Metis VS Espresso: Which decentralized sequencer represents the future?

Source: View on the chain

On one side is @MetisL2  Decentralized Sequencer construction is carried out against all difficulties, and on the other side is @EspressoSys  Based on modularization, the shared Sequencer solution is killing everyone. Some people may I would like to ask, in order to decentralize the sorter, what is the difference between the "self-operated store" that Metis is trying and the "outsourcing module" of Espresso?

In my opinion, the decentralization of the Sequencer sorter is only a prerequisite for the formation of layer2 basic infra. The real challenge lies in the construction of a decentralized ecosystem. Come on, let me talk about my understanding.

Recently, Metis is promoting the decentralization process of Sequencer, and community testing is also in full swing. At one time, the TPS of the Metis main network was raised to Ethereum Top3.

This is actually just a stress test of the decentralized Sequencer technology layer to ensure that it can operate stably after its official launch, such as multi-Sequencer rotation mechanism, single Sequencer evil punishment design, Sequencer MEV problems, Sequencer miner ecology Participation issues and more.

1) POS block production rights: Metis has introduced a Sequencer pool. Sequencer nodes need to Stake 20,000 Metis blocks in turn. The higher participation threshold ensures the node’s operational capabilities and special communication conditions. It provides instant repair capabilities and also deters the possibility of nodes doing evil at a single point.

2) Community stress test: After launching the Sepolia test network, Metis launched a community stress test in January. At present, the test network has conducted more than 600 sequencing rotations and over 3 million transactions to test The stability of the block generation of the Sequencer rotation mechanism pave the way for the official launch of the decentralized Sequencer mainnet.

3) MEV processing: If the centralized Sequencer can perform MEV clipping operations through arbitrary sorting, but in a decentralized environment, an independent node is needed to monitor the sorting and block-producing behavior of each Sequencer. Metis There are dedicated L2 Ranger nodes on the blockchain to perform random checks on the blocks generated by the Sequencer to prevent the transaction order from being reversed.

However, purely relying on random inspections to supervise and punish is somewhat passive. In the later stages of the MEV confrontation, the Ranger node can become a special MEV pipeline similar to MEV-Boost, using MEV to control the "out of control" of MEV?

4) LSD pledge mining ecological implementation: @ENKIProtocol   The LSD protocol is about to be launched in the Metis ecosystem. Its goal is to solve the problem of circulation and application of Reward after mining by Sequencer miners in the entire decentralized economy. Obviously, as the main source of output of $METIS, whether miners are willing to pledge their tokens for lending and mining is very important to the subsequent development of the entire $METIS ecosystem.

In short, in my opinion, the technical issues of decentralized Sequencer are not complicated. What is rare is a set of robust decentralized ecological construction, support, benefit distribution and incentive models behind it.

It should be said that technical issues are just the tip of the iceberg of Sequencer decentralization. A series of comprehensive services and experiences centered on market, ecology, operations, users, etc. are the core of the decentralized Sequencer project. The highlight.

The above describes the challenges existing in the advancement of Metis decentralized Sequencer. Based on this, let’s sort out Espreeso’s shared Sequencer solution. It will be easy to understand.

To put it simply: @EspressoSys   provides a modular Sequencer shared component for Rollups, which is equivalent to interspersing a Sequencer Mempool preprocessing process.

Originally, users calling RPC to request to send transactions will be directly submitted to the Sequencer processing center. Now they can share the Sequencer center through Espresso, and Espresso will process these transaction sorting requests in batches, and then return the status to layer2 after batching. .

Espresso has a Hotshot equity proof system, which is equivalent to Metis’ POS staking system, to ensure the decentralization of Sequencer and the design of corresponding reward and punishment mechanisms; moreover, Espresso will most likely cooperate with Eigenlayer and outsource Ethereum’s Validators provide secure consensus. In addition, its Espresso DA function is also a highlight.

What’s interesting is that if a user submits a cross-chain request, it must go through Espresso’s shared Sequencer. In theory, Espresso can realize transaction status conversion in the Mempool environment, thereby reducing cross-chain submission status and main issues. Transaction losses in network interactions. (There will be a lot of room for lightweight imagination)

In short, Espresso and Celestia play in a similar way, abstracting the decentralized Sequencer requirements that layer 2 needs into a shared sequencer "commodity" to modularize The highly adaptable approach continues to penetrate into the layer2 ecosystem.

This is a clever approach based on modular thinking. Due to its advantages such as low cost, speed and efficiency, it is bound to attract a large number of layer2s who demand "one-click chaining".

However, in my opinion, modularization is a double-edged sword. Moderate modularization will significantly reduce development costs, but if it is excessive, it will cause subsequent ecological development to lack source vitality:

1) The lightweight attributes of modular components can help Layer2 developers quickly start chains at low cost. When DA capabilities and Sequencers become outsourced commodities, the number and replacement speed of Layer2 chains will also accelerate.

I have written before that when the number of layer2 projects emerges in batches, comprehensive soft power such as brand, market, and ecology will become the core of layer2. At that time, the key will be whether a layer 2 will be provided with DA by Ethereum, whether there will be an independent and reliable decentralized Sequencer, and whether there will be a mature, stable and growth-potential ecosystem. The modularization of components will accelerate the hierarchical differentiation of the layer2 market, and the head layer2 and the middle and tail layer2 will take completely different paths; (OP and ZK can use modularization to drive the implementation of their Stack strategy)

2) After the DA capability is outsourced, if the Sequencer is outsourced, the Prover system will also be outsourced in the future. The only trump cards that layer 2 can compete with are marketing and operation and maintenance.

There is a paradox here. It is easy to understand that DA has the ability to give up Ethereum and choose low-cost Celestia. If layer2 hands over the core Sequencer tax collection rights to outsourcing, how should the handling fees be collected, and what ecological projects should How to motivate and support will pose a series of subsequent operation and maintenance problems.

This kind of layer 2 is very much like an "empty shelf". It is glamorous on the surface, but in fact it lacks an independent decentralized economic system framework, and its market ecology is difficult to develop. For example, you can use molds to quickly assemble a set of mobile homes. But if you want to build a 100-meter building, give up the steel frame structure and try using molds again?

Of course, I do not deny the potential "variable" space that modular paradigms such as Celestia and Espresso bring to the entire layer1 and layer2 ecology. They lower the technical threshold and can expand the market ecosystem. When it gets big and there are many chains, innovation entities will be plentiful and the market will produce some high-quality projects. There must be a difference in value and meaning between short-term and long-term.

In the final analysis, no matter how fast layer2 launches the chain, the ones that survive must be the super Mass Adoption economy that can continue to bring effective expansion of users, projects, funds, etc. to the main network.