title: An HTTP Status Code for Indicating Hints abbrev: Early Hints docname: draft-ietf-httpbis-early-hints-latest date: 2017 category: exp
ipr: trust200902 area: General workgroup: HTTP keyword: Internet-Draft
stand_alone: yes pi: [toc, docindent, sortrefs, symrefs, strict, compact, comments, inline]
ins: K. Oku
name: Kazuho Oku
organization: DeNA Co., Ltd.
email: [email protected]
normative: RFC2119: RFC7230: RFC7231: RFC7540:
informative: Preload: title: Preload author: ins: I. Grigorik date: 2016-09-16 target: https://w3c.github.io/preload/
--- abstract
This memo introduces an informational status code for HTTP that can be used for indicating hints to help a client start making preparations for processing the final response.
--- note_Note_to_Readers
Discussion of this draft takes place on the HTTP working group mailing list ([email protected]), which is archived at https://lists.w3.org/Archives/Public/ietf-http-wg/.
Working Group information can be found at https://httpwg.github.io/; source code and issues list for this draft can be found at https://github.com/httpwg/http-extensions/labels/early-hints.
--- middle
Most if not all of the web pages processed by a web browser contain links to external resources that need to be fetched prior to rendering the documents. Therefore, it is beneficial to send such links as early as possible in order to minimize the time spent until the browser becomes possible to render the document. Link header of type "preload" ([Preload]) can be used to indicate such links within the response headers of an HTTP response.
However, it is not always possible for an origin server to send a response immediately after receiving a request. In fact, it is often the contrary. There are many deployments in which an origin server needs to query a database before generating a response. It is also not unusual for an origin server to delegate a request to an upstream HTTP server running at a distant location.
The dilemma here is that even though it is preferable for an origin server to send some headers as soon as it receives a request, it cannot do so until the status code and the headers of the final HTTP response is determined.
HTTP/2 ([RFC7540]) push can be used as a solution to the issue, but has its own limitations. The resources that can be pushed using HTTP/2 are limited to those belonging to the same origin. Also, it is impossible to send only the links of the resources using HTTP/2 push. Sending HTTP responses for every resource is an inefficient way of using bandwidth, especially when a caching server exists as an intermediary.
This memo defines a status code for sending an informational response ([RFC7231], section 6.2) that contains headers that are likely to be included in the final response. A server can send the informational response containing some of the headers to help the client start making preparations for processing the final response, and then run time-consuming operations to generate the final response. The informational response can also be used by an origin server to trigger HTTP/2 push at an caching intermediary.
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 [RFC2119].
This informational status code indicates the client that the server is likely to send a final response with the headers included in the informational response.
A server MUST NOT include Content-Length, Transfer-Encoding, or any hop-by-hop headers ([RFC7230], section 6.1) in the informational response using the status code.
A client MAY speculatively evaluate the headers included in the informational response while waiting for the final response. For example, a client may recognize the link header of type preload and start fetching the resource. However, the evaluation MUST NOT affect how the final response is processed; the client must behave as if it had not seen the informational response. A client MUST NOT process the headers included in the response as if they belonged to the informational response.
An intermediary MAY drop the informational response. It MAY send HTTP/2 ([RFC7540]) push responses using the information found in the informational response.
Clients may have issues handling Early Hints, since informational response is rarely used for requests not including an Expect header ([RFC7231], section 5.1.1).
An HTTP/1.1 client that mishandles the informational response as a final response is likely to consider all the responses to the succeeding requests sent over the same connection to be part of the final response. Such behavior may constitute a cross-origin information disclosure vulnerability in case the client multiplexes requests to different origins onto a single persistent connection.
Therefore, a server might refrain from sending Early Hints over HTTP/1.1 unless when the client is known to handle informational responses correctly.
HTTP/2 clients are less likely to suffer from incorrect framing since handling of the response headers does not affect how the end of the response body is determined.
If Early Hints is standardized, the HTTP Status Codes Registry should be updated with the following entries:
- Code: 103
- Description: Early Hints
- Specification: this document
Thanks to Tatsuhiro Tsujikawa for coming up with the idea of sending the link headers using an informational response.
- None yet.
- Forbid processing the headers of a 103 response as part of the informational response.