Modular Blockchain - Ethereum as an Engineered Solution for the "World Computer"

Author: IOBC Capital

The Modularization Trend of Blockchain

Looking at the development trend of Crypto in 2022, it seems a bit far-fetched to create a new L1 public chain at this time; the narrative of modular blockchain cannot be ignored.

After The Merge, Ethereum's development path increasingly leans towards the direction of modular blockchain. The main difference between modular blockchain and monolithic blockchain lies in the fact that monolithic blockchain implements four functions—execution, settlement, consensus, and data availability—on a single base consensus layer, while modular blockchain divides these functions among multiple modules. In fact, Ethereum is not the only one planning a modular architecture: Celestia, which first proposed the idea of modular blockchain, is building a data availability layer for Rollups based on the Cosmos ecosystem; Tezos is also embracing a Rollup-centric roadmap; NEAR is designing data availability sharding. This article mainly discusses Ethereum's modularization trend.

The congestion status of Ethereum has already reflected the drawbacks of monolithic blockchain—poor scalability, lack of customization, and high costs. The problem with monolithic blockchain is that it must perform many different tasks at the consensus layer, and optimizing for only one of these functions cannot effectively improve the blockchain's performance. Figuratively speaking, a monolithic blockchain is like a barrel made of four wooden planks; its capacity (performance) depends on the shortest plank. As long as there is one attribute that is a shortcoming, the overall performance will be limited, and the "blockchain trilemma" restricts the possibility of achieving extreme development in all attributes simultaneously. Therefore, simply expanding based on the monolithic blockchain idea cannot solve Ethereum's predicament.

Modular hybrid scaling: layer1 (data sharding) + layer2 (rollups)

In fact, modular blockchain is essentially a hybrid scaling solution. At the sixth Global Blockchain Summit, Vitalik's speech was titled "The Rise of Ethereum's Layer 2 Protocol Ecosystem." He stated that the Ethereum ecosystem does not solely rely on Layer 1 scaling or Layer 2 scaling, but adopts a hybrid scaling approach. The essence of modular blockchain is similar to hybrid scaling of layer1 and layer2.

Ethereum's Modular Architecture

Ethereum's modular architecture design is mainly divided into four layers: execution layer, settlement layer, consensus layer, and data availability layer. In many cases, the industry also refers to the execution layer and settlement layer collectively as the execution layer, and the consensus layer and data availability layer collectively as the consensus layer.

Execution Layer: Responsible for processing on-chain transactions, executing on-chain orders, and verifying the execution of transfers and smart contracts, primarily focusing on Rollups. After the modular blockchain reaches a certain stage, users typically interact with the blockchain based on the execution layer, including signing transactions, deploying smart contracts, and transferring assets. The execution layer addresses the scalability of the blockchain.
Settlement Layer: The settlement layer is used to verify the execution results of the execution layer, such as Rollups, resolve disputes, and settle state commitments.

Consensus Layer: The consensus layer reaches consensus on the validity of state transitions by downloading and executing the contents of blocks through a full node network, thus providing ordering and finality, and validating block production through a PoS mechanism.

Data Availability Layer: Ensures that transaction data can be used (ensuring storage, verifiability, and availability). It is necessary to publish and store the data required to verify the validity of state transitions in this layer. In the event of malicious block proposers withholding transaction data, the data from the data availability layer can be used for verification.

In the foreseeable short to medium term after The Merge, Ethereum's settlement layer, consensus layer, and data availability layer are unified. The future Danksharding will transform Ethereum L1's data sharding into a data availability engine, with the beacon chain serving as the consensus layer, and the original Ethereum mainnet becoming an execution layer, with more execution layers being L2 Rollups. Additionally, the industry has already begun exploring customized L3 as an extension of the execution layer based on the current L2.

If we say that Ethereum is still just a theoretical "world computer," then modular blockchain is the engineering solution for Ethereum to become a "world computer."

Ethereum's Future Plans

As is well known, The Merge is about transitioning from PoW to PoS, merging the Beacon Chain with the original Ethereum main chain. Beyond The Merge, Ethereum is also simultaneously advancing The Surge, The Verge, The Purge, and The Splurge. The order of these upgrades is still uncertain, as they are independent and proceed in parallel.

The Surge is about introducing sharding, which will allow the Ethereum network to achieve large-scale scaling through sharding.

The Verge is about Verkle Trees, which help optimize storage on Ethereum and reduce node size. This upgrade will focus on optimizing storage through Verkle Trees, a mathematical proof that is an upgrade of Merkle Trees. By reducing the amount of data that validators need to store on their computers to operate, node sizes will shrink, allowing more users to become validators. This will further decentralize the network and enhance security.

The Purge will reduce the hard disk space required by validators by eliminating historical data and technical debt. This means simplified storage, thereby reducing network congestion.

The Splurge consists of a series of tweaks to the Ethereum network, including various smaller upgrades to make the Ethereum network smoother.
Vitalik stated that after completing these five key phases, Ethereum could achieve 100,000 TPS, truly becoming the "world computer" he initially envisioned.

Although the naming of these five parallel key phases is catchy, it may still be quite difficult to understand Ethereum's specific plans for the next three to four years. If we extract some of the more critical and specific upgrade events, it may provide a clearer view of Ethereum's modularization trend:

1. Proto-danksharding (EIP-4844)

Proto-danksharding is a proposal to implement most of the logic and foundational rules that make up the complete Danksharding specification (e.g., transaction format, verification rules, etc.), but this stage has not yet implemented any sharding. In the Proto-danksharding phase, all validators and users still need to directly verify the availability of complete data.

The main feature introduced by Proto-danksharding is a new transaction type called "blob-carrying transactions." Blob-carrying transactions are similar to regular transactions, with the main difference being that they also carry additional data called a blob. The blob is approximately 128kb, which is much cheaper than similarly sized calldata. However, EVM execution cannot access blob data; the EVM can only view the commitment to the blob.

Currently, the block size of Ethereum is determined by gas capacity. After the implementation of EIP-4844, the number of blobs will become another dimension determining block size. A blob is a binary data structure, approximately 128kb in size, and Ethereum blocks limit the number of blobs that can fit in each block, targeting 8 blobs, with a maximum of 16, thus each block will add an additional 1-2MB (1288-12816) of storage space.

Blobs are mainly used to store Layer 2 data, whereas previously Layer 2 data storage was achieved through calldata. With the introduction of blobs, the space available for storage within blocks will significantly increase. However, due to the large size of blob data, if each block adds an extra 1MB of blob data, the Ethereum blockchain will accumulate several TB of data in a month. To address the rapid increase in data volume, these blob data will be stored offline and automatically deleted after 30 days.

Since blob data does not compete with the gas usage of existing Ethereum transactions, it can still achieve significant scaling effects. To simply understand the Proto-Danksharding proposal of EIP-4844, it can be seen as—Ethereum layer 1 maintains a block size of 1MB while storing Layer 2 data in the form of blobs through a short-term, offline storage method for 30 days to achieve scaling effects.

2. Danksharding

Danksharding is a new sharding design proposed for Ethereum. The previously planned sharding was state sharding, but later it was decided to implement a Rollup-centric roadmap, adopting a modular hybrid scaling solution of layer1 (data sharding) + layer2 (rollups), which implements data sharding. Data sharding essentially embodies the idea of modular blockchain, dividing Ethereum into multiple data shards, each connected to one or more Rollups, with Rollups serving as the execution layer, and Ethereum as the consensus layer and data availability layer.

The core mechanisms introduced by Danksharding mainly include: PBS and DAS.

PBS (Proposer Builder Separation) refers to the separation of block proposers and block builders when constructing blocks. The proposer proposes the block, and the builder bids for the transaction ordering rights and computes the block header. The proposer packages the transactions based on the builder's computation results and writes the block header into the block to complete block production. Before the introduction of the PBS mechanism, block proposers (miners before The Merge, validators after The Merge) could maximize their mining profits by viewing which transactions were in the mempool and adopting some strategies to gain MEV opportunities. The introduction of the PBS mechanism, with its role separation and the bidding mechanism for builder ordering rights, can to some extent address the MEV issue, with the final MEV profits being shared among all validators in the network. Additionally, PBS helps solve the synchronization issues between sharding and the beacon chain, as well as the Ethereum network's censorship resistance issues.

DAS (Data Availability Sampling) is an effective method to address blockchain state explosion. It allows validating nodes to check block availability. By using DAS checks, light clients can verify whether a block has been published by only downloading some randomly selected blocks. Since DAS can perform parallel verification of block data, even if the number of data shards (Data Sharding) is large in the future, it will not increase the burden on individual validating nodes; rather, it will stimulate more validating nodes to join, thus ensuring sufficient decentralization of validating nodes.

Ultimately, Danksharding can achieve centralized block production for Ethereum through PBS, decentralized verification through DAS, and possess a certain degree of censorship resistance, thereby ensuring that Ethereum becomes a scalable consensus layer and data availability layer, capable of accommodating more Rollups in the execution layer. (PS: Centralized block production and decentralized verification are also concepts proposed by Vitalik in Endgame regarding the future development of Ethereum.)

Conclusion

I have always felt that the founding team of Ethereum is very passionate, and many details make me believe they will remain true to their original intentions and forge ahead.

Among the various upgrades of Ethereum, three upgrades have left a deep impression on me: the Byzantine hard fork at block height 4,370,000, the Constantinople hard fork at block height 7,280,000, and the Istanbul network upgrade at block height 9,069,000.

Interestingly, Byzantine, Constantinople, and Istanbul are all names of the same city. This city spans Europe and Asia, bordered to the north by the Golden Horn, to the south by the Sea of Marmara, to the east by the Aegean coast of the Anatolian Peninsula, and only connected to land on the west. Napoleon once eloquently stated about this city: "If the world were a country, its capital would surely be Istanbul." This ancient city has formed a subtle connection with the blockchain world because of Ethereum, and the naming of these three upgrades conveys a message—Ethereum remains consistent.

Perhaps Ethereum's path towards a modular blockchain will not be so fast, but it is certain that whether it is the grand themes of The Merge, The Surge, The Verge, The Purge, The Splurge aimed at achieving 100,000 TPS in five phases, or the specific key upgrades Proto-danksharding and Danksharding, the ultimate goal is to promote Ethereum to realize its original intention of becoming a "world computer."