HTTP Client HintsGoogleilya@igvita.comhttps://www.igvita.com/
Applications and Real-Time
HTTPContent NegotiationAn increasing diversity of Web-connected devices and software capabilities has created a need to deliver optimized content for each device.This specification defines a set of HTTP request header fields, colloquially known as Client Hints, to address this. They are intended to be used as input to proactive content negotiation; just as the Accept header field allows user agents to indicate what formats they prefer, Client Hints allow user agents to indicate device and agent specific preferences.Discussion of this draft takes place on the HTTP working group mailing list
(ietf-http-wg@w3.org), which is archived at https://lists.w3.org/Archives/Public/ietf-http-wg/.Working Group information can be found at http://httpwg.github.io/; source
code and issues list for this draft can be found at https://github.com/httpwg/http-extensions/labels/client-hints.There are thousands of different devices accessing the web, each with different device capabilities and preference information. These device capabilities include hardware and software characteristics, as well as dynamic user and client preferences.One way to infer some of these capabilities is through User-Agent (Section 5.5.3 of ) header field detection against an established database of client signatures. However, this technique requires acquiring such a database, integrating it into the serving path, and keeping it up to date. However, even once this infrastructure is deployed, user agent sniffing has numerous limitations:User agent detection cannot reliably identify all static variablesUser agent detection cannot infer any dynamic client preferencesUser agent detection requires an external device databaseUser agent detection is not cache friendlyA popular alternative strategy is to use HTTP cookies () to communicate some information about the user agent. However, this approach is also not cache friendly, bound by same origin policy, and imposes additional client-side latency by requiring JavaScript execution to create and manage HTTP cookies.This document defines a set of new request header fields that allow user agent to perform proactive content negotiation (Section 3.4.1 of ) by indicating device and agent specific preferences, through a mechanism similar to the Accept header field which is used to indicate preferred response formats.Client Hints does not supersede or replace the User-Agent header field. Existing device detection mechanisms can continue to use both mechanisms if necessary. By advertising its capabilities within a request header field, Client Hints allows for cache friendly and proactive content negotiation.The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in .This document uses the Augmented Backus-Naur Form (ABNF) notation of with the list rule extension defined in , Appendix B. It includes by reference the DIGIT rule from and the OWS and field-name rules from .A Client Hint request header field is a HTTP header field that is used by HTTP clients to indicate configuration data that can be used by the server to select an appropriate response. Each one conveys client preferences that the server can use to adapt and optimize the response.Clients control which Client Hints are sent in requests, based on their default settings, user configuration and/or preferences. Implementers might provide user choice mechanisms so that users may balance privacy concerns with bandwidth limitations. Implementations specific to certain use cases or threat models might avoid transmitting these headers altogether, or limit them to secure contexts or authenticated sessions. Implementers should be aware that explaining the privacy implications of passive fingerprinting or network information disclosure may be challenging.The client and server, or an intermediate proxy, can use an opt-in mechanism to negotiate which fields should be reported to allow for efficient content adaption.When presented with a request that contains one or more client hint headers, servers can optimize the response based upon the information in them. When doing so, and if the resource is cacheable, the server MUST also generate a Vary response header field (Section 7.1.4 of ), and optionally Key (), to indicate which hints can affect the selected response and whether the selected response is appropriate for a later request.Further, depending on the hint used, the server can generate additional response header fields to convey related values to aid client processing. For example, this specification defines “Content-DPR” response header field that needs to be returned by the server when the “DPR” hint is used to select the response.Servers can advertise support for Client Hints using the Accept-CH header field or an equivalent HTML meta element with http-equiv attribute ().For example:When a client receives Accept-CH, or if it is capable of processing the HTML response and finds an equivalent HTML meta element, it can treat it as a signal that the application is interested in receiving specified request header fields that match the advertised field-values; subresource requests initiated as a result of processing the response from the server that includes the Accept-CH opt-in can include the request header fields that match the advertised field-values.For example, based on Accept-CH example above, a user agent could append DPR, Width, and Viewport-Width header fields to all subresource requests initiated by the page constructed from the response.Servers can ask the client to remember an origin-wide Accept-CH preference for a specified period of time to enable delivery of Client Hints on all subsequent requests to the origin, and on subresource requests initiated as a result of processing a response from the origin.The field-value indicates that the Accept-CH preference should be considered stale after its age is greater than the specified number of seconds.For example, based on the Accept-CH and Accept-CH-Lifetime example above, a user agent could persist an origin-wide Accept-CH preference for up to 86400 seconds (1 day). Then, if a request is initiated to the same origin before the preference is stale (e.g. as a result of a navigation to the origin, or fetching a subresource from the origin) the client could append the requested header fields (DPR and Viewport-Width in this example) to the request and any subresource requests initiated as a result of processing a response from same origin.When selecting an optimized response based on one or more Client Hints, and if the resource is cacheable, the server needs to generate a Vary response header field () to indicate which hints can affect the selected response and whether the selected response is appropriate for a later request.Above example indicates that the cache key needs to include the DPR header field.Above example indicates that the cache key needs to include the DPR, Width, and Downlink header fields.Client Hints MAY be combined with Key () to enable fine-grained control of the cache key for improved cache efficiency. For example, the server can return the following set of instructions:Above example indicates that the cache key needs to include the value of the DPR header field with three segments: less than 1.5, 1.5 to less than 2.5, and 4.0 or greater.Above example indicates that the cache key needs to include the value of the Width header field and be partitioned into groups of 320: 0-320, 320-640, and so on.Above example indicates that the cache key needs to include the (Mbps) value of the Downlink header field with six segments: less than 0.5, 0.5 to less than 1.0, 1.0 to less than 3.0, 3.0 to less than 5.0, 5.0 to less than 10; 10 or higher.The “DPR” request header field is a number that indicates the client’s current Device Pixel Ratio (DPR), which is the ratio of physical pixels over CSS px (Section 5.2 of ) of the layout viewport (Section 9.1.1 of ) on the device.If DPR occurs in a message more than once, the last value overrides all previous occurrences.The “Content-DPR” response header field is a number that indicates the ratio between physical pixels over CSS px of the selected image response.DPR ratio affects the calculation of intrinsic size of image resources on the client - i.e. typically, the client automatically scales the natural size of the image by the DPR ratio to derive its display dimensions. As a result, the server MUST explicitly indicate the DPR of the selected image response whenever the DPR hint is used, and the client MUST use the DPR value returned by the server to perform its calculations. In case the server returned Content-DPR value contradicts previous client-side DPR indication, the server returned value MUST take precedence.Note that DPR confirmation is only required for image responses, and the server does not need to confirm the resource width as this value can be derived from the resource itself once it is decoded by the client.If Content-DPR occurs in a message more than once, the last value overrides all previous occurrences.The “Width” request header field is a number that indicates the desired resource width in physical px (i.e. intrinsic size of an image). The provided physical px value is a number rounded to the smallest following integer (i.e. ceiling value).If the desired resource width is not known at the time of the request or the resource does not have a display width, the Width header field can be omitted. If Width occurs in a message more than once, the last value overrides all previous occurrences.The “Viewport-Width” request header field is a number that indicates the layout viewport width in CSS px. The provided CSS px value is a number rounded to the smallest following integer (i.e. ceiling value).If Viewport-Width occurs in a message more than once, the last value overrides all previous occurrences.The “Downlink” request header field is a number that indicates the client’s maximum downlink speed in megabits per second (Mbps).If Downlink occurs in a message more than once, the minimum value should be used to override other occurrences.The “Save-Data” request header field consists of one or more tokens that indicate client’s preference for reduced data usage, due to high transfer costs, slow connection speeds, or other reasons.This document defines the “on” sd-token value, which is used as a signal indicating explicit user opt-in into a reduced data usage mode on the client, and when communicated to origins allows them to deliver alternate content honoring such preference - e.g. smaller image and video resources, alternate markup, and so on. New token and extension token values can be defined by updates to this specification.For example, given the following request header fields:The server knows that the device pixel ratio is 2.0, that the intended display width of the requested resource is 160 CSS px (320 physical pixels at 2x resolution), and that the viewport width is 320 CSS px.If the server uses above hints to perform resource selection for an image asset, it must confirm its selection via the Content-DPR response header to allow the client to calculate the appropriate intrinsic size of the image response. The server does not need to confirm resource width, only the ratio between physical pixels and CSS px of the selected image resource:The Content-DPR response header field indicates to the client that the server has selected resource with DPR ratio of 1.0. The client can use this information to perform additional processing on the resource - for example, calculate the appropriate intrinsic size of the image resource such that it is displayed at the correct resolution.Alternatively, the server could select an alternate resource based on the maximum downlink speed advertised in the request header fields:The server knows that the client’s maximum downlink speed is 0.384Mbps (GPRS EDGE), and it can use this information to select an optimized resource - for example, an alternate image asset, stylesheet, HTML document, media stream, and so on.The request header fields defined in this specification expose information that is already available to Web applications in the browser runtime itself (e.g., using JavaScript and CSS). For example, the application can obtain viewport width, image display width, and device pixel ratio via JavaScript, or through the use of CSS media queries and unique resource URLs even if JavaScript is disabled. However, servers that gather this information through such mechanisms are typically observable (e.g., you can see that they’re using JavaScript to gather it), whereas servers’ use of the header fields introduced by this specification is not observable. Section 2.1 discusses potential mitigations.For example, sending Client Hints on all requests can make information about the user’s environment available to origins that otherwise did not have access to this data, which may or may not be the desired outcome - e.g. this may enable an image optimization service to deliver a tailored asset, and it may reveal same information about the user to other origins that may not have had access to it before. Similarly, sending highly granular data, such as image and viewport width may help identify users across multiple requests. Restricting such field values to an enumerated range, where the user agent advertises a threshold value that is close but is not an exact representation of the current value, can help mitigate the risk of such fingerprinting.Implementers ought to provide mechanisms and policies to control how and when such hints are advertised. For example, they could require origin opt-in via Accept-CH; clear remembered opt-in, as set by Accept-CH-Lifetime, when site data, browsing history, browsing cache, or similar, are cleared; restrict delivery to same origin subrequests; limit delivery to requests that already carry identifying information (e.g. cookies); modify delivery policy when in an “incognito” or a similar privacy mode; enable user configuration and opt in, and so on.This document defines the “Accept-CH”, “DPR”, “Width”, and “Downlink” HTTP request fields, “Content-DPR” HTTP response field, and registers them in the Permanent Message Header Fields registry.Header field name: Accept-CHApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: Accept-CH-LifetimeApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: Content-DPRApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: DownlinkApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: DPRApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: Save-DataApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: Viewport-WidthApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsHeader field name: WidthApplicable protocol: HTTPStatus: standardAuthor/Change controller: IETFSpecification document(s): of this documentRelated information: for Client HintsKey words for use in RFCs to Indicate Requirement LevelsIn many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements.Augmented BNF for Syntax Specifications: ABNFInternet technical specifications often need to define a formal syntax. Over the years, a modified version of Backus-Naur Form (BNF), called Augmented BNF (ABNF), has been popular among many Internet specifications. The current specification documents ABNF. It balances compactness and simplicity with reasonable representational power. The differences between standard BNF and ABNF involve naming rules, repetition, alternatives, order-independence, and value ranges. This specification also supplies additional rule definitions and encoding for a core lexical analyzer of the type common to several Internet specifications. [STANDARDS-TRACK]Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and RoutingThe Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document provides an overview of HTTP architecture and its associated terminology, defines the "http" and "https" Uniform Resource Identifier (URI) schemes, defines the HTTP/1.1 message syntax and parsing requirements, and describes related security concerns for implementations.Hypertext Transfer Protocol (HTTP/1.1): Semantics and ContentThe Hypertext Transfer Protocol (HTTP) is a stateless \%application- level protocol for distributed, collaborative, hypertext information systems. This document defines the semantics of HTTP/1.1 messages, as expressed by request methods, request header fields, response status codes, and response header fields, along with the payload of messages (metadata and body content) and mechanisms for content negotiation.Hypertext Transfer Protocol (HTTP/1.1): CachingThe Hypertext Transfer Protocol (HTTP) is a stateless \%application- level protocol for distributed, collaborative, hypertext information systems. This document defines HTTP caches and the associated header fields that control cache behavior or indicate cacheable response messages.HTML5CSS Values and Units Module Level 3Cascading Style Sheets Level 2 Revision 1 (CSS 2.1) SpecificationHTTP State Management MechanismThis document defines the HTTP Cookie and Set-Cookie header fields. These header fields can be used by HTTP servers to store state (called cookies) at HTTP user agents, letting the servers maintain a stateful session over the mostly stateless HTTP protocol. Although cookies have many historical infelicities that degrade their security and privacy, the Cookie and Set-Cookie header fields are widely used on the Internet. This document obsoletes RFC 2965. [STANDARDS-TRACK]The Key HTTP Response Header FieldThe 'Key' header field for HTTP responses allows an origin server to describe the secondary cache key (RFC 7234, Section 4.1) for a resource, by conveying what is effectively a short algorithm that can be used upon later requests to determine if a stored response is reusable for a given request. Key has the advantage of avoiding an additional round trip for validation whenever a new request differs slightly, but not significantly, from prior requests. Key also informs user agents of the request characteristics that might result in different content, which can be useful if the user agent is not sending request header fields in order to reduce the risk of fingerprinting.Issue 168 (make Save-Data extensible) updated ABNF.Issue 163 (CH review feedback) editorial feedback from httpwg list.Issue 153 (NetInfo API citation) added normative reference.Issue 200: Moved Key reference to informative.Issue 215: Extended passive fingerprinting and mitigation considerations.Changed document status to experimental.Issue 239: Updated reference to CR-css-values-3Issue 240: Updated reference for Network Information APIIssue 241: Consistency in IANA considerationsIssue 250: Clarified Accept-CHIssue 284: Extended guidance for Accept-CHIssue 308: Editorial cleanupIssue 306: Define Accept-CH-LifetimeNone yet.