Partnered With:

Briefly analyze the Layer2 scheme Arbitrum

Read this article in 18 Minutes

Arbitrum is getting a lot of attention as a Layer2 solution on Ethereum.

Original title: "A Simple Understanding of Arbitrum"

Original source: Blue Fox Notes

Recently, both Uniswap and Sushiswap are planning to use Arbitrum's L2 solution, which has made Arbitrum the focus of the Ethereum L2 solution. So, how to understand Arbitrum in a nutshell.

Arbitrum is also the L2 solution in the Rollups series

There are a number of Layer2 schemes on Ethereum, but one of the most notable ones is the Rollups family. The Optimisitc Rollups and ZK Rollups are described in Blue Fox Notes. See the article "Understanding ZK Rollup and Optimistic Rollup: Important Expansion Directions for Ethereum".

Arbitrum is also a solution to the Rollups family. These schemes basically put security on the Ethereum chain and computing and storage under the chain. In simple terms, participants submit their transaction data to the Ethereum chain, where users can view transactions, but transactions are calculated and stored off the chain. Arbitrum regularly sends Ethereum a hash containing the full state that occurs on Arbitrum, and this hash is placed on the chain as a result, achieving immutability and finality.

Relatively speaking, Arbitrum is similar to Optimistic Rollups in terms of frame thinking. In terms of result verification, both of them belong to the category of fraud proof, while ZK Rollups belongs to the category of encryption validity proof.

On the premise, the fraud proof class Rollups assumes that the results on the commit chain are available, and anyone can extract and execute the transaction data of Layer1, comparing the states, to check if the contract was executed correctly. Encryption proof class Rollups takes a mathematical approach and achieves validity through zero-knowledge proof, which is inherently impossible to fake. Therefore, the proof of encryption validity belongs to the active proof, while the proof of fraud belongs to the passive proof. When people do not agree with the result, there will be a dispute, which needs arbitration.

Overall, ZK Rollups is more secure and has a faster implementation finality, which means it has a faster pull time and is better suited for scenarios such as money transfers. However, both the schemes of Optimistic Rollup and Arbitrum have a challenge period, which can only be confirmed and withdrawn after the end of the challenge period. This is the shortcoming of both schemes. However, it is computationally expensive for ZK Rollups to generate proof of encryption validity, and it will take some time for it to mature.

In the long term, ZK Rollpus may be the ultimate solution, but it will take longer for ZK Rollups to mature and land, while the need for Ethereum to expand is more urgent, as can be seen from the rise of this smart chain. The slow landing of ZK Rollups gives opportunities for Optimistic Rollups and Arbitrum.

Therefore, from the perspective of short and medium term landing, fraud proves that Optimistic Rollups and Arbitrum Rollups have a faster landing chance. The fraud proof itself is "optimist", presupposing that all presenters are good people unless proven guilty. As long as a verifier is present, proof of fraud can be submitted and it is safe. Therefore, the assumption that there is an honest verifier becomes important. Only one verifier is required to submit an online proof of fraud, with a one-week (or more) window for proof of fraud, which also results in slow final confirmation times and long withdrawal times.

In summary, the Rollups for both fraud proof and encryption validity proof classes have their advantages and disadvantages. ZK Rollups are non-interactive and solve problems through a mathematical approach, while Optimistic Rollups and Arbitrum Rollups are both interactive. In this sense, Optimistic Rollups and Arbitrum Rollups are similar, but they differ in specific paths.

Arbitrum is a multi-round interactive Rollups solution

Seen from the above, Arbitrum and Optimistic are both Rollups schemes in the scope of fraud proof. The core problems of both schemes are: how to verify the correctness of the results; Will someone send malicious proof; Therefore, both have challenge mechanics. The verifier can send assertions to the chain and dispute them. If the assertion is false, it loses the margin asset it pledged.

The main difference between Arbitrum and Optimism is the way in which differences are resolved. What happens when a validator submits a ROLLUP block to L1 and someone thinks it is incorrect?

Arbitrum uses multiple rounds of interactive protocol to resolve disputes, breaking down large disputes into small ones until the most critical step is found, and then using the Ethereum contract to determine if it is correct. By breaking down disputes, Arbitrum attempts to achieve a more efficient solution.

Participants in Arbitrum can pledge margin, asserting that a state will eventually be confirmed. If the assertion is incorrect, the user's pledge security is forfeited. Two users pledge on different squares, which means that only one of them is true. To prevent someone from attacking, the claimant of the dispute needs to pledge the security asset to prevent it from doing evil.

In the initial design, the Arbitrum protocol dealt with one disputed assertion at a time. A disputed claim is made by a participant and there is a challenge period during which anyone can challenge the disputed claim. If no one challenges it, then the disputed assertion shall be deemed valid.

But there are bottlenecks in this single-line propulsion model. Activating a dispute assertion one at a time is inefficient and the VM process is limited. At the same time, the wrongdoer can deliberately slow down the VM by challenging disputed assertions. Even if you have to pay out the money pledged, you can delay.

Therefore, in the new design, Arbitrum can handle multiple disputed assertions simultaneously, rather than linearly. It is more difficult for the perpetrator to slow down the process. At present, Arbitrum adopts the method of simultaneous processing of multiple disputed assertions. One pledge can process one claim at a time, while different pledgers can process multiple assertions simultaneously.

As the assertion is recognized as valid, the state of its VM moves forward. In general, people choose to act honestly because it ensures that their collateral will not be forfeited. There is no need to pledge their assets to the wrong branch unless there is a reason for malicious attack.

In order to implement the trustless feature, any honest participant in the design of Arbitrum can promote VM correctness and progress. If a participant always pledges on the right branch, he or she will win all the disputed claims. If the others disagree, they will only lose their pledge. Only honest participants can get away with it.

(Alice and Bob pledge on different branches, ED&NBSP from Arbitrum; Felten)

If Alice's assertion in the figure above proves to be correct, then the margin asset pledged by Bob will be lost, but Alice will gain.

In a single-round interactive Rollups, where the assertion includes the result of each call, the challenger points to the specific call in the assertion that has the wrong result. The on-chain contract simulates the challenged call and checks for errors. If there is an error, the entire claim is cancelled and the assertor loses the security for the pledge. If there is no successful challenge after the challenge window period, the assertion is accepted and is final.

In Rollups with multiple rounds of interaction, there is also a challenge window, where there are multiple rounds of interaction between the assertor and the challenger, and the on-chain contract acts as a referee to decide who is at fault. The main consideration of multi-round interaction design is to minimize on-chain effort for dispute resolution. Minimize the amount of work on the chain through multiple rounds of interaction between the challenger and the assertor.

Single-round and multi-round interactions are essentially a balance between on-chain costs and dispute resolution time. Both single-round and multi-round interactive Rollups need to write all calls to the contract and their data on the chain, differing in what they need to commit to the chain as part of an assertion. Single-round interaction requires a full call on the chain to be simulated, which is more expensive in terms of cost. The multi-round interaction can be subdivided, which can narrow the scope of disputes, and the data written to the chain is relatively less, which makes the cost of its execution on the chain lower. However, it is not without its drawbacks, and its determination increases the number of rounds and the time.

In addition, Arbitrum's multi-round interaction mode may face delayed attacks. Because the system is not trusted. Malicious attackers can slow down the process, but they do so at a cost, of course, in the forfeiture of their collateral. It is also important to note that the malicious actor cannot prevent the honest actor from continuously building the honest branch, nor can it prevent the honest actor from getting final confirmation on the honest branch. What they can do is to postpone confirmation of the honest branch on the chain by pledging the margin on the wrong branch. They can make multiple disputed claims, and if there are enough honest participants, they can tackle the disputed claims simultaneously and capture the attacker's pledge money. In other words, the more honest participants, the more assets the attacker needs to pledge to delay the attack, and the more difficult the attack.

Furthermore, in Rollups with multiple rounds of interaction, a valid "assertion" that can be achieved becomes a "trustless finality" where anyone can be confident that its final confirmation is inevitable. To ensure their own final validation, the user can participate in the protocol to defend the assertion, even if acting alone, to facilitate its final validation to occur.

If there is no challenge, multi-round and single-round interactions will be confirmed at a similar rate, and if there is a challenge, multi-round disputes will take longer. In this respect, single-round interaction has advantages. A single round of interaction does not allow confirmation to delay an attack, while the benefit of multiple rounds of interaction is smaller on-chain data that can handle contracts (computation and storage) that breach Ethereum gas limits.

There is a certain period of time in the final validation of the Rollups for the fraud proof class. As a result, transfers through Arbitrum do not achieve the same fast finality as the ZK Rollps solution. The possible solution is through third party intervention. The third party can transfer its funds to the user immediately, but the user also has to pay an early withdrawal fee. In general, third parties assume that an undetermined transfer is certain to occur, and the third party will pledge the margin on an honest outcome. Because anyone can transfer funds to users who need quick withdrawals. Well, there is a market, too, where fees are in equilibrium, not too high or too low.

EVM compatibility of Arbitrum

Arbitrum is trying to be fully compatible with Ethereum, compatible with EVM, without having to rewrite the program. Fully compatible with Ethereum, and this is where we see the power in the smart chain. This is the most important step for it to successfully embed into the current ecosystem of Ethereum.

The dApp developers use Solidity language to write their smart contracts. These smart contracts are compiled into the Arbitrum VM and can be run on Arbitrum Rollups.

The progress of the Arbitrum

Arbitrum Technologies is currently launching Arbitrum One. The Arbitrum contract is deployed on the Ethereum mainnet and has started to accept the project. It is currently in the testing phase of the mainnet. According to Arbitrum, more than 250 teams have applied to join, and it is open to all developers. The next few weeks will see more and more Defi projects enter the Arbitrum ecosystem.

Currently, some important players in the Ethereum ecosystem, such as Etherscan, are also supporting the Arbitrum One chain.

To taifang and the landing of L2

Interestingly, Arbitrum has no native token, uses ETH as fuel, and pays for everything with ETH, which is good for ETH, regardless of how much it needs ETH in the future.

For now, the scalability of Ethereum requires a combination of short and long term approaches to meet its rapidly growing needs, especially the explosive growth of DEFI and NFT. In the long term, L2 is likely to be the ZK Rollups winner, but due to its high cost and delayed landing, it may take a long time to digest. Arbitrum and Optimistic Rollups may be landed earlier. Although they take a long time to extract, they are not without solutions. In addition, when you grow up in practice, new ideas may also be born.

Aside from the gradual arrival of L2, with the possible launch of EIP-1559 in July this year, and the possible PoS by the end of the year, 2021 promises to be an unusual year for Ethereum.