There has been a proliferation recently of new top-level domains (TLDs) issued by various blockchain-based naming companies. We’ve also received requests over the years for the Ethereum Name Service (ENS) to create new ENS-native TLDs (beyond .ETH) but intentionally haven’t done so. Both of these things raise the questions: What does it mean for a blockchain-based naming system to be a good citizen in the global namespace? And what approach is best for users and the long-term success of blockchain-based naming?
This is a topic to which we’ve given a lot of thought, and we’d like to explain our position to the community. We believe blockchain technology represents a paradigm shift in how naming is accomplished on the Internet, but building good technology isn’t even half the battle. Human-readable naming is chiefly a task of social coordination, so implementing it correctly is essential for avoiding unnecessary conflict and achieving long-term success.
In this post, I will explain why Internet naming requires careful social coordination, why ENS doesn’t create more TLDs, but how we plan on nonetheless expanding the namespace available for use on ENS.
In other words, this is our current thinking on what it means to be a responsible citizen in the global namespace.
This is a long post! But we think it’s important, so let’s dive in.
Human readable domain names are by their very nature socially contingent. This is of course unlike computer generated machine identifiers, which can be automatically generated without social significance.
The difference between two IPv6 addresses is highly meaningful to computers but socially meaningless; few people know their device’s IP address, and neither do people care. The difference between human-readable domain names, on the other hand, can carry enormous social meaning. Human readable names are not socially interchangeable, and their particularity means it matters a great deal who gets what name.
Further, there can only be one owner of any particular human readable domain name on the Internet. Imagine if “google.com” took you to different websites owned by different companies depending on what browser you were using or which ISP you had. For naming on the Internet to be useful, there needs to be one globally agreed upon namespace.
But who gets what name? And under what terms? And when someone creates a new name, how is that communicated to the rest of the Internet, ensuring no one else will also try to use that name? Therefore, creating a new TLD is not so much a technological question but a social one. It’s easy technologically for anyone to create new TLDs, whether they are using blockchain technology or not; we could easily create thousands of new TLDs and start selling names from them to users. The question is whether the names would be accepted by the rest of the Internet in a reliable way so that they would be useful long-term to users. Avoiding the name collision problem of different parts of the Internet using the same name in conflicting ways requires getting everyone on the Internet to agree to the same naming rules.
In other words, for there to be one globally agreed upon namespace, you need close social coordination between representatives and stakeholders from the whole world.
This is why ICANN exists. The Internet Corporation for Assigned Names and Numbers is a non-profit organization with the purpose of coordinating the Internet’s Domain Name System. ICANN meetings are open to the public (attendance is free and sessions are often broadcast live on the Internet) and include thousands of people from around the world.
To those already scoffing at the mention of ICANN: There are of course legitimate reasons to criticize ICANN, as is the case for any human organization. The point here is the role ICANN plays is essential to the smooth running of the Internet we all take for granted.
Among other things, the ICANN community representing the whole world has developed rules and procedures for the creation of new TLDs to ensure there is only one owner of a particular TLD on the Internet, that everyone has a chance to apply for a given TLD, and that a TLD has the best chance of ending up in the hands of the organization that makes the most sense. These coordination mechanisms are critical to maintaining, as the ICANN motto puts it, “one world, one Internet.”
The ENS-native TLD: .ETH
When ENS launched in May of 2017, it did so with the newly created TLD .ETH. We did not claim .ETH through the normal ICANN process, and here’s why.
ENS was experimental technology. It was a side project of a few folks working at the Ethereum Foundation, and it wasn’t clear if it would work (the DAO had just blown up less than a year earlier) or if people would really use it. In fact, cognizant of their responsibilities and to limit the possible damage of failure, only .ETH names that were 7 characters or longer were issued. (This is because shorter names are rarer and therefore more valuable. Now that we are confident in the ENS smart-contracts, we are currently releasing 3–6 character .ETH names with a one-time auction, after which all remaining names will be available for instant registration.)
The process for claiming a new TLD through ICANN wasn’t open at the time (and still isn’t), and, even if it was, it can take years and cost hundreds of thousands of dollars. So waiting to getting approval from ICANN wasn’t practical.
In other words, ENS needed some room to experiment.
Well, it turns out there is in fact a massive demand for a naming system with the decentralization and censorship-resistance that blockchains provide. Even though .ETH names were tricky to acquire (RIP your lost ETH due to not revealing your bid; that auction process has since been replaced by instant registrations), required the use of ETH and a Web3 browser, had no marketing, was only known by a fairly small group of technology nerds… there are now more than 310k .ETH names registered (not counting sub-domains) in just two and a half years, with millions of dollars spent in their acquisition (none of which, by the way, went to the ENS developers in the original deposit-and-return system). There are also 35 wallets that have already integrated ENS (or committed to integrate soon), more than any other blockchain-based naming service, as well as native integration in the Opera browser (Brave coming soon), in MetaMask for decentralized websites, and more.
To put this in context: if you compare .ETH to the new DNS gTLDs released through the ICANN process eight years ago, .ETH is one of the most successful new gTLDs on the Internet.
Being Good Citizens
So we did some experimentation outside of the normal procedures. But that doesn’t mean we don’t want to be good citizens now that we’re growing up.
ENS spun off from the Ethereum Foundation last year with a grant, creating its own non-profit organization. We chose the non-profit model because we view ENS as a public good, a basic piece of infrastructure not only for the Ethereum ecosystem but potentially for the future of the whole Internet.
We changed the deposit-and-return fee model to a yearly spent fee model to better guard against name squatting, choosing prices we thought were the lowest possible so as not to deter legitimate users while still deterring large-scale squatting (most .ETH names cost $5 USD per year).
Before making any changes we had months of public discussions, including on what to do with the funds raised. You can read more about our thinking on this issue here and here, but ultimately we concluded the funds were best used to support the long-term improvement, maintenance, and promotion of ENS. This would reduce our dependency on grants and make it more likely ENS can continue to serve the Internet long-term.
Given this model, it also means we could potentially bring in more money if we created more TLDs from which to sell names. And in fact there are several venture-backed for-profit businesses, inspired by the success of ENS, doing just that: selling blockchain-based names with their self-created TLDs (sometimes basing their system off the ENS code base), with little regard for how these names will fit (or not) into the global namespace long-term.
We have intentionally refrained from doing that. This is a case in which being a non-profit has helped us look beyond short-term payoffs and think about what’s best for users and the Internet long-term.
We believe creating additional TLDs outside of the ICANN process and selling these names to customers is not only damaging to the long-term success of blockchain-based naming, it’s also bad for users.
The more blockchain-based TLDs that are created outside ICANN’s process — not simply for experimentation but in a flagrant unwillingness to respect the procedures developed by the global Internet community through hard-won experience — the more unnecessary antagonism is created with that global Internet community. Some people may want to “fight ICANN,” but that means fighting the global Internet community that ICANN embodies. We don’t think that’s a good use of our time and resources.
But wantonly creating new TLDs outside of ICANN is especially bad for users, for two reasons: First, this unnecessary antagonism means it will likely take longer for most Internet users to enjoy the benefits of blockchain-based naming, such as better security, censorship-resistance, and programmability, et al.
Second, and perhaps most importantly, it means the blockchain-based names sold to users might not ever be integrated into the Internet’s global namespace. This means the names could potentially be given out to different people in the future, rendering their original purchase worthless. For example, if we created the TLD .WALLET and started issuing names from that to users, the next time ICANN opens up registrations for new gTLDs, someone else could (benignly or maliciously) acquire .WALLET and start selling the same names — but for use on the much larger and more entrenched Domain Name System of the Internet. How much will a user’s name be worth if in most of the places where names are used on the Internet someone else owns their name? In the very least, the confusion will likely lead to unnecessary mistakes, which we all know in the cryptocurrency world can be costly.
The success of our experiment with .ETH has proven the value of the technology. Moving forward, we want to be as responsible as we can. This includes possibly seeking to register .ETH through the normal ICANN process if and when the opportunity presents itself in order to protect our users. We may also apply for other new TLDs through the ICANN process that we could make available on ENS. And we have been investing in building relationships with the ICANN community, attending their meetings, and being an ambassador for blockchain technology and how it can benefit naming.
To members of the ICANN community who may be reading this: we come as friends!
Expanding the Namespace Available on ENS
But does that mean users of ENS are stuck with only .ETH names for the time being? ENS provides naming for many things beyond the Ethereum ecosystem, such as IPFS websites, Tor .onion addresses, text records, other cryptocurrencies, and we even have a project to store traditional DNS records on ENS. Inasmuch as the TLD .ETH represents Ethereum, this would seem very limiting.
Our solution to this is the integration of the current DNS namespace for use on ENS.
We have devised a system that uses DNSSEC to create cryptographic proofs to allow DNS domain name owners to claim their respective names for use on ENS.
For example, the Ethereum Foundation owns the DNS name “ethereum.org”; with our system they could also have an ENS record for “ethereum.org” (not ethereum.eth, which is a separate name). In this way, the Ethereum Foundation could use “ethereum.org” both for their normal website (using DNS) and for receiving cryptocurrency payments (using ENS).
This already works for .XYZ names, as well as in a special way for names on .LUXE, .KRED, and .ART. And soon we will be rolling out this functionality to all DNSSEC-enabled DNS TLDs, which is most DNS TLDs and includes all the major ones.
In this way, we support and respect the current DNS namespace already used by the whole Internet, give domain name owners new functionality with the ENS infrastructure, and expand the usefulness of ENS.
The future is bright for blockchain-based naming infrastructure, provided it is done responsibly. We believe ENS is the next evolution in how naming is accomplished on the Internet, one that can evolve with and support the current namespace and Internet community.