Ethereum Rollups and Data Availability

2026-07-28

Ethereum Rollups and Data Availability

Rollups move computation off Ethereum, but they must still publish their transaction data somewhere everyone can check. That "somewhere" is the data availability problem, and how a rollup solves it decides whether it is truly secured by Ethereum or only loosely attached to it.

Ethereum Rollups and Data Availability: key points at a glance

Why rollups need data availability

Rollups process transactions off-chain, then post a compressed record back to Ethereum. For anyone to verify that record — or to rebuild the rollup's state if its operator disappears — the underlying data must actually be available to download. If the data is withheld, users cannot prove what they own. Publishing the data is exactly what lets a rollup inherit Ethereum's security instead of asking you to trust its operator.

The old way: expensive calldata

Before 2024, rollups wrote their data into ordinary Ethereum transactions, called calldata. It worked, but it was costly: rollup users were paying for permanent storage they did not need, and those fees passed straight through to every transaction. Data availability, not computation, was the main thing making Layer-2 expensive.

Blobs and EIP-4844

The Dencun upgrade in March 2024 introduced EIP-4844, also called proto-danksharding, adding a new data type known as a blob. Each blob is a roughly 128 KB packet attached to a block but kept separate from normal execution, and it is automatically deleted after about 18 days — long enough for anyone to check, short enough to stay cheap. Because blobs gave rollups a dedicated, temporary data lane, Layer-2 fees fell by over 90 percent almost overnight.

Widening the data lane

Blobs launched with a target of three per block and a cap of six. Later upgrades kept widening that lane: Pectra in 2025 raised the targets, and Fusaka in late 2025 added PeerDAS, which lets nodes confirm data is available by randomly sampling small pieces instead of downloading everything. Follow-up "blob parameter" forks pushed the per-block count higher still, on the road toward the long-term goal of full danksharding.

Off-chain data availability and validiums

Not every rollup posts its data to Ethereum. Some designs, often called validiums, publish data to a separate network — such as Celestia, EigenDA, or Avail — or to their own committee. This is cheaper and higher-capacity, but it trades away part of Ethereum's guarantee: if that outside layer fails or withholds data, your funds can be stuck. Where the data lives is the single biggest security decision a rollup makes.

The bottom line

Data availability is the quiet foundation under every rollup. Posting data to Ethereum through blobs is what makes a rollup genuinely Ethereum-secured, while cheaper off-chain options accept weaker guarantees for lower cost. When you judge a Layer-2, look past the fees and ask where its data goes — that answer tells you who you are really trusting.

Disclaimer: This article is educational content from Bitbase Academy, provided for informational purposes only. It is not investment, trading, tax, or financial advice. Written as of July 2026; rely on the latest official information.

References

[1] Ethereum.org, "Danksharding" ethereum.org

[2] thirdweb, "Ethereum Blob Space Explained" thirdweb.com

Related Articles

More