Written by: Haotian
A friend asked me to analyze the newcomers besides the four major layer2 kings, such as Linea, Polygon zkEVM, Mantle, Metis, etc. After a rough look, I found that the core differences of most layer2s are in two points: 1) Does Data Availability rely on the Ethereum mainnet? 2) What is the affinity of EVM?
Whether EVM is equivalent is easy to understand, which determines the low threshold for developers to transfer to the ecosystem; this article discusses from a popular science perspective, what exactly is Data Availability? Why is the DA layer the core layer of layer2? What trade-offs have third-party DAs such as Celestia and Eigenlayer made?
Let me first explain in simple terms what Data Availability is?
Simply put, Data Availability means that all nodes in the blockchain network can obtain the availability of all historical data generated by the system. Why is this important? Because the blockchain is distributed and recorded on different nodes. If some key data is lost or hidden, the security of the entire system will be destroyed.
For example, if only some nodes know the details of a transaction and other nodes cannot obtain it, then the transaction cannot be correctly verified by all validators. For another example, to query the balance of an account, all related historical transactions must be traversed to calculate the final result. If some transaction data is missing in the middle, the balance result will be wrong.
Therefore, Data Availability ensures that any node can obtain all historical data, so that it can perform correct verification and query, ensuring that data in the blockchain system will not be lost or hidden, and all nodes can obtain complete data. If the DA layer cannot work properly, for example, Optimism, its 7-day challenge period will not be able to proceed, because if the Calldata before and after the batch cannot be read normally, the Rollup contract cannot be confirmed. In the final analysis, DA is to ensure the security of the blockchain ledger system.
Let’s talk about the pros and cons of using Ethereum as the DA layer in layer2
Using Ethereum as the DA layer of Layer 2 can indeed take advantage of Ethereum's security and data integrity, which is the biggest advantage. Otherwise, there is no restriction on malicious tampering and disordered accounting of layer 2 transactions, and such a layer 2 network is naturally difficult to be recognized by the mainstream. However, Ethereum's own block capacity and transaction throughput are limited, and it may be difficult to support a high TPS Layer 2 network in the long run.
This serious mismatch in capacity means that a large number of Layer 2 transactions are waiting in line for block confirmation, and DA verification cannot be completed in time. It also increases the competition for Ethereum block space in Layer 2, pushing up gas fees.
Therefore, relying solely on the Ethereum mainnet as the DA layer will definitely become a bottleneck for the development of layer2 in the long run. Therefore, some mainstream DA solutions have emerged in the process of layer2 development:
1) Compress state data to L1. For example, zkSync submits zkSNARK proof to the main network, and a large amount of original transaction data is maintained by L2.
2) A third-party public chain platform such as Celestia provides a DA layer solution, and the data is submitted to the third-party distributed network for verification and the result status is synchronized to L1;
3) Cancun was upgraded to add additional Blob storage space, which was encoded and verified using the KZB tool.
Third-party DA modular solution trade-offs
When it comes to third-party DA modular solutions, the most well-known projects are Celestai and Eigenlayer (as well as Ethstorage).
The Celestia network itself can handle a large transaction volume, ensuring that the large amount of data generated by applications on layer 2 can guarantee the validity of DA. Only the transaction status is uploaded to Ethereum in batches, so it is a highly modular DA solution.
Eigenlayer separates transaction processing and data storage. Transactions are processed quickly in rollup, but integrity data is stored in the distributed network built by Eigenlayer. Rollup only uploads key data such as block headers and Merkle tree roots to Ethereum. When it is necessary to read complete data, the Rollup contract will read the data of the Eigenlayer network in real time.
Currently, both modular DA solutions are quite popular. For example, application chains in the Cosmos ecosystem such as Eclipse and Cevmos use Celestia as the DA layer, while layer2 networks such as Mantle use Eigenlayer as the DA layer.
It should be said that the modular DA solution is a DA solution that is more oriented towards the future. After all, it embeds a third-party external consensus and is not as strong as Ethereum's DA consensus. That is why some people have proposed to expel projects that adopt third-party DA solutions from the layer2 echelon.
Essentially, it depends on whether you trust this externally embedded DA solution that can enhance capacity and performance. In my opinion, in the short term, Ethereum DA is definitely more popular, but in the long term, especially in the era of multi-chain applications, this type of third-party modular DA solution may also come out with strong consensus.
After briefly discussing the impact of Data Availability on layer2, the ideas and frameworks of the various layer2s that have emerged are very clear. The core of the OP and ZK Rollup methods is the difference in the way of verifying the accuracy of data status, and whether to use a third-party modular DA is to solve the capacity limitation of Ethereum as a DA. The additional expansion of Blob after the Cancun upgrade is a compromise solution.
Based on this premise, when we look at various cutting-edge layer2 solutions, we will have a basic cognitive framework for judging them.
