JSON for RESTful Provisioning Protocol (RPP)

Introduction The RESTful Provisioning Protocol (RPP) defines a set of data objects for managing registry resources, such as domain names, contacts, hosts, organisations, and users. The data model is defined in independently of any particular representation format. This document defines the JSON representation of those data objects. JSON has emerged as the de facto standard data format for modern RESTful APIs. Its widespread adoption across tools, libraries, and developer communities makes it well suited as the primary representation format for RPP. This document provides the normative rules and JSON Schema definitions required for implementations to produce and consume RPP messages in JSON. The separation between the abstract data model and its concrete JSON representation ensures that the protocol's semantic foundation remains stable while enabling the adoption of JSON across diverse deployment environments.

Motivation The RESTful Provisioning Protocol (RPP) introduces a new provisioning mechanism that aligns more closely with modern cloud infrastructure, enhancing the scalability of server deployments. While RESTful protocols do not mandate a specific media type for resource description, the widespread adoption of JSON in web services has established it as the de facto standard for modern APIs. The increasing availability of tools, software libraries, and a skilled workforce has led several registries to adopt JSON for data exchange within their API ecosystems. Registries supporting JSON can offer a unified API ecosystem that extends beyond domain name and IP address provisioning, maintaining a consistent technology stack, data formats, and developer experience. JSON's syntax, known for its straightforwardness and minimal verbosity, significantly eases the tasks of writing, reading, and maintaining code. This simplicity is especially advantageous for the rapid comprehension and integration of provisioning APIs. The lightweight nature of JSON can result in faster processing and data transfers, a critical aspect in high-volume transaction environments such as domain registration. Enhanced API response times can lead to more efficient domain lookups, registrations, and updates. JSON parsing is typically fast and well-supported by standard libraries, contributing to improved system performance amid frequent interactions between RPP clients and servers. However, the absence of a standardised JSON format for domain provisioning has led to the emergence of TLD-specific implementations that lack interoperability, increasing the development effort required for integration. Similarly, at the registrar level, the absence of standards has resulted in numerous incompatible API implementations provided to clients and resellers. Standardising a JSON format for domain provisioning within the RPP framework could mitigate these challenges, reducing fragmentation and simplifying integration efforts across the domain registration industry.

Terminology In this document, the following terminology is used. Terms of the RPP data model, such as Data Object, Component Object, Process Object, Common Data Type, data element, cardinality, mutability, Identifier, Unique Identifier, Input Data Element, Parameter, Visibility, Direct Access, Input, Output, Reference, Owner Data Object, and This Object, are used as defined in . RPP Data Objects - The abstract data model definitions of the objects provisioned and exchanged with RPP, as specified in . RESTful Provisioning Protocol - A RESTful protocol for provisioning heterogeneous database objects. JSON Schema - A vocabulary that allows annotation and validation of JSON documents, as described in . EPP Compatibility Profile - A set of additional constraints defined in that a server MUST adhere to when supporting both RPP and EPP concurrently.

Conventions Used in This Document The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in . JSON is case sensitive. Unless stated otherwise, JSON specifications and examples provided in this document MUST be interpreted in the character case presented. The examples in this document assume that request and response messages are properly formatted JSON documents. Indentation and white space in examples are provided only to illustrate element relationships and for improving readability, and are not REQUIRED features of the protocol. All JSON Schema definitions in this document use JSON Schema draft 2020-12 , and where not provided with a $schema keyword, the following default applies:

JSON Representation Rules This section defines the normative rules for representing the RPP data model in JSON. The data model defines all primitive types, common data types, component objects, resource objects, and associations independently of any concrete representation format. The rules in this section specify how those abstract definitions map to JSON and JSON Schema version 2020-12.

Primitive Type Mappings RPP primitive types MUST be represented in JSON as follows:

RPP Primitive Type	JSON Type	Notes
String	`string`	Unicode character sequence
Integer	`number`	Whole number, positive or negative
Boolean	`boolean`	`true` or `false`
Decimal	`string`	Base-10 fractional value with exact representation within a defined precision. Number is encoded into string same as `"number"` in without exponent part `"ext"`.
Date	`string`	Full-date as per , e.g. `"2025-10-27"`
Timestamp	`string`	Date-time in UTC as per , e.g. `"2025-10-27T09:42:51Z"`
URL	`string`	Uniform Resource Locator as per
Binary	`string`	Base64-encoded binary data

Undefined Values A data element can have no defined value, for example when the value is NULL in the server's database or has never been set. The JSON null value must not be used to represent an undefined value, a JSON Schema type such as "type": "string" does not allow null. An undefined value of a primitive type data element must not be rendered, the property must be omitted from the JSON object. An array data element with cardinality 0+ that has no values may be rendered as an empty array ([]) or may be omitted from the JSON object. Consumers must treat both representations as equivalent. Example: a domain name for which the primitive type data element expiryDate is undefined and not rendered, the array data element nameservers has no values, with nameservers rendered as an empty array: The same domain name with the nameservers array omitted is equivalent:

Cardinality Rules The cardinality of each data element in the RPP data model MUST be represented as follows in JSON: A data element with cardinality 1 (exactly one) MUST be represented as a JSON property and MUST be present in the containing JSON object. The element MUST be listed under required in the corresponding JSON Schema. A data element with cardinality 0-1 (zero or one) MUST be represented as an optional JSON property. The element MUST NOT be listed under required in the corresponding JSON Schema. When absent, the element MUST be omitted from the JSON object (not represented as null). A data element with cardinality 0+ (zero or more) MUST be represented as an optional JSON array. When no values are present, the property MUST be omitted or represented as an empty array. A data element with cardinality 1+ (one or more) MUST be represented as a required JSON array with "minItems": 1 and the element MUST be listed under required in the corresponding JSON Schema.

Mutability Rules Data elements in the RPP data model carry a mutability attribute: create-only, read-write, or read-only. Which of them an object representation contains follows from the Input and Output of the operation:

Response content contains data elements of every mutability. A data element is left out only when it is not present on the object, when its Visibility is onDemand and the client did not request it, or when authorisation or server policy withholds it.
Request content for an object being created contains its create-only and read-write data elements. For a Component Object or an Aggregation object creation can take place in create or update operation on the containing object, when a value is set or added.
Request content for an existing object being modified, i.e. when both the Reference and the Object of the Input are This Object, contains its read-write data elements.

Input Data Elements and Parameters are not data elements of the object, have no mutability, and are not subject to the following rules. Data elements with mutability read-only MUST be annotated with "readOnly": true in the JSON Schema and MUST NOT be part of any request view. Data elements with mutability create-only MUST be annotated with "writeOnly": true in request schemas and MUST NOT be part of the update view. Data elements with mutability read-write have no additional annotation. A server MUST reject a request whose content contains a data element that is not allowed in that request, such as a read-only data element or a create-only data element of an existing object. With the enrichment required by , such a request fails JSON Schema validation.

Association Rules The RPP data model defines several association types between objects, the following rules specify their JSON representations. An Aggregation represents a relationship between two independent objects, where one object references another. A Composition represents a parent-child relationship where the child object is embedded within the parent object and cannot exist independently.

Labelled Associations Some associations between objects carry a string label that provides additional context for the relationship. The label is not an identifier of the target object, but rather a descriptor of the association itself. Labelled associations can occur in both aggregations and compositions. When representing labelled associations in JSON, the property label MUST be included alongside the reference to the target object. A property with the name object MUST be used to contain the reference to the target object, which can be either a limited representation containing at minimum the Unique Identifier of the target object for aggregations or an embedded object for compositions.

Aggregation An Aggregation[Type] represents a relationship between two independent objects. The referenced object is identified by its Unique Identifier. When the cardinality allows more than one target, it MUST be represented as a JSON array. Each element of the array is a limited representation of the referenced object, as specified below. Aggregation[Type] with cardinality 0+ or 1+ MUST be represented as a JSON array of embedded objects. Each object in the array MUST include the data elements of the referenced object type that are relevant to the context (at minimum the Unique Identifier). Other data elements of the referenced object type MAY be included as needed to provide additional context for the client, but are not required. The JSON Schema MUST allow for the presence of these additional fields. Example: domain nameservers (Aggregation[Host Data Object]) in a read response, returning a limited object representation containing only the Unique Identifier hostName:

Composition A Type represents a parent-child relationship where the child's lifecycle is bound to the parent and the child cannot exist independently of the parent. In JSON, the child object MUST be fully embedded within the parent object.

Labelled Aggregation A LabelledAggregation[Type] is a relationship between two independent objects where each association carries a string label. Multiple associations with the same label are allowed. LabelledAggregation[Type] with cardinality 0+ MUST be represented as a JSON array of objects. Each object in the array MUST contain a label property (string) alongside the identifier of the referenced object. The object MUST include at least the Unique Identifier of the referenced object type. Other data elements of the referenced object type MAY be included as needed to provide additional context for the client, but are not required. The JSON Schema MUST allow for the presence of these additional fields. Example: domain contacts (LabelledAggregation[Contact Object]):

Dictionary Aggregation A DictionaryAggregation[Type] is a relationship between two independent objects where each association carries a unique string label that serves as a dictionary key. DictionaryAggregation[Type] MUST be represented as a JSON object where each key is the unique label and the corresponding value is the referenced object, the object MUST include at least the Unique Identifier of the referenced object type. Other data elements of the referenced object type MAY be included as needed to provide additional context for the client, but are not required. The JSON Schema MUST allow for the presence of these additional fields. Example: domain contacts keyed by unique role (DictionaryAggregation[Contact Object]):

Labelled Composition A LabelledComposition[Type] is a parent-child relationship where each embedded child carries a string label. Multiple instances with the same label are allowed. LabelledComposition[Type] with cardinality 0+ MUST be represented as a JSON array of embedded objects. Each object in the array MUST contain a label property alongside the data elements of the composed type. Example: remarks (LabelledComposition[Remark Object]):

Dictionary Composition A DictionaryComposition[Type] is a parent-child relationship where each embedded child carries a unique string label used as a dictionary key. DictionaryComposition[Type] MUST be represented as a JSON object where each key is the unique label and the corresponding value is the fully embedded child object. Example: remarks keyed by scope (DictionaryComposition[Remark Object]):

Object Identifier Rules When a resource or component object is referenced by identifier (for example in an aggregation), the identifier MUST be represented as a JSON string using the value of the object's primary identifier data element. When a resource or component object is embedded (as in a composition), all data elements of the object MUST be represented as properties of a JSON object according to the rules of this section.

External Type Embedding Rules RPP data objects may include data elements whose types are defined by external specifications (e.g., JSContact in ). When embedding a value of an externally defined type into a JSON representation, the following rules apply, in order of preference: If the external type has a native JSON representation (i.e., the external specification defines how the type is represented using the JSON format), the value MUST be embedded as a valid JSON element using the native JSON representation defined by that specification. If the external type does not have a native JSON representation but has a native UTF-8 encoded text representation, the value MUST be embedded as a JSON string using the UTF-8 encoding. If the external type has neither a native JSON nor a native UTF-8 text representation, the binary representation of the value MUST be Base64-encoded (, Section 4) and embedded as a JSON string.

Process Data Embedding Rules As defined in , a uniform interface operation MAY require process data in addition to the object representation data. This section defines how such process data MUST be represented in JSON when transmitted together with the object representation in a single request body. Process data accompanying a uniform interface operation MUST be represented using the processes data element of the owning resource object (see Processes Object), following the Aggregation rules for the corresponding process type (for example createProcess for creation-specific inputs). Each embedded process object MUST be a valid JSON representation of the process object. This applies uniformly regardless of whether the resource object already exists. Example: a domain create request body carrying a two-year registration period as process data embedded in a createProcess process object under processes:

JSON Schema Definition Rules Each RPP component object and resource object MUST have a corresponding JSON Schema definition. Object definitions MUST be placed in the $defs keyword of the JSON Schema document. Identifier fields MUST use "type": "string" in JSON Schema. Enumeration constraints on string fields MUST be expressed using the "enum" keyword in JSON Schema. Example (Transfer Status enum): Each JSON Schema definition for an RPP object MUST include a "required" array listing all data elements with cardinality 1 or 1+.

Undeclared Properties JSON Schema definitions for extendible RPP objects MUST NOT use "additionalProperties": false or "unevaluatedProperties": false. However, before validation, schemas on every property level MUST be enriched with "unevaluatedProperties": false property to prevent the presence of undeclared properties in JSON instances. JSON Schemas for Object type MAY use "additionalProperties": true to allow for free key definition.

Type Property Every RPP object representation MUST include a "@type" property whose value is the object's identifier as registered in the IANA RPP Data Object Registry. This property enables identification and allows clients and servers to unambiguously determine the type of an object. The "@type" property MUST be included in the JSON Schema "properties" object for each RPP object definition with a "const" constraint fixing the value to the object's registered identifier. The "@type" property MUST be listed in the "required" array of the corresponding JSON Schema definition. Example (Domain Name Data Object):

Authorisation Information in Requests When a transfer request or other operation requires authorisation information (e.g., EPP-style authinfo), the client MUST NOT include the authInfo object in the JSON request body. Instead, the client MUST convey the authorisation information using the RPP-Authorization HTTP request header as defined in . Servers MUST reject any request that includes an authInfo object in the JSON body with an appropriate error response.

JSON Schema Composition Rules The JSON Schema definitions of all RPP objects MUST be located in a single shared $defs object at the top level of the RPP schema document. A schema that refers to another definition MUST use the $ref keyword with a JSON Pointer fragment to the definition in the shared $defs object. The key of a definition in the shared $defs object MUST be formed as follows, using the Identifiers of the data model unchanged:

Data Object, Process Object, or Component Object: its Identifier, followed by a . and the name of the view. An object that extends another object uses its own Identifier, not the Identifier of the object it extends.

The view names are:

create: the request content of the Create operation.
read: the object representation returned in responses.
update: the request content of the Update operation.
reference: the limited representation of an aggregated object (see Aggregation), containing @type and the object's Unique Identifier.
query: the response content of the Query operation, typically an array of read views.
The Identifier of the operation: the request content of any other operation.
The Identifier of the operation followed by .output: the response content of an operation, other than Read and Query, whose output is not an existing view of the object.

A Component Object has a read view. A Component Object that has at least one data element with mutability create-only or read-write also has a create view. A Component Object has no other views. A view that holds request content MUST reference the create view of an embedded Component Object. A view that holds response content MUST reference the read view of an embedded Component Object. Each key MUST be unique within the shared $defs object. Common Data Types and types defined by an external specification are not subject to this rule. They MAY use a key that is not derived from an RPP Identifier. A schema that refers to another component object or data object MUST use the $ref keyword with a JSON Pointer fragment to the definition in the shared $defs object, for example { "$ref": "#/$defs/period.create" }. The final RPP schema document, containing the shared $defs object with all sub-schemas, MUST have an $id keyword whose value is an absolute URI without a fragment, serving as the stable identifier of the complete RPP schema document. Example (structure of the final RPP schema document):

Status Objects Each status object has a status values table whose "Set By" column identifies the entity that adds and removes each status value: client, server, or client, server. In this section, a client-settable status value is a value whose "Set By" column contains client. The read view of a status object lists all of its status values in the "enum" of "label". Its create view lists only the client-settable status values in the "enum" of "label". When a status object has no client-settable status value, the create and update views of the Data Object MUST NOT contain the status data element. When the status values of a status object state that a client is not required to provide a status value when creating or updating the object, the status data element MUST NOT be listed under "required" and MUST NOT carry "minItems" in the create and update views, notwithstanding and . When the status values of a status object state that a server MUST assign at least one status value to the object, the status data element MUST be listed under "required" in the read view, with "minItems": 1 when it is an array.

Inheritance RPP Data Objects and Component Objects MAY inherit the data elements and operations of another object. The JSON Schema for an object must include all the inherited data elements and constraints from its base object, using a flat structure where all properties are defined at the top level of the schema, rather than nested under an "allOf" or "anyOf" construct. Example: a Shape Object with the data elements name (cardinality 1) and colour (cardinality 0-1), and a Circle Object that extends it with the additional data element radius (cardinality 1). JSON Schema of the base Shape Object: JSON Schema of the Circle Object, which defines the inherited name and colour properties at the top level alongside its own radius property, without referencing the Shape Object schema using "allOf": Example Circle Object instance:

RPP Profiles and Validation RPP profiles, such as the EPP Compatibility Profile, may impose additional constraints on top of the base RPP data model. These additional constraints MUST be enforced by implementations through validation rules that go beyond what can be expressed in JSON Schema. Such validation rules MUST be clearly documented in the profile specification and implemented by both clients and servers when operating under that profile. For example, the EPP Compatibility Profile requires that certain fields be present in specific object types, and that certain identifier fields conform to EPP syntax rules. These constraints cannot be fully captured in JSON Schema and therefore require additional validation logic in implementations.

Update Rules JSON representations for update operations follow specific rules to ensure consistency and correctness. There are two types of update operations: full updates and partial updates.

Full A full update operation replaces the entire state of a resource object with the new representation provided in the request, all required properties MUST be included in the new representation.

Partial A partial update operation allows clients to update specific properties of a resource object, while leaving other properties unchanged. A partial update MUST use JSON Patch , extended with a match property for matching array elements. The match property is not part of ; it is introduced and defined in this document (see ). The request body of a partial update is a JSON array of patch operations, see . Each patch operation consists of an op (the operation), a path (a JSON Pointer to the property the operation targets), a value (the new value, for add and replace), and optionally a match (to select array elements). When the path points to an existing primitive type property, the add operation has the same semantics as replace and will overwrite the existing value.

Matching The match property is used to identify elements in an array of objects. It is a JSON object that contains key-value pairs corresponding to the properties of the array elements. The server uses the match property to locate the elements in the array that should be removed or replaced. The match property MUST NOT be used when the path does not point to an array, and MUST NOT be used for an array of primitive types. An array of primitive types is changed by replacing the array as a whole, or by adding a value without match. The following rules apply to matching and depend on the type of the property in the match object:

A primitive type property (e.g. string, number or boolean) matches if the target object property has the same type and the same value.
An array of primitive type property (e.g. array of string, number or boolean) matches if the target array contains elements of the same type and both sets of values are identical. The ordering of elements does not affect matching.
An object type is considered a match when the value of every property in the match object matches the value of the property with the same name in the target object. Properties of the target object that are not named in the match object are not compared.
The server MUST NOT match nested data structures. The match property is only used for matching elements in the immediate array specified by the path property of the JSON Patch operation.
The server MUST reject a match that is not a JSON object, or that contains a value that cannot be compared using the rules above.

Example domain object for matching an array element with a specific key-value pair: Example match object for matching the array element with the "tech" label and the contact ID "REG-200" in the contacts array of the domain object above: Example object for matching an array element that has a property containing an array of primitive types, here the tld property of the elements of the idnLgr array: Example match object for matching the elements of the idnLgr array above for which the tld array contains the value "org" and "example". The order of the values in tld is ignored and only the first element matches, because the second element does not contain "org":

Processing Steps The server MUST process the patch operations in the order in which they appear in the request body. A partial update is atomic, meaning that if any of the operations fails or is rejected, the server MUST reject the complete request with an appropriate error response that identifies the patch operation that failed. For each patch operation the server MUST perform the following steps. shows the steps for a single patch operation, shows the handling of a primitive type value or an object, and shows the handling of an array.

Check the operation: the op must be one of "add", "remove" or "replace". JSON Patch operations other than these are not allowed.
Resolve the path: the path must point to an existing property in the JSON object or a non-existing but allowed property defined in the JSON Schema of the object.
Determine the type of the target that the path points to:
- A primitive type value (for example a string, number or boolean), or an object: continue with step 4.
- An array: continue with step 5.
Apply the operation to a primitive type value or an object:
- For "remove", the field is removed.
- For "add" and "replace", the value property must be present.
  - If the target is a primitive type value, the type of value must match the type of the target field.
  - If the target is an object, value must be a valid object, the @type property of value must match the @type property of the target object or be a supertype of the target object's @type.
  - The field is then set to value.
Apply the operation to an array:
- If match is present, the operation targets the elements identified by match:
  - The "add" operation is not allowed to be used together with match.
  - If no element matches, the operation fails.
  - For "remove", all matching elements are removed from the array.
  - For "replace", if more than one element matches, the operation is ambiguous and fails. If exactly one element matches, it is replaced by value.
- If match is absent, the operation targets the array property as a whole:
  - For "add", value is appended to the array as a new element.
  - For "replace", the entire array is replaced with the array provided in value.
  - For "remove", the array property is removed entirely.
After the last operation has been applied, the resulting resource object must satisfy the JSON Schema and the constraints of the resource object that cannot be enforced by the JSON Schema.

If a step fails because no element matches, the server MUST return an error response indicating that the specified element could not be found.

Partial update, steps for each patch operation REJECT | | "replace"? | | +---------------+---------------+ | | yes | v | +-------------------------------+ | | Does path point to an |--- no --> REJECT | | existing field or element? | | +---------------+---------------+ | | yes | v | +-------------------------------+ | | Does path point to an array? | | +-------+---------------+-------+ | | no | yes | v v | +---------------+ +---------------+ | | Simple value | | Array | | | or object | | (figure 3) | | | (figure 2) | | | | +-------+-------+ +-------+-------+ | | | | +--------+--------+ | | | v | +-------------------------------+ | | Another patch operation? |--- yes ------------+ +---------------+---------------+ | no v +-------------------------------+ | Is the resulting object valid |--- no --> REJECT | for the JSON Schema and the | | constraints? | +---------------+---------------+ | yes v Apply all changes ]]>

Partial update, primitive type value or object REJECT +---------------+---------------+ | no v +-------------------------------+ | Is op "remove"? |--- yes --> Remove the field +---------------+---------------+ | no ("add" or "replace") v +-------------------------------+ | Is value present? |--- no ---> REJECT +---------------+---------------+ | yes | v +---------------+ | Simple value | | and same type | --- yes ---> Set the field to value | | +-------+-------+ | no | v +---------------+ | Object and | | @type matches | --- yes ---> Set the field to value | | +-------+-------+ | | no v REJECT ]]>

Partial update, array continue at (B) +---------------+---------------+ | no v +-------------------------------+ | Is op "add"? |--- yes --> Append value as +---------------+---------------+ a new element | no v +-------------------------------+ | Is op "remove"? |--- yes --> Remove the property +---------------+---------------+ (REJECT if the property has cardinality 1+) | no ("replace") v +-------------------------------+ | Does value violate the schema | | of the array elements? |--- yes --> REJECT +---------------+---------------+ | no v Replace the array with value (B) Match present: the matching elements are targeted +-------------------------------+ | Is op "add"? |--- yes --> REJECT +---------------+---------------+ | no v +-------------------------------+ | Is match a JSON object whose | | values can be compared? |--- no ---> REJECT +---------------+---------------+ | yes v +-------------------------------+ | Does any element match? |--- no ---> REJECT +---------------+---------------+ | yes v +-------------------------------+ | Is op "remove"? |--- yes --> Remove all +---------------+---------------+ matching elements | no ("replace") v +-------------------------------+ | Does more than one element | | match? |--- yes --> REJECT (ambiguous) +---------------+---------------+ | no v +-------------------------------+ | Is value valid for the | | element type? |--- no ---> REJECT +---------------+---------------+ | yes v Replace the element with value ]]>

Schema A partial update request body MUST be a JSON array of patch operation objects. Each operation MUST include an op and a path property. The value property MUST be present for add and replace operations. The match property MAY be present when the path points to an array property, to identify the array elements to operate on. When match is absent, the operation targets the array property as a whole (see step 5 in ). The value property is intentionally left unconstrained ({}) in the JSON Schema below, since the JSON type it MUST hold depends on the target path: it can be a JSON object (e.g. when replacing a single object property or a single matched array element), a JSON array (e.g. when replacing an array property as a whole), or a primitive type such as a string, number, or boolean.

Examples Example JSON document used by partial update examples below: Example of a partial update request to change the registrant of a domain name from "JD-1234" to "JD-1235": Example of a partial update request to change the status of a domain name from "ok" to "clientUpdateProhibited": Example of a partial update request for adding an additional tech contact to a domain name: Example of a partial update request replacing the entire nameservers array of a domain name in a single operation, with no "match" property present:

JSON Schema Definitions This section provides normative JSON Schema definitions for RPP component objects and resource objects. All schemas use JSON Schema draft 2020-12 .

Common Data Types Schemas This section defines shared data types that are based on the primitive data types above and are reused across multiple data object definitions.

Identifier Identifiers are character strings with a specified minimum length, a specified maximum length, and a specified format outlined in . Identifiers for certain object types MAY have additional constraints imposed either by server policy, object-specific specifications, or both. In JSON, an Identifier MUST be represented as a string value conforming to the roidType syntax defined in : one to eight word characters optionally preceded by an underscore, followed by a hyphen and one or more word characters, with a maximum total length of 89 characters.

Client Identifier Client identifiers are character strings with a specified minimum length, a specified maximum length, and a specified format. Client identifiers use the clIDType syntax described in . In JSON, a Client Identifier MUST be represented as a string value.

Phone Number Telephone number syntax is derived from structures defined in . Telephone numbers described in this specification are character strings that MUST begin with a plus sign ("+", ASCII value 0x002B), followed by a country code defined in , followed by a dot (".", ASCII value 0x002E), followed by a sequence of digits representing the telephone number. An optional "x" (ASCII value 0x0078) separator with additional digits representing extension information can be appended to the end of the value. In JSON, a Phone Number MUST be represented as a string value conforming to the pattern described above.

Component Objects Schemas

Period Object Read schema: Create schema:

Provisioning Metadata Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

upClientId and upDate MUST NOT be present if the object has never been modified.
trDate MUST NOT be present if the object has never been transferred.
In the EPP Compatibility Profile, repositoryId MUST be provided.

Read schema:

Status Objects Each RPP Data Object type that carries a status uses a type-specific status object. The create view of a status object holds only the status values that a client can set (see ). The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

A server MUST keep the status values it has set when it processes a create or update request; their absence from the "status" of the request does not remove them. A server MAY alter or override client-settable status values, subject to local server policy.
A server MUST reject a request whose "status" array contains a combination of label values prohibited by the status exclusion rules for that object type.
A "status" array MUST NOT contain the same "label" value more than once.

Extensions MAY define additional "label" values for a status object; such values are added to the "enum" of the corresponding status object in the effective schema, as described in .

Domain Name Status Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

"due": Servers MAY restrict the ability of clients to set or update this value.
Label values with the prefix "rgp" MUST only appear when the server supports the Redemption Grace Period (RGP) for the domain name.
The labels "rgpRedemptionPeriod", "rgpPendingRestore", and "rgpPendingDelete" MUST only appear together with the "pendingDelete" label.

Read schema: Create schema (status values that a client can set):

Host Status Object Read schema: Create schema (status values that a client can set):

Contact Status Object Read schema: Create schema (status values that a client can set):

Organisation Role Status Object Read schema: Create schema (status values that a client can set):

Organisation Status Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

The "hold" and "terminated" labels are server-managed when the organisation has no parent organisation; otherwise they MAY be client-managed, subject to server policy.
The "status" array MUST contain exactly one of the labels "pendingCreate", "ok", "hold", or "terminated".

Read schema: Create schema (status values that a client can set):

User Status Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

The "pending" label is server-managed; "suspended" and "deactivated" are server-managed and MAY additionally be client-managed by a client with permission to manage users.

Read schema: Create schema (status values that a client can set):

DNS Resource Record Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

name MUST be a syntactically valid DNS host name in zone file string representation. Both absolute FQDNs (with trailing dot) and relative host names are allowed, as well as the @ symbol representing the domain name itself.
type MUST be a valid string representation of a DNS resource record type as defined in . Values MUST be converted to lowercase. Allowed values MAY be further constrained by server policy.
In the EPP Compatibility Profile (), the following record types MUST be supported: ns, a, and aaaa. With DNSSEC Extension , ds and dnskey MUST additionally be supported.
class, if present, MUST be chosen from Section 3.2.4 (CLASS values) of . A client SHOULD omit this element; the server MUST assume IN as the default.
The fields within rdata MUST match the expected structure for the given record type (see RDATA structures below).

RDATA structures required in the EPP Compatibility Profile:

NS records (, Section 3.3.11): nsdname
A records (, Section 3.4.1): address
AAAA records (, Section 2.2): address
DS records (, Section 5, with DNSSEC Extension): keyTag, algorithm, digestType, digest
DNSKEY records (, Section 2, with DNSSEC Extension): flags, protocol, algorithm, publicKey

All rdata property names MUST be written in camelCase and all values MUST use the string data type. Read schema: Create schema:

DNS Operational Controls Object The DNS Operational Controls Object contains operational control parameters that a client MAY use to influence server-side DNS behaviour for a set of DNS records. A server MAY ignore these values, e.g. for policy reasons. This structure is aligned with . Read schema: Create schema:

DNS Data Object The DNS Data Object is a container for DNS resource records and associated operational controls for a provisioned object. This structure groups DNS records together with control parameters that influence server-side DNS behaviour. It is aligned with . The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

In the EPP Compatibility Profile with DNSSEC Extension , records of type ds and dnskey MUST be supported in addition to ns, a, and aaaa. A server MUST support either ds or dnskey or both. If provided with only dnskey, a server MUST calculate the DS record.

Read schema: Create schema:

Authorisation Information Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

method MUST be one of the values registered in the IANA RPP Authorization Method Registry as defined in . In the EPP Compatibility Profile, this value MUST be "authinfo" for standard password-based authorisation.
The Authorisation Information Object is immutable. When authorisation information changes, a new instance MUST be created rather than modifying the existing one. The value of authdata MAY be omitted from read responses, depending on the method and server policy.

Read schema: Create schema:

Disclose Object The Disclose Object is used by the disclose data element of the Contact Data Object. Read schema:

JSContact Card Object The Contact Data Object uses version 2.0 of JSContact to represent contact information. The Contact object is defined below according to the RPP JSContact Profile [TODO Add Ref]. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

addresses[*].countryCode MUST be a valid two-character ISO 3166-1 alpha-2 code when present.
localizations MUST be fully expanded; nested PatchObject-style keys (e.g., "addresses/addr/full") are NOT allowed.
When phones[*].features is absent, the number MUST be treated as a voice number.

Restore Report Object The Restore Report Object contains the redemption grace period restore report submitted by the sponsoring client as required by the RGP process (). The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

At least one and at most two statements MUST be provided.
restoreTime MAY be omitted when the restore report is submitted inline within the restore request in a single-step process.
In the EPP Compatibility Profile, restoreTime MUST be present as defined in .
In the EPP Compatibility Profile, exactly two statements MUST be present as defined in .

Read schema: Create schema:

Organisation Role Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

type MUST only use values registered in the IANA "EPP Organization Role Values" registry ().

Read schema: Create schema:

Extension Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

version MUST use the Semantic Versioning format.
url MUST be a valid URL and accessible to the client.

Read schema:

Discovery

Version Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

version MUST use the Semantic Versioning format.

Read schema:

Environment Object Read schema:

IDN LGR Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

name MUST be a valid LGR name as listed in the registry.

Read schema:

Notice Object Read schema:

Processes Object The Processes Object is a read-only container grouping the currently active Process Objects on the owning resource object, keyed by process type. Each key holds an array of the corresponding Process Object's read representation, per the Aggregation rules. Read schema:

Availability The Availability Objects are the output of the Availability operation. The Base Availability Object is extended by Data Object specific Availability Objects in order to allow independent extensibility.

Base Availability Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

reason MUST only be present when available is false.

Read schema:

Domain Name Availability Object Read schema:

Contact Availability Object Read schema:

Host Availability Object Read schema:

Messages All message objects are generated by the server and are read-only from the perspective of the receiving client. The Base Message Object is extended by all other message objects defined in this section. The JSON Schema of each message object repeats the data elements it inherits from the Base Message Object as top-level properties. The value of the @type property identifies the concrete message object and enables a client to determine which data elements to expect.

Base Message Object The Base Message Object contains the data elements common to all message objects. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

objectType MUST only be used when objectId is present.
objectId MUST be a valid unique identifier for the object type specified in objectType or empty when objectType is not present or empty.

Read schema:

Review Response Object Read schema:

Transfer Request Message Object Read schema:

Transfer Outcome Message Object Read schema:

Expiration Deletion Message Object Read schema:

Auto-Renewal Message Object Read schema:

Service Notice Message Object The Service Notice Message Object extends the Base Message Object. It is not linked to a specific object or process, so objectType, objectId and process are normally absent. Read schema:

Process Object Schemas

Base Process Object The Base Process Object is extended by all Process Objects, which repeat its data elements as top-level properties. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

processId MUST be unique within the scope of the owning resource object.

Read schema:

Create Process Object The Create Process Object is extended by object-specific create processes and is submitted only within the processes data element of the owning resource object's create request. Create request schema (create-only and read-write properties): Read response schema (read-only properties): Reference schema (identifier only):

Renew Process Object Renew request schema (create-only properties): Renew response schema (read-only properties): Reference schema (identifier only):

Transfer Process Object

Create Create request schema (create-only and read-write properties):

Read Read response schema (read-write and read-only properties):

Approve The Approve operation MUST always have a request body, which may be an empty JSON object. The server can identify the transfer process to be approved based on the request URL and the authenticated client's permissions. There is no Approve request schema defined in this specification. The Approve response schema MUST equal the read response schema.

Cancel The Cancel operation, which is similar to a delete operation, has no request body, the server can identify the transfer process to be cancelled based on the request URL and the authenticated client's permissions. Because the Cancel operation has no request body, it is not possible to extend the request schema with additional properties. The Cancel response schema MUST equal the read response schema or an empty response.

Reject Reject request schema: Reject response schema MUST equal the read response schema.

Reference Reference schema (identifier only):

Restore Process Object The Restore Process Object represents the current state of a restore request for an object that has entered the Redemption Grace Period (RGP). It is returned as a response for all restore operations. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

requestDate MUST NOT be present if no restore request has been submitted yet.
reportDate MUST NOT be present if no restore report has been accepted yet.
reportDueDate MUST NOT be present when restoreStatus is not "pendingRestore".

Create request schema (create-only and read-write properties): Read response schema (read-write and read-only properties): The Update request applies only to servers that support or require submission of a restore report. Update request schema (read-write properties): Reference schema (identifier only):

Domain Name Data Object The Domain Name Data Object represents a domain name and its associated provisioning data. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

name MUST be a fully qualified domain name conforming to the syntax described in . Servers MAY restrict allowable domain names to a specific namespace for which they are authoritative. The implicit trailing dot MUST NOT be included.
name MUST use the ASCII Compatible Encoding (ACE) A-label form when the domain name is internationalised, as defined in .
uName, when present, MUST be normalised to Unicode Normalization Form C (NFC) as defined in , and MUST convert to the exact value of name using the procedure described in .
lgr, when present, MUST identify a valid Label Generation Ruleset (LGR) registered in the registry, and MUST be present whenever uName is present.

Availability The response of the Availability operation is the Domain Name Availability Object, see .

Create Create request schema (create-only and read-write properties):

Create Process The Domain Create Process Object extends the Create Process Object and is submitted only within the processes data element of the Domain Name create request. Create request for Domain Create Process Object schema (create-only and read-write properties): Read response for Domain Create Process Object schema (read-write and read-only properties): Reference for Domain Create Process Object schema (identifier only):

Read Read response schema (read-write and read-only properties):

Update Update request schema (read-write properties):

Transfer Create request for Domain Transfer Process Object schema (create-only and read-write properties): Read response for Domain Transfer Process Object schema (read-write and read-only properties): Both Approve and Delete (Cancel) operations have no request body, the server can identify the transfer process to be approved or cancelled based on the request URL and the authenticated client's permissions. Approve and Delete response schemas are the read schema of domainTransferProcess. Reject request Domain Object schema: Reject response schema is the read schema of domainTransferProcess. Reference for Domain Transfer Process Object schema (identifier only):

Reference Reference schema (identifier only):

Contact Data Object This document uses version 2.0 of JSContact for the JSON representation of Contact Data Object contact information. The contact's name, postal address, phone numbers, email addresses, and other contact details are encapsulated in a JSContact Card object embedded in the contactInfo property.

JSContact Profile for RPP Since JSContact is a general-purpose representation of contact data, this document defines a restricted usage profile for use within RPP, see [TODO: ref to RPP JSContact profile here: https://github.com/SIDN/ietf-rpp-jscontact-profile]. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

card.name.full MUST be provided in the EPP Compatibility Profile.
card.addresses.addr MUST be provided in the EPP Compatibility Profile, containing at least countryCode and a "locality" kind component.
card.emails.email.address MUST be provided in the EPP Compatibility Profile.
card.addresses[*].countryCode MUST be a valid two-character ISO 3166-1 alpha-2 code when present.

Availability The response of the Availability operation is the Contact Availability Object, see .

Create Create request schema (create-only and read-write properties):

Read Read response schema (read-write and read-only properties):

Update Update request schema (read-write properties):

Reference Reference schema (identifier only):

Host Data Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

hostName MUST be a syntactically valid fully qualified host name.
If the host name is subordinate to a domain for which the server is authoritative, the superordinate domain MUST already exist in the server.

Availability The response of the Availability operation is the Host Availability Object, see .

Create Create request schema (create-only and read-write properties):

Read Read response schema (read-write and read-only properties):

Update Update request schema (read-write properties):

Reference Reference schema (identifier only):

Organisation Data Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

id MUST be a server-unique identifier conforming to the Identifier syntax.
status MUST always contain at least one value. pendingCreate, ok, hold, and terminated are mutually exclusive. ok MAY only be combined with linked.
parent, if present, MUST reference a known Organisation Object. Circular parent references MUST be rejected.
contacts keys MUST be contact type values registered in the IANA "EPP Organisation Contact Types" registry ().

Create Create request schema (create-only and read-write properties):

Read Read response schema (read-write and read-only properties):

Update Update request schema (read-write properties):

Reference Reference schema (identifier only):

Message Data Object The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

data is polymorphic: it MUST contain an instance of a message object that directly or indirectly extends the Base Message Object. The concrete message object is identified by the @type of the value used for the data property.

Create Create request schema. Only the server is authorised to create Message Data Objects, so the representation also carries the server-managed data elements id, creationDate and status.

Read Read response schema (read-only properties):

Query Query response schema (array of message read representations):

Reference Reference schema (identifier only):

User Data Object The User Data Object represents a user linked to an Organisation Object. The lifecycle of a User Object is bound to its owning Organisation Object. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations: - The "status" array MUST contain exactly one of the labels "active", "pending", "suspended", or "deactivated".

Create Create request schema (create-only and read-write properties):

Read Read response schema (read-write and read-only properties):

Update Update request schema (read-write properties):

Reference Reference schema (identifier only):

Discovery Data Object The Discovery Data Object describes the capabilities of an RPP server. The server provides a single Discovery Data Object, which has no Unique Identifier. All data elements are read-only, and the only supported operation is Read. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

datetime MUST be expressed in UTC using the Z suffix and MUST NOT use a numeric UTC offset.
Each object value in links MUST be an absolute URL using the HTTPS scheme, and the document MUST be accessible to the client without authentication.
Each label in links MUST be a link relation type registered in the IANA "Link Relation Types" registry.
Each value in objects MUST be a valid object identifier as defined in the RPP Data Object Registry.

Read Read response schema (read-only properties).

Profile Data Object The Profile Data Object represents a named set of protocol features and versions published by the server. All data elements are read-only, and the only supported operation is Read. The following constraints cannot be expressed in JSON Schema and MUST be enforced by implementations:

name of a standard profile MUST use the RPP URN namespace, for example urn:ietf:params:rpp:profile:{profile_name}. The name of a private profile MUST be globally unique, the use of reverse domain name notation is RECOMMENDED.
A profile MUST NOT list the same extension name more than once, and the server MUST support all extensions listed in the profile.
The profile referenced by baseProfile MUST NOT itself have a baseProfile, and a profile MUST NOT be its own base profile.
Only one profile MUST have defaultProfile set to true.

Read Read response schema (read-only properties):

Reference Reference schema (identifier only):

Extension Framework Extensions MUST employ an additive schema composition method, allowing new optional or required properties to be added to an existing base object schema. New properties MUST not conflict with existing definitions and MUST NOT redefine existing properties from the base schema. To support multiple extensions, a mechanism for mapping extensions to their respective base objects is defined using the rpp:extends property within the extension schema. Extensions SHOULD be independent of one another, meaning that an extension should not rely on the presence of another extension to function correctly. This ensures modularity and reduces the risk of conflicts between extensions. An extension schema MUST define a unique identifier using a top-level $id to ensure that references to the schema are unambiguous and correctly resolved. For IETF standardised extensions, using a stable and unique URI as the $id is RECOMMENDED. An IANA registry for RPP JSON extensions is requested in this document, standardised JSON extensions SHALL be registered in this registry. Ensuring that each extension has a unique and stable identifier that can be referenced reliably across different implementations. This also promotes the reuse of common extensions and reduces duplication of effort. It is RECOMMENDED to use the prefix "ext." for every extension sub-schema defined, for example #/$defs/ext.domainFoo. TODO what objects are eligible for extensions?

Extension Mapping Each extension MUST define a mapping to every base object it applies to, this is done by using a special rpp:extends property within the extension schema. For example, this mapping where the base object https://rpp.example/rpp/schema.json#/$defs/domainName.create is extended by the local definition #/$defs/ext.domainFoo:

Effective Schema Extensions are additive and independent, the effective schema for a base object MUST account for all applicable extensions to ensure that a validator can determine whether an object fully conforms to the combined set of constraints. When one or more extensions are applied to a base object, the effective schema for that object is obtained by composing the base object's schema with the schemas contributed by each extension using the allOf keyword, the server uses the rpp:extends metadata property to resolve the mapping between the base object and its extensions correctly. The resulting schema is a self-contained schema that includes everything from both the base specification and the extensions, allowing clients to validate objects against a single comprehensive schema. The following is an example of an extension JSON Schema, which adds new properties to both the domain and contact base objects: The effective schema is a distinct schema resource from the base schema it composes and therefore MUST be assigned its own $id, different from the $id of the base schema and of any other effective schema, to avoid ambiguous or circular references. It is RECOMMENDED to use the prefix "effective." for every effective sub-schema defined, for example #/$defs/effective.domainFoo. Domain Name effective schema, which includes the full schema of the base domain object along with any applicable extensions:

Examples This section provides examples that follow the JSON representation rules and JSON Schema definitions specified in the previous sections. The examples illustrate typical request and response messages for the RPP Data Objects defined in .

Domain Name

Availability Example domain availability response for a domain name that can be created: Example domain availability response for a domain name that cannot be created:

Create Example domain create request: Example domain create response from a server with RGP support:

Create Process Data Example Create Process Object without object-specific create inputs, as returned in a read response: Example Create Process Object for a two-year domain registration period, as embedded above in processes.createProcess:

Read Example domain read response: Example Read response for an internationalised domain name, showing the ACE name, the corresponding uName, and the lgr identifying the Label Generation Ruleset (LGR) used for validation:

Update Example domain update request (read-write properties): Example domain update response:

Delete The domain delete operation takes the domain name as the resource identifier in the request. No request body is required. Example domain delete response (minimal, server may return full representation):

Renew The renew operation creates a Renew Process Object. The optional expiryDate in the request is the current expiry date, sent by the client for server-side validation to prevent duplicate renewals. The response returns a Renew Process Object containing the new expiry date. Example domain renew request: Example domain renew response:

Transfer Request Authorisation information for the transfer MUST be conveyed using the RPP-Authorization HTTP header (see ), not in the JSON request body. Example domain transfer request (pull transfer): Example domain transfer response (Transfer Process Object):

Transfer Query Example domain transfer query response (Transfer Process Object):

Transfer Approve The Transfer approve operation has a request body with an empty JSON object, the server can identify the transfer process to be approved based on the request URL and the authenticated client's permissions. Example domain transfer approve request with empty JSON object: The response is the Transfer Process Object, similar to the Create (Transfer Request) and Read (Transfer Query) responses.

Transfer Cancel The Transfer cancel operation has an empty request body, the server can identify the transfer process to be cancelled based on the request URL and the authenticated client's permissions. The response is the Transfer Process Object, similar to the Create (Transfer Request) and Read (Transfer Query) responses.

Transfer Reject Example domain transfer reject request: The transfer reject response is the Transfer Process Object, similar to the Create (Transfer Request) and Read (Transfer Query) responses.

Restore Request Example domain restore request (without inline report; object transitions to pendingRestore state): Example domain restore response (Restore Process Object, server requires a report): Example domain restore request with inline restore report (single-step; object restored immediately): Example domain restore response with inline report (Restore Process Object, immediately restored):

Restore Report Example domain restore report request: Example domain restore report response (Restore Process Object):

Read (Restore Query) The Read (Restore Query) operation takes no request body (Parameters: None). Example domain restore query response (Restore Process Object, object in pendingRestore state): Example domain restore query response (Restore Process Object, object restored):

Contact

Availability Example contact availability response for a contact that cannot be created:

Create Example contact create request: Example contact create response:

Read Example contact read response:

Update Example contact update request: Example contact update response:

Delete The contact delete operation takes the contact identifier as the resource identifier. No request body is required.

Transfer Request Authorisation information for the transfer MUST be conveyed using the RPP-Authorization HTTP header (see ), not in the JSON request body. Example contact transfer request (pull transfer): Example contact transfer response (Transfer Process Object):

Transfer Query Example contact transfer query response (Transfer Process Object):

Transfer Cancel / Approve Transfer cancel and approve responses return the Transfer Process Object. The response structure is the same as the Transfer Query response above. The trStatus value reflects the outcome of the operation (e.g. "clientCancelled", "clientRejected", or "clientApproved").

Transfer Reject Example contact transfer reject request: The transfer reject response is the Transfer Process Object, similar to the Create (Transfer Request) and Read (Transfer Query) responses.

Host

Availability Example host availability response for a host that can be created:

Create Example host create request: Example host create response:

Read Example host read response:

Update Example host update request:

Delete The host delete operation takes the host name as the resource identifier. No request body is required. The server SHOULD reject the request if the host object is associated with any domain name objects.

Restore Request Example host restore request (without inline report; object transitions to pendingRestore state): Example host restore request response (Restore Process Object, server requires a report): Example host restore request with inline restore report (single-step; object restored immediately): Example host restore response with inline report (Restore Process Object, immediately restored):

Restore Report Example host restore report request: Example host restore report response (Restore Process Object):

Read (Restore Query) The Read (Restore Query) operation takes no request body (Parameters: None). Example host restore query response (Restore Process Object, object in pendingRestore state): Example host restore query response (Restore Process Object, object restored):

Organisation

Create Example organisation create request: Example organisation create response:

Read Example organisation read response:

Update Example organisation update request:

Delete The organisation delete operation takes the organisation identifier as the resource identifier. No request body is required. The server MUST reject the request if the organisation object is associated with any other objects.

Reference Example organisation reference (used when referencing an organisation from another object):

User

Create Example user create request: Example user create response:

Read Example user read response:

Update Example user update request (suspend the user):

Delete The user delete operation takes the user identifier as the resource identifier in the context of the owning organisation. No request body is required.

Reference Example user reference (used when referencing a user from an organisation object):

Message

Create Example message create request using a Transfer Request Message Object, notifying the recipient organisation that a transfer of the domain name transfer-test.example is pending and requires action: Example message create response:

Read Example message read response: Example message read response with a Review Response Object: Example message read response with a Transfer Outcome Message Object: Example message read response with an Auto-Renewal Message Object: Example message read response with an Expiration Deletion Message Object: Example message read response with a Service Notice Message Object, which is not linked to a specific object:

Query There is no request message for the message query operation. Example message query response, containing a Transfer Request Message Object for the domain transfer-test.example:

Discovery

Read There is no request message for the discovery read operation. Example discovery read response:

Profile

Read There is no request message for the profile read operation. Example profile read response: Example profile read response for a profile that extends a base profile:

IANA Considerations

RPP Media Type (application/rpp+json) The IANA is requested to add the "application/rpp+json" media type to the "Media Types" registry, following the template in

RPP JSON Schema Extensions Registry The IANA is requested to create a new registry for RPP JSON Schema extensions, this registry will be used to register standardised JSON Schema extensions to the JSON representation of RPP JSON objects, as defined in the of this document. Fields to be registered:

name: The name of the extension, for example "RPP example JSON extension".
id: The unique $id URI of the extension JSON Schema, for example "https://www.iana.org/assignments/rpp-json-extensions/example-extension.json".
version: The version of the extension, for example "1.0".
RFC: The RFC number for the extension specification, for example "RFC 1234".
description: A human-readable description of the extension and its intended use.

The "RESTful Provisioning Protocol (RPP)" registry group is defined in . TODO there already exists a registry for RPP extensions, defined in the rpp-core document, consider whether this new registry is necessary or if it should be merged with the existing one.

Internationalization Considerations Internationalized Domain Names (IDN) are represented using the paired name/uName data elements described in the Domain Name Data Object. The name property, which addresses the resource and is used for object identity, MUST always contain the ACE A-label form; the uName property carries the Unicode U-label form for display and MUST be Unicode Normalization Form C (NFC) as defined in . Because JSON strings are Unicode character sequences, implementations MUST take the following precautions to avoid producing or misinterpreting uName (and any other Unicode-bearing string value) in a way that breaks common JSON consumers, in particular JavaScript, whose native strings are sequences of UTF-16 code units rather than Unicode code points:

A string value MUST NOT contain an unpaired (lone) UTF-16 surrogate code point. Although , Section 8.2, permits such values in JSON text, they do not represent a well-formed Unicode scalar value and MAY fail to round-trip when re-encoded as UTF-8 by non-JavaScript consumers.
Consumers MUST NOT use a UTF-16 code-unit count (such as JavaScript's String.prototype.length) to validate or enforce length constraints on uName. Domain name label length restrictions are defined in terms of the name (A-label) octet count; scripts using supplementary-plane code points are represented as surrogate pairs in UTF-16, so a code-unit count over-counts the number of Unicode code points.
Consumers MUST NOT rely on native string equality (e.g. JavaScript's ===) to compare uName values for identity purposes, since two strings that differ only in Unicode normalisation form compare unequal even though they represent the same name; name MUST be used instead for identity comparison.
Servers MUST emit name in lowercase. Because ACE (A-label) domain names are case-insensitive per DNS comparison rules but JSON/JavaScript string comparison is case-sensitive, clients performing equality checks on name MUST lowercase both operands first rather than relying on native string equality.

Security Considerations If a server accepts a uName value without strictly verifying both that it is already Unicode Normalization Form C (NFC) and that it converts, via the ToASCII procedure of , to the exact supplied name value, an attacker could cause a displayed uName to diverge from the name actually resolved in the DNS. Servers MUST reject any create or update request where uName is not already NFC-normalised or does not convert to the supplied name; servers MUST NOT silently normalise or correct the value on the client's behalf. TODO

Acknowledgments TODO

Change History

Version draft-ietf-rpp-json-00 to draft-ietf-rpp-json-01

Added section "JSON Schema Composition Rules", including and read/create views for Component Objects. (Issue #102)
Replaced the JSON Schema definitions the generic Status Object with type-specific objects in .
Added schemas and examples for the Availability Objects, service messages, and for the Discovery and Profile Data Objects. (Issues #133, #82, #92)
Added IANA registration for the "application/rpp+json" media type. (Issue #84)
Added . (Issue #78)
Added support for Internationalized Domain Names (IDN), see . (Issue #86)
Added section "Inheritance". (Issue #88)
Added section describing the representation of undefined values. (Issue #97)
Rewrote the partial update processing in , added replacement of a whole array, and added flow figures. (Issue #91, #77)
Aligned the JSContact constraints with the RPP JSContact profile. (Issue #76)
Made the JSON Schema definitions normative. (Issue #75)
Added the Disclose Object and Process Object (Base, Create, Restore) schemas, and removed redundant Process alias and Transfer response schemas.
Aligned all examples, including those in , with the updated JSON Schema definitions.
Generalised the Mutability Rules to derive request and response content from operation Input and Output, and changed handling of disallowed request data elements from ignore to reject. (Issue #112 M2)
Replaced the primary identifier with the Unique Identifier in Association Rules.
Editorial: Removed rule numbering and replaced references to rule numbers with section references.
Editorial: Updated the Abstract and Introduction to cover all defined Data Objects.

Version draft-wullink-rpp-json-02 to draft-ietf-rpp-json-00

Added "Update Rules" section, describing update requests (Issue #54)
Added schema and examples for Transfer approve/reject/cancel operations (Issue #28)
Added Organisation and User Object JSON schemas and examples. (Issue #57)
Added schema and examples for the Renew Process Object. (Issue #45)
Added representation for embedding of process data in uniform interface operations (especially create). (Issue #60)
Added JSContact as External Type Embedding (Issue #43).

Version 01 to 02

Adjust rule 19 and all schemas to match it.

Version 00 to 01

Updated all examples and schemas to be based on RPP Data Object and no longer on EPP XML schemas. (Issue #15)
Updated labelled and dictionary aggregation rules (Issue #17)
Added required "@type" property to all JSON Schema definitions. (Issue #20)
Updated all example domain names to use the .example TLD. (Issue #26)