Dialogue with Offchain Labs engineers: How does Arbitrum solve the problems faced by Rollup in ModuLar Oriented?

Original title: How does Arbitrum find its ecoLogicaL niche in ModuLar Oriented?

Original Author: Vision, MetastoneGroup

first level title

Topic 1: The development history of Arbitrum

Vision: At present, there are some hot discussions about Arbitrum on the whole network. Can Jason briefly introduce the whole development process of Arbitrum from its establishment to today.

Jason：Okay, let me express my opinion on the whole process of Arbitrum. First of all, we launched one of our testnets in May 21. After the launch in May, this testnet has not yet been opened to the public. I just invited some project parties to come to our deployment, experience some of our network experience, and at the same time make some improvements. Finally, we launched our first-generation mainnet on August 31, 2021, which is the mainnet deployed by a technology stack we now call cLassic. Although the technology stack is relatively early, the technology stack has actually implemented interactive fraud proofs, and all related functions of Optimistic Rollups have been deployed on it. At that time, it was already a completely EVM-compatible, Optimistic Rollups solution, so after we released it, we saw a very high chain activity.

Jason：At the same time, of course, after we went online, it may be because the technology stack is relatively original, so we may have encountered some problems in capacity expansion at that time. For example, during the Odyssey in June 2022, because of insufficient throughput due to gas , so the increase in gas will be more, and we have also noticed these problems. Therefore, in the same year, on August 31, 2022, our second-generation technology-Arbitrum Nitro was launched. The second-generation technology stack was directly upgraded from the first-generation mainnet, and on the nitro mainnet. After the line, our throughput has expanded by about seven times than before. For example, based on nitro, our fastest output time per second is about 0.25 seconds per block, and the maximum limit of each block is 32 million gas, which is basically what can be achieved in the current EVM A very large order of magnitude.

Jason：first level title

Topic 2: About State Expansion

Vision: At present, the entire blockchain needs to carry a relatively large-scale user volume. In addition to solving the throughput problem of transaction execution, there is also a problem that everyone is generally concerned about, which is state expansion. For hundreds of millions of users, let them hold many types of such assets and application states on the chain, and the problem of state explosion may also appear. I would like to ask Jason what is his point of view on state expansion?

Jason：First of all, Ethereum itself is also doing state expiration to solve this problem, and then actually including Layer 2, let me first explain a mechanism of L2, in fact, we put the DA of this transaction on L1, and then this DA is actually It is the order that includes the input of the transaction and the transaction. Because everyone knows that EVM is now a single-threaded execution environment. Since it is a single-threaded execution environment, then we can know that if the order of input and input is already known, then its output can only be unique. It is also one of the reasons why L2 can put DA in L1, and then at the same time, it can directly read DA to synchronize the L2 network. That is to say, if our state sum on Layer 2 is large enough, we can actually introduce these state expiration Yes, and after these states expire, if you want to query later, you can query on some other hosting parties, or run a node yourself, read L1 and re-execute this DA, you can regain this state up.

Jason：first level title

Topic 3: Centralization of DAC

Vision: The next question is that Celestia currently focuses on DA, that is, data feasibility. Celestia uses DAS, which is to conduct a sampling test on the data. In addition to the Rollup expansion solution, there is also an off-chain solution. Validium relies on a third-party committee for data availability, and then I learned that Arbitrum Nova also uses this DAC method. I would like to ask Jason about the logic of Arbitrum's design at the beginning. Another question is, will there be a centralization problem due to the small number of committees on the community side?

Jason：One of Nova's technology stacks is based on our Anytrust. This technology is actually the same code base as Arbitrum Nitro, which is based on our Arbitrum One. It's just that the modes they start are different, and the differences under the same code base mode. Then for data availability, Nitro will directly upload this DA to L1, but for Anytrust, it will upload this DA to DAC, and after DAC receives this, it will sign the data and issue a certificate , and then send it back to the sequencer, and the sequencer uploads the certificate to L1. This is a basic execution logic of Anytrust.

Jason：Going back to the question just now, that is, will DAC introduce a serious problem of centralization. First of all, in terms of DAC, we have been continuously expanding the data of our DAC, which is a node scale. It can be seen that there are many DACs in our current state, not only Web3, but also some large companies in Web2. There are nodes running our DAC. In addition to the DAC committee, we actually have a Mirror mechanism, which means that if you are not a DAC, but if you want to participate in this network, you can also run a normal mirror. In this case, you can synchronize some stored data from other DACs. The data, and then other people can not only get the data from the DAC, but also get the data from the mirror.

Jason：This is actually a double insurance, avoiding the situation where the DAC cannot obtain the DA when the data is lost, and the DAC will submit the certificate to Rollup. The advantage is that it greatly reduces the DA uploaded by the sequencer to L1. this overhead. Because we only need a certificate to upload now, compared to the size of the DA, it is actually very, very small, so it can also greatly reduce the gas billing of the network under the anytrust mechanism. At the same time, if you actually pay attention to the situation of the Ethereum GoerLi test network two days ago, you will know that the Ethereum GoerLi was due to some network activities, so the gas fee of goerLi will be very high. At that time, in fact, the ETH above GoerLi was already priced because of the Layer 0 bridge.

Jason：If you compare the gas overhead of Anytrust at that time, you can find that the overhead of Anytrust is even lower than the gas consumption of an Ethernet test network at that time. So this is a great advantage of Nova, that is to say, the gas overhead of our Anytrust is very, very low. You can even reach a price of less than 1 cent if it is an ordinary transfer of eth. It is a price that some other networks cannot achieve at present.

Jason：At the same time, regarding Anytrust, if DAC is doing evil, because we have a mechanism that is fall back to Rollups, we can roll back to Rollup in an emergency, and we found that the number of DAC signatures in any trust is lower than our assumption After the threshold, the network will stop putting DA into DAC, but put it into L1 through the same mechanism as nitro.

Jason：first level title

Topic 4: About the future of RaaS

Vision: Another problem is that what I know so far is that in the optimism solution mechanism, op has launched a service called Layer 2 Rollup as a Service based on op stack. In fact, its design logic is similar to Starkex customizing a DApp chain type service for other application parties, such as games or sociaLfi. Arbitrum also carried out the ecological attributes of Arbitrum one and Abitrum Nova at the beginning of the design. Separately, One is responsible for derivative services like defi and other chains, and Nova is aimed at games or socialLfi. I would like to ask, for example, when we designed this logic, did we also consider such a direction of RaaS? Then what do you think about the future roLL up as a service, or the direction of such an ecological application chain?

Jason：In fact, regarding this issue, we changed the authorization mechanism of Arbitrum last month, that is, in March, which means that after that, you can deploy a new Layer 3 network on Arbitrum without permission. , the L3 network can be deployed by choosing Nitro or Anytrust. These can be deployed directly without our permission. But if you want to deploy directly on Layer 1, that is to say, deploy a new Layer 2, you need to apply to our Arbitrum DAO, and then you need to be approved by DAO to run.

Jason：first level title

Topic 5: About ArbitrumNova and ArbitrumOne

Vision: Considering the entire Arbitrum, the current ecology is that the leaders of various projects are mainly defi projects, including the more famous projects GMX and RDNT, so the current activities on the entire chain are mainly concentrated on Arbitrum One. I would like to ask Compared with one, is there any good project that Nova can experience? At present, the retention rate of gamefi and socialLfi users is not particularly high. I would like to ask Jason about the users of Nova carrying gamefi and socialLfi What kind of point of view do you hold on retention?

Jason：OK, because I was doing technology here, so I may not be so professional in terms of ecology, so I can answer this technically. In fact, I just mentioned Nova just now, it is a dominant execution environment with low gas price, in fact, it just meets some needs of some high-throughput Gamefi. Because in Nova, some DApps need high throughput, because each transaction needs to pay gas, so the lower the gas price, the better, so this also fits its needs well.

Jason：first level title

Topic 6: Decentralized Sequencer and anti-censorship issues

Vision: I recently learned that the zkSync era mainnet was launched around March 24, but their Matter Labs said that the entire zk-Layer 2 sequencer is still not decentralized enough, and I also learned some time ago A developer of Starkware, he used Polkadot's substrate to create a decentralized Starknet sequencer. I would like to ask what is Arbitrum's current view on the over-centralization of sorting? Then and what is the way to solve this problem?

Jason：First of all, the sorting centralization is actually different from the problems caused by other node centralization. The sorter does not have the ability to do evil, because the sorter simply packs the transaction and uploads it to L1. Your sorter has no way to sign the state, or make any guarantee for the state Yes, he simply aggregates the transactions, sorts the transactions and then synchronizes the sorted transactions to other full nodes in the network and uploads them to L1.

Jason：That is to say, a transaction broadcast by the sequencer does not broadcast the final state of the transaction. It will not express any subjective opinion, but simply sort the transaction, and then many people may It will be said that if there is any review of this transaction when sorting, or if some transactions are deliberately rejected to be carried out, of course, we can solve this problem.

Jason：We also opened a deLay inbox on L1 at the same time. In the deLay inbox, we added a function, that is, you can directly send a transaction of L2 in the deLay inbox. When you find that the sorter is reviewing you transaction, or in the case of deliberately rejecting or intentionally performing MEV on your transaction, you can package your transaction into L2 in that way, so that you can avoid the sequencer from doing some Some potential mechanisms for doing evil. So here we can see that if the sequencer does evil, at most the transaction will be censored, but if it is censored, there is also a mechanism to avoid it.

Jason：first level title

Topic 7: About the future of We2 development increment

Vision: The last question is whether arbitrum will have any technical plans that everyone can look forward to in the new year after completing the entire ecological incentive. You can elaborate on the technical roadmap in detail.

Jason：Well, in fact, there is no need to wait for 2024. At the end of this year, we will soon launch the styLus technology, which is implemented by adding Wasm to this virtual machine. And after joining WASM, you can use WASM, that is, you can deploy and deploy your contracts through C++ or rust, and then if you use these languages ​​to deploy contracts at the same time, it will be the same as the contract deployed by EVM It is the same state tree that is shared. If the contract is deployed in the form of WASM, the efficiency can be increased to 10 times. If you use EVM and now use some libraries, or use some other components, you can also give up the current components, and then deploy new components through WASM, and make a call in this contract.

Jason：This is also a strategy of EVM pLus that we have been mentioning, that is to say, in the future, we will not only be perfectly compatible with EVM, but also support some functions other than EVM, and then make our scalability higher , a situation with better user experience. Then before going online at the same time, that is, in the near future, we will also deploy a new devnet, which is the developer network, and then provide developers with an experience.

Jason：If the blockchain wants to access the next level of user scale, it must support the next level of developers to enter your ecology. And in fact, if many, many Web 3 invite Web 2 now, some developers who join the current Web 3 are not used to soLidity, but through WASM, they can use a language they are already familiar with to write this contract. In this way, we It can also host more developers, and is also compatible with current Web3 developers.

Vision：OK, thanks Jason. Today, it not only answered some questions of Arbitrum, but also answered some more important questions in this direction of blockchain Medura Oriented.