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/rfc9536.¶

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

The protocol described in this specification aims to extend the RDAP query capabilities and response to enable reverse search based on the relationships defined in RDAP between an object class for search and a related object class. The reverse search based on the domain-entity relationship is treated as a particular case of such a generic model.¶

RDAP providers willing to implement this specification should carefully consider its implications on the efficiency (see Section 10), the security (see Section 13), and the compliance with privacy regulations (see Section 12) of their RDAP service.¶

2. Reverse Search Path Segment Specification

A generic reverse search path is described by the syntax:¶

{searchable-resource-type}/reverse_search/{related-resource-type}?<search-condition>¶

The path segments are defined as follows:¶

"searchable-resource-type":

It MUST be one of the resource types for search defined in Section 3.2 of [RFC9082] (i.e., "domains", "nameservers", and "entities") or a resource type extension.¶

"related-resource-type":

It MUST be one of the resource types for lookup defined in Section 3.1 of [RFC9082] (i.e., "domain", "nameserver", "entity", "ip", and "autnum") or a resource type extension.¶

"search-condition":

A sequence of "property=search pattern" predicates separated by the ampersand character ('&', US-ASCII value 0x0026).¶

While related-resource-type is defined as having one of a number of different values, the only reverse searches defined in this document are for a related-resource-type of "entity". Reverse searches for the other resource types specified in [RFC9082] and resource type extensions may be defined by future documents.¶

3. Reverse Search Definition

Based on the content of Section 2, defining a reverse search means to define the triple <searchable resource type, related resource type, property> and the mapping with the corresponding RDAP object member. The mapping is done through the use of a JSONPath expression [RFC9535]. Reverse searches are registered in the "RDAP Reverse Search" registry (see Section 11.2.3), whereas reverse search mappings are registered in the "RDAP Reverse Search Mapping" registry (see Section 11.2.4). The reason for having two registries is that it may be possible for a single type of reverse search to rely on different members, depending on the server's configuration (see Section 5).¶

All of the reverse searches defined by this document (see Section 8) have property names that are the same as the name of the RDAP object member that is the subject of the search. For example, the reverse search with the property name "fn" relies on the value of the "fn" member inside the jCard of an entity object. However, it is not necessary that these two names be the same. In particular, remapping of searches as part of the deprecation of an existing member (see Section 5) will typically lead to a member with a different name being used for the search.¶

Servers MUST NOT provide or implement reverse searches or reverse search mappings that are not registered with IANA.¶

4. Reverse Search Properties Discovery

Servers complying with this specification MUST extend the help response [RFC9083] with the "reverse_search_properties" member that contains an array of objects with the following mandatory child members:¶

"searchableResourceType":

the searchable resource type of the reverse search query, as defined in Section 2¶

"relatedResourceType":

the related resource type of the reverse search query, as defined in Section 2¶

"property":

the reverse search property used in the predicate of the reverse search query, as defined in Section 2¶

An example of the help response including the "reverse_search_properties" member is shown in Figure 2¶

5. Reverse Search Properties Mapping

To permit clients to determine the member used by the server for a reverse search, servers MUST detail the mapping that is occurring by adding the "reverse_search_properties_mapping" member to the topmost object of a reverse search response. This data structure is included in the search response, rather than in the help response, because it may differ depending on the query that is sent to the server.¶

Documents that deprecate or restructure RDAP responses such that a registered reverse search is no longer able to be used MUST either note that the relevant reverse search is no longer available (in the case of deprecation) or describe how to continue supporting the relevant search by adding another mapping for the reverse search property (in the case of restructuring).¶

The "reverse_search_properties_mapping" member contains an array of objects with the following mandatory child members:¶

"property":

the reverse search property used in the predicate of the current query, as defined in Section 2¶

"propertyPath":

the JSONPath expression of the object member (or members) corresponding to the reverse search property¶

The searchable and the related resource types are derived from the query, so there is no need to include them in addition to the property in this member.¶

This member MUST be included for all properties used in the search, regardless of whether that property has multiple registered mappings as at the time of the search, because new mappings may be registered at any time.¶

When applied to an object, the JSONPath expression MUST produce a list of values, each of which is a JSON number or string.¶

An example of a reverse search response including the "reverse_search_properties_mapping" member is shown in Figure 3.¶

6. Reverse Search Response Specification

Reverse search responses use the formats defined in Section 8 of [RFC9083], which correspond to the searchable resource types defined in Section 2.¶

7. Reverse Search Query Processing

To process a reverse search, the server returns the objects from its data store that are of type searchable-resource-type and that match each of the predicates from the search conditions. To determine whether an object matches a predicate, the server:¶

• applies the mapping it uses for the reverse search property to the object in order to generate a list of values, each of which MUST be a JSON number or string and¶
• checks whether the search pattern matches one or more of those values.¶

A search pattern matches a value where it equals the string representation of the value or where it is a match for the value in accordance with the partial string matching behavior defined in Section 4.1 of [RFC9082].¶

Objects are only included in the search results if they satisfy all included predicates. This includes predicates that are for the same property; in such a case, it is necessary for the related object to match against each of those predicates.¶

Servers MUST return an HTTP 501 (Not Implemented) [RFC9110] response to inform clients of unsupported reverse searches.¶

Based on their policy, servers MAY restrict how predicates are used to make a valid search condition by returning a 400 (Bad Request) response when a problematic request is received.¶

A given reverse search or reverse search mapping MAY define additional or alternative search behavior past that set out in this section.¶

8. Reverse Searches Based on Entity Details

Since an entity can be associated with any other object class in RDAP, the most common kind of reverse search is one based on an entity's details. Such reverse searches arise from the query model by setting the related resource type to "entity".¶

By selecting a specific searchable resource type, the resulting reverse search aims at retrieving all the objects (e.g., all the domains) that are related to any entity object matching the search conditions.¶

This section defines the reverse search properties servers SHOULD support for the domain, nameserver, entity-searchable resource types, and entity-related resource type:¶

Reverse search property:

role¶

RDAP member path:

$.entities[*].roles¶

Reference:

Section 10.2.4 of [RFC9083]¶

Reverse search property:

handle¶

RDAP member path:

$.entities[*].handle¶

Reference:

Section 5.1 of [RFC9083]¶

Reverse search property:

fn¶

RDAP member path:

$.entities[*].vcardArray[1][?(@[0]=='fn')][3]¶

Reference:

Section 6.2.1 of [RFC6350]¶

Reverse search property:

email¶

RDAP member path:

$.entities[*].vcardArray[1][?(@[0]=='email')][3]¶

Reference:

Section 6.4.2 of [RFC6350]¶

The presence of a predicate on the reverse search property "role" means that the RDAP response property "roles" MUST contain at least the specified role.¶

The last two properties are related to jCard elements [RFC7095], but the field references are to vCard [RFC6350], since jCard is the JSON format for vCard.¶

Examples of reverse search paths based on the domain-entity relationship are presented in Figure 1.¶

 /domains/reverse_search/entity?handle=CID-40*&role=technical

 /domains/reverse_search/entity?fn=Bobby*&role=registrant

 /domains/reverse_search/entity?handle=RegistrarX&role=registrar

An example of the help response including the supported reverse search properties is shown in Figure 2.¶

   {
     "rdapConformance": [
       "rdap_level_0",
       "reverse_search"
     ],
     ...
     "reverse_search_properties": [
       {
         "searchableResourceType": "domains",
         "relatedResourceType": "entity",
         "property": "fn"
       },
       {
         "searchableResourceType": "domains",
         "relatedResourceType": "entity",
         "property": "handle"
       },
       {
         "searchableResourceType": "domains",
         "relatedResourceType": "entity",
         "property": "email"
       },
       {
         "searchableResourceType": "domains",
         "relatedResourceType": "entity",
         "property": "role"
       }
     ],
     ...
   }

An example of a response including the mapping that is occurring for the first reverse search in Figure 1 is shown below.¶

   {
     "rdapConformance": [
       "rdap_level_0",
       "reverse_search"
     ],
     ...
     "reverse_search_properties_mapping": [
       {
         "property": "handle",
         "propertyPath": "$.entities[*].handle"
       },
       {
         "property": "role",
         "propertyPath": "$.entities[*].roles"
       }
     ],
     ...
   }

9. RDAP Conformance

Servers complying with this specification MUST include the value "reverse_search" in the rdapConformance property of the help response [RFC9083] and any other response including the "reverse_search_properties_mapping" member. The information needed to register this value in the "RDAP Extensions" registry is described in Section 11.1.¶

10. Implementation Considerations

To limit the impact of processing the search predicates, servers are RECOMMENDED to make use of techniques to speed up the data retrieval in their underlying data store, such as indexes or similar. In addition, risks with respect to performance degradation or result set generation can be mitigated by adopting practices used for standard searches, e.g., restricting the search functionality, limiting the rate of search requests according to the user's authorization, truncating and paging the results [RFC8977], and returning partial responses [RFC8982].¶

12. Privacy Considerations

The search functionality defined in this document may affect the privacy of entities in the registry (and elsewhere) in various ways; see [RFC6973] for a general treatment of privacy in protocol specifications. Registry operators should be aware of the trade-offs that result from implementing this functionality.¶

Many jurisdictions have laws or regulations that restrict the use of "personal data", per the definition in [RFC6973]. Given that, registry operators should ascertain whether the regulatory environment in which they operate permits implementation of the functionality defined in this document.¶

In those cases where this functionality makes use of sensitive information, the information MUST only be accessible to authorized users under a lawful basis.¶

Since reverse search requests and responses could contain Personally Identifiable Information (PII), reverse search functionality MUST be available over HTTPS only.¶

Providing reverse search in RDAP carries the following threats as described in [RFC6973]:¶

• Correlation¶
• Disclosure¶
• Misuse of data¶

Therefore, RDAP providers need to mitigate the risk of those threats by implementing appropriate measures supported by security services (see Section 13).¶

13. Security Considerations

Security services that are required to provide controlled access to the operations specified in this document are described in [RFC7481]. A non-exhaustive list of access control paradigms an RDAP provider can implement is presented in Appendix A.¶

As an additional measure to enforce security by preventing reverse searches to be accessed from unauthorized users, the RDAP providers may consider physically separating the reverse search endpoints from the other ones by configuring a proxy routing the reverse searches to a dedicated backend server and leveraging further security services offered by other protocol layers, such as digital certificates and IP allow-listing.¶

Finally, the specification of the relationship within the reverse search path allows the RDAP servers to implement different authorization policies on a per-relationship basis.¶

Appendix A. Paradigms to Enforce Access Control on Reverse Search in RDAP

Access control can be implemented according to different paradigms introducing increasingly stringent rules. The paradigms listed below leverage the capabilities that are either built in or provided as extensions by the OpenID Connect [OIDCC]:¶

Role-Based Access Control (RBAC):

Access rights are granted depending on roles. Generally, this is done by grouping users into fixed categories and assigning static grants to each category. A more dynamic approach can be implemented by using the OpenID Connect "scope" claim.¶

Purpose-Based Access Control (PBAC):

Access rules are based on the notion of purpose, being the intended use of some data by a user. It can be implemented by tagging a request with the usage purpose and making the RDAP server check the compliance between the given purpose and the control rules applied to the data to be returned.¶

Attribute-Based Access Control (ABAC):

Rules to manage access rights are evaluated and applied according to specific attributes describing the context within which data are requested. It can be implemented within an out-of-band process by setting additional OpenID Connect claims that describe the request context and make the RDAP server check for compliance between the given context and the control rules that are applied to the data to be returned.¶

Time-Based Access Control (TBAC):

Data access is allowed for a limited time only. It can be implemented by assigning users temporary credentials linked to access grants with limited scopes.¶

With regard to the privacy threats reported in Section 12, correlation and disclosure can be mitigated by minimizing both the request features and the response data based on user roles (i.e., RBAC). Misuse can be mitigated by checking for the purpose of the request (i.e., PBAC). It can be accomplished according to the following approaches:¶

Full Trust:

The registry trusts the fairness of an accredited user. The requestor is always legitimized to submit their requests under a lawful basis. Additionally, they can be required to specify the purpose as either a claim of their account or a query parameter. In the former case, the purpose is assumed to be the same for every request. In the latter case, the purpose must be one of those associated to the user.¶

Zero Trust:

The registry requires documents that assess whether the requestor is legitimized to submit a given request. It can be implemented by assigning the requestor a temporary OpenID account linked to the given request (i.e., TBAC) and describing the request through a set of claims (i.e., ABAC). The association between the temporary account and the claims about the request is made by an out-of-band application. In so doing, the RDAP server is able to check that the incoming request is consistent with the request claims linked to the temporary account.¶

The two approaches can be used together:¶

• The former is suitable for users carrying out a task in the public interest or exercising their official authority (e.g., an officer of a cybercrime agency). Similarly, registrars can submit reverse searches on their domains and contacts based on their contractual relationship with the domain holders. In this case, the query results can be restricted to those pertaining to a registrar by adding an implicit predicate to the search condition.¶
• The latter can be taken to allow domain name dispute resolution service providers to request information in defense of the legitimate interests of complainants.¶

Authors' Addresses