Qué es lo que realmente hereda una capa 2 de su capa 1

TheCoinrise Media · 2026-08-24T04:18:30.000Z
Una capa 2 hereda dos cuestiones específicas y mecánicas de su capa 1: suficiente información publicada para que cualquiera pueda reconstruir lo que ocurrió, y una forma exigible para que la cadena base compruebe o dispute el resultado. No hereda una garantía general de que sea tan segura como la cadena base en conjunto; quién ejecuta el secuenciador, quién tiene las claves de actualización y si el sistema de disputas o de pruebas está realmente en funcionamiento son preguntas separadas sobre las que la cadena base no tiene ninguna voz. La propia documentación de capa 2 de Ethereum lo plantea como una prueba de herencia: describe una capa 2 como una cadena que extiende Ethereum e hereda sus garantías de seguridad, según ethereum.org. Esa herencia es lo que separa un rollup de una sidechain, que ejecuta su propio conjunto de validadores y su propio presupuesto de seguridad en lugar de derivar la seguridad desde Ethereum, según eco.com.
A layer 2 inherits two specific, mechanical things from its layer 1: enough data posted where anyone can reconstruct what happened, and an enforceable way for the base chain to check or dispute the result. It does not inherit a blanket guarantee that it is as safe as the base chain overall – who runs the sequencer, who holds the upgrade keys, and whether the dispute or proof system is actually live are separate questions the base chain has no say over.
Ethereum’s own layer 2 documentation frames this as an inheritance test, describing a layer 2 as a chain that extends Ethereum and inherits its security guarantees, according to ethereum.org. That inheritance is what separates a rollup from a sidechain, which runs its own validator set and its own security budget rather than deriving security from Ethereum, according to eco.com.
The two things inheritance actually means
Every rollup design splits a blockchain’s normal job – ordering transactions, storing data, and executing the resulting state change – into an offchain part and an onchain part, according to Cube Exchange’s explainer on optimistic rollups (21 March 2026). Execution moves off Ethereum. What stays on Ethereum, or gets posted there, is the transaction data and a claim about the resulting state. Ethereum does not replay every L2 transaction as it happens; it holds the evidence needed to replay it if someone disputes the claim, per Cube Exchange.
That is the whole mechanism. Security does not come from Ethereum re-checking everything. It comes from the fact that anyone can force a check, because the data is there and the dispute path is open, per Cube Exchange.
Worked example: an honest batch and a tampered one
Cube Exchange lays out a hypothetical: a rollup sequencer collects 10,000 user transactions, orders them, executes them offchain, and posts the batch data to Ethereum along with a claimed new state root. Ethereum does not immediately replay those transactions. It opens a challenge window. Any watcher can download the posted data, re-execute the batch independently, and compare results. If the sequencer was honest, watchers get the same answer and do nothing, and the batch finalizes once the window passes, according to Cube Exchange.
Now change one fact in that same example: the sequencer slips in an invalid transaction that credits itself extra tokens. A watcher replaying the batch notices the claimed state root does not match what they computed, and challenges it on Ethereum. That triggers the dispute process – in modern optimistic rollups, an interactive back-and-forth that narrows the disagreement down to a single disputed step rather than replaying the whole batch, which Cube Exchange describes as a way to keep the eventual onchain check cheap even though the batch itself was large.
StarkWare’s own explainer (18 June 2024) describes the same design in blunter terms: optimistic rollups assume every batch sent to the base layer is valid “unless proven otherwise,” per StarkWare, backed by a dispute window StarkWare describes as up to seven days for anyone to submit a fraud proof.
Why the data itself has to reach L1
None of this works if challengers cannot get the underlying transaction data. Cube Exchange’s comparison of data-posting methods notes that Ethereum-based rollups publish batch data either as calldata, which lands directly onchain at higher cost, or in blobs, a cheaper and separate fee lane introduced by EIP-4844. A design that skips onchain data entirely and relies only on hashes or an external data-availability layer weakens the guarantee, because a challenger has nothing to reconstruct execution from, according to Cube Exchange.
EIP-4844 changed the economics of this, not the security model. Before it, per eco.com (26 May 2026), L2s paid Ethereum’s ordinary calldata price to post their data, and that cost dominated user fees. Blobs, a dedicated cheaper lane sized for rollup data, cut L2 fees by roughly an order of magnitude, according to eco.com. Cube Exchange makes the same point from the technical side: publishing to blobs reduces user fees without changing the basic optimistic-rollup security model, because commitments to the blob data remain available onchain even as the underlying blob data itself is propagated separately, per Cube Exchange.
The scale of the original problem is worth naming. StarkWare’s 18 June 2024 post puts pre-scaling Ethereum throughput at around 15 transactions per second with fees that it says reached up to $40 per transaction, against a Visa network StarkWare describes as capable of roughly 65,000 transactions per second but typically running near 1,700. An academic survey by Song, Qu and Wei (arXiv, 19 June 2024) makes a related point about the base layer itself: a blockchain’s low throughput compared with centralized systems traces to its consensus mechanism, which requires every node to verify transactions before a block is added, unlike a centralized processor that simply handles requests on a single server.
Optimistic vs ZK: two different enforcement paths
ZK rollups solve the same inheritance requirement differently. Instead of assuming a batch is correct and waiting out a challenge window, a ZK rollup submits a cryptographic validity proof alongside its state update, and an Ethereum verifier confirms that proof without re-executing the transactions, according to StarkWare. Because the proof is checked immediately rather than disputed over days, StarkWare says ZK rollup withdrawals settle within hours rather than the up-to-seven-day window optimistic designs use. Both approaches still rely on the same two-part inheritance: data has to reach L1 in some usable form, and L1 has to be the place where correctness is ultimately settled – by proof in one case, by dispute in the other.
Where inheritance breaks: sidechains and validiums
Not everything marketed as a layer 2 clears this bar. Ethereum’s documentation draws a hard line: sidechains and validiums “do not derive their security or data availability from the main chain,” even when they bridge assets to and from Ethereum. eco.com’s 2026 primer names Polygon PoS and Gnosis Chain as sidechains in this category, each running its own validator set and its own security budget rather than posting data back to Ethereum for anyone to check. Validiums sit in between: eco.com describes them as posting a validity proof to Ethereum but keeping the underlying transaction data on a separate data-availability layer, which lowers cost but means state can only be reconstructed if that separate layer stays honest and available.
What ‘inherits security’ leaves out
Ethereum’s own layer 2 page hedges its claim carefully: it says layer 2s inherit Ethereum’s security only “in an ideal world,” and adds that many of the projects are “still young and somewhat experimental,” with none as battle-tested as Ethereum mainnet itself. That is the gap between the mechanism and the marketing phrase. Data availability and an enforceable dispute or proof path are the parts the base chain actually guarantees. Who controls the sequencer that orders transactions, who holds the keys that can upgrade the rollup’s contracts, and whether a chain’s fraud-proof or validity-proof system is fully live and permissionless are separate, chain-specific facts that the L1 inheritance test says nothing about. eco.com’s own 2026 primer points readers to L2Beat’s risk framework for exactly this reason, describing it as scoring each chain on correctness guarantees, data availability, and the exit game a user relies on if a sequencer misbehaves.
What this page does not tell you
This page explains the mechanism shared across rollups generally. It does not check the live security status of any specific chain named here – Arbitrum, Optimism, Base, zkSync Era, Starknet, or any other. eco.com’s 2026 primer states that both Arbitrum and Optimism run permissionless fraud proofs and that Base has been the fastest-growing chain on either side, but those are claims from a single source in this evidence set, made in the course of promoting its own cross-chain routing product, and they are not independently corroborated here. None of the sources in this evidence pack document, per chain, who currently controls a given rollup’s upgrade keys or how decentralized its sequencer is – both of which sit outside what L1 inheritance covers regardless of how mature the chain’s fraud-proof or prover system is. The seven-day challenge window and the multi-hour ZK finality figures are described in the sources as typical or up-to figures, not as a specific parameter for any named chain today. The $40-per-transaction and 15-transaction-per-second figures come from StarkWare’s 18 June 2024 post and describe a historical high-fee condition on Ethereum, not a current baseline. Readers who need the present-day, chain-specific answer to whether an L2’s security is actually live should check L2Beat’s risk framework directly, which is what both ethereum.org and eco.com point to rather than asserting that status themselves.
Sources
Every fact above is attributed to one of these reports. Where they disagree, the article says so.
TheCoinrise reference desk
TheCoinrise reference desk
TheCoinrise reference desk
TheCoinrise reference desk
TheCoinrise reference desk
TheCoinrise reference desk
The post What a layer 2 actually inherits from its layer 1 appeared first on TheCoinrise.com.