Why is the path to decentralization of the L2 sorter harder than reaching the blue sky?

What Makes Achieving Decentralization of the L2 Sorter More Challenging Than Reaching the Blue Sky?

Author: Haotian

Previously, many articles have discussed the stagnation of decentralized Sequencer solutions such as OP-Rollup and ZK-Rollup. Does this mean that decentralized Sequencers are completely unsolvable?

With this question in mind, I researched @MetisDAO, which claims to have developed the first decentralized Sequencer in a layer2 project. I found that the decentralized technology logic of Sequencers is not difficult, but what is challenging is “power decentralization”.

The use of both “hard distributed” consensus and “soft distributed” social consensus in Sequencer solutions seems to make sense, but they have completely different End Games! Why?

• Navigating the Cryptocurrency Rollercoaster: Strategy as an Art Form
• Exclusive Interview with Injective Asia Growth Lead Why has it become the dark horse of L1 in 2023?
• Inscription Heat’ brings pressure testing, can the public chains other than Bitcoin handle such a surge of wealth and prosperity?

On the Ethereum mainnet, users submit transactions to the Mempool, which are then collected and broadcasted by Searchers and eventually sorted and packaged into blocks by miner Builders based on GAS fees. This means that the transaction you submit may be affected by MEV, and you need to follow the Nounce queuing mechanism. The Gas you pay is the fee required for EVM execution, resulting in transactions being expensive, slow, and easily frontrun on the mainnet.

By analogy, the Layer2 Sequencer is like a closed Mempool. Users submit transactions to the Sequencer, which is responsible for sorting them and batching them into blocks to be submitted to the mainnet. This means that in Layer2, the transaction ordering is completely decided by the Sequencer, and there is no Nounce queuing. The Gas you pay is the shared cost of the batch of transactions.

In theory, this means that Layer2 transactions are fast and cheap. After receiving transactions, the Sequencer will sort them fairly based on the transaction Nonce, reception time, Gas price, or random algorithm. However, due to the excessive power held by the Sequencer, there may be concerns about MEV and the insertion of malicious transactions.

Currently, the mainstream solution in the Rollup market is a solution where the Sequencer is initially centralized but has an “optimistic” approach.

On one hand, because the Sequencer is a core component of Layer2, if a distributed Sequencer is used from the beginning, the potential risks of failure could affect user experience. Therefore, most projects tend to adopt centralized operation in the early stages to ensure security and stability.

On the other hand, when the project reaches a certain scale and has high transaction demand, the Sequencer gains supreme power. It not only controls the pricing of Gas, but also directly profits from it. By this time, the market has grown to a certain size and is less willing to further promote decentralization and relinquish power. This is the reason why the decentralization process of Layer2’s four Sequencer kings is slow.

However, they all independently chose a strategic architecture based on Stack. This is a strategy that opens up the core technology code and attempts to share key technology components.

By sharing the core Sequencer, they create a Superchain multi-chain composition architecture, and then they use methods such as MPC multisignature and governance voting to achieve a socialized “transparent and decentralized” consensus. It’s like the Monkey King appointing several managers, but when it comes to major decisions, we still have to listen to the Monkey King’s opinions, right?

This type of soft distributed Sequencer solution is based on a broad perspective and is suitable for markets that have grown to a certain scale, where everyone has absolute trust in the center of power and the premises that certain trust frictions (malicious behavior, challenges) have been desensitized, such as OP Stack.

So, what is the decentralized Sequencer solution provided by @MetisDAO?

Simply put, it is a “hard distributed” solution for building distributed Sequencer nodes.

Multiple Sequencer nodes on layer2 form a Sequencer pool, and only nodes that have staked 20,000 Metis tokens have the right to produce blocks. Users can also stake their tokens to choose a Sequencer node, enabling them to share the rewards while the Sequencer mines blocks.

In addition, to prevent malicious behavior by Sequencers, Metis introduces L2 Rangers, which are responsible for random sampling of blocks and verifying them against their original state roots, including checking if the order of transactions has been swapped or if any malicious transactions have been inserted, among other things. Performing this verification also earns mining rewards. If a malicious Sequencer is discovered, the system will slash the assets staked by the malicious node, and the verification node will receive a portion of the forfeited assets.

To incentivize Sequencers, they join the POS equity queue, and Verifiers strengthen their sampling and verification work. Through this incentive and punishment model, a decentralized Sequencer system is constructed.

This “hard distributed” architecture has multiple entities operating Sequencers, and there is a transparent reward and penalty mechanism, which serves as a means to constrain the individual rights of Sequencers and ensure fair sequencing.

That’s it!

Both soft distributed and hard distributed approaches are essentially strategic methods.

The Stack open-source framework is more suitable for the expansion of monolithic entities, relying more on intangible trust assets such as market position and brand reputation. On the other hand, the decentralized POS Sequencer is relatively more suitable for modular small entities to surpass the competition as the most powerful consensus in blockchain technology is the consensus model of node mining.

In the context of the layer2 market of Rollup As A Service, the Stack framework can lower the development cost of layer2 for project teams, but the “trust” issues it accumulates need to be borne collectively by all participants.

In contrast, the directly hard distributed technical consensus focuses on Sequencer problems with clearer boundaries. It is more suitable for a situation where the overall layer2 technical progress is slow, and there is negative public opinion. It can provide a more eye-catching starting point and growth prospects.

Of course, I have also looked at other decentralized Sequencer technical consensus solutions such as @EspressoSys, @AstriaOrg, @radius_xyz, etc.

Their logic is similar to Metis’, where the single Sequencer is designed as multiple entities. However, when it comes to off-chain consensus in this part, Esprsso utilizes Eigenlayer that extends the validation capabilities of Ethereum Validators based on Restaking.

Astria, on the other hand, adopts a more modular approach, providing a fast modular Sequencer solution for layer2 developers;

Radius, meanwhile, uses encryption to prevent misconduct and MEV risks in all transactions entering the Sequencer pool. (Opportunity to analyze further)

In conclusion, in my opinion, the decentralization issue of Sequencer will not stop at the Stack strategy of the four Layer2 titans. Although there is a risk of consensus overload in the decentralized solution of technical consensus, it has more prospects compared to the social consensus governance of MPC multi-signature.

We will continue to update Blocking; if you have any questions or suggestions, please contact us!