RDAP Extensions

Background The Registration Data Access Protocol (RDAP) defines a uniform means to access data from Internet operations registries, specifically Domain Name Registries (DNRs), Regional Internet Registries (RIRs), and other registries serving Internet Number Resources (INRs). RDAP queries are defined in and RDAP responses are defined in . RDAP contains a means to define extensions for queries not found in and responses not found in . RDAP extensions are also described in . This document describes the requirements for RDAP extension definition and use, clarifying ambiguities and defining additional semantics and options that were previously implicit.

Summary of Updates This document updates , , and in the following areas:

provides additional guidance on the purpose of RDAP extension identifiers, including their registration for the purpose of avoiding namespace collisions, as well as for signaling server policy/behavior.
clarifies the usage of extension identifiers in RDAP URLs and formally defines their usage with child path segments and query parameters.
clarifies the usage of extension identifiers in responses, including for new object classes and search result arrays.
documents behavior considerations for extension implementers.
documents behavior considerations for extension authors, covering redirects (), referrals (), and versioning ().
documents existing extensions that were not registered in accordance with the requirements from the relevant RDAP documents (, , and ).
and provide further guidance on registrations into the IANA RDAP registries and recommendations for the expert review processes for these registries.

Document Terms 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 when, and only when, they appear in all capitals, as shown here.

Identifiers

Purpose describes the identifier used to signify RDAP extensions and the IANA registry into which RDAP extensions are to be registered. When in use in RDAP, extension identifiers are prepended to URL path segments, URL query parameters, and JSON object member names (herein further referred to as "JSON names"). They are also included in the "rdapConformance" member of each response that relies on the extension, so that clients can determine the extensions being used by the server for that response. The "/help" response returns an "rdapConformance" member containing the identifiers for all extensions used by the server. The main purpose of the extension identifier is to act as a namespace, preventing collisions between elements from different extensions. Additionally, implementers and operators can use the extension identifiers to find extension definitions via an IANA registry.

Profile and Marker Extensions While the RDAP extension mechanism was created to extend RDAP queries and/or responses, extensions can also be used to signal server policy (for example, specifying the conditions of use for existing response structures). Extensions that are primarily about signaling server policy are often called "profiles". Some extensions exist to denote the usage of values placed into an IANA registry, such as the IANA RDAP registries, or the usage of extensions for specifications used in RDAP responses, such as extended vCard/jCard properties. Such extensions exist to "mark" these usages and are often called "marker" extensions. For example, an extension may be used to signal desired processing of a "rel" attribute in a "links" array, where the "rel" value is registered in the Link Relations Registry (https://www.iana.org/assignments/link-relations/link-relations.xhtml): When defining the usage of link relations, extensions should specify the media types expected to be used with those link relations. Regardless of the category of the extension, its usage may also leverage the appearance of its identifier in the "rdapConformance" array. Clients may use the "/help" query as defined in to discover the extensions in use by the server.

Multiple Identifiers in Single Extension Extension specifications have customarily defined only one extension identifier. However, there is no explicit limit on the number of extension identifiers that may be defined in a single extension specification. The main reason for defining multiple identifiers is to reserve multiple namespaces in URLs or responses: see e.g. .

Syntax In brief, RDAP extension identifiers start with an alphabetic character and may contain alphanumeric characters and "_" (underscore) characters. This formulation was explicitly chosen to allow compatibility with variable names in programming languages and transliteration with XML. RDAP extension identifiers have no explicit structure, and are opaque insofar as no inner-meaning can be "seen" in them. RDAP extensions MUST NOT define an extension identifier that when prepended to an underscore character may collide with an existing extension identifier. For example, if there were a pre-existing identifier of "foo_bar", another extension could not define the identifier "foo". Likewise, if there were a pre-existing identifier of "foo_bar", another extension could not define the identifier "foo_bar_buzz". However, an extension could define "foo" if there were a pre-existing definition of "foobar", and vice versa. For this reason, usage of an underscore character in RDAP extension identifiers is NOT RECOMMENDED. Implementers should be aware that many existing extension identifiers do contain underscore characters. does not explicitly state that extension identifiers are case-sensitive. This document updates the formulation in to explicitly note that extension identifiers are case-sensitive, and extension identifiers MUST NOT be registered where a new identifier is a mixed-case version of an existing identifier. For example, if "lunarNIC" is already registered as an identifier, a new registration with "lunarNic" (note the lowercase if "ic" in "Nic") would not be allowed.

Usage in Requests

Usage in Paths describes the use of extension identifiers in formulating URLs to query RDAP servers. The extension identifiers are to be prepended to the path segments they use. For example, if an extension uses the identifier "foobar", then the path segments used in that extension are prepended with "foobar_". If the "foobar" extension defines paths "fizz" and "fazz", the URLs for this extension would be like so: While describes the extension identifier as a prepended string to a path segment, it does not describe the usage of the extension identifier as a path segment which may have child path segments. This document updates to allow the usage of extension identifiers as path segments which may have child path segments. For example, if the "foobar" extension defines the child paths "fizz" and "fazz", the URLs for this extension would be like so: Extensions defining new URL paths MUST explicitly define the expected responses for each new URL path. New URL paths may return existing object classes or search results as defined in , object classes or search results defined by the extension (see and below), or object classes or search results from other extensions.

Usage in Query Parameters Although describes the use of URL query strings, it does not define their use with extensions. instructs servers to ignore unknown query parameters. Therefore, the use of query parameters, whether prefixed with an extension identifier or not, is not supported by and . Despite this, there are several extensions that do specify query parameters. This document updates with regard to the use of RDAP extension identifiers in URL query parameters. When an RDAP extension defines query parameters to be used with a URL path that is not defined by that RDAP extension, those query parameter names SHOULD be constructed in the same manner as URL path segments (that is, extension identifier + '_' + parameter name). (See section regarding when an extension identifier may be omitted.) When an RDAP extension defines query parameters to be used with a URL path defined by that RDAP extension, prefixing of query parameters is not required. In this situation, the URL path operates as a namespace for the query parameters, so there is no risk of collision with parameters defined elsewhere. See and for other guidance on the use of query parameters, and see and regarding constraints on the usage of query parameters.

Usage in Responses

Basic Requirements describes the use of extension identifiers in the JSON returned by RDAP servers. Just as in URLs, the extension identifier is prepended to JSON names to create a namespace so that the JSON name from one extension will not collide with the JSON name from another extension. Just as with unknown query parameters in URLs, clients are to ignore unknown JSON names. The example given in is as follows: In this example, the extension identified by "lunarNIC" is prepended to the member names of both a JSON string and a JSON array. As requires the use of the "rdapConformance" data structure, and the "objectClassName" string is required of all object class instances, the complete example from above would be:

Child JSON Values Prefixing of the extension identifier is not required for children of a prefixed JSON object defined by an RDAP extension. The following example shows this use with a JSON object: Here the JSON name "lunarNIC_author" will separate the JSON from other extensions that may have an "author" structure. But the JSON contained within "lunarNIC_author" need not be prepended, as collision is avoided by the use of "lunarNIC_author".

Object Classes in Extensions As described in and , an extension may define new paths in URLs. If the extension describes the behavior of an RDAP query using that path to return an instance of a new class of RDAP object, the JSON names are not required to be prepended with the extension identifier as described in . However, the extension MUST define the value for the "objectClassName" string which is used by clients to evaluate the type of the response. To avoid collisions with object classes defined in other extensions, the value for the "objectClassName" MUST be prepended with the extension identifier, in the same way as for URL paths, query parameters, and JSON names: It is RECOMMENDED that object class names comprise lowercase ASCII characters, and that the "_" (underscore) character be used as a word separator. Though "objectClassName" is a string and does define one object class name with a space separator (i.e. "ip network"), usage of the space character or any other character that requires URL-encoding is NOT RECOMMENDED. When object class names are also used in URLs, extensions MUST specify that the names are to be URL-encoded as defined in if the object class names contain any characters requiring URL-encoding.

Search Results in Extensions As described in and , an extension may define new paths in URLs. If the extension describes the behavior of an RDAP query using the path to return a new RDAP search result, the JSON name of the search result MUST be prepended with the extension identifier (to avoid collision with search results defined in other extensions). If the search result contains object class instances defined by the extension, each instance MUST have an "objectClassName" string as defined in . For example:

Bare Extension Identifiers Some RDAP extensions define only one JSON value and do not prefix it with their RDAP extension identifier, instead using the extension identifier as the JSON name for that JSON value. That is, the extension identifier is used "bare" and not appended with an underscore character and subsequent names. Consider the example in . Using the bare extension identifier pattern, that example could be written as: Usage of a bare extension identifier contravenes the guidance in . This document updates to explicitly allow this pattern. Along similar lines, an extension may define a single new object class, and use the extension's identifier as the object class name.

rdapConformance Population offers the following guidance on including extension identifiers in the "rdapConformance" member of an RDAP response: A strict interpretation of this wording where "construction of the response" refers to the JSON structure only would rule out the use of extension identifiers, which are in common use in RDAP. This document updates the guidance. For responses to queries other than "/help", a response MUST include in the "rdapConformance" array only those extension identifiers necessary for a client to deserialize the JSON and understand the semantic meaning of the content within the JSON, and each extension identifier MUST be free from conflict with the other identifiers with respect to their syntax and semantics. Note that this document does not update the guidance from regarding "/help" responses and the "rdapConformance" array. When a server implementation supports multiple extensions, it is RECOMMENDED that the server also support and return versioning information such as that defined by .

Camel Casing The styling convention used in for JSON names is often called "camel casing", in reference to the hump of a camel. In this style, the first letter of every word, except the first word, composing a name is capitalized. This convention was adopted to visually separate the namespace from the name, with an underscore between them. Extension authors SHOULD use camel casing for JSON names defined in extensions.

Identifier Omission Though all RDAP extensions are to be registered in the IANA RDAP Extensions Registry, there is an implicit two-class system of extensions that comes from the ownership of the RDAP specifications by the IETF: extensions created by the IETF and extensions not created by the IETF. In the perspective of how extension identifiers are used as namespace separators, extensions created by the IETF are not required to use the extension identifier as a prefix in requests and responses, as the IETF can coordinate its own activities to avoid name collisions. In practice, most extensions owned by the IETF do use extension identifiers as prefixes in their requests and responses. RDAP extensions not defined by the IETF MUST use the extension identifier as a prefix in accordance with this document, , , and . RDAP extensions defined by the IETF SHOULD use the extension identifier as a prefix or as a bare extension identifier (see ). IETF-defined RDAP extensions that do not follow this guidance MUST describe why it is not being followed.

Extension Implementer Considerations

Redirects describes the use of redirects in RDAP. Redirects are prominent in the discovery of authoritative RIR servers, as the process outlined in , which uses IANA allocations, does not account for transfers of resources between RIRs. instructs servers to ignore unknown query parameters. As it relates to issuing URLs for redirects, servers MUST NOT blindly copy query parameters from a request to a redirect URL as query parameters may contain sensitive information, such as security credentials, not relevant to the target server of the URL. Following the advice in , servers SHOULD only place query parameters in redirect URLs when it is known by the origin server (the server issuing the redirect) that the target server (the server referenced by the redirect) can process the query parameter and the contents of the query parameter are appropriate to be received by the target.

Extension Author Considerations

Redirects As it is unlikely that every server in a cross-authority, redirect scenario will be upgraded to process every new extension, extensions should not rely on query parameters alone to convey information about a resource, as query parameters are not guaranteed to survive a redirect. This does not mean extensions are prohibited from using query parameters, but rather that the use of query parameters must be applied for the scenarios appropriate for the use of the extension. Therefore, extensions SHOULD NOT rely on query parameters when the extension is to be used in scenarios requiring clients to find authoritative servers, or other scenarios using redirects among servers of differing authorities. Extensions MAY use query parameters in scenarios where the client has a priori knowledge of the authoritative server to which queries are to be sent, and will be sending queries to that server directly. Searches () are an example scenario where a client will be operating in this way. In general, extension authors should be mindful of situations requiring clients to directly handle redirects at the RDAP layer. Some clients may not be utilizing HTTP libraries that provide such an option, and many HTTP client libraries that do provide the option do not provide it as a default behavior. Additionally, requiring clients to handle redirects at the RDAP layer adds complexity to the client in that additional logic must be implemented to handle redirect loops, parameter deconfliction, and URL encoding. The guidance given in exists to simplify clients, especially those constructed with shell scripts and HTTP command-line utilities.

Referrals It is common in the RDAP ecosystem to link from one RDAP resource to another, such as can be found in domain registrations in gTLD DNRs. These are typically conveyed in the link structure defined in and use the "application/rdap+json" media type. For example: Extensions MUST explicitly define any required behavioral changes to the processing of referrals. If an extension does not make any provision in this respect, clients MUST assume the information provided by referrals requires no additional processing or modification to use in the dereferencing of the referral.

Extension Versioning As stated in , RDAP extensions are opaque, and they possess no explicit version despite the fact that some extension identifiers include trailing numbers. That is, RDAP extensions without an explicitly-defined versioning scheme are opaquely versioned. For example, "fizzbuzz_1" may be the successor to "fizzbuzz_0", but it may also be an extension for a completely separate purpose. Only consultation of the definition of "fizzbuzz_1" will determine its relationship with "fizzbuzz_0". Additionally, "fizzbuzz_99" may be the predecessor of "fizzbuzz_0". If a future RFC defines a versioning scheme, an RDAP extension definition MUST explicitly denote its compliance with that scheme.

Backwards-Compatible Changes If an RDAP extension author wants to publish a new version of an extension that is backwards-compatible with the previous version, then one option is for the new version of the extension to define a new identifier, as well as requiring that both the previous identifier and the new identifier be included in the "rdapConformance" array of responses. That way, clients relying on the previous version of the extension will continue to function as intended, while clients wanting to make use of the newer version of the extension can check for the new identifier in the response. This approach can be used for an arbitrary number of new backwards-compatible versions of a given extension. For an extension with a large number of backwards-compatible successor versions, this may lead to a large number of identifiers being included in responses. An extension author may consider excluding older identifiers from the set required by new successor versions, based on data about client use/support or similar.

Backwards-Incompatible Changes With the current extension model, an extension with a backwards-incompatible change is indistinguishable from a new, unrelated extension. Implementers of such changes should consider the following:

whether the new version of the extension can be provided alongside the old version of the extension, so that a service can simply support both during a transition period;
whether some sort of client signaling should be supported, so that clients can opt for the old or new version of the extension in responses that they receive (see for an example of how this might work); and
whether the extension itself should define how versioning is handled within the extension documentation.

Extension Specification Content The primary purpose of an RDAP extension specification is to aid in the implementation of RDAP clients. Extension authors should consider the following content guidelines:

Examples of RDAP JSON should be generously given, especially in areas of the specification which may be complex or difficult to describe with prose.
Normative references, i.e. references to materials that are required for the interoperability of the extension, should be stable and non-changing.
Extension specifications should strongly consider making the use of HTTPS with RDAP mandatory if appropriate.

Extension Definitions Extensions must be documented in an RFC or in some other permanent and readily available reference, in sufficient detail that interoperability between independent implementations is possible. Though RDAP gives each extension its own namespace, the definition of an extension may reuse definitions found in the base RDAP specification or in any other registered extension. notes that the extension identifiers provide a "hint" to the client as to how to interpret the response. This wording does not intentionally restrict the extension to defining only JSON values within the extension's namespace. Therefore, an extension may define the use of its own JSON values together with the use of JSON values from other extensions or RDAP specifications. As with the and extensions, the extension may simply signal policy applied to previously-defined RDAP structures.

Existing Extension Registrations The following extensions have been registered with IANA, but do not comply with the requirements set out in the base specifications, as clarified by this document:

Extension identifier: fred
- RDAP conformance value: fred_version_0
- Field/path prefix: fred
Extension identifier: artRecord
- RDAP conformance value: artRecord_level_0
- Field/path prefix: artRecord
Extension identifier: platformNS
- RDAP conformance value: platformNS_level_0
- Field/path prefix: platformNS
Extension identifier: regType
- RDAP conformance value: regType_level_0
- Field/path prefix: regType

Client authors should be aware that responses that make use of these extensions may require special handling on the part of the client. Also, while these extensions will be retained in the registry, future extensions that are similarly non-compliant will not be registered. To avoid any confusion with the operation of the existing entries, an extension registration that attempts to use one of the RDAP conformance values given in this section as an extension identifier (and so as an RDAP conformance value also) will be rejected.

IANA Considerations

RDAP Extensions Registry defines the RDAP Extensions Registry (https://www.iana.org/assignments/rdap-extensions/rdap-extensions.xhtml). This document does not change the RDAP Extensions Registry nor its purpose. However, this document does update the procedures to be used by its expert reviewers. The RDAP Extensions Registry should have as a minimum three expert reviewers and ideally four or five. An expert reviewer assigned to the review of an RDAP extension registration must have another expert reviewer double-check any submitted registration. Expert reviewers are to use the following criteria for extensions defined in this document, , , and . The following is a summary checklist:

Does the extension define an extension identifier following the naming conventions described in and ? For any recommendations regarding naming conventions (guidance given using RECOMMENDED, SHOULD, etc.), does the extension describe the need for departing from the established convention?
If the extension defines new queries, does it clearly describe the expected results of each new query?
Does the extension follow the JSON naming requirements as described in ?
If the extension is a newer version of an older extension, does the extension specification clearly describe if it is backwards-compatible (see ) or backwards-incompatible (see )?
If the extension registers new values in an IANA registry used by RDAP, does it describe how a client is to use those values?

As noted in , any new registration that is a case variant of an existing registration MUST be rejected. RDAP clients SHOULD match values in this registry using case-insensitive matching. Extension authors are encouraged but not required to seek an informal review of their extension by sending a request for review to regext@ietf.org.

RDAP JSON Values Registry defines the RDAP JSON Values Registry in IANA. This registry contains values to be used in the JSON values of RDAP responses. Registrations into this registry may occur in IETF-defined RDAP extensions or via requests to the IANA. Authors of RDAP extensions not defined by the IETF MAY register values in this registry via requests to the IANA. This document does not change the RDAP JSON Values Registry nor its purpose. However, this document does update the procedures for registrations and the processes to be used by its expert reviewers. In addition to the registration of values, RDAP extensions defined by the IETF and other IETF specifications MAY define additional value types (the "type" field). These specifications MUST describe the specific JSON field to be used for each new value type. defines the criteria for the values. Of these, criteria two states:

Values must be strings. They should be multiple words separated by single space characters. Every character should be lowercased. If possible, every word should be given in English and each character should be US-ASCII.

All registrations SHOULD meet these requirements. However, there may be scenarios in which it is more appropriate for the values to follow other requirements, such as for values also used in other specifications or documents. In all cases, it should be understood that additional registrations of RDAP JSON values occurring after the specification of the value's type in the registry may not be recognized by clients, and therefore either ignored or passed on to users without processing. Designated experts MUST reject any registration that is a duplicate of an existing registration, and all registrations are to be considered case-insensitive. That is, any new registration that is a case variant of an existing registration should be rejected. RDAP clients SHOULD match values in this registry using case-insensitive matching. Definitions of new types (see above) MAY additionally constrain the format of values for those new types beyond the specification of this document and . Designated experts MUST evaluate registrations with those criteria. The RDAP JSON Values Registry should have as a minimum three expert reviewers and ideally four or five. An expert reviewer assigned to the review of an RDAP JSON values registration must have another expert reviewer double-check any submitted registration. Expert reviewers are to use the criteria defined in .

Security Considerations describes the usage of query parameters and describes the restrictions extensions must follow to use them. instructs servers to ignore unknown query parameters. As it relates to issuing URLs for redirects, servers MUST NOT blindly copy query parameters from a request to a redirect URL as query parameters may contain sensitive information, such as security credentials or tracking information, not relevant to the target server of the URL. Following the advice in , servers SHOULD only place query parameters in redirect URLs when it is known by the origin server (the server issuing the redirect) that the target server (the server referenced by the redirect) can process the query parameter and the contents of the query parameter are appropriate to be received by the target.

Privacy Considerations describes the usage of query parameters and describes the restrictions extensions must follow to use them. As query parameters have been known to be used to subvert the privacy preferences of users in HTTP-based protocols, server MUST NOT blindly copy query parameters from a request to a redirect URL as described in and extensions MUST follow the constraints of query parameter usage as defined in .

Acknowledgments The following individuals have provided feedback and contributions to the content and direction of this document: James Gould, Daniel Keathley, and Ties de Kock.