Difference between Modular & Monolithic Frameworks, and Issues Related to Rollups

Key Takeaways

• The difference between modular and monolithic blockchains is whether they divide up the task of processing transactions or do it all at once.

• A monolithic blockchain handles consensus, computation, persistence, and data availability - the roles of today's blockchains - all at once, while a modular blockchain splits these roles between two or more layers.

• Both monolithic and modular blockchains have their advantages and disadvantages, so there's no one methodology that's superior.

• People have different definitions and ideas about what a modular blockchain is, so there is still a lot of learning to be done.

If you're new to the blockchain industry, you might think that modular blockchains are the way blockchains have always been designed. However, they're a relatively recent framework. Rather, the traditional way blockchains have worked is Monolithic. We'll get into the differences between modular and monolithic blockchains in more detail later, but as a quick refresher, the former is where all work is distributed among different blockchains, while the latter is where all work is done in a single shard. Based on this simple analogy, you may feel that the comparison is unfair because modular blockchains sound more efficient in every way, but when you dig deeper into the details, there are clear advantages and disadvantages to both design approaches. Therefore, in this article, I will analyze and compare the differences, advantages, and drawbacks of modular and monolithic blockchains to help readers understand them more quickly(and easily, of course) when they come across these words in KBW(Korea Blockchain Week).

1. Monolithic Blockchain

"Monolithic blockchain" may sound like a grandiose term, but most of the blockchains we have seen so far actually have a monolithic structure. For example, Ethereum, which is now at the center of modular blockchain ecosystem, has a monolithic blockchain structure, and apart from Ethereum, blockchains such as Sei, Solana, Aptos, and Sui have a monolithic blockchain structure. So what is the structure of a monolithic blockchain? The blockchain is basically responsible for consensus, execution, settlement, and data availability, and the detailed description of each role is as follows (these roles will be applied later when explaining modular blockchains as well):

1. Consensus: Responsible for determining the order of transactions to be included in newly produced blocks.

2. Execution: The layer that literally executes transactions, and is responsible for computing network state.

3. Settlement (as we'll see below, mentioning a settlement layer is now a bit controversial. Forgive me mentioning this James Prestwich): acts as a hub for the execution layer, validating transactions and allowing liquidity to be shared between compute layers built on a same settlement layer.

4. Data Availability: It is responsible for ensuring that all transactional data used in the course of computation is accessible, storing that data and ensuring that it is always downloadable.

1.1 Advantages of Monolithic Blockchains

Monolithic blockchains address all of the features I mentioned above in one place, and the advantages of this structure are very clear:

• Independence: One blockchain handles consensus-execution-settlement-data availability, so there are no dependencies on other external blockchains (Although there may be some dependencies on external blockchains when oracles and interoperability come into play, this discussion is limited to dependencies that are necessary for transactions to occur.). Modular blockchains require the premise that each blockchain has a different function, so the division of labor requires that all of them do a good job in order to achieve a complete transaction, but monolithic blockchains do not need to care about the status of other blockchains as long as they do their job well.

• Compatibility: Since monolithic blockchains process all transactions on a single shard, all applications can be run on a single blockchain smoothly, "assuming that it achieves scalability and decentralization at the same time". Modular blockchains, on the other hand, require assets on Layer 1 to be moved to Layer 2(need bridge which I will explain about this later) in order to use applications on Layer 2. Monolithic is designed to handle various applications even if they run on a single blockchain, so users can use various applications without any hassle.

• Simplicity: Monolithic blockchains only need the validator that manages them at the infrastructure level, whereas modular blockchains are structurally and politically complex, with many different layer 2 blockchains (and now layer 3) and many different stakeholders involved in them. Monolithic chains are structurally and politically simpler than modular blockchains. This sometimes creates a user-friendly environment and sometimes a developer-friendly environment.

I consider these three things to be advantages of monolithic blockchains, and I still believe that the "ideal blockchain structure" is monolithic. Keep in mind that modular blockchains were created not because people like modularity, but because monolithic blockchains had trouble scaling(Monolithic blockchain still has this problem).

1.2 Drawbacks of Monolithic Blockchains

On the other hand, however, there are still many problems with monolithic blockchains, which is why modular blockchain structures are gaining traction. So, what are the problems with monolithic blockchains?

• Scalability limitations: As I write this, numerous consensus algorithms are being developed with the aim of handling as many transactions as possible on a single shard. However, unlike a modular approach where an infinite number of layer 2 chains can theoretically be added, achieving scalability technically is very challenging. This is because validators must fulfill many roles and always maintain a single shard.

• Monolithic governance: Of course, this may change later on depending on how the governance design is done, but so far the governance of monolithic blockchain has not been diversified. Also, in the case of general-purpose blockchains, governance decisions need to be made very carefully, because if done incorrectly, governance decisions can make the blockchain sector-biased (e.g., a governance decision for DeFi applications can be a bad decision for NFT applications). For these reasons, governance on monolithic blockchains is often not as active as it could be.

• Security: Monolithic blockchains must establish their own security measures, making it difficult for everyone to achieve strong security. Additionally, poor governance can lead to a hard fork that separates security levels, which can have a significant negative impact on the ecosystem.

• Generality: Aiming for general-purpose blockchains may result in an environment where not all applications can work well together. To address this issue, application-specific chains using the Cosmos SDK have emerged as a solution. While modularity allows for layer 2 or layer 3 chains to become application-specific chains, there are clear limitations in the case of monolithic blockchains.

So far, we've discussed the advantages and disadvantages of monolithic blockchains, but what about modular blockchains, which have emerged as an alternative to monolithic blockchains?

2. Modular Blockchain

Modular blockchains divide some or all of the four functions mentioned above. Celestia is a prime example of such blockchains. It will be presenting a session at KBW. While many people may think of Ethereum as the first modular blockchain, it is, in fact, a monolithic structure. Celestia, on the other hand, was built for modularity from the ground up, making it the first modular blockchain.

As shown in the figure above, Celestia is a blockchain that is responsible solely for data availability and consensus. This approach provides flexibility in terms of settings and execution layers. Of course, modular blockchains have their pros and cons.

2.1 Advantages of Modular Blockchains

• Scalability: Assuming that scalability is achieved through layer 2 rollups, scalability can be nearly infinite. Rollups bundle transactions from the rollup chain and send them to the main chain as a single transaction. Therefore, creating more layer 2s and attaching them to the main chain theoretically solves the scalability problem. Additionally, rollups provide better interoperability by building a stable bridge between computation layers that share a settlement layer.

• Flexibility:In a modular blockchain structure, rollup chains can be seen as parasitic on the main chain in terms of structure, but they can also have their own independence. Rollup chains can choose their own programming or smart contract language, even if you choose Ethereum as your main chain. This means that you can choose other virtual machines and smart contract languages besides EVM and Solidity, and define the purpose of the chain accordingly.

• Flexible governance: Modular blockchains partition functions across chains, allowing for the creation of efficient governance structures tailored to fit the purpose of each chain. For instance, the computation layer, the layer facing the actual application, can employ a more aggressive governance approach to ensure efficient application development. Conversely, the consensus layer, where security is paramount, should minimize governance intervention to maintain the overall neutrality of the modular blockchain. This flexibility is impossible to achieve in a monolithic blockchain, where everything must be solved on a single shard.

2.2 Disadvantages of Modular Blockchains

However, even modular blockchains, which have infinite scalability and flexibility while leveraging the security of the main chain, have their drawbacks:

• Complexity: Scaling to infinity means that there can be many different rollup chains. However, managing multiple shards or chains in a modular structure can be quite complex, unlike a single shard where everything is done in one place. This complexity is not only an issue for users, but also for builders.

• Dependencies: At the end of the day, modular blockchains are also seen as a disadvantage because they depend on a main chain and are not completely independent. Additionally, the interdependencies that come with the division of labor can also be seen as a disadvantage of the modular structure. All layers, such as execution, settlement, consensus, and data availability, must work well together for a single complete transaction. This means that modular blockchains cannot create a complete transaction on their own.

• Incompatibility: Of course, if they share a settlement layer (Jaems Prestwich has told people to stop using the word settlement layer, but I'll use it for simplicity's sake), then interoperability between chains will be more natural, but there will be many cases where they don't, and if they can't communicate well, it's hard to expect rollup chains to be compatible with monolithic blockchains. Monolithic blockchains are easier to work with because all transactions literally occur on a single shard. This is why Solana emphasizes composability between smart contracts.

3. Recent discussions around modular blockchains

As I summarized above, modular blockchains are not a perfect improvement over the existing system (monolithic blockchains). Modular blockchains also have many drawbacks, and many researchers and developers are working on improving them. It's interesting to note that even though the concept of modular blockchains has been around for a few years now, there is still some disagreement about how they work. This is when Jon Charbonneau, a well-known researcher from Delphi Digital, wrote a series of articles summarizing his thoughts on rollups, which have draw a lot of attention in the industry. Of course, before Jon, James Prestwich's "Stop saying 'settlement layer'" was essentially the beginning of the discussion about rollups and modular blockchains. Their argument, in a nutshell, is that there is a disconnect between how modular blockchains and rollups work in theory and how they work in practice, and that they are still a fiction that exists only in theory. James Prestwich is a developer known for being a vocal critic of the blockchain industry, and last year he claimed that there is no such thing as settlement layer, that in fact all rollups do not rely on a chain, which they call a "settlement layer", to ensure finality, and that the settlement layer is just a marketing term(also known as a buzzword). To summarize their argument, here's what they said:

3.1 Bridges are not rollups: The 7-day dispute period is a fiction.

When considering optimistic rollups, it's important to note that they have a 7-day dispute period for finalizing transactions.This helps secure the network by invalidating any malicious behavior committed by the sequencer. That's why when they "promote" themselves, they claim that their transactions are confirmed after 7 days. However, this is technically false. In fact, in an optimistic rollup, rollup blocks are confirmed when they are validated by the rollup nodes. More strictly speaking, the 7-day period they're referring to is the period during which the bridge that connects the rollup chain to the main chain confirms the transaction (unfortunately, the bridge is not the rollup). Of course, the 7-day dispute period is only valid for assets that come from bridge contracts, as the enderlying chain can act as a source of truth through verification, but for rollup-native assets, the 7-day dispute period is a myth.

3.2 The social sovereignty that Sovereign Rollup talks about also exists in classic rollups.

There are two types of rollups: "Sovereign Rollups" and classic rollups (Jon and James refer to classic rollups as Smart Contract Rollups or Enshrined Bridge Rollups). We call it an Enshrined Bridge Rollup because usually when people talk about the difference between a classic rollup and a sovereign rollup, they emphasize the part about being able to hard fork, since the classic rollups have a built-in bridge to a specific network (in the case of Arbitrum or Optimism, it would be a bridge to Ethereum),the difference is that Sovereign Rollups can be hard forked if the rollup community itself agrees to it, while classic rollups cannot be forked without upgrading the contract (bridge contract) of the main chain that the rollup is on. However, as mentioned in 3.1, given that bridges are not rollups, a classic rollup can fork itself without upgrading the main chain (layer 1) smart contract. However, since embedded bridges cannot be forked (and the assets on them cannot be taken with them), they can be forked as long as they accept the financial risk of not being able to take the bridge with them. In essence, there is no difference between a classic rollup and a Sovereign Rollup (in terms of social sovereignty).

3.3 This is not to say that bridges are not important.

Reading the arguments in 3.1 and 3.2, it might sound like we're saying "bridges don't matter", but that's not true at all, and bridges still play a very important role in the structure of a rollup. The nodes in the rollup still define the rollup (which is why it can be forked without a bridge), but that doesn't mean they can fork the collateral assets on the bridge. In the case of Arbitrum, bridge also has the ability to include your transaction if the sequencer censors it. Of course, this censorship resistance is only possible if the rollup is a "canonical chain" from the perspective of the bridge, but it's not meaningless (even if the chain is canonical from the perspective of the rollup, it may actually be meaningless if the rollup node has already forked and the forked chain is different from what the rollup considers to be a canonical chain). However, this situation is very unlikely to happen, which is why the bridge's censorship resistance is meaningful.

Why are these discussions happening? It's very simple. It's to clear up some of the misconceptions people have about the phenomenon of bridges appearing to be rollups when they're not. Kelvin from OP Labs pointed out that this misconception blinds people to the difference between what bridges actually look like (rollups) and their true nature (bridges are not rollups, and the social consensus is that rollups are possible without bridges).Finally, he cautioned against the "words" that the industry often uses. It doesn't matter if it's L1, L2, or L99, what matters is how much more trustworthy and usable this system of trust called blockchain can be.

4. Conclusion: Modular Blockchain is a key theme of KBW 2023

In this article, we've covered the definitions of monolithic and modular blockchains, their pros and cons, and the recent industry debate around modular blockchains. KBW2023 will feature a wide variety of Layer 2 projects, and this article will help you understand what a modular blockchain is, as well as the recent debate around Classic Rollup (or Smart Contract Rollup, or Enshrined Rollup) and Sovereign Rollup, so you can better understand the various rollup-related sessions at KBW2023.

In fact, in the case of monolithic blockchains, there is not much disagreement on the definition, whereas in the case of modular blockchains, people still disagree on the details. Also, in terms of development, it seems that monolithic blockchains are very different in that the most important thing is the development of consensus (how to reduce the communication cost between nodes and allow more nodes to participate in various regions), while modular blockchains need to be defined and developed in various parts (what is L2, what is the authority of the rollup node, how to subdivide the layers, how to define the scope of governance). However, this also means that we have a lot to contribute to the modular blockchain space, so it will be very beneficial to keep up with the discussions around modular blockchain going forward.

Finally, don't focus too much on "words". It doesn't matter if something is L1, L2, or L3. As James Prestwich once said, it's not what you call it, it's how you evolve it that matters. As a researcher, I wholeheartedly support the competition and evolution of the monolithic and modular frameworks.

Thanks to Kate for designing the graphics for this article.