导语：排序器（Sequencer）用于解决以太坊网络中的可扩展性和性能限制。它的引入旨在通过将链上交易进行批量处理和排序，将大部分计算和数据存储的工作转移到Layer2来减轻以太坊主链的负担。去中心化排序器的天花板在于其所依赖的底层协议以及网络环境。尽管去中心化的排序器可以提高系统的安全性和韧性，但仍存在一些限制和挑战。
排序器是Rollup网络的核心组成部分，它负责接收交易、排序交易、执行交易和提交交易数据等关键操作。如果网络的单一排序器出现故障或不可用，则整个网络将停止处理交易。然而，很多现有的Rollup方案仅有一个单一的排序器，导致它们的去中心化程度远不如某些中心化的替代Layer 1方案。因此，去中心化排序器的重要性不言而喻，而有潜力的去中心化排序器方案应当通过设计和实施上的改进，有效提升系统的去中心化性质。
排序器的重要性
现有的Rollup方案包括基于零知识证明（ZK）的Rollup和基于乐观执行的Rollup。这些方案在提高可扩展性方面相比单体式的Layer 1方案有着更好的设计。然而，它们也存在一些问题：
零知识证明（ZK）Rollup的问题：
- 计算复杂性：使用ZK证明来验证交易的正确性和合法性需要大量的计算资源和时间。这可能导致交易处理的延迟和高昂的计算成本。
- 可验证性的依赖：ZK Rollup依赖于外部可验证性，即需要外部监督者验证ZK证明的正确性。这可能引入信任问题和中心化风险。
乐观执行的Rollup的问题：
- 可逆性：乐观执行的Rollup在交易执行过程中采用乐观的假设，即假设所有交易都是有效且没有冲突的。但如果存在冲突或无效的交易，整个系统可能需要回滚和重新执行，导致一定的不确定性和处理时间延迟。
- 前置运营（MEV）：乐观执行的Rollup可能面临前置运营问题，即交易执行顺序的滥用，导致操纵交易和不公平的交易优先级。
These issues limit the performance and security of existing Rollup schemes, and may affect user experience and usability. The key to solving these problems is to improve the performance and decentralization of Rollup by introducing new designs such as sorters. The role of the sorter is mainly to improve throughput and compress transaction data. The sorter can sort the incoming transactions to improve the efficiency and throughput of transaction processing. By sorting transactions according to certain rules, conflicts and competition among transactions can be reduced, and the processing capacity of the overall system can be improved. The sorter can also compress transactions, packaging multiple transactions into a single transaction, thereby reducing the size of transaction data. This compression can reduce the cost of on-chain storage and transmission, and increase the efficiency of the overall system.
Problems with centralized sorters
Most of the existing Rollup solutions operate their own centralized sorter, because it is more convenient and lower cost. However, the disadvantages of centralized sorters are also obvious, which may lead to problems such as reviewing transactions, charging excessive fees, prioritizing transactions in transaction execution (front-end operations), or generating bad MEV (maximizing extractable value).
Source: Bing Ventures
We believe that the key to solving the centralized sorter problem is to promote technological innovation and explore decentralized sorter solutions. As mentioned earlier, while current centralized orderers play an important role in improving throughput and performance, we should actively pursue safer, more censorship-resistant, and more decentralized alternatives. These schemes may require trade-offs between security, performance, and degree of decentralization, and are difficult to design and implement, but they promise to solve the problems faced by current centralized orderers.
The road to decentralization
At present, the decentralized sorter technology still needs to be improved, and possible directions include more effective sorting algorithms, more efficient verification mechanisms, and smarter sorter designs. Below, we summarize some of the routes that we consider useful attempts to take. There is no doubt that decentralized orderer technology will continue to improve and evolve over time. This could include higher throughput, faster acknowledgments, lower latency, and improved security and composability.
- Proof of Authority (POA) : This scheme authorizes a group of entities to take turns acting as orderers in a PoA system. It effectively improves censorship resistance and has the lowest latency, but still has the risk of a single point of failure.
- Based on Rollups : This scheme allows anyone to submit L2 batches to the data availability layer (DA), and then the DA layer decides the final block (proposer). It has the advantage of inheriting the activity and censorship resistance of the DA layer, but it may leak revenue and be affected by MEV, and the confirmation speed is slow.
- Decentralized Validator Technology (DVT) : This scheme distributes ordering responsibilities to a cluster, with each node in the cluster signing independent attestations using a fractional share of their validator keys. This scheme is flexible and can be combined with other solutions, but it will add some delay.
- Shared sorter : This scheme allows multiple Rollups to choose to enter a shared sorter, which can simultaneously process transactions on chain A and chain B, and provide strong economic security and real-time censorship resistance for the sorting layer. Shared orderers have multi-chain network effects, but are still limited by L1's data and transaction ordering throughput.
- Building a new set of sorters : This scheme creates a decentralized set of sorters permissionlessly through the use of token incentives. It has the advantage of increasing the utility of the token, but there may be a delay, and the implementation barrier may be difficult for lesser-known Rollups.
Although decentralized sorters have many advantages, the various current solutions have their own trade-offs, and each has its own advantages and disadvantages.
Source: Bing Ventures
potential opportunity
Now we jump out, can decentralized orderers solve Ethereum’s problems once and for all? Has it buried some hidden dangers for Ethereum in some places?
First of all, for a decentralized sorter solution that relies on the underlying blockchain protocol (L1), its performance and scalability are naturally limited by the L1 protocol itself. If the underlying protocol has a bottleneck in processing transactions and consensus, then even if the orderer is highly decentralized and responsive, the performance of the entire system will be limited.
Secondly, the impact of the network environment on the decentralized sorter is also an important consideration. The synchronization and stability of the network will directly affect the activity and security of the sequencer. In an asynchronous network, the orderer will lose liveness, that is, it will not be able to process transactions in a timely manner. Whereas the network is more synchronized, the sequencer can be more reliably kept alive.
Well, this means that with the development of the decentralized sorter, its related infrastructure will become a potential investment opportunity. This includes technology providers providing orderer services, security audit firms, cross-chain solution providers, and governance and engagement platforms, among others. These infrastructures will hopefully improve the issues we mentioned. It should be emphasized that the decentralized sorter is a solution explored by the Ethereum community to improve system performance and scalability, but it is not the only solution. Additional technologies and improvements will emerge in the future to further improve the performance of the Ethereum network:
- Multi-chain interoperability : With the emergence of various blockchains and Layer 2 solutions, multi-chain interoperability will become an important aspect of decentralized orderers. Future sorters may need to be able to process transactions on multiple chains at the same time and achieve atomic composability to provide a smoother user experience and more powerful functions.
- Preventing MEV and improving user protection : Future sequencers may take steps to reduce the impact of MEV and provide better user protection mechanisms. This may include the adoption of random ordering mechanisms, reasonable transaction fee mechanisms, and better privacy protection measures.
- Enhanced Governance and Participation Mechanisms : In order to ensure the fairness and security of decentralized sorters, future sorters may introduce stronger governance and participation mechanisms. This is achieved through voting by token holders, election of validators, and decentralized decision-making by participants. A more open and transparent governance mechanism can promote community participation and drive the development of the system.
All in all, we expect more business model innovations to emerge as decentralized orderers develop. This could include different transaction fee models, orderer-based data services and on-chain applications, etc. Innovative business models will provide more economic incentives for the sequencer, thereby promoting its wide application and sustainable development.