Should Ethereum Layer 2s Urgently Decentralize Their Sequencers?

After Layer 2 network Linea halted block production this week, many demanded the sequencers be decentralized. But is it necessary?

The recent temporary halt of block production on Consensys’ Linea network following a $6.8 million exploit on the Velocore DEX, has renewed the debate about decentralizing sequencers in Ethereum Layer 2 (L2) solutions. Amidst this, on Wednesday, Optimism passed a proposal to implement fraud proofs. Although this wouldn’t help decentralize the sequencer, it would propel the blockchain to what’s referred to as “Stage 1” decentralization on layer 2s.

Sequencers in layer 2 solutions are entities that order and batch transactions before submitting them to the layer 1 blockchain, enabling more transactions per second at a lower cost. However, the centralization of these sequencers has raised concerns about security and censorship.

Learn more: What Are Sequencers in Layer 2 Protocols?

Steven Goldfeder, CEO of Offchain Labs, which develops Ethereum’s top L2, Arbitrum, told Unchained that it’s important to properly contextualize when discussing sequencers: “The term ‘sequencer’ means different things in different protocols.” The sequencer’s role is to order transactions, thus the definition. While this centralized component allows for faster transaction processing, it could censor by delaying transactions, though it cannot alter account balances or post illegitimate transactions. This safeguard is ensured by a network of validators that can challenge any incorrect actions via a fraud-proof mechanism. 

Therefore, according to Goldfeder, while the sequencer is centralized, the validation process is not, involving entities like Offchain Labs, Consensys, and even Google. However, Goldfeder highlighted that the “holy grail” is that everyone can validate. To achieve that, in the next couple weeks, they expect to roll out Arbitrum Bold, a new protocol for managing disputes on Arbitrum chains that enables permissionless validation. Goldfeder suggests that while decentralizing sequencers is an important problem, it is a “bit smaller than what a lot of people might expect.”

Christine Kim, VP of research at Galaxy Digital, believes that decentralization is a spectrum and that fault-proof mechanisms alone do not ensure decentralization. “Rollups in general are significantly less decentralized and secure than the base layer, Ethereum.” However, she thinks that having a working fault-proof mechanism does not mean that a rollup is more decentralized than another because “working fault proofs is not the only mechanism that rollups need to achieve higher levels of decentralization and resiliency,” she told Unchained via email.

Haseeb Qureshi, managing partner at Dragonfly, concurs, noting that while Arbitrum has fault proofs, they are currently whitelist-only. “The important thing about sequencers is not decentralizing so much as making the sequencers more robust. That almost certainly comes first,” he told Unchained.

Even though fault proofs are not the only measure of decentralization for L2s, of the top five L2s by TVL, only Arbitrum has a fraud system in place. Optimism, Base, Blast and Mantle, currently lack fraud-proof mechanisms, meaning their sequencers have more power over transaction outcomes. “Basically they have a dictator,” said Goldfeder.

“We need fault proofs on OP Stack, and then we need fault proofs to be reliable enough that anybody should be able to submit them, and move away from security councils and multisigs,” Qureshi said.

Kim argues that while optimistic rollups should work towards decentralizing their sequencers, they must also build working fault proofs, robust governance mechanisms, reduce reliance on admin controls, lower fees, and improve interoperability with other rollups. “A decentralized sequencer may be one of the hardest initiatives for a rollup to work towards but certainly not the only important one in the near-term to improve a rollup’s decentralization and resiliency,” she added.

Learn more: A Beginner’s Guide to Ethereum Layers: Data Availability, Consensus, and Execution