Rollup

On this page 7

A rollup is a blockchain scaling design that executes transactions outside a base chain and then posts the transaction data and resulting state commitments back to that chain, so that the base layer guarantees the data is available and the results can be verified or challenged. Rollups are the main form of layer 2 blockchain on Ethereum and the dominant approach to blockchain scaling in its ecosystem.

The idea is to move expensive work, namely execution, off the base chain while keeping the two things that make a blockchain trustworthy on it: public data and a final arbiter of correctness. Because thousands of transactions share the cost of one batch posted to the base layer, the cost per transaction falls dramatically.

How a rollup processes transactions

Users send transactions to a sequencer, the node that orders them and executes them in the rollup’s own virtual machine. The sequencer gives users a fast soft confirmation, usually within a second or two, then compresses many transactions into a batch and publishes it to the base chain together with a state root, a cryptographic summary of all balances and contract storage after the batch.

A bridge contract on the base chain holds the assets deposited into the rollup and accepts state roots. Its rules decide when a withdrawal is final: either after a challenge period passes without a valid fraud proof, or immediately once a validity proof is verified. Because the full data is public, anyone can rebuild the rollup state and, in a mature design, force a withdrawal even if the sequencer stops.

The components of a rollup and their roles are summarized below.

ComponentRoleWhere it runs
SequencerOrders and executes transactionsOff-chain, often one operator
Batch posterPublishes compressed data to the base chainOff-chain, posts on-chain
Bridge contractHolds deposits and processes withdrawalsBase chain
Proof systemFraud proofs or validity proofsProver off-chain, verifier on-chain
Full nodesRecompute state from public dataAnyone, off-chain
Main rollup components

Optimistic rollups and ZK-rollups

Rollups split into two families by how they prove correctness. An optimistic rollup assumes each state root is valid and lets anyone dispute it during a challenge window of about seven days; a successful fraud proof reverts the invalid result. A ZK-rollup attaches a validity proof, a succinct cryptographic proof that the new state follows correctly from the old one, which the base chain checks before accepting it.

AspectOptimistic rollupZK-rollup
Correctness modelValid unless challengedProven valid before acceptance
Native withdrawal timeAbout seven daysMinutes to hours, after proof
Off-chain costLow; disputes are rareHigh; proof generation is heavy
EVM compatibilityMature and close to exactImproving; varies by design
Security assumptionAt least one honest challengerSound proof system and circuits
Optimistic and ZK-rollups compared

The seven-day delay applies to the canonical bridge only. Liquidity providers offer fast exits by paying users on the base chain at once and waiting out the window themselves, for a fee.

Data availability and rollup costs

The defining requirement of a rollup is data availability on the base layer. Since the Dencun upgrade in 2024, Ethereum rollups post data in blobs, chunks of about 128 kilobytes that are priced in their own fee market and pruned after roughly 18 days, long enough for anyone to download and verify them. Blob space cut the data cost of rollup transactions by a large factor, and the rollup fee now consists mainly of execution plus blob data.

Designs that keep data elsewhere are related but weaker. The table shows how the data location changes the security model.

DesignData locationSecurity consequence
RollupBase chainInherits base layer availability
ValidiumOff-chain committeeFunds can freeze if data is withheld
Rollup on an external data layerSeparate data availability layerDepends on that layer’s security
SidechainIts own chainOwn validators; no base layer guarantees
Where transaction data is stored

Rollups compared with other scaling approaches

A rollup is not the only way to scale, but it offers the strongest inheritance of base layer security for general-purpose smart contracts. A sidechain runs its own consensus and only bridges to the main chain, a state channel suits repeated payments between fixed parties, and blockchain sharding splits the base layer itself.

ApproachBest suited forMain limitation
RollupGeneral applications and DeFiSequencer centralization today
State channelFrequent payments between known partiesParticipants must stay online
SidechainCheap throughput with own rulesSeparate and weaker security
ShardingBase layer capacityProtocol complexity
Scaling approaches at a glance

Ethereum’s roadmap treats these approaches as complementary: rollups provide execution, while upgrades such as danksharding expand the data capacity that rollups consume.

Risks and maturity of rollups

Most rollups still rely on training wheels: a single sequencer, upgradeable contracts controlled by a multisignature group and, in some cases, permissioned proofs. Evaluating a rollup means checking how far it has moved from these controls toward full trust minimization.

  • Sequencer censorship or downtime: a single operator can delay transactions; a forced inclusion path through the base chain is the safeguard.
  • Upgrade keys: a council that can upgrade the bridge instantly can, in principle, change the rules over user funds.
  • Bridge risk: assets on a rollup are only as safe as its bridge contract, which is a high-value target.
  • Fragmentation: liquidity and users spread across many rollups, which complicates moving assets between them.

A common misconception is that every rollup is as secure as Ethereum itself. That is true only when the proof system is live and permissionless, users can exit without the operator, and upgrades come with an exit window long enough for users to leave before new rules take effect.

The simple explanation

A rollup is like a busy branch office that handles customers quickly and sends a bundled report to head office at regular intervals. Head office does not redo every transaction, but it keeps the full report on file so anyone can check it.

If the branch made a mistake or cheated, the report reveals it and head office can overrule the branch. Customers get speed and low cost, while the final word still rests with the main blockchain.

How it is used in practice

A payments startup builds a tipping feature for an online creator platform, where typical tips are between one and five dollars. On Ethereum mainnet, a single stablecoin transfer can cost more than the tip itself during busy periods, so the team deploys its contracts on an optimistic rollup instead.

Tips settle in about a second with fees of a fraction of a cent, and creators withdraw weekly. For withdrawals the startup uses a fast bridge service rather than waiting the seven-day challenge window, and it monitors the rollup’s forced inclusion mechanism as a fallback. The outcome is a micro-payment product that remains viable without giving up settlement on Ethereum.