Understanding the innovations and practical significance of the "data availability layer" proposed by Polygon and Celestia

Written by: Sungjae Han, Blockchain investment firm Genesis Block Ventures

Translated by: Lu Jiangfei

Recently, Polygon announced the launch of a new data availability layer called "Avail," which quickly attracted widespread attention from the crypto community.

The data availability layer technology originated from a project called "Celestia" (formerly known as LazyLedger), but due to its complexity, most people are reluctant to spend too much time understanding it. In this article, I attempt to explain the concept of "data availability infrastructure" in simple terms and why this architecture is necessary.

First, let's explore the issue of data availability. Data availability problems arise when a block producer publishes a valid block header but deliberately conceals the underlying transaction data. This issue occurs mainly because light nodes rely solely on block headers for verification, making them susceptible to deception and accepting invalid blocks.

Full nodes cannot generate data availability proofs for light nodes due to missing block data. Similarly, full nodes cannot generate fraud proofs for invalid blocks, which means light nodes must either verify data availability themselves or assume that most data is honest and trustworthy.

This is why most nodes must download all transaction data and verify the data availability of blocks. However, doing so imposes scalability limitations and issues on network scaling solutions. For instance, as the complexity of verification increases, sharding, rollups, and block sizes will also increase.

So, how do Avail and Celestia address this issue? In fact, they have built a blockchain where on-chain consensus is used solely to verify transaction ordering and data availability, without the need for transaction execution/verification.

To help you better understand this, let's first look at what a conventional blockchain architecture looks like. Generally, a blockchain architecture consists of the following three layers:

1. Execution Layer

2. Consensus Layer

3. Data Availability Layer

For blockchains like Ethereum, which are widely used, these three layers are coupled into a single architecture—the Ethereum Virtual Machine (EVM) is part of the consensus rules and protocol. Projects like Celestia and Avail actually provide a "separate" data availability layer and a consensus layer, allowing different blockchains/transaction execution environments to benefit from them.

Celestia and Avail can be used as minimal, pluggable consensus and data availability layers, validating only data availability and transaction ordering, which provides greater flexibility for modular blockchain network architectures.

Moreover, separating the data availability layer and consensus layer enhances scalability, as they can efficiently perform DA checks using mathematical data availability proofs. (Note: Celestia uses erasure coding, while Avail uses the KZG commitment scheme.)

In this way, the execution consensus bottleneck can be effectively resolved, as the data availability layer only needs to execute data availability proofs without needing to verify the state and validity of blocks through execution.

More importantly, as a "separate" layer, it means that the execution environment is "sovereign independent." For application-specific blockchains, Celestia and Avail are "pluggable," allowing them to benefit from security without needing to bootstrap their own set of validators/consensus.

In other words, these blockchains can not only scale their networks more easily but also gain the security provided by Celestia/Avail (data availability layer and consensus layer).

However, since the data availability layer and consensus layer do not verify transaction validity, some form of "execution" must be performed to handle transaction execution and verification/dispute resolution. Celestia proposes using Optimistic Rollups as a solution while suggesting adding Rollup support to the Cosmos SDK. Essentially, this solution establishes an execution layer above the data availability layer and consensus layer.

Overall, the "data availability layer and consensus layer" environment is very similar to the future scenario that ETH 2.0 aims to construct. For example, Vitalik Buterin wrote in the rollup-centric Ethereum roadmap: "Everyone can handle a single high-security execution shard and a scalable data availability layer."

If Polygon's new tools can build a reliable execution layer and a stable and convenient execution layer on off-chain scaling solutions, then this scalable data availability layer technology seems very promising.

In conclusion, Avail and Celestia provide a new blockchain architecture that not only allows for modular design but also enables application-specific sovereign chains to benefit from shared security.