Interview: Pan Zhixiong
Interviewed teams: DeGate, DODO, EthSign, imToken, Math Wallet, MCDEX, WePiggy
As one of the earliest EVM-compatible Rollup expansion solutions launched, Arbitrum has opened up access to the mainnet to developers. It has been running for exactly one month. Many development teams at home and abroad have been actively trying and deploying it. The overall maturity of the network is optimistic, but some components, such as the block explorer, are still in their early stages and may affect the development of the project.
The two-layer network is called Arbitrum One to distinguish it from the Arbitrum technology. Although it has been deployed on the Ethereum mainnet, in order to ensure security, the team still calls it the beta version of the mainnet (Beta), and only developers can deploy it in advance.
Apart from Arbitrum, Optimism is a technical solution similar to them. However, since the restricted main network was launched in the first quarter and it chose to cooperate with Synthetix, it has not yet conducted a wider test, and it has also postponed the launch time once, expressing It will be officially launched in July.
The other two Rollup expansion solutions based on zero-knowledge proofs may be launched later. Matter Labs said that zkSync 2.0 will be launched in August, and its compatibility is worse than that of the previous two. It does not support some uncommonly used EVM opcodes; and StarkWare’s StarkNet solution launched the test network only this month, and the main network is estimated to be At the earliest it will be the end of this year.
In order to understand the current state of Arbitrum One more intuitively, Lianwen interviewed 7 teams that are connecting to the Arbitrum network, including DeFi, applications, wallets and other upstream and downstream teams. It is hoped that through the different perspectives of these developers, a more comprehensive state of Arbitrum's first mainnet can be restored.
On the whole, these development teams have a very good evaluation of the maturity and completeness of Arbitrum One. Most of the tools and infrastructure have already supported or will migrate to support soon (such as Chainlink and The Graph).
The focus of everyone's complaints is that the current block browser is not perfect, which may affect the speed of development. However, Etherscan is the most widely used block browser in the industry. They have added more functions to Arbitrum One's block browser. It shouldn't be too difficult. In addition, the difference in Gas calculation method has also been mentioned many times, and developers need to pay attention to this part of the difference.
In addition, there is another very interesting situation. Since Arbitrum One adopts a mechanism similar to fraud proof, which is the top priority of this network security feature, it is thought that developers will pay attention to the various tests and methods of fraud proof of the network itself. Practice, but everyone did not pay attention. However, in the early days, the official or trusted third party may be the maintainer of the network, so there should be no problem with security.
Finally, regarding the official opening time of the Arbitrum One main network, everyone’s assessments are also quite different. The most optimistic team believes that the current network is complete enough to open, but the more cautious development team thinks that it may take up to 6 months. .
In addition to the above conclusions, this interview mainly discussed these topics:
Is the EVM compatibility really the same as the official announcement, and what is the overall migration project?
How stable is the network? Can development tools and infrastructure be migrated without a trace?
What is the actual performance and cost performance of the Arbitrum network?
How long is it estimated that the main network will be fully opened?
Interview: Pan Zhixiong
Interviewed teams: DeGate, DODO, EthSign, imToken, Math Wallet, MCDEX, WePiggy
As one of the earliest EVM-compatible Rollup expansion solutions launched, Arbitrum has opened up access to the mainnet to developers. It has been running for exactly one month. Many development teams at home and abroad have been actively trying and deploying it. The overall maturity of the network is optimistic, but some components, such as the block explorer, are still in their early stages and may affect the development of the project.
The two-layer network is called Arbitrum One to distinguish it from the Arbitrum technology. Although it has been deployed on the Ethereum mainnet, in order to ensure security, the team still calls it the beta version of the mainnet (Beta), and only developers can deploy it in advance.
Apart from Arbitrum, Optimism is a technical solution similar to them. However, since the restricted main network was launched in the first quarter and it chose to cooperate with Synthetix, it has not yet conducted a wider test, and it has also postponed the launch time once, expressing It will be officially launched in July.
The other two Rollup expansion solutions based on zero-knowledge proofs may be launched later. Matter Labs said that zkSync 2.0 will be launched in August, and its compatibility is worse than that of the previous two. It does not support some uncommonly used EVM opcodes; and StarkWare’s StarkNet solution launched the test network only this month, and the main network is estimated to be At the earliest it will be the end of this year.
In order to understand the current state of Arbitrum One more intuitively, Lianwen interviewed 7 teams that are connecting to the Arbitrum network, including DeFi, applications, wallets and other upstream and downstream teams. It is hoped that through the different perspectives of these developers, a more comprehensive state of Arbitrum's first mainnet can be restored.
On the whole, these development teams have a very good evaluation of the maturity and completeness of Arbitrum One. Most of the tools and infrastructure have already supported or will migrate to support soon (such as Chainlink and The Graph).
The focus of everyone's complaints is that the current block browser is not perfect, which may affect the speed of development. However, Etherscan is the most widely used block browser in the industry. They have added more functions to Arbitrum One's block browser. It shouldn't be too difficult. In addition, the difference in Gas calculation method has also been mentioned many times, and developers need to pay attention to this part of the difference.
In addition, there is another very interesting situation. Since Arbitrum One adopts a mechanism similar to fraud proof, which is the top priority of this network security feature, it is thought that developers will pay attention to the various tests and methods of fraud proof of the network itself. Practice, but everyone did not pay attention. However, in the early days, the official or trusted third party may be the maintainer of the network, so there should be no problem with security.
Finally, regarding the official opening time of the Arbitrum One main network, everyone’s assessments are also quite different. The most optimistic team believes that the current network is complete enough to open, but the more cautious development team thinks that it may take up to 6 months. .
In addition to the above conclusions, this interview mainly discussed these topics:
Is the EVM compatibility really the same as the official announcement, and what is the overall migration project?
How stable is the network? Can development tools and infrastructure be migrated without a trace?
What is the actual performance and cost performance of the Arbitrum network?
How long is it estimated that the main network will be fully opened?
Question 1: When you migrated to Arbitrum, did the original smart contracts on Ethereum L1 need to be adjusted? What is the overall engineering effort for this part of the migration?
MCDEX: The smart contract did not adjust the code when migrating. For block and time, a review may be required. Since our tokens are issued on L1, we have some cross-chain communication needs and need to write code.
DODO: No adjustments required, very little work. Because of DODO's multi-chain strategy, when it first migrated BSC from Ethereum, it prepared a lot of scripts. However, some work is required on the front-end products, because DODO's products are relatively complex, with many caches and data middle layers, and the migration of these contents took some time.
WePiggy: When migrating to Arbitrum, the original smart contracts on Ethereum L1 basically do not need to be adjusted. It is just a version upgrade of the OpenZeppelin contract to the latest version.
DeGate: There is basically no need to adjust the smart contract, and our contract has not encountered it. If you use block and gas related data, please refer to the following differences:
《Differences from Solidity on Ethereum》（https://developer.offchainlabs.com/docs/solidity_support#differences-from-solidity-on-ethereum）
imToken: The amount of engineering is not high, because it is EVM compatible, so it is relatively smooth. But the workload is instead on the environment and system monitoring, requiring an additional set of system maintenance and monitoring.
MathWallet: Let’s focus on the smart contract cross-chain of assets. Arbitrum provides a default Token Bridging mechanism. All ERC20 and ERC721 assets on L1 have a default automatic mapping contract on Arbitrum, which does not require the project party to redeploy ERC20 Or ERC721 contract, very convenient, I believe this mechanism will migrate a large number of assets on L1 to Arbitrum L2.
EthSign: Currently undeployable except requiring reconfiguration of Truffle network settings. It is preliminarily determined that Arbitrum's EVM is not compatible with OpenZeppelin contracts (such as the most basic Ownable), and the logic that runs normally on other networks will be directly reverted during deployment.
Question 2: Is the infrastructure on Arbitrum similar to that on Ethereum L1, such as development tools, IDE, Chainlink, and The Graph? How stable is it, have you encountered any issues?
WePiggy: We started tracking the development progress of Arbitrum from the Kovan3 version of Arbitrum, and experienced Kovan3, Kovan4, Kovan5 and other versions in the middle. When we participated in the test, we found some problems one after another, such as: the block synchronization problem between L1 and L2, the ETH contract transfer problem on L2, and actively gave feedback to the Arbitrum development team, which was also recognized by them.
MCDEX: The development tools are completely consistent, which is a very amazing experience. Chainlink and TheGraph are in the process of migrating to the Arbitrum mainnet. The only problem is that the information given by explorer is not enough, but it is not a big problem. etherscan is also being deployed and will improve the explorer experience.
DODO: Not yet. Development tools and IDEs actually fully reuse L1. The Graph belongs to the middle layer, and its stability needs to be tested (because the main network has not yet been opened to a large number of users). Chainlink belongs to other project parties, and it is not fully launched yet, and is in the debugging stage.
DeGate: Because it is compatible with Solidity, the development tools and IDE can be used universally, but there may be problems with the old versions of some tools, such as the failure of contract deployment in Truffle. Chainlink and The Graph have not yet been fully deployed.
imToken: There is no problem with using Hardhat as the development tool, and the IDE has no effect on the chain. I haven’t tried Chainlink because I didn’t use the oracle.
In terms of stability, because Arbitrum's test network has been switched, and there was even instability before, so some turbulence was encountered during the process, but it was much better afterwards. On the whole, I feel that Explorer has not yet supported by Etherscan, so the amount of information is not very sufficient, but it is usable. Recently, both Alchemy and Infura support Arbitrum nodes, which is relatively convenient.
EthSign: The development tools and L1 experience are basically the same, and the stability cannot be tested because it cannot be deployed.
Question 3: From the perspective of L2 performance, have you evaluated the performance and cost of Arbitrum at the current stage?
DeGate: We believe that the cost of Arbitrum will come more from the uplink cost of calldata for a long period of time, and its performance bottleneck is also here. Based on the current upper limit of 15 million gas in an ether block, 30% of which is occupied by Arbitrum, the average gas cost of callldata in L2 transactions is 4500 gas (an ERC-20 transfer is about 1800 gas) to estimate its throughput at 1000 transaction per L1 block, that is, 71 transactions/s.
MCDEX: The cost is about 1/100, depending on the specific smart contract function. The calculation method of ArbGas is quite different from that of L1, which developers need to adapt to.
DODO: There are already a lot of theoretical analysis on the market, and the actual situation can only be judged after a large number of users flood in.
WePiggy: From our current tests on the test network and the main network: the GasLimit of Arbitrum is relatively high, but the GasPrice is very low. In this way, the overall cost is actually much smaller than that of L1, and the performance is even faster, basically in seconds transaction.
imToken: No at present, because this requires interactive testing, but we only test the part of deploying ourselves, and do not interact with other projects.
Question 4: Have you observed the data on the chain, has Arbitrum’s fraud proof design been practiced, and what is the effect?
DODO: We have not observed fraud proofs on the mainnet, but believe that the Arbitrum team has conducted a complete exercise.
imToken: No, you may need to run a node to observe it.
MCDEX: Haven't tried it.
WePiggy: We have no experience participating in actual drills in this regard.
DeGate: Haven't actually observed the challenge process.
Question 5: As a mainnet only for developers, how long do you estimate that the current Arbitrum will be completely open to use? What infrastructure do you think this ecosystem still lacks?
DODO: We think it can be opened now, but the Arbitrum team obviously hopes to be better prepared. After all, this is a highly anticipated debut. In terms of infrastructure, we have no shortage of it, but the stability needs to be tested, and we hope it can handle huge traffic.
WePiggy: Although we have officially deployed the contracts of the WePiggy protocol on the Arbitrum One network for more than half a month, considering the implementation and improvement of infrastructure such as browsers, cross-chain bridges and oracles, it is estimated that the development team of Arbitrum still needs more It will take one month to officially open the network to ordinary users.
MCDEX: It is difficult to judge the opening time. There is nothing missing in the infrastructure as it is all moving. At present, most of the infrastructure of Ethereum has officially announced that it will enter the Arbitrum ecosystem. We can only silently expect them to complete the development, deployment and testing work as soon as possible, and it is not easy to rush them. I personally think that the infrastructure that is still missing is: a more developer-friendly SDK, more spare nodes, etc.
DeGate: It feels like it will take 3 to 6 months. It is still in the early stage of system improvement, and there is a lack of more trusted third-party verification nodes participating in governance, because not every developer can run a relatively heavy verification node.
Math Wallet: At present, there are still some basic tools in the internal testing stage, such as the Arbitrum version of Etherscan. The current official browser is not easy to use. For example, the lack of contract verification function may cause some security problems. From the point of view of node operation and basic functions, we tested that there are basically no problems. Math Wallet has also completed the wallet support of Arbitrum, and it will be released to ordinary users at the same time after it is fully opened.
EthSign: What is most lacking is actually a way to quickly get tokens. The current bridge speed is too slow, and it takes dozens of minutes for the token to cross the bridge. Second, incompatibilities with OpenZeppelin-like industry-standard contracts need to be fully fixed.
imToken：
A more complete explorer;
A gateway or solution that can have fast withdrawal;
At present, the project parties are relatively developing and deploying independently, which needs to be tested interactively.
