Equilibrium Infra Bulletin #16: Move L2 on Ethereum, Off-chain Computation Is All You Need, and Exploring Virtual Blockchain Design
Equilibrium Labs builds the state-of-the-art of decentralized infrastructure. We are a global team of ~30 people who tackle challenges around security, privacy, and scaling.
This newsletter allows us to share more about what we read, what excites us, and what we think is relevant to the space. In addition, you will get a glimpse into the organization and our culture. You can also find us on Telegram.
If you have 2-mins to spare, we would appreciate feedback through this form!
The winners for the Token Terminal Pro subscriptions have now been drawn, so check your email to find out if you were among the lucky ones!
Research, Articles, and Industry News:
📚 Lumio: First Move L2 on Ethereum - Recommended by Joakim:
Lumio is an optimistic rollup built on the OP Stack that is compatible with both MoveVM and EVM while using Ethereum for settlement and DA. Developers can deploy their EVM or Aptos Move VM codebases in just a few lines, but cross-VM calls require more effort as the two VMs operate in isolated modes without shared memory or resources. It currently uses EVM as the primary VM, while Move transactions are included in EVM transactions as call data.
In the end state, Lumio aims to be both execution- and settlement-layer agnostic. This means it could work as an L2 for various blockchains or VMs, such as Aptos or Celestia. Lumio is built by the Pontem team who have been core contributors to Move since its inception at Meta. The same team is also behind Liquidswap (DEX) and Pontem Wallet on Aptos.
This announcement fits into the broader theme of rollups experimenting with newer VMs (other than the EVM) and alternative DA solutions. While Ethereum remains appealing as a settlement layer, other VMs than the EVM can offer better performance and/or improved developer experience. Recent examples include Eclipse (SVM with Celestia for DA) and Fluent (zkWASM with EigenDA/ NearDA). We believe that this trend will only accelerate going forward.
Key Takeaway: Lumio brings MoveVM to the OP stack, in a hybrid optimistic rollup that supports both the MoveVM and the EVM. While it currently settles on Ethereum, the team aims to make Lumio execution- and settlement-layer agnostic so that it could utilize any VM and settle on layers other than Ethereum. Currently, only whitelisted users can try out the Lumio testnet, but they will gradually open it to a wider user base.
📚 Off-Chain Compute Is All You Need (by Emperor and Krane) - Recommended by Olli:
Computation over on-chain data is expensive if done directly on blockchains. Co-processors enable moving some of this computation off-chain while guaranteeing correct execution either cryptographically (using ZKPs) or with cryptoeconomic incentives (optimistically with slashing). For a more in-depth overview of co-processors, see Infra Bulletin #12.
Given that both co-processors and rollups move execution off-chain, how do they compare? While there are similarities, a few differences between the two are:
Rollups maintain a state that’s different from the base chain, whereas co-processors don’t have a state of their own. Instead, they use the state of the base chain and take on-chain data as input, execute the computation off-chain, and return the result (along with a proof of computation). This is the main difference.
Rollups also batch several transactions together to amortize costs, whereas co-processors are typically used for complicated tasks with proving done on a per-transaction basis. In other words, co-processors enable more complex computation but don’t increase throughput. Thinking of ZK-based co-processors more specifically, this means they only benefit from the verifiability aspect of ZKPs and not compression (validity rollups get both).
Co-processors can focus on specific tasks, whereas rollups are typically more general-purpose and require circuits that can prove any program that targets the virtual machine (zkEVMs, for example, aim to support any contract that targets the EVM).
A Booster rollups sit somewhere between a co-processor and a full rollup. Booster rollups execute transactions as if they are executed on L1 with access to the full L1 state, but they also have storage of their own.
The difference between coprocessors and rollups is not as clear as it seems at first glance.
While the article focuses on ZK co-processors, the computational overhead of ZKP-based systems is still orders of magnitude higher. Sreeram from Eigenlayer offers some calculations and frameworks to think around the tradeoffs between ZK, optimistic, and cryptoeconomic co-processors.
Key Takeaway: To increase the efficiency of blockchains and enable more complex computations, we need to break loose of the current paradigm of all validators re-executing transactions on-chain and instead embrace the model where one executes and generates a proof, but all verify the proof. This is typically termed on-chain vs off-chain computation. Co-processors enable more complex computation over on-chain data, but they only solve half of the scaling equation as they don’t increase throughput (in its current form at least). However, if you add things like transaction batching to coprocessors - the lines between co-processors and what we today think of as “rollup” start to blur. Will the two eventually converge? Probably not fully as there are use cases for both, but it seems plausible that the gap between the two will narrow.
📚 Exploring Virtual Blockchain Design - Recommended by Hannes:
This article by Polymer (building an IBC hub on Ethereum as a rollup on the OP-stack) expands on a post by Joel at Placeholder VC that defines rollups as virtual blockchains. Just like web2 infrastructure runs on simulations of servers (virtual machines) rather than directly on physical servers, it’s plausible that web3 infrastructure follows a similar path to scaling with blockchains powering multiple rollups (virtual blockchains) on top. Deploying a rollup is much simpler than building a new chain from scratch with all surrounding infrastructure.
Polymer defines application as the part of the blockchain that determines the state transition function and server as the part that deals with consensus or settlement logic. While most virtual chains today can be modeled as a single application and single server, it doesn’t need to be this way. Similar to the cloud native movement in web2, a virtual chain server can service multiple applications simultaneously that provide different functions (or we can even have multiple servers as in the graph below).
A visualization of a multi-application, multi-server system What role does Polymer play in all this? In the context of the Ethereum ecosystem, Polymer aims to serve as the connection layer between different rollups as well as give access to the growing IBC network. Note that in this framework, Polymer (a rollup built using the OP stack) also plays the role of an application that specializes in providing IBC connectivity to other applications (rollups) on Ethereum. While some solutions aim to connect rollups within the same ecosystem (such as zkSync’s hyperbridge), Polymer aims to bring seamless interconnectivity between all rollups. Near’s fast finality layer has a similar end goal but takes a different approach to the problem.
Key Takeaway: While web3 aims to improve existing systems, using analogies from web2 can help contextualize current developments and where we might be heading. Thinking of rollups as virtual blockchains (analogous to virtual machines in web2) and in the context of application/server split is one example of this. Another interesting point raised in the original Placeholder post is that modularity represents web3’s shift from capex to opex, which means that costs scale with growth instead of being front-loaded by large capital raises before launch. This means entrepreneurs can iterate faster and applications can scale cheaply, all while requiring startups to raise less capital upfront.
Personal recommendations from our team:
📚Reading: Eastern Standard Tribe by Cory Doctorow - A fun short story from Doctorow that takes place in a world where online "tribes" are held together by a common time zone, even though members might physically be distributed around the world. Also touches on the theme of remote working, way before COVID-19 made it mainstream (the book was published in 2004). Free to read!
🎧Listening: Chaconne by J.S.Bach - Moran Wasser interprets this beautifully on an 11-string guitar 🎶
💡Other: Proof of Cat - A novel approach to generating randomness for cryptographic systems, utilizing the inherent unpredictability and entropy of domestic felines, also known as cats. One concern raised internally is how random cats are, given that they sleep most of the day? 🐈💤