2021 witnessed a boom of the Ethereum ecosystem. To solve the related network congestion problem and high Gas fees, multiple teams introduced Layer 2 solutions. Today, several of those protocols are already up and running, with prominent examples such as Rollups and Plasma. At the same time, Layer 2s are still at an early stage, and innovated.
Given so many Layer 2 solutions, it is difficult to find the most suitable one for any specific use case. This is why this article will walk you through leading Layer 2 scaling solutions, in terms of how they work and their pros as well as cons.
| Layer 2 Scaling Solutions | |||||
| Solutions | Mode of operation | Advantage | Disadvantage | Ecosystem | |
|
Rollup |
ZK Rollup |
Computation is done off-chain before the compressed data and its availability proofs are uploaded to the mainnet. |
Transactions and proofs can be verified instantly after being submitted to the mainnet. Transaction data is stored on the mainnet with a relatively high level of security and decentralization. Less data is included for high scalability. |
A great volume of data availability computation is needed. The transaction order can be decided by node operators. |
zkSync, StarkNet |
|
Optimistic Rollup |
Computation is done off-chain before the compressed data is submitted to the mainnet.
Optimistic Rollup assumes the data is valid by default.
Yet, it is equipped with a fraud-proving scheme for the disputed transactions. |
Compatible with EVM and Solidity.
Transaction data is stored in the mainnet with a high level of security.
No computation is needed allowing higher scalability. |
The challenge period is relatively long with a fraud-proofing mechanism.
More data will be produced given a disputed transaction.
The transaction order can be decided by node operators. |
Arbitrum, Optimism |
|
Plasma |
It is preferable to compress transactions into a block on the Plasma chain for computation and storage. After that, the transactions will be settled in the mainnet with a fraud-proofing mechanism for the disputed ones. |
High trading speed with low costs given fewer interactions with the mainnet.
Strong security under no supervision.
|
Don’t support general-purpose computing and many types of trades. Plasma relies on operators for data storage. Don’t support rollback leading to unavailable data. Delayed withdrawal caused by the challenge period. |
OMG, Polygon, Gluon, LeapDAO |
|
State Channels |
Based on multisig contracts, state channels allow participants to carry out a great volume of off-chain transactions with only two of those trades being submitted to the mainnet. Now, there are state channels and payment channels.
|
High efficiency without waiting for blocks to be packed. Theoretically, an infinite TPS can be achieved given high throughput. Off-chain trades can be settled immediately for channel participants. Suitable for large traders. |
Relatively high time and capital costs for settlement and channel establishment. Funds need to be locked in payment channels. Not accessible for all with limited use cases. |
Celer, Raiden, Liquidity |
|
Validium |
Computation is done off-chain before merkle root is uploaded to the mainnet. Also, transaction completeness is secured through availability proofs. But transaction data will not be saved on the mainnet.
|
No withdrawal delay. Resistant to some economic attacks for fraud proofing systems. Securing data validity through availability proofs.
|
Using exclusive languages with limited support for general-purpose computing and smart contracts. High computing power is needed to generate ZK proofs. Don’t support rollback leading to unavailable data. |
DeversiFi, ZKSwap, zkPorter （under development ） |
|
Through the comparison of different solutions above, it can be found that Rollup focuses more on security and decentralization while state channels are more scalability-centered. On-chain transaction data are not available since they are not saved on the mainnet for Plasma and Validium using different proofing mechanisms to secure data validity.
Hence, each solution should be matched with its suitable use cases based on its focus before a solution with great security, decentralization, and scalability emerges.