Status of This Memo

This is an Internet Standards Track document.¶

This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 7841.¶

Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at https://www.rfc-editor.org/info/rfc9606.¶

Copyright Notice

Copyright (c) 2024 IETF Trust and the persons identified as the document authors. All rights reserved.¶

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶

Table of Contents

1. Introduction

Historically, DNS clients communicated with recursive resolvers without needing to know anything about the features supported by these resolvers. However, more and more recursive resolvers expose different features that may impact delivered DNS services (privacy preservation, filtering, transparent behavior, etc.). DNS clients can discover and authenticate encrypted DNS resolvers provided by a local network, for example, using the Discovery of Network-designated Resolvers (DNR) [RFC9463] and the Discovery of Designated Resolvers (DDR) [RFC9462]. However, these DNS clients can't retrieve information from the discovered recursive resolvers about their capabilities to feed the resolver selection process. Instead of depending on opportunistic approaches, DNS clients need a more reliable mechanism to discover the features that are configured on these resolvers.¶

This document fills that void by specifying a mechanism that allows communication of DNS resolver information to DNS clients for use in resolver selection decisions. For example, the resolver selection procedure may use the retrieved resolver information to prioritize privacy-preserving resolvers over those that don't enable QNAME minimisation [RFC9156]. Another example is when a DNS client selects a resolver based on its filtering capability. For instance, a DNS client can choose a resolver that filters domains according to a security policy using the Blocked (15) Extended DNS Error (EDE) [RFC8914]. Alternatively, the client may have a policy not to select a resolver that forges responses using the Forged Answer (4) EDE. However, it is out of the scope of this document to define the selection procedure and policies. Once a resolver is selected by a DNS client, and unless explicitly mentioned, this document does not interfere with that resolver's DNS operations.¶

Specifically, this document defines a new resource record (RR) type for DNS clients to query the recursive resolvers. The initial information that a resolver might want to expose is defined in Section 5. That information is scoped to cover properties that are used to infer privacy and transparency policies of a resolver. Other information can be registered in the future per the guidance in Section 8.2. The information is not intended for end-user consumption.¶

2. Terminology

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶

This document makes use of the terms defined in [RFC8499]. The following additional terms are used:¶

Encrypted DNS:

Refers to a DNS scheme where DNS exchanges are transported over an encrypted channel between a DNS client and server (e.g., DNS over HTTPS (DoH) [RFC8484], DNS over TLS (DoT) [RFC7858], or DNS over QUIC (DoQ) [RFC9250]).¶

Encrypted DNS resolver:

Refers to a DNS resolver that supports any encrypted DNS scheme.¶

Reputation:

Defined as "the estimation in which an identifiable actor is held, especially by the community or the Internet public generally" per Section 1 of [RFC7070].¶

3. Retrieving Resolver Information

A DNS client that wants to retrieve the resolver information may use the RR type "RESINFO" defined in this document. The content of the RDATA in a response to a query for RESINFO RR QTYPE is defined in Section 5. If the resolver understands the RESINFO RR type, the RRset MUST have exactly one record. Invalid records MUST be silently ignored by DNS clients. RESINFO is a property of the resolver and is not subject to recursive resolution.¶

A DNS client can retrieve the resolver information using the RESINFO RR type and the QNAME of the domain name that is used to authenticate the DNS resolver (referred to as the Authentication Domain Name (ADN) in DNR [RFC9463]).¶

If the Special-Use Domain Name "resolver.arpa", defined in [RFC9462], is used to discover an encrypted DNS resolver, the client can retrieve the resolver information using the RESINFO RR type and QNAME of "resolver.arpa". In this case, a client has to contend with the risk that a resolver does not support RESINFO. The resolver might pass the query upstream, and then the client can receive a positive RESINFO response from either a legitimate DNS resolver or an attacker.¶

The DNS client MUST set the Recursion Desired (RD) bit of the query to 0. The DNS client MUST discard the response if the AA flag in the response is set to 0, indicating that the DNS resolver is not authoritative for the response.¶

If a group of resolvers is sharing the same ADN and/or anycast address, then these instances SHOULD expose a consistent RESINFO.¶

4. Format of the Resolver Information

The resolver information record uses the same format as DNS TXT records. The format rules for TXT records are defined in the base DNS specification (Section 3.3.14 of [RFC1035]) and are further elaborated in the DNS-based Service Discovery (DNS-SD) specification (Section 6.1 of [RFC6763]). The recommendations to limit the TXT record size are discussed in Section 6.1 of [RFC6763].¶

Similar to DNS-SD, the RESINFO RR type uses "key/value" pairs to convey the resolver information. Each key/value pair is encoded using the format rules defined in Section 6.3 of [RFC6763]. Using standardized key/value syntax within the RESINFO RR type makes it easier for future keys to be defined. If a DNS client sees unknown keys in a RESINFO RR type, it MUST silently ignore them. The same rules for the keys, as defined in Section 6.4 of [RFC6763], MUST be followed for RESINFO.¶

Resolver information keys MUST either be defined in the IANA registry (Section 8.2) or begin with the substring "temp-" for names defined for local use only.¶

5. Resolver Information Keys/Values

The following resolver information keys are defined:¶

qnamemin:

The presence of this key indicates that the DNS resolver supports QNAME minimisation [RFC9156] to improve DNS privacy. Note that, per the rules for the keys defined in Section 6.4 of [RFC6763], if there is no '=' in a key, then it is a boolean attribute, simply identified as being present, with no value.¶

The presence of this key indicates that the DNS resolver is configured to minimise the amount of privacy-sensitive data sent to an authoritative name server.¶

This is an optional attribute.¶

exterr:

If the DNS resolver supports the EDE option defined in [RFC8914] to return additional information about the cause of DNS errors, the value of this key lists the possible EDE codes that can be returned by this DNS resolver. A value can be an individual EDE or a range of EDEs. Range values MUST be identified by "-". When multiple non-contiguous values are present, these values MUST be comma-separated.¶

Returned EDEs (e.g., Blocked (15), Censored (16), and Filtered (17)) indicate whether the DNS resolver is configured to reveal the reason why a query was filtered/blocked when such an event happens. If the resolver's capabilities are updated to include new similar error codes, the resolver can terminate the TLS session, prompting the client to initiate a new TLS connection and retrieve the resolver information again. This allows the client to become aware of the resolver's updated capabilities. Alternatively, if the client receives an EDE for a DNS request, but that EDE was not listed in the "exterr", the client can query the resolver again to learn about the updated resolver's capabilities to return new error codes. If a mismatch still exists, the client can identify that the resolver information is inaccurate and discard it.¶

This is an optional attribute.¶

infourl:

A URL that points to the generic unstructured resolver information (e.g., DoH APIs supported, possible HTTP status codes returned by the DoH server, or how to report a problem) for troubleshooting purposes. The server that exposes such information is called "resolver information server".¶

The resolver information server MUST support only the content-type "text/html" for the resolver information. The DNS client MUST reject the URL as invalid if the scheme is not "https". Invalid URLs MUST be ignored. The URL MUST be treated only as diagnostic information for IT staff. It is not intended for end-user consumption as the URL can possibly provide misleading information.¶

This key can be used by IT staff to retrieve other useful information about the resolver and also the procedure to report problems (e.g., invalid filtering).¶

This is an optional attribute.¶

New keys can be defined as per the procedure defined in Section 8.2.¶

6. An Example

Figure 1 shows an example of a published resolver information record.¶

resolver.example.net. 7200 IN RESINFO qnamemin exterr=15-17
                      infourl=https://resolver.example.com/guide

As mentioned in Section 3, a DNS client that discovers the ADN "resolver.example.net" of its resolver using DNR will issue a query for RESINFO RR QTYPE for that ADN and will learn that:¶

• the resolver enables QNAME minimisation,¶
• the resolver can return Blocked (15), Censored (16), and Filtered (17) EDEs, and¶
• more information can be retrieved from "https://resolver.example.com/guide".¶

7. Security Considerations

DNS clients communicating with discovered DNS resolvers MUST use one of the following measures to prevent DNS response forgery attacks:¶

1. Establish an authenticated secure connection to the DNS resolver.¶
2. Implement local DNSSEC validation (Section 10 of [RFC8499]) to verify the authenticity of the resolver information.¶

It is important to note that, of these two measures, only the first one can apply to queries for "resolver.arpa".¶

An encrypted resolver may return incorrect information in RESINFO. If the client cannot validate the attributes received from the resolver, which will be used for resolver selection or displayed to the end user, the client should process those attributes only if the encrypted resolver has sufficient reputation according to local policy (e.g., user configuration, administrative configuration, or a built-in list of reputable resolvers). This approach limits the ability of a malicious encrypted resolver to cause harm with false claims.¶

8. IANA Considerations

Acknowledgments

This specification leverages the work that has been documented in [RESINFO].¶

Thanks to Tommy Jensen, Vittorio Bertola, Vinny Parla, Chris Box, Ben Schwartz, Tony Finch, Daniel Kahn Gillmor, Eric Rescorla, Shashank Jain, Florian Obser, Richard Baldry, and Martin Thomson for the discussion and comments.¶

Thanks to Mark Andrews, Joe Abley, Paul Wouters, and Tim Wicinski for the discussion on RR formatting rules.¶

Special thanks to Tommy Jensen for the careful and thoughtful Shepherd review.¶

Thanks to Johan Stenstam and Jim Reid for the dns-dir reviews, Ray Bellis for the RRTYPE allocation review, Arnt Gulbrandsen for the ART review, and Mallory Knodel for the gen-art review.¶

Thanks to Éric Vyncke for the AD review.¶

Thanks to Gunter Van de Velde, Erik Kline, Paul Wouters, Orie Steele, Warren Kumari, Roman Danyliw, and Murray Kucherawy for the IESG review.¶

Authors' Addresses