RFC 8552 | DNS AttrLeaf | March 2019 |
Crocker | Best Current Practice | [Page] |
Formally, any DNS Resource Record (RR) may occur under any domain name. However, some services use an operational convention for defining specific interpretations of an RRset by locating the records in a DNS branch under the parent domain to which the RRset actually applies. The top of this subordinate branch is defined by a naming convention that uses a reserved node name, which begins with the underscore character (e.g., "_name"). The underscored naming construct defines a semantic scope for DNS record types that are associated with the parent domain above the underscored branch. This specification explores the nature of this DNS usage and defines the "Underscored and Globally Scoped DNS Node Names" registry with IANA. The purpose of this registry is to avoid collisions resulting from the use of the same underscored name for different services.¶
This memo documents an Internet Best Current Practice. ¶
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 BCPs 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/rfc8552. ¶
Copyright (c) 2019 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 Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. ¶
The core Domain Name System (DNS) technical specifications ([RFC1035] and [RFC2181]) assign no
semantics to domain names or their parts, and no constraints upon
which resource record (RR) types are permitted to be stored under
particular names [RFC1035] [RFC2181].
Over time, some leaf node names, such as
www
and ftp
,
have come to imply support for particular services, but this is a
matter of operational convention rather than defined protocol
semantics.
This freedom in the basic technology has permitted a wide range of
administrative and semantic policies to be used -- in parallel. DNS
data semantics have been limited to the specification of particular
resource record types on the expectation that new resource record
types would be added as needed. Unfortunately, the addition of new resource record types has proven
extremely challenging, with significant adoption and use barriers
occurring over the life of the DNS.
¶
As an alternative to defining a new RR TYPE, some DNS service enhancements call for using an existing resource record type but specifying a restricted scope for its occurrence. Scope is meant as a static property, not one dependent on the nature of the query. It is an artifact of the DNS name. That scope is a leaf node containing the specific resource record sets that are formally defined and constrained. Specifically: ¶
Because the DNS rules for a "host" (host name) do not allow use of the underscore character, the underscored name is distinguishable from all legal host names [RFC0952]. Effectively, this convention for naming leaf nodes creates a space for the listing of "attributes" -- in the form of resource record types -- that are associated with the parent domain above the underscored sub-branch. ¶
The scoping feature is particularly useful when generalized
resource record types are used -- notably
TXT
, SRV
,
and URI
[RFC1035] [RFC2782] [RFC6335] [RFC7553]. It provides
efficient separation of one use of them from others. Absent this
separation, an undifferentiated mass of these RRsets is returned
to the DNS client, which then must parse through the internals of
the records in the hope of finding ones that are relevant. Worse,
in some cases, the results are ambiguous because a record type
might not adequately self-identify its specific purpose. With
underscore-based scoping, only the relevant RRsets are
returned.¶
A simple example is DomainKeys Identified Mail (DKIM) [RFC6376], which
uses _domainkey
to define a place
to hold a TXT record containing signing information for the
parent domain.¶
This specification formally defines how underscored names are used as "attribute" enhancements for their parent domain names. For example, the domain name "_domainkey.example." acts as an attribute of the parent domain name "example.". To avoid collisions resulting from the use of the same underscored names for different applications using the same resource record type, this document establishes the "Underscored and Globally Scoped DNS Node Names" registry with IANA. Use of such node names, which begin with an underscore character, is reserved when they are the underscored name closest to the DNS root; as in that case, they are considered "global". Underscored names that are farther down the hierarchy are handled within the scope of the global underscored node name. ¶
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. ¶
Some resource record types are used in a fashion that can create scaling problems if an entire RRset associated with a domain name is aggregated in the leaf node for that name. An increasingly popular approach, with excellent scaling properties, places the RRset under a specially named branch, which is in turn under the node name that would otherwise contain the RRset. The rules for naming that branch define the context for interpreting the RRset. That is, rather than: ¶
domain-name.example / RRset¶
the arrangement is: ¶
_branch.domain-name.example / RRset¶
A direct lookup to the subordinate leaf node produces only the desired record types, at no greater cost than a typical DNS lookup.¶
As defined in [RFC1034], the DNS uses names organized in a tree-structured or hierarchical fashion. A domain name might have multiple node names that begin with the underscore character (e.g., "_name"). A global underscored node name is the one that is closest to the root of the DNS hierarchy, also called the highest level or topmost. In the presentation convention described in Section 3.1 of [RFC1034], this is the rightmost name beginning with an underscore. In other presentation environments, it might be positioned differently. To avoid concern for the presentation variations, the qualifier "global" is used here.¶
DNS wildcards interact poorly with underscored names in two ways:¶
Since wildcards are only interpreted as leaf names, one cannot create the equivalent of a wildcard name for prefixed names. A name such as label.*.example.com is not a wildcard. ¶
Conversely, a wildcard such as *.example.com can match any name including an underscored name. So, a wildcard might match an underscored name, returning a record that is the type controlled by the underscored name but is not intended to be used in the underscored context and does not conform to its rules. ¶
Originally, different uses of underscored node names
developed largely without coordination. For TXT records, there is
no consistent, internal syntax that permits distinguishing among
the different uses. In the case of the SRV RR and URI RR,
distinguishing among different types of use was part of the
design (see [RFC2782] and [RFC7553]). The
SRV and URI specifications serve as templates, defining RRs that
might only be used for specific applications when there is an
additional specification. The template definition included
reference to two levels of tables of names from which
underscored names should be drawn. The lower-level (local scope)
set of _service
names is defined in
terms of other IANA tables, namely any table with symbolic names.
The upper-level (global scope) SRV naming field is
_proto
, although its pool of names is
not explicitly defined. ¶
The aggregate effect of these independent efforts was a long list of underscored names that were reserved without coordination, which invites an eventual name-assignment collision. The remedy is this base document and a companion document ([RFC8553]), which define a registry for these names and attempt to register all those already in use as well as to direct changes to the pre-registry specifications that used global underscored node names. ¶
A registry for global DNS node names that begin with an underscore is defined here. The purpose of the "Underscored and Globally Scoped DNS Node Names" registry is to avoid collisions resulting from the use of the same underscored name for different applications. ¶
An underscored name defines the scope of use for specific resource record types, which are associated with the domain name that is the "parent" to the branch defined by the underscored name. A given name defines a specific, constrained context for one or more RR TYPEs, where use of such record types conforms to the defined constraints. ¶
Structurally, the registry is defined as a single, flat table of RR TYPEs, under node names beginning with underscore. In some cases, such as for use of an SRV record, the full scoping name might be multi-part, as a sequence of underscored names. Semantically, that sequence represents a hierarchical model, and it is theoretically reasonable to allow reuse of a subordinate underscored name in a different, global underscored context; that is, a subordinate name is meaningful only within the scope of the global underscored node name. Therefore, they are ignored by this "Underscored and Globally Scoped DNS Node Names" registry. This registry is for the definition of highest-level -- that is, global -- underscored node name used.¶
NAME |
---|
_service1 |
_protoB._service2 |
_protoB._service3 |
_protoC._service3 |
_useX._protoD._service4 |
_protoE._region._authority |
Only global underscored node names are registered in the "Underscored and Globally Scoped DNS Node Names" registry. From the example above, that would mean _service1, _service2, _service3, _service 4, and _authority would be listed in the IANA registry. ¶
That is, if a scheme using a global underscored node name has one or more subordinate levels of underscored node naming, the namespaces from which names for those lower levels are chosen are controlled by the parent underscored node name. Each registered global underscored node name owns a distinct, subordinate namespace.¶
This section provides guidance for specification writers, with a basic template they can use, to register new entries in the "Underscored and Globally Scoped DNS Node Names" registry. The text can be added to specifications using RR TYPE / _NODE NAME combinations that have not already been registered:¶
RR Type | _NODE NAME | Reference |
---|---|---|
{RR TYPE} | _{DNS global node name} | {citation for the document making the addition.} |
IANA has established the "Underscored and Globally Scoped DNS Node Names" registry. This section describes the registry, the definitions, the initial entries, the use of_ta and _example, and the use of [RFC8126] as guidance for expert review. IANA has also updated the "Enumservices Registrations" registry with a pointer to this document.¶
The "Underscored and Globally Scoped DNS Node Names" registry includes any DNS node name that begins with the underscore character ("_", ASCII 0x5F) and is the underscored node name closest to the root; that is, it defines the highest level of a DNS branch under a "parent" domain name. ¶
A registry entry contains: ¶
Each RR TYPE that is to be used with a _Node Name MUST have a separate registry entry. ¶
The initial entries in the registry are as follows:¶
RR Type | _NODE NAME | Reference |
---|---|---|
* | _example | Section 4.1.4 |
NULL | _ta-* {Section 4.1.3} | [RFC8145] |
OPENPGPKEY | _openpgpkey | [RFC7929] |
SMIMEA | _smimecert | [RFC8162] |
SRV | _dccp | [RFC2782] |
SRV | _http | [RFC4386] |
SRV | _ipv6 | [RFC5026] |
SRV | _ldap | [RFC4386] |
SRV | _ocsp | [RFC4386] |
SRV | _sctp | [RFC2782] |
SRV | _sip | [RFC5509] |
SRV | _tcp | [RFC2782] |
SRV | _udp | [RFC2782] |
SRV | _xmpp | [RFC3921] |
TLSA | _dane | [RFC7671] |
TLSA | _sctp | [RFC6698] |
TLSA | _tcp | [RFC6698] |
TLSA | _udp | [RFC6698] |
TXT | _acme-challenge | [RFC8555] |
TXT | _dmarc | [RFC7489] |
TXT | _domainkey | [RFC6376] |
TXT | _mta-sts | [RFC8461] |
TXT | _spf | [RFC7208] |
TXT | _sztp | [ZEROTOUCH] |
TXT | _tcp | [RFC6763] |
TXT | _udp | [RFC6763] |
TXT | _vouch | [RFC5518] |
URI | _acct | [RFC6118] |
URI | _dccp | [RFC7566] |
URI | _email | [RFC6118] |
URI | _ems | [RFC6118] |
URI | _fax | [RFC6118] |
URI | _ft | [RFC6118] |
URI | _h323 | [RFC6118] |
URI | _iax | [RFC6118] |
URI | _ical-access | [RFC6118] |
URI | _ical-sched | [RFC6118] |
URI | _ifax | [RFC6118] |
URI | _im | [RFC6118] |
URI | _mms | [RFC6118] |
URI | _pres | [RFC6118] |
URI | _pstn | [RFC6118] |
URI | _sctp | [RFC6118] |
URI | _sip | [RFC6118] |
URI | _sms | [RFC6118] |
URI | _tcp | [RFC6118] |
URI | _udp | [RFC6118] |
URI | _unifmsg | [RFC6118] |
URI | _vcard | [RFC6118] |
URI | _videomsg | [RFC6118] |
URI | _voice | [RFC6118] |
URI | _voicemsg | [RFC6118] |
URI | _vpim | [RFC6118] |
URI | _web | [RFC6118] |
URI | _xmpp | [RFC6118] |
Under the NULL RR Type, the entry _ta-*
denotes all node names beginning with the string
_ta-*
. It does NOT refer to a DNS
wildcard specification.¶
This section provides guidance for expert review of registration requests in the "Underscored and Globally Scoped DNS Node Names" registry.¶
The review is for the purposes of ensuring that:¶
For the purposes of this expert review, other matters of the specification's technical quality, adequacy, or the like are outside of scope. ¶
The following note has been added to the "Enumservice Registrations" registry:¶
This memo raises no security issues.¶
Thanks go to Bill Fenner, Dick Franks, Tony Hansen, Martin Hoffmann, Paul Hoffman, Peter Koch, Olaf Kolkman, Murray Kucherawy, John Levine, Benno Overeinder, and Andrew Sullivan for diligent review of the (much) earlier draft versions. For the later enhancements, thanks to Stephane Bortzmeyer, Alissa Cooper, Bob Harold, Joel Jaeggli, Benjamin Kaduk, Mirja Kuehlewind, Warren Kumari, John Levine, Benno Overeinder, Eric Rescorla, Adam Roach, Petr Spacek, Ondrej Sury, Paul Vixie, Tim Wicinski, and Paul Wouters.¶
Special thanks to Ray Bellis for his persistent encouragement to continue this effort, as well as the suggestion for an essential simplification to the registration model.¶