Ethereum’s Endgame: Modularity

In his Endgame article, Vitalik described what he believed to be the ultimate form of Ethereum, spending a lot of space describing a new Ethereum built with Rollups and DA.

Original article: "Exploring a Breakthrough - Blockchain Modularity" by Codex Labs

Review: ECN

It has been 5 years since Vitalik Buterin proposed The Blockchain Trilemma. Both the ETH-core community and developers of other public chains have been trying to find the key to unlock the next generation of blockchain.

In his Endgame article, Vitalik described what he believed to be the ultimate form of Ethereum, in which he spent a lot of space describing a new Ethereum built with Rollups and DA. This undoubtedly points out, to some extent, the path to breakthrough for Ethereum in the next decade - modularization.

As we know, both Ethereum and new public chains have already started their own modularization journey, but their "answers" are different, and they have formed completely different technical routes.

1. Blockchain Layering

In fact, Ethereum itself has divided the blockchain into multiple layers, but in the 1.0 era, nodes assumed the responsibilities of all layers. We can follow Ethereum's thinking and split the blockchain into four layers:

• Consensus Layer: determines the order, validity, and consistency of transactions between nodes

• Data Availability Layer: Ensure that transaction data can be used (guarantee storage, verifiability and availability)

• Settlement Layer: Settlement of status commitments

• Execution Layer: Computing State Transitions

Returning to the narrative logic of Ethereum being the "world computer", and comparing the modules of the blockchain to the computer structure, then:

• The consensus layer of the blockchain ⇒ maintains the consistency algorithm of the distributed computer cluster

• The execution layer of the blockchain ⇒ the computer's operating system (the environment that actually executes instructions)

• The blockchain’s DA layer ⇒ the computer’s memory (for short-term data storage and access)

• The settlement layer of the blockchain ⇒ the CPU of the computer (the hardware ensures the correctness of instruction execution)

It is worth mentioning that the EigenLayr project has very innovatively further split the consensus layer. He understands the role of the PoS consensus layer as responsible for maintaining the consistency of the distributed system and maintaining the trust foundation of the network through mechanisms such as staking. This has resulted in two sub-layers:

• Trust layer: It is composed of the bottom-level validator nodes, or the hardware entities that are pledged to join the Ethereum network.

• Consistency layer: connects each node to maintain the consistency of this huge distributed system

It is the distinction between these levels (modules) that provides the basis for the modular development of blockchain, resulting in various disassembly schemes for the main chain.

2. Project Mapping

In the layered logic we mentioned above, trust is gradually transferred from the consensus (trust) end to the execution layer, and the core starting point of the separation module is to reduce the burden on the main network without damaging the original trust system as much as possible.

2.1 Secured Rollups

Obviously, the first thing to be disassembled is the execution layer at the end of the trust system. Therefore, various Rollups based on Ethereum came into being. Whether it is optimistic Rollups based on fraud proof or zk-Rollups based on validity proof, the existing solutions are to hand over the execution layer to the off-chain, while settlement, DA and consensus are handed over to Layer1 to continue to undertake. This is of course the most intuitive solution.

We call it Secured Rollups, which can be understood as Rollups protected by Layer 1. The name comes from the description of Scailing in the official Ethereum document:

For some solutions the layer 2 instance then batches them into groups before anchoring them to layer 1, after which they are secured by layer 1 and cannot be altered. The details of how this is done vary significantly between different layer 2 technologies and implementations.

However, "disassembly" often means that the problem of "connection" will surface. Almost all mainstream Rollups now have centralized power over transaction sequencing. Of course, as an important component connecting the execution layer and the main network, the centralization of the Sequencer will certainly help improve system efficiency, but the "dark forest" formed by the high concentration of transaction sequencing power makes people inevitably fall into concerns about MEV attacks (Shutter Network is committed to de-trusting the Sequencer by introducing threshold encryption technology).

2.2 Sovereign Rollups

Rollups on Ethereum currently publish their blocks directly to an EVM-based smart contract, also known as a bridge contract. This contract effectively implements an on-chain light client for Rollups that receives block headers and processes fraud or validity proofs.

In this model, we believe that Ethereum is regarded as a coupled settlement layer of Rollup, where Rollup is the "baby chain" of Ethereum rather than an independent chain with its own rights.

To make an inappropriate analogy, Ethereum Rollups that only separate the execution layer are like "laborers" who only "move bricks" (calculations) without any right to participate in the final game (settlement). L1 still has the final say on how and in what order the bricks are stacked.

Celestia's design obviously gains more power for Rollups. It defines a new way of module segmentation: Sovereign Rollups.

Sovereign Rollups on Celestia do not publish their blocks into smart contracts, but directly to the chain as raw data. The Celestia consensus and data availability layer does not interpret or perform any computations on Rollup blocks, nor does it run on-chain light clients for Rollups.

2.3 Danksharding

Whether it is Danksharding or the original data sharding solution, they are essentially hoping to solve the problem of the DA layer. In the context of modular blockchain, Danksharding obviously improves the Ethereum mainnet's ability to bear the data availability layer.

Danksharding uses Data Availability Sampling (DAS) technology to implement "decentralized validators", which is very similar to Celestia's solution.

Although we know that the ultimate form of the Ethereum consensus layer will still bear the responsibilities of settlement and data availability, the decentralization of verification work achieved by DAS technology greatly reduces the cost of participating in network verification, thereby further unleashing the performance of the main network.

After the Danksharding fork, the Ethereum mainnet assumes part of the power to review data availability, and Rollups, in addition to the original responsibilities of the execution layer, also assumes the power of the DA layer.

2.4 EigenLayr

We mentioned above that EigenLayr is a further subdivision of the consensus layer. It separates the trust layer of Ethereum from the entire network and attempts to build an independent layer on the trust layer - EigenLayer. This layer relies on the Ethereum trust layer network (or Ethereum giant

Large pledged capital) to provide infrastructure/middleware service providers with a platform on which they can freely build consensus and applications.

Among them, how to connect the Ethereum trust layer and EigenLayer is the key to the entire design. EigenLayr cleverly proposed a new concept - Restaking. Staking nodes on Ethereum can voluntarily choose to join EigenLayr. On the other hand, middleware/infrastructure service providers who have more requirements on the performance level can also choose to build their own networks on EigenLayr.

EigenLayr's separation plan is equivalent to "borrowing" the credit (consensus) foundation of Ethereum. On this basis, we can develop endless associations, whether it is to build a new sovereign chain or to build an infrastructure network. Behind these possibilities is the huge credit foundation of the Ethereum main network.

2.5 Arweave

We can see that in the foreseeable future of Ethereum's modularization, the main network must still maintain control over the consensus layer, or in other words, the consensus layer, as the soul of Ethereum, is the "holy grail" that brings together participants and value.

As a modular blockchain, Arweave is clearly unique in that it breaks away from the narrative logic of Ethereum and builds a design paradigm based on storage consensus.

In Arweave's design, the main network acts as the paper tape of the entire Turing machine, recording these states and the transactions that modify each state at the bottom layer, thus achieving the separation of storage and computing resources.

It can be said that Arweave redefines our modular segmentation. It uses the DA layer to ensure consensus, while execution and settlement are left to the outside of the chain.

Of course, the price is: in order to obtain the validity guarantee of the latest status data, users need to download all historical status data and perform a verification calculation before executing off-chain calculations locally, which obviously brings efficiency problems (KYVE Network is committed to solving the inefficiency problem caused by this part of the work).

3. Summary

Looking at the exploration of these blockchain modularization paths, although there are many different paths, there are still traces to follow. At its core, the driving force behind different route choices often comes from different understandings of market demand.

The bull market triggered by DeFi has created an urgent need for Ethereum scalability and the composability of different chains. Based on these needs, many solutions to alleviate the pressure on the main network have come on the scene. If the narrative logic of future blockchain networks is still built on financial primitives, then perhaps ETH2.0 with Rollups as the core will be the visible end of the road to modularization.

What if we jump out of the constraints of the centralized financial context and try to imagine the broader applications of blockchain? Large-scale games, real-world asset transactions, decentralized social networks, etc. Then we may need a new modular solution to fully unleash the performance of each module.

On the other hand, for the development of the public chain itself, consensus is the core value of the public chain. As we mentioned above, no matter how Ethereum develops, it will not abandon its consensus layer in exchange for performance.

The road to blockchain modularization is not just about blindly expanding performance. More importantly, it is necessary to return to the properties of the blockchain itself, obtain value from consensus, and then feedback consensus through value, thereby building a stronger consensus.