Have you ever typed nickname.eth in your browser and felt confused? Am I missing something, or is ENS at fault for this not resolving? The reality is that domains are controlled by a single root entity called ICANN, which hasn’t offered new top-level domains for over a decade. There are plenty of reasons why decentralized domains haven’t succeeded so far.
ICANN recently announced they’ve opened applications for new gTLD domains after a decade-long pause (since 2012). This made me realize we already have solid infrastructure in place to become a registrar, which could be an excellent use case for a global singleton (e.g., could solve IRR poor sync).
We currently run authoritative DNS servers as part of IBP geodns, and this software could be expanded. We’ve also secured IP resources for two IPv4/IPv6 anycast ranges to provide more resilient services. Anycast would enable low-latency public DNS (similar to 1.1.1.1/8.8.8.8) for resolving domains. However, resolving domains without ICANN authority violates several RFCs (particularly 8244) and could risk our ASN/IP resources.
The cost of owning a gTLD starts around $185K plus potential auction fees, with ~$25K/year ongoing - a modest amount from a marketing budget for gaining visibility/accessibility to apps in all public DNS. The required infrastructure should be relatively straightforward to set up (plenty of Web2 code exists).
Technical requirements to become a gTLD registrar:
Registry-Registrar Protocol (EPP) System (rs-epp-client):
EPP server implementation supporting TLS
Connection management for registry communications
Command/response handling for domain operations
XML processing capabilities
Authoritative DNS Infrastructure (IBP-geodns/PowerDNS):
At least two geographically separated nameservers
DNSSEC key management system (built-in/plugin in PowerDNS?)
We could make a compelling case for how cryptographically secure self-ownership on a feeless, high-performance chain would innovate in the stagnant DNS space.
What do peers think - is this something we should pursue? If I recall correctly, a referendum made Parity the registrar of .dot nicknames on People’s chain, and the hierarchical structure of nickname registrars suggests some consideration for accepting ICANN as root and obtaining a gTLD. After a decade of waiting, we now have the opportunity to make it happen.
If we would succeed in this process and receive rights to function as registrar for .dot grand top level domain from ICANN, we could expand upcoming nickname service on people chain to also allow setting traditional domain records.
for example you would rent rights to own thebicsam.dot and set following record
A thebicsam.dot 169.169.169.69
and now you could host your website and when you wrote it on browser it would work on browser or anything that uses traditional DNS.
How this all works is that usually your internet service provider runs DNS server that communicates to your pc what ip address each domain name links to. reason ENS domains are not working because those are not accepted by ICANN and require ability to access blockchain data instead of being relayed in DNS servers.
Internet as we know it is very permissioned and hierarchial structure where ICANN(domains) and IANA(IP-address/ASN/routes) hold root rights giving lesser rights down the tree to operators. But because perhaps even intentional design flaws, it is not as bad as many crypto folks often make you feel tho.
Regardless of IANA’s root rights they could not really shut down internet at their will or take already allocated resources even back from RIRs. AFRINIC(african RIR) tried to revoke Cloud Innovation(ISP) ipv4 ranges after ISP was leasing/selling them forward to other regions at ~300x profits. regardless of revoking resources there was barely no influence to routes way BGP works(& poor level of RKPI adaption). Bunch of legacy /8 allocation ranges(usmilitary/IBM/apple) also predates RIRs and are also uncensorable by IANA/RIRs.
This is largely the reason why nation states have chosen to target ISPs through forcing vendors(f.g. CALEA) to produce wiretap-ready router/switches.
Very kool, this will match the cloud narrative, right? I will support this prosal 100%, we need recruit best talent and fund them with good high salary to make this happen
The gTLD project seems like a great idea. It raises an intriguing question: considering the current expenses of treasury financing, what kind of ISP could be established with an annual budget of $1 million? Specifically, how large and which tier could such an ISP operate at?
Would it be feasible to secure agreements with tier-1 backbone providers and gradually develop a tier-1 network fully operated by the Polkadot economy? Owning critical internet infrastructure could unlock numerous possibilities—imagine MBone-like overlays repurposed for entirely new goals.
What we are suggesting is moving forward with establishing ISPs (Internet Service Providers). This includes public IP ranges, packet-switching hardware, peer agreements for connectivity, physical locations, and other necessary components, all funded by the treasury. The aim is to expand operations and progressively become a tier-1 provider, effectively owning critical internet infrastructure. This context forms the basis for the questions posed.
to be frank as much as that industry is in need of disruption, I don’t think that becoming a tier-1 backbone provider is realistic at all. it’s quite sad how centralized the ISP business circle jerk is, but also for that reason it is extremely difficult to disrupt. even big tech companies like Amazon have not been able to obtain T1 status even though they probably run most internet traffic via their own subsea/submarine cables and ~600 multiterabit pops globally.
what i do think matters to the polkadot network in terms of resiliency is for validators to obtain their own asn and ipv6 resources to increase the nakamoto coefficient rate in terms of ISPs in the network. obtaining these resources costs around ~1k/y membership fee depending on RIR, as well as requiring the ability to manage multihomed networking. it’s probably not economically and technically viable for most validators, but still a good north star to target for.
seems like this isp topic is confusing folks so let’s just focus on talking about generic top level domains.
What annual budget of $1 million? The OP said $25K/y. Even $100K/y would be wasteful.
it’s good to understand here that obtaining an asset like gtld is a good opportunity to generate revenue through registration sales as well as gain visibility, traction and legitimacy. second level domain names like polkadot.network are usually sold for around 10 to 20 usd with yearly recurring payments (premium 2-4 letter names even more) and there’s a decent chance that dotdot domains could gain traction and be resold by second level domain vendors like namecheap/porkbun/godaddy.
there is probably quite a bit of pressure/weight of compliance and responsibility for moderation/censorship as the registrar in charge. and I think hardest question for me is that if we proceed to apply, which entity should be in charge and mailbox for all possible copyright/trademark rolling and how to separate it to not influence our unstoppable nicknames.
No one optimistic about Polkadot should think that investing in becoming a backbone provider could generate returns better than investing in “core” projects…
I agree about the registrar. I meant if $100K was required for registrar + backbone provider business - it’d take years to build it so it’d mean $500-$1,000 K over years. The odds of recouping those costs would seem slim to me.
Massive support, as long as the technical side can be done.
About the legal side of things, who will be the legal entity responsible/owner/front face of the new .dot domain/gTLD?
The Polkadot Community Foundation? or an alternative legal figure should be the legal holder of that asset?
Also, some automatic payment rails to the ICANN need to be studied as they probably won’t support crypto payments any time soon.
DNS records are a good thing to put on-chain since they rely on centralized authority servers that sometimes go down. Having dealt with ICANN in the past some things really get stuck in bureaucracy, there are several companies that still are waiting for permissions on their registered domains. Have you asked ICANN how long it would take to process the .dot application?
A good case study on financial successful dns authorities is the Anguilla Country with there .ai domain(this is a cTLD and not a gTLD doe…). As the sales of .ai domains covers a huge chunk of the island nations budget.
