Skip to content

Analysis · Rollups

“Secured by Ethereum”: what a rollup inherits and what it does not

Data availability and settlement on L1 are genuinely inherited. Liveness, ordering and the safety of the upgrade keys are not, and those are the properties that decide what happens on a bad day.

By Etherelysium Editorial TeamUpdated 18 November 20263 min read

“Secured by Ethereum”: what a rollup inherits and what it does not
Share X LinkedIn

Almost every L2 describes itself as secured by Ethereum. The phrase is not false, but it folds several different properties into one, and the ones it leaves out tend to matter most when something goes wrong. Separating them is the fastest way to compare rollups honestly.

What is genuinely inherited

Data availability, if the data is on L1

A rollup that publishes its transaction data to Ethereum lets anyone rebuild its state without the operator’s help. That is the property that separates a rollup from a sidechain with its own validator set. Since EIP-4844 the data usually travels in blobs, which Ethereum nodes are required to serve for about 18 days. The guarantee is that the data was published and could be downloaded; long-term storage is someone else’s job. Our data availability entry explains why this is the foundation.

Settlement rules

The rollup’s bridge contract on Ethereum decides which state roots are accepted and which withdrawals are paid. An operator cannot simply mint itself funds on L1, because the contract will reject a state that does not pass its checks: a fraud-proof game for optimistic rollups, a validity proof for ZK rollups.

What is not inherited

Liveness

Ethereum keeps producing blocks whether or not a given rollup does. Most rollups run a single sequencer; if it halts, the L2 halts. That is an operational fact, not a theoretical one.

Censorship resistance and ordering

A sequencer can decline to include a transaction and decides ordering within a batch. Ethereum’s documentation on optimistic rollups describes the mitigation: users can submit a transaction directly to the rollup’s L1 contract once it has been delayed too long. The ZK rollup page describes a direct-submission route for forcing an exit. Whether such a path exists, how long its delay is, and whether anyone has used it in production are the questions that matter.

The upgrade keys

The largest real risk on many rollups is neither the proof system nor the sequencer but the authority that can replace the bridge or verifier contracts. If a multisig can swap the code that checks proofs, the proof is only as strong as that multisig. A delay between announcing and executing an upgrade does not prevent a bad upgrade; it gives users time to leave before it lands, which turns a fatal risk into a survivable one.

Optimistic and ZK, briefly

Optimistic rollups assume a posted state is valid and allow a challenge window, roughly seven days in Ethereum’s general description. Their safety rests, in the words of the documentation, on “the existence of one honest node” able to dispute an invalid claim. ZK rollups prove each state transition, so withdrawals can complete once the proof verifies, but proving is costly and the documentation warns that hardware requirements “may reduce the number of participants that can force the chain to make progress”. Neither design is better in general; both sit downstream of the upgrade-key question.

The Settlement Test

  • What settles: transaction data and state commitments on Ethereum. An L2 transaction is economically final for L1 purposes once its batch is in a finalized Ethereum block and the rollup’s proof or challenge condition has been met.
  • What it costs: the L2 execution fee plus a share of the L1 data fee; for exits, either the challenge-window wait or a liquidity provider’s fee to skip it. Current readings are on our gas and fees page.
  • What breaks: a halted sequencer stops the chain until forced inclusion is used; a compromised or careless upgrade authority can change the rules for everyone at once.

Questions to ask, in order

Is the data posted to Ethereum or to an outside committee? Who runs the sequencer and what happens when it stops? Is there a forced-inclusion or forced-exit path, with what delay? Who can upgrade the contracts, at what threshold, and with how much notice? Only then: which proof system, and is it live or still on a roadmap? Our guides to L2 security models and sequencer centralisation go deeper, and the token pages for Arbitrum, Optimism and ZKsync show how each project describes itself.

Primary sources

Educational content, human-edited and possibly AI-assisted. It describes how systems work and is not financial, legal or tax advice. Report an error via our corrections page.

More in Rollups

Related articles

See all