Subject: Re: [IANA #1494826] Request for new DNS Resource Record Type (UTXO, 
code 260)
Dear Paul,
Thank you for your thorough review and for raising these critical architectural 
questions. Your reference to RFC 7929 (DANE for OpenPGP) is highly relevant, 
and it highlights exactly why we must be precise about what this RRtype is—and 
is not—designed to do.
We would like to clarify the design intent and address your concerns directly:
1. Mapping to Individual Users (The RFC 7929 Challenge)
You are absolutely correct that DNS is not designed to be a user-level identity 
database. The UTXO RRtype is not intended to map a single domain to a 
monolithic user wallet. Instead, the proposed Data RR acts as a "network-native 
discovery anchor" for cryptographic endpoints and application-layer delegation 
policies.
To handle individual users or specific AI Agents, the architecture leverages 
DNS hierarchical namespaces via subdomains. For example:
• _utxo.example.test → Points to an organization-level AgentPolicyEnvelope 
(authorization constraints).
• agent-01.example.test → Points to a specific agent's vLEI DID or payment 
endpoint.
• alice.example.test → Points to an individual user's endpoint.
The DNS record only provides the lightweight pointer. All user-specific 
authentication, KYC, and delegation verification occur at the application layer 
(e.g., W3C Wallet Layer), not in the DNS resolver. DNS remains a neutral, 
cacheable transport layer.
2. UTXO Informative Reference & Payload Size
We acknowledge the lack of context. UTXO stands for "Unspent Transaction 
Output," a cryptographic security model used to bind ownership of a blockchain 
endpoint.
To address your concern about data volume, the Data RR is strictly designed as 
an opaque but lightweight carrier. The RDATA consists of a CBOR-encoded 
Endpoint Identifier and an optional VRF (Verifiable Random Function) Proof. It 
does not store raw blockchain data or KYC information. The payload size is 
intentionally constrained to tens to hundreds of bytes (comparable to 
DNSKEY/RRSIG records) to ensure it remains a cache-friendly, lookup-oriented 
pointer and avoids UDP truncation/TCP fallback.
3. Clarifying the Draft Examples
The example.test in the current draft was merely a placeholder for standard DNS 
documentation practices (RFC 2606). We will update the IETF draft (in the 
upcoming -02 revision) to include explicit examples of subdomains, multi-agent 
scenarios, and an informative reference explaining the UTXO model to better 
illustrate the granular mapping capability and intended DNS usage.
We appreciate your expertise and will explicitly discuss the boundary between 
DNS resolution and wallet-level verification, as well as the exact payload 
constraints, in the next revision of the draft.
Best regards,
Guorong Tian
CoCa Foundation

来自阿里邮箱Android客户端------------------------------------------------------------------
发件人:Foundation<[email protected]>
日 期:2026年10月24日 00:23:22
收件人:Paul Wouters<[email protected]>
抄 送:<[email protected]>; dnsop<[email protected]>
主 题:回复:[DNSOP] RRTYPE Allocation Request: UTXO (proposed code 260) & 
Supplementary Architecture Evidence

Subject: Re: [IANA #1494826] Request for new DNS Resource Record Type (UTXO, 
code 260)
Dear Paul,
Thank you for your thorough review and for raising these critical architectural 
questions. Your reference to RFC 7929 (DANE for OpenPGP) is highly relevant, 
and it highlights exactly why we must be precise about what this RRtype is—and 
is not—designed to do.
We would like to clarify the design intent and address your concerns directly:
1. Mapping to Individual Users (The RFC 7929 Challenge)
You are absolutely correct that DNS is not designed to be a user-level identity 
database. The UTXO RRtype is not intended to map a single domain to a 
monolithic user wallet. Instead, the proposed Data RR acts as a "network-native 
discovery anchor" for cryptographic endpoints and application-layer delegation 
policies.
To handle individual users or specific AI Agents, the architecture leverages 
DNS hierarchical namespaces via subdomains. For example:
• _utxo.example.test → Points to an organization-level 



来自阿里邮箱Android客户端------------------------------------------------------------------
发件人:Paul Wouters<[email protected]>
日 期:2026年10月23日 21:43:31
收件人:Foundation<[email protected]>
抄 送:<[email protected]>; dnsop<[email protected]>
主 题:Re: [DNSOP] RRTYPE Allocation Request: UTXO (proposed code 260) & 
Supplementary Architecture Evidence

On Fri, 23 Oct 2026, Foundation wrote:

> This is a formal request for the allocation of a new DNS Resource Record Type 
> parameter under the terms of RFC 5395. The proposed RRTYPE mnemonic is "UTXO" 
> and the
> proposed code is 260.

I've read through the information and left very confused about how the
UTXO RRtype is supposed to map to individual user wallets. DNS is
extremely tricky to map to individual users, see for example

https://www.rfc-editor.org/info/rfc7929/

The only example I see in the draft is mapping the entire domain, eg
"example.test" having 1 UTXO wallet ?

I would also like to see at least an informative reference to UTXO, so I
have some idea of what and how much data is being put into this proposed
RRtype to better evaluate its planned DNS usage.

Paul

_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to