From congestion to rollups: the scaling story
The Merge chapter separated consensus from execution capacity. Growing application use poses a different problem: how can more people complete useful operations without requiring every participant to do unlimited work? A larger block is one design choice, but the resources needed to verify it also matter. Scaling is a distribution-of-work question as well as a transaction-count question.
Rollups group activity and use a base chain for specified settlement and data functions. The Ethereum optimistic-rollup guide describes batched execution and a challenge process; the validity-rollup guide describes proof-based state acceptance. These architectural explanations do not certify every deployment's administrators, bridge or emergency route.
The historical connection is a learning sequence, not a claim that rollup research began after the Merge. Application demand, execution research and consensus development overlapped. Separating them helps explain why one upgrade cannot solve every bottleneck.
| Stage | Cost or delay to record | Dependency to identify |
|---|---|---|
| Enter | Funding, approval and bridge costs. | Correct asset representation and supported destination. |
| Execute | Application fee, network fee and failed attempts. | Contract permissions and transaction ordering. |
| Settle | Proof or dispute process and base-chain finality. | Data availability and the exact acceptance rules. |
| Exit | Withdrawal delay, liquidity fee and final conversion. | A usable asset at the destination. |
The optional guide below compares an invented $12 direct operation with an L2 route costing $7 to enter, $0.40 per operation and $6 to return. The L2 round trip costs $13.40 for one operation and $15 for five. The outcome depends on the whole journey, not only on the small execution fee. Open the guide to follow the arithmetic and settlement stages.
A sequencer orders transaction requests in a rollup system. Its receipt is one milestone; it may not be the final point at which the user can withdraw.
Check your reasoning: a sequencer shows “complete.” Which later milestones may remain?
Depending on the system and route, data publication, proof or dispute resolution, base-chain finality and usable withdrawal may still remain. Read the actual deployment's definitions before treating a receipt as final settlement.
Detailed guide: rollups, sidechains, settlement and total-cost calculations
Reviewed and updated: October 19, 2026.
A layer 1 blockchain establishes its own transaction history and consensus rules. A layer 2 system processes activity while relying on a base chain for specified security and settlement functions. The practical comparison is not simply fast versus slow: it is what has been confirmed, which assumptions remain and what it costs to complete the whole route.
Rollups and sidechains are different designs
Ethereum's sidechain documentation describes a sidechain as an independent blockchain with its own consensus and a bridge to Ethereum. It does not inherit Ethereum's security merely because it uses a similar virtual machine or a familiar address format.
Rollups move execution away from the base chain while using it to verify or resolve state claims and provide required data under the rollup design. Optimistic rollups use a challenge process; validity rollups use proofs. A specific deployment can add administrator, upgrade, sequencer and bridge assumptions beyond this broad classification.
Five stages behind a completed-looking transaction
| Stage | What to establish |
|---|---|
| Local receipt | Has an operator accepted or executed the transaction? |
| Data publication | Is the required transaction/state data available through the specified mechanism? |
| State acceptance | Has a validity proof been accepted or the relevant dispute process completed? |
| Base-chain finality | Has the settlement transaction reached the base chain's finality conditions? |
| Usable withdrawal | Can the user access the intended asset at the destination? |
A fast receipt is useful feedback, but it should not be described as equivalent to every later stage. Ethereum's validity-rollup guide ties settlement to proof acceptance on layer 1. Producing a proof, including its verification transaction and waiting for the underlying chain still take time. Validity proofs do not mean every withdrawal is instantaneous.
An optimistic system can impose a challenge period on its canonical withdrawal path. A liquidity provider may offer an earlier exit for a fee, but that adds a service and its conditions. Compare the actual route rather than combining the speed of one route with the security assumptions of another.
Worked example: low execution fees versus total cost
Use these invented costs: a direct L1 operation costs $12; entering an L2 costs $7, each L2 operation costs $0.40 and returning costs $6. Five L1 operations cost $60. The L2 round trip costs $7 + 5 × $0.40 + $6 = $15.
For only one operation, however, the same L2 route costs $13.40, compared with $12 directly. Under these fixed assumptions, two or more operations make the L2 route cheaper. Real prices change, and an existing L2 balance can remove the assumed entry step. These are arithmetic examples, not current network quotes.
Include approvals, swaps, bridge charges and the destination asset's usability when relevant. A token with the same ticker on two networks may have different issuer or bridge arrangements. A successful arrival is not enough if the receiving service does not support that asset representation.
A checklist for comparing a named network
- Identify the exact network and bridge rather than relying on the L2 label.
- Read data-availability, proof or dispute, and emergency-exit documentation.
- Identify who can upgrade contracts, pause operations or influence transaction ordering.
- Separate sequencer confirmation, settlement and withdrawal estimates.
- Calculate costs from the user's actual starting asset to the final usable balance.
State channels provide another scaling approach with their own participant and dispute requirements; they should not be collapsed into the rollup category. The right explanation follows the mechanism. No architectural label alone guarantees lower cost, unrestricted access or the complete absence of additional trust assumptions.
Sources and scope
Optional worked case: Follow an OP Mainnet withdrawal from the source receipt to Ethereum delivery. The case adds a stage-by-stage evidence guide, timing example and fee budget.
Continue reading: Bitcoin scaling approaches · Comparing complete trade routes
Editorial update: Kept sidechains separate from rollups; distinguished receipts, proof/dispute settlement and withdrawals; replaced fee promises with a hypothetical round-trip cost comparison.