Interview with John Adler of Celestia Labs on Sovereign Radio

Summary of the Interview with John Adler by Laura Shin at the last Modular Summit 3.0

📺Read the full interview here:

Laura Shin begins by introducing John Adler, Director of Protocol Engineering and co-founder of Celestia Labs.

Laura: Celestia has taken off since its launch last year. What are some of the accomplishments you’re especially proud of? And what have been some of the surprises since launch?

John: Yeah, I would say there are two achievements that I’m particularly proud of. One of them has already happened and the other one is about to happen. The first one is that a large number of people have been running light nodes on various devices around the world, even in limited internet radiuses. I think this is a huge milestone in the development of blockchains in general, in terms of how they are used. Because, other than Bitcoin, no other blockchain except Celestia has light clients that people can run on such cheap hardware and with so few bandwidth restrictions.

Laura: So a light client doesn’t contain the entire blockchain….

John: Correct. A light client doesn’t download the entire blockchain, it only downloads block headers. But more importantly, Celestia light clients don’t just download block headers, they also perform data availability sampling. This is kind of the “secret sauce” that allows light clients to be as secure as full nodes. They are not only light clients, but they also perform data availability sampling.

Some light clients, as I mentioned, are very popular in Bitcoin. Satoshi, in his design of Bitcoin, dedicated an entire section to light clients, called SPV (Simplified Payment Verification) clients back then. Basically, no other blockchain except Celestia has prioritized this type of client as a core part.

The second milestone, which is in progress but close to completion, is what we call the Lemongrass Hard Fork . It is Celestia’s first hard fork that introduces a number of new features and improvements, and lays the groundwork for future upgrade paths. This is the first network upgrade expected in the next month or so. I think it’s a great achievement for the entire team to be able to implement such a major change to the protocol, and to do so in a way that is reminiscent of the style of Ethereum’s Hard Forks, allowing for optionality and authority, and allowing users to choose whether they want to follow this new network upgrade or not.

Laura: Let’s take a step back and think back to before launch. What were the main problems you were trying to solve with Celestia?

John: The idea for Celestia, which was called Lazy Ledger back then , was created by my co-founder Mustafa Al-Bassam and me. It was a way to scale blockchains. The scalability problem is pretty generic, but at the time, a lot of people were thinking about how to scale blockchains by separating certain functions. For example, rollups were starting to gain popularity. The idea was to submit all the data to Ethereum, and verify the execution externally. This separation between the base layer, which is responsible for data availability, and the execution layer is crucial.

But even with these solutions, there were very significant restrictions on how much data the base layer could handle at the time. Ethereum had many limitations, which made continuing to use the base layer extremely expensive, albeit cheaper than other alternatives. Separating data availability from execution allows the base layer to handle only data availability, which is much more scalable and can reduce costs for rollups.

Laura: What was your experience before starting Celestia?

John: Before Celestia, I worked as a research engineer at ConsenSys , where I focused on Plasma and rollups. There I developed the idea of ​​optimistic rollups. After this, I developed Lazy Ledger because I saw that this was really the way blockchains should scale, through this modular approach. Mustafa approached me one day to ask if I wanted to start this new blockchain, and I thought, “Yeah, this is exactly how I think blockchains should scale.”

Laura: Well, Celestia is working on some new stuff now. There’s a new primitive called Lazy Bridging , which uses bead abstraction. Can you explain what it is?

John: Yes. Lazy Bridging is a ZK account application. First, I need to explain how it works in the context of Celestia. Celestia is a blockchain that only performs data availability and not execution. Unlike Ethereum, it does not have a smart contract execution environment. This means that you cannot directly verify that a rollup has been executed correctly, nor can you make deposits and withdrawals of funds from the rollup on the base layer.

This is similar to Bitcoin, which uses predicates on UTXOs to define spending conditions, Celestia can perform verifications instead of executions. This offers performance and security benefits. The idea is that, by not needing a generic virtual machine, only verification of specific proofs is performed. This verification logic is framed within ZK accounts, allowing for great flexibility without the need for execution on the base layer.

ZK accounts on Celestia function as a special type of account where changes require a valid ZK proof to be published and verified on Celestia. The proof is completely generic and what is verified is independent of what happens off-chain. This means that complex logic can be had, even without an Ethereum Virtual Machine, operating completely off-chain. Celestia’s base layer only needs to verify the proofs, allowing the account to update its state based on the verified transaction.

Laura: Can you give an example of how this applies?

John: For example, if someone wants to move a token from Solana to Celestia, they could use Lazy Bridging to issue a ZK proof that verifies the transaction. Celestia’s base layer only needs to verify this proof, without needing to know the specific logic of the transaction. This makes the process more secure and efficient. Additionally, by combining ZK accounts with account abstraction, atomic transactions can be performed between multiple rollups, such as withdrawing from one and returning to another, without the need to execute smart contracts on the base layer.

This offers a lot of flexibility without the need to add generic execution to the base layer. You can have all the functionality of a generic blockchain system like Solana or Ethereum virtual machine without complicating the base layer.

Laura: That reminds me, I don’t know if you remember, but when you had Google Authenticator, you had to update each code individually every time you got a new phone. Now, they update automatically when you change phones.

John: I don’t use Google Authenticator, but I can imagine. It’s a great example of how something that used to be a big pain is now much easier. In a blockchain system, it’s not only a pain, it’s also very expensive because you have to sign multiple transactions. With this type of key storage, you can do a key rotation with a single transaction.

Laura: And I imagine that using ZK accounts in that way allows you to keep the base layer lightweight and avoid a lot of hard forks to update functionality, you just need to update these external accounts.

John: Exactly. There is a lot of flexibility because the ZK accounts in the rollup can have arbitrarily complex logic. You could have the EVM completely operating off-chain and only need to verify proofs on the base layer.

Laura: Does this also address the fragmentation issues you mentioned earlier?

John: Yes, exactly. By allowing asset transactions and transfers to be done securely and without the need to trust third parties, fragmentation is significantly reduced. This allows for greater interoperability between different chains and rollups within the Celestia ecosystem, creating a more cohesive and secure environment for users.

Laura: But this only applies to state transitions and not necessarily to external assets, correct?

John: Exactly, this only applies to state transitions and not automatically to external assets. For example, if you have an NFT on Solana and want to move it into the Celestia ecosystem, you need a protocol like Wormhole. This can cause fragmentation as users must rely on the custodial power of Wormhole bridge operators and other bridge operators to move funds across chains.

Lazy Bridging allows for the creation of bridges that do not require trust in a committee or external validators. Everything is based on mathematically verified ZK proofs. This eliminates the need to trust third parties for custody of funds, significantly reducing the risks of hacks and security issues we have seen with other bridges. Thus, fragmentation is minimized by allowing chains and rollups within the Celestia ecosystem to interact securely and trustlessly.

Laura: How does this work in practice?

John: You can have a two-part bridge system: one part in the Celesita base layer that verifies the ZK proof, and another part in the rollup that implements the transaction logic. For example, if a valid ZK proof indicates a withdrawal, the base layer can process that withdrawal without knowing the specific logic. This allows for great flexibility and security, eliminating the need for a committee or external validators.

This would solve the bridge hacking issues we’ve seen, as long as the ZK logic is implemented correctly. You don’t need to trust a third party to secure keys or hold funds. Funds are essentially held by the state machine itself, following rules defined by ZK proofs. This allows for a trustless bridge — you just need to trust that your full Celestia node is operating correctly and that the state machine is running correctly.

Laura: Obviously, as you just mentioned, sharding is a problem that is solved. But we have also seen a lot of hacks with bridges. Does this solve that in any way, or at least minimize the centralization issues that led to those hacks?

John: Correct. Yes, as long as the ZK logic of the circuit is implemented correctly. You don’t have to trust a third party to secure your keys or hold custody of funds. The funds are essentially held in custody by the state machine itself and the rules of the state machine.

So you don’t trust anyone else to hold the funds. Obviously, if the state machine is hacked, similar to what would happen if there was a major bug in the virtual machine, you would expect the community to adopt a fix. Of course, you can’t force the community to adopt a fix, but they usually do. In the rare cases where there are bugs in the underlying virtual machine, the community usually comes together to fix it.

State machines are usually quite simple, which reduces the chance of errors compared to smart contracts, which can be very complex and difficult to debug. While it doesn’t completely eliminate problems, this system would significantly reduce the security issues we’ve seen with bridges in the past.

Laura: To make it a little more concrete for listeners, could you contrast the current experience with bridges compared to what it would be like with Lazy Bridging?

John: Yes, of course. The current experience involves choosing a specific bridge provider to move funds between chains. This can cause fragmentation issues, as different providers have different rules and levels of security. Additionally, there can be liquidity and custody issues, as well as risks of hacks.

In the future, with Lazy Bridging, there would be a canonical bridge that provides better security. Similar to how the canonical bridge between Ethereum and Arbitrum works, there would be no need to wrap assets through a specific provider. Assets would simply move from one chain to another without complicated prefixes or suffixes.

Laura: It definitely sounds a lot easier. Celestia is also helping to enable sovereign rollups . Can you talk about what these are and how they differ from smart contract rollups?

John: Yes. Sovereign rollups are a concept we are very excited about. Unlike traditional rollups, which rely on smart contracts on the base layer to verify correctness, sovereign rollups allow users of the rollup to determine its correctness. In sovereign rollups, correctness is determined by node operators and the community of the rollup itself. This is different from traditional rollups like Optimism on Ethereum, where correctness relies on Ethereum validators and the community to fix bugs.

In sovereign rollups, you don’t need a multisig committee for updates. They can be updated independently by the rollup community, without relying on the state machine or the base layer community.

Laura: What else is on that Celestia roadmap?

John: It’s divided into three areas. One is abundance, where we’re talking about a lot of block space. The other is verifiability, which means people can run light nodes. And finally adoption by a large number of users.

Big blocks : In our roadmap, we want to reach 1 Gigabyte blocks every 12 seconds, which is a magnitude increase compared to the current 2 Megabyte blocks every 12 seconds. This represents a significant increase in performance compared to other blockchains.

Verifiability: We want users to be able to run light nodes very cheaply, so that everyone has access to high assurance of data availability. We are developing light nodes that can be run in the browser, using WebAssembly (Wasm). This allows users to verify not only the header string, but also the availability of data directly in their browser without needing to download any additional software.

This means that eventually, wallets will be able to connect to these light nodes in the browser, instead of relying on centralized providers like Infura for Ethereum.

We are also working on several improvements such as header pruning, lock running, header drilling … , which will reduce the cost of running a light node. This will make it more accessible to a large number of users. Also, as the number of light nodes performing data availability sampling increases, we will be able to increase the block size while maintaining the same security.

Laura: That definitely sounds like a glimpse into the future. It’s been really exciting chatting with you. Thank you so much for sharing your current work with Celestia and your vision for the future.

John: Thank you, Laura.

Twitter | Medium | LinkedIn | Discord | Telegram | cumulo.pro