practical level, centralized sorters represent single points of failure that can impact the effectiveness of cryptographic rollups and pose a threat to the recovery capability of cryptographic rollups. Not to mention the possibility of extracting a significant amount of MEV, some of which may be hidden from users and extracted in private mempools. The possibility of censorship (even temporary) and delayed transactions is also a concern to bear in mind, especially when considering the industry’s desire for strong growth. Interoperability of cryptographic rollups is equally important, particularly when people hold a view that the future of the crypto industry revolves around cryptographic rollups. If there are more and more cryptographic rollups in the market, whether for specific applications or other purposes, these cryptographic rollups should be able to communicate and collaborate seamlessly with each other. Otherwise, how can we achieve a Web2-type user experience?
The future will definitely bring challenges, and some large companies may prefer to create their own proprietary solutions rather than using shared ordering networks. One way to address this issue is for shared ordering networks to solve the problems of value accumulation and income distribution through economic mechanisms, as sharing a sorter among many companies will eventually achieve strong network effects.
This topic will become more intense in the coming months, and we believe that there will be many new participants entering the market, whether in the rollup field or in the shared sorter field. It will be very interesting to observe the choices of different projects. We look forward to closely following this trend.