In-depth essays · #5 / 50 · L2 Rollups, Federations & Interoperability

What Sequencers Do—and Why Based Rollups Like Taiko Hand Ordering Back to L1

Inside the design trade-offs between sub-second UX, sequencer revenue, and L1-grade censorship resistance

Press send in an L2 wallet and your transaction confirms in under a second, even though Ethereum L1 produces a block only every 12 seconds. That instant feedback comes from the L2’s sequencer—the traffic controller that orders transactions—and how to decentralize that role is one of the defining debates in Ethereum scaling.

1. How a Sequencer Works: Ordering Transactions and Issuing Soft Confirmations

A sequencer receives incoming L2 transactions, orders them into blocks, executes the state change, and hands users an immediate soft confirmation before compressing the batch and posting it to Ethereum L1.

Most major L2s today run a single sequencer operated by the core team. With no multi-node consensus round needed for ordering, a single sequencer can produce blocks every 250 milliseconds to 1 second (or 10 milliseconds on MegaETH) while capturing the spread between L2 execution fees and L1 blob costs.

2. The Trade-Offs of a Single Sequencer: Liveness and MEV

Relying on one sequencer introduces two bottlenecks: if the sequencer goes offline, normal L2 block production pauses until it recovers (though users on Stage 1 rollups can still force transactions through L1), and a single operator could theoretically censor or reorder transactions for MEV.

Projects address this either by rotating a decentralized sequencer set or, as Unichain does, moving block building inside Trusted Execution Environments (TEEs) so anyone can verify that priority ordering rules were respected.

3. Based Rollups: Letting Ethereum L1 Validators Sequence L2 Blocks

A more radical design, championed by researcher Justin Drake and brought to mainnet by Taiko in May 2024, is the ‘based rollup’. A based rollup eliminates the separate L2 sequencer altogether and lets Ethereum L1 block proposers include and order L2 blocks directly alongside L1 blocks.

By piggybacking on L1 proposers, a based rollup inherits Ethereum’s full liveness and censorship resistance from day one and opens the door to synchronous composability between L1 and L2.

4. A Spectrum from 10ms Real-Time Sequencers to L1-Sequenced Rollups

Because L1 slots occur every 12 seconds, based rollups pair L1 sequencing with ‘based preconfirmations’ so users still get fast UX. In Ethereum City, watching MegaETH’s high-frequency trucks alongside Taiko’s L1-anchored blocks lets you compare both ends of this design spectrum.

Related districts in the city

Related glossary terms

This essay is an explainer written to help you read the city; it is not investment advice. Facts on this page are as of 2026-10 and may change. If you spot a mistake, please tell us through the contact page.