By: Raye Hadi
Table of Contents:
-
Introduction
-
Team
-
Based vs Non-Based Sequencing
-
Hotshot + Design
-
Conclusion
Introduction:
After embracing the switch to a rollup centric future, Ethereum is finally tackling one of its major issues: scalability. With the deployment of numerous adjacent rollups racing to solve these scalability woes, the promise of improved performance finally seems attainable. However, this progress is not without its own drawbacks. Not only has the complex structure of a multi-rollup system raised concerns for its propensity to fragment Ethereum’s ecosystem by dividing up applications and liquidity; it has also introduced new centralization risks. This divergence from the credible neutrality and composability Ethereum was built for is the tradeoff many L2s have had to make in their pursuit of increased throughput. To some, this seems extremely irrational, as relying on a trusted third party (in the form of a centralized sequencer) to enable faster execution nullifies the value of using a permissionless platform. Lacking any unification mechanisms, these rollups splinter the Ethereum ecosystem, adding complexity and difficulties for both users and developers alike, thus eroding composability. The value of Ethereum’s settlement layer comes from its credible neutrality and thorough decentralization. Recognition of this value is what has elevated Ethereum to the ~$460 billion market cap it boasts today.
For some time we have seen a growing desire to transact on Ethereum in a lower latency, higher throughput fashion. Though rollups have proven capable of realizing this desire, they come with their own costs . As development continues, there will be immense value in a middleware that allows users to leverage the low latency, high throughput benefits of rollups, without the inherent centralization risks and fragmentation that their current implementation causes. This is the ambition of Espresso, a shared sequencer layer built to bring web3 security to web2 performance.
Team:
Founded by Yale cryptographer and assistant professor, Ben Fisch, Espresso boasts a highly skilled team of cryptographers and engineers. Initially starting with aspirations for a highly scalable and decentralized L1, the Espresso team realized the best way to achieve this ambition was to build on Ethereum. Espresso recognized the potential of rollups but also their tendency to undermine the benefits of using Ethereum in the first place. From there, they worked to develop a shared sequencer to help rollups avoid such tradeoffs and preserve credible neutrality, security, and interoperability without sacrificing performance.
Based vs Non-Based Sequencing:
Today, most rollups use what is called “non-based sequencing”. This means they rely on one external sequencer to order transactions from the mempool into blocks. This method is highly performant as the need to rely on the processing capacity of the L1 is bypassed. However, dependency on a sole entity for ordering and block building comes with risks. Due to the lack of competition, the sole sequencer has the ability to set fees at monopolistic levels and censor users by excluding their transactions from the block. To combat this, many rollups implement an “escape hatch”, allowing users to force exit their funds to the L1 in the case of malicious activity by the sequencer. The alternative approach is a method known as “based sequencing”. This rollup method is very decentralized, as ordering and block building is performed on the L1 and only execution is performed on the L2. Notably, this method relies on the L1’s processing capacity for transaction ordering. This makes based sequencing less performant than non-based sequencing, but more secure and decentralized. With the main purpose of a rollup being better performance, most implementations go the route of the former and accept the potential risks of a centralized sequencer. Compounding this problem, each rollup’s external sequencer works in isolation, meaning every rollup’s transactions are fetched and ordered independently, with no inter-rollup capabilities. This makes interoperability a nightmare, leading to the fragmentation concerns covered prior. Observing these two options, Espresso devised a decentralized shared sequencer layer powered by a consensus protocol to give rollups the best of both worlds.
HotShot + Design:
Espresso’s shared sequencer is designed with distinct features to optimize for decentralization and interoperability on the sequencer level without sacrificing scalability. By using a PoS consensus protocol called HotShot, Espresso is able to leverage optimistic responsiveness to run a high throughput model (more centralized) in favorable conditions (low latency, high bandwidth, nodes behaving honestly) and automatically shift to a more robust model (more decentralized) in unfavorable conditions (high latency, low bandwidth, nodes behaving maliciously). This can be thought of as leveraging a typical “Web2” architecture such as routing traffic through a few high-bandwidth nodes in favorable conditions, and leveraging a typical “Web3” architecture such as a traditional gossip based model in unfavorable conditions. From this HotShot gets the best of both worlds, automatically adjusting between decentralization and speed to optimize performance. Espresso uses a similar method for data availability (DA) solutions. Utilizing a small rotating committee of random nodes to sample data (DAS) in favorable conditions, but instituting a reliable yet slower backup path that uses erasure coding and DAS to fall back on in unfavorable conditions.
Additionally, Espresso has explored partnerships with EigenLayer to allow L1 validators to participate in their validator set through restaking. This would bolster both decentralization and security of the shared sequencer while aligning the economic incentives of Espresso validators with Ethereum mainnet. In terms of interoperability, the architecture of a shared sequencer already enables cross-rollup transactions. This is because every rollup integrated with the shared sequencer sends transactions to the same mempool where they are fetched and ordered before being sent to their respective L2. This means transactions between rollups can be executed seamlessly, as assets are transferred out of one rollup (effectively debiting its ledger) and credited to another. This allows for a coherent movement of funds across rollups before the consolidated final state is recorded on mainnet. To further enable interoperability, Espresso uses proposer builder separation (PBS), to make sure transactions are sequenced in a manner that respects inter-rollup dependencies. This allows for atomicity and effectively combines liquidity between rollups, reducing slippage, improving price discovery, and maximizing capital efficiency. When all these features are combined, Espresso can be regarded as a one stop solution for scaling in an Ethereum aligned way.
Conclusion:
Addressing a crucial problem in Ethereum’s transition to a modular design, Espresso essentially kills two birds with one stone by tackling issues with credible neutrality and composability. Both are key components to Ethereum’s value proposition and as rollups grow in influence, both problems are exacerbated without a sufficient solution. It is unclear what Espresso’s revenue model will look like, but from an investment perspective it is not difficult to see that in the successful execution of the rollup centric vision; a middleware that the majority of rollups deploy for sequencing will be an incredibly valuable module. The current rollup market sits at a TVL of ~$37.5 billion, hosting ~145 TPS compared to Ethereum’s ~16 TPS at the time of writing. Aggregating this activity through a singular shared sequencer would not only generate significant value at the shared sequencer layer, but provide overall benefits to the entirety of Ethereum.
Additionally, with the execution of the Dencun upgrade (EIP-4844) we are due to see much more utility and demand for L2s. However, convincing the majority of rollups to all aggregate their activity through a single shared sequencer will be difficult, as open sourced markets epitomize competition. Projects like Polygon’s Agg layer and alternative shared sequencers are already looking to provide their own solutions. This is why it will be extremely important for Espresso to grow adoption and quickly. Espresso has already run two testnets, showcasing integration with three large players in the rollup space: the Gibraltar testnet with Arbitrum, and the Cortado testnet with Polygon zkEVM and Optimism. Early onboarding efforts like these are important, especially in terms of building a first mover advantage and the network effects that come with it. While the protocol Espresso hopes to build is ambitious, the quality of their founding team and their track record of successful testnets are both strong indicators of their ability to deliver. The shared sequencer layer will be a risky and competitive landscape. Any project able to offer seamless developer integration and onboard a large share of the rollup market early will be hard to dethrone. If shared sequencing is to take off, then exposure to this layer will be imperative for investors in the modular ecosystem. For Espresso, the founding team, innovative architecture, proven testnets, and early start provide a solid foundation. It is definitely a project to keep an eye on and possesses significant tailwinds that could make it a staple in the shared sequencer conversation going forward.
Sources: