HTTP Client HintsGoogleilya@igvita.comhttps://www.igvita.com/Googleyoav@yoav.wshttps://blog.yoav.ws/
Applications and Real-Time
HTTPContent NegotiationHTTP defines proactive content negotiation to allow servers to select the appropriate response for a given request, based upon the user agent’s characteristics, as expressed in request headers. In practice, clients are often unwilling to send those request headers, because it is not clear whether they will be used, and sending them impacts both performance and privacy.This document defines an Accept-CH response header that servers can use to advertise their use of request headers for proactive content negotiation, along with a set of guidelines for the creation of such headers, colloquially known as “Client Hints.”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. Applications that want to allow the server to optimize content delivery and user experience based on such capabilities have, historically, had to rely on passive identification (e.g., by matching User-Agent () header field against an established database of client signatures), used HTTP cookies and URL parameters, or use some combination of these and similar mechanisms to enable ad hoc content negotiation.Such techniques are expensive to setup and maintain, are not portable across both applications and servers, and make it hard to reason for both client and server about which data is required and is in use during the negotiation:User agent detection cannot reliably identify all static variables, cannot infer dynamic client preferences, requires external device database, is not cache friendly, and is reliant on a passive fingerprinting surface.Cookie based approaches are not portable across applications and servers, impose additional client-side latency by requiring JavaScript execution, and are not cache friendly.URL parameters, similar to cookie based approaches, suffer from lack of portability, and are hard to deploy due to a requirement to encode content negotiation data inside of the URL of each resource.Proactive content negotiation () offers an alternative approach; user agents use specified, well-defined request headers to advertise their capabilities and characteristics, so that servers can select (or formulate) an appropriate response.However, proactive content negotiation requires clients to send these request headers prolifically. This causes performance concerns (because it creates “bloat” in requests), as well as privacy issues; passively providing such information allows servers to silently fingerprint the user agent.This document defines a new response header, Accept-CH, that allows an origin server to explicitly ask that clients send these headers in requests. It also defines guidelines for content negotiation mechanisms that use it, colloquially referred to as Client Hints.Client Hints mitigate the performance concerns by assuring that clients will only send the request headers when they’re actually going to be used, and the privacy concerns of passive fingerprinting by requiring explicit opt-in and disclosure of required headers by the server through the use of the Accept-CH response header.This document defines the Client Hints infrastructure, a framework that enables servers to opt-in to specific proactive content negotiation features, which will enable them to adapt their content accordingly. However, it does not define any specific features that will use that infrastructure. Those features will be defined in their respective specifications.The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”, “SHOULD NOT”, “RECOMMENDED”, “NOT RECOMMENDED”, “MAY”, and “OPTIONAL” in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here.This document uses the Augmented Backus-Naur Form (ABNF) notation of with the list rule extension defined in . 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 server preferences. The client and server can use an opt-in mechanism outlined below to negotiate which fields should be sent to allow for efficient content adaption, and optionally use additional mechanisms to negotiate delegation policies that control access of third parties to same fields.Implementers should be aware of the passive fingerprinting implications when implementing support for Client Hints, and follow the considerations outlined in “Security Considerations” section of this document.When presented with a request that contains one or more client hint header fields, 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 () 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.Servers can advertise support for Client Hints using the mechnisms described below.The Accept-CH response header field or the equivalent HTML meta element with http-equiv attribute () indicate server support for particular hints indicated in its value.Accept-CH is a Structured Header . Its value MUST be an sh-list () whose members are tokens (). Its ABNF is:For example:When a client receives an HTTP response advertising support for provided list of Clients Hints, it SHOULD process it as origin () opt-in to receive Client Hint header fields advertised in the field-value, for subsequent same-origin requests.The opt-in MUST be delivered over a secure transport.The opt-in SHOULD be persisted and bound to the origin to enable delivery of Client Hints on subsequent requests to the server’s origin, and MUST NOT be persisted for an origin that isn’t HTTPS.For example, based on the Accept-CH example above, which is received in response to a user agent navigating to “https://example.com”, and delivered over a secure transport: a user agent SHOULD persist an Accept-CH preference bound to “https://example.com” and use it for user agent navigations to “https://example.com” and any same-origin resource requests initiated by the page constructed from the navigation’s response. This preference SHOULD NOT extend to resource requests initiated to “https://example.com” from other origins.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 Sec-CH-Example header field.Above example indicates that the cache key needs to include the Sec-CH-Example and Sec-CH-Example-2 header fields.The request header fields defined in this document, and those that extend it, expose information about the user’s environment to enable proactive content negotiation. Such information may reveal new information about the user and implementers ought to consider the following considerations, recommendations, and best practices.Transmitted Client Hints header fields SHOULD NOT provide new information that is otherwise not available to the application via other means, such as using HTML, CSS, or JavaScript. Further, sending highly granular data, such as image and viewport width may help identify users across multiple requests. Reducing the set of field values that can be expressed, or restricting them to an enumerated range where the advertised value is close but is not an exact representation of the current value, can improve privacy and reduce risk of linkability by ensuring that the same value is sent by multiple users. However, such precautions can still be insufficient for some types of data, especially data that can change over time.Implementers ought to consider both user and server controlled mechanisms and policies to control which Client Hints header fields are advertised:Implementers SHOULD restrict delivery of some or all Client Hints header fields to the opt-in origin only, unless the opt-in origin has explicitly delegated permission to another origin to request Client Hints header fields.Implementers MAY provide user choice mechanisms so that users may balance privacy concerns with bandwidth limitations. However, implementers should also be aware that explaining the privacy implications of passive fingerprinting to users may be challenging.Implementations specific to certain use cases or threat models MAY avoid transmitting some or all of Client Hints header fields. For example, avoid transmission of header fields that can carry higher risks of linkability.Implementers SHOULD support Client Hints opt-in mechanisms and MUST clear persisted opt-in preferences when any one of site data, browsing history, browsing cache, or similar, are cleared.This document defines the “Accept-CH” HTTP response field, and registers it 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 HintsAugmented BNF for Syntax Specifications: ABNFHypertext Transfer Protocol (HTTP/1.1): Message Syntax and RoutingHypertext Transfer Protocol (HTTP/1.1): Semantics and ContentHypertext Transfer Protocol (HTTP/1.1): CachingThe Web Origin ConceptHTML5Key words for use in RFCs to Indicate Requirement LevelsAmbiguity of Uppercase vs Lowercase in RFC 2119 Key WordsStructured Headers for HTTPHTTP State Management MechanismThe Key HTTP Response Header FieldHTTP Representation VariantsClient Hints may be combined with Variants response header field to enable fine-grained control of the cache key for improved cache efficiency. Features that define Client Hints will need to specify the related variants algorithms as described in .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-LifetimeIssue 361: Removed DownlinkIssue 361: Moved Key to appendix, plus other editorial feedbackIssue 372: Scoped CH opt-in and delivery to secure transportsIssue 373: Bind CH opt-in to originIssue 524: Save-Data is now defined by NetInfo spec, droppingPR 775: Removed specific features to be defined in other specificationsIssue 761: Clarified that the defined headers are response headers.Issue 730: Replaced Key reference with Variants.Issue 700: Replaced ABNF with structured headers.PR 878: Removed Accept-CH-Lifetime based on feedback at IETF 105Thanks to Mark Nottingham, Julian Reschke, Chris Bentzel, Ben Greenstein, Tarun Bansal, Roy Fielding, Vasiliy Faronov, Ted Hardie, Jonas Sicking, Martin Thomson, and numerous other members of the IETF HTTP Working Group for invaluable help and feedback.