Op-Ed: Application-specific rollups are a prime solution for growing the Ethereum network’s processing power.
By Felipe Argento and Brandon J. Isaacson
The scalability issue at a glance
To understand why monolithic blockchains struggle to scale, we need to get a bit into the weeds. When one chain is handling every function — consensus, data availability, and execution — a single set of nodes is called upon to verify each and every action, by every DApp and user in the network. This can overload the system very quickly.
As usage increases, DApps and users start competing for the chain’s finite blockspace, which becomes an unmanageably scarce resource. The result is that DApps and users who can’t win blockspace bidding wars are excluded from participating.
The “shared” rollup approach
If we think of Ethereum as one large, shared computer, Ethereum shifted its focus to scaling how much data its blocks can hold (let’s view this as “upgrading its hard drive”), while delegating computational scalability (“upgrading its processor”) to rollups projects.
The first major wave of rollups adoption began in 2021. The launch of Layer 2 (L2) projects like Arbitrum and Optimism was based on “shared” rollups architectures. The rollups from these projects are “shared” in the sense that each DApp on the protocol shares space inside a single rollup with other DApps deployed on the same L2.
The “shared” rollup approach brings us back to square one — albeit with some buffer time.
The arrival of appchains
In search of maximum scalability, customizability, and fee predictability, projects, including Cartesi, are starting to move beyond shared rollup architectures. They’re finding that building on application-specific rollups (commonly referred to as appchains) can bring better results for computational scalability.
Similar to shared rollups, application-specific rollups act as off-chain execution layers that inherit security and censorship guarantees from the Ethereum base layer. But now, instead of sharing space inside a single rollup, each DApp has its own dedicated rollup to process off-chain tasks.
This setup not only solves the issue of bidding wars amongst applications, but also provides significant gains in computational scalability. With only one application per rollup, each DApp can now benefit from full, unshared computing power. Rather than competing in a zero-sum game for validator computing power, each DApp can have its own high-performing rollup chain.
These gains in computing power open the design space for developers in a very meaningful way. DApps deployed on application-specific rollups can now begin to more closely mimic traditional software applications in terms of programmability, cost efficiency, fee predictability, and user experience.
Moving forward: the cone of innovation
But for Web3 to reach its full innovative potential, rollups projects must continue to push the boundaries of decentralized computation. The figure below helps us visualize how scaling data and computation together can pave the way for previously impossible decentralized applications:
Web3’s Cone of Innovation created by Cartesi core contributors
On the x-axis, we see that data availability improves with the implementation of EIP4844 and sharding. On the y-axis, computational capacity scales as we move from a monolithic L1 blockchain, to shared rollups, to application-specific rollups.
(The applications and their positioning inside the cone aren’t intended to be taken as gospel. Rather, the figure is meant to provide an intuitive outlook on the growing horizon of decentralized applications.)
The main takeaway is that gains in data availability can’t be fully utilized without concurrent gains in computational capacity, and vice-versa. Both need to grow together.
Application-specific rollups are a prime solution for growing the Ethereum network’s processing power.
Felipe Argento is an advisor at the Cartesi Foundation, focusing on Layer 2 research and application. Felipe joined the Cartesi project in 2018, leading a team of blockchain engineers and making notable contributions to the architecture, design, and on-chain implementation of Cartesi Rollups. Before joining Cartesi, Felipe was a software engineer focusing on blockchain applications for clean energy, carbon credit, and energy futures in a partnership with Brazilian energy companies. Felipe speaks Portuguese and English.
Brandon J. Isaacson is a lawyer and board member at the Cartesi Foundation. Prior to joining the Cartesi project, Brandon was an associate at Latham & Watkins, specializing in the securities and regulatory sector, before launching his own law practice. Over the past decade, Brandon has represented clients in regulatory investigations, enforcement proceedings, and securities and ICO litigations, as well as advised startups on regulatory issues associated with the launch of disruptive emerging technologies.
Cartesi is an app-specific rollup protocol with a virtual machine that runs Linux distributions, creating a richer and broader design space for DApp developers. Cartesi Rollups offer a modular scaling solution, deployable as L2, L3, or sovereign rollups, while maintaining strong base layer security guarantees.