Scaling Ethereum - Rollups

Rollups have emerged as the prominent solution for blockchain scalability to the extent that Ethereum released its roadmap with rollups as the choice scalability tool. A modular blockchain, rollups are one of the more popular second-layer solutions whose sole task is completing transaction executions off-chain. The growth of modular chains has provided developers with a wide array of choices in their dApp deployment, forging the path for greater innovation, interoperability, and scalability.

A crucial element within the discussion around Ethereum scalability is EVM compatibility which, for developers, is advantageous due to its ease of porting smart contracts on Ethereum to the rollup chain and the integration of existing Ethereum tooling (such as code libraries, deployment infrastructure, and programming languages), tools which have been thoroughly tested over the years. Developers can also have access to liquidity already on that chain, a pre existing user base, and cross-compatible dApps. There exist many EVM-compatible blockchains on which developers can deploy their dApps to ease Mainnet congestion (ex. BNB Chain, Polygon, and Avalanche), but these sidechains lack Ethereum’s security measures as well as have a tendency to compromise on either decentralization or security.

With rollups however, given that the transaction data is posted on-chain after being executed off-chain, it allows rollups to inherit Ethereum’s native security, evidenced by the fact that any reversal of a rollup transaction also reverses on Ethereum Mainnet, publicly viewable by anyone on the blockchain. Primarily, compatibility with Ethereum infrastructure means developers can have fewer barriers when building and deploying dApps.

General-Purpose vs. Application-Specific Rollups

In the realm of rollups, developers are faced with two distinct paradigms: general-purpose and application-specific. Each supports varied approaches to addressing the network’s throughput limitations; in fact, the tradeoffs can be characterized as choosing between adaptability and optimization. Generalized rollups offer what a general-purpose blockchain like Ethereum offers, just at a lower cost; some projects are now being deployed directly on layer 2s, bypassing Mainnet entirely. These solutions, such as Arbitrum One, Optimism, and Starknet, inherently share their universal flexibility with a wide range of dApps that might find beneficial the attention and liquidity already available on the general-purpose ecosystem. Consider a general-purpose rollup. For a dApp that requires oracles, it would be sensible for that project to apply various sources of data to enhance the robustness of price reliability, which a general-purpose rollup can provide due to its wide ranging support for interconnectivity.

On the other hand, optimization is realized through modularity, as in, application-specific rollups allow developers to have control over gas costs, the ability to affect on-chain congestion, broader token utility, customized development environments, and a tailored experience for users. Therefore, these rollups are able to optimize functionalities within a defined set of use cases, fulfilling unique requirements from their specialized approach, all while assuming the security practices of its base layer. The scalability problem could also be framed as the challenges around computation and data scalability. Dissecting the problem further, these concepts, while at the user level may have the reputation of impacting gas costs, means understanding that the solutions lie in addressing the predicament in a rather highly designed manner. Essentially, it would behoove developers to recognize the computation scalability disadvantages that general-purpose rollups have through their use of a shared network where other dApps are vying for computation power and data availability. Some may regard application-specific rollups to be simpler to implement as the cascading effects of decisions and updates are felt at just the application level rather than across the entire system.

However, this is not to say that application-specific rollups are easier in general to implement; flexibility is the price to pay for specialization. If necessary, developers must adopt the specific programming logic of that app, resulting in an overall increase in engineering effort. As is the case when engineers are faced with the dilemma of choice, it is highly contingent upon the developers, their experiences, and the intended use cases.

The aforementioned ideas of attention and liquidity are suitable for certain use cases, namely gaming, DeFi, social apps, and NFTs. For instance, gaming (ex. Immutable X, Ancient8) exists in an environment that is self-contained and self-referential; dApps aim to cultivate their own users and utilize their own capital, while needing the scaling solution to counteract the heavy computation loads that gaming mandates to ensure network functionality and a smooth user experience. Moreover, such dApps gain from having greater control (i.e. optimization) more so than the breadth of services that general-purpose rollups have available. DeFi is another area that can participate in the utility innate to application-specific rollups: a lowered gas cost barrier to encourage more users; incredibly diverse and specific protocols; and scaling methods to mitigate high transaction volumes.

Ultimately, as blockchain infrastructure moves toward a decidedly optimizable and tailored landscape, application-specific rollups can be the catalyst in distributing dApps in a decentralized manner, creating a market for cross-platform communication protocols.

Optimistic Rollups vs. Zero-Knowledge Rollups

There is no discussion around rollups without briefly mentioning the two prevalent methods of transaction data submission: optimistic and zero-knowledge (zk), both of which can be employed by general-purpose and application-specific rollups. Even more so than being EVM-compatible, optimistic rollups (ex. Arbitrum One, Optimism) are EVM-equivalent, which, for developers, is desirable for its interoperability across multiple networks, programming familiarity, and market reach – all culminating as functional parity. However, the drawback is that due to its innate optimism, the entire security model depends on the existence of at least one honest node; otherwise malicious users can act without consequence.

Zk rollups, highly proficient at providing stronger privacy and security guarantees, are top of mind for many in the scalability space because by virtue of their cryptographic properties, the computation load is significantly more difficult to solve. A hindrance to the development of dApps on the zk rollup layer (and why zk rollups are purely application-specific at this point), there currently is no EVM-equivalent zk rollup available as EVM-compatibility is an undeniably more challenging property to achieve due to the complex and resource-intensive nature of circuits which are the fundamental building block for implementing cryptographic operations.

Nevertheless, both optimistic and zk rollups face the challenge of centralized sequencers, an issue to which many practitioners are proposing solutions. Many see zk rollups being a more applicable use case for scaling, largely due to their quicker transaction finality and validity proofs secured by trustless cryptographic mechanisms. Ethereum also continues to design Solidity with layer 2 requirements in mind, in an effort to advance towards a zk-based EVM.

The Future of Rollups

Continued expansion of zk rollups and zk technology in general is undoubtedly developing quickly. With the focus on bridging Ethereum and its infrastructure to layer 2 rollups, zkEVMs are a means to scaling Ethereum via validity proofs to support general composable dApps. Furthermore, as rollups in general continue to improve, so should the research and applications around security risks to which they may still be vulnerable. Amongst the burgeoning landscape, it is expected that the model of zk rollups is extrapolated as rollups as a service or rollups on demand. Exemplified by Ethereum’s rollup-centric roadmap, the future of scalability could revolve around zk rollups as the technology advances towards grander utility.

The ethos around enabling scalability on Ethereum can be summarized as producing compatible yet specialized tools that enable further cross-chain functionality. The more scaling options there are, the larger the cumulative effect that these solutions bring to Ethereum. Modularity, combined with other innovative infrastructure initiatives, contributes to the vision of realizing the blockchain trilemma.

References

Argento, F., de Moura, E., Teixeira, A. (2022, October 11). Application-Specific Rollups. https://medium.com/cartesi/application-specific-rollups-e12ed5d9de01

Ethereum. (2023, October 25). Layer 2. https://ethereum.org/layer-2

Ethereum. (2023, October 25). A digital future on a global scale. https://ethereum.org/roadmap/vision

Ethereum. (2024, January 25). Scaling. https://ethereum.org/en/developers/docs/scaling

Kim, C. (2023). Scaling Blockchains: The Modularity Thesis. [Whitepaper]. Galaxy Research. https://assets.ctfassets.net/h62aj7eo1csj/6wtWCTyWpkvOES781X2bLd/ba7a4aefd66eebcde1979856639e84bb/GLXY_2023_Whitepaper_Scaling-Blockchains_v01.pdf

Koegl, A., Meghji, Z., Pellegrino, D., Gorzny, J., & Derka, M. (2023). Attacks on Rollups. Proceedings of the 4th International Workshop on Distributed Infrastructure for the Common Good. https://doi.org/10.1145/3631310.3633493

L2BEAT (2024). https://l2beat.com/scaling/summary

Levin, A. (2023, June 30). App-Specific Rollups: A Trade-Off Between Connectivity and Control. https://variant.fund/articles/app-specific-rollups-trade-off-connectivity-control/

nanfengpo. (2022, July). Rollup as Service: Opportunities and Challenges. [Online forum post]. Eth Research. https://ethresear.ch/t/rollup-as-a-service-opportunities-and-challenges/13051

Thibault, L. T., Sarry, T., & Hafid, A. A. (2022). Blockchain scaling using Rollups: A comprehensive survey. IEEE Access, 10, 93039–93054. https://doi.org/10.1109/access.2022.3200051

Zhang, Y. (2022, March 14). zkEVM. https://hackmd.io/@yezhang/S1_KMMbGt#Why-possible-now