Source: ENS Blog; Translated by: Deng Tong, Golden Finance
I'm Esk3nder.eth, Head of Product at ENS Labs. I'm excited to tell you about our plans for ENSv2, announced in May 2024. As a quick recap, ENSv2 is our ambitious plan to scale ENS to Layer 2 while reimagining the protocol from the ground up. Since 2017, ENS has become a cornerstone of crypto infrastructure, with millions of .eth names and thousands of ecosystem integrations. But ENSv2 is not just about L2 scaling; it's about leveraging seven years of experience and technological advances to make ENS more accessible, decentralized, and more powerful.
We hope this update gives you some insight into our research process so far, and where we are currently in our ENSv2 L2 scaling journey. As always, we are here to answer any questions and feedback - thank you for your patience and support!
Our research journey
The ENSv2 project began with a fundamental question: "If we reimagined ENS from scratch today, what would it look like?" Frustrated by the limitations of the mainnet user experience, we began investigating the ecosystem. We spoke with builders across the space, and after productive discussions both externally and internally within the team, we unanimously agreed to extend ENS to Layer 2.
Over the past few months, we’ve immersed ourselves in the L2 ecosystem, consulting with a wide variety of stakeholders — from L2 teams and RaaS providers to infrastructure companies and early-stage startups. We invited many of them to an “ENS Unconference” during EthCC to discuss the technical barriers and opportunities for ENSv2. We also uploaded the talk to Youtube for those who want to follow along at home! Throughout this process, we’ve been focusing on six key criteria (listed below in no particular order) that act as our North Star, helping us formalize our process as we explore the Layer 2 landscape.
EVM Compatibility
CCIP Read Support
Open Source
Exit to L1
Sequencer Decentralization
Finality
Where We Stand: Three Promising Options
Ok, let’s get straight to the point! Well, we’ve narrowed down our focus to three main options, in no particular order:
Public ZK Chain
This option involves extending ENS directly onto an existing public ZK chain, such as Scroll and Taiko. Alternatively, we could leverage an Optimistic L2 with ZK proofs, similar to what Succinct recently demonstrated at Frontiers (OP+SP1). This approach requires minimal chain development and long-term maintenance costs, but sacrifices some independence of ENS governance from the public chain.
Our own ZK chain instance
This approach involves deploying our own instance of an existing ZK chain stack. A recent example of this approach is Status’ recent announcement that they are working with Linea to launch their own instance of the Linea stack. This option provides teams with more flexibility while still benefiting from mature technology. The flip side is the effort of developing, launching, and maintaining a separate L2 instance.
zkVM (special purpose rollup)
This approach involves developing a custom zkVM (zero-knowledge virtual machine) optimized specifically for ENS operations. We are exploring this option in collaboration with Axiom, who are helping us develop a zkVM prototype. By optimizing the VM, you can make the ZK integration more efficient. This reduces the cost of proofs and allows you to generate proofs more frequently, significantly reducing finality time. However, it represents the biggest departure from our current architecture and requires the most extensive development and migration effort.
The Road Not Taken
You may have noticed that there are no purely optimistic chains in our list of options. While chains like Arbitrum and Base currently dominate layer 2 economic activity, having proven their product-market fit and proven technology stacks, this omission is intentional and ties into our sixth criterion: finality. In this context, finality refers to an agreed-upon checkpoint where all previous blocks are considered irreversible.
In optimistic rollups, there is no immediate proof of validity when state is published to mainnet. This initiates a “challenge window” — typically 7 days — during which the proposed state can be challenged before it is considered final. In contrast, ZK rollups use “ZK proofs” to immediately prove the validity of every transaction that has been executed, batched, and then sent to L1 for final settlement. ZK finality tends to be within 2-24 hours compared to the time of the “challenge window,” which is typically set at 7 days. This rapid finality is critical for ENS to prevent name ownership rollbacks, a principle set out in the ENS DAO constitution. For a more in-depth look at the differences between optimistic and ZK rollups, StarkWare has published an excellent explainer post.
However, we are not completely dismissing optimistic technology. As seen in our first option, hybrid solutions combining optimistic chains with ZK proofs are still under consideration. As we continue our research and development, we remain open to innovation across the L2 ecosystem that aligns with our core standards and values.
Next step
We have learned a lot over the past few months and would like to thank all the teams and individuals who have contributed to this process. Thanks to their help, we are starting to move from theory to practice. The next steps are as follows:
Benchmarking and testing: We know our three main paths and will test their performance, cost-effectiveness, and scalability so that we can use concrete data to guide our decisions.
Technical improvements: Our team is working on improvements to the ENSv2 architecture, including a hierarchical registration system and an updated universal resolver.
Funding Proposal: To support this critical work, we are preparing a funding proposal for the ENS DAO that will broadly outline the resources needed to implement ENSv2.
Community Engagement: Your input is invaluable as we embark on this journey, and we will continue to host community calls and AMA sessions to gain insights and hear your ideas.
Final proposal: We aim to have a final proposal for L2 policy for ENSv2 in the coming months.
