Introduction to Modular Blockchains
Earlier architectural views classified the key components of a blockchain protocol into distinct layers of modules relatd to infrastructure, data, network, consensus, clients and applications. Public blockchains have augmented themselves into nascent economies powered by network effects. Thus key components of a blockchain protocol can be summarised into following layers from an economic point of view.
Execution layer — the layer to process transactions and generate state transition functions. Users interact with the blockchain through this layer by signing transactions, deploying smart contracts and transferring assets. We can consider rollups as modular blockchains that specialise in execution layer tasks.
Settlement layer — the layer to verify transactions, bridge assets, confirm proofs and resolve disputes. This layer typically exist in the layer 1 blockchain.
Consensus layer — the layer to provide transaction ordering and finality through a network of full nodes downloading and executing the contents of block and reaching consensus on the validity of transactions.
Data availability layer — The data required to verify the state transition is published and stored in this layer. Even if malicious block producers withhold transaction data, this layer should continue to provide transaction data with ease of verifiability. Innovations like data sampling, data dispersal and data compression techniques are getting introduced in this layer.
Modularity is achieved when we can divide a monolithic system into smaller components which can be enhanced, exchanged or even replaced. Each module encapsulates specific functionalities or features, allowing for independent development, deployment, and upgrades. Unlike the monolithic counterparts, where all components are tightly integrated, modular blockchains encourage modularity, enabling developers to optimise and customise different aspects of the system.
Through the adoption of modular design approaches in execution, consensus, and data availability layers, blockchains are able to enhance their throughput while maintaining the intrinsic properties that make the network robust and resilient.
Rollups are worth mentioning in this context. Rollups are essentially modular blockchains that publish their blocks on a monolithic blockchain. Unlike other scalability approaches, rollups support arbitrary computation. Rollups have three main components namely — sequencers, rollup full nodes and rollup light nodes.
Sequencers are nodes that receive rollup transactions from block proposers and combine them into a block that they send to L1. Rollup full nodes obtain blocks from L1, process and validate all transaction data by verifying that the root is correct. If a block contains invalid transactions it is discarded. In this way, Sequencers cannot create valid blocks containing invalid transactions. Rollup light nodes cannot compute the new state themselves, but verify that it is valid using techniques such as invalidity proofs or validity proofs. According to Luca Donno of University of Bologna, rollups can be considered to scalable as the amortized cost of transactions decreases as the number of users increases, as the cost of ensuring blockchain validity grows sub-linearly relative to the cost of verifying transactions individually
In this model, block production transitions to specialised and scalable producers for better throughput, while the block verification happens through a decentralised network. Ethereum engineering has become a brave new world of probabilistic proofs and distributed data structures thanks to the introduction of rollups. There is an interesting concept known as fractal scaling introduced by StarkWare which allows application specific rollups to be deployed on L2s. This becomes a further focused network economy such that all L3 applications on L2 can achieve interoperability and network effect on L2 itself.
Thus modular blockchains have carved out a niche and nascent space for themselves powered by the finesse of engineering and ecosystem effects.
