Bankless: Shared Sequencer for More Decentralized Rollups

Original Title: 《The Tech Speeding Up the 'Progressive Decentralization' of Rollups》

Author: 563, Bankless

Compiled by: Deep Tide TechFlow

The current Rollup architecture still relies on centralized sequencers to handle transaction ordering, which poses risks of censorship and downtime, making a simple and convenient decentralized solution urgently needed.

In this article, Bankless researcher 563 proposes a solution—Shared Sequencers (SS)—and compares the advantages of SS relative to existing designs.

If you are a DeFi enthusiast like me, you will appreciate a good Rollup. Swapping, lending, and trading—all for just a few cents—was an almost frictionless experience that seemed unimaginable a few years ago. However, sometimes using Rollups feels more like a centralized chain like BNB Chain rather than Ethereum.

While lightning-fast transaction confirmations and low fees have attracted DeFi users to Rollups, most L2s are still struggling with decentralization.

Although Crypto Twitter trolls love to point fingers at Solana… Ethereum L2s are not without their flaws either.

These chains often have single points of failure in their sequencers.

You may have heard that Arbitrum and Optimism plan to "decentralize" their sequencers, but how will this be achieved? And (more importantly) why haven't they done it yet?

Current State of Rollup Sequencers

When you sign a transaction on a Rollup chain, it gets queued in the Rollup's memory pool along with other transactions. The sequencer then orders these transactions to build and execute blocks.

Sequencers are unique components of Rollups; they can efficiently package transactions and provide users with the assurance that their transactions will be quickly sent to Layer 1.

Today, most Rollups rely on a single internal sequencer to order transactions. This model not only increases the probability of downtime but also creates trust assumptions. The reliance on centralized participants makes Rollup traders vulnerable to censorship—this is a significant point of contention among us decentralization advocates. If for any reason the sequencer does not want you to transact on their chain, you may have little recourse to access your funds.

While the seemingly obvious solution is to create a decentralized role, many Rollup teams have prioritized increasing market share. These teams believe their time is better spent building developer tools, onboarding new partners, and engaging with the community. Establishing the necessary infrastructure to support a decentralized sequencer set takes time and has little impact on the average retail user—let alone the fact that these centralized sequencers have been earning fees.

For these reasons, if you ask most projects when they plan to decentralize their sequencers, the answer is likely to be "coming soon." However, Shared Sequencers provide an elegant solution to address some obvious needs.

Solution = Shared Sequencers

The concept behind Shared Sequencers (SS) is actually quite simple. They propose establishing a decentralized network of sequencer nodes that Rollups can collectively plug into, replacing their single centralized sequencer.

The network is no longer dominated by a single sequencer; instead, a new leader is chosen for each round of ordering, effectively eliminating the active and censorship-related single points of failure introduced by centralized sequencers.

Rollups that can "plug in" and "unplug" from these networks provide a clean transition and allow for healthy competition in the SS space. If an SS network mistreats its users or extracts too much value for any reason, its Rollup clients can easily turn to better options among its competitors.

Thus, the SS network offers "decentralization as a service," rather than zkSync or Starknet having to bootstrap a whole new set of sequencers from scratch.

The ability for many Rollups to plug into the SS network brings exciting possibilities for atomic transactions. Since SS can simultaneously order transactions from multiple Rollups, cross-Rollup operations become much simpler. Imagine being able to arbitrage ETH between Arbitrum and Optimism—ensuring that your buy on Arbitrum only succeeds if your sell on Optimism is successful. This effectively allows users to cross-margin between Rollups, potentially solving the prevalent liquidity fragmentation issue in today's DeFi.

In summary, Shared Sequencers have several advantages over current designs:

• Censorship resistance;
• Cross-Rollup atomic swaps;
• Vitality (low/no downtime) guarantees;
• Plug-and-play solutions for existing and new Rollups.

So—are there any drawbacks? Understandably, new technologies come with new hurdles. The main concerns expressed by the community boil down to value capture and incentive distribution. For instance, SS might extract value from Layer 1 and/or aggregate MEV into the dominant SS network.

While these issues seem solvable, developers need to ensure that incentives do not stray too far from the Layer 1 chain. If enough value does not flow to the base layer, dishonest behavior may arise. For example, if the Rollup sequencer network is obtaining more than its fair share of MEV, it may be more profitable for malicious validators to fork the Rollup contract rather than manage it honestly.

Having a pre-built decentralized sequencer network is a game-changer for entrepreneurs. In many upcoming exciting new zero-knowledge and application-specific Rollups, they are more inclined to focus on optimizing their products rather than having to build an entire set of sequencers.

Competitive Landscape

Teams like Astria, Espresso, and Flashbots are at the forefront of this technology, working to help Rollups move toward a decentralized future.

Astria fully embraces a modular narrative, and its shared sequencing project is no exception. Astria has spun out from the Celestia team and will utilize Celestia's data availability layer, planning to "test" their SS by launching Astria EVM.

Espresso's SS will use HotShot's custom consensus mechanism, promising higher throughput and faster finality than other designs. They also plan to interact securely with Ethereum validators through restaking protocols like EigenLayer.

The Flashbots team has never been absent from any narrative; they are working on SUAVE—aimed at making the block generation process more trustless and collaborative. Their novel "preference" design offers users an innovative way to interact with a cross-chain builder network—where each builder competes for users' attention.

This project will take time (possibly years) to develop, and in the end, Vitalik may be right when he says, "Block production is centralized, but block validation is trustless and highly decentralized." But it's always worth trying for decentralization… right?

These tools bring us closer to the decentralized mass adoption that we, as aspiring cyberpunks, support. Therefore, even if "coming soon" turns into "not coming," Shared Sequencers can still provide the much-needed solution for Rollups.