Redfish DMTF Logo

Document Identifier: DSP0266

Date: 2024-08-01

Version: 1.20.2

Redfish Specification






Supersedes: 1.20.1

Document Class: Normative

Document Status: Published

Document Language: en-US
Copyright Notice
Copyright © 2015-2024 DMTF. All rights reserved.

DMTF is a not-for-profit association of industry members dedicated to promoting enterprise and systems management and interoperability. Members and non-members may reproduce DMTF specifications and documents, provided that correct attribution is given. As DMTF specifications may be revised from time to time, the particular version and release date should always be noted.

Implementation of certain elements of this standard or proposed standard may be subject to third party patent rights, including provisional patent rights (herein "patent rights"). DMTF makes no representations to users of the standard as to the existence of such rights, and is not responsible to recognize, disclose, or identify any or all such third party patent right, owners or claimants, nor for any incomplete or inaccurate identification or disclosure of such rights, owners or claimants. DMTF shall have no liability to any party, in any manner or circumstance, under any legal theory whatsoever, for failure to recognize, disclose, or identify any such third party patent rights, or for such party's reliance on the standard or incorporation thereof in its product, protocols or testing procedures. DMTF shall have no liability to any party implementing such standard, whether such implementation is foreseeable or not, nor to any patent owner or claimant, and shall have no liability or responsibility for costs or losses incurred if a standard is withdrawn or modified after publication, and shall be indemnified and held harmless by any party implementing the standard from any and all claims of infringement by a patent owner for such implementations.

For information about patents held by third-parties which have notified DMTF that, in their opinion, such patent may relate to or impact implementations of DMTF standards, visit http://www.dmtf.org/about/policies/disclosures.php.

This document's normative language is English. Translation into other languages is permitted.

CONTENTS

Foreword

DMTF's Redfish Forum develops the Redfish standard.

DMTF is a not-for-profit association of industry members that promotes enterprise and systems management and interoperability. For information about DMTF, see DMTF.

This version supersedes version 1.20.1. For a list of the changes, see ANNEX A (informative) Change log.

Acknowledgments

DMTF acknowledges the following individuals for their contributions to the Redfish standard, including this document and Redfish schemas, interoperability profiles, and message registries:

Introduction

Redfish is a standard that uses RESTful interface semantics to access a schema-based data model to conduct management operations. It is suitable for a wide range of devices, from stand-alone servers to composable infrastructures and large-scale cloud environments.

The initial Redfish scope targeted servers. DMTF and its alliance partners expanded that scope to cover most data center IT equipment and other solutions, and both in- and out-of-band access methods.

Additionally, DMTF and other organizations that use Redfish as part of their industry standard or solution have added educational material.

This document defines the RESTful interface protocol and the various concepts and services necessary to implement a Redfish interface. The definition of the schema based data model and standard messages for the Redfish interface are covered separately in the following documents:

1 Scope

This specification defines the required protocols, data model, behaviors, and other architectural components for an interoperable, multivendor, remote, and out-of-band capable interface. This interface meets the cloud-based and web-based IT professionals' expectations for scalable platform management. While large and hyperscale environments are the primary focus, clients can use the specification for individual system management.

The specification defines the required elements for all Redfish implementations, and the optional elements that system vendors and manufacturers can choose. This specification also defines at which points an implementation can provide OEM-specific extensions.

The specification sets normative requirements for Redfish services and associated materials, such as Redfish schema files. In general, the specification does not set requirements for Redfish clients but indicates how a client can successfully and effectively access and use a Redfish service.

The specification does not require that implementations of the Redfish interfaces and functions require particular hardware or firmware.

2 Normative references

The following documents are referred to in the text in such a way that some or all of their content constitutes requirements of this document. For dated references, only the edition cited applies. For undated references, the latest edition of the referenced document (including any amendments) applies.

3 Terms, definitions, symbols, and abbreviated terms

Some terms and phrases in this document have specific meanings beyond their typical English meanings. This clause defines those terms and phrases.

The terms "shall" ("required"), "shall not", "should" ("recommended"), "should not" ("not recommended"), "may", "need not" ("not required"), "can" and "cannot" in this document are to be interpreted as described in ISO/IEC Directives, Part 2, Clause 7. The terms in parenthesis are alternatives for the preceding term, for use in exceptional cases when the preceding term cannot be used for linguistic reasons. Note that ISO/IEC Directives, Part 2, Clause 7 specifies additional alternatives. Occurrences of such additional alternatives shall be interpreted in their normal English meaning.

The terms "clause", "subclause", "paragraph", and "annex" in this document are to be interpreted as described in ISO/IEC Directives, Part 2, Clause 6.

The terms "normative" and "informative" in this document are to be interpreted as described in ISO/IEC Directives, Part 2, Clause 3. In this document, clauses, subclauses, or annexes labeled "(informative)" do not contain normative content. Notes and examples are always informative elements.

The term "deprecated" in this document is to be interpreted as material that is not recommended for use in new development efforts. Existing and new implementations may use this material, but they should move to the favored approach. Deprecated material may be implemented in order to achieve backwards compatibility. Deprecated material should contain references to the last published version that included the deprecated material as normative material and to a description of the favored approach. Deprecated material may be removed from the next major version of the specification.

This document defines these additional terms:

3.1 Hardware terms

3.1.1 baseboard management controller (BMC)

embedded device or service

Note 1 to entry: Typically, an independent microprocessor or system-on-chip with associated firmware in a computer system that completes out-of-band systems monitoring and management-related tasks.

3.1.2 IPMI

Intelligent Platform Management Interface

3.1.3 KVM-IP

keyboard, video, mouse redirection over IP

3.1.4 NIC

network interface controller

3.1.5 PCI

Peripheral Component Interconnect

3.1.6 PCIe

Peripheral Component Interconnect Express

3.2 Web development terms

3.2.1 CORS

cross-origin resource sharing

3.2.2 CRUD

basic Create, Read, Update, and Delete operations that any interface can support

3.2.3 CSRF

cross-site request forgery

3.2.4 HTTP

Hypertext Transfer Protocol

3.2.5 HTTPS

Hypertext Transfer Protocol Secure

Note 1 to entry: TLS secures HTTP.

3.2.6 hypermedia API

API that enables you to navigate through URIs that a service returns

3.2.7 IP

Internet Protocol

3.2.8 JSON

JavaScript Object Notation

3.2.9 message

complete HTTP-formatted or HTTPS-formatted request or response

Note 1 to entry: In the REST-based Redfish protocol, every request results in a response.

3.2.10 OData

Open Data Protocol (OData), as defined in OData Version 4.0 Part 1: Protocol

3.2.11 OData service document

resource that provides information about the service root for generic OData clients

3.2.12 operation

HTTP POST, GET, PUT, PATCH, HEAD, and DELETE request methods that map to generic CRUD operations

3.2.13 parent resource

parent to another resource if the initial segment of the resource URI is the same as the URI of the other resource, but is at least one level higher

Note 1 to entry: For example, /redfish/v1/Chassis/A88 is a parent resource of /redfish/v1/Chassis/A88/Assembly.

3.2.14 property

name-value pair in a Redfish-defined request or response

Note 1 to entry: A property can be any valid JSON data type.

3.2.15 request

message from a client to a service

3.2.16 response

message from a service to a client in response to a request message

3.2.17 subscription

registration of a destination to receive events

3.2.18 TCP

Transmission Control Protocol

3.2.19 TLS

Transport Layer Security

3.2.20 XSS

cross-site scripting

3.3 Redfish terms

3.3.1 action

a custom operation specific to one or more resources other than one of the standard CRUD operations

Note 1 to entry: For more information, see the Actions clause.

3.3.2 collection

see resource collection

3.3.3 event

data structure that corresponds to one or more alerts that have occurred

Note 1 to entry: For more information, see the Eventing clause.

3.3.4 excerpt

subset of data that is copied from one resource and presented in another resource

Note 1 to entry: An excerpt provides data in convenient locations without duplication of entire resources.

3.3.5 member

single resource instance in a resource collection

3.3.6 Redfish client

communicates with a Redfish service and accesses one or more of the service's resources or functions

3.3.7 Redfish protocol

discovers, connects to, and inter-communicates with a Redfish service

3.3.8 Redfish schema

a set of human and machine-readable documents that define Redfish resources using one or more of the supported schema definition languages

3.3.9 Redfish service

implementation of the protocols, resources, and functions that deliver the interface that this specification defines and its associated behaviors for one or more managed systems

Note 1 to entry: Also known as the service.

3.3.10 resource

URI-addressable Redfish data structure

3.3.11 resource collection

set of similar resources where the number of instances can shrink or grow

3.3.12 resource tree

tree structure of resources accessible through a well-known starting URI

Note 1 to entry: A client can discover the available resources on a Redfish service by following the resource hyperlinks from the base of the tree.

3.3.13 resource type

set of definitions for properties and actions contained within a resource and documented in the Redfish schema files

3.3.14 service root

starting-point resource for locating and accessing the other resources and associated metadata that make up an instance of a Redfish service

3.3.15 subordinate resource

is subordinate to another resource if the initial segment of the resource URI is the same as the URI of the other resource, but is at least one level deeper

Note 1 to entry: For example, /redfish/v1/Chassis/A88/Assembly is a subordinate resource of the Chassis resource named A88.

3.3.16 task

representation of a long-running operation

3.3.17 task monitor

opaque service-generated URI that the client that initiates the request can use to monitor an asynchronous operation

4 Typographical conventions

The following typographical convention indicates deprecated material:

DEPRECATED

Deprecated material appears here.

END DEPRECATED

In places where this typographical convention cannot be used, such as tables or figures, the "DEPRECATED" label is used alone.

5 Overview

Redfish is a management standard that uses a data model representation with a RESTful interface.

Being RESTful, Redfish is easier to use and implement.

Being model-oriented, it can express the relationships between components and the semantics of the Redfish services and components within them. The model is also easy to extend.

By requiring JSON representation, Redfish enables easy integration with programming environments. It is also easy to interpret by humans.

An interoperable Redfish schema defines this model, which is freely available and published in OpenAPI YAML, OData CSDL, and JSON Schema formats.

5.1 Goals

As an architecture, data model, and set of protocols that enable a client to access Redfish services, Redfish has these goals.

Table 1 describes these goals:

Table 1 — Redfish goals
Goal Purpose
Scalable Can scale on stand-alone machines or racks of equipment.
Flexible Can implement through existing hardware or entirely as a software service.
Extensible Can easily add new and vendor-specific capabilities to the data model.
Backward-compatible Can add capabilities while preserving investments in earlier implementations.
Interoperable Provides consistent functionality across multiple vendor implementations.
Standards-based Built on ubiquitous and secure protocols. Leverages other standards where applicable.
Simple Easy-to-use without the need for highly specialized programming skills or systems knowledge.
Lightweight Designed to reduce complexity and implementation costs. Minimizes the required footprint for implementations.

5.2 Design tenets

To deliver these goals, Redfish:

5.3 Limitations

Redfish minimizes the need for clients to complete upgrades by using strict versioning and forward-compatibility rules, and separation of the protocols from the data model. However, Redfish does not guarantee that clients never need to update their software. For example, clients might need to upgrade to manage new system or component types, or update the data model.

Interoperable does not mean identical. Many elements of Redfish are optional. Clients should be prepared to discover the optional elements by using the built-in discovery methods.

The resource tree reflects the topology of the system and its devices. Consequently, different hardware or device types result in different resource trees, even for identical systems from the same manufacturer. References between resources may result in a graph instead of a tree. Clients that traverse the resource tree should provide logic to avoid infinite loops.

Additionally, not all Redfish resources use simple REST read-and-write semantics. Different use cases may follow other types of client logic. For example, clients cannot simply read user credentials or certificates from one service and write them to another service.

Finally, the hyperlink values between resources and other elements can vary across implementations. Clients should not assume that they can reuse hyperlinks across different Redfish service instances.

5.4 Additional design background and rationale

5.4.1 REST-based interface

Redfish exposes many service applications as RESTful interfaces. This document defines a RESTful interface.

Redfish defines a RESTful interface because it:

5.4.2 Data-oriented

The Redfish data model is developed by focusing on the contents of the payload. By concentrating on the contents of the payload first, Redfish payloads are easily mapped to schema definition languages and encoding types. The data model is defined in various schema languages, including OpenAPI YAML, OData CSDL, and JSON Schema.

5.4.3 Separation of protocol from data model

Redfish separates the protocol operations from the data model and versions the protocol independently from the data model. This enables clients to extend and change the data model as needed without requiring the protocol version to change.

5.4.4 Hypermedia API service root

Redfish has a single service root URI and clients can discover all resources through referenced URIs. The hypermedia API enables the discovery of resources through hyperlinks.

5.4.5 OpenAPI v3.0 support

The OpenAPI v3.0 provides a rich ecosystem of tools for using RESTful interfaces that meet the design requirements of that specification. Starting with Redfish Specification v1.6.0, the Redfish schemas support the OpenAPI YAML file format and URI patterns that conform to the OpenAPI Specification were defined. Conforming Redfish services that support the Redfish protocol version v1.6.0 or later implement those URI patterns to enable use of the OpenAPI ecosystem.

For details, see OpenAPI Specification v3.0.

5.4.6 OData conventions

With the popularity of RESTful APIs, there are nearly as many RESTful interfaces as there are applications. While following REST patterns helps promote good practices, due to design differences between the many RESTful APIs there few common conventions between them.

To provide for interoperability between APIs, OData defines a set of common RESTful conventions and annotations. Redfish follows OData conventions for describing schema, URL conventions, and definitions for typical properties in a JSON payload.

5.5 Service elements

5.5.1 Synchronous and asynchronous operation support

Some operations can take more time than a client typically wants to wait. For this reason, some operations can be asynchronous at the discretion of the service. The request portion of an asynchronous operation is no different from the request portion of a synchronous operation.

To determine whether an operation was completed synchronously or asynchronously, clients can review the HTTP status codes. For more information, see the Asynchronous operations clause.

5.5.2 Eventing mechanism

Redfish enables clients to receive messages outside the normal request and response paradigm. The service uses these messages, or events, to asynchronously notify the client of a state change or error condition, usually of a time critical nature.

This specification defines two styles of eventing:

For information, see the Eventing clause.

5.5.3 Actions

Actions are Redfish operations that do not easily map to RESTful interface semantics. These types of operations may not directly affect properties in the Redfish resources. The Redfish schema defines certain standard actions for common Redfish resources. For these standard actions, the Redfish schema contains the normative language on the behavior of the action.

5.5.4 Service discovery

While the service itself is at a well-known URI, clients need to discover the network address of the service. Like UPnP, Redfish uses SSDP for discovery. A wide variety of devices, such as printers and client operating systems, support SSDP. It is simple, lightweight, IPv6 capable, and suitable for implementation in embedded environments.

For more information, see the Discovery clause.

5.5.5 Remote access support

Remote management functionality typically includes access mechanisms for redirecting operator interfaces such as serial console, keyboard video and mouse (KVM-IP), command shell, or command-line interface, and virtual media. While these mechanisms are critical functionality, they cannot be reasonably implemented as a RESTful interface.

Therefore, this standard does not define the protocols or access mechanisms for those services but encourages implementations that leverage existing standards. However, the Redfish schema includes resources and properties that enable client discovery of these capabilities and access mechanisms to enable interoperability.

5.6 Security

The challenge of remote interface security is to protect both the interface and exchanged data. To accomplish this, Redfish provides authentication and encryption. As part of this security, Redfish defines and requires minimum levels of encryption.

For more information, see the Security details clause.

6 Protocol details

In this document, the Redfish protocol refers to the RESTful mapping to HTTP, TCP/IP, and other protocol, transport, and messaging layer aspects. HTTP is the application protocol that transports the messages and TCP/IP is the transport protocol. The RESTful interface is a mapping to the message protocol.

The Redfish protocol is designed around a web service-based interface model, which provides network and interaction efficiency for both user interface (UI) and automation usage. Specifically, the protocol can leverage existing tool chains.

Table 2 describes the items that the Redfish protocol uses:

Table 2 — Redfish protocol
Item Description
HTTP methods Maps to common CRUD operations.
Actions Expands operations beyond CRUD-type operations.
Media types Negotiates the type of data sent in the message body.
HTTP status codes Indicates the success or failure of the server's request.
Error responses Returns more information than HTTP status codes.
TLS Secures messages. See Security details.
Asynchronous semantics Manages long-running operations.

A Redfish interface shall be exposed through a web service endpoint implemented by using HTTP version 1.1. See RFC7230, RFC7231, and RFC7232.

A Redfish interface may additionally be exposed through a web service endpoint implemented by using HTTP version 2.0. See RFC7540.

The subsequent clauses describe how the Redfish interface uses and adds constraints to HTTP to ensure interoperability of Redfish implementations.

6.1 Universal Resource Identifiers

A Universal Resource Identifier (URI) identifies a resource, including the service root and all Redfish resources.

A GET operation on a URI returns a representation of the resource with properties and hyperlinks to associated resources. The service root URI is well known and is based on the protocol version.

To discover the URIs to additional resources, extract the associated resource hyperlinks from earlier responses. The hypermedia API enables the discovery of resources through hyperlinks.

Redfish considers the RFC3986-defined scheme, authority, service root, and version, and unique resource path component parts of the URI.

For example, this URI:

https://mgmt.vendor.com/redfish/v1/Systems/1

Contains these component parts:

In a URI:

In an implementation:

For the absolute URI definition, see RFC3986.

For example, a POST operation may return the /redfish/v1/Systems/2 URI in the Location header of the response, which points to the POST-created resource.

Assuming that the client connects through the mgmt.vendor.com appliance, the client accesses the resource through the https://mgmt.vendor.com/redfish/v1/Systems/2 absolute URI.

URIs that conform to RFC3986 may also contain the query, ?query, and frag, #frag, components. For information about queries, see Query parameters. When a URI includes a fragment (frag) to submit an operation, the server ignores the fragment.

If a property in a response references another property within a resource, use the RFC6901-defined URI fragment identifier representation format. If the property is a reference property in the schema, the fragment shall reference a valid resource identifier. For example, the following fragment identifies a property at index 0 of the Fans array in the /redfish/v1/Chassis/MultiBladeEncl/Thermal resource:

{
   "@odata.id": "/redfish/v1/Chassis/MultiBladeEncl/Thermal#/Fans/0"
}

For requirements on constructing Redfish URIs, see the resource URI patterns annotation clause.

6.2 HTTP methods

Table 3 describes the mapping of HTTP methods to the Redfish-supported operations. If the Required column contains Yes, a Redfish interface shall support the HTTP method. If the Required column contains No, a Redfish interface may support the HTTP method.

Table 3 — Mapping of HTTP methods to Redfish-supported operations
HTTP method Interface semantic Required
POST Create resource
Resource action
Eventing		 Yes
GET Retrieve resource Yes
PUT Replace resource No
PATCH Update resource Yes
DELETE Delete resource Yes
HEAD Retrieve resource header No
OPTIONS Retrieve header
Cross-origin resource sharing (CORS) pre-flight		 No

For HTTP methods that the Redfish service does not support or that Table 3 omits, the Redfish service shall return the HTTP 405 Method Not Allowed status code or the HTTP 501 Not Implemented status code.

6.3 HTTP redirect

HTTP redirect enables a service to redirect a request to another URL. Among other things, HTTP redirect enables Redfish resources to alias areas of the data model.

All Redfish clients shall correctly handle HTTP redirect.

The service for the redirected resource shall enforce the authentication and authorization requirements for the redirected resource.

6.4 Media types

Some resources may be available in more than one type of representation. The media type indicates the representation type.

In HTTP messages, the media type is specified in the Content-Type header. To tell a service to return the response through certain media types, the client sets the HTTP Accept header to a list of the media types.

To request compression in the response body, clients specify an Accept-Encoding request header.

6.5 ETags

To reduce unnecessary RESTful accesses to resources, the Redfish service should support the association of a separate entity tag (ETag) with each resource.

Because the service knows whether the new version of the object is substantially different, the service generates and provides the ETag as part of the resource payload.

The ETag mechanism supports both strong and weak validation. If a resource supports an ETag, it shall use the RFC7232-defined ETag.

This specification does not mandate a particular algorithm for ETag creation, but ETags should be highly collision-free.

An ETag can be:

If a client performs a PUT operation or PATCH operation to update a resource, it should include an ETag from a previous GET in the HTTP If-Match or If-None-Match header. Both strong and weak ETags are allowed in these headers. If a service supports the return of the ETag header on a resource, it may respond with the HTTP 428 Precondition Required status code if the If-Match or If-None-Match header is missing from the PUT or PATCH request for the same resource, as specified in RFC6585. Services should perform weak ETag comparison when verifying the ETag provided by the client in PUT or PATCH operations. Clients should treat ETags received from services as opaque values and not modify them when providing them in PUT or PATCH operations.

In addition to the return of the ETag property on each resource, a Redfish service should return the ETag header on:

The format of the ETag header is:

ETag: <string>

For responses to $expand requests:

For resources that update frequently, such as resources with a date-time configuration or a sensor reading, implementations might consider methods to reduce frequent ETag updates, such as rounding a property value when calculating an ETag for the resource. This might be important if clients want to use the If-Match header to prevent collisions with other clients. For example, if a client is attempting to modify a Manager resource and the DateTime property updates the ETag every second, a client performing a PATCH operation on the Manager resource with the If-Match header will likely contain an old ETag if DateTime is not handled specifically to reduce frequent ETag updates, causing the service to respond with the HTTP 412 Precondition Failed status code.

6.6 Protocol version

The protocol version is separate from the resources' version or the Redfish schema version that the resources support.

Each Redfish protocol version is strongly typed by using the URI of the Redfish service in combination with the resource obtained at that URI, called the ServiceRoot resource.

The root URI for this version of the Redfish protocol shall be /redfish/v1/.

The URI defines the major version of the protocol.

The RedfishVersion property of the ServiceRoot resource defines the protocol version, which includes the major version, minor version, and errata version of the protocol, as defined in the Redfish schema for that resource.

The protocol version is a string in the format:

<MajorVersion>.<MinorVersion>.<ErrataVersion>

where

Any resource that a client discovers through hyperlinks that the service root or any service root-referenced service or resource returns shall conform to the same protocol version that the service root supports.

A GET operation on the /redfish resource shall return this response body:

{
   "v1": "/redfish/v1/"
}

6.7 Redfish-defined URIs and relative reference rules

Table 4 describes the Redfish-defined URIs. Redfish services shall support URIs in the table if the Required column contains Yes. Redfish services should support URIs in the table if the Required column contains No.

Table 4 — Redfish-defined URIs
URI Returns Required
/redfish Version. A major update that does not preserve compatibility with earlier minor versions. Yes
/redfish/v1/ Redfish service root. Yes
/redfish/v1/odata Redfish OData service document. Yes
/redfish/v1/$metadata Redfish metadata document. Yes
/redfish/v1/openapi.yaml Redfish OpenAPI YAML document. No
/redfish/v1/Schemas/<SchemaFile> Local copy of a Redfish schema file, where <SchemaFile> is the file name of the local schema file. No
/redfish/v1/Registries/<RegistryFile> Local copy of a Redfish registry file, where <RegistryFile> is the file name of the local registry file. No
/redfish/v1/TaskService/TaskMonitors/<TaskMonitorId> Redfish task monitor, where <TaskMonitorId> is the identifier of the task monitor. No

In addition, Table 5 describes the URIs that services shall process without a trailing slash in one of these ways:

Table 5 — Redfish-defined URIs without trailing slashes
URI Associated Redfish-defined URI
/redfish/v1 /redfish/v1/
/redfish/ /redfish

All other Redfish service-supported URIs shall match the resource URI patterns definitions, except the supplemental resources that the @Redfish.Settings, @Redfish.ActionInfo, and @Redfish.CollectionCapabilities payload annotations reference. The client shall treat the URIs for these supplemental resources as opaque.

All Redfish-defined URIs and URIs starting with /redfish are reserved for future standardization by DMTF and DMTF alliance partners, except OEM extension URIs, which shall conform to the requirements of the OEM URIs clause.

All relative references that the service uses shall start with either:

For details, see RFC3986.

7 Service requests

This clause describes the requests that clients can send to Redfish services.

7.1 Request headers

Table 6 lists the HTTP-defined request headers and their requirements for Redfish services and clients.

For Redfish services:

For Redfish clients (sending the HTTP requests):

Table 6 — Request headers
Header Service requirement Client requirement Supported values Description
Accept Yes No RFC7231 Communicates to the server the media type or types that this client is prepared to accept.

Services shall support resource requests with Accept header values of application/json or application/json;charset=utf-8.

Services shall support XML metadata requests with Accept header values of application/xml or application/xml;charset=utf-8.

Services shall support OpenAPI YAML schema requests with Accept header values of application/yaml or application/yaml;charset=utf-8 or application/vnd.oai.openapi or application/vnd.oai.openapi;charset=utf-8.

Services shall support SSE requests with Accept header values of text/event-stream or text/event-stream;charset=utf-8.

Services shall support any request with Accept header values of application/*, application/*;charset=utf-8, */*, or */*;charset=utf-8.
Accept-Encoding No No RFC7231 Indicates whether the client can handle gzip-encoded responses. If a service cannot return an acceptable response to a request with this header, it shall respond with the HTTP 406 Not Acceptable status code. If the request omits this header, the service should not return gzip-encoded responses.
Accept-Language No No RFC7231 The languages that the client accepts in the response. If the request omits this header, uses the service's default language for the response.
Authorization Conditional Conditional RFC7617, RFC6749 Required for HTTP Basic authentication and OAuth 2.0.

A client can access unsecured resources without this header on systems that support Basic authentication.
Content-Length No No RFC7231 The size of the message body.

To indicate the size of the body, a client can use the Transfer-Encoding: chunked header.

If a service needs to use Content-Length and does not support Transfer-Encoding, it responds with the HTTP 406 Not Acceptable status code.
Content-Type Conditional Conditional RFC7231 The request format. Required for operations with a request body.

Services shall accept the Content-Type header set to either application/json or application/json;charset=utf-8.

It is recommended that clients use these values in requests because other values can cause an error.
Host Yes No RFC7230 Enables support of multiple origin hosts at a single IP address.
If-Match Conditional No RFC7232 To ensure that clients update the resource from a known state, PUT and PATCH requests for resources for which a service returns ETags shall support If-Match.

While not required for clients, it is highly recommended for PUT and PATCH operations.
If-None-Match No No RFC7232 A service only returns the resource if the current ETag of that resource does not match the ETag sent in this header.

If the ETag in this header matches the resource's current ETag, the GET operation returns the HTTP 304 Not Modified status code.
Last-Event-ID No No HTML5 SSE The event source's last id field from the SSE stream. Requests history event data.

See Server-sent events.
Max-Forwards No No RFC7231 Limits gateway and proxy hops.

Prevents messages from remaining in the network indefinitely.
OData-MaxVersion No No 4.0 The maximum OData version that an OData-aware client understands.
OData-Version Yes No 4.0 The OData version.

Services shall reject requests that specify an unsupported OData version.

If a service encounters an unsupported OData version, it should reject the request with the HTTP 412 Precondition Failed status code.
Origin Yes No Fetch Living Standard, 3.1. Origin header Enables web applications to consume a Redfish service while preventing CSRF attacks.
User-Agent Yes No RFC7231 Traces product tokens and their versions.

The header can list multiple product tokens.
Via No No RFC7230 Defines the network hierarchy and recognizes message loops.

Each pass inserts its own Via header.

Redfish services shall understand and be able to process the headers in Table 7 as defined by this specification if the Service requirement column contains Yes.

Table 7 — Request headers part 2
Header Service requirement Client requirement Supported values Description
X-Auth-Token Yes Conditional Opaque encoded octet strings Authenticates user sessions.
The token value shall be indistinguishable from random.
While services shall support this header, a client can access unsecured resources without establishing a session.

7.2 GET (read requests)

7.2.1 GET (read requests) overview

The GET operation retrieves resources from a Redfish service. Clients make a GET request to the individual resource URI. Clients may obtain the resource URI from published sources, such as the OpenAPI document, or from a resource identifier property in a previously retrieved resource response, such as the links property.

The service shall return the resource representation using one of the media types listed in the Accept header, subject to the requirements of the media types. If the Accept header is absent, the service shall return the resource's representation as application/json. Services may but are not required to support the convention of retrieving individual properties within a resource by appending a segment containing the property name to the URI of the resource.

If supported by the service, clients can perform a conditional GET operation by specifying an If-None-Match request header that contains the ETag of the resource.

7.2.2 Resource collection requests

Clients retrieve a resource collection by making a GET request to the resource collection URI. The response includes the resource collection's properties and an array of its members.

No requirements are placed on implementations to return a consistent set of members when a series of requests that use paging query parameters are made over time to obtain the entire set of members. These calls can result in missed or duplicate elements if multiple GET requests use paging to retrieve the Members array instances.

A subset of the members can be retrieved using client paging query parameters.

A service might not be able to return all of the contents of a resource collection request in a single response body. In this case, the response can be paged by the service. If a service pages a response to a resource collection request, the following rules shall apply:

7.2.3 Service root request

The root URL for Redfish version 1.x services shall be /redfish/v1/.

The service returns the ServiceRoot resource, as defined by this specification, as a response for the root URL.

Services shall not require authentication to retrieve the service root and /redfish resources.

7.2.4 OData service and metadata document requests

Redfish services expose two OData-defined documents at specific URIs to enable generic OData clients to navigate the Redfish service.

7.3 Query parameters

7.3.1 Query parameter overview

To paginate, retrieve subsets of resources, or expand the results in a single response, clients can include the query parameters. Some query parameters apply only to resource collections.

Services:

Services shall return:

Services should return:

Services shall process query parameters in this order:

Table 8 describes the query parameters:

Table 8 — Query parameters
Query parameter Description and example
excerpt Returns a subset of the resource's properties that match the defined Excerpt schema annotation.

If no Excerpt schema annotation is defined for the resource, the entire resource is returned.

Example:

https://resource?excerpt
$expand=<string> Returns a hyperlink and its contents in-line with retrieved resources, as if a GET call response was included in-line with that hyperlink.

See The $expand query parameter.

Example:

https://resource?$expand=*($levels=3)

https://resourcecollection?$expand=.($levels=1)
$filter=<string> Applies to resource collections. Returns a subset of collection members that match the $filter expression.

See The $filter query parameter.

Example:

https://resourcecollection?$filter=SystemType eq 'Physical'
only Applies to resource collections. If the target resource collection contains exactly one member, clients can use this query parameter to return that member's resource.

If the collection contains either zero members or more than one member, the response returns the resource collection, as expected.

Services should return the HTTP 400 Bad Request and an error response with the QueryCombinationInvalid message from the Base Message Registry if only is being combined with other query parameters.

Example:

https://resourcecollection?only
$select=<string> Returns a subset of the resource's properties that match the $select expression.

See The $select query parameter.

Example:

https://resource?$select=SystemType,Status
$skip=<integer> Applies to resource collections. Returns a subset of the members in a resource collection, or an empty set of members if the $skip value is greater than or equal to the member count. This paging query parameter defines the number of members in the resource collection to skip.

Example:

https://resourcecollection?$skip=5
$top=<integer> Applies to resource collections. Defines the number of members to show in the response.

Minimum value is 0, though a value of 0 returns an empty set of members.

Example:

https://resourcecollection?$top=30

Services may support OEM-defined query parameters. OEM-defined query parameter names shall not contain characters that conflict with syntax for query parameter parsing, such as &. OEM-defined query parameters shall be in the form:

OEM-<OemIdentifier>-<ParameterName>

where

For example, if Contoso defined a StatusOnly parameter, the query parameter would be OEM-Contoso-StatusOnly.

7.3.2 The $expand query parameter

The $expand query parameter enables a client to request a response that includes not only the requested resource, but also includes the contents of the subordinate or hyperlinked resources. The definition of this query parameter follows the OData Protocol Specification.

The $expand query parameter has a set of possible options that determine which hyperlinks in a resource are included in the expanded response.

Table 9 describes the Redfish-supported options for the $expand query parameter. The service may implement some of these options but not others. Any other supported syntax for $expand is outside the scope of this specification.

Table 9 — The $expand query parameter options
Option Description Example
asterisk (*) Shall expand all hyperlinks, including those in payload annotations, such as @Redfish.Settings, @Redfish.ActionInfo, and @Redfish.CollectionCapabilities. https://resource?$expand=*
$levels Number of levels the service should cascade the $expand operation.

The default level shall be 1.

For example, $levels=2 expands both the hyperlinks in the current resource (level 1), and the hyperlinks in the resulting expanded resources (level 2).		 https://resourcecollection?$expand=.($levels=2)
period (.) Shall expand all hyperlinks not in any links property instances of the resource, including those in payload annotations, such as @Redfish.Settings, @Redfish.ActionInfo, and @Redfish.CollectionCapabilities. https://resourcecollection?$expand=.
tilde (~) Shall expand all hyperlinks found in all links property instances of the resource. https://resourcecollection?$expand=~

Examples of $expand usage include:

When services execute $expand, they should include all supported properties in the requested resources.

When clients use $expand, they should be aware that the payload may increase beyond what can be sent in a single response.

If a client does not have the privileges required to access referenced resources in an expansion request, the service shall not expand the referenced resources. Services should not fail the request if the client has the privilges to access the URI of the requested resource.

If a service cannot return the payload due to its size, it shall return the HTTP 507 Insufficient Storage status code.

If a service cannot return the payload corresponding to an individual member of a resource collection, but can return other members, it should:

Responses may automatically include expanded resources due to the expanded resource annotation on reference properties, such as the Members property in the LogEntryCollection resource. $expand operates in parallel to the usage of this schema annotation. For example, since Members is automatically expanded, the following three requests produce identical responses:

The following example expands the RoleCollection resource with the level set to 1:

{
    "@odata.id": "/redfish/v1/AccountService/Roles",
    "@odata.type": "#RoleCollection.RoleCollection",
    "Name": "Roles Collection",
    "Members@odata.count": 3,
    "Members": [
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/Administrator",
            "@odata.type": "#Role.v1_1_0.Role",
            "Id": "Administrator",
            "Name": "User Role",
            "Description": "Admin User Role",
            "IsPredefined": true,
            "AssignedPrivileges": [ "Login", "ConfigureManager", "ConfigureUsers",
                                    "ConfigureSelf", "ConfigureComponents" ]
        },
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/Operator",
            "@odata.type": "#Role.v1_1_0.Role",
            "Id": "Operator",
            "Name": "User Role",
            "Description": "Operator User Role",
            "IsPredefined": true,
            "AssignedPrivileges": [ "Login", "ConfigureSelf", "ConfigureComponents" ]
        },
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/ReadOnly",
            "@odata.type": "#Role.v1_1_0.Role",
            "Id": "ReadOnly",
            "Name": "User Role",
            "Description": "ReadOnly User Role",
            "IsPredefined": true,
            "AssignedPrivileges": [ "Login", "ConfigureSelf" ]
        }
    ]
}

7.3.3 The $select query parameter

The $select query parameter indicates that the implementation should return a subset of the resource's properties that match the $select expression. If a request omits the $select query parameter, the response returns all properties by default. The definition of this query parameter follows the OData Protocol Specification.

The $select expression shall not affect the resource itself.

The $select expression defines a comma-separated list of properties to return in the response body.

The syntax for properties in objects or properties in arrays of objects shall be the object and property names concatenated with a slash (/). For arrays, the expression shall not contain the array index.

An example of $select usage is:

GET /redfish/v1/Systems/1?$select=Name,SystemType,Status/State

For successful responses, when services execute $select, they shall return all requested properties of the referenced resource that are supported. Services shall omit unsupported properties requested in the $select query parameter from the response. If a requested property is an object, the service shall return the entire object. The @odata.id and @odata.type properties shall be in the response payload and contain the same values as if $select was omitted. If the @odata.context property is supported, it shall be in the response payload and should be in the context property recommended format. If the @odata.etag property is supported, it shall be in the response payload and contain the same values as if $select was omitted. Services may include object-level messages or messages for requested properties in the response payload.

For unsuccessful responses, the response body shall contain an error response regardless of the properties requested in the $select query parameter.

Any other supported syntax for $select is outside the scope of this specification.

7.3.4 The $filter query parameter

The $filter parameter enables a client to request a subset of the resource collection's members based on the $filter expression. The definition of this query parameter follows the OData Protocol Specification.

The $filter query parameter defines a set of properties and literals with an operator.

A literal value can be:

If the literal value does not match the data type for the specified property, the service should reject $filter requests with the HTTP 400 Bad Request status code.

The $filter section of the OData ABNF Components Specification contains the grammar for the allowable syntax of the $filter query parameter, with the additional restriction that only built-in filter operations are supported.

Table 10 lists the Redfish-supported values for the $filter query parameter. Any other supported syntax for $filter is outside the scope of this specification.

Table 10 — The $filter query parameter options
Value Description Example
() Precedence grouping operator. (Status/State eq 'Enabled' and Status/Health eq 'OK') or SystemType eq 'Physical'
and Logical and operator. ProcessorSummary/Count eq 2 and MemorySummary/TotalSystemMemoryGiB gt 64
eq Equal comparison operator. ProcessorSummary/Count eq 2
ge Greater than or equal to comparison operator. ProcessorSummary/Count ge 2
gt Great than comparison operator. ProcessorSummary/Count gt 2
le Less than or equal to comparison operator. MemorySummary/TotalSystemMemoryGiB le 64
lt Less than comparison operator. MemorySummary/TotalSystemMemoryGiB lt 64
ne Not equal comparison operator. SystemType ne 'Physical'
not Logical negation operator. not (ProcessorSummary/Count eq 2)
or Logical or operator. ProcessorSummary/Count eq 2 or ProcessorSummary/Count eq 4

When evaluating expressions, services shall use the following operator precedence:

If the service receives an unsupported $filter query parameter, it shall reject the request and return the HTTP 501 Not Implemented status code.

The HEAD method differs from the GET method in that it shall not return message body information.

However, the HEAD method completes the same authorization checks and returns all the same meta information and status codes in the HTTP headers as a GET method.

Services may support the HEAD method to:

Services may support the HEAD method to verify resource accessibility.

Services shall not support any other use of the HEAD method.

The HEAD method shall be idempotent in the absence of outside changes to the resource.

Services shall reject HEAD requests that contain query parameters. Services should return the HTTP 400 Bad Request status code if provided with a query parameter in a HEAD request.

7.5 Data modification requests

7.5.1 Data modification requests overview

To create, modify, and delete resources, clients issue the following operations:

The following clauses describe the success and error response requirements common to all data modification requests.

7.5.2 Modification success responses

For POST (create) operations, the response from the service, after the create request succeeds, should be one of these responses:

For PATCH (update), PUT (replace), and DELETE (delete) operations, the response from the service, after successful modification, should be one of the following responses:

For details on successful responses to action requests, see POST (action).

7.5.3 Modification error responses

If the resource exists but does not support the requested operation, services shall return the HTTP 405 Method Not Allowed status code.

Otherwise, if the service returns a client 4XX or service 5XX status code, the service encountered an error and the resource shall not have been modified or created as a result of the operation.

7.6 PATCH (update)

To update a resource's properties, the service shall support the PATCH method.

The request body defines the changes to make to one or more properties in the resource that the request URI references. The PATCH request does not change any properties that are not in the request body. Services may accept a PATCH method with an empty JSON object, which indicates that the service should make no changes to the resource.

For resources that allow for properties to not be updated immediately, clients can perform PATCH requests to a designated settings resource. For more information, see the Settings resource clause.

See the Modification success responses clause for behavior when the PATCH operation is successful.

If supported by the service, clients can perform a conditional PATCH operation by specifying an If-Match or If-None-Match request header that contains the ETag of the resource.

Services may reject the update on certain properties based on their own policies and, in this case, not make the requested update. Services should not require clients to provide properties other than those the client is attempting to modify, including the Oem property.

A partial success of a PATCH operation occurs when a modification request for multiple properties results in at least one property updated successfully, but one or more properties could not be updated. In these cases, the service shall return the HTTP 200 OK status code and a resource representation with extended information that lists the properties that could not be updated. Examples include:

If all properties in the update request are read-only, unknown, or unsupported, but the resource can be updated, the service shall return the HTTP 400 Bad Request status code and an error response with messages that show the non-updatable properties.

The service shall ignore OData annotations in the request body, such as the resource identifier, type, and ETag properties, except for the conditions listed below. If the update request only contains OData annotations, the service shall return the HTTP 400 Bad Request status code and an error response with the NoOperation message from the Base Message Registry (preferred) or a modification success response, except for the conditions listed below.

In the absence of outside changes to the resource, the PATCH operation should be idempotent, although the original ETag value may no longer match.

7.7 PATCH on array properties

The Array properties clause describes the three styles of array properties in a resource.

Within a PATCH request, the service shall accept null to remove an element, and accept an empty object {} to leave an element unchanged. Array properties that use the fixed or variable length style remove those elements, while array properties that use the rigid style replace removed elements with null elements. A service may indicate the maximum size of an array by padding null elements at the end of the array sequence.

When processing a PATCH request, the order of operations shall be:

A PATCH request with fewer elements than in the current array shall remove the remaining elements of the array.

For example, a fixed length-style Flavors array indicates that the service supports a maximum of six elements, by padding the array with null elements, with four populated.

{
   "Flavors": ["Chocolate", "Vanilla", "Mango", "Strawberry", null, null]
}

A client could issue the following PATCH request to remove Vanilla, replace Strawberry with Cherry, and add Coffee and Banana to the array, while leaving the other elements unchanged.

{
   "Flavors": [{}, null, {}, "Cherry", "Coffee", "Banana"]
}

After the PATCH operation, the resulting array is:

{
   "Flavors": ["Chocolate", "Mango", "Cherry", "Coffee", "Banana", null]
}

7.8 PUT (replace)

To completely replace a resource, services may support the PUT method. The service may add properties to the response resource that the client omits from the request body, the resource definition requires, or the service normally supplies.

The PUT operation should be idempotent in the absence of outside changes to the resource, with the possible exception that the operation might change ETag values.

See the Modification success responses clause for behavior when the PUT operation is successful.

If supported by the service, clients can perform a conditional PUT operation by specifying an If-Match or If-None-Match request header that contains the ETag of the resource.

Services may reject requests that do not include properties that the resource definition (schema) requires.

7.9 POST (create)

To create a resource, services shall support the POST method on resource collections.

The POST request is submitted to the resource collection to which the new resource will belong. See the Modification success responses clause for behavior when the POST operation is successful.

The body of the create request contains a representation of the object to create. The service may ignore any service-controlled properties, such as Id, which would force the service to overwrite those properties. Additionally, the service shall set the Location header in the response to the URI of the new resource.

7.10 DELETE (delete)

To remove a resource, the service shall support the DELETE method. Resources subordinate to the resource removed by a DELETE method are typically removed, as the contents of subordinate resources are dependent on the parent resource. In some cases, related resources may also be relocated in the resource tree based on their definition and usage. Other resources in the resource tree may also be removed or incur side effects of a resource removal.

See the Modification success responses clause for behavior when the DELETE operation is successful.

7.11 POST (action)

Services shall support the POST method as a way for clients to send actions to resources.

To request actions on a resource, send the HTTP POST method to the URI of the action. The target property in the resource's Actions property shall contain the URI of the action. The URI of the action shall be in the format:

<ResourceUri>/Actions/<QualifiedActionName>

where

To determine the available actions and the valid parameter values for those actions, clients can query a resource directly.

Clients provide parameters for the action as a JSON object within the request body of the POST operation. For information about the structure of the request and required parameters, see the Actions property clause. Some parameter information may require that the client examine the Redfish schema that corresponds to the resource.

If the action request does not contain all required parameters, the service shall return the HTTP 400 Bad Request status code. If the action request contains unsupported parameters, the service shall ignore the unsupported parameters or return the HTTP 400 Bad Request status code. If an action does not have any required parameters, the service shall accept an empty JSON object in the HTTP body for the action request.

Note: Previous versions of the specification allowed a service to reject action requests with an empty JSON object in the HTTP body if the client is not providing parameters for the action request. Services implementing version v1.20.0 or higher are required to accept an empty JSON object in an action request when the client is not providing parameters for the action request.

Table 11 describes the HTTP status codes and additional information that the service shall return a response to a successful POST (action) request:

Table 11 — POST (action) status codes
To indicate HTTP status code Additional information
Success, and the action's schema definition does not contain an action response. 200 OK An error response, with a message that indicates success or any additional relevant messages. If the action was successfully processed and completed without errors, warnings, or other notifications for the client, the service should return the Success message from the Base Message Registry in the code property in the response body.
Success, and the action's schema definition contains an action response. 200 OK The response body conforms to the action response defined in the schema.
A new resource was created, and the action's schema definition does not contain an action response. 201 Created A Location response header set to the URI of the created resource. An error response, with a message that indicates success or any additional relevant messages. If the action was successfully processed and completed without errors, warnings, or other notifications for the client, the service should return the Success message or Created message from the Base Message Registry in the code property in the response body.
A new resource was created, and the action's schema definition contains an action response. 201 Created A Location response header set to the URI of the created resource. The response body conforms to the action response defined in the schema.
Additional time is required to process. 202 Accepted A Location response header set to the URI of a task monitor.
Success, and the action's schema definition does not contain an action response. 204 No Content No response body.

If an action requested by the client has no effect, such as a reset of a ComputerSystem where the ResetType parameter is set to On and the ComputerSystem is already On, the service should respond with the HTTP 200 OK status code and return the NoOperation message from the Base Message Registry.

If an error was detected and the action request was not processed, the service shall return an HTTP 4XX or HTTP 5XX status code. The response body, if provided, shall contain an error response that describes the error or errors.

Example successful action response:

{
   "error": {
      "code": "Base.1.8.Success",
      "message": "Successfully Completed Request",
      "@Message.ExtendedInfo": [{
         "@odata.type": "#Message.v1_1_1.Message",
         "MessageId": "Base.1.8.Success",
         "Message": "Successfully Completed Request",
         "Severity": "OK",
         "MessageSeverity": "OK",
         "Resolution": "None"
      }]
   }
}

7.12 Operation apply time

Services may accept the @Redfish.OperationApplyTime annotation in the following request bodies:

This annotation enables the client to control when an operation is carried out.

For example, if the client wants to delete a particular Volume resource, but can only safely do so when a reset occurs, the client can use this annotation to instruct the service to delete the Volume on the next reset.

If multiple operations are pending, the service shall process them in the order in which the service receives them.

Services that support the @Redfish.OperationApplyTime annotation for create operations on a resource collection and delete operations on members of a resource collection shall include the @Redfish.OperationApplyTimeSupport response annotation for the resource collection.

The following example is a response for a resource collection that supports the @Redfish.OperationApplyTime annotation in requests to create new members and delete existing members:

{
   "@odata.id": "/redfish/v1/Systems/1/Storage/SATAEmbedded/Volumes",
   "@odata.type": "#VolumeCollection.VolumeCollection",
   "Name": "Storage Volume Collection",
   "Description": "Storage Volume Collection",
   "Members@odata.count": 2,
   "Members": [{
      "@odata.id": "/redfish/v1/Systems/1/Storage/SATAEmbedded/Volumes/1"
   }, {
      "@odata.id": "/redfish/v1/Systems/1/Storage/SATAEmbedded/Volumes/2"
   }],
   "@Redfish.OperationApplyTimeSupport": {
      "@odata.type": "#Settings.v1_3_3.OperationApplyTimeSupport",
      "SupportedValues": ["Immediate", "OnReset"]
   }
}

In the previous example, a client can annotate their create request body on the VolumeCollection itself, or a delete operation on the Volumes within the VolumeCollection.

The following sample request deletes a Volume on the next reset:

DELETE /redfish/v1/Systems/1/Storage/SATAEmbedded/Volumes/2 HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
   "@Redfish.OperationApplyTime": "OnReset"
}

Services that support the @Redfish.OperationApplyTime annotation for an action shall include the @Redfish.OperationApplyTimeSupport response annotation for the action.

The following example is a response for a ComputerSystem resource that supports the @Redfish.OperationApplyTime annotation in the reset action request:

{
   "@odata.id": "/redfish/v1/Systems/1",
   "@odata.type": "#ComputerSystem.v1_5_0.ComputerSystem",
   "Actions": {
      "#ComputerSystem.Reset": {
         "target": "/redfish/v1/Systems/1/Actions/ComputerSystem.Reset",
         "ResetType@Redfish.AllowableValues": ["On", "ForceOff", "ForceRestart",
            "Nmi", "ForceOn", "PushPowerButton"],
         "@Redfish.OperationApplyTimeSupport": {
            "@odata.type": "#Settings.v1_3_3.OperationApplyTimeSupport",
            "SupportedValues": ["Immediate", "AtMaintenanceWindowStart"],
            "MaintenanceWindowStartTime": "2017-05-03T23:12:37-05:00",
            "MaintenanceWindowDurationInSeconds": 600,
            "MaintenanceWindowResource": {
               "@odata.id": "/redfish/v1/Systems/1"
            }
         }
      }
   },
   ...
}

In the previous example, a client can annotate their reset action request body on the ComputerSystem in the payload.

The following sample request completes a reset at the start of the next maintenance window:

POST /redfish/v1/Systems/1/Actions/ComputerSystem.Reset HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
   "ResetType": "ForceRestart",
   "@Redfish.OperationApplyTime": "AtMaintenanceWindowStart"
}

Services that support the @Redfish.OperationApplyTime annotation for the JSON part of a multipart HTTP POST operation shall include the @Redfish.OperationApplyTimeSupport response annotation for the property that specifies the URI of the multipart POST operation.

The following example is a response for an UpdateService resource that supports the @Redfish.OperationApplyTime annotation in the JSON part of a multipart request:

{
   "@odata.id": "/redfish/v1/UpdateService",
   "@odata.type": "#UpdateService.v1_8_0.UpdateService",
   "MultipartHttpPushUri": "/redfish/v1/UpdateService/update-multipart",
   "MultipartHttpPushUri@Redfish.OperationApplyTimeSupport": {
      "@odata.type": "#Settings.v1_3_3.OperationApplyTimeSupport",
      "SupportedValues": ["Immediate", "AtMaintenanceWindowStart"],
      "MaintenanceWindowStartTime": "2017-05-03T23:12:37-05:00",
      "MaintenanceWindowDurationInSeconds": 600,
      "MaintenanceWindowResource": {
         "@odata.id": "/redfish/v1/UpdateService"
      }
   },
   ...
}

Services that support the @Redfish.OperationApplyTime annotation for a resource collection, action, or multipart HTTP POST operation shall create a task, and respond with the HTTP 202 Accepted status code with a Location header set to the URI of a task monitor, if the client's request body contains @Redfish.OperationApplyTime in the request.

The Settings Redfish schema defines the structure of the @Redfish.OperationApplyTimeSupport object and the @Redfish.OperationApplyTime annotation value.

7.13 Deep operations

Implementations may support operations that modify the current resource as well as subordinate resources. These operations are known as deep operations. They give the client the ability to modify more than one resource with a single operation.

Table 12 describes the types of deep operations that this specification defines:

Table 12 — Deep operations
Operation Description Example
Deep PATCH (update) Modify a resource and one or more subordinate resources. Modify a ComputerSystem resource as well as subordinate Storage and NetworkInterface resources.
Deep POST (create) Create one or more resources in a resource collection and optionally subordinate resources. Create multiple ManagerAccount resources.

The body of deep operations contains the resource being modified as well as the subordinate resources being modified. This resource can be a collection or a single instance. These resources could be subordinate resources, subordinate resource collections, or subordinate members of resource collections. The client can omit properties from the request such as those it does not want to modify or that the service controls. Requests that include references to multiple instances, such as members of a collection, shall include the Members property as part of the request body.

To determine which members of subordinate resource collections are to be modified by a deep PATCH, services shall use the @odata.id property provided by the client to identify the member of the resource collection to be modified.

Clients may provide the @odata.etag property in subordinate resources being modified by a deep PATCH. If the request contains the If-Match or If-None-Match header, the service shall compare the ETag in the request header with the ETag of the resource specified by the URI. If this check passes, the operation can proceed using the @odata.etag values contained in the body of the subordinate resources. In this case, the operation on each subordinate resource shall be completed independently, where some subordinate values that pass the condition check proceed and the resources that fail do not proceed. In this case, annotated extended information shall be included in the subordinate resource representation of the response.

Failure semantics for deep operations are similar to that of other operations of similar type. If any properties in a deep PATCH operation succeeded, the result is a 200 OK with the results returned in the response, and the service should include extended information indicating warnings or errors. For a deep POST operation, if any member of the collection was created then a 201 Created shall be returned, and any members that were not created should have extended information in their place holders with sufficient identifying information, such as returning all of the properties provided in the POST request body for that member, as well as extended information indicating why the creation was not successful. When sending a deep POST request, the value of the Location header shall be that of one of the URIs created and should be that of one of the least subordinate URIs, such as that of a ComputerSystem resource and not one of the devices subordinate to the ComputerSystem resource.

If the request body for a deep operation contains resources that are not modifiable, but no modifications are requested for those resources, services shall not treat this as a modification request for those resources. For example, if the service root is not modifiable, meaning PATCH is not accepted on the resource, a client is allowed to provide the service root in the deep operation request body if there are no modifications to the service root.

Deep POST shall not be allowed on the SessionCollection resource.

The following deep PATCH example modifies two members of the RoleCollection resource:

PATCH /redfish/v1/AccountService/Roles.Deep HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
    "Members": [
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/OperatorRestricted",
            "AssignedPrivileges": [ "Login", "ConfigureComponents" ]
        },
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/ReadOnlyRestricted",
            "AssignedPrivileges": [ "Login" ]
        }
    ]
}

The following deep POST example creates two members in the RoleCollection resource:

POST /redfish/v1/AccountService/Roles.Deep HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
    "Members": [
        {
            "RoleId": "OperatorRestricted",
            "AssignedPrivileges": [ "Login", "ConfigureComponents" ]
        },
        {
            "RoleId": "ReadOnlyRestricted",
            "AssignedPrivileges": [ "Login" ]
        }
    ]
}

The following deep POST example creates one member in the EventDestinationCollection resource and creates a Certificate resource subordinate to the new EventDestination resource:

POST /redfish/v1/EventService/Subscriptions.Deep HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
    "Members": [
        {
            "Destination": "https://listener1.contoso.com",
            "Protocol": "Redfish",
            "Certificates": {
                "Members": [
                    {
                        "CertificateType": "PEM",
                        "CertificateString": "----- BEGIN CERTIFICATE ...."
                    }
                ]
            }
        }
    ]
}

The following deep PATCH example modifies the asset tag and BIOS settings of a ComputerSystem resource:

PATCH /redfish/v1/Systems/47832.Deep HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
    "AssetTag": "Inventory Tag 12394783431",
    "Bios": {
        "@odata.id": "/redfish/v1/Systems/47832/Bios",
        "@Redfish.Settings": {
            "@odata.type": "#Settings.v1_3_3.Settings",
            "SettingsObject": {
                "@odata.id": "/redfish/v1/Systems/47832/Bios/SD",
                "Attributes": {
                    "AdminPhone": "(123) 456-789",
                    "BootMode": "Uefi"
                }
            }
        }
    }
}

The following example shows a deep PATCH with ETags in the request:

PATCH /redfish/v1/AccountService/Roles.Deep HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
If-Match: <Collection ETag>
OData-Version: 4.0

{
    "Members": [
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/OperatorRestricted",
            "@odata.etag": "W/\"ABCDEFG\"",
            "AssignedPrivileges": [ "Login", "ConfigureComponents" ]
        },
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/ReadOnlyRestricted",
            "@odata.etag": "W/\"ABCDEFG\"",
            "AssignedPrivileges": [ "Login" ]
        }
    ]
}

The following example response shows a partial failure of a deep PATCH where the ETag provided in the request for the Role resource named ReadOnlyRestricted was incorrect:

HTTP/1.1 200 OK
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
ETag: <Resource collection ETag>
OData-Version: 4.0

{
    "Members": [
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/OperatorRestricted",
            "@odata.etag": "W/\"ABCDEFG\"",
            "AssignedPrivileges": [ "Login", "ConfigureComponents" ]
        },
        {
            "@odata.id": "/redfish/v1/AccountService/Roles/ReadOnlyRestricted",
            "@Message.ExtendedInfo": [
                {
                    "@odata.type": "#Message.v1_1_1.Message",
                    "MessageId": "Base.1.8.PreconditionFailed",
                    "RelatedProperties": [ "/AssignedPrivileges" ]
                }
            ]
        }
    ]
}

8 Service responses

This clause describes the responses that Redfish services can return to clients.

8.1 Response headers

HTTP defines headers for use in response messages. Table 13 defines those headers and their requirements for Redfish services:

Table 13 — Response headers
Header Required Supported values Description
Access-Control-Allow-Origin No Fetch Living Standard, 3.2.3. HTTP responses Prevents or allows requests based on originating domain. Prevents CSRF attacks.
Allow Yes POST, PUT, PATCH, DELETE, GET, HEAD Shall be returned with the HTTP 405 (Method Not Allowed) status code to indicate the valid methods for the request URI. Shall be returned with any GET or HEAD operation to indicate the other allowable operations for this resource.
Cache-Control Yes RFC7234 Shall be supported and indicates whether a response can or cannot be cached.
Content-Encoding No RFC7231 Encoding used to compress the message body.
Content-Length No RFC7231 Size of the message body. An optional means of indicating size of the body uses Transfer-Encoding: chunked, that does not use the Content-Length header. If a service does not support Transfer-Encoding and needs Content-Length instead, the service shall respond with the HTTP 411 Length Required status code.
Content-Type Yes RFC7231 The message body's representation type.

Services shall specify a Content-Type of application/json when returning resources as JSON.

Services shall specify a Content-Type of application/xml when returning metadata as XML.

Services shall specify a Content-Type of application/yaml or application/vnd.oai.openapi when returning OpenAPI schema as YAML.

Services shall specify a Content-Type of text/event-stream when returning an SSE stream.

;charset=utf-8 shall be appended to the Content-Type if specified in the chosen media-type in the Accept header for the request.
ETag Conditional RFC7232 An identifier for a specific version of a resource, often a message digest. The ETag header shall be included on responses to GETs of ManagerAccount resources.
Link Yes RFC8288 Link headers shall be returned, as described in the Link headers clause.
Location Conditional RFC7231 URI of a newly created resource. Shall be returned upon creation of a resource. Location and X-Auth-Token shall be included on responses that create user sessions.
Max-Forwards No RFC7231 Limits gateway and proxy hops. Prevents messages from remaining in the network indefinitely.
OData-Version Yes 4.0 OData version of the payload to which the response conforms.
Retry-After No RFC7231, Section 7.1.3 Informs a client how long to wait before requesting the task information again.
Server No RFC7231 A product token and its version. Multiple product tokens may be listed.

Note: Previous versions of the Specification marked this header as required. This has been changed because no use cases for requiring it have been identified.
Via No RFC7230 Defines the network hierarchy and recognizes message loops. Each pass inserts its own Via header.
WWW-Authenticate Conditional RFC7617 Provides a challenge to clients, such as directing a client to perform HTTP Basic authentication, when authentication headers in the request are missing or invalid. Services shall return this header if the HTTPBasicAuth property in the AccountService resource contains Enabled or is not implemented.
X-Auth-Token Yes Opaque encoded octet strings Contains the authentication token for user sessions. The token value shall be indistinguishable from random.

The Link header provides metadata information on the accessed resource in response to a HEAD or GET request. The metadata information can include hyperlinks from the resource and JSON Schemas that describe the resource.

The following example shows the Link headers for a ManagerAccount with an Administrator role, in addition to a Settings annotation:

Link: </redfish/v1/AccountService/Roles/Administrator>; path=/Links/Role
Link: <http://redfish.dmtf.org/schemas/Settings.json>
Link: </redfish/v1/JsonSchemas/ManagerAccount.v1_0_2.json>; rel=describedby

A Link header containing rel=describedby shall be returned on GET and HEAD requests for Redfish resources. If the referenced JSON Schema is a versioned schema, it shall match the version contained in the value of the @odata.type property returned in this resource.

A Link header satisfying annotations should be returned on GET and HEAD requests for Redfish resources.

8.3 Status codes

HTTP defines status codes that appear in responses. The status codes themselves provide general information about how the request was processed, such as whether the request was successful, if the client provided bad information, or the service encountered an error when processing the request.

Note: For security implications of extended errors, See Security details.

Table 14 lists HTTP status codes that have meaning or usage defined for a Redfish service, or are otherwise referenced by this specification. Other codes may be returned by the service as appropriate, and their usage is implementation-specific. For usage and additional requirements imposed by this specification, see the Description column.

Table 14 — HTTP status codes
HTTP status code Description
200 OK Request completed successfully and includes a representation in its body.
201 Created Request to create a resource completed successfully. The Location header shall be set to the canonical URI for the newly created resource. For POST (create) requests, the response body may include a representation of the newly created resource. For POST (action) requests, the response body shall include the action response.
202 Accepted Request has been accepted for processing but the processing has not been completed. The Location header shall be set to the URI of a task monitor that can later be queried to determine the status of the operation. If a response body is provided, it shall contain a representation of the Task resource.
204 No Content Request succeeded, but no response body is provided.
301 Moved Permanently Requested resource resides under a different URI.
302 Found Requested resource resides temporarily under a different URI.
304 Not Modified Service has made a conditional GET request where access is allowed but the resource content has not changed. Certain request headers, such as If-None-Match, initiate conditional requests to save network bandwidth if no change has occurred. See HTTP 1.1, sections 14.25 and 14.26.
400 Bad Request Request could not be processed because it contains invalid information, such as an invalid input field, or is missing a required value. The response body shall return an extended error as defined in the Error responses clause.
401 Unauthorized Authentication credentials included with this request are missing or invalid. Additional details are described in the Sensitive data clause.
403 Forbidden Service recognized the credentials in the request but those credentials do not possess authorization to complete this request. This code is also returned when the user credentials provided need to be changed before access to the service can be granted. For details, see the Security details clause.
404 Not Found Request specified a URI of a resource that does not exist. Additional details are described in the Sensitive data clause.
405 Method Not Allowed HTTP verb in the request, such as DELETE, GET, HEAD, POST, PUT, or PATCH, is not supported for this request URI. The response shall include an Allow header that provides a list of methods that the resource identified by the URI in the client request supports. Additional details are described in the Sensitive data clause.
406 Not Acceptable Accept header was specified in the request and the resource identified by this request cannot generate a representation that corresponds to one of the media types in the Accept header.
409 Conflict Creation or update request could not be completed because it would cause a conflict in the current state of the resources that the platform supports. For example, a conflict occurred due to an attempt to set multiple properties that work in a linked manner by using incompatible values.
410 Gone Requested resource is no longer available at the service and no forwarding address is known. This condition is expected to be considered permanent. Clients with hyperlink editing capabilities should delete references to the URI in the client request after user approval. If the service does not know or cannot determine whether the condition is permanent, client should use the HTTP 404 Not Found status code. This response is cacheable unless otherwise indicated.
411 Length Required Request did not use the Content-Length header to specify the length of its content but perhaps used the Transfer-Encoding: chunked header instead. The addressed resource requires the Content-Length header.
412 Precondition Failed Precondition check, such as check of the OData-Version, If-Match, or If-None-Match header, failed.
413 Payload Too Large Request payload, or a part in a multipart request, is larger than the maximum size the service supports.
415 Unsupported Media Type Request specifies a Content-Type for the body that is not supported.
428 Precondition Required Request did not provide the required precondition, such as an If-Match or If-None-Match header.
431 Request Header Field Too Large Service is unwilling to process the request because either an individual header field or the collection of all header fields are too large.
500 Internal Server Error Service encountered an unexpected condition that prevented it from fulfilling the request. The response body shall return an extended error as defined in the Error responses clause.
501 Not Implemented Service does not currently support the functionality required to fulfill the request. This response is appropriate when the service does not recognize the request method and cannot support the method for any resource.
503 Service Unavailable Service currently cannot handle the request due to temporary overloading or maintenance of the service. A service may use this response to indicate that the request URI is valid but the service is performing initialization or other maintenance on the resource. A service may also use this response to indicate that the service itself is undergoing maintenance, such as finishing initialization steps after reboot of the service.
507 Insufficient Storage Service cannot build the response for the client due to the size of the response.

8.4 OData metadata responses

8.4.1 OData metadata responses overview

OData metadata describes resources, resource collections, capabilities, and service-dependent behavior to generic OData consumers with no specific understanding of this specification. Clients are not required to request metadata if they already have sufficient understanding of the target service. For example, clients are not required to request metadata to request and interpret a JSON representation of a resource that this specification defines.

A client can access the OData metadata at the /redfish/v1/$metadata URI.

A client can access the OData service document at the /redfish/v1/odata URI.

8.4.2 OData $metadata

The OData metadata describes top-level service resources and resource types according to OData Common Schema Definition Language. The OData metadata is represented as an XML document with an Edmx root element in the http://docs.oasis-open.org/odata/ns/edmx namespace with an OData version attribute set to 4.0.

The service shall use the application/xml or application/xml;charset=utf-8 MIME types to return the OData metadata document as an XML document.

<edmx:Edmx xmlns:edmx="http://docs.oasis-open.org/odata/ns/edmx" Version="4.0">
  <!-- edmx:Reference and edmx:Schema elements go here -->
</edmx:Edmx>

8.4.2.1 Referencing other schemas

The OData metadata should include the namespaces for each of the Redfish resource types, along with the RedfishExtensions.v1_0_0 namespace. Dynamic clients that reference the OData metadata document leverage schema definitions that are referenced to understand the definitions of the resources in the service. However, there are cases where it might not be practical to maintain an accurate document, such as when resources are dynamically discovered by the service through devices that support Redfish Device Enablement.

These references shall use either:

<edmx:Reference Uri="http://redfish.dmtf.org/schemas/v1/ServiceRoot_v1.xml">
  <edmx:Include Namespace="ServiceRoot"/>
  <edmx:Include Namespace="ServiceRoot.v1_0_0"/>
</edmx:Reference>

...

<edmx:Reference Uri="http://redfish.dmtf.org/schemas/v1/VirtualMedia_v1.xml">
  <edmx:Include Namespace="VirtualMedia"/>
  <edmx:Include Namespace="VirtualMedia.v1_0_0"/>
</edmx:Reference>
<edmx:Reference Uri="http://redfish.dmtf.org/schemas/v1/RedfishExtensions_v1.xml">
  <edmx:Include Namespace="RedfishExtensions.v1_0_0" Alias="Redfish"/>
</edmx:Reference>

The service's OData metadata document shall include an EntityContainer that defines the top-level resources and resource collections.

8.4.2.2 Referencing OEM extensions

The OData metadata document may reference additional schema documents that describe OEM-specific extensions that the service uses.

For example, the OData metadata document may reference custom types for additional resource collections.

<edmx:Reference Uri="http://contoso.org/Schema/CustomTypes">
  <edmx:Include Namespace="CustomTypes"/>
</edmx:Reference>

8.4.3 OData service document

The OData service document serves as a top-level entry point for generic OData clients. More information about the OData service document can be found in the OData JSON Format Specification.

{
   "@odata.context": "/redfish/v1/$metadata",
   "value": [{
      "name": "Service",
      "kind": "Singleton",
      "url": "/redfish/v1/"
   }, {
      "name": "Systems",
      "kind": "Singleton",
      "url": "/redfish/v1/Systems"
   }, ...]
}

The service shall use the application/json MIME type to return the OData service document as a JSON object.

The JSON object shall contain the @odata.context context property set to /redfish/v1/$metadata.

The JSON object shall include a value property set to a JSON array that contains an entry for the service root and each resource that is a direct child of the service root.

Table 15 describes the properties that each JSON object entry includes:

Table 15 — JSON object properties
Property Description
name User-friendly resource name of the resource.
kind Type of resource. Value is Singleton for all cases defined by Redfish.
url Relative URL for the top-level resource.

8.5 Resource responses

Services use the application/json MIME type to return resources and resource collections as JSON payloads. A service shall not break responses for a single resource into multiple results.

The format of these payloads is defined by the Redfish schema. For rules about the Redfish schema and how it maps to JSON payloads, see the Data model and Schema definition languages clauses.

8.6 Error responses

HTTP status codes often do not provide enough information to enable deterministic error semantics. For example, if a client makes a PATCH call and some properties do not match while others are not supported, the HTTP 400 Bad Request status code does not tell the client which values are in error. Error responses provide the client more meaningful and deterministic error semantics.

To provide the client with as much information about the error as possible, a Redfish service may provide multiple error responses in the HTTP response. Additionally, the service may provide Redfish standardized errors, OEM-defined errors, or both, depending on the implementation's ability to convey the most useful information about the underlying error.

Table 16 describes the properties in the extended error response, which is a single JSON object:

Table 16 — Error properties
Property Description
code String. Defines a MessageId from the message registry. See the MessageId format clause for the format of MessageId.
message Displays a human-readable error message that corresponds to the message in the message registry.
@Message.ExtendedInfo Displays an array of message objects. Describes one or more error messages.

See the Schema definition languages clause for references to the schema definitions of the error response payload.

The @Message.ExtendedInfo property should be present in all error responses. If the @Message.ExtendedInfo property is present, all information necessary to process the error should be provided in the @Message.ExtendedInfo property. Clients should look for the @Message.ExtendedInfo property for error processing first, and fallback on the code and message properties if @Message.ExtendedInfo is not present.

The following sample error response contains two messages in the @Message.ExtendedInfo property that describe two different errors. The message described by the code and message properties do not provide actionable information for the client.

{
  "error": {
    "code": "Base.1.8.GeneralError",
    "message": "A general error has occurred. See Resolution for information on how to resolve the error.",
    "@Message.ExtendedInfo": [{
      "@odata.type": "#Message.v1_1_1.Message",
      "MessageId": "Base.1.8.PropertyValueNotInList",
      "RelatedProperties": [
        "/IndicatorLED"
      ],
      "Message": "The value Red for the property IndicatorLED is not in the list of acceptable values.",
      "MessageArgs": ["Red",
        "IndicatorLED"
      ],
      "Severity": "Warning",
      "MessageSeverity": "Warning",
      "Resolution": "Choose a value from the enumeration list that the implementation can support and resubmit the request if the operation failed."
    }, {
      "@odata.type": "#Message.v1_1_1.Message",
      "MessageId": "Base.1.8.PropertyNotWritable",
      "RelatedProperties": [
        "/SKU"],
      "Message": "The property SKU is a read only property and cannot be assigned a value.",
      "MessageArgs": ["SKU"],
      "Severity": "Warning",
      "MessageSeverity": "Warning",
      "Resolution": "Remove the property from the request body and resubmit the request if the operation failed."
    }]
  }
}

9 Data model

One of the key tenets of Redfish is the separation of protocol from the data model. This separation makes the data both transport and protocol agnostic. By concentrating on the data transported in the payload of the protocol (in HTTP, it is the HTTP body), Redfish can also define the payload in any encoding and the data model is intended to be schema-language agnostic. While Redfish uses the JSON data-interchange format, Redfish provides a common encoding type that ensures property naming conventions that make development easier in JavaScript, Python, and other languages. This encoding type helps the Redfish data model be more easily accessible in modern tools and programming environments.

The data model allows an OEM to extend the model by adding an OEM resource or extending a resource.

This clause describes common data model, resource, and Redfish schema requirements.

9.1 Resources

A resource is a single entity accessed at a specific URI. Services use the application/json MIME type to return resources as JSON payloads.

Each resource shall be strongly typed, defined by a resource type in a Redfish schema document, and identified in the response payload by the value of the type identifier property.

Responses for a single resource shall contain the following properties:

Responses may also contain other properties defined within that resource type. Responses shall not include any properties not defined by that resource type.

9.2 Resource types

A resource type defines the set of properties that may be returned in the response payload of a Redfish resource request. Each resource type is documented in a Redfish schema document, and those documents are known collectively as the Redfish schema. The resource type may also include definitions for actions available for that resource.

Resource types are named to match the contents and purpose of the resource that they define. For example, the Circuit resource type defines the properties and actions related to a single electrical circuit. Resource types provide global uniqueness for definitions across multiple schema files and allow for schema files to reference each other. Resource types may be defined by OEMs to extend the Redfish schema, and should follow the naming rules specified by the OEM resource types clause.

9.3 Resource collections

A resource collection is a set of resources that share the same schema definition. Services use the application/json MIME type to return resource collections as JSON payloads.

Resource collection responses shall contain the following properties:

Responses for resource collections may contain the following properties:

Responses for resource collections shall not contain any other properties with the exception of payload annotations.

9.4 OEM resources

OEMs and other third parties can extend the Redfish data model by creating additional resource types. Extending the data model is accomplished by defining an OEM resource type, and schema file, for each resource type, and creating hyperlinks to connect instances of new resources to the resource tree.

Companies, OEMs, and other organizations may also use the Oem property in resources, the links property, and the actions property to define additional properties, hyperlinks, and actions for standard Redfish resource types.

While the information and semantics of these extensions are outside of the standard, the schema representing the data, the resource itself, and the semantics around the protocol shall conform to the requirements in this specification. OEMs are encouraged to follow the design tenets and naming conventions in this specification when defining OEM resources or properties.

9.5 Common data types

9.5.1 Primitive types

Table 17 describes the primitive data types for properties and action parameters in the data model:

Table 17 — Primitive data types
Type Description
Boolean A variable with a value of true or false.
Number A number with optional decimal point or exponent. Number properties may restrict the representation to an integer or a number with decimal point.
String A sequence of characters enclosed with double quotes (").
Array A comma-separated set of the previous types enclosed with square braces ([ and ]). See the Array properties clause.
Object A set of properties enclosed with curly braces ({ and }). See the Structured properties clause.
Null null value, which the service uses when it is unable to determine the property's value due to an error or other temporary condition, or if the schema has requirements for using null for other special conditions.

When receiving values from the client, services should support other valid representations of the data in the specified JSON type. In particular, services should support valid integer and decimal values in exponential notation and integer values that contain a decimal point with no non-zero trailing digits.

9.5.2 Enumerations

Enumerations are frequently used in Redfish to promote readability and interoperability, especially compared to the use of string values when used for similar purposes. Enumerations aren't optimal in all cases. Properties with two values that are likely to not have additional values should consider the boolean type if the true and false values can be described by the property name. The following design tenets apply to enumerations:

9.5.3 Empty string values

String properties should return an empty string ("") for properties configured by a user or external service that have not been set to an initial value. This allows client software to identify the property as supported by the service and avoids the use of null, which indicates an error condition. For example, the AssetTag property must be set by the end user, and therefore would return an empty string ("") until assigned a value by the user, while a failure to read the stored AssetTag value due to a non-volatile memory error would return null. To improve interoperability, implementations should avoid the use of filler strings, such as N/A or <Empty>, to represent a value not set by a user.

9.5.4 GUID and UUID values

Globally Unique Identifier (GUID) and Universally Unique Identifier (UUID) values are unique identifier strings and shall use the RFC4122-defined format:

([0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12})

9.5.5 Date-Time values

Date-Time values are strings according to the ISO 8601 extended format, including the time offset or UTC suffix.

Date-Time values shall use the format:

<YYYY>-<MM>-<DD>T<hh>:<mm>:<ss>[.<SSS>](Z|((+|-)<HH>:<MM>))

where

For example, 2015-03-13T04:14:33+06:00 represents March 13, 2015 at 4:14:33 with a +06:00 time offset.

When the time of day is unknown or serves no purpose, the service shall report 00:00:00Z for the time of day value.

9.5.6 Duration values

Duration values are strings according to the ISO 8601 duration format, with the exception of not expressing a representation for years, months, weeks, or negative values. Duration values shall use the format:

P[<d>D][T[<h>H][<m>M][<s>[.<f>]S]]

where

Each field is optional and can contain more than one digit.

For example, Table 18 describes the following durations:

Table 18 — Durations
Value Duration
P90D Ninety days.
P3D Three days.
PT6H Six hours.
PT10S Ten seconds.
PT0.001S 0.001 seconds.
PT1H30M One hour and 30 minutes.

DEPRECATED: Duration values shall use the format: P[<y>Y][<m>M][<w>W][<d>D][T[<h>H][<m>M][<s>[.<f>]S]]. This definition allows for specifying years, months, and weeks. ISO 8601 does not specify an exact value for the duration of a year or of a month, which introduces interoperability challenges.

9.5.7 Reference properties

Reference properties provide a reference to another resource in the data model. Reference properties are JSON objects that contain an @odata.id property. The @odata.id property value is the URI of the referenced resource.

9.5.8 Non-resource reference properties

Non-resource reference properties provide a URI to services or documents that are not Redfish-defined resources. These properties shall include the Uri or URI term in their property name and shall be of type string. For example, AssemblyBinaryDataUri in the Assembly schema. The access protocol, request headers, response headers, and data format of the referenced URI may be defined in schema for that property. Non-resource reference properties that refer to local HTTP/S targets shall follow the Redfish protocol, including use of Redfish sessions and access control, unless otherwise specified by the property definition in schema.

9.5.9 Array properties

Array properties contain a set of values or objects, and appear as JSON arrays within a response body. Array elements shall all contain values of the same data type.

Table 19 describes the array types, regardless of the data type of the elements:

Table 19 — Array types
Array type Description
Fixed length Contains a static number of elements. The property definition sets or the implementation chooses the size of the array.
Variable length Contains a variable number of elements. The array size is not specified and the size varies among instances. The array size may change. This array style is the most common style.
Rigid The array index is meaningful. When elements are added to or removed from the array, the elements do not change their position, or index, in the array. An element that is removed from a rigid array shall be replaced by a null element and all other elements shall remain at their current index.

Empty elements in a rigid array property shall be represented by null elements. Any array property that uses this style shall indicate the rigid style in the long description of its schema definition.

Services may pad an array property with null elements at the end of the sequence to indicate the array size to clients. This practice is useful for small fixed-length arrays, and for variable or rigid arrays with a restrictive maximum size. Services should not pad array properties if the maximum array size is not restrictive. For example, an array property typically populated with two elements, that a service limits to a maximum of 16 elements, should not pad the array with 14 null elements.

9.5.10 Structured properties

Structured properties are JSON objects within a response body.

Some structured properties inherit from the Resource.v1_0_0.ReferenceableMember definition. Structured properties that follow this definition shall contain the MemberId and resource identifier properties.

Because the definition of structured properties can evolve over time, clients need to be aware of the inheritance model that the different structured property definitions use.

For example, the Location property definition in the Resource schema has gone through several iterations since the Resource.v1_1_0 type was introduced, and each iteration inherits from the earlier version so that existing references in other schemas can leverage the additions.

Structured property references need to be resolved for both local and external references.

A local reference is a resource that has a structured property in its own schema, such as ProcessorSummary in the ComputerSystem resource. In these cases, the type property for the resource is the starting point for resolving the structured property definition.

To find the latest applicable version, clients can step the version of the resource backwards.

For example, if a service returns #ComputerSystem.v1_4_0.ComputerSystem as the resource type, a client can step backwards from ComputerSystem.v1_4_0, to ComputerSystem.v1_3_0, to ComputerSystem.v1_2_0, and so on, until it finds the ProcessorSummary structured property definition.

An external reference is a resource that has a property that references a definition found in a different schema, such as the Location property in the Chassis resource.

In these cases, clients can use the latest version of the external schema file as a starting point to resolve the structured property definition.

For example, if the latest version of the Resource schema is 1.6.0, a client can go backward from Resource.v1_6_0, to Resource.v1_5_0, to Resource.v1_4_0, and so on, until it finds the Location structured property definition.

9.5.11 Message object

9.5.11.1 Overview

A message object provides additional information about an object, property, or error response.

Table 20 describes the properties of the message object, which is a JSON object:

Table 20 — Message object properties
Property Type Required Defines
MessageId String Yes Error or message. Do not confuse this value with the HTTP status code. Clients can use this code to access a detailed message from a message registry.
Message String No Human-readable error message that indicates the semantics associated with the error. This shall be the complete message, and not rely on substitution variables.
RelatedProperties An array of JSON pointers No Properties in a JSON payload that the message describes.
MessageArgs An array of strings No Substitution parameter values for the message. If the parameterized message defines a MessageId, the service shall include the MessageArgs in the response.
MessageSeverity String (enumeration) No Severity of the error. Services can replace the value of the MessageSeverity property defined in the message registry with a value more applicable to the implementation.
Severity String No Severity of the error. Services can replace the value of the Severity property defined in the message registry with a value more applicable to the implementation.

DEPRECATED: This property has been deprecated in favor of MessageSeverity.
Resolution String No Recommended actions to take to resolve the error. Services can replace the value of the Resolution property defined in the message registry with a service-defined resolution.

Each instance of a message object shall contain at least a MessageId, together with any applicable MessageArgs, or a Message property that defines the complete human-readable error message.

A MessageId identifies a specific message that a message registry defines.

9.5.11.2 MessageId format

The MessageId property value shall be in the format:

<MessageRegistryPrefix>.<MajorVersion>.<MinorVersion>.<MessageKey>

where

To search the message registry for a message, the client can use the MessageId.

The message registry approach has advantages for internationalization because the message registry can be translated easily and is lightweight for implementations because large strings need not be included with the implementation.

The use of GeneralError from the Base Message Registry as a MessageId in ExtendedInfo is discouraged. If no better message exists or the ExtendedInfo array contains multiple messages, use GeneralError from the Base Message Registry only in the code property of the error object.

When an implementation uses GeneralError from the Base Message Registry in ExtendedInfo, the implementation should include a service-defined value for the Resolution property with this error to indicate how to resolve the problem.

9.6 Properties

9.6.1 Properties overview

Every property included in a Redfish response payload shall be defined in the schema for that resource. The following attributes apply to all property definitions:

This clause also contains a set of common properties across all Redfish resources. The property names in this clause shall not be used for any other purpose.

9.6.2 Resource identifier (@odata.id) property

Registry resources in a response may include an @odata.id property. All other resources and resource collections in a response shall include an @odata.id property. The value of the identifier property shall be the resource URI.

9.6.3 Resource type (@odata.type) property

All resources and resource collections in a response shall include an @odata.type type property. To support generic OData clients, all structured properties in a response should include an @odata.type type property.

The value of the type property for resources and structured properties shall be in the format:

#<ResourceType>.<Version>.<TermName>

where

An example of a resource type value is #ComputerSystem.v1_0_0.ComputerSystem, where ComputerSystem.v1_0_0 denotes the version 1.0.0 of the ComputerSystem resource type, and the specific type is ComputerSystem.

The value of the type property for resource collections shall be in the format:

#<ResourceType>.<ResourceType>

where

An example of a resource collection type value is #ComputerSystemCollection.ComputerSystemCollection for the ComputerSystemCollection resource collection.

9.6.4 Resource ETag (@odata.etag) property

ETags enable clients to conditionally retrieve or update a resource. Resources should include an @odata.etag property. For a resource, the value shall be the ETag.

9.6.5 Resource context (@odata.context) property

Responses for resources and resource collections may contain an @odata.context property that describes the source of the payload.

If the @odata.context property is present, it shall be the context URL that describes the resource, according to OData Protocol.

The context URL for a resource should be in the format:

/redfish/v1/$metadata#<ResourceType>.<ResourceType>

where

For example, the following context URL specifies that the results show a single ComputerSystem resource:

{
   "@odata.context": "/redfish/v1/$metadata#ComputerSystem.ComputerSystem",
   ...
}

The context URL for a resource may be in one of the other formats that OData Protocol specifies.

9.6.6 Id

The Id property of a resource uniquely identifies the resource within the resource collection that contains it. The value of Id shall be a string that is unique across a resource collection. Since URIs are constructed from the value of the Id property, the value shall not contain any RFC1738-defined unsafe characters.

9.6.7 Name

The Name property conveys a human-readable moniker for a resource. The data type of the Name property shall be string. The value of Name is NOT required to be unique across resource instances within a resource collection.

9.6.8 Description

The Description property conveys a human-readable description of the resource. The data type of the Description property shall be string.

9.6.9 MemberId

The MemberId property uniquely identifies an element within an array, where a reference property can reference the element. The value of MemberId shall be a string that is unique across the array.

9.6.10 Count () property

The count property defines the total number of resource, or members, that are available in a resource collection. The count property shall be named Members@odata.count and its value shall be the total number of members available in the resource collection. The $top or $skip query parameters shall not affect this count. If the number of members available in the resource collection is reduced due to filtering, such as in response to the $filter query parameter, the count should be the total number of members available in the resource collection after the filter is applied.

9.6.11 Members

The Members property of a resource collection identifies the members of the collection. The Members property is required and shall be returned in the response for any resource collection. The Members property shall be an array of JSON objects named Members. The Members property shall not be null. Empty collections shall be an empty JSON array.

The next link (Members@odata.nextLink) property value shall be an opaque URL to a resource, with the same @odata.type, which contains the next set of partial members from the original operation. The next link property shall only be present if the number of members in the resource collection is greater than the number of members returned, and if the payload does not represent the end of the requested resource collection.

The next link property may contain the $skiptoken query parameter. The $skiptoken query parameter shall contain an opaque value that allows a client to reference the next page of the collection. Clients shall not interpret or construct $skiptoken values.

The next link property value should contain control information provided by the client in query parameters that affect the size of the collection and properties returned in the response, such as the $filter and $expand query parameters. For example, if a client performs a request to /redfish/v1/Chassis/1/Sensors?$filter=Status/Health eq 'OK'&$expand=., some appropriate values include:

The Members@odata.count property value is the total number of resources available if the client enumerates all pages of the resource collection.

The Links property represents the hyperlinks associated with the resource, as defined by that resource's schema definition. All associated reference properties defined for a resource shall be nested under the links property. All directly (subordinate) referenced properties defined for a resource shall be in the root of the resource. There are some exceptions to these rules, such as where ease of expansion or deep operations is beneficial to the user.

The links property shall be named Links and contain a property for each related resource.

To navigate vendor-specific hyperlinks, the Links property shall also include an Oem property.

A reference to a single resource is a JSON object that contains a single resource identifier property. The name of this reference is the name of the relationship. The value of this reference is the URI of the referenced resource.

{
   "Links": {
      "ManagedBy": {
         "@odata.id": "/redfish/v1/Chassis/Encl1"
      }
   }
}

A reference to a set of zero or more related resources is an array of JSON objects. The name of this reference is the name of the relationship. Each element of the array is a JSON object that contains a resource identifier property with the value of the URI of the referenced resource.

{
   "Links": {
      "Contains": [{
         "@odata.id": "/redfish/v1/Chassis/1"
      }, {
         "@odata.id": "/redfish/v1/Chassis/Encl1"
      }]
   }
}

9.6.14 Actions property

The Actions property contains the actions supported by a resource.

9.6.14.1 Action representation

Each supported action is represented as a property nested under Actions. The unique name that identifies the action is used to construct the property name.

This property name shall be in the format:

#<ResourceType>.<ActionName>

where

The client may use this fragment to identify the action definition in the referenced Redfish schema document.

The property for the action is a JSON object and contains the following properties:

The OData JSON Format Specification defines the target and title properties.

To specify the list of supported values for a parameter, the service may include the @Redfish.AllowableValues annotation.

For example, the following property defines the Reset action for a ComputerSystem:

{
   "#ComputerSystem.Reset": {
      "target": "/redfish/v1/Systems/1/Actions/ComputerSystem.Reset",
      "title": "Computer System Reset",
      "ResetType@Redfish.AllowableValues": ["On", "ForceOff", "GracefulRestart",
         "GracefulShutdown", "ForceRestart", "Nmi", "ForceOn",
         "PushPowerButton"]
   },
   ...
}

Given this, the client could invoke a POST request to /redfish/v1/Systems/1/Actions/ComputerSystem.Reset with the following body:

POST /redfish/v1/Systems/1/Actions/ComputerSystem.Reset HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
   "ResetType": "On"
}

The resource may provide a separate @Redfish.ActionInfo resource to describe the parameters and values that a particular instance or implementation supports. Use the @Redfish.ActionInfo annotation to specify the ActionInfo resource, which contains a URI to the @Redfish.ActionInfo resource for the action. For details, see the Action info annotation clause.

9.6.14.2 Action responses

Response payloads for actions may contain a JSON body that is described by the schema definition for the action. See the Schema definition languages clause for the representation of these definitions. Actions that do not define a response body may provide an error response in the response payload.

Clients can discover the schema definition of the action response based on the property name of the action found in GET responses for resources containing the action. For example, if an action is named #ContosoComputerSystem.Reset, a client can find the action response definition, if there is one available, by locating the Reset action definition found in the ContosoComputerSystem schema.

9.6.15 Oem

The Oem property is used for extending standard resources with OEM extensions.

9.6.16 Status

The Status property represents the status of a resource. The Status property shall follow the definition for Status in the Resource schema.

By having a common representation of status, clients can depend on consistent semantics. The Status property is capable of indicating the current state, health of the resource, and the health of subordinate resources.

9.7 Naming conventions

The Redfish interface is intended to be easily readable and intuitive. Thus, consistency helps the consumer understand the use of a newly discovered property. While consistency is no substitute for the normative information in the Redfish Specification and Redfish schema, the following naming rules help with readability and client usage. In general, names in Redfish are designed and intended to be human-readable and convey the meaning of the name, in context, without the need to consult schema definitions or other documentation.

9.7.1 Naming rules

Standard Redfish schema and registries defined and published by DMTF as well as those created by others and republished by DMTF shall follow a set of naming conventions. These conventions are intended to ensure consistent naming and eliminate naming collisions. For schema files, the resource type is used to construct the type property and the schema file name.

Standard Redfish properties follow similar naming conventions, and should use a common definition when defined in multiple schemas across the Redfish data model. This consistency enables code re-use across resources and increases interoperability. New resource definitions should leverage existing property definitions whenever possible.

The general Redfish naming rules for resource types, registries, properties, and action parameters are as follows:

Exceptions are allowed for the following cases:

Enumeration values follow the above rules and exceptions. Additionally, enumeration values:

For properties that have units or other special meaning, append a unit identifier to the name. Examples include:

9.7.2 URI naming rules

The following rules apply to Redfish schema-defined URIs:

Starting with Redfish Specification v1.18.0, in cases where the URI segment does not follow one or more of the previous rules, the schema definition for the reference property for the URI shall specify the URI segment annotation to define the URI segment to append.

9.8 Extending standard resources

9.8.1 Extending standard resources overview

In the context of this clause, the OEM term refers to any company, manufacturer, or organization that provides or defines an extension to DMTF-published schema and functionality for Redfish. All Redfish-specified resources include an empty structured Oem property. The value of this predefined placeholder can encapsulate one or more OEM-specified object properties, which can contain OEM-specific property definitions.

9.8.2 OEM property format and content

Each property contained within the Oem property shall be an OEM-specified JSON object. The name of each object property shall uniquely identify the OEM or organization that defines the properties contained by that object. The OEM-specified object naming clause describes this naming convention.

The OEM-specified object shall include a type property if the object:

To avoid naming collisions with other OEM definitions, the type property should start with the defining organization's unique OEM identifier, including possible subdivisioning. Separator underscores (_) may be excluded from the type property and schema file name for improved readability. The Schema definition languages clause contains file name construction rules, such as handling version information. See OEM property examples for examples.

The Oem property can simultaneously hold multiple OEM-specified objects, including objects for more than one company or organization.

The definition of any other properties that are contained within the OEM-specified object, along with the functional specifications, validation, or other requirements for that content is OEM-specific and outside the scope of this specification. While there are no Redfish-specified limits on the size or complexity of the elements within an OEM-specified object, it is intended it is typically used for only a small number of simple properties that augment the Redfish resource. If a large number of objects or a large quantity of data compared to the size of the Redfish resource is to be supported, the OEM should consider creating a subordinate resource for their extensions.

9.8.3 OEM-specified object naming

The OEM-specified object properties within the Oem property are named by using a unique OEM identifier. There are two specified forms for the identifier. The identifier shall be either an ICANN-recognized domain name (including the top-level domain suffix), with all dot (.) separators replaced with underscores (_), or an IANA-assigned Enterprise Number prefixed with "EID_."

DEPRECATED: The identifier shall be either an ICANN-recognized domain name including the top-level domain suffix, or an IANA-assigned Enterprise Number prefixed with EID:.

Organizations that use .com domain names may omit the .com suffix. For example, Contoso.com would use Contoso instead of Contoso_com, but Contoso.org would use Contoso_org. The domain name portion of an OEM identifier shall be considered to be case-insensitive. That is, the text Contoso_biz, contoso_BIZ, conTOso_biZ, and so on all identify the same OEM.

The OEM identifier portion of the object name may be followed by an underscore (_) and any additional string to enable further subdivisions of OEM-specified objects as desired. For example, Contoso_xxxx or EID_412_xxxx. The form and meaning of any text that follows the trailing underscore is completely OEM-specific. OEM-specified extension suffixes may be case-sensitive, depending on the OEM. Generic client software should treat such extensions, if present, as opaque and not try to parse nor interpret the content.

There are organizations with which DMTF has a working relationship and have registered their OEM namespace directly in the specification to allow extensions of the ICANN domain name requirements above. The following organization OEM namespaces shall be considered reserved:

This suffix could be used in many ways, depending on OEM need. For example, the Contoso company may have a Research sub-organization, in which case the OEM-specified property name might be extended to Contoso_Research. Alternatively, it can identify a unique resource type for a functional area, geography, subsidiary, and so on.

The OEM identifier portion of the name typically identifies the company or organization that created and maintains the schema for the property. However, this practice is not a requirement. The identifier is only required to uniquely identify the party that is the top-level manager of a resource type to prevent collisions between OEM property definitions from different vendors or organizations. Consequently, the organization for the top of the resource type may be different from the organization that provides the definition of the OEM-specified property. For example, Contoso may allow one of their customers, such as CustomerA, to extend a Contoso product with certain CustomerA proprietary properties. In this case, although Contoso allocated the name Contoso_CustomerA, it could be CustomerA that defines the content and functionality within that resource type. In all cases, OEM identifiers should not be used except with permission or as specified by the identified company or organization.

9.8.4 OEM resource types

Companies, OEMs, and other organizations can define additional resources and link to them from an Oem property in a standard Redfish resource, preferably from the Oem property within the Links property. To avoid naming collisions with current or future standard Redfish resource types or schema files, the defining organization's unique OEM identifier, including possible subdivisioning, should be prepended to the OEM resource type name with an optional underscore (_) as separator. This unique OEM identifier should follow the same naming as defined in the OEM-specified object naming clause. The name of the OEM resource type, including the unique OEM identifier, should also be prepended to the file name of OEM schema file that specifies the OEM resource type. Separator underscores (_) may be excluded from the OEM resource type name and schema file name for improved readability. The Schema definition languages clause contains other file name construction rules, such as handling version information.

For example, OEM resource type ContosoDrive or Contoso_CustomerA_Drive would not conflict with the standard Redfish Drive resource type, or conflict with another OEM's drive-related definition.

9.8.5 OEM registries

Companies, OEMs, and other organizations can define additional registries and utilize them in message objects, privileges or for BIOS attributes. To avoid naming collisions with current or future standard Redfish message registries, the defining organization's unique OEM identifier, including possible subdivisioning, should be prepended to the registry name with an optional underscore (_) as separator. This unique OEM identifier should follow the same naming as defined in the OEM-specified object naming clause. Separator underscores (_) may be excluded from the OEM registry name for improved readability. The OEM registry name, including the unique OEM identifier, should also be used to construct the registry file name as defined in the Registry file naming clause.

For example, OEM registry name ContosoDriveEvent or Contoso_CustomerB_DriveEvent would not conflict with a possible future standard Redfish DriveEvent message registry name, or conflict with another OEM's drive-related registry name.

9.8.6 OEM URIs

To avoid URI collisions with other OEM resources and future Redfish standard resources, the URIs for OEM resources within the Redfish resource tree shall be in the form:

<BaseUri>/Oem/<OemIdentifier>/<ResourcePath>

where

For example, if Contoso defined a new ContosoAccountServiceMetrics OEM resource type to be linked through the Oem property at the /redfish/v1/AccountService URI, the OEM resource has the /redfish/v1/AccountService/Oem/Contoso/AccountServiceMetrics URI. If Contoso uses a subdivision of their OEM identifier such as Contoso_CustomerA the OEM resource has the URI /redfish/v1/AccountService/Oem/Contoso_CustomerA/AccountServiceMetrics or /redfish/v1/AccountService/Oem/ContosoCustomerA/AccountServiceMetrics.

9.8.7 OEM property examples

The following fragment shows examples of naming and the Oem property as it might appear when accessing a resource. The example shows that the OEM identifiers can be of different forms, that OEM-specified content can be simple or complex, and that the format and usage of extensions of the OEM identifier is OEM-specific.

{
    "Oem": {
        "Contoso": {
            "@odata.type": "#ContosoAnvil.v1_2_1.AnvilTypes1",
            "Slogan": "Contoso anvils never fail",
            "Disclaimer": "* Most of the time"
        },
        "Contoso_biz": {
            "@odata.type": "#ContosoBizEngine.v1_1_0.RelatedSpeed",
            "Speed": "Plaid"
        },
        "EID_412": {
            "@odata.type": "#AdatumPowerExtensions.v1_0_1.PowerInfoExt",
            "ReadingInfo": {
                "Accuracy": "5",
                "IntervalSeconds": "20"
            }
        },
        "Contoso_CustomerA": {
            "@odata.type": "#ContosoCustomerASling.v1_0_0.SlingPower",
            "AvailableTargets": [ "Rabbit", "Duck", "Runner" ],
            "LaunchPowerOptions": [ "Low", "Medium", "Eliminate" ],
            "LaunchPower": "Eliminate",
            "Target": "Rabbit"
        }
    },
    ...
}

9.8.8 OEM actions

OEM-specific actions appear in the JSON payload as properties of the Oem object, nested under an Actions property.

The name of the property that represents the action, which shall follow the form:

#<OEMSchemaName>.<Action>

where

{
   "Actions": {
      "Oem": {
         "#Contoso_ABC_ComputerSystem.Ping": {
            "target": "/redfish/v1/Systems/1/Actions/Oem/Contoso_ABC_ComputerSystem.Ping",
            "@Redfish.ActionInfo": "/redfish/v1/Systems/1/Oem/Contoso_ABC/PingActionInfo"
         },
         "#ContosoCustomerAComputerSystem.CustomPing": {
            "target": "/redfish/v1/Systems/1/Actions/Oem/ContosoCustomerAComputerSystem.CustomPing",
            "@Redfish.ActionInfo": "/redfish/v1/Systems/1/Oem/ContosoCustomerA/CustomPingPingActionInfo"
         }
      }
   },
   ...
}

The URI of the OEM action in the target property shall be in the form:

<ResourceUri>/Actions/Oem/<OEMSchemaName>.<Action>

where

The URI to the ActionInfo resource should be in the form:

<ResourceUri>/Oem/<OemIdentifier>/<Action>ActionInfo

where

9.9 Payload annotations

9.9.1 Payload annotations overview

Resources, objects within a resource, and properties may include additional annotations as properties with the name, in the format:

[<PropertyName>]@<Namespace>.<TermName>

where

Services shall limit the annotation usage to the odata, Redfish, and Message namespaces. The OData JSON Format Specification defines the odata namespace. The Redfish namespace is an alias for the RedfishExtensions.v1_0_0 namespace.

The client can get the definition of the annotation from the OData metadata document, the HTTP Link header, or may ignore the annotation entirely, but should not fail reading the resource due to unrecognized annotations, including new annotations that the Redfish namespace defines.

9.9.2 Allowable values for strings

Services may use the @Redfish.AllowableValues annotation to specify the list of allowable values for a string property in modification requests or action parameter. The values of the annotation should only include those values that are both supported by the service and currently available as valid values for the particular instance of the property or action parameter. The annotation shall contain a JSON array of strings that define the allowable values for the property or action parameter.

The following example shows that the FavoriteFruit property supports four values.

{
    "FavoriteFruit": "Kiwi",
    "FavoriteFruit@Redfish.AllowableValues": [ "Orange", "Pineapple", "Kiwi", "Starfruit" ]
}

9.9.3 Allowable patterns for string values

Services may use the @Redfish.AllowablePattern annotation to specify a pattern that describes the valid values for a string property in modification requests or action parameter. The annotation shall contain a regular expression that describes the supported pattern for the property or action parameter. If a pattern is specified in the schema definition for this property, this annotation may further restrict the allowable values, but shall not allow values which would violate the schema-defined pattern. Services shall not use the @Redfish.AllowablePattern for enumerations.

The following example shows the AssetTag property allows alphanumeric characters, colons, and dashes, and allows a string length up to 31 characters.

{
    "AssetTag": "22HOU-34566",
    "AssetTag@Redfish.AllowablePattern": "^[\\w:-]{0,31}$"
}

9.9.4 Allowable values for numbers and durations

Services may use the @Redfish.AllowableNumbers annotation to specify one or more ranges of allowable values and an optional incremental step value between valid values for a numeric property in modification requests or action parameter or duration property or action parameter. The annotation shall contain an array of strings, each specifying a range of values and an optional step value. Each element in the array shall contain a number or a duration in the format:

<min>:<max>:<step>

where

If the value does not contain : characters, the value specifies a single supported value.

If the value specifies a range and <min> is omitted, the minimum supported value shall be assumed to be a value of zero.

If the value specifies a range and :<step> is omitted, no step is defined for the supported range.

The following example shows the usage of the @Redfish.AllowableNumbers annotation for different properties. PacketSizeBytes supports a range of 1024 to 65536 in increments of 256. TemperatureThresholdCelsius supports a range of 0 to 50 with no step restrictions. ClockSpeedMHz supports 800, 1150, or 1600 to 5000 in increments of 100. TimeoutDuration supports a duration between 5 minutes and 24 hours, in 5 minute increments.

{
    "PacketSizeBytes": 2048,
    "PacketSizeBytes@Redfish.AllowableNumbers": [ "1024:65536:256" ],
    "TemperatureThresholdCelsius": 37,
    "TemperatureThresholdCelsius@Redfish.AllowableNumbers": [ ":50" ],
    "ClockSpeedMHz": 2200,
    "ClockSpeedMHz@Redfish.AllowableNumbers": [ "800", "1150", "1600:5000:100" ],
    "TimeoutDuration": "PT2H",
    "TimeoutDuration@Redfish.AllowableNumbers": [ "PT5M:PT24H:PT5M" ]
}

9.9.5 Extended information

The following clauses describe the methods of providing extended information:

9.9.5.1 Extended object information

To specify object-level status information, services may annotate a JSON object with the @Message.ExtendedInfo annotation.

{
  "@odata.id": "/redfish/v1/Managers/1/SerialInterfaces/1",
  "@odata.type": "#SerialInterface.v1_0_0.SerialInterface",
  "Name": "Managed Serial Interface 1",
  "Description": "Management for Serial Interface",
  "Status": {
    "State": "Enabled",
    "Health": "OK"
  },
  "InterfaceEnabled": true,
  "SignalType": "Rs232",
  "BitRate": "115200",
  "Parity": "None",
  "DataBits": "8",
  "StopBits": "1",
  "FlowControl": "None",
  "ConnectorType": "RJ45",
  "PinOut": "Cyclades",
  "@Message.ExtendedInfo": [{
    "MessageId": "Base.1.8.PropertyDuplicate",
    "Message": "Indicates that a duplicate property was included in the request body.",
    "RelatedProperties": [
      "/InterfaceEnabled"
    ],
    "Severity": "Warning",
    "MessageSeverity": "Warning",
    "Resolution": "Remove the duplicate property from the request body and resubmit the request if the operation failed."
  }]
}

The property contains an array of message objects.

9.9.5.2 Extended property information

Services may use @Message.ExtendedInfo, prepended with the name of the property to annotate an individual property in a JSON object with extended information:

{
   "@odata.id": "/redfish/v1/Managers/1/SerialInterfaces/1",
   "@odata.type": "#SerialInterface.v1_0_0.SerialInterface",
   "Name": "Managed Serial Interface 1",
   "Description": "Management for Serial Interface",
   "Status": {
      "State": "Enabled",
      "Health": "OK"
   },
   "InterfaceEnabled": true,
   "SignalType": "Rs232",
   "BitRate": 115200,
   "Parity": "None",
   "DataBits": 8,
   "StopBits": 1,
   "FlowControl": "None",
   "ConnectorType": "RJ45",
   "PinOut": "Cyclades",
   "PinOut@Message.ExtendedInfo": [{
      "MessageId": "Base.1.8.PropertyValueNotInList",
      "Message": "The value Contoso for the property PinOut is not in the list of acceptable values.",
      "Severity": "Warning",
      "MessageSeverity": "Warning",
      "Resolution": "Choose an enumeration list value that the implementation supports. Resubmit the request if the operation failed."
   }]
}

9.9.6 Action info annotation

The @Redfish.ActionInfo term within the action representation conveys the parameter requirements and allowable values on parameters for actions. This term contains a URI to the ActionInfo resource.

The URI to the ActionInfo resource should be in the form:

<ResourceUri>/<Action>ActionInfo

where

Example #ComputerSystem.Reset action with the @Redfish.ActionInfo annotation and resource:

{
   "Actions": {
      "#ComputerSystem.Reset": {
         "target": "/redfish/v1/Systems/1/Actions/ComputerSystem.Reset",
         "@Redfish.ActionInfo": "/redfish/v1/Systems/1/ResetActionInfo"
      }
   },
   ...
}

The ResetActionInfo resource contains a more detailed description of the parameters and the supported values. This resource follows the ActionInfo schema definition.

{
   "@odata.id": "/redfish/v1/Systems/1/ResetActionInfo",
   "@odata.type": "#ActionInfo.v1_0_0.ActionInfo",
   "Id": "ResetActionInfo",
   "Name": "Reset Action Info",
   "Parameters": [{
      "Name": "ResetType",
      "Required": true,
      "DataType": "String",
      "AllowableValues": ["On", "ForceOff", "ForceRestart", "Nmi",
         "ForceOn", "PushPowerButton"]
   }]
}

9.9.7 Settings and settings apply time annotations

See the Settings resource clause.

9.9.8 Operation apply time and operation apply time support annotations

See the Operation apply time clause.

9.9.9 Maintenance window annotation

The settings apply time and operation apply time annotations enable an operation to be performed during a maintenance window. The @Redfish.MaintenanceWindow term at the root of a resource configures the start time and duration of a maintenance window for a resource.

The following example body for the /redfish/v1/Systems/1 resource configures the maintenance window to start at 2017-05-03T23:12:37-05:00 and last for 600 seconds.

{
   "@odata.id": "/redfish/v1/Systems/1",
   "@odata.type": "#ComputerSystem.v1_5_0.ComputerSystem",
   "@Redfish.MaintenanceWindow": {
      "@odata.type": "#Settings.v1_3_3.MaintenanceWindow",
      "MaintenanceWindowStartTime": "2017-05-03T23:12:37-05:00",
      "MaintenanceWindowDurationInSeconds": 600
   },
   ...
}

9.9.10 Collection capabilities annotation

Resource collections may contain a collection capabilities annotation. The @Redfish.CollectionCapabilities term at the root of a resource collection shows what properties a client is allowed to use in a POST request for creating a resource.

The following ComputerSystemCollection example body contains the collection capabilities annotation. The UseCase property contains the ComputerSystemComposition value, and the CapabilitiesObject property contains the /redfish/v1/Systems/Capabilities value. The resource at /redfish/v1/Systems/Capabilities describes the POST request format for creating a ComputerSystem resource for compositions.

{
   "@odata.id": "/redfish/v1/Systems",
   "@odata.type": "#ComputerSystemCollection.ComputerSystemCollection",
   "Name": "Computer System Collection",
   "Members@odata.count": 0,
   "Members": [],
   "@Redfish.CollectionCapabilities": {
      "@odata.type": "#CollectionCapabilities.v1_1_0.CollectionCapabilities",
      "Capabilities": [{
         "CapabilitiesObject": {
            "@odata.id": "/redfish/v1/Systems/Capabilities"
         },
         "UseCase": "ComputerSystemComposition",
         "Links": {
            "TargetCollection": {
               "@odata.id": "/redfish/v1/Systems"
            }
         }
      }]
   }
}

The CapabilitiesObject resource follows the same schema for the resource that the resource collection contains. It contains annotations to show which properties the client can use in the POST request body. Services may omit properties marked as required for the resource referenced by the CapabilitiesObject property, but may provide annotations for the property to express POST request body requirements. Table 21 describes the CapabilitiesObject resource annotations. These annotations describe which properties are required, optional, or if other rules are associated with the properties.

Table 21 — CapabilitiesObject resource annotations
Annotation Description
<PropertyName>@Redfish.RequiredOnCreate Required in the POST request body.
<PropertyName>@Redfish.OptionalOnCreate Not required in the POST request body.
<PropertyName>@Redfish.SetOnlyOnCreate Cannot be modified after the resource is created.
<PropertyName>@Redfish.UpdatableAfterCreate Can be modified after the resource is created.
<PropertyName>@Redfish.AllowableValues Can be set to any of the listed values.
@Redfish.RequestedCountRequired Required in the POST request body for the corresponding object to indicate the number of requested object instances.

Used for composition requests.
@Redfish.ResourceBlockLimits Indicates restrictions regarding quantities of ResourceBlock resources of a given type in the POST request body.

Used for composition requests.

Example CapabilitiesObject resource:

{
   "@odata.id": "/redfish/v1/Systems/Capabilities",
   "@odata.type": "#ComputerSystem.v1_8_0.ComputerSystem",
   "Id": "Capabilities",
   "Name": "Capabilities for the system collection",
   "Name@Redfish.RequiredOnCreate": true,
   "Name@Redfish.SetOnlyOnCreate": true,
   "Description@Redfish.OptionalOnCreate": true,
   "Description@Redfish.SetOnlyOnCreate": true,
   "HostName@Redfish.OptionalOnCreate": true,
   "HostName@Redfish.UpdatableAfterCreate": true,
   "Links@Redfish.RequiredOnCreate": true,
   "Links": {
      "ResourceBlocks@Redfish.RequiredOnCreate": true,
      "ResourceBlocks@Redfish.UpdatableAfterCreate": true
   },
   "@Redfish.ResourceBlockLimits": {
      "MinCompute": 1,
      "MaxCompute": 1,
      "MaxStorage": 8
   }
}

9.9.11 Requested count and allow over-provisioning annotations

Table 22 describes the @Redfish.RequestedCount and @Redfish.AllowOverprovisioning annotations.

Clients use these annotations in composition requests to define the number of resource to allocate and to indicate whether the Redfish service can provision more resources than the client requests:

Table 22 — RequestCount and AllowOverprovisioning annotations
Annotation Description
@Redfish.RequestedCount Number of requested resources.
@Redfish.AllowOverprovisioning Boolean. If true, the service may provision more resources than the @Redfish.RequestedCount annotation requests. Default is false.

Example client request for at least four and possibly more Processor resources:

{
   "Processors": {
      "Members": [{
         "@Redfish.RequestedCount": 4,
         "@Redfish.AllowOverprovisioning": true
      }]
   },
   ...
}

9.9.12 Zone affinity annotation

The zone affinity annotation is used by clients in composition requests to indicate the components for the composition come from the specified resource zone. The @Redfish.ZoneAffinity term in the request body contains the value of the Id property of the requested resource zone.

Example client request for components to be allocated from the resource zone with the Id property containing 1:

{
   "@Redfish.ZoneAffinity": "1",
   ...
}

9.9.13 Supported certificates annotation

Resource collections of type CertificateCollection should contain a supported certificates annotation. The @Redfish.SupportedCertificates term at the root of a resource collection shows the different certificate formats allowed in the resource collection.

Example CertificateCollection that only supports PEM style certificates:

{
   "@odata.id": "/redfish/v1/Managers/BMC/NetworkProtocol/HTTPS/Certificates",
   "@odata.type": "#CertificateCollection.CertificateCollection",
   "Name": "Certificate collection",
   "Members@odata.count": 1,
   "Members": [{
      "@odata.id": "/redfish/v1/Managers/BMC/NetworkProtocol/HTTPS/Certificates/1"
   }],
   "@Redfish.SupportedCertificates": ["PEM"]
}

9.9.14 Deprecated annotation

Services may annotate properties with @Redfish.Deprecated if the schema definition has the property marked as deprecated.

Example deprecated property:

{
   "VendorID": "0xABCD",
   "VendorID@Redfish.Deprecated": "This property has been deprecated in favor of ModuleManufacturerID.",
   ...
}

9.9.15 Writable properties annotation

Services may annotate resources or objects with @Redfish.WriteableProperties to list the properties in the resource or object whose value may be changed by a modification request. This annotation shall only list those properties that are defined as read-write in schema, are implemented as read-write by the service, and can be modified given the current configuration of the resource. When present at the root of a resource, the properties listed shall include only writable properties at the root level. Writable properties within objects shall be shown in additional annotation instances within the object. For arrays of objects, if the annotation appears in only the first element of the array, the values shall apply to all elements of the array.

Note: While "writable" is the preferred spelling of the word, the name of the annotation in payloads contains "writeable".

Example writable properties annotations that show writable properties at the root level and an array of objects within each element:

{
    "UserName": "John Smith",
    "Alias": "Major Goofball",
    "FavoriteFruit": "Pineapple",
    "@Redfish.WriteableProperties": [ "Alias", "FavoriteFruit" ],
    "Pets": [
        {
            "PetName": "Waffles",
            "PetType": "Donkey",
            "Friendly": true,
            "@Redfish.WriteableProperties": [ "Friendly" ]
        },
        {
            "PetName": "Fluffy",
            "PetType": "Snake",
            "Friendly": false
        }
    ]
}

9.10 Settings resource

A settings resource represents the future intended state of a resource. Some resources have properties that can be updated and the updates take place immediately. However, some properties need to be updated at a future point in time, such as after a system reset. While the active resource represents the current state, the settings resource represents the future intended state.

For resources that support a future intended state, the response shall contain a property with the @Redfish.Settings payload annotation. When a settings annotation is used, the following conditions shall apply:

The settings resource shall contain the properties that are updated at a future point in time. For resources that support a future intended state, Table 23 describes the behavior of supported properties in the resource and settings resource that a service should support.

Table 23 — Active resource and settings resource property behavior
Property Active resource behavior Settings resource behavior
Read-only, required. Returned in the resource response to a GET request. Returned in the settings resource response to a GET request.
Read-only, not required. Returned in the resource response to a GET request. Not returned in the settings resource response to a GET request.
Writable, updates immediately, but not at a future point in time. Active value returned in the resource response to a GET request.
Modification requests change the active value immediately.		 Not returned in the settings resource response to a GET request.
Modification requests are rejected.
Writable, updates immediately or at a future point in time. Active value returned in the resource response to a GET request.
Modification requests change the active value immediately.		 Future value returned in the settings resource response to a GET request if a future value is pending, otherwise not returned.
Modification requests change the future value.
Writable, updates at a future point in time, but not immediately. Active value returned in the resource response to a GET request.
Modification requests are rejected.		 Future value returned in the settings resource response to a GET request.
Modification requests change the future value.

The @Redfish.Settings annotation includes several properties that help clients monitor when the service has consumed the active resource and determine the success or failure of applying the values.

The following active resource example body supports a settings resource. A client can use the SettingsObject property to locate the URI of the settings resource.

{
   "@Redfish.Settings": {
      "@odata.type": "#Settings.v1_3_3.Settings",
      "SettingsObject": {
         "@odata.id": "/redfish/v1/Systems/1/Bios/SD"
      },
      "Time": "2017-05-03T23:12:37-05:00",
      "ETag": "\"A89B031B62\"",
      "Messages": [{
         "MessageId": "Base.1.8.PropertyNotWritable",
         "RelatedProperties": ["/Attributes/ProcTurboMode"]
      }]
   },
   ...
}

If a service enables a client to indicate when to apply settings:

In the following example request, the client indicates that the settings resource values are applied on reset during the specified maintenance window:

{
   "@Redfish.SettingsApplyTime": {
      "@odata.type": "#Settings.v1_3_3.PreferredApplyTime",
      "ApplyTime": "InMaintenanceWindowOnReset",
      "MaintenanceWindowStartTime": "2017-05-03T23:12:37-05:00",
      "MaintenanceWindowDurationInSeconds": 600
   },
   ...
}

9.11 Special resource situations

9.11.1 Overview

Resources need to exhibit common semantic behavior whenever possible. This can be difficult in some situations discussed in this clause.

9.11.2 Absent resources

Resources may be absent or their state unknown at the time a client requests information about that resource. For resources that represent removable or optional components, absence provides useful information to clients because it indicates a capability, such as an empty PCIe slot, DIMM socket, or drive bay, that would not be apparent if the resource simply did not exist.

This also applies to resources that represent a limited number of items or unconfigured capabilities within an implementation, but this usage should be applied sparingly and should not apply to resources limited in quantity due to arbitrary limits. For example, an implementation that limits SoftwareInventory to a maximum of 20 items should not populate 18 absent resources when only two items are present.

For resources that provide useful data in an absent state and where the URI is expected to remain constant, such as when a DIMM is removed from a memory socket, the resource should exist and should return the Absent value for the State property in the Status object.

In this circumstance, any required properties that have no known value shall contain null. Properties whose support is based on the configuration choice or the type of component installed, and therefore unknown while in the absent state, should not be returned. Likewise, subordinate resources for an absent resource should not be populated until their support can be determined. For example, the Power and Thermal resources under a Chassis resource should not exist for an absent Chassis.

Client software should be aware that when absent resources are later populated, the updated resource may represent a different configuration or physical item, and previous data, including read-only properties, obtained from that resource may be invalid. For example, the Memory resource shows details about a single DIMM socket and the installed DIMM. When that DIMM is removed, the Memory resource remains as an absent resource to indicate the empty DIMM socket. Later, a new DIMM is installed in that socket, and the Memory resource represents data about this new DIMM, which could have completely different characteristics.

9.11.3 Transiently unavailable resources

Resources may be transiently unavailable due to the underlying configuration of a device. For example, a network adapter might not support management protocols while in a low power state.

For resources that provide useful data in a transiently unavailable state and where the URI is expected to remain constant, such as when a network adapter is powered off and is unable to respond to management requests, the resource should exist and should return the StandbyOffline value for the State property in the Status object.

In this circumstance, any required properties that have no known value shall contain null. Properties whose support is based on the configuration choice or the type of component installed, and therefore unknown while in the transiently unavailable state, should not be returned. Likewise, subordinate resources for a transiently unavailable resource should not be populated until their support can be determined.

9.12 Registries

Registry resources assist the client in interpreting Redfish resources beyond the Redfish schema definitions. To get more information about a resource, event, message, or other item, use an identifier to search registries. This information can include other properties, property restrictions, and the like. Registries are themselves resources.

Table 24 describes the types of registries that Redfish supports:

Table 24 — Registries
Registry Description See
Attribute Determines the semantics of each property within the Attributes object in a BIOS or BIOS settings resource. Because BIOS information can vary from platform to platform, Redfish cannot define a fixed schema for these values. Attribute registries should be assigned unique identifiers to allow users to match a given registry with compatible products.

This registry contains both property descriptions and other information, such as data type, allowable values, and user menu information.
Message Constructs a message from a MessageId and other message information to present to an end user. The messages in these registries appear in both eventing and error responses to operations.

This registry is the most common type of registry.
Privilege Maps the resources in a Redfish service to the privileges that can complete specified operations against those resources.

A client can use this information to:
  • Determine which roles should have specific privileges.
  • Map accounts to those roles so that the accounts can complete operations on Redfish resources.
 Privilege model

9.13 Schema annotations

9.13.1 Schema annotations overview

The schema definitions of the data model use schema annotations to provide additional documentation for developers. This clause describes the different types of schema annotations that the Redfish data model uses. For information about how each of the annotations are implemented in their respective schema languages, see the Schema definition languages clause.

9.13.2 Description annotation

The description annotation can be applied to any type, property, action, or parameter to provide a description of Redfish schema elements suitable for end users or user interface help text.

All schemas that are published or republished by DMTF's Redfish Forum shall include a description annotation on the following schema definitions:

9.13.3 Long description annotation

The long description annotation can be applied to any type, property, action, or parameter to provide a formal, normative specification of the schema element.

When the long descriptions in the Redfish schema contain normative language, the service shall be required to conform with the statement.

All schemas that are published or republished by DMTF's Redfish Forum shall include a long description annotation on the following schema definitions:

9.13.4 Resource capabilities annotation

The resource capabilities annotation can be applied to resources and resource collections to express the different type of HTTP operations a client can invoke on the given resource or resource collection.

All schemas that are published or republished by DMTF's Redfish Forum for resources and resource collections shall include resource capabilities annotations.

9.13.5 Resource URI patterns annotation

The resource URI patterns annotation expresses the valid URI patterns for a resource or resource collection.

The strings for the URI patterns may use { and } characters to express parameters within a given URI pattern. Items between the { and } characters are treated as identifiers within the URI for given instances of a Redfish resource. Clients interpret this as a string to be replaced to access a given resource. A URI pattern may contain multiple identifier terms to support multiple levels of nested resource collections. The identifier term in the URI pattern shall match the Id string property for the corresponding resource, or the MemberId string property for the corresponding object within a resource. The process for forming the strings that are concatenated to form the URI pattern are in the URI naming rules clause.

The following string is an example URI pattern that describes a ManagerAccount resource: /redfish/v1/AccountService/Accounts/{ManagerAccountId}

Using the previous example, {ManagerAccountId} is replaced by the Id property of the corresponding ManagerAccount resource. If the Id property for a ManagerAccount resource is John, the full URI for that resource is /redfish/v1/AccountService/Accounts/John.

The URI patterns are constructed based on the formation of the resource tree. When constructing the URI pattern for a subordinate resource, the URI pattern for the current resource is used and appended. For example, the RoleCollection resource is subordinate to AccountService. Because the URI pattern for AccountService is /redfish/v1/AccountService, the URI pattern for the RoleCollection resource is /redfish/v1/AccountService/Roles.

In some cases, the subordinate resource is found inside of a structured property of a resource. In these cases, the name of the structured property appears in the URI pattern for the subordinate resource. For example, the CertificateCollection resource is subordinate to the ManagerNetworkProtocol resource from the HTTPS property. Because the URI pattern for ManagerNetworkProtocol is /redfish/v1/Managers/{ManagerId}/NetworkProtocol, the URI pattern for the CertificateCollection resource is /redfish/v1/Managers/{ManagerId}/NetworkProtocol/HTTPS/Certificates.

All schemas that are published or republished by DMTF's Redfish Forum for resources and resource collections shall be annotated with the resource URI patterns annotation.

All Redfish resources and Redfish resource collections implemented by a service shall match the URI pattern described by the resource URI patterns annotation for their given definition.

9.13.6 Additional properties annotation

The additional properties annotation specifies whether a type can contain additional properties outside of those defined in the schema. Types that do not support additional properties shall not contain properties beyond those described in the schema.

9.13.7 Permissions annotation

The permissions annotation specifies whether a client can modify the value of a property, or if the property is read-only.

A service can implement a modifiable property as read-only.

The value of a write-only property, such as Password, cannot be read, and shall be null in responses.

All schemas that are published or republished by DMTF's Redfish Forum shall include a permissions annotation for all properties that are not structured properties.

9.13.8 Required annotation

The required annotation specifies whether a service needs to support a property. Required properties shall be annotated with the required annotation. All other properties are optional.

9.13.9 Required on create annotation

The required on create annotation specifies that a property is required to be provided by the client on creation of the resource. Properties not annotated with the required on create annotation are not required to be provided by the client on a create operation.

9.13.10 Units of measure annotation

In addition to following the naming rules, properties representing units of measure shall be annotated with the units of measure annotation to specify the units of measurement for the property.

The value of the annotation shall be a string that contains the case-sensitive "(c/s)" symbol of the unit of measure as listed in the Unified Code for Units of Measure (UCUM), unless the symbolic representation does not reflect common usage. If the unit in common usage is not available in UCUM, curly braces should wrap the value, such as {value}, to follow UCUM parsing rules. For example, RPM is commonly used to report fan speeds in revolutions-per-minute, and the preferred representation in UCUM is {rev}/min, but the value {RPM} is acceptable. For units with prefixes, the case-sensitive (c/s) symbol for the prefix as listed in UCUM should be prepended to the unit symbol. For example, the mebibyte (1024^2 bytes), which has the UCUM Mi prefix and By symbol, would use MiBy as the value for the annotation. For values that also include rate information, such as megabits per second, the rate unit's symbol should be appended and use a slash (/) character as a separator. For example, Mbit/s.

DEPRECATED: Previous versions of this specification recommended RPM as a commonly used unit of measure for certain properties. New recommendations are provided that follow UCUM parsing rules.

9.13.11 Expanded resource annotation

The expanded resource annotation can be applied to a reference property to specify that the default behavior for the service is to include the contents of the related resource or resource collection in responses. This behavior follows the same semantics of the expand query parameter with a level of 1.

Reference properties annotated with this term shall be expanded by the service, even if not requested by the client. A service may page resource collections.

9.13.12 Owning entity annotation

The owning entity annotation can be applied to a schema to specify the name of the entity responsible for development, publication, and maintenance of a given schema.

9.13.13 Deprecated annotation

The deprecated annotation specifies if a property, enumeration, or other schema element has been deprecated. Schema elements marked as deprecated contain a schema version that shows when the element was deprecated, as well as text that specifies the favored approach.

The deprecated annotation also specifies if resource URI patterns have been deprecated. Deprecated resource URI patterns shall also be included in the resource URI patterns annotation.

Existing and new implementations may use deprecated schema elements or URIs, but they should move to the favored approach. Deprecated schema elements may be implemented to achieve backwards compatibility. Deprecated schema elements may be removed from the next major version of the schema.

9.13.14 URI segment annotation

The URI segment annotation can be applied to a reference property to specify the segment appended to the URI of the resource when constructing the URI of a subordinate resource if the segment differs from the property name. For more information, see the URI naming rules clause.

9.13.15 URI annotation

The URI annotation can be applied to a string property to indicate the property value contains a URI. This is used for cases where a service provides a hyperlink to a URI outside of the Redfish model or where a reference property is not used to avoid undesireable expansion.

9.14 Versioning

As stated previously, a resource can be an individual entity or a resource collection, which acts as a container for a set of resources.

A resource collection does not contain any version information because it defines a single Members property, and the overall collection definition never grows over time.

A resource has both unversioned and versioned definitions.

References from other resources use the unversioned definition of a resource to ensure no version dependencies exist between the definitions. The unversioned definition of a resource contains no property information about the resource.

The versioned definition of a resource contains a set of properties, actions, and other definitions associated with the resource. The version of a resource follows the format:

v<X>.<Y>.<Z>

where

9.15 Localization

The creation of separate localized copies of Redfish schemas and registries is allowed and encouraged. Localized schema and registry files may be submitted to DMTF for republication in the Redfish schema repository.

Property names, parameter names, and enumeration values in the JSON response payload are never localized but translated copies of those names may be provided as additional annotations in the localized schema for use by client applications. A separate file for each localized schema or registry shall be provided for each supported language. The English-language versions of Redfish schemas and registries shall be the normative versions, and alterations of meaning due to translation in localized versions of schemas and registries shall be forbidden.

Schemas and registries in non-English languages shall use the appropriate schema annotations to identify their language. Descriptive property, parameter, and enumeration text not translated into the specified language shall be removed from localized versions. This removal enables software and tools to combine normative and localized copies, especially for minor schema version differences.

10 File naming and publication

For consistency in publication and to enable programmatic access, all Redfish-related files shall follow a set of rules to construct the name of each file. The Schema definition languages clause describes the file name construction rules, while the following clauses describe the construction rules for other file types.

10.1 Registry file naming

Redfish message registry files, privilege registry files, and BIOS attribute registry files shall use the registry name to construct the file name, in this format:

<RegistryName>.<MajorVersion>.<MinorVersion>.<Errata>.json

For example, the file name of the Base Message Registry v1.0.2 is Base.1.0.2.json.

The registry name should be unique to avoid conflict with other registry files. The clause OEM registries describes registry name to use for OEM registry files.

10.2 Profile file naming

The document that describes a profile follows the Redfish schema file naming conventions. The file name format for profiles shall be:

<ProfileName>.v<MajorVersion>_<MinorVersion>_<Errata>.json

For example, the file name of the BasicServer profile v1.2.0 is BasicServer.v1_2_0.json. The file name shall include the profile name and version, which matches those property values within the document.

10.3 Dictionary file naming

The binary file describing a Redfish Device Enablement dictionary follows the Redfish schema file naming conventions for the schema definition language that the dictionary is converted from. Because a single dictionary file contains all minor revisions of the schema, only the major version appears in the file name. The file names for Dictionaries shall be formatted as:

<DictionaryName>_v<MajorVersion>.dict

For example, the file name of the Chassis dictionary v1.2.0 is Chassis_v1.dict.

10.4 Localized file naming

Localized schemas and registries shall follow the same file naming conventions as the English language versions. When multiple localized copies are present in a repository and which have the same file name, files in languages other than English shall be organized into sub-folders named to match the ISO 639-1 language code for those files. English language files may be duplicated in an en sub-folder for consistency.

10.5 DMTF Redfish file repository

All Redfish schemas, registries, dictionaries, and profiles published or republished by DMTF's Redfish Forum are available from the DMTF website for download. Programs may use the following durable URLs to access the repository. Programs incorporating remote repository access should implement a local cache to reduce latency, program requirements for Internet access and undue traffic burden on DMTF's website.

Organizations creating Redfish-related files such as OEM schemas, Redfish interoperability profiles, or message registries are encouraged to use the form at https://redfish.dmtf.org/redfish/portal to submit those files to DMTF for republication in DMTF's Redfish file repository.

Table 25 describes how files are organized on the site:

Table 25 — Redfish file repository
URL Folder contents
redfish.dmtf.org/schemas Current (most recent minor or errata) release of each schema file in CSDL, JSON Schema, and/or OpenAPI formats.
redfish.dmtf.org/schemas/v1 Durable URL for programmatic access to all v1.xx schema files. Every v1.xx minor or errata release of each schema file in CSDL, JSON Schema, OpenAPI formats.
redfish.dmtf.org/schemas/v1/{code} Durable URL for programmatic access to localized v1.xx schema files. Localized schemas are organized in sub-folders using the two-character ISO 639-1 language code as the {code} segment.
redfish.dmtf.org/schemas/archive Sub-folders contain schema files specific to a particular version release.
redfish.dmtf.org/registries Current (most recent minor or errata) release of each registry file.
redfish.dmtf.org/registries/v1 Durable URL for programmatic access to all v1.xx registry files. Every v1.xx minor or errata release of each registry file.
redfish.dmtf.org/registries/v1/{code} Durable URL for programmatic access to localized v1.xx registry files. Localized schemas are organized in sub-folders using the two-character ISO 639-1 language code as the {code} segment.
redfish.dmtf.org/registries/archive Sub-folders contain registry files specific to a particular version release.
redfish.dmtf.org/profiles Current release of each Redfish interoperability profile (.json) file and associated documentation.
redfish.dmtf.org/profiles/v1 Durable URL for programmatic access to all v1.xx Redfish interoperability profile (.json) files.
redfish.dmtf.org/profiles/archive Sub-folders contain profile files specific to a particular profile version or release.
redfish.dmtf.org/dictionaries Durable URL for programmatic access to all v1.xx Redfish Device Enablement dictionary files.
redfish.dmtf.org/dictionaries/v1 Durable URL for programmatic access to all v1.xx Redfish Device Enablement dictionary files.
redfish.dmtf.org/dictionaries/archive Sub-folders contain dictionary files specific to a particular version release.

11 Schema definition languages

Individual resources and their dependent types and actions are defined within a Redfish schema document. This clause describes how these documents are constructed in the following formats:

11.1 OData Common Schema Definition Language

11.1.1 OData Common Schema Definition Language overview

OData Common Schema Definition Language (CSDL) is an XML schema format defined by the OData CSDL Specification. The following clause describes how Redfish uses CSDL to describe resources and resource collections.

11.1.2 File naming conventions for CSDL

Redfish CSDL schema files shall be named using the resource type name for the schema, followed by _v and the major version of the schema. Because a single CSDL schema file contains all minor revisions of the schema, only the major version appears in the file name. The file name shall be formatted as:

<ResourceType>_v<MajorVersion>.xml

For example, version 1.3.0 of the Chassis schema is Chassis_v1.xml.

11.1.3 Core CSDL files

Table 26 describes the core CSDL files:

Table 26 — Core CSDL files
File Description
RedfishError_v1.xml Payload definition of the Redfish error response.
RedfishExtensions_v1.xml All definitions for Redfish types and annotations.
Resource_v1.xml All base definitions for resources, resource collections, and common properties, such as Status.

11.1.4 CSDL format

The outer element of the OData schema representation document shall be the Edmx element, and shall have a Version attribute with a value of 4.0.

<edmx:Edmx xmlns:edmx="http://docs.oasis-open.org/odata/ns/edmx" Version="4.0">
  <!-- edmx:Reference and edmx:DataServices elements go here -->
</edmx:Edmx>

The Referencing other CSDL files and CSDL data services clauses describe the items that are found within the Edmx element.

11.1.4.1 Referencing other CSDL files

CSDL files may use Reference tags to reference types defined in other CSDL documents.

The Reference element uses the Uri attribute to specify a CSDL file. The Reference element also contains one or more Include tags that specify the Namespace attribute containing the types to be referenced, along with an optional Alias attribute for that namespace.

Type definitions generally reference the OData and Redfish namespaces for common type annotation terms. Redfish CSDL files shall contain the Alias attribute on the following namespaces:

<edmx:Reference
  Uri="http://docs.oasis-open.org/odata/odata/v4.0/cs01/vocabularies/Org.OData.Core.V1.xml">
  <edmx:Include Namespace="Org.OData.Core.V1" Alias="OData"/>
</edmx:Reference>
<edmx:Reference
  Uri="http://docs.oasis-open.org/odata/odata/v4.0/os/vocabularies/Org.OData.Measures.V1.xml">
 <edmx:Include Namespace="Org.OData.Measures.V1" Alias="Measures"/>
</edmx:Reference>
<edmx:Reference Uri="http://redfish.dmtf.org/schemas/v1/RedfishExtensions_v1.xml">
  <edmx:Include Namespace="RedfishExtensions.v1_0_0" Alias="Redfish"/>
  <edmx:Include Namespace="Validation.v1_0_0" Alias="Validation"/>
</edmx:Reference>
<edmx:Reference Uri="http://redfish.dmtf.org/schemas/v1/Resource_v1.xml">
  <edmx:Include Namespace="Resource"/>
  <edmx:Include Namespace="Resource.v1_0_0"/>
</edmx:Reference>

11.1.4.2 CSDL data services

Define structures, enumerations, and other definitions in CSDL within a namespace. Use a Schema tag to define the schema and use the Namespace attribute to declare the name of the namespace.

Redfish uses namespaces to differentiate different versions of the schema. CSDL enables structures to inherit from other structures, which enables newer namespaces to define only the changes. The Elements of CSDL namespaces clause describes this behavior.

Namespaces containing unversioned resource and resource collection definitions shall use the resource type to name the namespace, in this format:

<ResourceType>

For example, the unversioned namespace of the Chassis resource is Chassis.

Namespaces containing versioned resource definitions shall use the resource type to name the namespace, in this format:

<ResourceType>.v<MajorVersion>_<MinorVersion>_<Errata>

For example, the version 1.3.0 namespace of the Chassis resource is Chassis.v1_3_0.

The Schema element is a child of the DataServices element, which is a child of the Edmx element:

<edmx:DataServices>
  <Schema xmlns="http://docs.oasis-open.org/odata/ns/edm" Namespace="MyTypes.v1_0_0">
    <!-- Type definitions for version 1.0.0 of MyTypes go here -->
  </Schema>
  <Schema xmlns="http://docs.oasis-open.org/odata/ns/edm" Namespace="MyTypes.v1_1_0">
    <!-- Type definitions for version 1.1.0 of MyTypes go here -->
  </Schema>
</edmx:DataServices>

11.1.5 Elements of CSDL namespaces

The following clauses describe the definitions within each namespace:

11.1.5.1 Qualified names

Many definitions in CSDL use references to qualified names. CSDL defines this as a string in the form:

<Namespace>.<TypeName>

where

For example, if a reference is made to MyType.v1_0_0.MyDefinition, the definition can be found in the MyType.v1_0_0 namespace with an element named MyDefinition.

11.1.5.2 Entity type and complex type elements

Use the EntityType and ComplexType tags to define the entity type and complex type elements, respectively. These elements define a JSON structure and their set of properties by defining property elements and navigation property elements within the EntityType or ComplexType tags.

All entity types and complex types shall contain a Name attribute, which specifies the name of the definition.

Entity types and complex types may contain a BaseType attribute, which specifies a qualified name. When the BaseType attribute is present, all definitions of the referenced BaseType are available to the entity type or complex type being defined.

All resources and resource collections shall be defined with the entity type element. Resources inherit from Resource.v1_0_0.Resource, and resource collections inherit from Resource.v1_0_0.ResourceCollection.

All structured properties shall be defined with the complex type element or entity type element. Most structured properties are defined with the complex type element. Some use the entity type element that inherits from Resource.v1_0_0.ReferenceableMember. The entity type element enables references to be made by using the Navigation Property element, whereas the complex type element does not allow for this usage.

Example entity type and complex type element:

<EntityType Name="TypeA" BaseType="Resource.v1_0_0.Resource">
  <Annotation Term="OData.Description" String="Entity description."/>
  <Annotation Term="OData.LongDescription" String="Entity normative description."/>
  <!-- Property and navigation property definitions go here -->
</EntityType>

<ComplexType Name="PropertyTypeA">
  <Annotation Term="OData.Description" String="Structured property description."/>
  <Annotation Term="OData.LongDescription" String=Structured property normative description."/>
  <!-- Property and navigation property definitions go here -->
</ComplexType>

11.1.5.3 Action element

Use the Action tag to define the action element. This element defines an action that can be performed on a resource.

All Redfish actions shall be defined with the action element. All action elements shall contain a Name attribute, which specifies the name of the action. The action shall be represented in payloads as the qualified name of the action, preceded by #.

In Redfish, all action elements shall contain the IsBound attribute that is always set to true, which indicates that the action appears as a member of a structured type.

The action element shall contain one or more Parameter tags that specify the Name and Type of each parameter.

Because all action elements in Redfish use the IsBound="true" term, the first parameter is called the binding parameter and specifies the structured type to which the action belongs. All Redfish actions shall contain a binding parameter. The binding parameter shall be one of the following complex type elements:

The remaining Parameter elements shall describe additional parameters to be passed to the action. The term Nullable="false" in a parameter shall indicate the parameter is required in the action request body.

<Schema xmlns="http://docs.oasis-open.org/odata/ns/edm" Namespace="MyType">
  <Action Name="MyAction" IsBound="true">
    <Parameter Name="Thing" Type="MyType.Actions"/>
    <Parameter Name="Parameter1" Type="Edm.Boolean"/>
    <Parameter Name="Parameter2" Type="Edm.String" Nullable="false"/>
  </Action>

  <ComplexType Name="Actions">
    ...
  </ComplexType>

  ...

</Schema>

Some action parameters may specify a type that is defined by an entity type element. In these cases, the parameter in the request is a reference object to a resource within the service.

11.1.5.4 Action element for OEM actions

OEM-specific actions shall be defined by using the action element with the binding parameter set to the OemActions complex type for the resource. For example, the following definition defines the OEM #Contoso.Ping action for a ComputerSystem.

<Schema xmlns="http://docs.oasis-open.org/odata/ns/edm" Namespace="Contoso">
  <Action Name="Ping" IsBound="true">
    <Parameter Name="ComputerSystem" Type="ComputerSystem.v1_0_0.OemActions"/>
  </Action>
</Schema>

11.1.5.5 Action with a response body

A response body for an action shall be defined using the ReturnType tag within an action element. For example, the following definition defines the GenerateTicket action with a response that contains the definition specified by GenerateTicketResponse.

<Action Name="GenerateTicket" IsBound="true">
  <Parameter Name="ExampleResource" Type="ExampleResource.v1_0_0.Actions"/>
    ...
  <ReturnType Type="ExampleResource.v1_0_0.GenerateTicketResponse" Nullable="false"/>
</Action>

<ComplexType Name="GenerateTicketResponse">
  <Annotation Term="OData.AdditionalProperties" Bool="false"/>
  <Annotation Term="OData.Description" String="The response body for GenerateTicket."/>
  <Property Name="TicketId" Type="Edm.String" Nullable="false">
    <Annotation Term="OData.Permissions" EnumMember="OData.Permission/Read"/>
    <Annotation Term="OData.Description" String="The ticket identifier."/>
    <Annotation Term="Redfish.Required"/>
  </Property>
</ComplexType>

Using the above example, the following payload is an example response for the GenerateTicket action.

{
   "TicketId": "40478281bd0f6b9e7131db6c4f673438"
}

11.1.5.6 Property element

Properties of resources, resource collections, and structured properties are defined using the property element. The Property tag defines a property element inside entity type and complex type elements.

All property elements shall contain a Name attribute, which specifies the name of the property.

All property elements shall contain a Type attribute specifies the data type. The Type attribute shall be one of the following names or types:

Table 27 describes the primitive data types:

Table 27 — Primitive data types
Type Meaning
Edm.Boolean True or False.
Edm.DateTimeOffset Date-time string.
Edm.Decimal Number, optionally containing a decimal point.
Edm.Double Number, optionally containing a decimal point and optionally containing an exponent.
Edm.Duration Duration string.
Edm.Guid GUID/UUID string.
Edm.Int64 Signed 64-bit integer.
Edm.String UTF-8 string.

Property elements may specify a Nullable attribute. If the attribute is false, the property shall not contain null. If the attribute is true or absent, the property may contain null.

Example property element:

<Property Name="Property1" Type="Edm.String" Nullable="false">
  <Annotation Term="OData.Description" String="Property1 description."/>
  <Annotation Term="OData.LongDescription" String="Property1 normative description."/>
  <Annotation Term="OData.Permissions" EnumMember="OData.Permission/Read"/>
  <Annotation Term="Redfish.Required"/>
  <Annotation Term="Measures.Unit" String="Watts"/>
</Property>

Reference properties of resources, resource collections, and structured properties are defined using the navigation property element. The NavigationProperty tag defines a navigation property element inside entity type and complex type elements.

All navigation property elements shall contain a Name attribute, which specifies the name of the property.

All navigation property elements shall contain a Type attribute specifies the data type. The Type attribute shall be a qualified name that references an entity type element. The Collection term with the qualified name shall indicate the property is an array.

Navigation property elements may specify a Nullable attribute. If the attribute is false, the property shall not contain null. If the attribute is true or absent, the property may contain null.

Unless the reference property is to be expanded, all navigation properties in Redfish shall contain the OData.AutoExpandReferences annotation element to show that the reference is always available.

Example navigation property element:

<NavigationProperty Name="RelatedType" Type="MyTypes.TypeB">
  <Annotation Term="OData.Description" String="RelatedType description."/>
  <Annotation Term="OData.LongDescription" String="RelatedType normative description."/>
  <Annotation Term="OData.AutoExpandReferences"/>
</NavigationProperty>

11.1.5.8 Enum type element

Use the EnumType tag to define the enum type element. This element defines a set of enumeration values, which may be applied to one or more properties.

All enum type elements shall contain a Name attribute, which specifies the name of the set of enumeration values.

Enum type elements shall contain Member tags that define the members of the enumeration. The Member tags shall contain a Name attribute that specifies the string value of the member name.

<EnumType Name="EnumTypeA">
  <Annotation Term="OData.Description" String="EnumTypeA type description."/>
  <Annotation Term="OData.LongDescription" String="EnumTypeA type normative description."/>
  <Member Name="MemberA">
    <Annotation Term="OData.Description" String="The description of MemberA"/>
  </Member>
  <Member Name="MemberB">
    <Annotation Term="OData.Description" String="The description of MemberB"/>
  </Member>
</EnumType>

11.1.5.9 Annotation element

Annotations in CSDL are expressed using the Annotation tag. Any schema element in CSDL may contain annotations.

The following examples show how each Redfish schema annotation is expressed in CSDL.

Example description annotation:

<Annotation Term="OData.Description" String="The console color."/>

Example long description annotation:

<Annotation Term="OData.LongDescription"
  String="This property shall contain the console color."/>

Example additional properties annotation:

<Annotation Term="OData.AdditionalProperties"/>

Example permissions annotation (read-only):

<Annotation Term="OData.Permissions" EnumMember="OData.Permission/Read"/>

Example permissions annotation (read/write):

<Annotation Term="OData.Permissions" EnumMember="OData.Permission/ReadWrite"/>

Example permissions annotation (write-only, null in responses):

<Annotation Term="OData.Permissions" EnumMember="OData.Permission/Write"/>

Example required annotation:

<Annotation Term="Redfish.Required"/>

Example required on create annotation:

<Annotation Term="Redfish.RequiredOnCreate"/>

Example units of measure annotation:

<Annotation Term="Measures.Unit" String="MiBy"/>

Example expanded resource annotation:

<Annotation Term="OData.AutoExpand"/>

Example insert capabilities annotation (showing POST is not allowed):

<Annotation Term="Capabilities.InsertRestrictions">
  <Record>
    <PropertyValue Property="Insertable" Bool="false"/>
  </Record>
</Annotation>

Example update capabilities annotation (showing PATCH and PUT are allowed):

<Annotation Term="Capabilities.UpdateRestrictions">
  <Record>
    <PropertyValue Property="Updatable" Bool="true"/>
    <Annotation Term="OData.Description" String="The desired speed can be changed."/>
  </Record>
</Annotation>

Example delete capabilities annotation (showing DELETE is allowed):

<Annotation Term="Capabilities.DeleteRestrictions">
  <Record>
    <PropertyValue Property="Deletable" Bool="true"/>
    <Annotation Term="OData.Description" String="The resource can be deleted."/>
  </Record>
</Annotation>

Example resource URI patterns annotation:

<Annotation Term="Redfish.Uris">
  <Collection>
    <String>/redfish/v1/AccountService/Accounts/{ManagerAccountId}</String>
  </Collection>
</Annotation>

Example URI segment annotation:

<NavigationProperty Name="Tasks" Type="TaskService.TaskService" Nullable="false">
  <Annotation Term="OData.Permissions" EnumMember="OData.Permission/Read"/>
  <Annotation Term="OData.Description" String="The link to the task service."/>
  <Annotation Term="OData.AutoExpandReferences"/>
  <Annotation Term="Redfish.URISegment" String="TaskService"/>
</NavigationProperty>

Example owning entity annotation:

<Annotation Term="Redfish.OwningEntity" String="DMTF"/>

Example deprecated annotation:

<Annotation Term="Redfish.Revisions">
  <Collection>
    <Record>
      <PropertyValue Property="Kind" EnumMember="Redfish.RevisionKind/Deprecated"/>
      <PropertyValue Property="Version" String="v1_3_0"/>
      <PropertyValue Property="Description"
        String="This property has been deprecated in favor of ModuleManufacturerID."/>
    </Record>
  </Collection>
</Annotation>

Example deprecated annotation applied to resource URI patterns:

<Annotation Term="Redfish.DeprecatedUris">
  <Collection>
    <String>/redfish/v1/Chassis/{ChassisId}/Thermal</String>
  </Collection>
</Annotation>

Example URI annotation:

<Annotation Term="OData.IsURL"/>

11.2 JSON Schema

11.2.1 JSON Schema overview

The JSON Schema Specification defines a JSON format for describing JSON payloads. The following clause describes how Redfish uses JSON Schema to describe resources and resource collections.

11.2.2 File naming conventions for JSON Schema

Each Redfish JSON Schema file represents a single resource type.

Versioned Redfish JSON Schema files shall use the resource type to name the file, in this format:

<ResourceType>.v<MajorVersion>_<MinorVersion>_<Errata>.json

For example, version 1.3.0 of the Chassis schema is Chassis.v1_3_0.json.

Unversioned Redfish JSON Schema files shall use the resource type to name the file, in this format:

<ResourceType>.json

For example, the unversioned definition of the Chassis schema is Chassis.json.

11.2.3 Core JSON Schema files

Table 28 describes the core JSON Schema files:

Table 28 — Core JSON Schema files
File Description
odata-v4.json Definitions for common OData properties.
redfish-error.v1_0_0.json and its subsequent versions Payload definition of the Redfish error response.
redfish-schema-v1.json Extensions to the JSON Schema that define Redfish JSON Schema files.
Resource.json and its subsequent versions All base definitions for resources, resource collections, and common properties, such as Status.

11.2.4 JSON Schema format

Each JSON Schema file shall contain a JSON object to describe resources, resource collections, and other definitions for the data model.

Table 29 describes the JSON object, which contains the following terms:

Table 29 — JSON Schema format
Term Description
$id Reference to the URI where the schema file is published.
$ref For a schema file that describes a resource or resource collection, the reference to the structural definition of the resource or resource collection.
$schema URI to the Redfish schema extensions for JSON Schema. The value should be http://redfish.dmtf.org/schemas/v1/redfish-schema-v1.json.
copyright Copyright statement for the organization producing the JSON Schema.
definitions Structures, enumerations, and other definitions defined by the schema.
title For a schema file that describes a resource or resource collection, the matching type identifier for the resource or resource collection.

11.2.5 JSON Schema definitions body

This clause describes the types of definitions found in the definitions term of a Redfish JSON Schema file.

11.2.5.1 Resource definitions in JSON Schema

To satisfy versioning requirements, the JSON Schema representation of each resource shall have one unversioned schema file and a set of versioned schema files.

The unversioned definition of a resource shall contain an anyOf statement. This statement shall consist of an array of $ref terms, which point to the following definitions:

The unversioned definition of a resource shall contain the uris term to express the allowable URIs for the resource, and the deletable, insertable, and updatable terms to express the capabilities of the resource.

The following example shows an unversioned resource definition in JSON Schema:

{
  "ComputerSystem": {
    "anyOf": [{
      "$ref": "http://redfish.dmtf.org/schemas/v1/odata.v4_0_3.json#/definitions/idRef"
    }, {
      "$ref": "http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_0_0.json#/definitions/ComputerSystem"
    }, {
      "$ref": "http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_0_1.json#/definitions/ComputerSystem"
    }, {
      "$ref": "http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_6_0.json#/definitions/ComputerSystem"
    }],
    "deletable": true,
    "description": "The ComputerSystem schema represents a general purpose machine or system.",
    "insertable": false,
    "longDescription": "This resource shall represent resources that represent a computing system.",
    "updatable": true,
    "uris": [
      "/redfish/v1/Systems/{ComputerSystemId}"
    ]
  },
  ...
}

The versioned definition of a resource shall contain the property definitions for the given version of the resource.

11.2.5.2 Enumerations in JSON Schema

Table 30 describes the terms that constitute definitions for enumerations:

Table 30 — JSON Schema enumerations
Term Description
enum String array that contains the possible enumeration values.
enumDescriptions Object that contains the descriptions for each of the enumerations as name-value pairs.
enumLongDescriptions Object that contains the long descriptions for each of the enumerations as name-value pairs.
enumDeprecated Object that contains the deprecation guidance for each of the enumerations as name-value pairs.
enumVersionDeprecated Object that contains the deprecation version information for each of the enumerations as name-value pairs.
type Because all enumerations in Redfish are strings, the type term always has the string value.

The following example shows an enumeration definition in JSON Schema:

{
  "Flavors": {
    "enum": ["Lit", "Blinking", "Off"],
    "enumDescriptions": {
      "Blinking": "The Indicator LED is blinking.",
      "Lit": "The Indicator LED is lit.",
      "Off": "The Indicator LED is off."
    },
    "enumLongDescriptions": {
      "Blinking": "This value shall represent the Indicator LED is in a blinking state where the LED is being turned on and off in repetition.",
      "Lit": "This value shall represent the Indicator LED is in a solid on state.",
      "Off": "This value shall represent the Indicator LED is in a solid off state."
    },
    "type": "string"
  },
  ...
}

11.2.5.3 Actions in JSON Schema

Versioned definitions of resources shall contain a definition called Actions. This definition is a container with a set of terms that point to the different actions supported by the resource. The names of standard actions shall be in the form:

#<ResourceType>.<ActionName>

Example Actions definition:

{
   "Actions": {
      "additionalProperties": false,
      "description": "The available actions for this resource.",
      "longDescription": "This type shall contain the available actions for this resource.",
      "properties": {
         "#ComputerSystem.Reset": {
            "$ref": "#/definitions/Reset"
         }
      },
      "type": "object"
   },
   ...
}

Another definition within the same schema file shall describe the action itself. This definition shall contain a term called parameters to describe the client request body. It also shall contain property definitions for the target and title properties shown in response payloads for the resource.

The following example shows a definition of an action:

{
   "Reset": {
      "additionalProperties": false,
      "description": "This action resets the system.",
      "longDescription": "This action shall perform a reset of the ComputerSystem.",
      "parameters": {
         "ResetType": {
            "$ref": "http://redfish.dmtf.org/schemas/v1/Resource.json#/definitions/ResetType",
            "description": "The type of reset to be performed.",
            "longDescription": "This parameter shall define the type of reset to be performed."
         }
      },
      "properties": {
         "target": {
            "description": "Link to invoke action",
            "format": "uri",
            "type": "string"
         },
         "title": {
            "description": "Friendly action name",
            "type": "string"
         }
      },
      "type": "object"
   },
   ...
}

Action parameters may specify a type that is a resource definition. In these cases, the parameter in the request shall contain a reference object to a resource within the service.

11.2.5.4 OEM actions in JSON Schema

OEM-specific actions shall be defined by using an action definition in an appropriately named JSON Schema file. For example, the following definition defines the OEM #ContosoNetworkDevice.Ping action, assuming it's found in the versioned ContosoNetworkDevice JSON Schema file, such as ContosoNetworkDevice.v1_0_0.json.

{
   "Ping": {
      "additionalProperties": false,
      "parameters": {},
      "properties": {
         "target": {
            "description": "Link to invoke action",
            "format": "uri",
            "type": "string"
         },
         "title": {
            "description": "Friendly action name",
            "type": "string"
         }
      },
      "type": "object"
   },
   ...
}

11.2.5.5 Action with a response body

A response body for an action shall be defined using the actionResponse term within the action definition. For example, the following definition defines the GenerateTicket action with a response that contains the definition specified by #/definitions/GenerateTicketResponse.

{
   "GenerateTicket": {
      "actionResponse": {
         "$ref": "#/definitions/GenerateTicketResponse"
      },
      "parameters": {}
   },
   "GenerateTicketResponse": {
      "additionalProperties": false,
      "description": "The response body for GenerateTicket.",
      "properties": {
         "TicketId": {
            "description": "The ticket identifier.",
            "readonly": true,
            "type": "string"
         }
      },
      "required": ["TicketId"],
      "type": "object"
   }
}

In the previous example, the following payload is an example response for the GenerateTicket action.

{
   "TicketId": "40478281bd0f6b9e7131db6c4f673438"
}

11.2.6 JSON Schema terms

Table 31 describes the JSON Schema terms that Redfish uses to provide schema annotations for Redfish JSON Schema:

Table 31 — JSON Schema terms
JSON Schema term Related Redfish schema annotation
description
enumDescriptions		 Description
longDescription
enumLongDescriptions		 Long description
additionalProperties Additional properties
readonly
writeOnly		 Permissions
required Required
requiredOnCreate Required on create
units Units of measure
autoExpand Expanded resource
deletable
insertable
updatable		 Resource capabilities
uris Resource URI patterns
uriSegment URI segment
owningEntity Owning entity
deprecated
versionDeprecated
urisDeprecated		 Deprecated

11.3 OpenAPI

11.3.1 OpenAPI overview

The OpenAPI Specification defines a format for describing JSON payloads and the set of URIs a client can access on a service. The following clause describes how Redfish uses OpenAPI to describe resources and resource collections.

11.3.2 File naming conventions for OpenAPI schema

Each Redfish OpenAPI file represents a single resource type.

Versioned Redfish OpenAPI files shall be named using the resource type name for the schema, following the format:

<ResourceType>.v<MajorVersion>_<MinorVersion>_<Errata>.yaml

For example, version 1.3.0 of the Chassis schema is Chassis.v1_3_0.yaml.

Unversioned Redfish OpenAPI files shall use the resource type name to name the file, in this format:

<ResourceType>.yaml

For example, the unversioned definition of the Chassis schema is Chassis.yaml.

11.3.3 Core OpenAPI schema files

Table 32 describes the core OpenAPI schema files:

Table 32 — Core OpenAPI schema files
File Description
odata-v4.yaml Definitions for common OData properties.
openapi.yaml URI paths and their respective payload structures.
Resource.yaml and its subsequent versions All base definitions for resources, resource collections, and common properties, such as Status.

11.3.4 openapi.yaml

The openapi.yaml file is the starting point for clients to understand the construct of the service.

Table 33 describes the terms that the openapi.yaml file contains:

Table 33 — openapi.yaml terms
Term Description
components Global definitions. For Redfish, contains the format of the Redfish error response.
info Structure consisting of information about what the openapi.yaml is describing, such as the author of the file and any contact information.
openapi Version of OpenAPI the document follows.
paths URIs supported by the document, with possible methods, response bodies, and request bodies.

The service shall return the openapi.yaml file, if present in the Redfish service, as a YAML document by using either the application/yaml or application/vnd.oai.openapi MIME types. The service may append ;charset=utf-8 to the MIME type. Note that while the application/yaml type is in common use today, the application/vnd.oai.openapi type was recently defined and approved specifically to support OpenAPI. Implementations should use caution when selecting the MIME type as this specification may change in the future to reflect adoption of the OpenAPI-defined MIME type.

The paths term shall contain an array of the possible URIs. Each URI shall contain methods supported by the URI. Each method shall contain the possible response bodies and request bodies.

Example paths entry for a resource:

/redfish/v1/Systems/{ComputerSystemId}:
  get:
    parameters:
    - description: The value of the Id property of the ComputerSystem resource
      in: path
      name: ComputerSystemId
      required: true
      schema:
        type: string
    responses:
      '200':
        content:
          application/json:
            schema:
              $ref: http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_6_0.yaml#/components/schemas/ComputerSystem
        description: The response contains a representation of the ComputerSystem
          resource
      default:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RedfishError'
        description: Error condition

Example paths entry for an action:

/redfish/v1/Systems/{ComputerSystemId}/Actions/ComputerSystem.Reset:
  post:
    parameters:
    - description: The value of the Id property of the ComputerSystem resource
      in: path
      name: ComputerSystemId
      required: true
        type: string
    requestBody:
      content:
        application/json:
          schema:
            $ref: http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_6_0.yaml#/components/schemas/ResetRequestBody
      required: true
    responses:
      '200':
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RedfishError'
        description: The response contains the results of the Reset action
      '202':
        content:
          application/json:
            schema:
              $ref: http://redfish.dmtf.org/schemas/v1/Task.v1_4_0.yaml#/components/schemas/Task
        description: Accepted; a task has been generated
      '204':
        description: Success, but no response data
      default:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RedfishError'
        description: Error condition

11.3.5 OpenAPI file format

With the exception of openapi.yaml, each OpenAPI file shall contain a YAML object to describe resources, resource collections, or other definitions for the data model. Table 34 describes the terms that the YAML object contains:

Table 34 — YAML object terms
Term Description
components Structures, enumerations, and other definitions defined by the schema.
x-copyright Copyright statement for the organization producing the OpenAPI file.
title For a schema file that describes a resource or resource collection, the matching type identifier for the resource or resource collection.

11.3.6 OpenAPI components body

This clause describes the types of definitions that can be found in the components term of a Redfish OpenAPI file.

11.3.6.1 Resource definitions in OpenAPI

To satisfy versioning requirements, the OpenAPI representation of each resource shall have one unversioned schema file and a set of versioned schema files.

The unversioned definition of a resource shall contain an anyOf statement. This statement shall consist of an array of $ref terms, which point to the following definitions:

Example unversioned resource definition in OpenAPI:

ComputerSystem:
  anyOf:
  - $ref: http://redfish.dmtf.org/schemas/v1/odata.v4_0_3.yaml#/components/schemas/idRef
  - $ref: http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_0_0.yaml#/components/schemas/ComputerSystem
  - $ref: http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_0_1.yaml#/components/schemas/ComputerSystem
  - $ref: http://redfish.dmtf.org/schemas/v1/ComputerSystem.v1_6_0.yaml#/components/schemas/ComputerSystem
  description: The ComputerSystem schema represents a general purpose machine
    or system.
  x-longDescription: This resource shall be used to represent resources that represent
    a computing system.

The versioned definition of a resource shall contain the property definitions for the given version of the resource.

11.3.6.2 Enumerations in OpenAPI

Table 35 describes the terms in OpenAPI enumerations:

Table 35 — OpenAPI enumerations
Term Description
enum String array that contains the possible enumeration values.
type Because all enumerations in Redfish are strings, the type term always has the value string.
x-enumDescriptions Object that contains the descriptions for each of the enumerations as name-value pairs.
x-enumLongDescriptions Object that contains the long descriptions for each enumeration as a name-value pair.
x-enumDeprecated Object that contains the deprecation guidance for each of the enumerations as name-value pairs.
x-enumVersionDeprecated Object that contains the deprecation version information for each of the enumerations as name-value pairs.

Example enumeration definition in OpenAPI:

IndicatorLED:
  enum:
  - Lit
  - Blinking
  - 'Off'
  type: string
  x-enumDescriptions:
    Blinking: The Indicator LED is blinking.
    Lit: The Indicator LED is lit.
    'Off': The Indicator LED is off.
  x-enumLongDescriptions:
    Blinking: This value shall represent the Indicator LED is in a blinking state
      where the LED is being turned on and off in repetition.
    Lit: This value shall represent the Indicator LED is in a solid on state.
    'Off': This value shall represent the Indicator LED is in a solid off state.

11.3.6.3 Actions in OpenAPI

Versioned definitions of resources shall contain a definition called Actions. This definition is a container with a set of terms that point to the different actions supported by the resource. The names of standard actions shall be in the form:

#<ResourceType>.<ActionName>

Example Actions definition:

Actions:
  additionalProperties: false
  description: The available actions for this resource.
  properties:
    '#ComputerSystem.Reset':
      $ref: '#/components/schemas/Reset'
  type: object
  x-longDescription: This type shall contain the available actions for this resource.

Another definition within the same schema file shall describe the action itself. This definition shall contain property definitions for the target and title properties shown in response payloads for the resource.

The following example shows a definition of an action:

Reset:
  additionalProperties: false
  description: This action resets the system.
  properties:
    target:
      description: Link to invoke action
      format: uri
      type: string
    title:
      description: Friendly action name
      type: string
  type: object
  x-longDescription: This action shall reset the ComputerSystem.

The parameters for the action shall be defined in another definition with RequestBody appended to the name of the action. This gets mapped from the openapi.yaml file for expressing the POST method for the URI of the action.

The following example shows a definition of parameters of an action:

ResetRequestBody:
  additionalProperties: false
  description: This action resets the system.
  properties:
    ResetType:
      $ref: http://redfish.dmtf.org/schemas/v1/Resource.yaml#/components/schemas/ResetType
      description: The reset type.
      x-longDescription: This parameter shall define the type of reset to perform.
  type: object
  x-longDescription: This action shall reset the ComputerSystem.

11.3.6.4 OEM actions in OpenAPI

OEM-specific actions shall be defined by using an action definition in an appropriately named OpenAPI file. For example, the following definition defines the OEM #ContosoNetworkDevice.Ping action, assuming it's found in the versioned ContosoNetworkDevice OpenAPI file with a name, such as ContosoNetworkDevice.v1_0_0.yaml.

Ping:
  additionalProperties: false
  properties:
    target:
      description: Link to invoke action
      format: uri
      type: string
    title:
      description: Friendly action name
      type: string
  type: object
PingRequestBody:
  additionalProperties: false
  properties: {}
  type: object

11.3.7 OpenAPI terms used by Redfish

Table 36 describes the OpenAPI terms that Redfish uses to provide schema annotations for Redfish OpenAPI files:

Table 36 — OpenAPI terms used by Redfish
OpenAPI term Related Redfish schema annotation
description
x-enumDescriptions		 Description
x-longDescription
x-enumLongDescriptions		 Long description
additionalProperties Additional properties
readOnly
writeOnly		 Permissions
required Required
x-requiredOnCreate Required on create
x-units Units of measure
x-autoExpand Expanded resource
x-uriSegment URI segment
x-owningEntity Owning entity
deprecated
x-deprecatedReason
x-versionDeprecated		 Deprecated

11.4 Schema modification rules

Schema referenced from the implementation may vary from the canonical definitions of those schema defined by the Redfish schema or other entities, provided they adhere to the following rules. Clients should take this into consideration when attempting operations on the resources defined by schema.

Services that collect and surface schema files from managed devices, such as when aggregating other services, shall handle cases where different devices contain different modifications to the same schema file. This specification does not require a specific method to resolve these conflicts, but some options include:

12 Service details

12.1 Eventing

12.1.1 Eventing overview

This clause describes how to use the REST-based mechanism to subscribe to and receive event messages.

Note: For security implications of eventing, see the Security details clause.

The Redfish service requires a client or administrator to create subscriptions to receive events.

To create a subscription, use one of these methods:

12.1.2 POST to subscription collection

To locate the event service, the client traverses the Redfish service interface. The event service is located in the service root, as described in the ServiceRoot schema.

After the client discovers the service, they perform an HTTP POST on the resource collection URI for Subscriptions in the event service to subscribe to events. For the subscription body syntax, see the Redfish EventDestination schema. This request includes:

If the subscription request succeeds, the service shall return:

If the subscription request succeeds, the service should return:

After a subscription is registered with the service, clients begin receiving events. Clients do not receive events retroactively. The service does not retain historical events.

Services shall:

Services shall not:

Services may:

To unsubscribe from the events associated with this subscription, the client or administrator shall perform an HTTP DELETE request to the subscription's resource URI.

Subsequent requests to subscription resources that have been terminated respond with the HTTP 404 Not Found status code.

Some configurable properties define the behavior for all event subscriptions. For details, see the Redfish EventService schema.

DEPRECATED: Previous versions of the specification required services to reject subscription requests if RegistryPrefixes and MessageIds were in the same request.

12.1.3 Open an SSE connection

A service may support the ServerSentEventUri property in the EventService resource. If a client performs a GET request on the URI that the ServerSentEventUri contains, an SSE connection opens for the client. For details about this method, see the server-sent events Event service clause.

12.1.4 EventType-based eventing

DEPRECATED: EventType-based eventing is deprecated in the Redfish schema in favor of using RegistryPrefix and ResourceType.

DEPRECATED

Table 37 describes the types of events that Redfish generates:

Table 37 — EventType-based eventing
Event Occurs when Description
Life cycle Resources are created, modified, or destroyed.

Usually indicates that the resource and, optionally, its properties have changed.		 Not every modification of a resource results in an event. This behavior is similar to when ETags are changed and implementations might not send an event for every resource change.

For example, if an event is sent for every Ethernet packet that is received or each time that a sensor changes one degree, more events than fit in a scalable interface are generated.
Alert An event of some significance happens.

Depending on the resource, may be generated directly or indirectly.		 Usually adopts a message registry approach similar to extended error handling in that a MessageId is included.

An example of an alert event is, a chassis is opened, a button is pushed, a cable is unplugged, or a threshold exceeded.

These events usually do not correspond well to life cycle-type events. Therefore, alerts have their own category.
Metric report The telemetry service generates or updates a metric report. Generated as specified by the MetricReportDefinition resources found subordinate to the telemetry service. Can occur periodically, on demand, or when changes are detected in the metric properties.

For details, see the Redfish MetricReportDefinition schema.

END DEPRECATED

12.1.5 Subscribing to events

Table 38 describes the properties that a subscriber provides to subscribe to events and filter received messages:

Table 38 — Subscription properties
Property Description
MessageIds An array of allowable values for MessageId in an event.

An event is sent to the subscriber if the MessageId of the event is contained in this array or is found in a message registry referenced by RegistryPrefixes.

To not perform inclusive filtering based upon MessageId, provide an empty array.

The contents of the array should not include the major or minor version of the message registry. For example, instead of Resource.1.2.ResourceCreated, use Resource.ResourceCreated.
RegistryPrefixes An array of standard or OEM message registries containing the allowable values for MessageId in an event.

An event is sent to the subscriber if the MessageId is found in a message registry in this array or is contained in MessageIds.

To not perform inclusive filtering based upon the message registry of the MessageId, provide an empty array.

The contents of the array does not include the registry version. For example, instead of Base.1.5.0, use Base.
ResourceTypes An array of standard or OEM resource types.

An event is sent to the subscriber if the OriginOfCondition resource type matches one of the ResourceTypes values.

To not perform filtering based upon the resource type of the OriginOfCondition, provide an empty array.

The contents of the array does not include the schema version. For example, instead of Task.v1_2_0.Task, use Task.
OriginResources An array of URIs to resources.

An event is sent to the subscriber if the OriginOfCondition property matches one of the URIs listed in OriginResources. To include subordinate resources regardless of depth, set the SubordinateResources property to true.

To not perform filtering based upon the URI of the OriginOfCondition, provide an empty array.
ExcludeMessageIds An array of disallowed values for MessageId in an event.

An event is not sent to the subscriber if the MessageId of the event is contained in this array.

To not perform exclusive filtering based upon MessageId, provide an empty array.

The contents of the array should not include the major or minor version of the message registry. For example, instead of Resource.1.2.ResourceCreated, use Resource.ResourceCreated.
ExcludeRegistryPrefixes An array of standard or OEM message registries containing the disallowed values for MessageId in an event.

An event is not sent to the subscriber if the MessageId is found in a message registry in this array.

To not perform exclusive filtering based upon the message registry of the MessageId, provide an empty array.

The contents of the array does not include the registry version. For example, instead of Base.1.5.0, use Base.
EventFormatType The format that can be sent by using the EventFormatTypes property in the event service.

Represents the format of the payload sent to the event destination.

If the subscriber omits this value, the payload corresponds to the Event schema.

Clients can read the EventService resource of a service in order to determine what control properties in the previous table are supported.

12.1.6 Event formats

Table 39 describes the event formats:

Table 39 — Event formats
Event format Description
Metric report message objects Used when the telemetry service generates a new or updates an existing metric report. Metric report message objects sent to the specified client endpoint shall contain the properties, as described in the Redfish MetricReport schema.
Event message objects Used for all other types of events. Event message objects POSTed to the specified client endpoint shall contain the properties as described in the Redfish Event schema. Supports a message registry. In a message registry approach, a message registry lists the MessageIds in a well-known format. These MessageIds are terse in nature and thus they are much smaller than actual messages, making them suitable for embedded environments.

The registry also contains a message. The message itself can have arguments and default values for severity and recommended actions. The MessageId property follows the format defined in the MessageId format clause

Event messages may also have an EventGroupId property, which lets clients know that different messages may be from the same event. For instance, if a LAN cable is disconnected, they may get a specific message from one registry about the LAN cable being disconnected, another message from a general registry about the resource changing, perhaps a message about resource state change, and maybe more. For the client to determine whether these have the same root cause, these messages have the same value for the EventGroupId property.

12.1.7 OEM extensions

OEMs can extend both messages and message registries. Any individual message, per the MessageRegistry schema definition, define OEM sections. Thus, if OEMs wish to provide additional information or properties, use the OEM section.

OEMs shall not supply additional message arguments beyond those in a standard message registry. OEMs may substitute their own message registry for the standard registry to provide the OEM section within the registry but shall not change the standard values, such as messages, in such registries.

12.2 Asynchronous operations

Services that support asynchronous operations implement the TaskService and Task resources.

The TaskService resource describes the service that handles task operations. It contains a resource collection of zero or more Task resources. Each Task resource describes a long-running operation that is spawned when a request takes longer than a few seconds, such as when a service is instantiated.

The Task schema defines task structure, including the start time, end time, task state, task status, and zero or more task-associated messages.

Each task has a number of possible states, which are defined in the Task schema as the values for the TaskState property.

When a client issues a request that results in a long-running operation, the service returns the HTTP 202 Accepted status code and a Location header that contains a task monitor URI and, optionally, the Retry-After header that defines the amount of time that the client should wait before querying the status of the operation. The 202 Accepted response should include a response body. If a response body is provided, it shall contain a representation of the Task resource that represents the state of the operation.

The task monitor is an opaque, service-generated URI provided to the client that initiated the request. To query the status of an operation and determine when the operation has been completed and whether it succeeded, the client performs a GET request on the task monitor URI from the Location header of the response from the initial HTTP operation. The client should not include the application/http MIME type in the Accept header.

The task monitor URI should be in the format:

/redfish/v1/TaskService/TaskMonitors/<TaskMonitorId>

where

As long as the operation is in process, the service shall return the HTTP 202 Accepted status code when the client performs a GET request on the task monitor URI.

If a service supports cancellation of a task, the Allow header shall contain DELETE for the task monitor. To cancel the operation, the client may perform a DELETE request on the task monitor URI. The service determines when to delete the associated Task resource. The client may also perform a DELETE request on the Task resource to cancel the operation. Deleting the Task resource may invalidate the associated task monitor. A subsequent GET request on the task monitor URI returns either the HTTP 410 Gone or 404 Not Found status code.

In the unlikely event that a DELETE of the task monitor or Task resource returns the HTTP 202 Accepted status code, an additional task shall not be started and instead the client may monitor the existing Task resource for the status of the cancellation request. When the task finally completes cancellation, operations on the task monitor URI and Task resources shall return the HTTP 404 Not Found status code.

After the operation has been completed, the service shall update the TaskState in the Task resource with the appropriate value. In addition, the task monitor shall return:

If the initial operation fails, the response body shall contain an error response.

If the operation has been completed and the service has already deleted the task, the service may return the HTTP 410 Gone or 404 Not Found status code. This situation can occur if the client waits too long to read the task monitor.

To continue to get status information, the client can use the resource identifier from the 202 Accepted response to directly query the Task resource.

12.3 Resource tree stability

The resource tree, which is defined as the set of URIs and array elements within the implementation, should be consistent on a single service across device resets or power cycles, and should withstand a reasonable amount of configuration change, such as adding an adapter to a server.

The resource tree on one service might not be consistent across instances of devices. The client should traverse the data model and discover resources to interact with them.

Some resources might remain very stable from system to system, such as manager network settings. However, the architecture does not guarantee this stability.

12.4 Discovery

12.4.1 Discovery overview

Automatic discovery of managed devices supporting Redfish may be accomplished by using the Simple Service Discovery Protocol (SSDP). This protocol enables network-efficient discovery without resorting to ping-sweeps, router table searches, or restrictive DNS naming schemes. Use of SSDP is optional and, if implemented, shall enable the user to disable the protocol through the ManagerNetworkProtocol resource.

The objective of discovery is for client software to locate managed devices that conform to the Redfish Specification. Therefore, the primary SSDP functionality is incorporated in the M-SEARCH query. Redfish also follows the SSDP extensions and naming that UPnP uses, where applicable, so that systems that conform to the Redfish Specification can also implement UPnP without conflict.

12.4.2 UPnP compatibility

For compatibility with general-purpose SSDP client software, primarily UPnP, the service should use UDP port 1900 for all SSDP traffic. In addition, the Time-to-Live (TTL) hop count setting for SSDP multicast messages should default to 2.

12.4.3 USN format

The UUID in the USN field of the service shall equal the UUID property in the service root. If multiple or redundant managers exist, the UUID of the service shall remain static regardless of redundancy failover. The unique ID shall be in the canonical UUID format, followed by ::dmtf-org.

12.4.4 M-SEARCH response

The Redfish service Search Target (ST) is defined as:

urn:dmtf-org:service:redfish-rest:1

The managed device shall respond to M-SEARCH queries for Search Target (ST) of the Redfish service, as well as ssdp:all. For UPnP compatibility, the managed device should respond to M-SEARCH queries for Search Target (ST) of upnp:rootdevice.

The URN provided in the ST header in the reply shall use the redfish-rest: service name followed by the major version of the Redfish Specification. If the minor version of the Redfish Specification to which the service conforms is a non-zero value, that minor version shall be appended with and preceded by a colon (:).

For example, a service that conforms to a Redfish Specification v1.4 would reply with a redfish-rest:1:4 service.

The managed device shall provide clients with the AL header that points to the Redfish service root URL.

For UPnP compatibility, the managed device should provide clients with the Location header that points to the UPnP XML descriptor.

The response to an M-SEARCH multicast or unicast query shall use the following format:

HTTP/1.1 200 OK
CACHE-CONTROL:max-age=<MaxAgeSeconds>
ST:urn:dmtf-org:service:redfish-rest:1
USN:uuid:<ServiceUUID>::urn:dmtf-org:service:redfish-rest:1
AL:<ServiceRootURI>
EXT:

where

A service may provide additional headers for UPnP compatibility.

12.4.5 Notify, alive, and shutdown messages

Redfish devices may implement the additional UPnP-defined SSDP messages to announce their availability to software. If implemented, services shall allow the end user to disable the traffic separately from the M-SEARCH response functionality. This capability enables users to use the discovery functionality with minimal amounts of generated network traffic.

12.5 Server-sent events

12.5.1 General

Server-sent events (SSE), defined by the Web Hypertext Application Technology Working Group (WHATWG), enables a client to open a connection with a web service. The web service can continuously push data to the client, as needed.

Successful resource responses for SSE shall:

Unsuccessful resource responses for SSE shall:

A service may occasionally send a comment within a stream to keep the connection alive. Services shall separate events with blank lines. Blank lines should be sent as part of the end of an event, otherwise dispatch may be delayed in conforming consumers.

The following clauses describe how Redfish uses SSE in different Redfish data model contexts. For details about SSE, see the HTML5 Specification.

12.5.2 Event service

A service's implementation of the EventService resource may contain the ServerSentEventUri property. If a client performs a GET request on the URI specified by the ServerSentEventUri property, the service shall keep the connection open and conform to the HTML5 Specification until the client closes the socket. Service-generated events shall be sent to the client by using the open connection.

When a client opens an SSE stream for the event service, the service shall create an EventDestination resource in the Subscriptions collection for the event service to represent the connection. The Context property in the EventDestination resource shall be a service-generated opaque string.

The service shall delete the corresponding EventDestination resource when the connection is closed. The service shall close the connection if the corresponding EventDestination resource is deleted.

The service shall use the id field in the SSE stream to uniquely identify a payload in the SSE stream. The value of the id field is determined by the service. A service should accept the Last-Event-ID header from the client to allow a client to restart the event stream in case the connection is interrupted.

The service shall use the data field in the SSE stream based on the payload format. The SSE streams have these formats:

To reduce the amount of data returned to the client, the service should support the $filter query parameter in the URI for the SSE stream.

Note: The $filter syntax shall follow the format in the $filter query parameter clause.

The service should support these properties as filter criteria:

12.5.2.1 Event message SSE stream

The service shall use the data field in the SSE stream to include the JSON representation of the Event object.

The following example payload shows a stream that contains a single event with the id field set to 1, and a data field that contains a single Event object.

id: 1
data:{
data:    "@odata.type": "#Event.v1_6_0.Event",
data:    "Id": "1",
data:    "Name": "Event Array",
data:    "Context": "ABCDEFGH",
data:    "Events": [
data:        {
data:            "MemberId": "1",
data:            "EventType": "Alert",
data:            "EventId": "1",
data:            "Severity": "Warning",
data:            "MessageSeverity": "Warning",
data:            "EventTimestamp": "2017-11-23T17:17:42-0600",
data:            "Message": "The LAN has been disconnected",
data:            "MessageId": "Alert.1.0.LanDisconnect",
data:            "MessageArgs": [
data:                "EthernetInterface 1",
data:                "/redfish/v1/Systems/1"
data:            ],
data:            "OriginOfCondition": {
data:                "@odata.id": "/redfish/v1/Systems/1/EthernetInterfaces/1"
data:            },
data:            "Context": "ABCDEFGH"
data:        }
data:    ]
data:}

12.5.2.2 Metric report SSE stream

The service shall use the data field in the SSE stream to include the JSON representation of the MetricReport object.

The following example payload shows a stream that contains a metric report with the id field set to 127, and the data field containing the metric report object.

id: 127
data:{
data:    "@odata.id": "/redfish/v1/TelemetryService/MetricReports/AvgPlatformPowerUsage",
data:    "@odata.type": "#MetricReport.v1_3_0.MetricReport",
data:    "Id": "AvgPlatformPowerUsage",
data:    "Name": "Average Platform Power Usage metric report",
data:    "MetricReportDefinition": {
data:        "@odata.id": "/redfish/v1/TelemetryService/MetricReportDefinitions/AvgPlatformPowerUsage"
data:    },
data:    "MetricValues": [
data:        {
data:            "MetricId": "AverageConsumedWatts",
data:            "MetricValue": "100",
data:            "Timestamp": "2016-11-08T12:25:00-05:00",
data:            "MetricProperty": "/redfish/v1/Chassis/Tray_1/Power#/0/PowerConsumedWatts"
data:        },
data:        {
data:            "MetricId": "AverageConsumedWatts",
data:            "MetricValue": "94",
data:            "Timestamp": "2016-11-08T13:25:00-05:00",
data:            "MetricProperty": "/redfish/v1/Chassis/Tray_1/Power#/0/PowerConsumedWatts"
data:        },
data:        {
data:            "MetricId": "AverageConsumedWatts",
data:            "MetricValue": "100",
data:            "Timestamp": "2016-11-08T14:25:00-05:00",
data:            "MetricProperty": "/redfish/v1/Chassis/Tray_1/Power#/0/PowerConsumedWatts"
data:        }
data:    ]
data:}

12.6 Update service

12.6.1 Overview

This clause covers the mechanism for software updates by using the update service.

12.6.2 Software update types

Clients can use these methods to update software through the update service:

12.6.2.1 Simple updates

A service can support the SimpleUpdate action within the UpdateService resource. A client can perform a POST request on the action target URI to initiate a pull-based update, as defined by the UpdateService schema. After a successful POST, the service should return the HTTP 202 Accepted status code with the Location header set to the URI of a task monitor. Clients can use this task to monitor the progress and results of the update, which includes the progress of image transfer to the service, as described in the Asynchronous operations clause.

12.6.2.2 Multipart HTTP push updates

A service may support the MultipartHttpPushUri property within the UpdateService resource. A client can perform an HTTP or HTTPS POST request on the URI specified by this property to initiate a push-based update.

Table 40 describes the requirements of a multipart/form-data request body for an HTTP push software update. If the Required column contains Yes, the service shall require the client to provide the request body part. Services shall not require clients to provide request body parts that are not marked as required. If the Required column contains No, the service may allow the client to provide the request body part.

Table 40 — Multipart HTTP push updates
Request body part HTTP headers Header value and parameters Required Description
Action parameters JSON part Content-Disposition form-data; name="UpdateParameters" Yes JSON-formatted part for passing the action parameters. The value of the name field shall be "UpdateParameters". The format of the JSON shall follow the definition of the UpdateParameters object in the UpdateService schema.
Services may allow the inclusion of the @Redfish.OperationApplyTime property in the request body. See Operation apply time.
Content-Type application/json;charset=utf-8 or application/json Yes Media type format and character set of this request part.
Update file binary part Content-Disposition form-data; name="UpdateFile"; filename=string Yes Binary file to use for this software update. The value of the name field shall be "UpdateFile". The value of the filename field should reflect the name of the file as loaded by the client.
Content-Type application/octet-stream Yes Media type format of the binary update file.
OEM specific parts Content-Disposition form-data; name="OemXXXX" No Optional OEM part. The value of the name field shall start with "Oem. Content-Type is optional and depends on the OEM part type.

This example shows a multipart/form-data request to push an update image:

POST /redfish/v1/UpdateService/upload HTTP/1.1
Host: <host-path>
Content-Type: multipart/form-data; boundary=---------------------------d74496d66958873e
Content-Length: <computed-length>
Connection: keep-alive
X-Auth-Token: <session-auth-token>

-----------------------------d74496d66958873e
Content-Disposition: form-data; name="UpdateParameters"
Content-Type: application/json

{
   "Targets": ["/redfish/v1/Managers/1"],
   "@Redfish.OperationApplyTime": "OnReset",
   "Oem": {}
}

-----------------------------d74496d66958873e
Content-Disposition: form-data; name="UpdateFile"; filename="flash.bin"
Content-Type: application/octet-stream

<software image binary>

12.7 Outbound connections

12.7.1 Overview

There are cases where it may not be possible for a client to directly connect to a Redfish service. One common example is that the Redfish service is behind a firewall. Services may implement WebSocket-based outbound connections to establish a tunnel between the service and a remote client. If supported, services shall implement the OutboundConnectionCollection resource subordinate to the AccountService resource.

Clients may create new OutboundConnection resources to create WebSocket connections to remote clients. Remote clients initiate Redfish requests to the service by encoding Redfish HTTP requests in the WebSocket payload over this connection.

12.7.2 Establishing an outbound connection

The service shall establish a persistent RFC6455-defined WebSocket connection for each OutboundConnection resource where ConnectionEnabled contains true with the configured authentication mechanism to the address specified by the EndpointURI property. The service may apply proxy settings to the connection based on the configuration of the ManagerNetworkProtocol resource.

The service shall include the HTTP headers specified in the PreUpgradeHTTPHeaders property in the initial HTTPS request to establish the WebSocket, prior to the upgrade to a WebSocket. The service shall include the Sec-WebSocket-Protocol request header and it shall contain the value Redfish. The following example shows an initial HTTPS request from a service prior to the upgrade to a WebSocket:

GET /redfish HTTP/1.1
Host: clientapp.contoso.com:443
Upgrade: websocket
Connect: Upgrade
Sec-WebSocket-Protocol: Redfish
Sec-WebSocket-Key: abcdefhijklmnop==
Sec-WebSocket-Version: 13
PreUpgradeHeader1: PreUpgradeHeader1Value

When a connection is established with the remote client, the service shall create a Session resource with the role defined in the OutboundConnection resource.

If the connection terminates for any reason, any pending Redfish responses are dropped. Services shall re-establish outbound connections:

12.7.3 MTLS for an outbound connection

MTLS is the preferred authentication mechanism to establish a secure connection. This is specified when the Authentication property in the OutboundConnection resource contains MTLS. There are two sets of certificates in the OutboundConnection resource: Certificates and ClientCertificates.

The certificates found in the collection referenced by the Certificates property are the certificates for the remote client's WebSocket service endpoint, typically an HTTPS server with WebSockets enabled. Services shall verify the identity of the WebSocket service during TLS handshaking with the contents of this collection. Services may perform additional verification based on other factors, such as the configuration of the SecurityPolicy resource.

The certificates found in the collection referenced by the ClientCertificate property are the certificates the service provides to the remote client's WebSocket service during TLS handshaking. In some cases, the service may provide its IDEV-ID or LDEV-ID certificate. To verify the identity Redfish service during TLS handshaking, the remote client hosting the WebSocket will need copies of these certificates.

12.7.4 Handling Redfish requests over an outbound connection

Section 5.2 of RFC6455 contains the WebSocket packet definition.

To send a Redfish request, the remote client shall provide the entire HTTP request, including the method, path, version, and headers, in the "Payload data" portion of the WebSocket packet over the established connection. The remote client shall specify either the text (0x1) opcode or binary (0x2) opcode in the request. The request should specify the binary (0x2) opcode if the payload data is binary, such as a firmware image.

To send a Redfish response, the service shall provide the entire HTTP response, including the version, status code, status message, and headers, in the "Payload data" portion of the WebSocket packet over the established connection. The service shall provide the same opcode received in the request from the external client.

Services shall respond to ping (0x9) opcodes with a pong (0xA) opcode. Services shall send the ping (0x9) opcode to the remote client at the interval specified by the WebSocketPingIntervalMinutes property.

The following example shows the contents of "Payload data" for a remote client performing an HTTP GET request on the ServiceRoot resource:

GET /redfish/v1/ HTTP/1.1
Host: clientapp.contoso.com:443
Accept: application/json
Content-Type: application/json

The following example shows the contents of "Payload data" for a remote client performing an HTTP POST request on the Reset action for a ComputerSystem resource:

POST /redfish/v1/Systems/1/Actions/ComputerSystem.Reset HTTP/1.1
Host: clientapp.contoso.com:443
Accept: application/json
Content-Type: application/json
Content-Length: 27

{
    "ResetType": "On"
}

12.7.5 Closing an outbound connection

The service shall close the WebSocket connection to the remote client if:

13 Security details

13.1 Transport Layer Security (TLS) protocol

13.1.1 Transport Layer Security (TLS) protocol overview

Implementations shall support the Transport Layer Security (TLS) protocol v1.2 with RFC7525 recommendations or later. Implementations may remove support for older versions for TLS in favor of newer versions.

DEPRECATED: Previous versions of this specification allowed for TLS v1.1.

Implementations should support:

13.1.2 Cipher suites

Implementations shall only support cipher suites listed as "Recommended" in the TLS Cipher Suites table defined by the IANA TLS Parameters registry.

Cipher suites that are listed as mandatory in various RFCs, but are not "Recommended" in the TLS Cipher Suites table defined by the IANA TLS Parameters registry, shall not be supported.

Implementations should consider the support of pre-shared key ciphers suites listed as "Recommended" in the TLS Cipher Suites table defined by the IANA TLS Parameters registry, which enable authentication and identification without trusted certificates.

DEPRECATED

Implementations should support AES-256-based ciphers from the TLS suites.

Redfish implementations should consider the support of ciphers, such as the following ciphers, which enable authentication and identification without trusted certificates:

    TLS_PSK_WITH_AES_256_GCM_SHA384
TLS_DHE_PSK_WITH_AES_256_GCM_SHA384
TLS_RSA_PSK_WITH_AES_256_GCM_SHA384

The advantage of these recommended ciphers is:

AES-GCM is not only efficient and secure, but hardware implementations can achieve high speeds with low cost and low latency because the mode can be pipelined.

Additionally, Redfish implementations should support the following cipher:

TLS_RSA_WITH_AES_128_CBC_SHA

For more information, see RFC5487 and RFC5288.

END DEPRECATED

13.1.3 Certificates

Redfish implementations shall support replacement of the default certificate if one is provided.

Redfish implementations shall use certificates that conform to X.509-v3, as defined in RFC5280.

13.2 Sensitive data

Operations that contain sensitive data should use HTTPS only. For example, a SimpleUpdate action with a username and password should use HTTPS to protect the sensitive data.

Properties in service responses that represent sensitive data, such as passwords, shall be null.

Responses from URIs where the URI itself contains sensitive data in a URI segment may return the HTTP 404 Not Found status code instead of the HTTP 401 Unauthorized status code, the HTTP 403 Forbidden status code, or the HTTP 405 Method Not Allowed status code to prevent attackers from obtaining the sensitive data in the URI.

13.3 Authentication

13.3.1 Authentication overview

Services:

13.3.2 Authentication requirements

13.3.2.1 Resource and operation authentication requirements

Services shall authenticate all write requests to Redfish resources. For example:

Redfish resources shall not be available as unauthenticated, except for:

Services may reject requests to the previous resources if invalid credentials are provided by the client.

Note: This specification does not cover external services that are linked through external references. These services may have other security requirements.

13.3.2.2 HTTP header authentication requirements

An authentication header shall accompany every request that establishes a secure channel.

Services:

13.3.2.3 Authentication failure requirements

When authentication fails, extended error messages shall not provide privileged information.

13.3.3 HTTP Basic authentication

Services shall support HTTP Basic authentication, as defined by RFC7617, and shall use only connections that conform to TLS to transport the data between any third-party authentication service and clients.

All requests that use HTTP Basic authentication shall require HTTPS.

When multi-factor authentication is enabled, services shall reject HTTP Basic authentication for accounts that are not configured to bypass multi-factor authentication. Session-based authentication is required in this case.

Note: The IETF has highlighted security concerns with HTTP Basic authentication. While HTTPS is required for the usage of HTTP Basic authentication, there are other concerns implementers need to be aware of that RFC7617 documents. This functionality can be restricted or disabled with the HTTPBasicAuth property in the AccountService resource.

13.3.4 Redfish session login authentication

Service shall provide login sessions that conform with this specification.

Session management is determined by the implementation of the Redfish service, which includes orphaned session timeout and the management of the number of simultaneous open sessions.

13.3.4.1 Redfish login sessions

For improved performance and security, a client should use the session management interface to create a Redfish login session. The session service specifies the URI for session management.

To establish a session, find the URI in either:

Both URIs shall be the same.

{
   "SessionService": {
      "@odata.id": "/redfish/v1/SessionService"
   },
   "Links": {
      "Sessions": {
         "@odata.id": "/redfish/v1/SessionService/Sessions"
      }
   },
   ...
}

13.3.4.2 Session login

To create a Redfish session without an authentication header, perform an HTTP POST request on the session service's Sessions resource collection. The POST to create a session shall only be supported with HTTPS. If both HTTP and HTTPS are enabled, a POST request to create a session through the HTTP port should redirect to the HTTPS port. Include the following POST body:

POST /redfish/v1/SessionService/Sessions HTTP/1.1
Host: <host-path>
Content-Type: application/json;charset=utf-8
Content-Length: <computed-length>
Accept: application/json;charset=utf-8
OData-Version: 4.0

{
   "UserName": "<username>",
   "Password": "<password>"
}

Fields in brackets are placeholders for client-specific values.

When a multi-factor authentication type that requires tokens is enabled, services shall require the Token property in the POST request to the SessionCollection resource for accounts that are not configured to bypass multi-factor authentication. The service shall verify the provided token in addition to verifying the username and password. If the Token property is required and not provided by the client, but the client provided a valid UserName and Password combination, the service shall return the HTTP 401 Unauthorized status code and an error response with the AuthenticationTokenRequired message from the Base Message Registry. In addition, if the multi-factor authentication type uses a service-generated one-time passcode, the service shall also return the OneTimePasscodeSent message from the Base Message Registry, and send a one-time passcode to the configured delivery address for that account.

To verify that the request has been initiated from an authorized client domain, services should save the Origin header in reference to this session creation and compare it to subsequent requests using this session.

The response to the POST request to create a session shall include:

The following sample response shows a newly created session:

HTTP/1.1 201 Created
Location: /redfish/v1/SessionService/Sessions/1
X-Auth-Token: <session-auth-token>

{
   "@odata.id": "/redfish/v1/SessionService/Sessions/1",
   "@odata.type": "#Session.v1_0_0.Session",
   "Id": "1",
   "Name": "User Session",
   "Description": "User Session",
   "UserName": "<username>",
   "Password": null
}

The client that sends the session login request should save the session authentication token from the X-Auth-Token header and the contents of the Location header from the response of the login POST request.

To authenticate subsequent requests, the client sets the X-Auth-Token header to the session authentication token that the POST login request returns.

Note: The session ID differs from the session authentication token, as follows:

13.3.4.3 Session lifetime

Unlike some token-based methods that use token expiration times, Redfish sessions time out. As long as a client continues to send requests more frequently than the session timeout period, the session remains open and the session authentication token remains valid. If the session times out, it is automatically terminated.

13.3.4.4 Session termination or logout

When the client logs out, the Redfish session terminates. The session terminates through a DELETE request to the Session resource defined in either the Location header URI or the session ID in the response data.

This ability to DELETE a session through the Session resource enables an administrator with sufficient privileges to terminate other users' sessions from a different session.

When a session is terminated, the service shall not affect independent connections established originally by this session for other purposes, such as connections for server-sent events or transferring an image for the update service.

13.3.5 Client certificate authentication

If client certificate authentication is enabled, the service shall send a client certificate request during the Transport Layer Security (TLS) handshake. When the service obtains a client certificate during the TLS handshake, the service shall verify the certificate with the certificates in CertificateCollection resource referenced by the ClientCertificate property within the MFA property of the AccountService resource. The service shall check for certificate revocation before processing the request with any configured Online Certificate Status Protocol (OCSP) servers.

The RespondToUnauthenticatedClients property within the ClientCertificate property within the MFA property of the AccountService resource controls the response behavior when an invalid certificate is provided by the client.

13.4 Authorization

13.4.1 Authorization overview

The Redfish authorization subsystem controls which users have access to resources and the type of access that users have. It consists of two parts: the privilege model and the operation-to-privilege mapping.

The privilege model maps users to roles and maps roles to privileges. A privilege is a permission to complete an operation, such as read or write, within a defined management domain. For example, the ConfigureUsers privilege allows adding a user. A user is authorized to access a resource if they have the privileges required for that resource. The operation-to-privilege mapping defines which privileges are required to access any given operation.

Redfish allows vendors to extend the standard privilege model with OEM privileges and custom OEM roles. OEM privileges and custom roles participate in the privilege model the same as Redfish standard privileges and roles. Services may also allow clients to create custom roles. Restricted roles and restricted privileges allow vendors to further refine their authority model.

Services shall enforce the same privilege model for ETag-related activity as is enforced for the data being represented by the ETag. For example, the privilege required to read an ETag shall be the same as the privilege to read the data item that the ETag represents.

13.4.2 Privilege model

Each user shall be assigned exactly one role with the RoleId property in the ManagerAccount resource. The value of the RoleId property identifies a Role resource in the RoleCollection resource, where a role defines a set of privileges. A role shall be assigned to a user when a manager account is created. The client shall provide the RoleId property when creating a manager account to select one of the standard or custom roles.

Services shall provide information about all roles through the RoleCollection resource. The AssignedPrivileges and OemPrivileges arrays in the Role resource define a set of assigned privileges for the associated role. Two roles with the same privileges shall behave equivalently.

13.4.2.1 Roles

Redfish defines a set of standard roles, allows a service to define custom OEM roles, and allows client-defined custom roles.

A service shall support all of the standard roles in Table 41. The value of the Id and AssignedPrivileges properties in the Role resource for the standard roles shall contain the Role name and Assigned privileges column values, respectively. The AssignedPrivileges property for standard roles shall not be modifiable. The IsPredefined property for standard roles shall contain the value true.

Table 41 describes the standard roles:

Table 41 — Required standard roles
Role name Assigned privileges
Administrator Login, ConfigureManager, ConfigureUsers, ConfigureComponents, ConfigureSelf
Operator Login, ConfigureComponents, ConfigureSelf
ReadOnly Login, ConfigureSelf

A service may support one or more of the standard roles in Table 42. The value of the Id and AssignedPrivileges properties in the Role resource for the standard roles shall contain the Role name and Assigned privileges column values, respectively. The AssignedPrivileges property for standard roles shall not be modifiable. The IsPredefined property for standard roles shall contain the value true.

Table 42 describes the optional standard roles:

Table 42 — Optional standard roles
Role name Assigned privileges Description
SystemAdministrator Login, ConfigureSelf, AdministrateSystems Administrator for systems found in the systems collection. Able to manage boot configuration, keys, and certificates for systems.
SystemOperator Login, ConfigureSelf, OperateSystems Operator for systems found in the systems collection. Able to perform resets and configure interfaces.
StorageAdministrator Login, ConfigureSelf, AdministrateStorage Administrator for storage subsystems and storage systems found in the storage collection and storage system collection respectively.
StorageBackupOperator Login, ConfigureSelf, OperateStorageBackup Operator for storage backup functionality for storage subsystems and storage systems found in the storage collection and storage system collection respectively.

A service may define custom OEM roles. The IsPredefined property for OEM roles shall contain the value true. A service shall not allow users to modify predefined OEM roles. OEM role names should begin with a lowercase character or "Oem" followed by a vendor name to avoid conflict with future Redfish predefined role names.

A service may allow custom client-defined roles to be created, modified, and deleted. If allowed, a user can perform a POST request on the RoleCollection resource to create a role, indicating privileges in the AssignedPrivileges and OemPrivileges properties in the Role resource. A service may restrict which privileges are allowed. The IsPredefined property for client-defined roles shall contain the value false. A service shall not allow a client-defined role to be deleted while it is in use, for example, when it is assigned to a local user or an LDAP RemoteRoleMapping property.

The value of the RoleId property shall be unique across all roles within the RoleCollection resource.

Non-Redfish services, such as those enabled by the AccountTypes property within the ManagerAccount resource, should map the Redfish RoleId to their permission system. For example, an SSH user with Administrator as the value of the RoleId property could map to "root" for the SSH service. However, the privileges specified by the AssignedPrivileges and OemPrivileges do not necessarily map to non-Redfish services.

13.4.2.2 Restricted roles and restricted privileges

Restricted roles and restricted privileges are intended to prevent privilege escalation. Restricted roles and restricted privileges are not less functional, but their usage is restricted to particular users. For example, to have a security administrator have privileges that the administrator does not have, you need to ensure the administrator cannot escalate to the security administrator role. An implementation can help achieve this by restricting the Administrator role and providing an alternate administrator role that lacks the security privilege.

A service may restrict any role. The Restricted property for restricted roles shall contain the value true. When a standard role is restricted, services shall provide the AlternateRoleId property to reference a non-restricted custom role intended for clients to use as an alternate. Services may pre-define or create accounts that are configured with a restricted role.

Services shall not allow:

A service may restrict any privilege, including standard and OEM privileges. The RestrictedPrivileges and RestrictedOemPrivileges properties in the AccountService resource shall specify the restricted privileges. Services shall not allow custom roles to specify restricted privileges. Services may contain predefined roles that are configured with restricted privileges.

13.4.2.3 OEM privileges

OEM privileges allow a service to extend the privilege model by adding additional privileges to have additional control of what operations are allowed. It can be used when a standard privilege is overly broad.

A service may define OEM privileges and may include OEM privileges in any predefined role, including standard and custom OEM roles. The OemPrivileges property within the Role resource shall contain the OEM privileges that are assigned to the role. The OemPrivileges property in the Role resource for the predefined roles shall not be modifiable.

A service may allow OEM privileges to be assigned to client-defined roles.

13.4.3 Redfish service operation-to-privilege mapping

For every request that a client makes to a service, the service shall determine that the authenticated identity of the requester has the authorization to complete the requested operation on the resource in the request.

Using the role and privileges authorization model where an authenticated identity context is assigned a role and a role is a set of privileges, the service typically checks an HTTP request against a mapping of the authenticated requesting identity role and privileges to determine whether the identity privileges are sufficient to complete the operation in the request.

A service may perform additional checks based on the identity of the user and remove data from responses. For example, a service might restrict access for non-administrative users to only access their own ManagerAccount, Session, and EventDestination resources.

13.4.3.1 Why specify operation-to-privilege mapping?

Initial versions of the Redfish Specifications defined several role-to-privilege mappings for standardized roles and normatively identified several privilege labels but did not normatively detail what these privileges or how privilege-to-operations mappings could be specified or represented in a normative fashion.

The lack of a methodology to define which privileges are required to complete a requested operation against the URI in the request puts at risk the interoperability between service implementations that clients may encounter due to variances in privilege requirements between implementations.

Also, a lack of methodology for specifying and representing the operation-to-privilege mapping prevents the Redfish Forum or other governing organizations from normatively defining privilege requirements for a service.

13.4.3.2 Representing operation-to-privilege mappings

A service should provide a Privilege Registry in the registry collection. This registry represents the privileges required to complete HTTP operations against resources supported by the service.

The Privilege Registry is a JSON document that contains a Mappings array of where an individual entry exists for every resource type that the service supports.

The operation-to-privilege mapping is defined for every resource type and applies to every resource the service implements for the applicable resource type.

In several situations, specific resources or properties may have differing operation-to-privilege mappings than the resource type-level mappings. In these cases, the resource type-level mappings need to be overridden. The PrivilegeRegistry schema defines the methodology for resource type-level operation-to-privilege mappings and related overrides.

If a service provides a Privilege Registry, the service shall use the Redfish Forum's Privilege Registry definition as a base operation-to-privilege mapping definition for operations that the service supports to promote interoperability for Redfish clients.

13.4.3.3 Operation map syntax

An operation map defines the set of privileges required to complete an operation on a resource-type.

The mapped operations are GET, PUT, PATCH, POST, DELETE, and HEAD. A privilege mapping is defined for each operation, irrespective of whether the service or data model supports the operation on the resource-type.

The privilege labels may be the Redfish standardized labels that the PrivilegeType enumeration in the Privileges schema defines and they may be OEM-defined privilege labels. The required privileges for an operation are specified using logical AND and OR behavior. For more information, see the Privilege AND and OR syntax clause.

The following example defines the privileges required for various operations on the Manager resource. Unless the implementation defines mapping overrides to the OperationMap array, the specified operation-to-privilege mapping represents behavior for all Manager resources in a service implementation.

{
   "Entity": "Manager",
   "OperationMap": {
      "GET": [{
         "Privilege": ["Login"]
      }],
      "HEAD": [{
         "Privilege": ["Login"]
      }],
      "PATCH": [{
         "Privilege": ["ConfigureManager"]
      }],
      "POST": [{
         "Privilege": ["ConfigureManager"]
      }],
      "PUT": [{
         "Privilege": ["ConfigureManager"]
      }],
      "DELETE": [{
         "Privilege": ["ConfigureManager"]
      }]
   }
}

13.4.3.4 Mapping overrides syntax

Table 43 describes the operation-to-privilege mapping, which varies from the resource type-level mapping:

Table 43 — Mapping overrides syntax
Situation Description
Property override Property has different privilege requirements than the resource in which it resides. For example, the Password property in the ManagerAccount resource requires the ConfigureSelf or ConfigureUsers privilege to change, in contrast to the ConfigureUsers privilege required for the other properties in ManagerAccount resources. If multiple properties with the same name are present in a resource, the property override applies to all property instances.
Subordinate override Resource is used in context of another resource and the contextual privileges need to govern. For example, the privileges for PATCH operations on EthernetInterface resources depend on whether the resource is subordinate to the Manager resource, where ConfigureManager is required, or the ComputerSystem resource, where ConfigureComponents is required.
Resource URI override Resource instance has different privilege requirements for an operation than those defined for the resource type.

The overrides are defined in the context of the operation-to-privilege mapping for a resource type.

If multiple overrides are specified for a single resource type, the following precedence should be used for determining the appropriate override to apply:

13.4.3.5 Property override example

In the following example, the Password property on the ManagerAccount resource requires the ConfigureSelf or ConfigureUsers privilege to change, in contrast to the ConfigureUsers privilege required for the other properties in ManagerAccount resources:

{
   "Entity": "ManagerAccount",
   "OperationMap": {
      "GET": [{
         "Privilege": ["ConfigureManager"]
      }, {
         "Privilege": ["ConfigureUsers"]
      }, {
         "Privilege": ["ConfigureSelf"]
      }],
      "HEAD": [{
         "Privilege": ["Login"]
      }],
      "PATCH": [{
         "Privilege": ["ConfigureUsers"]
      }],
      "POST": [{
         "Privilege": ["ConfigureUsers"]
      }],
      "PUT": [{
         "Privilege": ["ConfigureUsers"]
      }],
      "DELETE": [{
         "Privilege": ["ConfigureUsers"]
      }]
   },
   "PropertyOverrides": [{
      "Targets": ["Password"],
      "OperationMap": {
         "PATCH": [{
            "Privilege": ["ConfigureUsers"]
         }, {
            "Privilege": ["ConfigureSelf"]
         }]
      }
   }]
}

13.4.3.6 Subordinate override

The Targets property in SubordinateOverrides lists a hierarchical representation for when to apply the override. In the following example, the override for an EthernetInterface resource is applied when it is subordinate to an EthernetInterfaceCollection resource, which in turn is subordinate to a Manager resource. If a client were to PATCH an EthernetInterface resource that matches this override condition, it requires the ConfigureManager privilege. Otherwise, the client requires the ConfigureComponents privilege.

{
   "Entity": "EthernetInterface",
   "OperationMap": {
      "GET": [{
         "Privilege": ["Login"]
      }],
      "HEAD": [{
         "Privilege": ["Login"]
      }],
      "PATCH": [{
         "Privilege": ["ConfigureComponents"]
      }],
      "POST": [{
         "Privilege": ["ConfigureComponents"]
      }],
      "PUT": [{
         "Privilege": ["ConfigureComponents"]
      }],
      "DELETE": [{
         "Privilege": ["ConfigureComponents"]
      }]
   },
   "SubordinateOverrides": [{
      "Targets": ["Manager", "EthernetInterfaceCollection"],
      "OperationMap": {
         "PATCH": [{
            "Privilege": ["ConfigureManager"]
         }],
         "POST": [{
            "Privilege": ["ConfigureManager"]
         }],
         "PUT": [{
            "Privilege": ["ConfigureManager"]
         }],
         "DELETE": [{
            "Privilege": ["ConfigureManager"]
         }]
      }
   }]
}

13.4.3.7 Resource URI override

The following example demonstrates the resource URI override syntax to define operation privilege variations for resource URIs.

The example defines both ConfigureComponents and OEMAdminPriv privileges as required to make a PATCH operation on the two resource URIs listed as targets.

{
   "Entity": "ComputerSystem",
   "OperationMap": {
      "GET": [{
         "Privilege": ["Login"]
      }],
      "HEAD": [{
         "Privilege": ["Login"]
      }],
      "PATCH": [{
         "Privilege": ["ConfigureComponents"]
      }],
      "POST": [{
         "Privilege": ["ConfigureComponents"]
      }],
      "PUT": [{
         "Privilege": ["ConfigureComponents"]
      }],
      "DELETE": [{
         "Privilege": ["ConfigureComponents"]
      }]
   },
   "ResourceURIOverrides": [{
      "Targets": ["/redfish/v1/Systems/VM6", "/redfish/v1/Systems/Sys1"],
      "OperationMap": {
         "GET": [{
            "Privilege": ["Login"]
         }],
         "PATCH": [{
            "Privilege": ["ConfigureComponents", "OEMSysAdminPriv"]
         }]
      }
   }]
}

13.4.3.8 Privilege AND and OR syntax

The array placement of the privilege labels in the OperationMap GET, HEAD, PATCH, POST, PUT, and DELETE operation element arrays define the logical combinations of privileges that are required to call an operation on a resource or property.

For OR logical combinations, the privilege label appears in the operation element array as individual elements.

The following example defines either Login or OEMPrivilege1 privileges that are required to perform a GET request.

{
   "GET": [{
      "Privilege": ["Login"]
   }, {
      "Privilege": ["OEMPrivilege1"]
   }]
}

For logical AND combinations, the privilege label appears in the Privilege property array in the operation element.

The following example defines both ConfigureComponents and OEMSysAdminPriv that are required to perform a PATCH request.

{
   "PATCH": [{
      "Privilege": ["ConfigureComponents", "OEMSysAdminPriv"]
   }]
}

13.4.4 Delegated authorization with OAuth 2.0

Services may support the RFC6749-defined OAuth 2.0 authorization framework.

13.4.4.1 OAuth 2.0 overview

The OAuth 2.0 authorization framework allows a client to obtain access to a resource server from a resource owner and an authorization server.

Clients request access from a resource owner and is given an authorization grant. The authorization grant is then provided to the authorization server and an access token is provided to the client. The client provides the access token to the resource server in order to access a protected resource.

A Redfish service is considered to be a resource server in the OAuth 2.0 authorization framework.

13.4.4.2 OAuth 2.0 data model requirements

Services that support OAuth 2.0:

13.4.4.3 OAuth 2.0 access tokens

Access tokens are the credentials the client provides to a service to access a protected resource. Clients provide the access token to the service in the Authorization request header as a bearer token.

Services that support OAuth 2.0 shall support receiving an RFC7519-defined JSON Web Token (JWT) in the Authorization request header.

JWTs are a compressed JSON structure that contain a JOSE Header, a set of claims that describe the type of access that is granted to a client, and a signature. Each component of a JWT is Base64URL-encoded and concatenated with a . to form the token string for the Authorization header.

Table 44 describes the JWT JOSE Header parameters and their requirements for services and clients. Any other parameters are outside the scope of this specification.

Services shall process the parameters in Table 44 if the Service requirement column contains Yes. Services should process other parameters.

The JWT provided by the client shall contain the parameters in Table 44 if the JWT requirement column contains Yes. The JWT provided by the client may omit other parameters.

Table 44 — OAuth 2.0 JWT JOSE Header parameters
Parameter Service requirement JWT requirement Description
typ Yes No Type of token. The string is case-insensitive. If not present, services shall assume the value is JWT.
alg Yes Yes Algorithm for the signature of the token. Services shall not accept the value none.

Table 45 describes the claims and their requirements for services and clients. Any other claims are outside the scope of this specification.

Services shall process the claims in Table 45 if the Service requirement column contains Yes. Services should process other claims.

The JWT provided by the client shall contain the claims in Table 45 if the JWT requirement column contains Yes. The JWT provided by the client may omit other claims.

Table 45 — OAuth 2.0 JWT claims
Claim Service requirement JWT requirement Description
iss Yes Yes Issuer of the token. Identifies the authorization server that signed the token.
sub Yes Yes Subject of the token. Identifies the client issued the token.
aud Yes Yes Audience of the token. Identifies the resource server intended to accept the token.
exp Yes No Expiration time of the token.
nbf Yes No "Not before" time of the token.
iat Yes No Issued time of the token.
jti Yes No Unique identifier of the token.
scope Yes Yes Type of access the token grants. See the Redfish OAuth2.0 scope usage clause.

Example JOSE Header:

{
    "typ": "JWT",
    "alg": "RS256"
}

Example JWT claims:

{
    "iss": "https://contoso.org/services/oauth2",
    "sub": "Joe Smith",
    "aud": "92384634-2938-2342-8820-489239905423",
    "exp": 1735707600,
    "scope": "Redfish.Role.Operator",
    "jti": "97d52311-5f55-4482-b947-8a70c326fdfd"
}

Example token encoded in the Authorization request header:

Authorization: Bearer mF_9.B5f-4.1JqM

Note: The previous example does not reflect a real JWT and is provided to show encoding in the Authorization request header.

13.4.4.4 Redfish OAuth2.0 scope usage

The value of the scope claim is expressed as a list of space-delimited, case-sensitive strings. Each value in the list describes a type of access that was granted to the client.

This specification defines two formats for values in the scope claim: Redfish roles and Redfish privileges. Other formats are outside the scope of this specification.

Redfish roles within the scope claim shall be in the form Redfish.Role.<RoleId> where <RoleId> is the identifier of the Redfish role granted to the client.

Redfish privileges within the scope claim shall be in the form Redfish.Privilege.<PrivilegeId> where <PrivilegeId> is the standard privilege or OEM privilege granted to the client.

Services shall ignore unsupported values in the scope claim. If the token provided by the client is valid, the service shall apply roles and privileges in the scope claim to the operation.

13.5 Account service

13.5.1 Account service overview

13.5.2 Password management

A Redfish service provides local user accounts through a collection of ManagerAccount resources located under the account service. The ManagerAccount resources enable users to manage their own account information, and for administrators to create, delete, and manage other user accounts.

When account properties are changed, the service may close open sessions for this account and require re-authentication.

13.5.3 Atomic password changes

If the RequireChangePasswordAction property in the AccountService resource contains true, services shall reject modifications to the Password property in ManagerAccount resources in PATCH or PUT operations. Clients are required to invoke the ChangePassword action in the ManagerAccount resource to update the password for a user account.

This functionality can be used to protect the service from session hijacking. The SessionAccountPassword parameter in the ChangePassword action allows the service to verify the client's credentials before changing the password of the account. If a user is changing their own password, the SessionAccountPassword parameter shall contain the current password of the account. If an administrator is performing the password change for a different user, the SessionAccountPassword parameter shall contain the administrator's password.

If the client is performing requests with HTTP Basic authentication, the SessionAccountPassword contains the same password encoded in the Authorization request header.

13.5.4 Password change required handling

The service may require that passwords assigned by the manufacturer be changed by the end user prior to accessing the service. In addition, administrators may require users to change their account's password upon first access.

The ManagerAccount resource contains a PasswordChangeRequired boolean property to enable this functionality. Resources that have the property set to true shall require the user to change the write-only Password property in that resource before access is granted. Manufacturers including user credentials for the service may use this method to force a change to those credentials before access is granted.

When a client accesses the service by using credentials from a ManagerAccount resource that has a PasswordChangeRequired value of true, the service shall allow:

When a client accesses the service by using credentials from a ManagerAccount resource that has a PasswordChangeRequired value of true, the service may allow GET operations on unauthenticated resources, such as the ServiceRoot resource.

For all other operations, the service shall respond with the HTTP 403 Forbidden status code and an error response with the PasswordChangeRequired message from the Base Message Registry.

13.6 Asynchronous tasks

Irrespective of which user or privileged context starts a task, services shall enforce the privileges described in the privilege registry required to perform operations on the Task resource.

13.7 Event subscriptions

Before pushing event data object to the destination, the service may verify the destination for identity purposes.

Services shall not send events to clients that do not have sufficient privilege to access the resource originating the event. For push-style eventing, this is determined by the client that performed an HTTP POST to the subscription collection. For server-sent events, this is determined by the client that opened the SSE connection.

14 Redfish Host Interface

The Redfish Host Interface Specification defines how software that runs on a host computer system can interface with a Redfish service that manages the host. For details, see DSP0270.

15 Redfish composability

A service may implement the CompositionService resource off of ServiceRoot to bind resources. One example is disaggregated hardware, which allows for independent components, such as processors, memory, I/O controllers, and drives, to be bound to create logical constructs that operate together. This enables a client to dynamically assign resources for an application.

A service that supports composability shall implement resource blocks, defined by the ResourceBlock schema, and resource zones, defined in the Zone schema, for the composition service. Resource blocks provide an inventory of components available to the client for building compositions. Resource zones describe the binding restrictions of the resource blocks that the service manages.

The resource zones within the composition service shall include the collection capabilities annotation in responses. The collection capabilities annotation allows a client to discover which resource collections in the service support compositions, the different composition request types allowed, how the POST request for the resource collection is formatted, and which properties are required.

A service that supports composability and client multi-tenancy shall:

15.1 Composition requests

15.1.1 Composition requests overview

A service that implements the composition service, as defined by the CompositionService schema, shall support one or more of the following types of composition requests:

A service that supports the removal of a composed resource shall support the DELETE method on the composed resource.

A service may implement the Compose action in the CompositionService resource for the above composition requests.

15.1.2 Specific composition

A specific composition is when a client identifies an exact set of resources in which to build a logical entity.

A service that supports specific compositions shall support a POST request that contains an array of hyperlinks to resource blocks. The schema for the resource being composed defines where the resource blocks are specified in the request.

The following example shows a ComputerSystem being composed with a specific composition request:

POST /redfish/v1/Systems HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
   "Name": "Sample Composed System",
   "Links": {
      "ResourceBlocks": [{
         "@odata.id": "/redfish/v1/CompositionService/ResourceBlocks/ComputeBlock0"
      }, {
         "@odata.id": "/redfish/v1/CompositionService/ResourceBlocks/DriveBlock2"
      }, {
         "@odata.id": "/redfish/v1/CompositionService/ResourceBlocks/NetBlock4"
      }]
   }
}

15.1.3 Constrained composition

A constrained composition is when a client has identified a set of criteria, or constraints, in which to build a logical entity. This includes criteria such as quantities of components, or characteristics of components. A service that supports constrained compositions shall support a POST request that contains the set of characteristics to apply to the composed resource. The specific format of the request is defined by the schema for the resource being composed. This type of request may include expanded elements of resources subordinate to the composed resource.

The following constrained composition request composes a ComputerSystem:

POST /redfish/v1/Systems HTTP/1.1
Content-Type: application/json;charset=utf-8
Content-Length: <computed length>
OData-Version: 4.0

{
   "Name": "Sample Composed System",
   "PowerState": "On",
   "BiosVersion": "P79 v1.00 (09/20/2013)",
   "Processors": {
      "Members": [{
         "@Redfish.RequestedCount": 4,
         "@Redfish.AllowOverprovisioning": true,
         "ProcessorType": "CPU",
         "ProcessorArchitecture": "x86",
         "InstructionSet": "x86-64",
         "MaxSpeedMHz": 3700,
         "TotalCores": 8,
         "TotalThreads": 16
      }]
   },
   "Memory": {
      "Members": [{
         "@Redfish.RequestedCount": 4,
         "CapacityMiB": 8192,
         "MemoryType": "DRAM",
         "MemoryDeviceType": "DDR4"
      }]
   },
   "SimpleStorage": {
      "Members": [{
         "@Redfish.RequestedCount": 6,
         "Devices": [{
            "CapacityBytes": 322122547200
         }]
      }]
   },
   "EthernetInterfaces": {
      "Members": [{
         "@Redfish.RequestedCount": 1,
         "SpeedMbps": 1000,
         "FullDuplex": true,
         "NameServers": ["names.redfishspecification.org"],
         "IPv4Addresses": [{
            "SubnetMask": "255.255.252.0",
            "AddressOrigin": "Dynamic",
            "Gateway": "192.168.0.1"
         }]
      }]
   }
}

15.1.4 Expandable resources

An expandable resource is when a service has a baseline composition that cannot be removed. Instead of a client making requests to create a composed resource, a client can only add or remove resources from the composed resource. A service that supports expandable resources shall support one or more of the update methods that the Updating a composed resource clause describes.

15.2 Updating a composed resource

A service that supports updating a composed resource shall provide one or more of the following methods to update composed resources:

16 Aggregation

Aggregation has been a Redfish concept since its inception. Redfish uses collection for services that can represent more than one system. As the scale of Redfish implementations increase, clients want to operate on Redfish resources in bulk.

Aggregation is the representation of Redfish resources from a variety of sources so that they can be managed, in whole or in part, by a Redfish client. Membership can be heterogeneous and arbitrary, but it is expected that most aggregate members are the same resource type, such as an aggregate of ComputerSystem resource, which is represented by an Aggregate resource where members of its Elements array are exclusively of type ComputerSystem. The Redfish service proxies on behalf of the aggregated components to provide common operations. The Redfish service is representing resources on behalf of the components and incoming operations must be tracked by the Redfish service before being accomplished by communicating with the individual resources. Thus, aggregation also allows a Redfish client to act on resources as a group using aggregates.

16.1 Classes of aggregators

16.1.1 Implicit and complex aggregators

There are at least two classes of Redfish aggregators:

16.1.2 Use cases

Several use cases make explicit aggregator representation necessary. What they have in common is the need for common functions for scalability. There are several classes of these common functions.

One use case is service-type functions. An example is a firmware update on a large number of systems. Rather than invoke actions on individual resources, it is more efficient for a client to specify to which resources to apply the image. In this case, a service already exists in the model so an aggregation service is not needed. Instead, the existing service must be augmented to enable the application of an image to a list of resources.

Another use case is common actions. Examples are the Reset or SetDefaultBootOrder actions. These actions are defined in the ComputerSystem schema, but the Redfish URI structure requires that the action occur on each ComputerSystem resource. Thus, an individual operation applies to each resource. It is more efficient for a client to send one action with the list of the resources to which to apply the action. For example, to reset one thousand systems, sending one thousand individual reset operations requires significant overhead as compared to sending a single operation with a list of one thousand systems to reset.

A final use case is changing an attribute on multiple members of a collection. An example is changing the boot order on a large number of systems. This use case requires one operation per system. However, assuming the resources are in the same collection, the deep PATCH operation meets the requirements of this use case.

16.2 Aggregation service

16.2.1 Aggregation service overview

The AggregationService resource represents the Redfish aggregation service, which provides aggregation functions.

The aggregation service contains the group actions that can apply to groups of resources. The AggregationService schema defines the common actions that a client can take on groups of resources. These actions take an array of resource URIs as one of the parameters to which the action applies. If all members of the resource array do not support the method, a 4xx status code shall be returned and the body shall contain an error response. If at least one member of the resource array successfully completed the action but others did not, the status code should be 200 OK with @Message.ExtendedInfo objects for the failed members.

The aggregation service also contains Aggregate, AggregationSource, and ConnectionMethod resources.

16.2.2 Aggregator requirements

By implementing the AggregationService resource and including an AggregationSourceCollection resource, a complex aggregator shall meet the following requirements:

16.2.3 Aggregates

The Aggregate resource is the grouping mechanism that clients use to indicate to the service that this group of resources can be treated the same for certain functions, such as the actions. Each aggregate contains the list of individual resources that are to be treated as a single unit for operations. For example, if a client wishes to express that a subset of the ComputerSystemCollection resource be treated as a single unit for certain operations like reset, reset boot order, or firmware update, it can express the aggregate as the target URI for the operation.

The Aggregate schema defines the common actions that a client can make on an aggregate. The Aggregate resource contains an Elements array that specifies the members of the aggregate. Actions that are supported on an aggregate but not supported on all Elements, such as a Reset action that is not supported on an individual member of the Elements array, are not silently skipped. If all members of the Elements array do not support the method, a 4xx status code shall be returned and the body shall contain an error response. If at least one member of the Elements array successfully completed the action, but others did not, the status code should be 200 OK with @Message.ExtendedInfo objects for the failed members.

16.2.4 Aggregation sources and connection methods

The aggregation service model also includes a definition for the information used to access the resources being represented by the aggregator. Two collections of resources are used to represent this. These are the AggregationSource and ConnectionMethod resources.

The AggregationSource resource represents the source of information for the resources being reflected by the aggregator. It typically represents a lower layer service provided by another manager. It contains information needed to access that source, such as the address and account information. It also has a reference to the ConnectionMethod resource used to access it.

The ConnectionMethod resource represents the protocol and other semantics required to communicate with the resources being aggregated. Examples of connection methods are Redfish, IPMI, and proprietary access methods. For methods such as IPMI, it's also possible to specify the variations and nuances from multiple vendors.

17 ANNEX A (informative) Change log

Version Date Description
1.20.2 2024-08-01 Updated the Registries clause to correct terminology for attribute registry files.
Added the /redfish/v1/Registries URI to the Redfish-defined URIs and relative reference rules clause.
Updated the Schema modification rules clause to give guidance for resolving multiple versions of the same modified schema.
Corrected terminology in the OEM actions clause to better describe how the action property name is constructed in payloads.
Updated the Naming rules clause to restrict the usage of ReturnType for action parameter names.
Updated the Password change required handling clause to clarify that the HTTP 403 Forbidden response body follows the error response format.
1.20.1 2024-04-03 Updated the Event subscriptions clause to clarify that privilege enforcement is required when sending events.
Updated the Query parameter overview clause to optionally allow for case-insensitive verification of query parameter names.
Updated the Naming rules clause to prevent naming collisions where the only difference is the casing of letters.
Updated the OEM property format and content clause to give guidance for constructing the value of @odata.type to avoid naming collisions.
Updated the Schema modification rules clause to allow for non-semantic text changes.
Added the Transiently unavailable resources clause to describe best practices for showing the status of a device that is in an unavailable state.
Clarified The $expand query parameter to show that HTTP 200 OK is the proper response code when a resource collection is partially expanded.
Updated the Status codes clause to remove the recommendation to avoid the HTTP 100 Continue status code since this is mandatory to support for chunked client requests from the HTTP 1.1 specifications.
1.20.0 2023-11-30 Added requirement to the Establishing an outbound connection clause to require including the Sec-WebSocket-Protocol HTTP header with the value Redfish.
Updated the POST (action) clause to require services to accept empty JSON objects for actions that do not have required parameters. This change might require modification to an implementation to support this version of the specification.
1.19.1 2023-11-30 Made several changes for style consistency, grammar, and general clarity. Except for the following additions, no normative changes were made. Any clarifications that inadvertently altered the normative behavior are considered errata and will be corrected in future revisions to the specification.
Clarified the Password change required handling clause to show that users can delete their own sessions.
Clarified the Password change required handling clause to show that performing GET operations on unauthenticated resources might be allowed by the service.
Updated the Action info annotation and OEM actions clauses to give guidance for constructing the URI for the ActionInfo resource.
Clarified the Property element clause to better describe the usage of Edm.Decimal and Edm.Double.
Updated the PATCH (update), POST (create), and POST (action) clauses to give guidance to not require clients to specify extraneous data in request bodies.
Clarified the Multipart HTTP push updates to show that clients are not required to provide an OEM form.
Updated the Next link (Members@odata.nextLink) property clause to give guidance for handling a client's original query parameters that affect the number of members returned in the collection.
Updated the Next link (Members@odata.nextLink) property clause to give guidance for using the $skiptoken query parameter.
Corrected the PATCH (update) clause to require responding with the HTTP 400 Bad Request with the NoOperation message or a modification success response when the client only gives OData annotations in the request body.
Clarified the Allowable values for strings, Allowable patterns for string values, Allowable values for numbers and durations clauses to show that these terms are used to express client modification restrictions on properties.
Updated The $expand query parameter to describe behavior when used with resources containing reference objects with the expanded resource annotation.
Updated the Response headers clause to show that the WWW-Authenticate is only required if HTTP Basic authentication is configured to advertise its capability.
Updated the HTTP Basic authentication clause to explain that the HTTPBasicAuth property in the AccountService resource can disable or restrict this functionality.
Clarified The $expand query parameter to describe the behavior when a client performs an expansion request, but does not have sufficient privilege to access referenced resources.
Clarified the Naming rules clause to describe naming pattern for string properties containing URIs.
Added the URI annotation clause to document how to indicate a string property is expected to contain a URI.
1.19.0 2023-08-03 Added the Atomic password changes clause to define behavior for services that require password changes with the ChangePassword action.
Updated the Password change required handling clause to allow for a client to invoke the ChangePassword action when a password change is required.
Added the Outbound connections clause to define how a service creates a WebSocket with a remote client through which Redfish requests are tunneled.
Updated the Session login clause to define specific messages in responses when a token is required in the request to create a session.
1.18.1 2023-08-03 Clarified the Asynchronous operations clause to state that HTTP 202 Accepted response bodies contain a Task resource or are absent.
Updated the ETags clause to recommend weak ETag comparisons in PUT or PATCH operations.
Updated the Naming rules clause to clarify all names are alphanumeric and start with letters.
Updated the Naming rules clause to clarify the naming style for enumeration values.
Updated the Deep operations clause to show that a deep POST can include different types of resources in the request.
Updated the Redfish-defined URIs and relative reference rules and Asynchronous operations clauses to give guidance on the recommended URI pattern for task monitors.
Updated The $expand query parameter to recommend that all properties in requested resources are included in responses.
1.18.0 2023-04-04 Added URI segment annotation clause to allow schema to describe when URIs do not meet expected naming rules.
1.17.1 2023-04-04 Updated example in M-SEARCH response to better explain the different values a service provides in an M-SEARCH response.
Updated Writable properties annotation to include an example.
Updated Modification success responses to provide links to appropriate sections for each operation. Clarified that response bodies for create operations are optional.
Updated the response table in POST (action) to use consistent terminology for action responses.
Updated Resource and operation authentication requirements to state that services can optionally reject requests to unauthenticated resources if the provided credentials are invalid.
Updated Action responses to provide guidance for finding the schema definition of the action response based on the action's name in response payloads.
Updated The $select query parameter to clarify that unsupported properties are omitted from responses.
Updated Response headers to make Access-Control-Allow-Origin an optional response header.
Clarified Duration values to state that negative durations are not allowed.
1.17.0 2022-12-08 Updated Deprecated annotation to allow for deprecating URIs.
Updated Protocol details to allow for optional HTTP 2.0 support.
Updated Permissions annotation to allow for write-only to be specified to enforce services respond with null for their value.
Updated HTTP Basic authentication and Redfish session login authentication to describe behavior when multi-factor authentication is enabled.
Added Client certificate authentication clause to allow for client certificates to be authenticated during TLS handshaking.
1.16.1 2022-12-08 Updated GUID and UUID values to correct the format for GUID and UUID properties.
Updated The $select query parameter to clarify the syntax of $select for arrays.
Updated Status codes to remove references to nonexistent HTTP headers.
Updated Sensitive data to clarify that the phrase "URIs containing sensitive data" is meant to reflect the URI itself rather than the response body.
Corrected the example RelatedProperties property in message objects throughout the specification to remove the leading # to meet syntax specified by RFC6901.
Updated MessageId format to better formalize the string tokens used to distinguish the different parts of MessageId values.
1.16.0 2022-08-04 Added optional standard roles to the Roles clause.
Added new Allowable patterns for string values, Allowable values for numbers and durations, and Writable properties annotation payload annotations.
Renamed allowable values payload annotation to Allowable values for strings.
1.15.2 2022-08-04 Updated URI naming rules to give an example URI when a hyperlink to a subordinate resource is not found at the root of the resource.
Updated Modification success responses to not imply a service deficiency if HTTP 204 No Content is returned.
Clarified the usage of the WWW-Authenticate header in the Response headers clause.
Updated Redfish service operation-to-privilege mapping to explain that the service can filter response data based on the user's identity beyond what is assigned as a privilege.
Updated Links to state that some design exceptions have been made for putting a hyperlink inside of Links or at the root of a resource.
Updated ETags to give guidance to implementers when managing resources that update frequently.
Updated Properties overview to state that all properties are required to have well-known types, with some exceptions.
Updated OEM-specified object naming to allow for exceptions to domain suffix rules for listed organizations.
Updated Asynchronous operations to better distinguish Task resources from task monitors to the reader.
Updated The $select query parameter to allow for object-level messages and messages for requested properties even if not specified by the client.
Clarified The $select query parameter that error responses act as if $select was not specified.
1.15.1 2022-04-07 Updated Query parameter overview to recommend the = character is ignored if specified with the only and excerpt query parameters.
Updated Multipart HTTP push updates to use consistent terminology for action parameters, and to correct the use of HTTP status code 413 instead of 412 to indicate a payload is too large to be processed by the service.
Updated Schema definition languages to use normative terminology throughout.
Clarified Asynchronous tasks to explain the allowable operations on a Task resource are controlled by the privilege registry.
Clarified ETags to describe the behavior of ETags during expansion.
Deprecated statement in POST to subscription collection with regards to keeping MessageIds and RegistryPrefixes as mutually exclusive.
Updated Subscribing to events to match the schema updates made to EventDestination.
1.15.0 2021-12-02 Updated the Operation apply time and Multipart HTTP push updates to expand the usage of the @Redfish.OperationApplyTimeSupport annotation to properties referencing URIs for multipart HTTP POST operations.
1.14.2 2021-12-02 Clarified that the Created message from the Base Message Registry is an acceptable response in the POST (action) clause when a new resource is created.
Clarified the Permissions annotation clause to describe requirements for schemas published by DMTF.
Updated the Units of measure annotation clause to describe encoding rules for units not covered by UCUM. Provided alternative recommendations for RPM units.
Clarified Collection capabilities annotation clause that properties marked as required are not required for the resource referenced by the CapabilitiesObject property.
Clarified that @Redfish.MaintenanceWindow property behavior in Task resources in the Asynchronous operations clause.
1.14.1 2021-10-06 Corrected various examples for @Redfish.OperationApplyTimeSupport, @Redfish.Settings, @Redfish.MaintenanceWindow, and @Redfish.SettingsApplyTime to include @odata.type.
Clarified the Link header clause to state the Link response header requirement only applies to resources.
1.14.0 2021-09-15 Extended Query parameter overview clause to define how OEM query parameters are constructed.
Added Delegated authorization with OAuth 2.0 clause to define how clients provide OAuth 2.0 tokens to a service as a method of authorization.
1.13.1 2021-08-04 Various clarifications to the Extending standard resources clause to better describe naming rules for OEM resources.
Added recommended URI for local schema files to the Redfish-defined URIs and relative reference rules clause.
Clarified the OData $metadata clause that any absolute or relative URI is allowed for referencing schema files.
Adding missing statement to the URI naming rules clause that Members cannot be used as the value of a URI segment for resource collections.
Added missing exceptions to the PATCH (update) clause for when @odata.id is to not be ignored by the service.
Clarified the ETags clause that both strong and weak ETags are allowed in If-Match and If-None-Match request headers.
Clarified the Deep operations clause to specify that services ignore resources in the payload if no modifications are requested.
Added the Enumerations clause to clarify the design patterns for creating enumerations.
Clarified the Id clause that HTTP unsafe characters are not permitted in the value of the Id property due to its usage in URI construction.
Clarified the Non-resource reference properties that these properties are strings containing URIs.
1.13.0 2021-04-08 Added client multi-tenancy behavior to the Redfish composability clause. This adds free pool, active pool, and composition reservation constructs to Redfish composability.
Added Compose action as a method of performing composition requests to the Redfish composability clause.
1.12.1 2021-04-08 International Organization for Standardization (ISO) updates:
Added paragraph numbering.
Added Foreword to the table of contents as an unnumbered heading, and placed Acknowledgments inside Foreword.
Made Scope a level-1 clause.
Normative references: Removed unused normative references and moved some references into Bibliography. The Bibliography lists, for information, those documents which are cited informatively in the document, as well as other information resources.
Changed Abstract to Introduction.
Corrected level-1 clauses to remove hanging paragraphs and to correct the occurrence of the single Use cases and Aggregator requirements sub-clauses.
Terms, definitions, symbols, and abbreviated terms:
  • Combined Symbols and abbreviated terms clause with Terms and definitions clause into Terms, definitions, symbols, and abbreviated terms clause.
  • Formatted the clause correctly.
  • Added the Hardware terms, Web development terms, and Redfish terms sub-clauses to this clause.
  • Removed may, shall, and should from definitions.
  • Removed these terms: managed system, Redfish event receiver, and Redfish provider.
  • Corrected definitions so none begin with an article.
Changed may to can or might where appropriate.
Changed one must to shall.
Added numbered captions to tables and changed occurrences of the following table to use precise references to the table numbers.
Fixed broken cross-references.
Corrected URIs in the deep PATCH example.
Fixed several query parameter examples where string values were not properly wrapped with single quotes.
Corrected Accept-Encoding usage to allow for encoded responses if the client does not provide the header to align with RFC7231.
Clarified usage of DELETE for the @Redfish.OperationApplyTimeSupport term.
Removed duplicative clauses for HTTP 405 Method Not Allowed usage in PATCH (update) in favor of more general clauses.
Replaced exception table in PATCH (update) in favor of text.
Moved error cases from response table in POST (action) to be with other text that describes error cases.
Added linkage in the description for HTTP 201 Created to reference response bodies for actions.
Added informative text regarding the usage of If-Match and If-Match-None headers in GET, PATCH, and PUT clauses.
Clarified the behavior of $select when an object property is selected.
Added introductory text to guide readers to other Redfish documents.
Clarified the ordering of processing query parameters.
Clarified that update restrictions for a resource can be modified to support writable OEM properties.
Clarified the Settings resource clause to show behavior of properties in the active resource and settings resource based on the service's capabilities.
Corrected behavior for usage of null based on the configuration of a resource and other special situations.
Clarified OEM naming rules for all OEM definitions to ensure names don't collide.
Removed the term "namespace" from all non-CSDL related clauses and replaced them with references to a new resource type term.
1.12.0 2020-12-01 Added introductory text to the Authorization clause.
Clarified usage of RoleId and how there are standard roles, custom OEM roles, and client-defined custom roles.
Added Restricted roles and restricted privileges to describe behavior for when roles and privileges are marked as restricted.
1.11.2 2020-12-01 Clarified that the Accept-Encoding header is used to request compression of response bodies.
Corrected the PATCH (update), PUT (replace), and DELETE (delete) clauses to leverage all normative statements for successful operations found in the Modification success responses clause.
Replaced RFC5988 reference with RFC8288.
Updated IETF links to use the "IETF Tools" site.
Clarified that insert capabilities is just for resource creation.
Fixed ETag examples to be RFC7234-conformant.
Clarified that OEM resources can have subordinate resources.
Replaced RFC4627 reference with RFC8259.
Replaced conflicting statements found in "HTTP redirect authentication requirements" with general clause for enforcing authentication and authorization at the target resource.
Clarified behavior of @odata.count when a collection is filtered.
Created standalone "MessageId format" clause.
Removed duplicative text found in the event format table and referenced the message object clauses as needed.
Corrected the response body specified for a PATCH operation containing read-only properties.
Added informative text in the intro to the Data model clause describing the methods for OEM extensions.
Clarified that sensitive data in URIs can be hidden from unauthorized users by returning HTTP 404 Not Found.
Added embedded links to the Location header entry in the response header table.
Corrected $select example in the The $select query parameter clause.
Corrected several embedded links to direct to the correct clause.
1.11.1 2020-08-04 Added missing clause requiring sensitive data to be returned as null.
Clarified that Resolution, Severity, and MessageSeverity in responses can be service-defined and not come from a message registry.
Relaxed schema rules to require description, long description, URI, and capabilities annotations only for schemas published or republished by DMTF.
Added clauses to Schema modification rules to allow for properties, actions, parameters, and URIs to be removed, descriptions to be modified, and pattern and length annotations to be added if not specified.
Relaxed rule for the OData metadata document to not require, but only recommend that all referenced namespaces are included in the document.
Added clause to clarify the usage of empty strings.
Clarified behavior of $skip when the value is greater than or equal to the number of members in a resource collection.
Corrected the minimum value for $top to align with OData.
Clarified behavior of PATCH for partial success scenarios.
Various clarifications and style fixes to the Aggregation clause.
Clarified that HEAD requests shall be rejected when a query parameter is provided.
Removed erroneous requirement for ETags to be strong.
1.11.0 2020-04-30 Added Aggregation clause.
Clarified that services are allowed use HTTP 501 Not Implemented for unsupported HTTP methods.
Clarified the normative semantics around the term "deprecated".
Clarified clauses describing the usage of null for properties versus not reporting a property.
1.10.0 2020-03-27 Restructured the Security details clause for ease of reading. Other than the changes listed below, no other changes were intended. Any clarifications that inadvertently altered the normative behavior are considered errata, and will be corrected in future revisions to the specification.
Deprecated TLS v1.1, and set the minimum TLS requirement to be TLS v1.2 with RFC7525 recommendations.
Deprecated existing cipher suites clause in favor of new clause to leverage IANA recommendations.
Added requirement for supporting the /redfish URI.
Added support for deep operations.
1.9.1 2020-03-27 Deprecated full ISO8601 duration format in favor of a simplified version that does not contain years, months, and weeks.
Added missing normative language for how actions with response bodies are defined in schema.
Added HTTP 201 Created as valid responses for actions.
Clarified the ~ operator for the $expand query parameter to expand hyperlinks found in all Links properties.
Clarified the * and . operators for the $expand query parameter to expand hyperlinks found in payload annotations, such as @Redfish.Settings.
Clarified usage of action parameters that point to resources; the expectation is a reference object pointing to the resource in question is passed by the client.
Clarified that DELETE on a resource likely deletes subordinate resources.
Clarified best practices for naming rules, in particular with regards to acronyms.
Clarified behavior for when individual members of a resource collection cannot be returned as part of a $expand request.
Clarified usage of @Message.ExtendedInfo in error responses and provided guidance for clients for handling error responses.
1.9.0 2019-12-06 Made change to no longer require the Server response header.
Added clause to Schema modification rules to allow for the addition of OEM URIs to standard resources.
Loosened requirements on @odata.type within Oem to not require it in arrays where the type is used repeatedly.
1.8.1 2019-12-06 Made many changes for style consistency, grammar, and general clarity. Except for the following additions, no normative changes were made. Any clarifications that inadvertently altered the normative behavior are considered errata, and will be corrected in future revisions to the Specification.
Clarified SSE with regards to requiring a blank line after each event.
Clarified order of precedence for resolving multiple operation overrides within the Privilege Registry.
Clarified cases for property overrides in the Privilege Registry where multiple objects in the same resource contain the same property name.
Updated references for HTTP Basic authentication to use RFC7617 instead of RFC7235.
Added text/event-stream, application/yaml, and application/vnd.oai.openapi usage to the Accept and Content-Type header table entries.
Added clause that provides guidance on service behavior when null is a property value in POST (create) operations.
Loosened requirements on SSE id based on client usage.
Added documentation for settings, settings apply time, operation apply time, operation apply time support, maintenance window, collection capabilities, requested count, allow over-provisioning, zone affinity, supported certificates, and deprecated terms to the Payload annotations clause.
Added clauses that document responses for actions with a response body defined in schema.
Clarified the allowable values payload annotation to show it can be used for both properties and action parameters.
1.8.0 2019-08-08 Added clause for using /redfish/v1/openapi.yaml as the well-known URI for the OpenAPI document.
Added clause that specifies non-resource reference properties with Uri in the name are accessed using Redfish protocol semantics.
Added SubordinateResources $filter parameter for SSE.
Added Update service clause that describes requirements for the SimpleUpdate action and the MultipartHttpPushUri property.
1.7.1 2019-08-08 Added statements about the owning entity annotation term and its usage in schema modifications.
Clarified SSE id from Id in an event payload and EventId within an event record.
Fixed recommended sequencing of the SSE id to be related to EventId within an event record.
Clarified that services are allowed to close sessions for an account when its password has changed.
Corrected the Password management clause to describe how a user can GET their respective account resources when a password change is required.
Clarified that registries are not required to return @odata.id.
Clarified that services should use HTTP 400 Bad Request for invalid query requests.
Clarified that services should use HTTP 400 Bad Request when the only query is being combined with other query parameters.
Clarified that services should use HTTP 400 Bad Request when query parameters are used on non-GET operations.
Added clause about how to construct enumeration values.
Clarified references to specific messages to also reference their message registry.
Added language about the construction of action names in payloads.
Added informative text for how OEM actions can be defined.
Added guidance for using HTTPS whenever sensitive data is being transmitted.
Added clause restricting the maximum size of an event payload to be 1MiB.
Clarified that auto expanded resource collections can use paging.
Clarified error response format for SSE.
Clarified that charset=utf-8 is not required within the Content-Type header for SSE.
Added clause about how URI patterns are constructed.
Added Excerpt term.
1.7.0 2019-05-16 Made many changes for style consistency, grammar, and general clarity. Except for the following additions, no normative changes were made. Any clarifications that inadvertently altered the normative behavior are considered errata, and will be corrected in future revisions to the Specification.
Added normative statements about how to handle array properties and PATCH operations on arrays.
Separated data model and schema language clauses.
Added clauses that describe how JSON Schema and OpenAPI files are formatted.
Added clause that describes the schema versioning methodology.
Added clause about how URI patterns are constructed based on the resource tree and property hierarchy.
Added dictionary file naming rules and repository locations.
Enhanced localization definitions and defined repository locations.
Added statement about SSE to the Eventing mechanism clause.
Added Constrained composition and Expandable resources clauses to Redfish Composability.
Added clause about requiring event subscriptions to be persistent across service restarts.
Added clause about persistence of tasks generated as a result of using @Redfish.OperationApplyTime across service restarts.
Added clause about using @Redfish.OperationApplyTime and @Redfish.MaintenanceWindow within task responses.
Removed @odata.context property from example payloads.
Added Password management clause to describe functional behavior for restricting access when an account requires a password change.
Added clause around the usage of the HTTP 403 Forbidden status code when an account requires a password change.
1.6.1 2018-12-13 Added clause about percent encoding being allowed for query parameters.
Changed $expand example to use SoftwareInventory instead of LogEntry.
Added clause about the use of a separator for multiple query parameters.
Fixed $filter examples to use / instead of . for property paths.
Clarified the usage of messages in a successful action response; provided an example.
Added clarification about services supporting a subset of HTTP operations on resources specified in schema.
Added clarification about services implementing writable properties as read only.
Added clarification about session termination not affecting connections opened by the session.
Added Redfish Provider term definition.
Updated JSON Schema references to point to Draft 7 of the JSON Schema Specification.
Added clarifications about scenarios for when a request to add an event subscription contains conflicting information and how services respond.
Removed language about ignoring the Links property in PATCH requests.
Clarified usage of ETags to show that a client is not supposed to PATCH @odata.etag when attempting to use ETag protection for a resource.
Clarified usage of the only query parameter to show it's not to be combined with $expand and not to be used with singular resources.
Clarified the usage of the HTTP status codes with task monitors.
Made various spelling and grammar fixes.
1.6.0 2018-08-23 Added methods of using $filter on the SSE URI for the event service.
Added support for the OpenAPI Specification v3.0. This allows OpenAPI-conforming software to access Redfish service implementations. This change might require modification to an implementation to support this version of the specification.
Added strict definitions for the URI patterns used for Redfish resources to support OpenAPI. Each URI is now constructed using a combination of fixed, defined path segments and the values of Id properties for resource collections. Also added restrictions on usage of unsafe characters in URIs. Implementations reporting support for Redfish v1.6.0 conform to these URI patterns.
Added support for creating and naming Redfish schema files in the OpenAPI YAML-based format.
Added URI construction rules for OEM extensions.
Changed ETag usage to require strong ETag format.
Added requirement for HTTP Allow header as a response header for GET and HEAD operations.
Added metric reports as a type of event that can be produced by a Redfish service. Added support for SSE streaming of metric reports in support of new telemetry service.
Added registry, resource, origin, or EventFormatType-based event subscription methods as detailed in the Specification and schema. Added an EventFormatType to enable additional payload types for subscription-based or streaming events. Deprecated EventType-based event subscription mechanism.
Added event message grouping capability.
Provided guidance for defining and using OEM extensions for messages and Message Registries.
Added excerpt and only query parameters.
Clarified requirements for resource collection responses, which includes required properties that were expected, but not listed explicitly in the Specification.
Changed the requirement for the @odata.context annotation to be optional.
Removed requirement for clients to include the OData-Version HTTP header in all requests.
1.5.1 2018-08-10 Added clarifications to required properties in structured properties derived from ReferenceableMembers.
Reorganized Eventing clause to break out the different subscription methods to differentiate pub-sub from SSE.
Removed statements referencing OData conformance levels.
Clarified terminology to explain usage of absolute versus relative reference throughout.
Clarified client-side HTTP Accept header requirements.
Added evaluation order for supported query parameters and clarified examples.
Clarified handling of annotations in response payloads when used with $select queries.
Clarified service handling of annotations in PATCH requests.
Clarified handling of various PATCH request error conditions.
Clarified ability to create resource collection members by POST operations to the resource collection or the Members array within the resource.
Corrected several examples to show required properties in payload.
Clarified usage of the Link header and values of rel=describedBy.
Clarified that the HTTP status code table only describes Redfish-specific behavior and that unless specified, all other usage follows the definitions within the appropriate RFCs.
Added entry for the HTTP 431 Request Header Fields Too Large status code.
Added statement that the HTTP 503 Service Unavailable status code can be used during reboot or reset of a service to indicate that the service is temporarily unavailable.
Clarified usage of the @odata.type annotation within embedded objects.
Added statements about the required Name, Id, and MemberId properties, and the common Description property, which have always been shown as required in schema files, but which the Specification did not mention.
Added guidance for the value of time-date properties when time is unknown.
Added the title property description in actions.
Clarified usage of the @odata.nextLink annotation at the end of resource collections.
Added additional guidance for naming properties and enumeration values that contain "OEM" or that include acronyms.
Corrected requirements for description and long description annotations.
Corrected name of ConfigureComponents in the Operation-to-privilege mapping clause.
Various typographical errors and grammatical improvements.
1.5.0 2018-04-05 Added support for server-sent eventing for streaming events to web-based GUIs or other clients.
Added @Redfish.OperationApplyTime annotation to provide a mechanism for specifying deterministic behavior for the application of Create, Delete or Action (POST) operations.
1.4.1 2018-04-05 Updated name of the forum from SPMF to Redfish Forum.
Consistently used the term, hyperlink.
Added example to clarify usage of $select query parameter with $expand, and clarified expected results when using AutoExpand. Corrected order of precedence for $filter parameter options.
Corrected terminology for OEM-defined actions removing "custom" in favor of OEM, and clarified that the action target property is always required for an action, along with its usage.
Corrected location header values for responses to data modification requests that create a task (Task resource vs. task monitor). Clarified error handling of DELETE operations on Task resources.
Removed references to obsolete and unused Privilege annotation namespace.
Clarified usage of the Base.1.0.GeneralError message in the Base Message Registry.
Added durable URIs for registries and profiles, and clarified intended usage for each folder in the repository. Added file naming conventions for registries and profiles, and clarified file naming for schemas.
Added statement to clarify that additional headers may be added to M-SEARCH responses for SSDP to enable UPnP compatibility.
Clarified assignment requirements for predefined or custom roles when new manager account instances are created, using the RoleId property.
1.4.0 2017-11-17 Added support for optional query parameters ($expand, $filter, and $select) on requests to enable more efficient retrieval of resources or properties from a Redfish service.
Clarified HTTP status and payload responses after successful processing of data modification requests. This includes POST operations to complete actions, and other POST, PATCH, or PUT requests.
Added entries for the HTTP 428 Precondition Required and 507 Insufficient Storage status codes to clarify the proper response to certain error conditions. Added reference links to the HTTP status code table throughout.
Updated the Abstract to reflect the current state of the specification.
Added reference to RFC6585 and clarified expected behavior when ETag support is used in conjunction with PUT or PATCH operations.
Added definition for Property term and updated text to use term consistently.
Added Client requirement column and information for HTTP headers on requests.
Clarified the usage and expected format of the @odata.context property value.
Added clause to describe how to revise structured properties and resolve their definitions in schema.
Added more descriptive definition for the settings resource. Added an example for the SettingsObject. Added description and example for using the @Redfish.SettingsApplyTime annotation.
Added Action example using the ActionInfo resource in addition to the simple @Redfish.AllowableValues example. Updated example to show a proper subset of the available enumerations to reflect a real-world example.
Added statement explaining the updates required to TaskState upon task completion.
1.3.0 2017-08-11 Added support for a service to optionally reject a PATCH or PUT operation if the If-Match or If-Match-None HTTP header is required by returning the HTTP 428 Precondition Required status code.
Added support for a service to describe when the values in the settings object for a resource are applied via the @Redfish.SettingsApplyTime annotation.
1.2.1 2017-08-10 Clarified wording of the Oem object definition.
Clarified wording of the Partial resource results clause.
Clarified behavior of a service when receiving a PATCH with an empty JSON object.
Added statement about other uses of the HTTP 503 Service Unavailable status code.
Clarified format of URI fragments to conform to RFC6901.
Clarified use of absolute and relative URIs.
Clarified definition of the target property as originating from OData.
Clarified distinction between hyperlinks and the links property.
Corrected the JSON example of the privilege map.
Clarified format of the @odata.context property.
Added clauses about the schema file naming conventions.
Clarified behavior of a service when receiving a PUT with missing properties.
Clarified valid values in the Accept header to include wildcards per RFC7231.
Corrected ConfigureUser privilege to be spelled ConfigureUsers.
Corrected the Session login clause to include normative language.
1.2.0 2017-04-14 Added support for the Redfish composability service.
Clarified service handling of the Accept-Encoding header in a request.
Improved consistency and formatting of example requests and responses throughout.
Corrected usage of the @odata.type property in response examples.
Clarified usage of the required annotation.
Clarified usage of SubordinateOverrides in the Privilege Registry.
1.1.0 2016-12-09 Added Redfish service operation-to-privilege mapping clause. This functionality enables a service to present a resource or even property-level mapping of HTTP operations to roles and privileges.
Added references to the Redfish Host Interface Specification (DSP0270).
1.0.5 2016-12-09 Errata release. Various typographical errors.
Corrected the use of collection, resource collection, and members throughout.
Added glossary entries for resource collection and members.
Corrected certificate requirements to reference definitions and requirements in RFC5280 and added a normative reference to RFC5280.
Clarified usage of the HTTP POST and PATCH operations.
Clarified usage of the HTTP status codes and error responses.
1.0.4 2016-08-28 Errata release. Various typographical errors.
Added example of an HTTP Link Header and clarified usage and content.
Added the Schema modification clause, which describes the allowed usage of the schema files.
Added recommendation to use TLS 1.2 or later, and to follow the SNIA TLS Specification. Added reference to the SNIA TLS Specification. Added additional recommended TLS_RSA_WITH_AES_128_CBC_SHA cipher suite.
Clarified that the Id property of a Role resource matches the role name.
1.0.3 2016-06-17 Errata release. Fixed the missing numbering in the table of contents and clauses. Corrected URL references to external specifications. Added missing normative references. Corrected typographical error in ETag example.
Clarified examples for @Message.ExtendedInfo to show arrays of messages.
Clarified that a POST to session service to create a new session does not require authorization headers.
1.0.2 2016-03-31 Errata release. Various typographical errors.
Corrected normative language for M-SEARCH queries and responses.
Corrected Cache-Control and USN format in M-SEARCH responses.
Corrected schema namespace rules to conform to OData namespace requirements and updated examples throughout the document to conform to this format. Specifically, <namespace>.<n>.<n>.<n> becomes <namespace>.v<n>_<n>_<n>. File naming rules for JSON Schema and CSDL (XML) schemas were also corrected to match this format and to enable future major (v2) versions to coexist.
Added clause that details the location of the schema repository and lists the durable URLs for the repository.
Added definition for the value of the Units annotation, using the definitions from the UCUM Specification. Updated examples throughout to use this standardized form.
Modified the naming requirements for Oem property naming to avoid future use of colon : and period . in property names, which can produce invalid or problematic variable names when used in some programming languages or environments. Both separators have been replaced with underscore (_), with colon (:) and period (.) usage now deprecated (but valid).
Removed duplicative or out-of-scope subclauses from the Security clause, which made unintended requirements on Redfish service implementations.
Added the requirement that property names in resource responses match the casing (capitalization) as specified in schema.
Updated normative references to current HTTP RFCs and added clause references throughout the document where applicable.
Clarified ETag header requirements.
Clarified that no authentication is required for accessing the service root.
Clarified description of retrieving resource collections.
Clarified usage of charset=utf-8 in the HTTP Accept and Content-Type headers.
Clarified usage of the Allow HTTP response header and added a table entry for the Retry-After header usage.
Clarified normative usage of the type property and context property, explaining the ability to use two URL forms, and corrected the @odata.context URL examples throughout.
Corrected inconsistent terminology throughout the resource collection response clause.
Corrected name of normative resource Members property (Members, not value).
Clarified that error responses may include information about multiple error conditions.
Corrected name of Measures.Unit annotation term as used in examples.
Corrected outdated reference to Core OData Specification in annotation term examples.
Added the Members property to the Common Redfish resource properties clause.
Clarified terminology and usage of the task monitor and related operations in the Asynchronous operations clause.
Clarified that implementation of the SSDP protocol is optional.
Corrected typographical error in the SSDP USN field's string definition (now ::dmtf-org).
Added the OPTIONS method to the allowed HTTP methods list.
Fixed nullablity in example.
1.0.1 2015-09-17 Errata release. Various grammatical corrections.
Clarified normative use of long description in schema files.
Clarified usage of the rel-describedby Link header.
Corrected text in example of "Select List" in OData context property.
Clarified Accept-Encoding request header handling.
Deleted duplicative and conflicting statement on returning extended error resources.
Clarified relative URI resolution rules.
Clarified USN format.
1.0.0 2015-08-04 Initial release.

18 Bibliography