Celestia AMA

Celestia is a minimalistic modular consensus and data network that enables anyone to effortlessly deploy their own blockchain with minimal overhead. By decoupling the consensus and application execution layers, Celestia modularizes the blockchain technology stack and unlocks new possibilities for Web3 builders. Celestia offers developers the freedom to design each application’s sovereign execution environment while still maintaining secure interoperability for all Celestia chains. With a current focus on EVMs and the Cosmos SDK, they will support all rollups in the future. 

Nick White, Co-founder of Harmony and COO of Celestia, joined us for an AMA on February 9th.

vVv: In this session, we chat about blockchain modularity and the implications it poses for all our on-chain processes. Nick, it’s a pleasure to meet you.

Nick: Nice to be here, guys. Thanks for hosting me so we can talk about modular blockchains and all that good stuff.

vVv: First, can you give us a high-level introduction to Celestia and your background?

Nick: I’m an engineer by training. I started with electrical engineering and studying things such as neural networks back in the day.  I discovered crypto in late 2016 and fell into the rabbit hole throughout 2017. In 2018, I finally made the jump from the AI industry into crypto. That’s when I co-founded a project called Harmony. The reason I was working on Harmony was that I believed that the core infrastructural problems in crypto needed to be solved. 

We needed scalable useful infrastructure for the rest of the crypto movement to flourish. At Harmony, we were trying to solve scalability with sharding and by building proof-of-stake. I had a great journey for two and a half years, but in late 2020, I discovered this new idea through a white paper called “LazyLedger.” That whitepaper introduced me to modular blockchains. I realized very quickly that modular blockchains were the answer to many problems in blockchain infrastructure, including scaling, bridging, flexibility, and sovereignty over execution environments. I was fully converted. I stepped back from Harmony and joined LazyLedger, now transformed into Celestia. At Celestia, we’re now building the first modular blockchain network where we provide data availability and consensus. This provides a scalable foundation of security for people to build blockchains on top of in the form of a roll-up. Modular blockchains were the answer to many problems in blockchain infrastructure, including scaling, bridging, flexibility, and sovereignty over execution environments.

vVv: Can you explain the differences between a modular and a monolithic blockchain approach?

Nick: Monolithic blockchains are chains everyone is familiar with, such as  Bitcoin, Ethereum, Solana, and Avalanche. Essentially every existing Layer-1 blockchain is a monolithic blockchain. Blockchains have a certain number of core functions to perform to be secure and usable. Monolithic blockchains perform these within one network, one set of nodes, and one protocol. They do all these things at the same time, which imposes constraints on aspects like scalability, application development, and many others. A modular blockchain, on the other hand, is when you return to first principles and realize that functions can separated. And specifically, that there’s no reason for bundling core functions together in one protocol. You can unbundle functions and delegate them to specialized protocols. 

In that case, you can reintegrate functions achieving the same functionality but with much better properties. You get increased scalability, flexibility, secure bridging, and many other new benefits that were not possible before modular architecture was invented.

vVv: Would an appropriate analogy be that the monolithic approach is akin to a single computer with simultaneous applications, whereas a modular approach is like AWS or other cloud-based services?

Nick: A more technical analogy that I really like to use is of a monolithic blockchain being a pre-packaged laptop or a stand-like desktop computer. It has a certain amount of memory and processing power, which will not change. That computer also comes preloaded with an operating system that defines what kind of applications you can build, and you’re stuck with that. Comparing that to a monolithic blockchain, you get a finite amount of blockspace and a finite amount of transactions it can process. It also comes with a preloaded virtual machine like the Ethereum or Solana VM, which also defines and restricts the kind of applications you can build. In the same analogy, a modular blockchain is where the hardware is not fixed or finite. It’s more like a cloud computer/computing type of system.

A data center, for example, is a bunch of raw computer parts that are all aggregated together. You can continue to add more and more capacity by adding more hardware and more machines. And unlike your laptop or desktop, there’s no fixed capacity. When there is more demand, the capacity can be increased. In a modular blockchain, the blockspace is also not finite. It can grow with the number of nodes within the network. That’s powerful because when demand increases, but the blockspace doesn’t, like in the case of Ethereum, this means that fees have to increase, which causes a bad user experience. The other important part of a modular blockchain, just like in a cloud computing setup, is that the computing elements are raw and don’t come preloaded with any operating system. As a developer, you can now run whatever operating system you want. In blockchains and with Celestia, that translates into flexibility and running any VM or application you want, even ones that people haven’t invented yet.

vVv: We think that’s impressive and necessary for the mass adoption of blockchain technology. Having that scalability and flexibility empowers developers to create high-performing and secure applications. Can you give our listeners more background on modular architecture and the different layers? What are the roles of Celestia’s execution, settlement, consensus, and data availability layers?

Nick: For this, I like to use an analogy in the context of a soccer game. In a blockchain, it’s important that there are a set of rules that are enforced and nobody can break them. In Bitcoin, you want to ensure no one can print more Bitcoin out of thin air, and in Ethereum, you want to ensure that the rules of the virtual machine and contracts are also enforced properly. In this soccer analogy, there are rules. If the ball goes out of bounds, you have to stop playing and the ball has to be retrieved. If you score a goal, you cannot be off-side, and only the keeper can use his hands to prevent a goal. Thus, there are a whole set of rules to make the game of soccer fair, and enforcing the rules is what we call execution in the blockchain environment. Execution is knowing what the state of the blockchain is like. It’s knowing who has what amount of tokens in their account, or in a soccer context, knowing who scored how many points. Execution is what ensures validity, and that the rules are followed.

The consensus layer ensures that the order of events is consistent across everyone. Everyone in the stadium that is watching the game needs to be able to agree on the order of events that happened. If that doesn’t happen, how do you know what’s fair and what’s not? If you’re not watching the exact same game, then it’s hard to agree on what the outcome is. The consensus layer ensures nobody can rewrite history.

Another important layer is data availability. This layer ensures that the people are actually able to see the game and see what’s happening. What if somebody tells you a point was scored? You have to be able to see it for yourself, and see the replay, or watch it happen live. The user/spectator has to see what happened to be able to verify that nobody broke the rules. There’s a class of attacks called data withholding attacks where someone will say something has happened, but they prevent you from seeing it and say you have to believe them. Data availability is making sure that the game or state of the blockchain is visible for everyone to see.

To extend this analogy into the monolithic versus modular discussion, I’d say that a monolithic blockchain is like a soccer game where there’s a limited number of players or games that can be played, as there is a limited amount of blockspace. When more and more people are trying to join the game, you will eventually run out of capacity. It also only knows how to play one game in one execution environment. Therefore, it can only play soccer, for example. In a modular blockchain, there’s no limit, and no fixed stadium size. The more people that are watching the game, the more nodes are running, and therefore the increase in participation capacity. A modular blockchain in the sport context would also not be limited to only one sport and therefore, not only soccer. You can play basketball, tennis, or any other sport that has yet to be invented. Modular blockchains can be compared to the Olympics as they can grow to a much greater scale and host a greater variety of different games that can be played.

vVv: In the blockchain space, we often hear about the blockchain trilemma. The modular blockchain approach could solve the trilemma regarding scalability, decentralization, and security. Can you expand on that? 

Nick: Absolutely. Since the early days of Bitcoin, there have been a lot of discussions and conversations on this very hot research topic of how to solve blockchain scalability. People have explored every trade-off in the space and eventually concluded that there was a trilemma. This meant that there was this kind of insurmountable tradeoff between security, decentralization, and scalability. It started to feel like that was true, and people started to accept that as the future of blockchain. This was the case until modular blockchains came about.

Mustafa (Al-Bassam), who invented modular blockchains and is the co-founder of Celestia, ended up solving this trilemma by taking a completely different approach. An approach that some people even thought realistic as they were designing and thinking about blockchains within a monolithic frame of mind. But, if you take a step back and start again from the first principles, then you can flip that trilemma on its head and break it open by building in a whole new design space. That’s where modular blockchains come into play. There are a lot of different reasons why this is the case. But, when you try to force everything into one protocol, the nodes have to do too many things at once. Those protocols can’t be specialized to achieve a certain function. By splitting the stack apart, you’re able to specialize and essentially break free from a lot of constraints.

There are two enabling technologies in this modular paradigm. First of all, there is the core insight of splitting execution into a separate layer, apart from the consensus and the data availability layer. But to make that scalable, there are two technologies. The first is data availability sampling, which solves the problem of scaling blockspace at the Layer-1.

You have to think about how to make it possible to add more and more transaction throughput without making it more expensive to run a node that verifies the blockchain. 

It’s very important, and it is the basis of blockchain where the end users can verify everything that happens. Mustafa co-authored a paper with Vitalik in 2019 describing a method called data availability sampling. With that, you can actually increase the size of a block, while enabling end users to verify it by randomly sampling from the block, rather than having to download the full block. Downloading a full block becomes very expensive as the block gets big. That allows them to verify with very minimal resources.That is how solving the first half of the puzzle can be done, which is scaling the underlying throughput of a layer-1 blockchain in a decentralized way. The second half is doing off-chain execution. In other words, securely running a Layer-2. This problem was solved by many different people, but there was a major breakthrough by another co-founder of Celestia, John Adler. He came up with a construction called a rollup, which outsources consensus and data availability to an underlying Layer-1, but uses proofs to ensure that the execution of the rules of that chain is followed. There are two primary proof approaches. One is fraud proofs, and the other one is validity proofs. A rollup uses those to ensure that people can verify that the rules were followed and no one broke them. With fraud proofs, someone will send you a message that you’ll be able to verify, telling you that it’s a fraudulent block and you shouldn’t trust it. Validity proofs, known in the zero-knowledge rollup context, makes sure that every new block that is mined, comes with a proof of validity. With this, you don’t have to trust anyone that it’s valid. You don’t have to re-execute, just verify it yourself. The combination of those two evolutions is what powers this new modular blockchain paradigm, and makes it so we can have these very scalable and secure blockchain systems.

vVv: In data availability sampling, can everything be performed on a simple smartphone, or is specific technical hardware required?

Nick: That’s a very good question because another major difference between modular and monolithic blockchains is that modular ones cut the cost to verify the blockchain. Even Ethereum, which people consider quite decentralized, requires good hardware to run a full node and verify the blockchain. It used to be a laptop, but at the moment you even need more computing power. In Solana’s case, you need a supercomputer. All these things make it expensive and technically complex to verify the chain, which is problematic because blockchain security relies on the fact that people are verifying the chain themselves. That’s what makes it decentralized and what forces the rules.

The cool thing about modular blockchains, thanks to rollups and data availability sampling, is that you can run a full node and verify the chain directly on very low-resource hardware. It can even be your mobile phone. We’ve already done demos and have documentation online of how you can run a Celestia node on your phone. At this moment, it has to be on an Android device but you can run a node and verify the chain directly. That’s something that is not possible on any monolithic chain because it requires too much processing power, bandwidth, and other resources. It’s a beautiful thing because anyone with a smartphone, which is most the people on the planet, will be able to be a first-class citizen of modular blockchains and verify the network themselves. That is what makes blockchains secure and what we are trying to achieve. It’s a key part of our vision.

vVv: Not many can say  that they are running a node on their smartphone. Does this imply a higher risk of cyber attacks as the entry cost is lower or are those nodes not incentivized? Can you elaborate on how that works?

Nick: Great question. These nodes on phones are not participating in consensus. The term “node” tends to overloaded the general discourse. I even wish it wasn’t a term. But I would say there are consensus nodes, and there are full nodes. Consensus nodes are the ones participating in voting on blocks and coming to a consensus on them. Conversely, you have the nodes that just observe the network and verify it. The whole point of consensus is to design Sybil-resistant. That’s why you have proof-of-stake and proof-of-work to solve the Sybil Attack problem. These nodes, however, do not participate in consensus. You’re not incentivized for these, meaning that you’re not earning staking rewards by running this kind of node on your phone. Your incentive for doing so is that you get to verify the chain and make sure that everything is valid. We plan to have these sorts of nodes embedded directly in your wallets and in your browsers so that users are running them by default. Right now, the default is that people use RPC endpoints like Infura when they are using Ethereum. Rather than running their own node as it’s too expensive, they are relying on a trusted third party that tells them the state of the chain. That’s fine, because we’re so early, but in the long-term, that’s not a sustainable path, nor is it in line with the true values of crypto.

vVv: The first time I heard about centralized node access on Ethereum and other chains, I was shocked to hear that no one was talking about this central point of failure. No one considers it until the network is down. And this was the exact scenario a few months ago with Infura when dApps suddenly stopped working because everyone was connected to the same nodes. It’s important to highlight these features of Celestia and the whole modular blockchain ecosystem. Going back to modular architecture, you’ve touched upon developers being able to build whatever they like. That brings up the question of how you can ensure the compatibility and interoperability of those different modules when they interact with each other.

Nick: A big part of what is still being worked on in the modular space is defining these interfaces between the different layers of the stack. We anticipate that there will be multiple data availability layers and many different kinds of execution environments. A lot of people and projects are actively building on execution and settlement layers in the Celestia ecosystem. 

We didn’t talk about settlement yet, which is important. As we build the stack, all those pieces must be easy to mix and match. The end vision resembles a stack of Legos, where everything fits together and can easily be swapped out. This way, you can combine the best features of all the components you want as a developer or an application builder. We’re taking a stab at that. But, it will ultimately be a community effort to define those things. For example, there’s already been some work on defining the sort of interface between consensus and execution in the Cosmos stack. They use an abstraction called the Application BlockChain Interface (ABCI) which turned out to be specific to Tendermint and runs an active consensus process. In a rollup execution layer, it might not need a similar sort of abstraction. It could even be more simple. Some cool ideas are floating around, but I think it will take some time before anything solidifies. In regards to Celestia and the interface with the data availability layer, we have an API that developers can just plug into and post data to Celestia. It’s a very simple API because it’s important to make it as easy to use as possible.

The end goal for Celestia is that developers don’t have to think about consensus or data availability. Celestia will provide that for them in a very frictionless, low-overhead, and cost-effective way. There are also possibilities, such as building a sovereign rollup, or building a rollup that plugs into a settlement layer, which then plugs into Celestia. There are a huge amount of different configurations that are still being built and explored. It will be an exciting couple of years as all these experiments are tested.

I’m not a maximalist, but people should be free to use what works best. That’s a big difference between the modular and the monolithic blockchain vision. In the monolithic one, these decisions are made for you, and all these chains and ecosystems argue why their system is the best. In a modular vision, it’s instead very user oriented. It’s about the different things the user can do and which tools and components they can pick from. They should be able to choose what is best for them. Nobody else except the end user/developer will know what will work best for them. It’s not like there is one chain to rule them all, on one execution layer. It’s all about the different features and trade-offs.

vVv: This reminds me of the Apple versus Microsoft discussion. You have the Apple approach, which is quite closed and defined, where you cannot do many modifications, but it runs as it should. The other approach is where you can exchange your hardware parts and modify whatever you like, often ending in blue screens of death and vulnerabilities. Do you see more risks of opening up attack vectors between those different layers because you have more interfaces than a monolithic approach?

Nick: I don’t think the interfaces will be an attack vector. But you do have to ensure that each component is secure in its own way. It should be designed as it claims to do. That will be increasingly important as more projects come out at varying stack layers. You could have someone who is “selling” you their rollup, but it turns out that their fraud proof system is not secure. Maybe you have a data availability layer where they make claims but don’t have key things like block reconstruction, or they have bad encoding proofs. You have to make sure that the underlying technology is sound and secure. But as I said, I don’t think interfaces will be a problem. The difference regarding the Apple versus Microsoft context is that the end user is a consumer who likes mixing and matching their own thing. The developer who’s building these things should be very informed on all the different components. In the same way, somebody who is building a computer should know what kind of circuit boards, battery components, and so on he selects as he is building out his computer. The modular blockchain stack is all about letting developers mix and match components to build the blockchain that the consumer will use.

vVv: If developers can mix and match, isn’t the settlement layer also an important part? What role does it play in the modular stack or in the rollups being built on top of it?

Nick: A settlement layer is something optional within the modular stack. It’s to be seen how important and adopted these settlement layers will be. Settlement layers provide a few different functionalities. One of the important ones is that for some types of rollups, such as the optimistic rollup, it can be useful to have a chain where you post your fraud proofs. With this, you end up having this interactive fraud-proof mechanism that needs a chain to host that interactive game, for example. A settlement layer provides that. It can be very important to the security of optimistic rollups.

Another thing settlement layers provide is liquidity bridging. The layer can be some sort of bridging hub where instead of each rollup having to have a bridge, it can be one bridge in between every rollupas a shared layer. By bridging the settlement layer, they’re one hop away from every other rollup. It could be a good place to pool liquidity or issue assets for example. Another cool thing about a settlement layer is that it can provide an exit mechanism for different rollups. If things turn out to go bad on your rollup, you can push the eject button. But, there’s a whole class of rollups that do not need settlement at all, called sovereign rollups. I expect those to become very popular in the long run. It all doesn’t really matter that much because it’s all modular and it’s all about which features are the most important.

vVv: Your co-founder and inventor of modular blockchains, Mustafa Al-Bassam, shares the same vision. What can you tell us about him?

Nick: Mustafa’s whole life story is extremely interesting. He was born in Iraq and emigrated to the UK when he was very young. He taught himself how to code at a very young age and very quickly learned how to hack. As a teenager, he ran a hacker group called LulzSec and was also involved in the hacker group called Anonymous. They hacked some entities such as the FBI, and different multinational companies like Sony and Fox, where they exposed things like corruption. It was a really important chapter in the history of hacktivism. The cool thing about Mustafa is that what motivated him was showing that technology can equal the playing field when power is being abused. He believes that technology can be used as a tool to fight corruption and raise up the people who are less advantaged. That kind of continues in his work for Celestia. His vision with Celestia is to empower anyone to build their own blockchain and build their own sovereign community. Ultimately, I think that is going to be one of the things that helps transition the world into a more equitable place, where people are not beholden to untrustworthy institutions or bureaucrats dictating the show. They’ll have the ability to sort of opt out and build their own system, thanks to modular blockchains and blockchain technology in general.

vVv: Something we also really like about Celestia, is the fact that you have such an all-star team. You mentioned John Adler, with whom we also had as a guest for an AMA session. Can you share a little bit about the founding story of Celestia?

Nick: It started when Mustafa published a white paper called LazyLedger. John replied very quickly and praised the white paper. They started chatting more and more and decided that it was something they wanted to build together. In parallel, Mustafa met a guy called Ismail, our CTO and co-founder. He had previously been working on the Cosmos ecosystem building Tendermint and a lot of the Cosmos software. The three of them got together and realized that they really aligned around the vision of a modular blockchain ecosystem. It was then that the LazyLedger white paper turned into an actual project. They started in 2019, which was in the depth of the bear market. It was extremely difficult to fundraise at that time, and everyone back then thought that Ethereum was the one chain to rule them all. It was hard for them to fundraise for a Layer-1 blockchain, especially one that had this crazy idea that most people didn’t understand. Even the word data availability was something most people didn’t know about. It might still be a relatively unknown term, but at least now most researchers and people who have been in the space long enough have at least heard about it. Back then, it was something nobody understood or wanted to invest in. They started very humbly and it was very hard to get going at first. Later, around 2020, I came into the picture when I read that LazyLedger white paper. I couldn’t stop thinking about it for two weeks straight; it was like electricity had zapped me. I was electrified with the beauty of the vision. I remember the first message I sent to Mustafa where I told him I had to build it with him. I needed to be part of history and what could maybe be the biggest idea since Ethereum. Eventually, I ended up joining full-time.

vVv: Glad you decided to join the team. It’s one of the most exciting things we’ve seen in the space in the past two years. Talking about the future of blockchain, there are many different opinions about the approaches and modularity. On one side you have people like Kyle Samani of Multicoin Capital who believe that only a few monolithic blockchains will prevail, and then you have people like Jake Zeppelin who are betting on the modular future. What is your stance on the future of blockchain? Will it only consist of modular blockchains, or can they coexist with monolithic ones?

Nick: I think it will be a gradual evolution and migration. There already are some modular blockchain projects running such as Arbitrum, Starkware and Optimism who are building in this new framework but they are currently in the minority. However, I expect more and more developer and user activity to migrate to modular blockchains because they will have so many advantages that these developers and users will care about. One reason is that they will make blockspace abundant with low fees. As a developer, you will also have freedom and flexibility, which is very appealing. It will enable people to easily build things like appchains (application-specific blockchains) and have a lot of advantages over building an application as a smart contract on a monolithic blockchain like Ethereum or Solana. For all those reasons, I expect modular blockchains to eventually dominate, but it will take time. I’m not an absolutist and I could see activity remaining on monolithic blockchains for years to come.

vVv: To wrap up this AMA, I want to inquire on behalf of our listeners, how they can support Celestia’s vision?

Nick: First of all, if there are any builders in the audience and you have ideas of applications you want to build, or you want to participate in the development of the core protocols or any other aspect of the modular stack, then we welcome you. That would be the most important thing we need. Right now, we have an engineering problem where we need to write lots of code and solve a lot of research problems. This modular movement is very technical in nature and will require a lot of hard work and engineering. Blockchain is still in the infrastructure phase, so the more infrastructure builders we have, the better. It is however starting to reach a point where modular blockchains are actually usable by application developers. We have some teams already doing that. If you’re interested, we would love to chat and build together with you. Nothing is more satisfying than seeing your stack get used. 

Beyond that, the general public, investors and enthusiasts have a lot of different things they can do. One is to spread the word about modular blockchains and educate other people about them. Many people are not aware of this approach and it’s a huge opportunity to get ahead of the curve. It also forces people to learn about the fundamentals of blockchain, and that’s a very important thing for people who invest in it.

Lastly, when the network goes live, you can run a light node. The amazing thing about Celestia is that running a node is not just verifying the network, it’s also helping secure it. By running a node, you’re going to add more security and also more throughput capacity to the network. It’s kind of like a BitTorrent network where the more people seeding content, the more content can be stored by the overall network. You can actively participate in Celestia, even as a normal node, which is not the case in monolithic blockchains.

vVv: Thank you so much for your time. If our listeners want to follow Celestia’s development and community, where can we find that information?

Nick: We are active on our main Twitter page called CelestiaOrg and you can also visit us on our website at Celestia.org. On the website, you’ll find links to our Discord, Telegram and Reddit. There are lots of ways to engage with us and the community. If you have any questions I would recommend you to ask them on Twitter and tag myself or CelestiaOrg. We would love to answer your questions. We also have a lot of resources on our website if you’re into learning more about the technology. There’s a section on our website called Learn Modular, which has a bunch of articles where you can deeper into understand how modular blockchains work, and what the different words mean. We also have a series of talks called Modular Insights, where we have different projects throughout the modular ecosystem presenting what they’re building. That’s also another way to get on the bleeding edge of the research and innovation. We also have Celestia spotlights where we interview the founders behind these projects. There is a lot of different content to consume and hopefully, people enjoy it, ask questions and spread the modular movement in their communities.

LISTEN TO THE SPOTIFY RECORDING BELOW FOR THE FULL CELESTIA AMA