A progression from centralized sequencers to based sequencing.
According to L2BEAT, the definition of a sequencer is: “a party responsible for ordering and executing transactions on the rollup. The sequencer verifies transactions, compresses the data into a block, and submits the data related to it to enable state reconstruction to Ethereum L1 as a single transaction. The data can be either transaction data or state diffs.”
Rollups work quite simply: users send transactions to an entity called a sequencer, which orders transactions, creates batches, and sends transactions to the base layer, usually Ethereum. Currently sequencers on most rollups are centralized, meaning that the centralized entity (typically the Labs corp contributing to the rollup) is responsible for inclusion in blocks and execution ordering of transactions in each block. Over the past few months, several projects have decided to launch as based rollups, such as Taiko’s Gwyneth and Puffer’s UniFi.
So, what is a based rollup?
Based Rollups use a sequencer election mechanism called "based sequencing"
And, what is based sequencing?
Based Sequencing is a sequencer election mechanism that chooses a sequencer from the set of current and future (known) sequencers of the base layer (L1). In the context of Ethereum, the set of current and future known sequencers are validators in the Ethereum proposer lookahead (and the current L1 proposer). Note that some based rollups may use a fallback sequencer election (so they aren't "based" for a few blocks).
Based sequencing involves giving more responsibilities to Ethereum validators. Currently, validator nodes only manage Ethereum transactions, but we can juice the Ethereum validator set with additional tasks, in exchange for greater rewards. Although Ethereum validators could earn additional rewards this does not necessarily come at the expense of the rollup. See the rollup podcast with Justin Drake for more on sequencer revenue, and the distinction between congestion and contention fees.
The based rollup approach is quite innovative, so let's understand why based sequencing is a significant improvement over centralized sequencing and traditional shared sequencing mechanisms.
Why should you care about based sequencing?
Based rollups still capture sequencing revenue.
Despite giving sequencing rights to the Ethereum validator set, based rollups still capture revenue from congestion fees, congestion MEV through EIP 1559, and general MEV through (re)selling proposing rights (like execution tickets) are possible mechanisms to capture rollup revenue. A model has been proposed by @gspasov and @danielkivanov for a sequencing design called Vanilla Based Sequencing. Their proposal directs gas fees to block proposers while allowing the rollups to capture a portion of the sequencing revenue. This model is achieved by embedding a commission fee over the L2 sequencer's balance increase. This fee aligns the protocol’s success with its revenue, benefiting both the rollup and the sequencers.
Better UX with “real” pre-confs
Okay, let’s start by breaking down what a pre-confirmation (pre-conf) is.
When you initiate a transaction (tx) on a rollup with a centralized sequencer, you are placing your trust in the entity that operates the sequencer. This entity, typically owned by the rollup Labs, provides some reasonable assurance to the user that the tx will eventually be included in an Ethereum block. The rollup provides you with a promise of inclusion, or inclusion pre-conf. Some Rollups may also include an “execution preconf”, so the result of the user's transaction is also guaranteed. We can think about (pre-)confirmations on a spectrum from less sure to more sure about tx settlement; inclusion pre-confs being the weakest assurance, execution pre-confs being next weakest but a bit stronger, and finality being the strongest assurance.
But, as users, we want assurances of inclusion and execution without relying on centralized sequencers!
Composability will improve the entire ecosystem.
Current rollup clusters cannot synchronously communicate between each other, nor with the Ethereum L1. Given the DeFi liquidity and user activity on Ethereum L1, this lack of composability presents a significant bottleneck. The lack of L2 composability has led to user and liquidity fragmentation, poor user experience, and very poor bridging experience leading to billions of dollars in lost funds due to bridge hacks and user error. The current state of Ethereum is fragmented.
Shared sequencers and chain abstraction are pushed as the solution to such fragmentation, but rollups on a shared sequencer are not necessarily synchronously composable; there are other requirements, one important requirement being that the block proposal process (sequence broadcast) for each domain is atomic. Additionally, this type of synchronicity is not enough for cross-chain contract calls such as flashloans.
The special thing about based rollups is that they are synchronously composable with the L1, as well as other based rollups.
Gwyneth Technical Design illustrating L2 <> L1 composability
Based rollups are not sovereign.
According to Celestia, the defining characteristic of blockchain sovereignty is forkability; “Sovereign rollups upgrade through forks like a layer 1 blockchain. New software versions are published, and nodes can choose to update their software to the latest version. If nodes disagree with the upgrade, they can stay on the old software. Providing a choice lets the community, those that run nodes, decide whether they agree with the new changes. They can’t be forced into accepting upgrades, even if most nodes upgrade. This feature, compared to smart contract rollups, is what makes sovereign rollups ‘sovereign’.”
An Ethereum based rollup uses based sequencing, which means it elects its sequencer from Ethereum's validators. This ties the rollup's operation directly to Ethereum's validator set and its proposer mechanisms. Since the sequencer election mechanism is anchored in Ethereum, a based rollup is inherently reliant on Ethereum, reinforcing the conclusion that it is not sovereign.
One caveat is that based rollups can be considered more sovereign than traditional dapps on Ethereum. Unlike dapps, which are restricted to the EVM, based rollups can implement custom rules, like an independent execution environment, or altVM. Additionally, fallback mechanisms for sequencing offer resilience that dApps cannot match, as dApps depend entirely on Ethereum's block proposers. While based rollups are not fully sovereign like sovereign rollups, their ability to manage state and transaction processing independently makes them more sovereign than traditional dApps, bridging the gap between full dependence and full autonomy.
Conclusion
As it stands, based rollups have led to many different opinions among protocol researchers.
Nevertheless, projects are opting to build based constructions for greater sovereignty, composability with L1, and better UX while still capturing revenue from their rollup networks. Plus, muh Ethereum alignment ;)
Thanks for reading.
P.S. Check out our podcast on Gwyneth, the based rollup from Taiko, for a deeper dive into being based.
Ressources: