Layer 2 Networks: Reference Table

Entry last verified September 2026. Trust assumptions change as networks decentralise; verify against the network’s own documentation before relying on this.

The designs

Design	How validity is established	Standard withdrawal	Data posted to L1
Optimistic rollup	Assumed valid, challengeable	About 7 days	Transaction data
Zero-knowledge rollup	Proven cryptographically	Hours	Proofs plus data
Validium	Proven cryptographically	Hours	Proofs only, data held off-chain
Sidechain	Own consensus	Varies	Nothing

The last two rows are frequently marketed alongside rollups and are different in a way that matters: if data is not available on the base layer, users cannot reconstruct state independently, and the security inherited from the base layer is partial or absent.

The questions that separate them

Who runs the sequencer? On most rollups today, a single operator orders transactions. If it stops, normal transacting stops. Every major rollup has a mechanism for forcing transactions through the base layer directly, which is slow and expensive and exists for exactly that case.

Who can upgrade the contracts? Most rollup contracts are upgradeable by a multi-signature wallet. Whether that wallet is controlled by a small team or a broader set, and whether there is a timelock, is the single most important trust question and is documented publicly for each network.

Is data available on the base layer? Determines whether users can exit without the operator’s cooperation.

How mature is the fraud proof or validity proof system? Some networks have deployed the full mechanism; others operate with it partially enabled.

What changed in 2024

Dedicated data capacity on Ethereum, introduced in the Dencun upgrade, priced rollup data separately from ordinary transactions. Rollup fees fell by a large multiple and decoupled from mainnet congestion.

The practical consequence is that fee differences between rollups are now small enough that the choice should be made on withdrawal characteristics, trust assumptions and ecosystem rather than on cost.

Independent tracking

Several independent sites maintain standardised assessments of each network’s trust assumptions, including sequencer arrangements, upgrade keys and proof system status.

Those are the authoritative source and they are updated as networks change. Any comparison written at a point in time, including this one, is a snapshot.

Practical route between a rollup and currency

Withdrawing through the canonical bridge is trust-minimised and slow on optimistic designs. Third-party fast bridges charge a premium for the seven days.

A third route is frequently cheapest and fastest: withdraw from the rollup to a venue that accepts deposits on that network, then withdraw to currency or to another chain. Which networks a venue supports per asset is published, including by a regulated European platform, and comparing that list against the bridge cost is worth doing before moving anything substantial.