Vitalik's online AMA summary, comprehensive interpretation of Ethereum Rollup Layer 2 technology

This article is sourced from Bihu and organized by Xingzou. On the evening of February 28, Bihu founder Guru and Ethereum co-founder Vitalik held an AMA on Ethereum Rollup Layer 2 technology.

Here is a summary of last night's content, Enjoy:

Guru: Let me first share why we chose this topic today. We are also conducting research on Rollup Layer 2 technology. I believe that Rollup Layer 2 technology will become a mainstream application scenario for Ethereum in the future. In the future, the Dapps that everyone uses will mainly operate on Layer 2 Rollups rather than Layer 1. This is a very important topic.

There are rumors that some Rollup Layer 2 networks will go live on Ethereum in March. It’s a good opportunity to share this with everyone during Bihu's third anniversary. Vitalik will answer questions about Layer 2 and Rollups in a face-to-face Q&A format.

Vitalik is not in China; he is sharing with us in the morning North American time, which is really commendable. Vitalik will try to communicate with everyone in Chinese throughout the session.

If anyone has questions before Vitalik arrives, I can answer some of them.

1. Will Layer 2 networks increase the value of Bitcoin?

The Layer 2 network of Ethereum is not closely related to Bitcoin. Of course, there is wBTC on Ethereum, but I personally feel it is not a very direct inertia system.

2. Which Rollup solution does Guru favor?

Let’s leave this question for Vitalik. What I have seen more is Optimistic Rollup. Its advantage is that applications on Layer 1 can be relatively easily migrated over. The code for DeFi or other applications on Layer 1 can basically be reused on Optimistic Rollup. Mature code and applications that have been audited on Layer 1 can be directly moved to Layer 2, and the ecosystem (established) will grow relatively quickly.

3. Will Ethereum's scalability impact Polkadot?

I believe that if Ethereum Layer 2 develops, it will put pressure on all public chains. Because currently, Ethereum's ecosystem is actually the best, and compared to other public chains, it is not a small margin.

The main issue with Ethereum now is throughput. I mentioned this in my last live stream with Dovey. At that time, many people did not think throughput was the biggest problem. In fact, when I communicated with many people in the industry, I felt that the biggest problem on Ethereum two years ago was indeed the throughput issue. Of course, many people in the Ethereum ecosystem also think so.

Ethereum's scalability has been in progress for a long time. Initially, it expanded through state channels, which led to the Lightning Network. The problem with state channels is that users have to be both service providers and users. State channels require users to observe the state themselves, and when they finish, they need to send the state back to the underlying public chain. In state channels, users are both users and service providers, which is definitely not right. It is unlikely to require users to both provide services and use them themselves.

Later, after state channels, Plasma emerged, where some nodes provide services. Users only need to care about usage. But Plasma has a problem: when data is not provided to you, how do you trust the Plasma nodes? How do you know if they are acting maliciously or not? This was a very difficult problem before. In Layer 2 data, when nodes do not provide you with data, you do not know what they are doing.

In the assumption of Plasma, if users still need to check the network's state. When users do not receive new data, they should escape from the Plasma network because they should assume that these Plasma nodes are acting maliciously. So Plasma also has a transition period of several hours or days, which is a challenge period. This is actually similar to the current Optimistic Rollup.

Ultimately, Plasma was not accepted by the Ethereum ecosystem because…

Vitalik Joins

Guru: Sorry, the Chinese community is really enthusiastic today. I will still make a brief opening statement.

Why did we choose this theme today? It is because we are also looking into Rollup research, and we are preparing to choose a Rollup on Ethereum Layer 2 to launch a decentralized trading protocol. During the research process, I increasingly feel that Layer 2 based on Rollup will become the mainstream of Ethereum's future.

It is possible that at some point in the future, when everyone uses Ethereum, they will mainly be using Ethereum's Layer 2. This kind of change is a qualitative change for Ethereum itself and for the entire ecosystem. I believe that Layer 2 technology based on Rollup will bring leapfrog changes to the entire industry.

So I am very optimistic about Layer 2 technology. This is also why we chose this topic today.

Everyone is very enthusiastic; sometimes I can't even see the questions in the comments. Today's theme will mainly be completed through a Q&A format. Everyone asks questions, and Vitalik answers. I believe everyone has seen many articles about Rollups recently.

Without further ado, let's get to the point. I will throw out a question to start. This question is also of great concern to many development teams that want to develop on Rollup Layer 2.

The security of Layer 2 Rollup is based on the data of Layer 1. For example, Optimistic Rollup is completed through economic challenges, while ZK Rollup is completed through zero-knowledge proofs.

1. My question is: How does the security of Rollup compare to Layer 1? For instance, if we develop MYKEY's smart wallet, can users deploy the smart wallet account directly on Layer 2 instead of Layer 1? For example, will the security of Optimistic Rollup be sufficient? Furthermore, in the future, if token contracts are deployed directly on Layer 2 instead of Layer 1?

Vitalik: This question is particularly important. The security of a Rollup involves two aspects. One is the security of the application, and the other is the security of the assets. Asset security means whether your coins or assets in the Rollup will be lost or whether someone will attack the Rollup to steal your coins; system security means that if you have an application on Layer 2, if the Rollup is attacked or has problems, even if the coins are still there, the application may not be able to do anything.

Asset security is basically not a big problem anymore. In ZK Rollup, there is no security issue at all. Each Rollup block has a zero-knowledge proof, so there is no possibility of putting an unverified block in. In Optimistic Rollup, there may be problems, but it is a matter of intervals. If there is a problematic block, and someone can discover it before a new block is generated, then there is no problem. So I think asset security is not a major issue.

The security of the system itself may be a bigger problem. Current Rollups are relatively simple. Current Rollups use a server to package blocks; if the server crashes, there will be a period of time when the Rollup cannot be used.

In the future, there are many ways to solve this problem. Many methods are being developed for decentralized servers; if one node does not send a block, the second or third can send a block. There are many methods. But current Rollups do not support these methods yet. We first need to solve the asset security issue, and the second phase of Rollup will address the systemic issues.

Guru: Can I understand it this way? At this stage, it is still not recommended to directly deploy, for example, smart wallet contracts solely on Layer 2, or to put coins directly on Layer 2? Perhaps in the second phase, it might be possible?

Vitalik: I think you can start putting some applications on Layer 2 now. Many blockchain applications do not need to run 24 hours a day; even if there are security issues, the losses will not be too great. For example, many non-financial applications can do this. For instance, using blockchain for some records, or NFTs can already be placed in Rollup.

The more dangerous things are particularly high-value DeFi items. If you put 100, 1000, or 10,000 coins in Rollup, there is not much problem. But if you start doing many transactions, possibly dealing with a billion every day, you may encounter server attacks. So my suggestion is that many things can be done on Rollup, but particularly high-value items should wait for a while.

2. Guru: Let's answer the first community question: Besides fast transfer speeds and low fees, are there other uses for Ethereum Layer 2?

Vitalik: Lower transaction fees are the most important.

3. Guru: The second question: Regarding using ETH as transaction fees to capture value, when a large number of transactions move to Rollup, can the value of ETH still be linked to the ecological value in the long term? I think this question is very good. Currently, some Layer 2 actually allow GAS payments with tokens other than ETH. I think this might be the question.

Vitalik: After Rollup, project parties will still pay transaction fees on the Ethereum chain. Users may have many ways (to pay their transaction fees on Layer 2 with other tokens), but for applications on Rollup, Ethereum nodes still need to send Ethereum transactions and still need to pay transaction fees on the Ethereum chain.

It can be confirmed that if the current transaction requires a fee of 100, after Rollup, it may only need to pay 1. Of course, different applications will have different efficiencies. Sometimes GAS may be reduced by 100 times, and sometimes by 500 times.

Because Rollup is more efficient and has lower fees, more transactions will occur on Rollup. If there are currently 1 million transactions on Ethereum every day, with each transaction requiring a fee of 100, if everyone moves to Rollup, there may be 100 million transactions daily, and even if each transaction has a fee of 1, the total transaction fees may be higher.

So ETH will still have fee value. ETH will also have other values. In short, Rollup does not threaten the value of ETH but will greatly help it.

4. Guru: This is a related question to the previous one: Will Layer 2 eventually siphon off the value of Layer 1? Will it develop into a new public chain?

Vitalik: It is always possible. But I think if that happens, the Ethereum community will not like it. The community will quickly move to other projects. Our community is currently quite strict. If you are doing a Rollup project, they will look at your code and details to know whether you are genuinely doing Rollup.

If it is genuine, they will go for it; if you claim to be Rollup but are not at all, they will know and will use other projects (voting with their feet).

So if a Rollup project starts by claiming to be Rollup but then moves to another chain (to create its own public chain), it poses a significant threat to the project itself.

The second question is that after we implement sharding, there will be a lot of shard data. The transaction fees for Rollup on the Ethereum chain may be lower. Therefore, I think there is not much greater value in turning a Rollup into a completely independent public chain. The connection value of the Ethereum ecosystem is greater; I think there is no benefit in leaving the Ethereum ecosystem.

Guru: If a Rollup turns into its own independent public chain, the community can actually fork it. The Layer 2 of Rollup is public data, and anyone can continue to run this Rollup.

Vitalik: Yes. A Rollup has a contract and a mechanism. If a Rollup project wants to create an independent public chain, the Rollup contract itself still exists. If Rollup users find that no new Rollup blocks are being generated, they can continue to run Rollup themselves.

5. Guru: The Layer 2 scalability solution is imminent, with projects like Loopring already online. Does Vitalik see potential in it?

Vitalik: I am particularly optimistic.

6. Guru: After the London plan is implemented in July, how much will each transaction actually cost?

Vitalik: Predictions are difficult. Because how much the transaction fee is depends on how many users there are, how well the technology for moving to Layer 2 networks works, and whether the GAS fees on the Ethereum network will be higher. The current GAS is consistently 12.5 million; it may increase a bit after the London fork due to ERP-2929 and increased security measures.

But as for how much it will increase, I don't know. After the London plan, transaction fees may be very high or may not be very high. If transaction fees are really too high, I believe that with the progress of Rollup in 4 to 6 months, many things can be done on Rollup, and transaction fees within Rollup will definitely be very low.

7. Guru: The next question: How do you compare the differences between Optimistic Rollup and Arbitrum Rollup?

Vitalik: Both teams are excellent. I think the biggest technical difference is that Arbitrum's mechanism directly manipulates EVM. Applications on Arbitrum can run the code directly. But Optimistic has its own EVM built into its contracts. So it does not run your EVM code directly but uses a built-in virtual machine, and contracts need to be executed in steps.

The advantage of Arbitrum's method is that it can support more VMs; it currently supports EVM and can support more programming languages in the future. But the downside is that in Optimistic, sending a transaction may be valid or invalid, and you can directly know the result. But on Arbitrum, the protocol requires you to first send a transaction, and then another user sends a transaction.

Transactions may need to be repeated three or five times, and then the built-in virtual machine needs to do asynchronous processing to determine based on the last step whether what you said is correct or what others did is correct. So this mechanism is more complex and requires more code, which makes me worry that using this method may have more security risks and issues. But this method is more universal and can support more programming languages.

Optimistic is simpler, and the code security is better. Arbitrum can support more VMs and more programming languages.

8. Guru: Will Optimistic Rollup and Arbitrum Rollup have their own native tokens in terms of economic incentives?

Vitalik: I don't know. I don't know how likely it is. We know that many projects have their own tokens, including Uniswap. But these two Rollup projects have not yet announced their economic incentive plans. So it is also possible that they will not have any tokens at all. We need to wait for them to announce what mechanisms they will have.

9. Guru: The next question: Currently, using the Optimistic Rollup solution is only temporary; will it eventually transition to the ZK Rollup solution?

Vitalik: ZK Rollup has a longer-term future because its mechanism and economic protocol will be simpler. But I do not know how long this long-term future will take; it could be 5 years or 10 years. The current technology of zero-knowledge proofs is relatively complex and novel. It will take a lot of time to develop a fully ZK virtual machine to ensure its security.

But once achieved in the distant future, ZK Rollup will have many advantages. I believe many applications will move to ZK Rollup.

10. Guru: If Ethereum 2.0 is launched too slowly, it may be surpassed by other public chains during this time. Why can't it be launched a bit faster?

Vitalik: Right now, short-term solutions to scalability issues really do not require 2.0; what is needed is Rollup.

Currently, Ethereum can support between 15 to 45 transactions per second depending on the type of transaction. After everyone moves to Rollup, the transaction volume may increase by 100 times. Depending on the specific transaction type, TPS may reach 1500 to 4000 transactions per second.

So even without sharding and Ethereum 2.0, Rollup technology can solve all current scalability issues. But in the long term, there are many public chains that can achieve TPS of 4000. However, if more people start using blockchain and it becomes more mainstream, we will definitely need more TPS. Because current mainstream centralized services require 10,000, 100,000, or more transactions per second. So in the long term, sharding is still particularly necessary.

The greatest value of 2.0 completion is the complete transition from POW to POS, allowing us to gain many advantages from POS. Our previous plan was to start with the 2.0 chain; we have already done this, and the POS chain has been running for three months with almost no issues. The second step in our original plan was sharding, and the third step was to merge the current Ethereum 1.0 chain with the 2.0 chain.

But now our plan is to first merge 1.0 and 2.0, and then do sharding. Because there are many communities (other public chains) that want to replace our ecosystem. Therefore, merging Ethereum's two chains back into one will completely resolve the issue of transitioning from POW to POS. The sharding issue can be delayed a bit and done later.

11. Guru: Currently, several Rollup networks have gone live. However, the funds migrating from users are not substantial. What methods can guide users to migrate?

Vitalik: There are currently two issues. The first issue is that many applications in the Ethereum community like to call smart contracts, such as DeFi projects. But currently, ZK Rollup does not support smart contracts; it only supports simple applications like issuing tokens and trading tokens. This is the first issue. When we have Rollups that support a complete EVM, I believe more users will migrate to Rollup.

The second issue is that the infrastructure ecosystem related to Rollup is still not very developed. For example, we have not yet solved the problem of transactions between different Rollups. If I have some tokens on ZKsync, how do I move the tokens to Loopring? First, I need to withdraw the tokens from ZKsync's Layer 2 to the Ethereum base layer, and then transfer them to Loopring's Layer 2 account. If this is done, will the transaction fees be particularly high?

Another issue is that many services do not yet support Rollup. For example, there are currently no exchanges that directly support ZKsync or Loopring for Dpos. But this will change; I already know that many exchanges are researching Rollup.

Vitalik: There will be many user experience challenges on Ethereum now. But I believe that many of these issues will be resolved in six months.

12. Guru: The next question comes from user skepticism. As the applications of Rollup Layer 2 increase, there will also be more Layer 2 projects, which will also become crowded. In the end, the issue of high transaction fees will still not be resolved.

Vitalik: If there is not enough space on Rollup, we will have sharding to expand the data space of Rollup. But I believe that the scalability of Layer 2 will not be an issue. When we need more scalability, the first step is to establish sharding, and then we can continuously upgrade the sharding; this is not something I am worried about.

What I am concerned about is that if the Gas fees on Layer 1 remain particularly high, there are many things you can implement in a Rollup, but if you need to conduct transactions across different Rollups, and you have to pay particularly high GAS fees, there will be unexpected situations that require you to call the base layer. At that time, transaction fees will become particularly high. This is a real issue. But we are also researching many methods to reduce this situation.

For example, one solution is that if a Rollup has issues, many users can simultaneously send a transaction, which can achieve calling the base layer transaction once to move users' accounts from one Rollup to another. This method can reduce the costs of Layer 1 network in unexpected situations. So there are many technical methods, but it will take a considerable amount of time to develop. But I believe we can achieve it.

13. Guru: The next question: Which Rollup projects do you favor?

Vitalik: I like all of them. Optimistic is good, Arbitrum is good, ZKsync and Loopring are also good. DeversiFi is not talked about much now, but I think this project is also very interesting. Currently, we have about five Rollup projects. But I know there are more teams working on Rollup.

14. Guru: I see many people asking questions related to EIP-1559. I think we can take some time to discuss your thoughts on EIP-1559. Many people are strongly opposed or strongly agree; there is controversy in the community. Please share your thoughts, Vitalik.

Vitalik: EIP-1559 is very important for the Ethereum ecosystem. It can solve many underlying issues for users. For example, users need to choose how high a transaction fee to pay. If they want to execute a transaction quickly, they need to pay a high transaction fee; if they choose a lower transaction fee, there is a risk that the transaction may take a minute, three minutes, an hour, or a day. EIP-1559 can solve such problems (making transaction fees more predictable).

EIP-1559 also has many other benefits. One frequently mentioned benefit is that most of the transaction fees will be burned. In this case, if the total amount of Ethereum transaction fees increases, the total amount of ETH will decrease. The economic model of Ethereum will become more sustainable.

Many people criticize Ethereum, believing that Bitcoin has a total supply cap of 21 million coins. But Ethereum's ETH has no cap, so they think ETH will be unstable. However, after EIP-1559, with more transactions, the total amount of ETH may decrease, or the total amount will only increase slightly (from inflation to moderate deflation—Xingzou's note). This change in the economic model is something many people are particularly looking forward to.

Many people were worried about EIP-1559 last year due to the risks associated with a new economic mechanism. But in the past year, we have conducted a lot of research, including our own research, reports from American game theory professors, and the EIP-1559 research from Filecoin, which everyone can see the results of. So by now, the risks of EIP-1559 have been greatly reduced.

15. Guru: Many people also asked about Ethereum's competitors and other public chain issues. I think they want to know how the competitive relationship between Ethereum Layer 2 technology and other smart contract platforms will change.

Vitalik: This is difficult to predict. I think the difference between Ethereum and many competing chains is that the TPS of competing chains will be higher. When they promote themselves, they will say that Ethereum's chain is slow, and our chain has higher TPS and is faster. But their choice has a downside: their chains tend to be more centralized.

On Ethereum's chain, if you have a computer, you can run your own node. But in these competing chains, the first issue is that they have fewer nodes, and the second issue is that the consensus algorithms are sometimes DPOS and sometimes POA, with very few nodes, possibly only 10, 20, or 100. Some applications do not require particular decentralization, and they may use other chains in the short term. But in the long term, when Ethereum has Rollup, Ethereum can achieve higher performance without sacrificing decentralized security and will have more scalability. I think when we have these Rollups, the competition between Rollups and other chains will be very interesting.

16. Guru: Let's ask the last question. If the layered network is very successful, is there still a need for sharding?

Vitalik: I think there is no need in the short term, but there is a need in the long term. The scalability brought by Rollup is limited. Each transaction on the Rollup side will still have some transactions occurring on the Ethereum chain (Layer 1). Some applications may improve performance by 100 times, 200 times, or 500 times. In the short term, this level of performance improvement is sufficient.

But in the long term, when more users start using blockchain, for example, if there are 500 million new Ethereum users, Rollup will not be enough. At that point, we need to combine both Rollup and sharding technologies.

Guru: Thank you, Vitalik, for answering many questions about Rollup and Layer 2. Thank you very much for getting up early to share in-depth with the Chinese community. I wish Ethereum a smooth progress in the Rollup Layer 2 network this year, bringing a qualitative leap to the Ethereum ecosystem. Also, thanks to Vitalik for his continuous hard work. I have gained a lot from this sharing.

Do you have any expectations for the construction of the Chinese community?

Vitalik: I hope that the expansion technologies like Rollup can receive mutual support from community users. Because the Ethereum ecosystem will face many challenges, including challenges from scalability, the complex process of a complete Ethereum 2.0 and POS, as well as security and privacy challenges. But we also have many excellent technologies and a great global community.

Therefore, I hope everyone can support and help each other so that we can build a better blockchain ecosystem that is accessible to more people.

Guru: We also hope that the entire Ethereum Chinese community can continue to unite as always, pushing exciting projects so that more people around the world can use them.

Disclaimer: The organizer of this article, "Xingzou," holds ETH spot and participates in Ethereum GPU mining.