Rollup Ecosystem Overview: What are the new designs beyond Optimistic Rollup and zk Rollup?

Written by: Alex Beckett

Compiled by: Amber

Since the concept was born in early 2019, Rollup has made significant progress. Today, there are two main competitors in the market, which have given rise to a plethora of imaginative design theories and hybrid evolutions. This article will review the development status of the Rollup ecosystem as of mid-2022.

Optimistic Rollup

Due to the first-mover advantage of the general Optimistic Rollup, it has captured most of the Rollup applications, with Arbitrum and Optimism being the benchmarks for "Rollup."

Ranked by TVL, Arbitrum currently holds the first position among Rollup applications, with a locked amount of 2.4 billion dollars. Although it has been running on the mainnet for over a year, there are still many aspects of the system that need testing. The main entity developing Arbitrum, Offchain Labs, operates a single centralized sequencer. While interactive fraud proofs are flexible, the centralized sequencer is the only whitelisted entity that can submit disputes to generate fraud proofs. From the user's perspective, although the gradual trend towards decentralization of the entire system is expected, they still need to place a significant amount of trust in the system.

The next-generation upgrade of the protocol, Nitro, has gone live on Devnet, which will fundamentally change the existing architecture by replacing the custom Arbitrum virtual machine with a WASM-Geth combination. Arbitrum's interactive fraud proofs will run on WASM, and nodes will have a codebase equivalent to Geth, along with some specific Rollup optimizations. In summary, Nitro will bring significant optimizations, improving performance while providing better EVM compatibility.

Optimism is the second-largest Optimistic Rollup, ranked third among all Rollups by TVL, with 469 million dollars locked. The operation of Optimism PBC is similar to that of Arbitrum, both having a single centralized sequencer. However, Optimism PBC has found a way to provide returns to the network using the profits from the sequencer to incentivize ecosystem development. The first round saw a donation of 1 million dollars to 58 public goods. Although centralized designs struggle to have a positive impact in a world that is fundamentally "decentralized," Optimism's attempts seem to have alleviated this issue to the greatest extent.

Currently, Optimism's design lacks fraud proofs, although gradual decentralization and necessary security upgrades are expected. However, almost all Rollups have immediate or delayed upgradability, so the security of Rollups ultimately depends on the upgrades of multi-signature designs. The upcoming Bedrock upgrade will transform Optimism's architecture to be similar to Arbitrum's. The current Optimism virtual machine will be replaced with a MIPS-Geth combination, where nodes are equivalent to regular Ethereum Geth nodes. New features will also include interactive fraud proofs, which are an upgrade from its original non-interactive fraud proofs. Bedrock will be another milestone for Optimism, adding many features and improvements to the system.

Additionally, Optimism has achieved a leap by conducting an experiment in non-elite governance. Decision-making is divided into two parts: token governance and user governance. Token governance uses tokens as voting credentials, while user governance follows a one-person-one-vote system. The most challenging aspect of such a system is the allocation of voting identities, which needs to minimize false attacks. If a user can obtain more than one voting identity, they can exert more influence on the "government" than ordinary users. Optimists point out that they will use non-transferable NFTs to represent citizenship; however, the threat of individuals exchanging private keys for more voting power still exists.

Fuel is another noteworthy competitor in the Optimistic Rollup space, taking a radically different approach from Arbitrum and Optimism. Fuel Labs is building a custom virtual machine for Fuel V2, using a Rust-based programming language. While EVM-compatible Rollups are particularly useful for the Ethereum development ecosystem, custom VMs can more easily maximize performance as they do not have to adhere to many established standards.

What I like most about Fuel V2 is its parallel transaction processing capability. Once the data volume reaches a significant scale, Rollups that implement parallel processing will have an advantage over those that do not. Notably, Fuel V1 is the first Optimistic Rollup on Ethereum mainnet and remains the only Rollup with permissionless sequencers and fraud proofs.

For Fuel, Optimistic Rollups are far superior to zk-Rollups. First, Optimistic Rollups have complete and unrestricted composability: zk-Rollups have an inherent disadvantage in composing smart contracts through zk circuits. The only zk-Rollup with universal composability is StarkNet, but it currently has limits on the number of TVL that can be supported by its permissioned smart contract deployments and bridges. All other zk-Rollups are specific applications or only facilitate token transfers.

Another point is due to EVM compatibility: zk circuits have inherent compatibility issues with certain types of cryptography in EVM, making zkEVM a challenging task. Optimistic Rollups have been running EVM compatibility on the mainnet for some time, and upgrades are expected to further achieve equivalence.

zk-Rollup

Recently, the Ethereum community has reached a consensus that zk-Rollups are the ultimate state of scalable Rollups. Therefore, compared to Optimistic Rollups, there seem to be more zk-Rollups actively in development, many of which are planned to launch within the next two years.

StarkNet is currently the only general-purpose, composable zk-Rollup on the mainnet. However, the system is still in its early alpha stage and has various limitations. The bridge between StarkNet and Ethereum has restrictions, and the upper limit will gradually increase in the future. Smart contract deployments on StarkNet are also in a whitelist mode. I believe this is mainly to reduce the risk of smart contract errors, as there may not be enough auditors to audit all the contracts developers want to deploy—StarkWare temporarily taking on this role is reasonable. Auditability is one of the general drawbacks of the new custom language, and the complexity of zk systems only exacerbates this.

zkSync is another major "zk-Rollup" competitor, with the recent testnet launch marking the introduction of the first instance of zkEVM in a live testnet. However, with the addition of zkPorter, zkSync 2.0 is no longer just an upgraded version. It will allow users to choose to publish their transaction data between zkPorter and Ethereum. While Ethereum is upgrading its data throughput through danksharding, volition serves as a great middle ground, providing users with better choices.

While most zk-Rollups prioritize scalability, privacy is another important aspect that zk-Rollups can achieve. Aztec is currently leading in the privacy space with their privacy token transfer zk-Rollup (zk.money), and they are the only Ethereum Rollup focused on privacy that I am aware of. Aztec is expected to launch its next iteration, Aztec connect, soon, enabling users to access Ethereum DeFi privately. This is a significant advancement compared to using applications like Tornado.cash, as Tornado.cash's privacy can only be achieved through mixing and cannot directly "hide" transactions.

zk-Rollups have become very complex, and adding privacy issues makes them even more complicated. zk-Rollups may never reach the state of privacy-composable smart contracts. Because of this, privacy may emerge through specific application chains, whether through zk-Rollups or Validiums built on top of zk-Rollups.

A variety of other zk-Rollups are also gradually emerging, including various projects from Scroll and Polygon. A significant distinction between zk-Rollups is the use of custom virtual machines or zkEVM execution environments. Its advantages and disadvantages are similar to those of Optimistic Rollups. However, zk-Rollups have more inherent complexities in implementing zkEVM. For this reason, there is a very strong case for adopting the route of custom virtual machines and languages, such as StarkNet and Cairo.

Sovereign Rollup

The last two categories of Rollups currently exist only in theory. The difference between Sovereign Rollup and common Rollups is that it has a fork choice rule that allows it to fork independently of its base layer. In contrast, a regular Rollup delegates its fork choice to its settlement layer, as it needs to ensure the correctness of the Rollup.

Sovereign Rollups are most prominent on DA layers like Celestia, where the correctness of Rollup transactions is not guaranteed by the DA layer. As a result, Rollup transactions on platforms like Celestia are inherently Sovereign transactions, as they must ensure the correctness of their transactions through fraud/validity proofs and fork choices. This should not be mistaken for consensus; the consensus provided by Celestia is a protocol for transaction ordering.

For optimistic Sovereign Rollups, transactions are considered correct, so Rollup nodes only need to download block data from Celestia. zk-Sovereign Rollups ensure correctness through validity proofs, which are distributed among Rollup nodes via a p2p network.

The importance of Sovereign Rollups is reflected in their forking capabilities, which allow Rollups to truly operate independently of their base layers.

Settlement Rollup

Settlement layers are a type of Sovereign Rollup specifically built for settlement. Importantly, the settlement layer is any blockchain that has a bidirectional trust-minimized bridge with the Rollup. This bridge allows tokens to be transferred bidirectionally between the Rollup layer and the settlement layer. Trust minimization is a property of the bridge, where communication relies solely on the honest minority assumption through verified data availability and fraud/validity proofs.

Like any settlement layer, the purpose of Settlement Rollups is to provide an environment for the "Rollup" to verify proofs, resolve disputes, and bridge tokens. Although, technically, "Rollups" above the settlement are hybrid, as they use off-chain DA through the data availability layer of the Settlement Rollup, making them a verification layer or an optimistic verification layer.

Hybrid

Validium

Validium is a hybrid zk-Rollup where transaction data is published off-chain, meaning it relies on an external environment in addition to the settlement layer used for validating validity proofs. StarkEx is the only instance of a validator that has gone live. StarkEx is a validator for specific applications, currently supporting three applications: Immutable X, Sorare, and DeversiFi. StarkEx also supports a zk-Rollup mode, which is used by the currently second-largest Rollup application by TVL, DyDx.

StarkEx employs a Data Availability Committee (DAC), composed of a group of trusted parties, to provide data availability for StarkEx's validators. While delegating data availability to a permissioned committee reduces security, it allows StarkEx's validators to offer cheaper transactions than zk-Rollups. Cost reductions are possible because publishing data to Ethereum is very expensive, which is also a major variable cost contributing to Rollup transaction fees.

Some security issues associated with Validium with a DAC can be mitigated by using an external data availability layer. The main increase in security comes from the cryptoeconomic security provided by the blockchain, where nodes can be imprisoned and slashed for dishonest activities. The effectiveness of this implementation is an interesting experiment in the cost-security trade-off of "Rollups."

Image source: Celestia

Optimistic Validium

Like Validium, Optimistic Validium is also a hybrid Optimistic Rollup where transaction data is published off-chain. There is currently no general consensus on what this specific hybrid should be called.

Metis is currently the only instance of Optimistic Validium. It is important to note that the security of Optimistic Validium is somewhat weaker than its corresponding Validium because data availability is necessary for generating fraud proofs and successfully resolving disputes. If a dispute arises and the data regarding state transitions is unavailable, then the fraud proof cannot prove fraud. As a result, if the off-chain data availability provider fails to provide data, funds can be stolen from Optimistic Validium.

Volition

By combining zk-Rollup and Validium, there is another hybrid Rollup called Volition. Volition allows users to choose between on-chain or off-chain data availability. This choice is made at the individual transaction level, with off-chain data representing cheaper fees and lower security, while on-chain data leads to higher fees and greater security. This gives users the freedom of choice provided by individual systems rather than explicitly seeking a link that fits their cost-security preferences.

Currently, zkSync 2.0 is the only publicly announced Volition in development. In zkSync 2.0, on-chain data is provided by Ethereum, while off-chain data is provided by their own dedicated PoS chain called zkPorter. However, due to the current lack of detailed information disclosure, it can only be speculated that StarkWare will provide a user-customizable option for StarkEx or StarkNet in the future.

Adamantium

Adamantium is a Validium where everyone must personally provide their data to the network. Individual transaction data is stored by individuals (off-chain), and they must remain online to prove the availability of data for each block. If a user is not online or fails to prove, their funds will automatically revert to the on-chain settlement layer. Although StarkWare proposed the design of Adamantium, no team has made further progress in this area. If this hypothesis can be successfully implemented in the future, it may be very attractive to users or entities that wish to have more personal control over their security.

Enshrined Rollup

Enshrined Rollup is a Rollup that is part of an existing blockchain. Simply put, it is an execution shard. The distinction between Enshrined Rollup and execution shards seen in Ethereum 2.0 proposals and other similar sharded blockchains is that execution shards are proposed as single shards. The global validator set will be divided into committees and assigned to a specific shard as a validator group. Execution shards will operate as independent blockchains, with their own execution, consensus, and data availability, but they will checkpoint back to the "beacon chain," similar to how sidechains checkpoint back to their chosen chains. Theoretically, an Ethereum Rollup will only perform execution and use the beacon chain to verify data availability and fraud/validity proofs.