# Rollups

- Main idea
- Store only the transaction data to layer1 and put all execution and storage off-chain
- Method
- Deploy smart contracts on layer1
- Types
- Optimistic rollups
- Fraud proof
- Idea
- Assume the posted data is valid (optimistic)
- If not valid, others can submit fraud proof
- There is a dispute resolution system to disincentivize bad actors
- Only when there is a dispute, the tx will be executed on layer1 to check its validity
- The one (Alice) who submits tx to layer1 needs to provide a bond (ETH)
- The one (Bob) who submits fraud proof also needs to provide a bond (ETH)
- If the fraud does exists, Alice’s bond will be slashed. Otherwise, Bob’s bond will be slashed
- Dispute resolution rarely happens because of the incentive mechanism
- Projects
- Optimism
- Uniswap, synthetix
- Arbitrum
- Uniswap, sushi, augur, aave, chainlink
- Difference
- Optimism
- Run whole tx on layer1 to check validity
- Handling tx ordering (MEV)
- The ordering of tx can be auctioned off
- Arbitrum
- Use an interactive multi-round model, narrow down the scope of dispute, and potentially execute a smaller portion of tx on layer1 to check validity
- Smart contracts can be larger than the max contract size on layer1
- Handling tx ordering (MEV)
- Running a centralized sequencer to order tx (plan to decentralize in the future)
- ZK rollups
- Validity proof
- Idea
- Use zero knowledge proof
- When submitting a tx batch, a zk-snark proof, which can be quickly verified by the layer1 smart contracts, is posted.
- Invalid zk-snark proof can be rejected straight away
- Projects
- Loopring, Hermez, Aztec (privacy focused), StarkWare (used by dyed)
- Differences
- Optimistic rollups
- Have to give enough time for ppl to submit fraud proofs before finalizing a tx
- Withdrawal from layer2 to layer1 is slow (7 days)
- Fast withdrawal that charges a small fee: Hop protocol, Connext
- 99% of the EVM smart contracts can be ported without making changes
- Zk rollups
- Fast withdrawal since the validity proof is submitted with the tx batch (already proven valid)
- Difficult to make it EVM compatible
- Zksync is working on this
- More computational heavy
- Nodes that compute zk proof have to be high spec machines
- hard to run for normal ppl
- Similarity
- Can scale ethereum from 15 TPS to 1000-4000 TPS
- Can even be higher if allocated more space on layer1
- If integrated with ETH 2.0 (save more data with sharding), can scale to 100k TPS
- Challenges
- Different rollup solutions are not compatible
- Fractional liquidity
- Split across layer1 and different layer2 solutions
- Notes
- If centralized exchanges enable easy withdrawal and deposit between rollup chains, the adoption can raise quickly
- Will rollup threaten side chains?
- No, side chains have lower fee than rollups

<iframe width="560" height="315" src="https://www.youtube.com/embed/7pWxCklcNsU" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>