Monolithic vs Modular Blockchains

Modularity breaks up a blockchain into its key parts

TLDR:

• Monolithic blockchains can be broken up into 4 key Modules: transaction execution, settlement, consensus and data storage.

• Monolithic vs Modular blockchains is all about trade-offs between security, simplicity and decentralisation vs scalability, flexibility and efficiency.

• Bitcoin for example is a better store of value since its a monolithic blockchain, making it a hardened, slow to change blockchain.

• Ethereum is moving to a more modular architecture with the Beacon chain for consensus and shards handling groups of transactions.

Welcome to the new people reading my Substack this week! If you’d like to catch up on the last couple of weeks, we covered DAOs, Stablecoins and Bitcoin ETFs.

This week I decided to look at one of the big topics being discussed at the moment, the topic of modular blockchains.

With the growth of attention on networks like Celestia, we’ve seen the conversation of monolithic vs modular blockchains grow too, so let’s dig in.

Monolithic vs Modular

Monolithic blockchains represent the traditional way to build blockchains. Bitcoin created the original monolithic blockchain architecture where the key aspects of the chain were built into a single unified layer, namely:

• Transaction Execution

• Settlement

• Consensus

• Data Storage

In other words, in a monolithic architecture every node in the network reaches consensus on every transaction, processes and settles every transaction, and stores the state of the entire blockchain’s data that’s come before.

In comparison, modular blockchains are a new emerging architecture that breaks up the traditional blockchain stack into separate modules, usually the four mentioned above.

This idea of breaking up a monolithic architecture into a modular one is a very common pattern seen in software and hardware architecture in general.

Almost all computing systems go through this type of evolution, where the initial versions are built in a monolithic way just to create an initial system. Then over time as this system becomes well established the focus becomes on optimisation, so developers will begin to break it up into its constituent pieces to refine them further.

Comparing Architectures

While breaking up monolithic architectures into modular ones to optimise them further is common in computing systems, it’s not without its challenges and also not without its downsides.

Monolithic blockchains have proven to stand the test of time, with Bitcoin being the original and surviving 15+ years at this point without any downtime, nor successful hacks or attacks. The fact there’s just a single layer handling everything makes it easier to interact with, and creates less potential points of failure.

Since every node maintains the entire state of the chain there’s a high level of security and robustness. This helps the blockchain to maintain its integrity and remain sufficiently decentralised, trustless, and permissionless.

On the flip side since every node processes every transaction there’s a real challenge with scalability and the amount of load the network can handle. Moreover, upgrades to these blockchains are very slow or in some cases impossible to reach consensus on since a single change can affect everything.

Modular architectures solve a lot of these problems. Firstly, since all the key aspects of the network are split up, they can scale a lot more easily and continually upgrade parts of the network without disrupting the whole.

Also, new ideas can be taken from other networks and inserted more readily into the original network. Plus a network can be more easily specialised for specific tasks.

Yet, this flexibility is not without its challenges. Firstly, modular blockchains are new and with this there’s more surface area for attacks, especially with the added complexity of different modules being built by different teams without a full understanding of how it all ties together.

Importantly, there’s also a key challenge on centralisation, as some modules could become controlled by a few dominant players if they become industry standards.

Comparing monolithic and modular architectures is all about trade-offs, these are between: security, simplicity and decentralisation on the monolithic side, versus scalability, flexibility and efficiency on the modular side.

It’s all tied up to the age-old “blockchain trilemma” that I discussed previously when writing about consensus algorithms. The constant challenge and trade-offs between scalability, decentralisation, and security.

Bitcoin

As we’ve already mentioned, the primary example of a monolithic blockchain is Bitcoin. Bitcoin created the very concept of a blockchain by integrating consensus, transaction processing, settlement, and data storage all in one, with its nodes processing everything.

Bitcoin has the best track record in crypto, its proved to be very strongly decentralised and has never been hacked. Together with its predictacle monetary policy, that will only ever create 21m Bitcoins in total, it has proven itself to be the best digital store of value there is and dominates the space in terms of market capitalistaion.

Bitcoin is an example of where a monolithic architecture is a ideal for its specific use-case, because it’s sufficiently hardened as a protocol, which is exactly what you want from a digital store of value! However, it’s also one of the slowest blockchains out there with slow block times and only 3-7 transactions per second, plus very expensive data storage costs.

Ethereum

Ethereum grew out of Bitcoin so it began monolithic. But now it’s transitioning over to a more modular architecture with Ethereum 2.0 as the developers try to grow transaction volume and reduce block space (ie. gas) costs.

Ethereum’s creation of the Beacon chain is a great example of a modular design as the consensus layer is split out from transaction processing, settlement and data storage into its own separate chain. Ethereum 2.0 plans to split things further through sharding, where shards will processing different chunks of transactions and data, so individual nodes will not process the entire network’s traffic.

Polkadot is another similar example, having split the consensus layer into the Relay chain, from the execution in its many parachains. These parachains can swap in different apps to execute transactions without affecting the overall network.

Celestia

Celestia’s been one of the big players recently in pushing the modular blockchain narrative. They provide a data availability layer that’s separate to consensus and transaction execution, and they’ve seen a lot of adoption by Layer 2s.

For many L2s the biggest cost is data storage. Celestia’s solution allows L2s to store data cheaply, so they only have to handle transaction execution. To acheive block finality they still need to store data on the L1, but this is just a summary of transactions, and is much smaller than the full data they keep on Celestia.

This setup means L2s can acheive high transaction throughput by focusing only execution, while they borrow security from the L1 and data availability from Celestia.

Future

Looking at the full picture it’s clear that there’s no right answer as to whether monolithic or modular chains are better. As with most questions with blockchains it all depends on the use-case.

So we’ll likely see a growth in modular blockchains without there being a detrimental impact on the growth of monolithic chains. Especially when considering that L2s in many cases can solve the same problems as a modular design without requiring a monolithic L1 to change.

Modular blockchains can already bring clear benefits but we may find the greatest benefits come in new use-cases that we’re not even yet aware of.

The goal of secure, scalable and sufficiently decentralised systems will continue to drive innovation on all fronts.