Crypto
2 min read

Data Availability

The guarantee that the data needed to verify a blockchain’s state has been published and can be retrieved by anyone. A core focus of modular blockchain design and projects like Celestia and EigenDA.

Why data availability matters

Verifying a blockchain's state requires access to the data behind it. If a block is proposed but its data isn't actually published — only a hash or commitment is shared — independent observers can't check whether the included transactions are valid. They have to trust the block producer.

This becomes especially important for Layer 2 rollups, which post compressed transaction data back to the underlying L1. If that data isn't available, users can't reconstruct rollup state to challenge fraud or exit if the rollup operator misbehaves.

DA in rollup design

Rollups split their architecture into several functions:

  • Execution — running transactions and computing new state.
  • Settlement — recording final state to the L1.
  • Consensus — agreeing on transaction ordering.
  • Data availability — making the underlying data accessible.

Original Ethereum rollups put DA on Ethereum itself: rollup data was posted as calldata to L1, with the full security guarantee that L1 data is permanent and accessible. EIP-4844 (Proto-Danksharding) introduced "blobs" — a cheaper data lane on Ethereum specifically designed for rollup DA, dramatically reducing rollup transaction costs.

Modular DA layers

A separate movement creates dedicated DA-only blockchains:

  • Celestia — first major DA-focused L1. Rollups can post data to Celestia at much lower cost than Ethereum, accepting Celestia's smaller validator set as a trade-off.
  • EigenDA — built on EigenLayer restaking. Provides DA secured by re-staked ETH, leveraging Ethereum's security pool without using Ethereum block space.
  • Avail — Polygon-affiliated DA layer using a similar model to Celestia.
  • Near DA — built on the Near Protocol.

Rollups using these alternative DA layers are often called "validiums" or "optimiums" depending on whether they use ZK or optimistic proofs for execution.

The DA trade-off

The choice of DA layer is a security decision:

  • Ethereum DA — strongest security; highest cost.
  • Dedicated DA layers (Celestia, EigenDA) — lower cost; smaller validator set; relatively new with less stress-tested track record.
  • Off-chain DA / data committees — cheapest; weakest security. Data committees are small groups of trusted operators who attest to data availability. If they collude, users can lose access to data.

For high-value applications, Ethereum DA remains the default. For consumer applications and games where transaction costs need to be very low, alternative DA layers offer real savings at the cost of additional security assumptions.

Why this matters now

The "modular blockchain" thesis — that execution, settlement, consensus, and DA should be specialized into separate layers — has become one of the dominant architectural arguments in 2024-2025. The bet is that specialized DA layers can scale beyond what a monolithic chain can, enabling rollups with much higher throughput than Ethereum-DA rollups.

The competing thesis — championed by Solana and Monad — is that monolithic chains can scale through parallel execution and don't need the modular architecture. Both bets have meaningful evidence; which approach wins long-term is an open question.