Rollup finality
An optimistic rollup is final on the main chain only after its challenge window, often 7 days. A rollup with a validity proof is final once the proof is verified, often within 1 hour. Pick the one that matches how fast users must exit.
See it on the board
The same traffic runs on two boards. The simulator measured both when this page was built.
| Measure | Goal | Optimistic rollup, 7-day window | Validity-proof rollup |
|---|---|---|---|
| p99 latency (Transfers) | under 5 s | 3.4 s Pass | 3.4 s Pass |
| Error rate (Transfers) | under 0.1% | 0% Pass | 0% Pass |
| Completion time (Withdrawals) | under 86,400 s | 172,800 s Fail | 1,800.01 s Pass |
| Lost writes | none | 0 Pass | 0 Pass |
| Monthly cost | 40 USD/mo | 40 USD/mo | |
| Level rules | none broken | Board is not complete Fail | Pass |
Key numbers
| Number | Value |
|---|---|
| Default block time | 12 s |
| Default confirmations | 1 blocks |
| Transfers per second, one rollup part | 10,000 ops/s |
Envelope keys · < 1 key service calls/s · limit 10,000/s
How it works
A rollup runs transfers cheaply off the main chain and posts a summary to it. The main chain must trust the summary, and two designs differ in how.
An optimistic rollup assumes that the summary is right. Anyone can challenge it during a window, often 7 days. A withdrawal to the main chain is final only after that window has ended, because until then a challenge could still prove a fraud.
A validity-proof rollup posts a cryptographic proof with each summary. The main chain verifies the proof, and a valid proof means the summary is right. Nobody needs to watch for fraud. A withdrawal is final once the proof is verified, often within an hour. Making the proof costs extra compute for the rollup operator.
In the mini board a payments app sends 300 transfers/s with p99 below 5 s. At midnight 500 users ask to withdraw their funds to the main chain, and every withdrawal must be final within 24 hours. In the run, the optimistic exit takes 48 hours, so it misses the limit. A proof-based exit meets it.
The optimistic design is often simpler and cheaper, and it can still work if a liquidity provider pays the user early and takes the delay.
Pick from the need of the exit, not from the name.
In the levels you choose the design from the limits for transfers and for exits.