source: draft-ietf-httpbis/latest/p1-messaging.xml @ 2116

Last change on this file since 2116 was 2116, checked in by fielding@…, 10 years ago

rewrite the sections on retrying requests and pipelining to resolve nonsense about non-idempotent sequences; partly addresses #426

  • Property svn:eol-style set to native
  • Property svn:mime-type set to text/xml
File size: 223.6 KB
1<?xml version="1.0" encoding="utf-8"?>
2<?xml-stylesheet type='text/xsl' href='../myxml2rfc.xslt'?>
3<!DOCTYPE rfc [
4  <!ENTITY MAY "<bcp14 xmlns=''>MAY</bcp14>">
5  <!ENTITY MUST "<bcp14 xmlns=''>MUST</bcp14>">
6  <!ENTITY MUST-NOT "<bcp14 xmlns=''>MUST NOT</bcp14>">
7  <!ENTITY OPTIONAL "<bcp14 xmlns=''>OPTIONAL</bcp14>">
8  <!ENTITY RECOMMENDED "<bcp14 xmlns=''>RECOMMENDED</bcp14>">
9  <!ENTITY REQUIRED "<bcp14 xmlns=''>REQUIRED</bcp14>">
10  <!ENTITY SHALL "<bcp14 xmlns=''>SHALL</bcp14>">
11  <!ENTITY SHALL-NOT "<bcp14 xmlns=''>SHALL NOT</bcp14>">
12  <!ENTITY SHOULD "<bcp14 xmlns=''>SHOULD</bcp14>">
13  <!ENTITY SHOULD-NOT "<bcp14 xmlns=''>SHOULD NOT</bcp14>">
14  <!ENTITY ID-VERSION "latest">
15  <!ENTITY ID-MONTH "January">
16  <!ENTITY ID-YEAR "2013">
17  <!ENTITY mdash "&#8212;">
18  <!ENTITY Note "<x:h xmlns:x=''>Note:</x:h>">
19  <!ENTITY caching-overview       "<xref target='Part6' x:rel='#caching.overview' xmlns:x=''/>">
20  <!ENTITY cache-incomplete       "<xref target='Part6' x:rel='#response.cacheability' xmlns:x=''/>">
21  <!ENTITY payload                "<xref target='Part2' x:rel='#payload' xmlns:x=''/>">
22  <!ENTITY media-type            "<xref target='Part2' x:rel='#media.type' xmlns:x=''/>">
23  <!ENTITY content-codings        "<xref target='Part2' x:rel='#content.codings' xmlns:x=''/>">
24  <!ENTITY CONNECT                "<xref target='Part2' x:rel='#CONNECT' xmlns:x=''/>">
25  <!ENTITY content.negotiation    "<xref target='Part2' x:rel='#content.negotiation' xmlns:x=''/>">
26  <!ENTITY diff-mime              "<xref target='Part2' x:rel='#differences.between.http.and.mime' xmlns:x=''/>">
27  <!ENTITY representation         "<xref target='Part2' x:rel='#representations' xmlns:x=''/>">
28  <!ENTITY HEAD                   "<xref target='Part2' x:rel='#HEAD' xmlns:x=''/>">
29  <!ENTITY header-allow           "<xref target='Part2' x:rel='#header.allow' xmlns:x=''/>">
30  <!ENTITY header-cache-control   "<xref target='Part6' x:rel='#header.cache-control' xmlns:x=''/>">
31  <!ENTITY header-content-encoding    "<xref target='Part2' x:rel='#header.content-encoding' xmlns:x=''/>">
32  <!ENTITY header-content-location    "<xref target='Part2' x:rel='#header.content-location' xmlns:x=''/>">
33  <!ENTITY header-content-range   "<xref target='Part5' x:rel='#header.content-range' xmlns:x=''/>">
34  <!ENTITY header-content-type    "<xref target='Part2' x:rel='#header.content-type' xmlns:x=''/>">
35  <!ENTITY header-date            "<xref target='Part2' x:rel='' xmlns:x=''/>">
36  <!ENTITY header-etag            "<xref target='Part4' x:rel='#header.etag' xmlns:x=''/>">
37  <!ENTITY header-expires         "<xref target='Part6' x:rel='#header.expires' xmlns:x=''/>">
38  <!ENTITY header-last-modified   "<xref target='Part4' x:rel='#header.last-modified' xmlns:x=''/>">
39  <!ENTITY header-mime-version    "<xref target='Part2' x:rel='#mime-version' xmlns:x=''/>">
40  <!ENTITY header-pragma          "<xref target='Part6' x:rel='#header.pragma' xmlns:x=''/>">
41  <!ENTITY header-proxy-authenticate  "<xref target='Part7' x:rel='#header.proxy-authenticate' xmlns:x=''/>">
42  <!ENTITY header-proxy-authorization "<xref target='Part7' x:rel='#header.proxy-authorization' xmlns:x=''/>">
43  <!ENTITY header-server          "<xref target='Part2' x:rel='#header.server' xmlns:x=''/>">
44  <!ENTITY header-warning         "<xref target='Part6' x:rel='#header.warning' xmlns:x=''/>">
45  <!ENTITY idempotent-methods     "<xref target='Part2' x:rel='#idempotent.methods' xmlns:x=''/>">
46  <!ENTITY safe-methods           "<xref target='Part2' x:rel='#safe.methods' xmlns:x=''/>">
47  <!ENTITY methods                "<xref target='Part2' x:rel='#methods' xmlns:x=''/>">
48  <!ENTITY OPTIONS                "<xref target='Part2' x:rel='#OPTIONS' xmlns:x=''/>">
49  <!ENTITY qvalue                 "<xref target='Part2' x:rel='#quality.values' xmlns:x=''/>">
50  <!ENTITY resource               "<xref target='Part2' x:rel='#resources' xmlns:x=''/>">
51  <!ENTITY status-codes           "<xref target='Part2' x:rel='' xmlns:x=''/>">
52  <!ENTITY status-1xx             "<xref target='Part2' x:rel='#status.1xx' xmlns:x=''/>">
53  <!ENTITY status-203             "<xref target='Part2' x:rel='#status.203' xmlns:x=''/>">
54  <!ENTITY status-3xx             "<xref target='Part2' x:rel='#status.3xx' xmlns:x=''/>">
55  <!ENTITY status-304             "<xref target='Part4' x:rel='#status.304' xmlns:x=''/>">
56  <!ENTITY status-4xx             "<xref target='Part2' x:rel='#status.4xx' xmlns:x=''/>">
57  <!ENTITY status-414             "<xref target='Part2' x:rel='#status.414' xmlns:x=''/>">
58  <!ENTITY iana-header-registry   "<xref target='Part2' x:rel='#header.field.registry' xmlns:x=''/>">
60<?rfc toc="yes" ?>
61<?rfc symrefs="yes" ?>
62<?rfc sortrefs="yes" ?>
63<?rfc compact="yes"?>
64<?rfc subcompact="no" ?>
65<?rfc linkmailto="no" ?>
66<?rfc editing="no" ?>
67<?rfc comments="yes"?>
68<?rfc inline="yes"?>
69<?rfc rfcedstyle="yes"?>
70<?rfc-ext allow-markup-in-artwork="yes" ?>
71<?rfc-ext include-references-in-index="yes" ?>
72<rfc obsoletes="2145,2616" updates="2817,2818" category="std" x:maturity-level="proposed"
73     ipr="pre5378Trust200902" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
74     xmlns:x=''>
75<x:link rel="next" basename="p2-semantics"/>
76<x:feedback template="{docname},%20%22{section}%22&amp;body=&lt;{ref}&gt;:"/>
79  <title abbrev="HTTP/1.1 Message Syntax and Routing">Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing</title>
81  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
82    <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
83    <address>
84      <postal>
85        <street>345 Park Ave</street>
86        <city>San Jose</city>
87        <region>CA</region>
88        <code>95110</code>
89        <country>USA</country>
90      </postal>
91      <email></email>
92      <uri></uri>
93    </address>
94  </author>
96  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
97    <organization abbrev="greenbytes">greenbytes GmbH</organization>
98    <address>
99      <postal>
100        <street>Hafenweg 16</street>
101        <city>Muenster</city><region>NW</region><code>48155</code>
102        <country>Germany</country>
103      </postal>
104      <email></email>
105      <uri></uri>
106    </address>
107  </author>
109  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
110  <workgroup>HTTPbis Working Group</workgroup>
114   The Hypertext Transfer Protocol (HTTP) is an application-level protocol for
115   distributed, collaborative, hypertext information systems. HTTP has been in
116   use by the World Wide Web global information initiative since 1990.
117   This document provides an overview of HTTP architecture and its associated
118   terminology, defines the "http" and "https" Uniform Resource Identifier
119   (URI) schemes, defines the HTTP/1.1 message syntax and parsing requirements,
120   and describes general security concerns for implementations.
124<note title="Editorial Note (To be removed by RFC Editor)">
125  <t>
126    Discussion of this draft takes place on the HTTPBIS working group
127    mailing list (, which is archived at
128    <eref target=""/>.
129  </t>
130  <t>
131    The current issues list is at
132    <eref target=""/> and related
133    documents (including fancy diffs) can be found at
134    <eref target=""/>.
135  </t>
136  <t>
137    The changes in this draft are summarized in <xref target="changes.since.21"/>.
138  </t>
142<section title="Introduction" anchor="introduction">
144   The Hypertext Transfer Protocol (HTTP) is an application-level
145   request/response protocol that uses extensible semantics and self-descriptive
146   message payloads for flexible interaction with network-based hypertext
147   information systems. This document is the first in a series of documents
148   that collectively form the HTTP/1.1 specification:
149   <list style="empty">
150    <t>RFC xxx1: Message Syntax and Routing</t>
151    <t><xref target="Part2" x:fmt="none">RFC xxx2</xref>: Semantics and Content</t>
152    <t><xref target="Part4" x:fmt="none">RFC xxx3</xref>: Conditional Requests</t>
153    <t><xref target="Part5" x:fmt="none">RFC xxx4</xref>: Range Requests</t>
154    <t><xref target="Part6" x:fmt="none">RFC xxx5</xref>: Caching</t>
155    <t><xref target="Part7" x:fmt="none">RFC xxx6</xref>: Authentication</t>
156   </list>
159   This HTTP/1.1 specification obsoletes and moves to historic status
160   <xref target="RFC2616" x:fmt="none">RFC 2616</xref>, its predecessor
161   <xref target="RFC2068" x:fmt="none">RFC 2068</xref>, and
162   <xref target="RFC2145" x:fmt="none">RFC 2145</xref> (on HTTP versioning).
163   This specification also updates the use of CONNECT to establish a tunnel,
164   previously defined in <xref target="RFC2817" x:fmt="none">RFC 2817</xref>,
165   and defines the "https" URI scheme that was described informally in
166   <xref target="RFC2818" x:fmt="none">RFC 2818</xref>.
169   HTTP is a generic interface protocol for information systems. It is
170   designed to hide the details of how a service is implemented by presenting
171   a uniform interface to clients that is independent of the types of
172   resources provided. Likewise, servers do not need to be aware of each
173   client's purpose: an HTTP request can be considered in isolation rather
174   than being associated with a specific type of client or a predetermined
175   sequence of application steps. The result is a protocol that can be used
176   effectively in many different contexts and for which implementations can
177   evolve independently over time.
180   HTTP is also designed for use as an intermediation protocol for translating
181   communication to and from non-HTTP information systems.
182   HTTP proxies and gateways can provide access to alternative information
183   services by translating their diverse protocols into a hypertext
184   format that can be viewed and manipulated by clients in the same way
185   as HTTP services.
188   One consequence of this flexibility is that the protocol cannot be
189   defined in terms of what occurs behind the interface. Instead, we
190   are limited to defining the syntax of communication, the intent
191   of received communication, and the expected behavior of recipients.
192   If the communication is considered in isolation, then successful
193   actions ought to be reflected in corresponding changes to the
194   observable interface provided by servers. However, since multiple
195   clients might act in parallel and perhaps at cross-purposes, we
196   cannot require that such changes be observable beyond the scope
197   of a single response.
200   This document describes the architectural elements that are used or
201   referred to in HTTP, defines the "http" and "https" URI schemes,
202   describes overall network operation and connection management,
203   and defines HTTP message framing and forwarding requirements.
204   Our goal is to define all of the mechanisms necessary for HTTP message
205   handling that are independent of message semantics, thereby defining the
206   complete set of requirements for message parsers and
207   message-forwarding intermediaries.
211<section title="Requirement Notation" anchor="intro.requirements">
213   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
214   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
215   document are to be interpreted as described in <xref target="RFC2119"/>.
218   Conformance criteria and considerations regarding error handling
219   are defined in <xref target="conformance"/>.
223<section title="Syntax Notation" anchor="notation">
224<iref primary="true" item="Grammar" subitem="ALPHA"/>
225<iref primary="true" item="Grammar" subitem="CR"/>
226<iref primary="true" item="Grammar" subitem="CRLF"/>
227<iref primary="true" item="Grammar" subitem="CTL"/>
228<iref primary="true" item="Grammar" subitem="DIGIT"/>
229<iref primary="true" item="Grammar" subitem="DQUOTE"/>
230<iref primary="true" item="Grammar" subitem="HEXDIG"/>
231<iref primary="true" item="Grammar" subitem="HTAB"/>
232<iref primary="true" item="Grammar" subitem="LF"/>
233<iref primary="true" item="Grammar" subitem="OCTET"/>
234<iref primary="true" item="Grammar" subitem="SP"/>
235<iref primary="true" item="Grammar" subitem="VCHAR"/>
237   This specification uses the Augmented Backus-Naur Form (ABNF) notation
238   of <xref target="RFC5234"/> with the list rule extension defined in
239   <xref target="abnf.extension"/>.  <xref target="collected.abnf"/> shows
240   the collected ABNF with the list rule expanded.
242<t anchor="core.rules">
243  <x:anchor-alias value="ALPHA"/>
244  <x:anchor-alias value="CTL"/>
245  <x:anchor-alias value="CR"/>
246  <x:anchor-alias value="CRLF"/>
247  <x:anchor-alias value="DIGIT"/>
248  <x:anchor-alias value="DQUOTE"/>
249  <x:anchor-alias value="HEXDIG"/>
250  <x:anchor-alias value="HTAB"/>
251  <x:anchor-alias value="LF"/>
252  <x:anchor-alias value="OCTET"/>
253  <x:anchor-alias value="SP"/>
254  <x:anchor-alias value="VCHAR"/>
255   The following core rules are included by
256   reference, as defined in <xref target="RFC5234" x:fmt="," x:sec="B.1"/>:
257   ALPHA (letters), CR (carriage return), CRLF (CR LF), CTL (controls),
258   DIGIT (decimal 0-9), DQUOTE (double quote),
259   HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line feed),
260   OCTET (any 8-bit sequence of data), SP (space), and
261   VCHAR (any visible <xref target="USASCII"/> character).
264   As a convention, ABNF rule names prefixed with "obs-" denote
265   "obsolete" grammar rules that appear for historical reasons.
270<section title="Architecture" anchor="architecture">
272   HTTP was created for the World Wide Web architecture
273   and has evolved over time to support the scalability needs of a worldwide
274   hypertext system. Much of that architecture is reflected in the terminology
275   and syntax productions used to define HTTP.
278<section title="Client/Server Messaging" anchor="operation">
279<iref primary="true" item="client"/>
280<iref primary="true" item="server"/>
281<iref primary="true" item="connection"/>
283   HTTP is a stateless request/response protocol that operates by exchanging
284   <x:dfn>messages</x:dfn> (<xref target="http.message"/>) across a reliable
285   transport or session-layer
286   "<x:dfn>connection</x:dfn>" (<xref target=""/>).
287   An HTTP "<x:dfn>client</x:dfn>" is a program that establishes a connection
288   to a server for the purpose of sending one or more HTTP requests.
289   An HTTP "<x:dfn>server</x:dfn>" is a program that accepts connections
290   in order to service HTTP requests by sending HTTP responses.
292<iref primary="true" item="user agent"/>
293<iref primary="true" item="origin server"/>
294<iref primary="true" item="browser"/>
295<iref primary="true" item="spider"/>
296<iref primary="true" item="sender"/>
297<iref primary="true" item="recipient"/>
299   The terms client and server refer only to the roles that
300   these programs perform for a particular connection.  The same program
301   might act as a client on some connections and a server on others.
302   We use the term "<x:dfn>user agent</x:dfn>" to refer to any of the various
303   client programs that initiate a request, including (but not limited to)
304   browsers, spiders (web-based robots), command-line tools, native
305   applications, and mobile apps.  The term "<x:dfn>origin server</x:dfn>" is
306   used to refer to the program that can originate authoritative responses to
307   a request. For general requirements, we use the terms
308   "<x:dfn>sender</x:dfn>" and "<x:dfn>recipient</x:dfn>" to refer to any
309   component that sends or receives, respectively, a given message.
312   HTTP relies upon the Uniform Resource Identifier (URI)
313   standard <xref target="RFC3986"/> to indicate the target resource
314   (<xref target="target-resource"/>) and relationships between resources.
315   Messages are passed in a format similar to that used by Internet mail
316   <xref target="RFC5322"/> and the Multipurpose Internet Mail Extensions
317   (MIME) <xref target="RFC2045"/> (see &diff-mime; for the differences
318   between HTTP and MIME messages).
321   Most HTTP communication consists of a retrieval request (GET) for
322   a representation of some resource identified by a URI.  In the
323   simplest case, this might be accomplished via a single bidirectional
324   connection (===) between the user agent (UA) and the origin server (O).
326<figure><artwork type="drawing">
327         request   &gt;
328    <x:highlight>UA</x:highlight> ======================================= <x:highlight>O</x:highlight>
329                                &lt;   response
331<iref primary="true" item="message"/>
332<iref primary="true" item="request"/>
333<iref primary="true" item="response"/>
335   A client sends an HTTP request to a server in the form of a <x:dfn>request</x:dfn>
336   message, beginning with a request-line that includes a method, URI, and
337   protocol version (<xref target="request.line"/>),
338   followed by header fields containing
339   request modifiers, client information, and representation metadata
340   (<xref target="header.fields"/>),
341   an empty line to indicate the end of the header section, and finally
342   a message body containing the payload body (if any,
343   <xref target="message.body"/>).
346   A server responds to a client's request by sending one or more HTTP
347   <x:dfn>response</x:dfn>
348   messages, each beginning with a status line that
349   includes the protocol version, a success or error code, and textual
350   reason phrase (<xref target="status.line"/>),
351   possibly followed by header fields containing server
352   information, resource metadata, and representation metadata
353   (<xref target="header.fields"/>),
354   an empty line to indicate the end of the header section, and finally
355   a message body containing the payload body (if any,
356   <xref target="message.body"/>).
359   A connection might be used for multiple request/response exchanges,
360   as defined in <xref target="persistent.connections"/>.
363   The following example illustrates a typical message exchange for a
364   GET request on the URI "":
367client request:
368</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
369GET /hello.txt HTTP/1.1
370User-Agent: curl/7.16.3 libcurl/7.16.3 OpenSSL/0.9.7l zlib/1.2.3
372Accept-Language: en, mi
376server response:
377</preamble><artwork type="message/http; msgtype=&#34;response&#34;" x:indent-with="  ">
378HTTP/1.1 200 OK
379Date: Mon, 27 Jul 2009 12:28:53 GMT
380Server: Apache
381Last-Modified: Wed, 22 Jul 2009 19:15:56 GMT
382ETag: "34aa387-d-1568eb00"
383Accept-Ranges: bytes
384Content-Length: <x:length-of target="exbody"/>
385Vary: Accept-Encoding
386Content-Type: text/plain
388<x:span anchor="exbody">Hello World!
392<section title="Implementation Diversity" anchor="implementation-diversity">
394   When considering the design of HTTP, it is easy to fall into a trap of
395   thinking that all user agents are general-purpose browsers and all origin
396   servers are large public websites. That is not the case in practice.
397   Common HTTP user agents include household appliances, stereos, scales,
398   firmware update scripts, command-line programs, mobile apps,
399   and communication devices in a multitude of shapes and sizes.  Likewise,
400   common HTTP origin servers include home automation units, configurable
401   networking components, office machines, autonomous robots, news feeds,
402   traffic cameras, ad selectors, and video delivery platforms.
405   The term "user agent" does not imply that there is a human user directly
406   interacting with the software agent at the time of a request. In many
407   cases, a user agent is installed or configured to run in the background
408   and save its results for later inspection (or save only a subset of those
409   results that might be interesting or erroneous). Spiders, for example, are
410   typically given a start URI and configured to follow certain behavior while
411   crawling the Web as a hypertext graph.
414   The implementation diversity of HTTP means that we cannot assume the
415   user agent can make interactive suggestions to a user or provide adequate
416   warning for security or privacy options.  In the few cases where this
417   specification requires reporting of errors to the user, it is acceptable
418   for such reporting to only be observable in an error console or log file.
419   Likewise, requirements that an automated action be confirmed by the user
420   before proceeding can be met via advance configuration choices,
421   run-time options, or simply not proceeding with the unsafe action.
425<section title="Intermediaries" anchor="intermediaries">
426<iref primary="true" item="intermediary"/>
428   HTTP enables the use of intermediaries to satisfy requests through
429   a chain of connections.  There are three common forms of HTTP
430   <x:dfn>intermediary</x:dfn>: proxy, gateway, and tunnel.  In some cases,
431   a single intermediary might act as an origin server, proxy, gateway,
432   or tunnel, switching behavior based on the nature of each request.
434<figure><artwork type="drawing">
435         &gt;             &gt;             &gt;             &gt;
436    <x:highlight>UA</x:highlight> =========== <x:highlight>A</x:highlight> =========== <x:highlight>B</x:highlight> =========== <x:highlight>C</x:highlight> =========== <x:highlight>O</x:highlight>
437               &lt;             &lt;             &lt;             &lt;
440   The figure above shows three intermediaries (A, B, and C) between the
441   user agent and origin server. A request or response message that
442   travels the whole chain will pass through four separate connections.
443   Some HTTP communication options
444   might apply only to the connection with the nearest, non-tunnel
445   neighbor, only to the end-points of the chain, or to all connections
446   along the chain. Although the diagram is linear, each participant might
447   be engaged in multiple, simultaneous communications. For example, B
448   might be receiving requests from many clients other than A, and/or
449   forwarding requests to servers other than C, at the same time that it
450   is handling A's request.
453<iref primary="true" item="upstream"/><iref primary="true" item="downstream"/>
454<iref primary="true" item="inbound"/><iref primary="true" item="outbound"/>
455   We use the terms "<x:dfn>upstream</x:dfn>" and "<x:dfn>downstream</x:dfn>"
456   to describe various requirements in relation to the directional flow of a
457   message: all messages flow from upstream to downstream.
458   Likewise, we use the terms inbound and outbound to refer to
459   directions in relation to the request path:
460   "<x:dfn>inbound</x:dfn>" means toward the origin server and
461   "<x:dfn>outbound</x:dfn>" means toward the user agent.
463<t><iref primary="true" item="proxy"/>
464   A "<x:dfn>proxy</x:dfn>" is a message forwarding agent that is selected by the
465   client, usually via local configuration rules, to receive requests
466   for some type(s) of absolute URI and attempt to satisfy those
467   requests via translation through the HTTP interface.  Some translations
468   are minimal, such as for proxy requests for "http" URIs, whereas
469   other requests might require translation to and from entirely different
470   application-level protocols. Proxies are often used to group an
471   organization's HTTP requests through a common intermediary for the
472   sake of security, annotation services, or shared caching.
475<iref primary="true" item="transforming proxy"/>
476<iref primary="true" item="non-transforming proxy"/>
477   An HTTP-to-HTTP proxy is called a "<x:dfn>transforming proxy</x:dfn>" if it is designed
478   or configured to modify request or response messages in a semantically
479   meaningful way (i.e., modifications, beyond those required by normal
480   HTTP processing, that change the message in a way that would be
481   significant to the original sender or potentially significant to
482   downstream recipients).  For example, a transforming proxy might be
483   acting as a shared annotation server (modifying responses to include
484   references to a local annotation database), a malware filter, a
485   format transcoder, or an intranet-to-Internet privacy filter.  Such
486   transformations are presumed to be desired by the client (or client
487   organization) that selected the proxy and are beyond the scope of
488   this specification.  However, when a proxy is not intended to transform
489   a given message, we use the term "<x:dfn>non-transforming proxy</x:dfn>" to target
490   requirements that preserve HTTP message semantics. See &status-203; and
491   &header-warning; for status and warning codes related to transformations.
493<t><iref primary="true" item="gateway"/><iref primary="true" item="reverse proxy"/>
494<iref primary="true" item="accelerator"/>
495   A "<x:dfn>gateway</x:dfn>" (a.k.a., "<x:dfn>reverse proxy</x:dfn>")
496   is a receiving agent that acts
497   as a layer above some other server(s) and translates the received
498   requests to the underlying server's protocol.  Gateways are often
499   used to encapsulate legacy or untrusted information services, to
500   improve server performance through "<x:dfn>accelerator</x:dfn>" caching, and to
501   enable partitioning or load-balancing of HTTP services across
502   multiple machines.
505   A gateway behaves as an origin server on its outbound connection and
506   as a user agent on its inbound connection.
507   All HTTP requirements applicable to an origin server
508   also apply to the outbound communication of a gateway.
509   A gateway communicates with inbound servers using any protocol that
510   it desires, including private extensions to HTTP that are outside
511   the scope of this specification.  However, an HTTP-to-HTTP gateway
512   that wishes to interoperate with third-party HTTP servers &MUST;
513   conform to HTTP user agent requirements on the gateway's inbound
514   connection and &MUST; implement the <x:ref>Connection</x:ref>
515   (<xref target="header.connection"/>) and <x:ref>Via</x:ref>
516   (<xref target="header.via"/>) header fields for both connections.
518<t><iref primary="true" item="tunnel"/>
519   A "<x:dfn>tunnel</x:dfn>" acts as a blind relay between two connections
520   without changing the messages. Once active, a tunnel is not
521   considered a party to the HTTP communication, though the tunnel might
522   have been initiated by an HTTP request. A tunnel ceases to exist when
523   both ends of the relayed connection are closed. Tunnels are used to
524   extend a virtual connection through an intermediary, such as when
525   Transport Layer Security (TLS, <xref target="RFC5246"/>) is used to
526   establish confidential communication through a shared firewall proxy.
528<t><iref primary="true" item="interception proxy"/>
529<iref primary="true" item="transparent proxy"/>
530<iref primary="true" item="captive portal"/>
531   The above categories for intermediary only consider those acting as
532   participants in the HTTP communication.  There are also intermediaries
533   that can act on lower layers of the network protocol stack, filtering or
534   redirecting HTTP traffic without the knowledge or permission of message
535   senders. Network intermediaries often introduce security flaws or
536   interoperability problems by violating HTTP semantics.  For example, an
537   "<x:dfn>interception proxy</x:dfn>" <xref target="RFC3040"/> (also commonly
538   known as a "<x:dfn>transparent proxy</x:dfn>" <xref target="RFC1919"/> or
539   "<x:dfn>captive portal</x:dfn>")
540   differs from an HTTP proxy because it is not selected by the client.
541   Instead, an interception proxy filters or redirects outgoing TCP port 80
542   packets (and occasionally other common port traffic).
543   Interception proxies are commonly found on public network access points,
544   as a means of enforcing account subscription prior to allowing use of
545   non-local Internet services, and within corporate firewalls to enforce
546   network usage policies.
547   They are indistinguishable from a man-in-the-middle attack.
550   HTTP is defined as a stateless protocol, meaning that each request message
551   can be understood in isolation.  Many implementations depend on HTTP's
552   stateless design in order to reuse proxied connections or dynamically
553   load-balance requests across multiple servers.  Hence, servers &MUST-NOT;
554   assume that two requests on the same connection are from the same user
555   agent unless the connection is secured and specific to that agent.
556   Some non-standard HTTP extensions (e.g., <xref target="RFC4559"/>) have
557   been known to violate this requirement, resulting in security and
558   interoperability problems.
562<section title="Caches" anchor="caches">
563<iref primary="true" item="cache"/>
565   A "<x:dfn>cache</x:dfn>" is a local store of previous response messages and the
566   subsystem that controls its message storage, retrieval, and deletion.
567   A cache stores cacheable responses in order to reduce the response
568   time and network bandwidth consumption on future, equivalent
569   requests. Any client or server &MAY; employ a cache, though a cache
570   cannot be used by a server while it is acting as a tunnel.
573   The effect of a cache is that the request/response chain is shortened
574   if one of the participants along the chain has a cached response
575   applicable to that request. The following illustrates the resulting
576   chain if B has a cached copy of an earlier response from O (via C)
577   for a request that has not been cached by UA or A.
579<figure><artwork type="drawing">
580            &gt;             &gt;
581       <x:highlight>UA</x:highlight> =========== <x:highlight>A</x:highlight> =========== <x:highlight>B</x:highlight> - - - - - - <x:highlight>C</x:highlight> - - - - - - <x:highlight>O</x:highlight>
582                  &lt;             &lt;
584<t><iref primary="true" item="cacheable"/>
585   A response is "<x:dfn>cacheable</x:dfn>" if a cache is allowed to store a copy of
586   the response message for use in answering subsequent requests.
587   Even when a response is cacheable, there might be additional
588   constraints placed by the client or by the origin server on when
589   that cached response can be used for a particular request. HTTP
590   requirements for cache behavior and cacheable responses are
591   defined in &caching-overview;. 
594   There are a wide variety of architectures and configurations
595   of caches deployed across the World Wide Web and
596   inside large organizations. These include national hierarchies
597   of proxy caches to save transoceanic bandwidth, collaborative systems that
598   broadcast or multicast cache entries, archives of pre-fetched cache
599   entries for use in off-line or high-latency environments, and so on.
603<section title="Conformance and Error Handling" anchor="conformance">
605   This specification targets conformance criteria according to the role of
606   a participant in HTTP communication.  Hence, HTTP requirements are placed
607   on senders, recipients, clients, servers, user agents, intermediaries,
608   origin servers, proxies, gateways, or caches, depending on what behavior
609   is being constrained by the requirement. Additional (social) requirements
610   are placed on implementations, resource owners, and protocol element
611   registrations when they apply beyond the scope of a single communication.
614   The verb "generate" is used instead of "send" where a requirement
615   differentiates between creating a protocol element and merely forwarding a
616   received element downstream.
619   An implementation is considered conformant if it complies with all of the
620   requirements associated with the roles it partakes in HTTP. Note that
621   SHOULD-level requirements are relevant here, unless one of the documented
622   exceptions is applicable.
625   Conformance applies to both the syntax and semantics of HTTP protocol
626   elements. A sender &MUST-NOT; generate protocol elements that convey a
627   meaning that is known by that sender to be false. A sender &MUST-NOT;
628   generate protocol elements that do not match the grammar defined by the
629   ABNF rules for those protocol elements that are applicable to the sender's
630   role. If a received protocol element is processed, the recipient &MUST; be
631   able to parse any value that would match the ABNF rules for that protocol
632   element, excluding only those rules not applicable to the recipient's role.
635   Unless noted otherwise, a recipient &MAY; attempt to recover a usable
636   protocol element from an invalid construct.  HTTP does not define
637   specific error handling mechanisms except when they have a direct impact
638   on security, since different applications of the protocol require
639   different error handling strategies.  For example, a Web browser might
640   wish to transparently recover from a response where the
641   <x:ref>Location</x:ref> header field doesn't parse according to the ABNF,
642   whereas a systems control client might consider any form of error recovery
643   to be dangerous.
647<section title="Protocol Versioning" anchor="http.version">
648  <x:anchor-alias value="HTTP-version"/>
649  <x:anchor-alias value="HTTP-name"/>
651   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate
652   versions of the protocol. This specification defines version "1.1".
653   The protocol version as a whole indicates the sender's conformance
654   with the set of requirements laid out in that version's corresponding
655   specification of HTTP.
658   The version of an HTTP message is indicated by an HTTP-version field
659   in the first line of the message. HTTP-version is case-sensitive.
661<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-version"/><iref primary="true" item="Grammar" subitem="HTTP-name"/>
662  <x:ref>HTTP-version</x:ref>  = <x:ref>HTTP-name</x:ref> "/" <x:ref>DIGIT</x:ref> "." <x:ref>DIGIT</x:ref>
663  <x:ref>HTTP-name</x:ref>     = <x:abnf-char-sequence>"HTTP"</x:abnf-char-sequence> ; "HTTP", case-sensitive
666   The HTTP version number consists of two decimal digits separated by a "."
667   (period or decimal point).  The first digit ("major version") indicates the
668   HTTP messaging syntax, whereas the second digit ("minor version") indicates
669   the highest minor version to which the sender is
670   conformant and able to understand for future communication.  The minor
671   version advertises the sender's communication capabilities even when the
672   sender is only using a backwards-compatible subset of the protocol,
673   thereby letting the recipient know that more advanced features can
674   be used in response (by servers) or in future requests (by clients).
677   When an HTTP/1.1 message is sent to an HTTP/1.0 recipient
678   <xref target="RFC1945"/> or a recipient whose version is unknown,
679   the HTTP/1.1 message is constructed such that it can be interpreted
680   as a valid HTTP/1.0 message if all of the newer features are ignored.
681   This specification places recipient-version requirements on some
682   new features so that a conformant sender will only use compatible
683   features until it has determined, through configuration or the
684   receipt of a message, that the recipient supports HTTP/1.1.
687   The interpretation of a header field does not change between minor
688   versions of the same major HTTP version, though the default
689   behavior of a recipient in the absence of such a field can change.
690   Unless specified otherwise, header fields defined in HTTP/1.1 are
691   defined for all versions of HTTP/1.x.  In particular, the <x:ref>Host</x:ref>
692   and <x:ref>Connection</x:ref> header fields ought to be implemented by all
693   HTTP/1.x implementations whether or not they advertise conformance with
694   HTTP/1.1.
697   New header fields can be defined such that, when they are
698   understood by a recipient, they might override or enhance the
699   interpretation of previously defined header fields.  When an
700   implementation receives an unrecognized header field, the recipient
701   &MUST; ignore that header field for local processing regardless of
702   the message's HTTP version.  An unrecognized header field received
703   by a proxy &MUST; be forwarded downstream unless the header field's
704   field-name is listed in the message's <x:ref>Connection</x:ref> header field
705   (see <xref target="header.connection"/>).
706   These requirements allow HTTP's functionality to be enhanced without
707   requiring prior update of deployed intermediaries.
710   Intermediaries that process HTTP messages (i.e., all intermediaries
711   other than those acting as tunnels) &MUST; send their own HTTP-version
712   in forwarded messages.  In other words, they &MUST-NOT; blindly
713   forward the first line of an HTTP message without ensuring that the
714   protocol version in that message matches a version to which that
715   intermediary is conformant for both the receiving and
716   sending of messages.  Forwarding an HTTP message without rewriting
717   the HTTP-version might result in communication errors when downstream
718   recipients use the message sender's version to determine what features
719   are safe to use for later communication with that sender.
722   An HTTP client &SHOULD; send a request version equal to the highest
723   version to which the client is conformant and
724   whose major version is no higher than the highest version supported
725   by the server, if this is known.  An HTTP client &MUST-NOT; send a
726   version to which it is not conformant.
729   An HTTP client &MAY; send a lower request version if it is known that
730   the server incorrectly implements the HTTP specification, but only
731   after the client has attempted at least one normal request and determined
732   from the response status or header fields (e.g., <x:ref>Server</x:ref>) that
733   the server improperly handles higher request versions.
736   An HTTP server &SHOULD; send a response version equal to the highest
737   version to which the server is conformant and
738   whose major version is less than or equal to the one received in the
739   request.  An HTTP server &MUST-NOT; send a version to which it is not
740   conformant.  A server &MAY; send a <x:ref>505 (HTTP Version Not
741   Supported)</x:ref> response if it cannot send a response using the
742   major version used in the client's request.
745   An HTTP server &MAY; send an HTTP/1.0 response to an HTTP/1.0 request
746   if it is known or suspected that the client incorrectly implements the
747   HTTP specification and is incapable of correctly processing later
748   version responses, such as when a client fails to parse the version
749   number correctly or when an intermediary is known to blindly forward
750   the HTTP-version even when it doesn't conform to the given minor
751   version of the protocol. Such protocol downgrades &SHOULD-NOT; be
752   performed unless triggered by specific client attributes, such as when
753   one or more of the request header fields (e.g., <x:ref>User-Agent</x:ref>)
754   uniquely match the values sent by a client known to be in error.
757   The intention of HTTP's versioning design is that the major number
758   will only be incremented if an incompatible message syntax is
759   introduced, and that the minor number will only be incremented when
760   changes made to the protocol have the effect of adding to the message
761   semantics or implying additional capabilities of the sender.  However,
762   the minor version was not incremented for the changes introduced between
763   <xref target="RFC2068"/> and <xref target="RFC2616"/>, and this revision
764   has specifically avoiding any such changes to the protocol.
768<section title="Uniform Resource Identifiers" anchor="uri">
769<iref primary="true" item="resource"/>
771   Uniform Resource Identifiers (URIs) <xref target="RFC3986"/> are used
772   throughout HTTP as the means for identifying resources (&resource;).
773   URI references are used to target requests, indicate redirects, and define
774   relationships.
776  <x:anchor-alias value="URI-reference"/>
777  <x:anchor-alias value="absolute-URI"/>
778  <x:anchor-alias value="relative-part"/>
779  <x:anchor-alias value="authority"/>
780  <x:anchor-alias value="path-abempty"/>
781  <x:anchor-alias value="path-absolute"/>
782  <x:anchor-alias value="port"/>
783  <x:anchor-alias value="query"/>
784  <x:anchor-alias value="uri-host"/>
785  <x:anchor-alias value="partial-URI"/>
787   This specification adopts the definitions of "URI-reference",
788   "absolute-URI", "relative-part", "port", "host",
789   "path-abempty", "path-absolute", "query", and "authority" from the
790   URI generic syntax.
791   In addition, we define a partial-URI rule for protocol elements
792   that allow a relative URI but not a fragment.
794<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="URI-reference"><!--exported production--></iref><iref primary="true" item="Grammar" subitem="absolute-URI"/><iref primary="true" item="Grammar" subitem="authority"/><iref primary="true" item="Grammar" subitem="path-absolute"/><iref primary="true" item="Grammar" subitem="port"/><iref primary="true" item="Grammar" subitem="query"/><iref primary="true" item="Grammar" subitem="uri-host"/><iref primary="true" item="Grammar" subitem="partial-URI"><!--exported production--></iref>
795  <x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in <xref target="RFC3986" x:fmt="," x:sec="4.1"/>&gt;
796  <x:ref>absolute-URI</x:ref>  = &lt;absolute-URI, defined in <xref target="RFC3986" x:fmt="," x:sec="4.3"/>&gt;
797  <x:ref>relative-part</x:ref> = &lt;relative-part, defined in <xref target="RFC3986" x:fmt="," x:sec="4.2"/>&gt;
798  <x:ref>authority</x:ref>     = &lt;authority, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2"/>&gt;
799  <x:ref>path-abempty</x:ref>  = &lt;path-abempty, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
800  <x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
801  <x:ref>port</x:ref>          = &lt;port, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.3"/>&gt;
802  <x:ref>query</x:ref>         = &lt;query, defined in <xref target="RFC3986" x:fmt="," x:sec="3.4"/>&gt;
803  <x:ref>uri-host</x:ref>      = &lt;host, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>&gt;
805  <x:ref>partial-URI</x:ref>   = relative-part [ "?" query ]
808   Each protocol element in HTTP that allows a URI reference will indicate
809   in its ABNF production whether the element allows any form of reference
810   (URI-reference), only a URI in absolute form (absolute-URI), only the
811   path and optional query components, or some combination of the above.
812   Unless otherwise indicated, URI references are parsed
813   relative to the effective request URI
814   (<xref target="effective.request.uri"/>).
817<section title="http URI scheme" anchor="http.uri">
818  <x:anchor-alias value="http-URI"/>
819  <iref item="http URI scheme" primary="true"/>
820  <iref item="URI scheme" subitem="http" primary="true"/>
822   The "http" URI scheme is hereby defined for the purpose of minting
823   identifiers according to their association with the hierarchical
824   namespace governed by a potential HTTP origin server listening for
825   TCP connections on a given port.
827<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http-URI"><!--terminal production--></iref>
828  <x:ref>http-URI</x:ref> = "http:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
831   The HTTP origin server is identified by the generic syntax's
832   <x:ref>authority</x:ref> component, which includes a host identifier
833   and optional TCP port (<xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>).
834   The remainder of the URI, consisting of both the hierarchical path
835   component and optional query component, serves as an identifier for
836   a potential resource within that origin server's name space.
839   If the host identifier is provided as an IP address,
840   then the origin server is any listener on the indicated TCP port at
841   that IP address. If host is a registered name, then that name is
842   considered an indirect identifier and the recipient might use a name
843   resolution service, such as DNS, to find the address of a listener
844   for that host.
845   The host &MUST-NOT; be empty; if an "http" URI is received with an
846   empty host, then it &MUST; be rejected as invalid.
847   If the port subcomponent is empty or not given, then TCP port 80 is
848   assumed (the default reserved port for WWW services).
851   Regardless of the form of host identifier, access to that host is not
852   implied by the mere presence of its name or address. The host might or might
853   not exist and, even when it does exist, might or might not be running an
854   HTTP server or listening to the indicated port. The "http" URI scheme
855   makes use of the delegated nature of Internet names and addresses to
856   establish a naming authority (whatever entity has the ability to place
857   an HTTP server at that Internet name or address) and allows that
858   authority to determine which names are valid and how they might be used.
861   When an "http" URI is used within a context that calls for access to the
862   indicated resource, a client &MAY; attempt access by resolving
863   the host to an IP address, establishing a TCP connection to that address
864   on the indicated port, and sending an HTTP request message
865   (<xref target="http.message"/>) containing the URI's identifying data
866   (<xref target="message.routing"/>) to the server.
867   If the server responds to that request with a non-interim HTTP response
868   message, as described in &status-codes;, then that response
869   is considered an authoritative answer to the client's request.
872   Although HTTP is independent of the transport protocol, the "http"
873   scheme is specific to TCP-based services because the name delegation
874   process depends on TCP for establishing authority.
875   An HTTP service based on some other underlying connection protocol
876   would presumably be identified using a different URI scheme, just as
877   the "https" scheme (below) is used for resources that require an
878   end-to-end secured connection. Other protocols might also be used to
879   provide access to "http" identified resources &mdash; it is only the
880   authoritative interface used for mapping the namespace that is
881   specific to TCP.
884   The URI generic syntax for authority also includes a deprecated
885   userinfo subcomponent (<xref target="RFC3986" x:fmt="," x:sec="3.2.1"/>)
886   for including user authentication information in the URI.  Some
887   implementations make use of the userinfo component for internal
888   configuration of authentication information, such as within command
889   invocation options, configuration files, or bookmark lists, even
890   though such usage might expose a user identifier or password.
891   Senders &MUST; exclude the userinfo subcomponent (and its "@"
892   delimiter) when an "http" URI is transmitted within a message as a
893   request target or header field value.
894   Recipients of an "http" URI reference &SHOULD; parse for userinfo and
895   treat its presence as an error, since it is likely being used to obscure
896   the authority for the sake of phishing attacks.
900<section title="https URI scheme" anchor="https.uri">
901   <x:anchor-alias value="https-URI"/>
902   <iref item="https URI scheme"/>
903   <iref item="URI scheme" subitem="https"/>
905   The "https" URI scheme is hereby defined for the purpose of minting
906   identifiers according to their association with the hierarchical
907   namespace governed by a potential HTTP origin server listening to a
908   given TCP port for TLS-secured connections <xref target="RFC5246"/>.
911   All of the requirements listed above for the "http" scheme are also
912   requirements for the "https" scheme, except that a default TCP port
913   of 443 is assumed if the port subcomponent is empty or not given,
914   and the TCP connection &MUST; be secured, end-to-end, through the
915   use of strong encryption prior to sending the first HTTP request.
917<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="https-URI"><!--terminal production--></iref>
918  <x:ref>https-URI</x:ref> = "https:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
921   Resources made available via the "https" scheme have no shared
922   identity with the "http" scheme even if their resource identifiers
923   indicate the same authority (the same host listening to the same
924   TCP port).  They are distinct name spaces and are considered to be
925   distinct origin servers.  However, an extension to HTTP that is
926   defined to apply to entire host domains, such as the Cookie protocol
927   <xref target="RFC6265"/>, can allow information
928   set by one service to impact communication with other services
929   within a matching group of host domains.
932   The process for authoritative access to an "https" identified
933   resource is defined in <xref target="RFC2818"/>.
937<section title="http and https URI Normalization and Comparison" anchor="uri.comparison">
939   Since the "http" and "https" schemes conform to the URI generic syntax,
940   such URIs are normalized and compared according to the algorithm defined
941   in <xref target="RFC3986" x:fmt="," x:sec="6"/>, using the defaults
942   described above for each scheme.
945   If the port is equal to the default port for a scheme, the normal form is
946   to elide the port subcomponent. When not being used in absolute form as the
947   request target of an OPTIONS request, an empty path component is equivalent
948   to an absolute path of "/", so the normal form is to provide a path of "/"
949   instead. The scheme and host are case-insensitive and normally provided in
950   lowercase; all other components are compared in a case-sensitive manner.
951   Characters other than those in the "reserved" set are equivalent to their
952   percent-encoded octets (see <xref target="RFC3986" x:fmt=","
953   x:sec="2.1"/>): the normal form is to not encode them.
956   For example, the following three URIs are equivalent:
958<figure><artwork type="example">
967<section title="Message Format" anchor="http.message">
968<x:anchor-alias value="generic-message"/>
969<x:anchor-alias value="message.types"/>
970<x:anchor-alias value="HTTP-message"/>
971<x:anchor-alias value="start-line"/>
972<iref item="header section"/>
973<iref item="headers"/>
974<iref item="header field"/>
976   All HTTP/1.1 messages consist of a start-line followed by a sequence of
977   octets in a format similar to the Internet Message Format
978   <xref target="RFC5322"/>: zero or more header fields (collectively
979   referred to as the "headers" or the "header section"), an empty line
980   indicating the end of the header section, and an optional message body.
982<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"><!--terminal production--></iref>
983  <x:ref>HTTP-message</x:ref>   = <x:ref>start-line</x:ref>
984                   *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
985                   <x:ref>CRLF</x:ref>
986                   [ <x:ref>message-body</x:ref> ]
989   The normal procedure for parsing an HTTP message is to read the
990   start-line into a structure, read each header field into a hash
991   table by field name until the empty line, and then use the parsed
992   data to determine if a message body is expected.  If a message body
993   has been indicated, then it is read as a stream until an amount
994   of octets equal to the message body length is read or the connection
995   is closed.
998   Recipients &MUST; parse an HTTP message as a sequence of octets in an
999   encoding that is a superset of US-ASCII <xref target="USASCII"/>.
1000   Parsing an HTTP message as a stream of Unicode characters, without regard
1001   for the specific encoding, creates security vulnerabilities due to the
1002   varying ways that string processing libraries handle invalid multibyte
1003   character sequences that contain the octet LF (%x0A).  String-based
1004   parsers can only be safely used within protocol elements after the element
1005   has been extracted from the message, such as within a header field-value
1006   after message parsing has delineated the individual fields.
1009   An HTTP message can be parsed as a stream for incremental processing or
1010   forwarding downstream.  However, recipients cannot rely on incremental
1011   delivery of partial messages, since some implementations will buffer or
1012   delay message forwarding for the sake of network efficiency, security
1013   checks, or payload transformations.
1016<section title="Start Line" anchor="start.line">
1017  <x:anchor-alias value="Start-Line"/>
1019   An HTTP message can either be a request from client to server or a
1020   response from server to client.  Syntactically, the two types of message
1021   differ only in the start-line, which is either a request-line (for requests)
1022   or a status-line (for responses), and in the algorithm for determining
1023   the length of the message body (<xref target="message.body"/>).
1026   In theory, a client could receive requests and a server could receive
1027   responses, distinguishing them by their different start-line formats,
1028   but in practice servers are implemented to only expect a request
1029   (a response is interpreted as an unknown or invalid request method)
1030   and clients are implemented to only expect a response.
1032<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="start-line"/>
1033  <x:ref>start-line</x:ref>     = <x:ref>request-line</x:ref> / <x:ref>status-line</x:ref>
1036   A sender &MUST-NOT; send whitespace between the start-line and
1037   the first header field. The presence of such whitespace in a request
1038   might be an attempt to trick a server into ignoring that field or
1039   processing the line after it as a new request, either of which might
1040   result in a security vulnerability if other implementations within
1041   the request chain interpret the same message differently.
1042   Likewise, the presence of such whitespace in a response might be
1043   ignored by some clients or cause others to cease parsing.
1046   A recipient that receives whitespace between the start-line and
1047   the first header field &MUST; either reject the message as invalid or
1048   consume each whitespace-preceded line without further processing of it
1049   (i.e., ignore the entire line, along with any subsequent lines preceded
1050   by whitespace, until a properly formed header field is received or the
1051   header block is terminated).
1054<section title="Request Line" anchor="request.line">
1055  <x:anchor-alias value="Request"/>
1056  <x:anchor-alias value="request-line"/>
1058   A request-line begins with a method token, followed by a single
1059   space (SP), the request-target, another single space (SP), the
1060   protocol version, and ending with CRLF.
1062<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-line"/>
1063  <x:ref>request-line</x:ref>   = <x:ref>method</x:ref> <x:ref>SP</x:ref> <x:ref>request-target</x:ref> <x:ref>SP</x:ref> <x:ref>HTTP-version</x:ref> <x:ref>CRLF</x:ref>
1065<iref primary="true" item="method"/>
1066<t anchor="method">
1067   The method token indicates the request method to be performed on the
1068   target resource. The request method is case-sensitive.
1070<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="method"/>
1071  <x:ref>method</x:ref>         = <x:ref>token</x:ref>
1074   The methods defined by this specification can be found in
1075   &methods;, along with information regarding the HTTP method registry
1076   and considerations for defining new methods.
1078<iref item="request-target"/>
1080   The request-target identifies the target resource upon which to apply
1081   the request, as defined in <xref target="request-target"/>.
1084   No whitespace is allowed inside the method, request-target, and
1085   protocol version.  Hence, recipients typically parse the request-line
1086   into its component parts by splitting on whitespace
1087   (see <xref target="message.robustness"/>).
1090   Unfortunately, some user agents fail to properly encode hypertext
1091   references that have embedded whitespace, sending the characters directly
1092   instead of properly encoding or excluding the disallowed characters.
1093   Recipients of an invalid request-line &SHOULD; respond with either a
1094   <x:ref>400 (Bad Request)</x:ref> error or a <x:ref>301 (Moved Permanently)</x:ref>
1095   redirect with the request-target properly encoded.  Recipients &SHOULD-NOT;
1096   attempt to autocorrect and then process the request without a redirect,
1097   since the invalid request-line might be deliberately crafted to bypass
1098   security filters along the request chain.
1101   HTTP does not place a pre-defined limit on the length of a request-line.
1102   A server that receives a method longer than any that it implements
1103   &SHOULD; respond with a <x:ref>501 (Not Implemented)</x:ref> status code.
1104   A server &MUST; be prepared to receive URIs of unbounded length and
1105   respond with the <x:ref>414 (URI Too Long)</x:ref> status code if the received
1106   request-target would be longer than the server wishes to handle
1107   (see &status-414;).
1110   Various ad-hoc limitations on request-line length are found in practice.
1111   It is &RECOMMENDED; that all HTTP senders and recipients support, at a
1112   minimum, request-line lengths of 8000 octets.
1116<section title="Status Line" anchor="status.line">
1117  <x:anchor-alias value="response"/>
1118  <x:anchor-alias value="status-line"/>
1119  <x:anchor-alias value="status-code"/>
1120  <x:anchor-alias value="reason-phrase"/>
1122   The first line of a response message is the status-line, consisting
1123   of the protocol version, a space (SP), the status code, another space,
1124   a possibly-empty textual phrase describing the status code, and
1125   ending with CRLF.
1127<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-line"/>
1128  <x:ref>status-line</x:ref> = <x:ref>HTTP-version</x:ref> <x:ref>SP</x:ref> <x:ref>status-code</x:ref> <x:ref>SP</x:ref> <x:ref>reason-phrase</x:ref> <x:ref>CRLF</x:ref>
1131   The status-code element is a 3-digit integer code describing the
1132   result of the server's attempt to understand and satisfy the client's
1133   corresponding request. The rest of the response message is to be
1134   interpreted in light of the semantics defined for that status code.
1135   See &status-codes; for information about the semantics of status codes,
1136   including the classes of status code (indicated by the first digit),
1137   the status codes defined by this specification, considerations for the
1138   definition of new status codes, and the IANA registry.
1140<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-code"/>
1141  <x:ref>status-code</x:ref>    = 3<x:ref>DIGIT</x:ref>
1144   The reason-phrase element exists for the sole purpose of providing a
1145   textual description associated with the numeric status code, mostly
1146   out of deference to earlier Internet application protocols that were more
1147   frequently used with interactive text clients. A client &SHOULD; ignore
1148   the reason-phrase content.
1150<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="reason-phrase"/>
1151  <x:ref>reason-phrase</x:ref>  = *( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
1156<section title="Header Fields" anchor="header.fields">
1157  <x:anchor-alias value="header-field"/>
1158  <x:anchor-alias value="field-content"/>
1159  <x:anchor-alias value="field-name"/>
1160  <x:anchor-alias value="field-value"/>
1161  <x:anchor-alias value="obs-fold"/>
1163   Each HTTP header field consists of a case-insensitive field name
1164   followed by a colon (":"), optional whitespace, and the field value.
1166<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="header-field"/><iref primary="true" item="Grammar" subitem="field-name"/><iref primary="true" item="Grammar" subitem="field-value"/><iref primary="true" item="Grammar" subitem="field-content"/><iref primary="true" item="Grammar" subitem="obs-fold"/>
1167  <x:ref>header-field</x:ref>   = <x:ref>field-name</x:ref> ":" <x:ref>OWS</x:ref> <x:ref>field-value</x:ref> <x:ref>BWS</x:ref>
1168  <x:ref>field-name</x:ref>     = <x:ref>token</x:ref>
1169  <x:ref>field-value</x:ref>    = *( <x:ref>field-content</x:ref> / <x:ref>obs-fold</x:ref> )
1170  <x:ref>field-content</x:ref>  = *( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
1171  <x:ref>obs-fold</x:ref>       = <x:ref>CRLF</x:ref> ( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1172                 ; obsolete line folding
1173                 ; see <xref target="field.parsing"/>
1176   The field-name token labels the corresponding field-value as having the
1177   semantics defined by that header field.  For example, the <x:ref>Date</x:ref>
1178   header field is defined in &header-date; as containing the origination
1179   timestamp for the message in which it appears.
1182<section title="Field Extensibility" anchor="field.extensibility">
1184   HTTP header fields are fully extensible: there is no limit on the
1185   introduction of new field names, each presumably defining new semantics,
1186   nor on the number of header fields used in a given message.  Existing
1187   fields are defined in each part of this specification and in many other
1188   specifications outside the core standard.
1189   New header fields can be introduced without changing the protocol version
1190   if their defined semantics allow them to be safely ignored by recipients
1191   that do not recognize them.
1194   New HTTP header fields ought to be be registered with IANA in the
1195   Message Header Field Registry, as described in &iana-header-registry;.
1196   A proxy &MUST; forward unrecognized header fields unless the
1197   field-name is listed in the <x:ref>Connection</x:ref> header field
1198   (<xref target="header.connection"/>) or the proxy is specifically
1199   configured to block, or otherwise transform, such fields.
1200   Other recipients &SHOULD; ignore unrecognized header fields.
1204<section title="Field Order" anchor="field.order">
1206   The order in which header fields with differing field names are
1207   received is not significant. However, it is "good practice" to send
1208   header fields that contain control data first, such as <x:ref>Host</x:ref>
1209   on requests and <x:ref>Date</x:ref> on responses, so that implementations
1210   can decide when not to handle a message as early as possible.  A server
1211   &MUST; wait until the entire header section is received before interpreting
1212   a request message, since later header fields might include conditionals,
1213   authentication credentials, or deliberately misleading duplicate
1214   header fields that would impact request processing.
1217   A sender &MUST-NOT; generate multiple header fields with the same field
1218   name in a message unless either the entire field value for that
1219   header field is defined as a comma-separated list [i.e., #(values)]
1220   or the header field is a well-known exception (as noted below).
1223   Multiple header fields with the same field name can be combined into
1224   one "field-name: field-value" pair, without changing the semantics of the
1225   message, by appending each subsequent field value to the combined
1226   field value in order, separated by a comma. The order in which
1227   header fields with the same field name are received is therefore
1228   significant to the interpretation of the combined field value;
1229   a proxy &MUST-NOT; change the order of these field values when
1230   forwarding a message.
1233  <t>
1234   &Note; In practice, the "Set-Cookie" header field (<xref target="RFC6265"/>)
1235   often appears multiple times in a response message and does not use the
1236   list syntax, violating the above requirements on multiple header fields
1237   with the same name. Since it cannot be combined into a single field-value,
1238   recipients ought to handle "Set-Cookie" as a special case while processing
1239   header fields. (See Appendix A.2.3 of <xref target="Kri2001"/> for details.)
1240  </t>
1244<section title="Whitespace" anchor="whitespace">
1245<t anchor="rule.LWS">
1246   This specification uses three rules to denote the use of linear
1247   whitespace: OWS (optional whitespace), RWS (required whitespace), and
1248   BWS ("bad" whitespace).
1250<t anchor="rule.OWS">
1251   The OWS rule is used where zero or more linear whitespace octets might
1252   appear. OWS &SHOULD; either not be generated or be generated as a single
1253   SP. Multiple OWS octets that occur within field-content &SHOULD; either
1254   be replaced with a single SP or transformed to all SP octets (each
1255   octet other than SP replaced with SP) before interpreting the field value
1256   or forwarding the message downstream.
1258<t anchor="rule.RWS">
1259   RWS is used when at least one linear whitespace octet is required to
1260   separate field tokens. RWS &SHOULD; be generated as a single SP.
1261   Multiple RWS octets that occur within field-content &SHOULD; either
1262   be replaced with a single SP or transformed to all SP octets before
1263   interpreting the field value or forwarding the message downstream.
1265<t anchor="rule.BWS">
1266   BWS is used where the grammar allows optional whitespace, for historical
1267   reasons, but senders &SHOULD-NOT; generate it in messages;
1268   recipients &MUST; accept such bad optional whitespace and remove it before
1269   interpreting the field value or forwarding the message downstream.
1271<t anchor="rule.whitespace">
1272  <x:anchor-alias value="BWS"/>
1273  <x:anchor-alias value="OWS"/>
1274  <x:anchor-alias value="RWS"/>
1276<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="OWS"/><iref primary="true" item="Grammar" subitem="RWS"/><iref primary="true" item="Grammar" subitem="BWS"/>
1277  <x:ref>OWS</x:ref>            = *( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1278                 ; optional whitespace
1279  <x:ref>RWS</x:ref>            = 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1280                 ; required whitespace
1281  <x:ref>BWS</x:ref>            = <x:ref>OWS</x:ref>
1282                 ; "bad" whitespace
1286<section title="Field Parsing" anchor="field.parsing">
1288   No whitespace is allowed between the header field-name and colon.
1289   In the past, differences in the handling of such whitespace have led to
1290   security vulnerabilities in request routing and response handling.
1291   A server &MUST; reject any received request message that contains
1292   whitespace between a header field-name and colon with a response code of
1293   <x:ref>400 (Bad Request)</x:ref>. A proxy &MUST; remove any such whitespace
1294   from a response message before forwarding the message downstream.
1297   A field value is preceded by optional whitespace (OWS); a single SP is
1298   preferred. The field value does not include any leading or trailing white
1299   space: OWS occurring before the first non-whitespace octet of the
1300   field value or after the last non-whitespace octet of the field value
1301   is ignored and &SHOULD; be removed before further processing (as this does
1302   not change the meaning of the header field).
1305   Historically, HTTP header field values could be extended over multiple
1306   lines by preceding each extra line with at least one space or horizontal
1307   tab (obs-fold). This specification deprecates such line
1308   folding except within the message/http media type
1309   (<xref target=""/>).
1310   Senders &MUST-NOT; generate messages that include line folding
1311   (i.e., that contain any field-value that matches the obs-fold rule) unless
1312   the message is intended for packaging within the message/http media type.
1313   Recipients &MUST; accept line folding and replace any embedded
1314   obs-fold whitespace with either a single SP or a matching number of SP
1315   octets (to avoid buffer copying) prior to interpreting the field value or
1316   forwarding the message downstream.
1319   Historically, HTTP has allowed field content with text in the ISO-8859-1
1320   <xref target="ISO-8859-1"/> charset, supporting other charsets only
1321   through use of <xref target="RFC2047"/> encoding.
1322   In practice, most HTTP header field values use only a subset of the
1323   US-ASCII charset <xref target="USASCII"/>. Newly defined
1324   header fields &SHOULD; limit their field values to US-ASCII octets.
1325   Recipients &SHOULD; treat other octets in field content (obs-text) as
1326   opaque data.
1330<section title="Field Limits" anchor="field.limits">
1332   HTTP does not place a pre-defined limit on the length of each header field
1333   or on the length of the header block as a whole.  Various ad-hoc
1334   limitations on individual header field length are found in practice,
1335   often depending on the specific field semantics.
1338   A server &MUST; be prepared to receive request header fields of unbounded
1339   length and respond with an appropriate <x:ref>4xx (Client Error)</x:ref>
1340   status code if the received header field(s) are larger than the server
1341   wishes to process.
1344   A client &MUST; be prepared to receive response header fields of unbounded
1345   length. A client &MAY; discard or truncate received header fields that are
1346   larger than the client wishes to process if the field semantics are such
1347   that the dropped value(s) can be safely ignored without changing the
1348   response semantics.
1352<section title="Field value components" anchor="field.components">
1353<t anchor="rule.token.separators">
1354  <x:anchor-alias value="tchar"/>
1355  <x:anchor-alias value="token"/>
1356  <x:anchor-alias value="special"/>
1357  <x:anchor-alias value="word"/>
1358   Many HTTP header field values consist of words (token or quoted-string)
1359   separated by whitespace or special characters. These special characters
1360   &MUST; be in a quoted string to be used within a parameter value (as defined
1361   in <xref target="transfer.codings"/>).
1363<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="word"/><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="tchar"/><iref primary="true" item="Grammar" subitem="special"><!--unused production--></iref>
1364  <x:ref>word</x:ref>           = <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref>
1366  <x:ref>token</x:ref>          = 1*<x:ref>tchar</x:ref>
1368  IMPORTANT: when editing "tchar" make sure that "special" is updated accordingly!!!
1369 -->
1370  <x:ref>tchar</x:ref>          = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*"
1371                 / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~"
1372                 / <x:ref>DIGIT</x:ref> / <x:ref>ALPHA</x:ref>
1373                 ; any <x:ref>VCHAR</x:ref>, except <x:ref>special</x:ref>
1375  <x:ref>special</x:ref>        = "(" / ")" / "&lt;" / ">" / "@" / ","
1376                 / ";" / ":" / "\" / DQUOTE / "/" / "["
1377                 / "]" / "?" / "=" / "{" / "}"
1379<t anchor="rule.quoted-string">
1380  <x:anchor-alias value="quoted-string"/>
1381  <x:anchor-alias value="qdtext"/>
1382  <x:anchor-alias value="obs-text"/>
1383   A string of text is parsed as a single word if it is quoted using
1384   double-quote marks.
1386<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-string"/><iref primary="true" item="Grammar" subitem="qdtext"/><iref primary="true" item="Grammar" subitem="obs-text"/>
1387  <x:ref>quoted-string</x:ref>  = <x:ref>DQUOTE</x:ref> *( <x:ref>qdtext</x:ref> / <x:ref>quoted-pair</x:ref> ) <x:ref>DQUOTE</x:ref>
1388  <x:ref>qdtext</x:ref>         = <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> /%x21 / %x23-5B / %x5D-7E / <x:ref>obs-text</x:ref>
1389  <x:ref>obs-text</x:ref>       = %x80-FF
1391<t anchor="rule.quoted-pair">
1392  <x:anchor-alias value="quoted-pair"/>
1393   The backslash octet ("\") can be used as a single-octet
1394   quoting mechanism within quoted-string constructs:
1396<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
1397  <x:ref>quoted-pair</x:ref>    = "\" ( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
1400   Recipients that process the value of a quoted-string &MUST; handle a
1401   quoted-pair as if it were replaced by the octet following the backslash.
1404   Senders &SHOULD-NOT; generate a quoted-pair in a quoted-string except where
1405   necessary to quote DQUOTE and backslash octets occurring within that string.
1407<t anchor="rule.comment">
1408  <x:anchor-alias value="comment"/>
1409  <x:anchor-alias value="ctext"/>
1410   Comments can be included in some HTTP header fields by surrounding
1411   the comment text with parentheses. Comments are only allowed in
1412   fields containing "comment" as part of their field value definition.
1414<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
1415  <x:ref>comment</x:ref>        = "(" *( <x:ref>ctext</x:ref> / <x:ref>quoted-cpair</x:ref> / <x:ref>comment</x:ref> ) ")"
1416  <x:ref>ctext</x:ref>          = <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / %x21-27 / %x2A-5B / %x5D-7E / <x:ref>obs-text</x:ref>
1418<t anchor="rule.quoted-cpair">
1419  <x:anchor-alias value="quoted-cpair"/>
1420   The backslash octet ("\") can be used as a single-octet
1421   quoting mechanism within comment constructs:
1423<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-cpair"/>
1424  <x:ref>quoted-cpair</x:ref>   = "\" ( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
1427   Senders &SHOULD-NOT; escape octets in comments that do not require escaping
1428   (i.e., other than the backslash octet "\" and the parentheses "(" and ")").
1434<section title="Message Body" anchor="message.body">
1435  <x:anchor-alias value="message-body"/>
1437   The message body (if any) of an HTTP message is used to carry the
1438   payload body of that request or response.  The message body is
1439   identical to the payload body unless a transfer coding has been
1440   applied, as described in <xref target="header.transfer-encoding"/>.
1442<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1443  <x:ref>message-body</x:ref> = *OCTET
1446   The rules for when a message body is allowed in a message differ for
1447   requests and responses.
1450   The presence of a message body in a request is signaled by a
1451   <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1452   field. Request message framing is independent of method semantics,
1453   even if the method does not define any use for a message body.
1456   The presence of a message body in a response depends on both
1457   the request method to which it is responding and the response
1458   status code (<xref target="status.line"/>).
1459   Responses to the HEAD request method never include a message body
1460   because the associated response header fields (e.g.,
1461   <x:ref>Transfer-Encoding</x:ref>, <x:ref>Content-Length</x:ref>, etc.),
1462   if present, indicate only what their values would have been if the request
1463   method had been GET (&HEAD;).
1464   <x:ref>2xx (Successful)</x:ref> responses to CONNECT switch to tunnel
1465   mode instead of having a message body (&CONNECT;).
1466   All <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, and
1467   <x:ref>304 (Not Modified)</x:ref> responses do not include a message body.
1468   All other responses do include a message body, although the body
1469   might be of zero length.
1472<section title="Transfer-Encoding" anchor="header.transfer-encoding">
1473  <iref primary="true" item="Transfer-Encoding header field" x:for-anchor=""/>
1474  <iref item="chunked (Coding Format)"/>
1475  <x:anchor-alias value="Transfer-Encoding"/>
1477   The Transfer-Encoding header field lists the transfer coding names
1478   corresponding to the sequence of transfer codings that have been
1479   (or will be) applied to the payload body in order to form the message body.
1480   Transfer codings are defined in <xref target="transfer.codings"/>.
1482<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
1483  <x:ref>Transfer-Encoding</x:ref> = 1#<x:ref>transfer-coding</x:ref>
1486   Transfer-Encoding is analogous to the Content-Transfer-Encoding field of
1487   MIME, which was designed to enable safe transport of binary data over a
1488   7-bit transport service (<xref target="RFC2045" x:fmt="," x:sec="6"/>).
1489   However, safe transport has a different focus for an 8bit-clean transfer
1490   protocol. In HTTP's case, Transfer-Encoding is primarily intended to
1491   accurately delimit a dynamically generated payload and to distinguish
1492   payload encodings that are only applied for transport efficiency or
1493   security from those that are characteristics of the selected resource.
1496   All HTTP/1.1 recipients &MUST; implement the chunked transfer coding
1497   (<xref target="chunked.encoding"/>) because it plays a crucial role in
1498   framing messages when the payload body size is not known in advance.
1499   If chunked is applied to a payload body, the sender &MUST-NOT; apply
1500   chunked more than once (i.e., chunking an already chunked message is not
1501   allowed).
1502   If any transfer coding is applied to a request payload body, the
1503   sender &MUST; apply chunked as the final transfer coding to ensure that
1504   the message is properly framed.
1505   If any transfer coding is applied to a response payload body, the
1506   sender &MUST; either apply chunked as the final transfer coding or
1507   terminate the message by closing the connection.
1510   For example,
1511</preamble><artwork type="example">
1512  Transfer-Encoding: gzip, chunked
1514   indicates that the payload body has been compressed using the gzip
1515   coding and then chunked using the chunked coding while forming the
1516   message body.
1519   Unlike <x:ref>Content-Encoding</x:ref> (&content-codings;),
1520   Transfer-Encoding is a property of the message, not of the payload, and
1521   any recipient along the request/response chain &MAY; decode the received
1522   transfer coding(s) or apply additional transfer coding(s) to the message
1523   body, assuming that corresponding changes are made to the Transfer-Encoding
1524   field-value. Additional information about the encoding parameters &MAY; be
1525   provided by other header fields not defined by this specification.
1528   Transfer-Encoding &MAY; be sent in a response to a HEAD request or in a
1529   <x:ref>304 (Not Modified)</x:ref> response (&status-304;) to a GET request,
1530   neither of which includes a message body,
1531   to indicate that the origin server would have applied a transfer coding
1532   to the message body if the request had been an unconditional GET.
1533   This indication is not required, however, because any recipient on
1534   the response chain (including the origin server) can remove transfer
1535   codings when they are not needed.
1538   Transfer-Encoding was added in HTTP/1.1.  It is generally assumed that
1539   implementations advertising only HTTP/1.0 support will not understand
1540   how to process a transfer-encoded payload.
1541   A client &MUST-NOT; send a request containing Transfer-Encoding unless it
1542   knows the server will handle HTTP/1.1 (or later) requests; such knowledge
1543   might be in the form of specific user configuration or by remembering the
1544   version of a prior received response.
1545   A server &MUST-NOT; send a response containing Transfer-Encoding unless
1546   the corresponding request indicates HTTP/1.1 (or later).
1549   A server that receives a request message with a transfer coding it does
1550   not understand &SHOULD; respond with <x:ref>501 (Not Implemented)</x:ref>.
1554<section title="Content-Length" anchor="header.content-length">
1555  <iref primary="true" item="Content-Length header field" x:for-anchor=""/>
1556  <x:anchor-alias value="Content-Length"/>
1558   When a message does not have a <x:ref>Transfer-Encoding</x:ref> header
1559   field, a Content-Length header field can provide the anticipated size,
1560   as a decimal number of octets, for a potential payload body.
1561   For messages that do include a payload body, the Content-Length field-value
1562   provides the framing information necessary for determining where the body
1563   (and message) ends.  For messages that do not include a payload body, the
1564   Content-Length indicates the size of the selected representation
1565   (&representation;).
1567<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
1568  <x:ref>Content-Length</x:ref> = 1*<x:ref>DIGIT</x:ref>
1571   An example is
1573<figure><artwork type="example">
1574  Content-Length: 3495
1577   A sender &MUST-NOT; send a Content-Length header field in any message that
1578   contains a <x:ref>Transfer-Encoding</x:ref> header field.
1581   A user agent &SHOULD; send a Content-Length in a request message when no
1582   <x:ref>Transfer-Encoding</x:ref> is sent and the request method defines
1583   a meaning for an enclosed payload body. For example, a Content-Length
1584   header field is normally sent in a POST request even when the value is
1585   0 (indicating an empty payload body).  A user agent &SHOULD-NOT; send a
1586   Content-Length header field when the request message does not contain a
1587   payload body and the method semantics do not anticipate such a body.
1590   A server &MAY; send a Content-Length header field in a response to a HEAD
1591   request (&HEAD;); a server &MUST-NOT; send Content-Length in such a
1592   response unless its field-value equals the decimal number of octets that
1593   would have been sent in the payload body of a response if the same
1594   request had used the GET method.
1597   A server &MAY; send a Content-Length header field in a
1598   <x:ref>304 (Not Modified)</x:ref> response to a conditional GET request
1599   (&status-304;); a server &MUST-NOT; send Content-Length in such a
1600   response unless its field-value equals the decimal number of octets that
1601   would have been sent in the payload body of a <x:ref>200 (OK)</x:ref>
1602   response to the same request.
1605   A server &MUST-NOT; send a Content-Length header field in any response
1606   with a status code of
1607   <x:ref>1xx (Informational)</x:ref> or <x:ref>204 (No Content)</x:ref>.
1608   A server &SHOULD-NOT; send a Content-Length header field in any
1609   <x:ref>2xx (Successful)</x:ref> response to a CONNECT request (&CONNECT;).
1612   Aside from the cases defined above, in the absence of Transfer-Encoding,
1613   an origin server &SHOULD; send a Content-Length header field when the
1614   payload body size is known prior to sending the complete header block.
1615   This will allow downstream recipients to measure transfer progress,
1616   know when a received message is complete, and potentially reuse the
1617   connection for additional requests.
1620   Any Content-Length field value greater than or equal to zero is valid.
1621   Since there is no predefined limit to the length of an HTTP payload,
1622   recipients &SHOULD; anticipate potentially large decimal numerals and
1623   prevent parsing errors due to integer conversion overflows
1624   (<xref target="attack.protocol.element.size.overflows"/>).
1627   If a message is received that has multiple Content-Length header fields
1628   with field-values consisting of the same decimal value, or a single
1629   Content-Length header field with a field value containing a list of
1630   identical decimal values (e.g., "Content-Length: 42, 42"), indicating that
1631   duplicate Content-Length header fields have been generated or combined by an
1632   upstream message processor, then the recipient &MUST; either reject the
1633   message as invalid or replace the duplicated field-values with a single
1634   valid Content-Length field containing that decimal value prior to
1635   determining the message body length.
1638  <t>
1639   &Note; HTTP's use of Content-Length for message framing differs
1640   significantly from the same field's use in MIME, where it is an optional
1641   field used only within the "message/external-body" media-type.
1642  </t>
1646<section title="Message Body Length" anchor="message.body.length">
1647  <iref item="chunked (Coding Format)"/>
1649   The length of a message body is determined by one of the following
1650   (in order of precedence):
1653  <list style="numbers">
1654    <x:lt><t>
1655     Any response to a HEAD request and any response with a
1656     <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, or
1657     <x:ref>304 (Not Modified)</x:ref> status code is always
1658     terminated by the first empty line after the header fields, regardless of
1659     the header fields present in the message, and thus cannot contain a
1660     message body.
1661    </t></x:lt>
1662    <x:lt><t>
1663     Any <x:ref>2xx (Successful)</x:ref> response to a CONNECT request implies that the
1664     connection will become a tunnel immediately after the empty line that
1665     concludes the header fields.  A client &MUST; ignore any
1666     <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1667     fields received in such a message.
1668    </t></x:lt>
1669    <x:lt><t>
1670     If a <x:ref>Transfer-Encoding</x:ref> header field is present
1671     and the chunked transfer coding (<xref target="chunked.encoding"/>)
1672     is the final encoding, the message body length is determined by reading
1673     and decoding the chunked data until the transfer coding indicates the
1674     data is complete.
1675    </t>
1676    <t>
1677     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a
1678     response and the chunked transfer coding is not the final encoding, the
1679     message body length is determined by reading the connection until it is
1680     closed by the server.
1681     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a request and the
1682     chunked transfer coding is not the final encoding, the message body
1683     length cannot be determined reliably; the server &MUST; respond with
1684     the <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1685    </t>
1686    <t>
1687     If a message is received with both a <x:ref>Transfer-Encoding</x:ref>
1688     and a <x:ref>Content-Length</x:ref> header field, the Transfer-Encoding
1689     overrides the Content-Length. Such a message might indicate an attempt
1690     to perform request or response smuggling (bypass of security-related
1691     checks on message routing or content) and thus ought to be handled as
1692     an error.  A sender &MUST; remove the received Content-Length field
1693     prior to forwarding such a message downstream.
1694    </t></x:lt>
1695    <x:lt><t>
1696     If a message is received without <x:ref>Transfer-Encoding</x:ref> and with
1697     either multiple <x:ref>Content-Length</x:ref> header fields having
1698     differing field-values or a single Content-Length header field having an
1699     invalid value, then the message framing is invalid and &MUST; be treated
1700     as an error to prevent request or response smuggling.
1701     If this is a request message, the server &MUST; respond with
1702     a <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1703     If this is a response message received by a proxy, the proxy
1704     &MUST; discard the received response, send a <x:ref>502 (Bad Gateway)</x:ref>
1705     status code as its downstream response, and then close the connection.
1706     If this is a response message received by a user agent, it &MUST; be
1707     treated as an error by discarding the message and closing the connection.
1708    </t></x:lt>
1709    <x:lt><t>
1710     If a valid <x:ref>Content-Length</x:ref> header field is present without
1711     <x:ref>Transfer-Encoding</x:ref>, its decimal value defines the
1712     expected message body length in octets.
1713     If the sender closes the connection or the recipient times out before the
1714     indicated number of octets are received, the recipient &MUST; consider
1715     the message to be incomplete and close the connection.
1716    </t></x:lt>
1717    <x:lt><t>
1718     If this is a request message and none of the above are true, then the
1719     message body length is zero (no message body is present).
1720    </t></x:lt>
1721    <x:lt><t>
1722     Otherwise, this is a response message without a declared message body
1723     length, so the message body length is determined by the number of octets
1724     received prior to the server closing the connection.
1725    </t></x:lt>
1726  </list>
1729   Since there is no way to distinguish a successfully completed,
1730   close-delimited message from a partially-received message interrupted
1731   by network failure, a server &SHOULD; use encoding or
1732   length-delimited messages whenever possible.  The close-delimiting
1733   feature exists primarily for backwards compatibility with HTTP/1.0.
1736   A server &MAY; reject a request that contains a message body but
1737   not a <x:ref>Content-Length</x:ref> by responding with
1738   <x:ref>411 (Length Required)</x:ref>.
1741   Unless a transfer coding other than chunked has been applied,
1742   a client that sends a request containing a message body &SHOULD;
1743   use a valid <x:ref>Content-Length</x:ref> header field if the message body
1744   length is known in advance, rather than the chunked transfer coding, since some
1745   existing services respond to chunked with a <x:ref>411 (Length Required)</x:ref>
1746   status code even though they understand the chunked transfer coding.  This
1747   is typically because such services are implemented via a gateway that
1748   requires a content-length in advance of being called and the server
1749   is unable or unwilling to buffer the entire request before processing.
1752   A user agent that sends a request containing a message body &MUST; send a
1753   valid <x:ref>Content-Length</x:ref> header field if it does not know the
1754   server will handle HTTP/1.1 (or later) requests; such knowledge can be in
1755   the form of specific user configuration or by remembering the version of a
1756   prior received response.
1759   If the final response to the last request on a connection has been
1760   completely received and there remains additional data to read, a user agent
1761   &MAY; discard the remaining data or attempt to determine if that data
1762   belongs as part of the prior response body, which might be the case if the
1763   prior message's Content-Length value is incorrect. A client &MUST-NOT;
1764   process, cache, or forward such extra data as a separate response, since
1765   such behavior would be vulnerable to cache poisoning.
1770<section anchor="incomplete.messages" title="Handling Incomplete Messages">
1772   A server that receives an incomplete request message, usually due to a
1773   canceled request or a triggered time-out exception, &MAY; send an error
1774   response prior to closing the connection.
1777   A client that receives an incomplete response message, which can occur
1778   when a connection is closed prematurely or when decoding a supposedly
1779   chunked transfer coding fails, &MUST; record the message as incomplete.
1780   Cache requirements for incomplete responses are defined in
1781   &cache-incomplete;.
1784   If a response terminates in the middle of the header block (before the
1785   empty line is received) and the status code might rely on header fields to
1786   convey the full meaning of the response, then the client cannot assume
1787   that meaning has been conveyed; the client might need to repeat the
1788   request in order to determine what action to take next.
1791   A message body that uses the chunked transfer coding is
1792   incomplete if the zero-sized chunk that terminates the encoding has not
1793   been received.  A message that uses a valid <x:ref>Content-Length</x:ref> is
1794   incomplete if the size of the message body received (in octets) is less than
1795   the value given by Content-Length.  A response that has neither chunked
1796   transfer coding nor Content-Length is terminated by closure of the
1797   connection, and thus is considered complete regardless of the number of
1798   message body octets received, provided that the header block was received
1799   intact.
1803<section title="Message Parsing Robustness" anchor="message.robustness">
1805   Older HTTP/1.0 user agent implementations might send an extra CRLF
1806   after a POST request as a lame workaround for some early server
1807   applications that failed to read message body content that was
1808   not terminated by a line-ending. An HTTP/1.1 user agent &MUST-NOT;
1809   preface or follow a request with an extra CRLF.  If terminating
1810   the request message body with a line-ending is desired, then the
1811   user agent &MUST; count the terminating CRLF octets as part of the
1812   message body length.
1815   In the interest of robustness, servers &SHOULD; ignore at least one
1816   empty line received where a request-line is expected. In other words, if
1817   a server is reading the protocol stream at the beginning of a
1818   message and receives a CRLF first, the server &SHOULD; ignore the CRLF.
1821   Although the line terminator for the start-line and header
1822   fields is the sequence CRLF, recipients &MAY; recognize a
1823   single LF as a line terminator and ignore any preceding CR.
1826   Although the request-line and status-line grammar rules require that each
1827   of the component elements be separated by a single SP octet, recipients
1828   &MAY; instead parse on whitespace-delimited word boundaries and, aside
1829   from the CRLF terminator, treat any form of whitespace as the SP separator
1830   while ignoring preceding or trailing whitespace;
1831   such whitespace includes one or more of the following octets:
1832   SP, HTAB, VT (%x0B), FF (%x0C), or bare CR.
1835   When a server listening only for HTTP request messages, or processing
1836   what appears from the start-line to be an HTTP request message,
1837   receives a sequence of octets that does not match the HTTP-message
1838   grammar aside from the robustness exceptions listed above, the
1839   server &SHOULD; respond with a <x:ref>400 (Bad Request)</x:ref> response. 
1844<section title="Transfer Codings" anchor="transfer.codings">
1845  <x:anchor-alias value="transfer-coding"/>
1846  <x:anchor-alias value="transfer-extension"/>
1848   Transfer coding names are used to indicate an encoding
1849   transformation that has been, can be, or might need to be applied to a
1850   payload body in order to ensure "safe transport" through the network.
1851   This differs from a content coding in that the transfer coding is a
1852   property of the message rather than a property of the representation
1853   that is being transferred.
1855<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1856  <x:ref>transfer-coding</x:ref>    = "chunked" ; <xref target="chunked.encoding"/>
1857                     / "compress" ; <xref target="compress.coding"/>
1858                     / "deflate" ; <xref target="deflate.coding"/>
1859                     / "gzip" ; <xref target="gzip.coding"/>
1860                     / <x:ref>transfer-extension</x:ref>
1861  <x:ref>transfer-extension</x:ref> = <x:ref>token</x:ref> *( <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> <x:ref>transfer-parameter</x:ref> )
1863<t anchor="rule.parameter">
1864  <x:anchor-alias value="attribute"/>
1865  <x:anchor-alias value="transfer-parameter"/>
1866  <x:anchor-alias value="value"/>
1867   Parameters are in the form of attribute/value pairs.
1869<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-parameter"/><iref primary="true" item="Grammar" subitem="attribute"/><iref primary="true" item="Grammar" subitem="value"/><iref primary="true" item="Grammar" subitem="date2"/><iref primary="true" item="Grammar" subitem="date3"/>
1870  <x:ref>transfer-parameter</x:ref> = <x:ref>attribute</x:ref> <x:ref>BWS</x:ref> "=" <x:ref>BWS</x:ref> <x:ref>value</x:ref>
1871  <x:ref>attribute</x:ref>          = <x:ref>token</x:ref>
1872  <x:ref>value</x:ref>              = <x:ref>word</x:ref>
1875   All transfer-coding names are case-insensitive and ought to be registered
1876   within the HTTP Transfer Coding registry, as defined in
1877   <xref target="transfer.coding.registry"/>.
1878   They are used in the <x:ref>TE</x:ref> (<xref target="header.te"/>) and
1879   <x:ref>Transfer-Encoding</x:ref> (<xref target="header.transfer-encoding"/>)
1880   header fields.
1883<section title="Chunked Transfer Coding" anchor="chunked.encoding">
1884  <iref primary="true" item="chunked (Coding Format)"/>
1885  <x:anchor-alias value="chunk"/>
1886  <x:anchor-alias value="chunked-body"/>
1887  <x:anchor-alias value="chunk-data"/>
1888  <x:anchor-alias value="chunk-ext"/>
1889  <x:anchor-alias value="chunk-ext-name"/>
1890  <x:anchor-alias value="chunk-ext-val"/>
1891  <x:anchor-alias value="chunk-size"/>
1892  <x:anchor-alias value="last-chunk"/>
1893  <x:anchor-alias value="trailer-part"/>
1894  <x:anchor-alias value="quoted-str-nf"/>
1895  <x:anchor-alias value="qdtext-nf"/>
1897   The chunked transfer coding modifies the body of a message in order to
1898   transfer it as a series of chunks, each with its own size indicator,
1899   followed by an &OPTIONAL; trailer containing header fields. This
1900   allows dynamically generated content to be transferred along with the
1901   information necessary for the recipient to verify that it has
1902   received the full message.
1904<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="chunked-body"><!--terminal production--></iref><iref primary="true" item="Grammar" subitem="chunk"/><iref primary="true" item="Grammar" subitem="chunk-size"/><iref primary="true" item="Grammar" subitem="last-chunk"/><iref primary="true" item="Grammar" subitem="chunk-ext"/><iref primary="true" item="Grammar" subitem="chunk-ext-name"/><iref primary="true" item="Grammar" subitem="chunk-ext-val"/><iref primary="true" item="Grammar" subitem="chunk-data"/><iref primary="true" item="Grammar" subitem="trailer-part"/><iref primary="true" item="Grammar" subitem="quoted-str-nf"/><iref primary="true" item="Grammar" subitem="qdtext-nf"/>
1905  <x:ref>chunked-body</x:ref>   = *<x:ref>chunk</x:ref>
1906                   <x:ref>last-chunk</x:ref>
1907                   <x:ref>trailer-part</x:ref>
1908                   <x:ref>CRLF</x:ref>
1910  <x:ref>chunk</x:ref>          = <x:ref>chunk-size</x:ref> [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1911                   <x:ref>chunk-data</x:ref> <x:ref>CRLF</x:ref>
1912  <x:ref>chunk-size</x:ref>     = 1*<x:ref>HEXDIG</x:ref>
1913  <x:ref>last-chunk</x:ref>     = 1*("0") [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1915  <x:ref>chunk-ext</x:ref>      = *( ";" <x:ref>chunk-ext-name</x:ref> [ "=" <x:ref>chunk-ext-val</x:ref> ] )
1916  <x:ref>chunk-ext-name</x:ref> = <x:ref>token</x:ref>
1917  <x:ref>chunk-ext-val</x:ref>  = <x:ref>token</x:ref> / <x:ref>quoted-str-nf</x:ref>
1918  <x:ref>chunk-data</x:ref>     = 1*<x:ref>OCTET</x:ref> ; a sequence of chunk-size octets
1919  <x:ref>trailer-part</x:ref>   = *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
1921  <x:ref>quoted-str-nf</x:ref>  = <x:ref>DQUOTE</x:ref> *( <x:ref>qdtext-nf</x:ref> / <x:ref>quoted-pair</x:ref> ) <x:ref>DQUOTE</x:ref>
1922                 ; like <x:ref>quoted-string</x:ref>, but disallowing line folding
1923  <x:ref>qdtext-nf</x:ref>      = <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / %x21 / %x23-5B / %x5D-7E / <x:ref>obs-text</x:ref>
1926   Chunk extensions within the chunked transfer coding are deprecated.
1927   Senders &SHOULD-NOT; send chunk-ext.
1928   Definition of new chunk extensions is discouraged.
1931   The chunk-size field is a string of hex digits indicating the size of
1932   the chunk-data in octets. The chunked transfer coding is complete when a
1933   chunk with a chunk-size of zero is received, possibly followed by a
1934   trailer, and finally terminated by an empty line.
1937<section title="Trailer" anchor="header.trailer">
1938  <iref primary="true" item="Trailer header field" x:for-anchor=""/>
1939  <x:anchor-alias value="Trailer"/>
1941   A trailer allows the sender to include additional fields at the end of a
1942   chunked message in order to supply metadata that might be dynamically
1943   generated while the message body is sent, such as a message integrity
1944   check, digital signature, or post-processing status.
1945   The trailer &MUST-NOT; contain fields that need to be known before a
1946   recipient processes the body, such as <x:ref>Transfer-Encoding</x:ref>,
1947   <x:ref>Content-Length</x:ref>, and <x:ref>Trailer</x:ref>.
1950   When a message includes a message body encoded with the chunked
1951   transfer coding and the sender desires to send metadata in the form of
1952   trailer fields at the end of the message, the sender &SHOULD; send a
1953   <x:ref>Trailer</x:ref> header field before the message body to indicate
1954   which fields will be present in the trailers. This allows the recipient
1955   to prepare for receipt of that metadata before it starts processing the body,
1956   which is useful if the message is being streamed and the recipient wishes
1957   to confirm an integrity check on the fly.
1959<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
1960  <x:ref>Trailer</x:ref> = 1#<x:ref>field-name</x:ref>
1963   If no <x:ref>Trailer</x:ref> header field is present, the sender of a
1964   chunked message body &SHOULD; send an empty trailer.
1967   A server &MUST; send an empty trailer with the chunked transfer coding
1968   unless at least one of the following is true:
1969  <list style="numbers">
1970    <t>the request included a <x:ref>TE</x:ref> header field that indicates
1971    "trailers" is acceptable in the transfer coding of the response, as
1972    described in <xref target="header.te"/>; or,</t>
1974    <t>the trailer fields consist entirely of optional metadata and the
1975    recipient could use the message (in a manner acceptable to the server where
1976    the field originated) without receiving that metadata. In other words,
1977    the server that generated the header field is willing to accept the
1978    possibility that the trailer fields might be silently discarded along
1979    the path to the client.</t>
1980  </list>
1983   The above requirement prevents the need for an infinite buffer when a
1984   message is being received by an HTTP/1.1 (or later) proxy and forwarded to
1985   an HTTP/1.0 recipient.
1989<section title="Decoding chunked" anchor="decoding.chunked">
1991   A process for decoding the chunked transfer coding
1992   can be represented in pseudo-code as:
1994<figure><artwork type="code">
1995  length := 0
1996  read chunk-size, chunk-ext (if any) and CRLF
1997  while (chunk-size &gt; 0) {
1998     read chunk-data and CRLF
1999     append chunk-data to decoded-body
2000     length := length + chunk-size
2001     read chunk-size and CRLF
2002  }
2003  read header-field
2004  while (header-field not empty) {
2005     append header-field to existing header fields
2006     read header-field
2007  }
2008  Content-Length := length
2009  Remove "chunked" from Transfer-Encoding
2010  Remove Trailer from existing header fields
2013   All recipients &MUST; be able to receive and decode the
2014   chunked transfer coding and &MUST; ignore chunk-ext extensions
2015   they do not understand.
2020<section title="Compression Codings" anchor="compression.codings">
2022   The codings defined below can be used to compress the payload of a
2023   message.
2026<section title="Compress Coding" anchor="compress.coding">
2027<iref item="compress (Coding Format)"/>
2029   The "compress" format is produced by the common UNIX file compression
2030   program "compress". This format is an adaptive Lempel-Ziv-Welch
2031   coding (LZW). Recipients &SHOULD; consider "x-compress" to be
2032   equivalent to "compress".
2036<section title="Deflate Coding" anchor="deflate.coding">
2037<iref item="deflate (Coding Format)"/>
2039   The "deflate" format is defined as the "deflate" compression mechanism
2040   (described in <xref target="RFC1951"/>) used inside the "zlib"
2041   data format (<xref target="RFC1950"/>).
2044  <t>
2045    &Note; Some incorrect implementations send the "deflate"
2046    compressed data without the zlib wrapper.
2047   </t>
2051<section title="Gzip Coding" anchor="gzip.coding">
2052<iref item="gzip (Coding Format)"/>
2054   The "gzip" format is produced by the file compression program
2055   "gzip" (GNU zip), as described in <xref target="RFC1952"/>. This format is a
2056   Lempel-Ziv coding (LZ77) with a 32 bit CRC.
2057   Recipients &SHOULD; consider "x-gzip" to be equivalent to "gzip".
2063<section title="TE" anchor="header.te">
2064  <iref primary="true" item="TE header field" x:for-anchor=""/>
2065  <x:anchor-alias value="TE"/>
2066  <x:anchor-alias value="t-codings"/>
2067  <x:anchor-alias value="t-ranking"/>
2068  <x:anchor-alias value="rank"/>
2070   The "TE" header field in a request indicates what transfer codings,
2071   besides chunked, the client is willing to accept in response, and
2072   whether or not the client is willing to accept trailer fields in a
2073   chunked transfer coding.
2076   The TE field-value consists of a comma-separated list of transfer coding
2077   names, each allowing for optional parameters (as described in
2078   <xref target="transfer.codings"/>), and/or the keyword "trailers".
2079   Clients &MUST-NOT; send the chunked transfer coding name in TE;
2080   chunked is always acceptable for HTTP/1.1 recipients.
2082<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TE"/><iref primary="true" item="Grammar" subitem="t-codings"/><iref primary="true" item="Grammar" subitem="t-ranking"/><iref primary="true" item="Grammar" subitem="rank"/>
2083  <x:ref>TE</x:ref>        = #<x:ref>t-codings</x:ref>
2084  <x:ref>t-codings</x:ref> = "trailers" / ( <x:ref>transfer-coding</x:ref> [ <x:ref>t-ranking</x:ref> ] )
2085  <x:ref>t-ranking</x:ref> = <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> "q=" <x:ref>rank</x:ref>
2086  <x:ref>rank</x:ref>      = ( "0" [ "." 0*3<x:ref>DIGIT</x:ref> ] )
2087             / ( "1" [ "." 0*3("0") ] )
2090   Three examples of TE use are below.
2092<figure><artwork type="example">
2093  TE: deflate
2094  TE:
2095  TE: trailers, deflate;q=0.5
2098   The presence of the keyword "trailers" indicates that the client is
2099   willing to accept trailer fields in a chunked transfer coding,
2100   as defined in <xref target="chunked.encoding"/>, on behalf of itself and
2101   any downstream clients. For chained requests, this implies that either:
2102   (a) all downstream clients are willing to accept trailer fields in the
2103   forwarded response; or,
2104   (b) the client will attempt to buffer the response on behalf of downstream
2105   recipients.
2106   Note that HTTP/1.1 does not define any means to limit the size of a
2107   chunked response such that a client can be assured of buffering the
2108   entire response.
2111   When multiple transfer codings are acceptable, the client &MAY; rank the
2112   codings by preference using a case-insensitive "q" parameter (similar to
2113   the qvalues used in content negotiation fields, &qvalue;). The rank value
2114   is a real number in the range 0 through 1, where 0.001 is the least
2115   preferred and 1 is the most preferred; a value of 0 means "not acceptable".
2118   If the TE field-value is empty or if no TE field is present, the only
2119   acceptable transfer coding is chunked. A message with no transfer coding
2120   is always acceptable.
2123   Since the TE header field only applies to the immediate connection,
2124   a sender of TE &MUST; also send a "TE" connection option within the
2125   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
2126   in order to prevent the TE field from being forwarded by intermediaries
2127   that do not support its semantics.
2132<section title="Message Routing" anchor="message.routing">
2134   HTTP request message routing is determined by each client based on the
2135   target resource, the client's proxy configuration, and
2136   establishment or reuse of an inbound connection.  The corresponding
2137   response routing follows the same connection chain back to the client.
2140<section title="Identifying a Target Resource" anchor="target-resource">
2141  <iref primary="true" item="target resource"/>
2142  <iref primary="true" item="target URI"/>
2143  <x:anchor-alias value="target resource"/>
2144  <x:anchor-alias value="target URI"/>
2146   HTTP is used in a wide variety of applications, ranging from
2147   general-purpose computers to home appliances.  In some cases,
2148   communication options are hard-coded in a client's configuration.
2149   However, most HTTP clients rely on the same resource identification
2150   mechanism and configuration techniques as general-purpose Web browsers.
2153   HTTP communication is initiated by a user agent for some purpose.
2154   The purpose is a combination of request semantics, which are defined in
2155   <xref target="Part2"/>, and a target resource upon which to apply those
2156   semantics.  A URI reference (<xref target="uri"/>) is typically used as
2157   an identifier for the "<x:dfn>target resource</x:dfn>", which a user agent
2158   would resolve to its absolute form in order to obtain the
2159   "<x:dfn>target URI</x:dfn>".  The target URI
2160   excludes the reference's fragment identifier component, if any,
2161   since fragment identifiers are reserved for client-side processing
2162   (<xref target="RFC3986" x:fmt="," x:sec="3.5"/>).
2166<section title="Connecting Inbound" anchor="connecting.inbound">
2168   Once the target URI is determined, a client needs to decide whether
2169   a network request is necessary to accomplish the desired semantics and,
2170   if so, where that request is to be directed.
2173   If the client has a response cache and the request semantics can be
2174   satisfied by a cache (<xref target="Part6"/>), then the request is
2175   usually directed to the cache first.
2178   If the request is not satisfied by a cache, then a typical client will
2179   check its configuration to determine whether a proxy is to be used to
2180   satisfy the request.  Proxy configuration is implementation-dependent,
2181   but is often based on URI prefix matching, selective authority matching,
2182   or both, and the proxy itself is usually identified by an "http" or
2183   "https" URI.  If a proxy is applicable, the client connects inbound by
2184   establishing (or reusing) a connection to that proxy.
2187   If no proxy is applicable, a typical client will invoke a handler routine,
2188   usually specific to the target URI's scheme, to connect directly
2189   to an authority for the target resource.  How that is accomplished is
2190   dependent on the target URI scheme and defined by its associated
2191   specification, similar to how this specification defines origin server
2192   access for resolution of the "http" (<xref target="http.uri"/>) and
2193   "https" (<xref target="https.uri"/>) schemes.
2196   HTTP requirements regarding connection management are defined in
2197   <xref target=""/>.
2201<section title="Request Target" anchor="request-target">
2203   Once an inbound connection is obtained,
2204   the client sends an HTTP request message (<xref target="http.message"/>)
2205   with a request-target derived from the target URI.
2206   There are four distinct formats for the request-target, depending on both
2207   the method being requested and whether the request is to a proxy.
2209<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-target"/><iref primary="true" item="Grammar" subitem="origin-form"/><iref primary="true" item="Grammar" subitem="absolute-form"/><iref primary="true" item="Grammar" subitem="authority-form"/><iref primary="true" item="Grammar" subitem="asterisk-form"/>
2210  <x:ref>request-target</x:ref> = <x:ref>origin-form</x:ref>
2211                 / <x:ref>absolute-form</x:ref>
2212                 / <x:ref>authority-form</x:ref>
2213                 / <x:ref>asterisk-form</x:ref>
2215  <x:ref>origin-form</x:ref>    = <x:ref>path-absolute</x:ref> [ "?" <x:ref>query</x:ref> ]
2216  <x:ref>absolute-form</x:ref>  = <x:ref>absolute-URI</x:ref>
2217  <x:ref>authority-form</x:ref> = <x:ref>authority</x:ref>
2218  <x:ref>asterisk-form</x:ref>  = "*"
2220<t anchor="origin-form"><iref item="origin-form (of request-target)"/>
2221   The most common form of request-target is the origin-form.
2222   When making a request directly to an origin server, other than a CONNECT
2223   or server-wide OPTIONS request (as detailed below),
2224   a client &MUST; send only the absolute path and query components of
2225   the target URI as the request-target.
2226   If the target URI's path component is empty, then the client &MUST; send
2227   "/" as the path within the origin-form of request-target.
2228   A <x:ref>Host</x:ref> header field is also sent, as defined in
2229   <xref target=""/>, containing the target URI's
2230   authority component (excluding any userinfo).
2233   For example, a client wishing to retrieve a representation of the resource
2234   identified as
2236<figure><artwork x:indent-with="  " type="example">
2240   directly from the origin server would open (or reuse) a TCP connection
2241   to port 80 of the host "" and send the lines:
2243<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2244GET /where?q=now HTTP/1.1
2248   followed by the remainder of the request message.
2250<t anchor="absolute-form"><iref item="absolute-form (of request-target)"/>
2251   When making a request to a proxy, other than a CONNECT or server-wide
2252   OPTIONS request (as detailed below), a client &MUST; send the target URI
2253   in absolute-form as the request-target.
2254   The proxy is requested to either service that request from a valid cache,
2255   if possible, or make the same request on the client's behalf to either
2256   the next inbound proxy server or directly to the origin server indicated
2257   by the request-target.  Requirements on such "forwarding" of messages are
2258   defined in <xref target="message.forwarding"/>.
2261   An example absolute-form of request-line would be:
2263<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2264GET HTTP/1.1
2267   To allow for transition to the absolute-form for all requests in some
2268   future version of HTTP, HTTP/1.1 servers &MUST; accept the absolute-form
2269   in requests, even though HTTP/1.1 clients will only send them in requests
2270   to proxies.
2272<t anchor="authority-form"><iref item="authority-form (of request-target)"/>
2273   The authority-form of request-target is only used for CONNECT requests
2274   (&CONNECT;).  When making a CONNECT request to establish a tunnel through
2275   one or more proxies, a client &MUST; send only the target URI's
2276   authority component (excluding any userinfo) as the request-target.
2277   For example,
2279<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2282<t anchor="asterisk-form"><iref item="asterisk-form (of request-target)"/>
2283   The asterisk-form of request-target is only used for a server-wide
2284   OPTIONS request (&OPTIONS;).  When a client wishes to request OPTIONS
2285   for the server as a whole, as opposed to a specific named resource of
2286   that server, the client &MUST; send only "*" (%x2A) as the request-target.
2287   For example,
2289<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2290OPTIONS * HTTP/1.1
2293   If a proxy receives an OPTIONS request with an absolute-form of
2294   request-target in which the URI has an empty path and no query component,
2295   then the last proxy on the request chain &MUST; send a request-target
2296   of "*" when it forwards the request to the indicated origin server.
2299   For example, the request
2300</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2304  would be forwarded by the final proxy as
2305</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2306OPTIONS * HTTP/1.1
2310   after connecting to port 8001 of host "".
2315<section title="Host" anchor="">
2316  <iref primary="true" item="Host header field" x:for-anchor=""/>
2317  <x:anchor-alias value="Host"/>
2319   The "Host" header field in a request provides the host and port
2320   information from the target URI, enabling the origin
2321   server to distinguish among resources while servicing requests
2322   for multiple host names on a single IP address.  Since the Host
2323   field-value is critical information for handling a request, it
2324   &SHOULD; be sent as the first header field following the request-line.
2326<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2327  <x:ref>Host</x:ref> = <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ; <xref target="http.uri"/>
2330   A client &MUST; send a Host header field in all HTTP/1.1 request
2331   messages.  If the target URI includes an authority component, then
2332   the Host field-value &MUST; be identical to that authority component
2333   after excluding any userinfo (<xref target="http.uri"/>).
2334   If the authority component is missing or undefined for the target URI,
2335   then the Host header field &MUST; be sent with an empty field-value.
2338   For example, a GET request to the origin server for
2339   &lt;; would begin with:
2341<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2342GET /pub/WWW/ HTTP/1.1
2346   The Host header field &MUST; be sent in an HTTP/1.1 request even
2347   if the request-target is in the absolute-form, since this
2348   allows the Host information to be forwarded through ancient HTTP/1.0
2349   proxies that might not have implemented Host.
2352   When a proxy receives a request with an absolute-form of
2353   request-target, the proxy &MUST; ignore the received
2354   Host header field (if any) and instead replace it with the host
2355   information of the request-target.  If the proxy forwards the request,
2356   it &MUST; generate a new Host field-value based on the received
2357   request-target rather than forward the received Host field-value.
2360   Since the Host header field acts as an application-level routing
2361   mechanism, it is a frequent target for malware seeking to poison
2362   a shared cache or redirect a request to an unintended server.
2363   An interception proxy is particularly vulnerable if it relies on
2364   the Host field-value for redirecting requests to internal
2365   servers, or for use as a cache key in a shared cache, without
2366   first verifying that the intercepted connection is targeting a
2367   valid IP address for that host.
2370   A server &MUST; respond with a <x:ref>400 (Bad Request)</x:ref> status code
2371   to any HTTP/1.1 request message that lacks a Host header field and
2372   to any request message that contains more than one Host header field
2373   or a Host header field with an invalid field-value.
2377<section title="Effective Request URI" anchor="effective.request.uri">
2378  <iref primary="true" item="effective request URI"/>
2379  <x:anchor-alias value="effective request URI"/>
2381   A server that receives an HTTP request message &MUST; reconstruct
2382   the user agent's original target URI, based on the pieces of information
2383   learned from the request-target, <x:ref>Host</x:ref> header field, and
2384   connection context, in order to identify the intended target resource and
2385   properly service the request. The URI derived from this reconstruction
2386   process is referred to as the "<x:dfn>effective request URI</x:dfn>".
2389   For a user agent, the effective request URI is the target URI.
2392   If the request-target is in absolute-form, then the effective request URI
2393   is the same as the request-target.  Otherwise, the effective request URI
2394   is constructed as follows.
2397   If the request is received over a TLS-secured TCP connection,
2398   then the effective request URI's scheme is "https"; otherwise, the
2399   scheme is "http".
2402   If the request-target is in authority-form, then the effective
2403   request URI's authority component is the same as the request-target.
2404   Otherwise, if a <x:ref>Host</x:ref> header field is supplied with a
2405   non-empty field-value, then the authority component is the same as the
2406   Host field-value. Otherwise, the authority component is the concatenation of
2407   the default host name configured for the server, a colon (":"), and the
2408   connection's incoming TCP port number in decimal form.
2411   If the request-target is in authority-form or asterisk-form, then the
2412   effective request URI's combined path and query component is empty.
2413   Otherwise, the combined path and query component is the same as the
2414   request-target.
2417   The components of the effective request URI, once determined as above,
2418   can be combined into absolute-URI form by concatenating the scheme,
2419   "://", authority, and combined path and query component.
2423   Example 1: the following message received over an insecure TCP connection
2425<artwork type="example" x:indent-with="  ">
2426GET /pub/WWW/TheProject.html HTTP/1.1
2432  has an effective request URI of
2434<artwork type="example" x:indent-with="  ">
2440   Example 2: the following message received over a TLS-secured TCP connection
2442<artwork type="example" x:indent-with="  ">
2443OPTIONS * HTTP/1.1
2449  has an effective request URI of
2451<artwork type="example" x:indent-with="  ">
2456   An origin server that does not allow resources to differ by requested
2457   host &MAY; ignore the <x:ref>Host</x:ref> field-value and instead replace it
2458   with a configured server name when constructing the effective request URI.
2461   Recipients of an HTTP/1.0 request that lacks a <x:ref>Host</x:ref> header
2462   field &MAY; attempt to use heuristics (e.g., examination of the URI path for
2463   something unique to a particular host) in order to guess the
2464   effective request URI's authority component.
2468<section title="Associating a Response to a Request" anchor="">
2470   HTTP does not include a request identifier for associating a given
2471   request message with its corresponding one or more response messages.
2472   Hence, it relies on the order of response arrival to correspond exactly
2473   to the order in which requests are made on the same connection.
2474   More than one response message per request only occurs when one or more
2475   informational responses (<x:ref>1xx</x:ref>, see &status-1xx;) precede a
2476   final response to the same request.
2479   A client that has more than one outstanding request on a connection &MUST;
2480   maintain a list of outstanding requests in the order sent and &MUST;
2481   associate each received response message on that connection to the highest
2482   ordered request that has not yet received a final (non-<x:ref>1xx</x:ref>)
2483   response.
2487<section title="Message Forwarding" anchor="message.forwarding">
2489   As described in <xref target="intermediaries"/>, intermediaries can serve
2490   a variety of roles in the processing of HTTP requests and responses.
2491   Some intermediaries are used to improve performance or availability.
2492   Others are used for access control or to filter content.
2493   Since an HTTP stream has characteristics similar to a pipe-and-filter
2494   architecture, there are no inherent limits to the extent an intermediary
2495   can enhance (or interfere) with either direction of the stream.
2498   Intermediaries that forward a message &MUST; implement the
2499   <x:ref>Connection</x:ref> header field, as specified in
2500   <xref target="header.connection"/>, to exclude fields that are only
2501   intended for the incoming connection.
2504   In order to avoid request loops, a proxy that forwards requests to other
2505   proxies &MUST; be able to recognize and exclude all of its own server
2506   names, including any aliases, local variations, or literal IP addresses.
2509<section title="Via" anchor="header.via">
2510  <iref primary="true" item="Via header field" x:for-anchor=""/>
2511  <x:anchor-alias value="pseudonym"/>
2512  <x:anchor-alias value="received-by"/>
2513  <x:anchor-alias value="received-protocol"/>
2514  <x:anchor-alias value="Via"/>
2516   The "Via" header field &MUST; be sent by a proxy or gateway in forwarded
2517   messages to indicate the intermediate protocols and recipients between the
2518   user agent and the server on requests, and between the origin server and
2519   the client on responses. It is analogous to the "Received" field
2520   used by email systems (<xref target="RFC5322" x:fmt="of" x:sec="3.6.7"/>).
2521   Via is used in HTTP for tracking message forwards,
2522   avoiding request loops, and identifying the protocol capabilities of
2523   all senders along the request/response chain.
2525<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Via"/><iref primary="true" item="Grammar" subitem="received-protocol"/><iref primary="true" item="Grammar" subitem="protocol-name"/><iref primary="true" item="Grammar" subitem="protocol-version"/><iref primary="true" item="Grammar" subitem="received-by"/><iref primary="true" item="Grammar" subitem="pseudonym"/>
2526  <x:ref>Via</x:ref>               = 1#( <x:ref>received-protocol</x:ref> <x:ref>RWS</x:ref> <x:ref>received-by</x:ref>
2527                          [ <x:ref>RWS</x:ref> <x:ref>comment</x:ref> ] )
2528  <x:ref>received-protocol</x:ref> = [ <x:ref>protocol-name</x:ref> "/" ] <x:ref>protocol-version</x:ref>
2529                      ; see <xref target="header.upgrade"/>
2530  <x:ref>received-by</x:ref>       = ( <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ) / <x:ref>pseudonym</x:ref>
2531  <x:ref>pseudonym</x:ref>         = <x:ref>token</x:ref>
2534   The received-protocol indicates the protocol version of the message
2535   received by the server or client along each segment of the
2536   request/response chain. The received-protocol version is appended to
2537   the Via field value when the message is forwarded so that information
2538   about the protocol capabilities of upstream applications remains
2539   visible to all recipients.
2542   The protocol-name is excluded if and only if it would be "HTTP". The
2543   received-by field is normally the host and optional port number of a
2544   recipient server or client that subsequently forwarded the message.
2545   However, if the real host is considered to be sensitive information,
2546   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
2547   be assumed to be the default port of the received-protocol.
2550   Multiple Via field values represent each proxy or gateway that has
2551   forwarded the message. Each recipient &MUST; append its information
2552   such that the end result is ordered according to the sequence of
2553   forwarding applications.
2556   Comments &MAY; be used in the Via header field to identify the software
2557   of each recipient, analogous to the <x:ref>User-Agent</x:ref> and
2558   <x:ref>Server</x:ref> header fields. However, all comments in the Via field
2559   are optional and &MAY; be removed by any recipient prior to forwarding the
2560   message.
2563   For example, a request message could be sent from an HTTP/1.0 user
2564   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2565   forward the request to a public proxy at, which completes
2566   the request by forwarding it to the origin server at
2567   The request received by would then have the following
2568   Via header field:
2570<figure><artwork type="example">
2571  Via: 1.0 fred, 1.1 (Apache/1.1)
2574   A proxy or gateway used as a portal through a network firewall
2575   &SHOULD-NOT; forward the names and ports of hosts within the firewall
2576   region unless it is explicitly enabled to do so. If not enabled, the
2577   received-by host of any host behind the firewall &SHOULD; be replaced
2578   by an appropriate pseudonym for that host.
2581   A proxy or gateway &MAY; combine an ordered subsequence of Via header
2582   field entries into a single such entry if the entries have identical
2583   received-protocol values. For example,
2585<figure><artwork type="example">
2586  Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2589  could be collapsed to
2591<figure><artwork type="example">
2592  Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2595   Senders &SHOULD-NOT; combine multiple entries unless they are all
2596   under the same organizational control and the hosts have already been
2597   replaced by pseudonyms. Senders &MUST-NOT; combine entries that
2598   have different received-protocol values.
2602<section title="Transformations" anchor="message.transformations">
2604   Some intermediaries include features for transforming messages and their
2605   payloads.  A transforming proxy might, for example, convert between image
2606   formats in order to save cache space or to reduce the amount of traffic on
2607   a slow link. However, operational problems might occur when these
2608   transformations are applied to payloads intended for critical applications,
2609   such as medical imaging or scientific data analysis, particularly when
2610   integrity checks or digital signatures are used to ensure that the payload
2611   received is identical to the original.
2614   If a proxy receives a request-target with a host name that is not a
2615   fully qualified domain name, it &MAY; add its own domain to the host name
2616   it received when forwarding the request.  A proxy &MUST-NOT; change the
2617   host name if it is a fully qualified domain name.
2620   A proxy &MUST-NOT; modify the "path-absolute" and "query" parts of the
2621   received request-target when forwarding it to the next inbound server,
2622   except as noted above to replace an empty path with "/" or "*".
2625   A proxy &MUST-NOT; modify header fields that provide information about the
2626   end points of the communication chain, the resource state, or the selected
2627   representation. A proxy &MAY; change the message body through application
2628   or removal of a transfer coding (<xref target="transfer.codings"/>).
2631   A non-transforming proxy &MUST; preserve the message payload (&payload;).
2632   A transforming proxy &MUST; preserve the payload of a message that
2633   contains the no-transform cache-control directive.
2636   A transforming proxy &MAY; transform the payload of a message
2637   that does not contain the no-transform cache-control directive;
2638   if the payload is transformed, the transforming proxy &MUST; add a
2639   Warning 214 (Transformation applied) header field if one does not
2640   already appear in the message (see &header-warning;).
2646<section title="Connection Management" anchor="">
2648   HTTP messaging is independent of the underlying transport or
2649   session-layer connection protocol(s).  HTTP only presumes a reliable
2650   transport with in-order delivery of requests and the corresponding
2651   in-order delivery of responses.  The mapping of HTTP request and
2652   response structures onto the data units of an underlying transport
2653   protocol is outside the scope of this specification.
2656   As described in <xref target="connecting.inbound"/>, the specific
2657   connection protocols to be used for an HTTP interaction are determined by
2658   client configuration and the <x:ref>target URI</x:ref>.
2659   For example, the "http" URI scheme
2660   (<xref target="http.uri"/>) indicates a default connection of TCP
2661   over IP, with a default TCP port of 80, but the client might be
2662   configured to use a proxy via some other connection, port, or protocol.
2665   HTTP implementations are expected to engage in connection management,
2666   which includes maintaining the state of current connections,
2667   establishing a new connection or reusing an existing connection,
2668   processing messages received on a connection, detecting connection
2669   failures, and closing each connection.
2670   Most clients maintain multiple connections in parallel, including
2671   more than one connection per server endpoint.
2672   Most servers are designed to maintain thousands of concurrent connections,
2673   while controlling request queues to enable fair use and detect
2674   denial of service attacks.
2677<section title="Connection" anchor="header.connection">
2678  <iref primary="true" item="Connection header field" x:for-anchor=""/>
2679  <iref primary="true" item="close" x:for-anchor=""/>
2680  <x:anchor-alias value="Connection"/>
2681  <x:anchor-alias value="connection-option"/>
2682  <x:anchor-alias value="close"/>
2684   The "Connection" header field allows the sender to indicate desired
2685   control options for the current connection.  In order to avoid confusing
2686   downstream recipients, a proxy or gateway &MUST; remove or replace any
2687   received connection options before forwarding the message.
2690   When a header field aside from Connection is used to supply control
2691   information for or about the current connection, the sender &MUST; list
2692   the corresponding field-name within the "Connection" header field.
2693   A proxy or gateway &MUST; parse a received Connection
2694   header field before a message is forwarded and, for each
2695   connection-option in this field, remove any header field(s) from
2696   the message with the same name as the connection-option, and then
2697   remove the Connection header field itself (or replace it with the
2698   intermediary's own connection options for the forwarded message).
2701   Hence, the Connection header field provides a declarative way of
2702   distinguishing header fields that are only intended for the
2703   immediate recipient ("hop-by-hop") from those fields that are
2704   intended for all recipients on the chain ("end-to-end"), enabling the
2705   message to be self-descriptive and allowing future connection-specific
2706   extensions to be deployed without fear that they will be blindly
2707   forwarded by older intermediaries.
2710   The Connection header field's value has the following grammar:
2712<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-option"/>
2713  <x:ref>Connection</x:ref>        = 1#<x:ref>connection-option</x:ref>
2714  <x:ref>connection-option</x:ref> = <x:ref>token</x:ref>
2717   Connection options are case-insensitive.
2720   A sender &MUST-NOT; send a connection option corresponding to a header
2721   field that is intended for all recipients of the payload.
2722   For example, <x:ref>Cache-Control</x:ref> is never appropriate as a
2723   connection option (&header-cache-control;).
2726   The connection options do not have to correspond to a header field
2727   present in the message, since a connection-specific header field
2728   might not be needed if there are no parameters associated with that
2729   connection option.  Recipients that trigger certain connection
2730   behavior based on the presence of connection options &MUST; do so
2731   based on the presence of the connection-option rather than only the
2732   presence of the optional header field.  In other words, if the
2733   connection option is received as a header field but not indicated
2734   within the Connection field-value, then the recipient &MUST; ignore
2735   the connection-specific header field because it has likely been
2736   forwarded by an intermediary that is only partially conformant.
2739   When defining new connection options, specifications ought to
2740   carefully consider existing deployed header fields and ensure
2741   that the new connection option does not share the same name as
2742   an unrelated header field that might already be deployed.
2743   Defining a new connection option essentially reserves that potential
2744   field-name for carrying additional information related to the
2745   connection option, since it would be unwise for senders to use
2746   that field-name for anything else.
2749   The "<x:dfn>close</x:dfn>" connection option is defined for a
2750   sender to signal that this connection will be closed after completion of
2751   the response. For example,
2753<figure><artwork type="example">
2754  Connection: close
2757   in either the request or the response header fields indicates that
2758   the connection &MUST; be closed after the current request/response
2759   is complete (<xref target="persistent.tear-down"/>).
2762   A client that does not support <x:ref>persistent connections</x:ref> &MUST;
2763   send the "close" connection option in every request message.
2766   A server that does not support <x:ref>persistent connections</x:ref> &MUST;
2767   send the "close" connection option in every response message that
2768   does not have a <x:ref>1xx (Informational)</x:ref> status code.
2772<section title="Establishment" anchor="persistent.establishment">
2774   It is beyond the scope of this specification to describe how connections
2775   are established via various transport or session-layer protocols.
2776   Each connection applies to only one transport link.
2780<section title="Persistence" anchor="persistent.connections">
2781   <x:anchor-alias value="persistent connections"/>
2783   HTTP/1.1 defaults to the use of "<x:dfn>persistent connections</x:dfn>",
2784   allowing multiple requests and responses to be carried over a single
2785   connection. The "<x:ref>close</x:ref>" connection-option is used to signal
2786   that a connection will not persist after the current request/response.
2787   HTTP implementations &SHOULD; support persistent connections.
2790   A recipient determines whether a connection is persistent or not based on
2791   the most recently received message's protocol version and
2792   <x:ref>Connection</x:ref> header field (if any):
2793   <list style="symbols">
2794     <t>If the <x:ref>close</x:ref> connection option is present, the
2795        connection will not persist after the current response; else,</t>
2796     <t>If the received protocol is HTTP/1.1 (or later), the connection will
2797        persist after the current response; else,</t>
2798     <t>If the received protocol is HTTP/1.0, the "keep-alive"
2799        connection option is present, the recipient is not a proxy, and
2800        the recipient wishes to honor the HTTP/1.0 "keep-alive" mechanism,
2801        the connection will persist after the current response; otherwise,</t>
2802     <t>The connection will close after the current response.</t>
2803   </list>
2806   A server &MAY; assume that an HTTP/1.1 client intends to maintain a
2807   persistent connection until a <x:ref>close</x:ref> connection option
2808   is received in a request.
2811   A client &MAY; reuse a persistent connection until it sends or receives
2812   a <x:ref>close</x:ref> connection option or receives an HTTP/1.0 response
2813   without a "keep-alive" connection option.
2816   In order to remain persistent, all messages on a connection &MUST;
2817   have a self-defined message length (i.e., one not defined by closure
2818   of the connection), as described in <xref target="message.body"/>.
2819   A server &MUST; read the entire request message body or close
2820   the connection after sending its response, since otherwise the
2821   remaining data on a persistent connection would be misinterpreted
2822   as the next request.  Likewise,
2823   a client &MUST; read the entire response message body if it intends
2824   to reuse the same connection for a subsequent request.
2827   A proxy server &MUST-NOT; maintain a persistent connection with an
2828   HTTP/1.0 client (see <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/> for
2829   information and discussion of the problems with the Keep-Alive header field
2830   implemented by many HTTP/1.0 clients).
2833   Clients and servers &SHOULD-NOT; assume that a persistent connection is
2834   maintained for HTTP versions less than 1.1 unless it is explicitly
2835   signaled.
2836   See <xref target="compatibility.with.http.1.0.persistent.connections"/>
2837   for more information on backward compatibility with HTTP/1.0 clients.
2840<section title="Retrying Requests" anchor="persistent.retrying.requests">
2842   Connections can be closed at any time, with or without intention.
2843   Implementations ought to anticipate the need to recover
2844   from asynchronous close events.
2847   When an inbound connection is closed prematurely, a client &MAY; open a new
2848   connection and automatically retransmit an aborted sequence of requests if
2849   all of those requests have idempotent methods (&idempotent-methods;).
2850   A proxy &MUST-NOT; automatically retry non-idempotent requests.
2853   A user agent &MUST-NOT; automatically retry a request with a non-idempotent
2854   method unless it has some means to know that the request semantics are
2855   actually idempotent, regardless of the method, or some means to detect that
2856   the original request was never applied. For example, a user agent that
2857   knows (through design or configuration) that a POST request to a given
2858   resource is safe can repeat that request automatically.
2859   Likewise, a user agent designed specifically to operate on a version
2860   control repository might be able to recover from partial failure conditions
2861   by checking the target resource revision(s) after a failed connection,
2862   reverting or fixing any changes that were partially applied, and then
2863   automatically retrying the requests that failed.
2866   An automatic retry &SHOULD-NOT; be repeated if it fails.
2870<section title="Pipelining" anchor="pipelining">
2871   <x:anchor-alias value="pipeline"/>
2873   A client that supports persistent connections &MAY; "<x:dfn>pipeline</x:dfn>"
2874   its requests (i.e., send multiple requests without waiting for each
2875   response). A server &MAY; process a sequence of pipelined requests in
2876   parallel if they all have safe methods (&safe-methods;), but &MUST; send
2877   the corresponding responses in the same order that the requests were
2878   received.
2881   A client that pipelines requests &MUST; be prepared to retry those
2882   requests if the connection closes before it receives all of the
2883   corresponding responses. A client that assumes a persistent connection and
2884   pipelines immediately after connection establishment &MUST-NOT; pipeline
2885   on a retry connection until it knows the connection is persistent.
2888   Idempotent methods (&idempotent-methods;) are significant to pipelining
2889   because they can be automatically retried after a connection failure.
2890   A user agent &SHOULD-NOT; pipeline requests after a non-idempotent method
2891   until the final response status code for that method has been received,
2892   unless the user agent has a means to detect and recover from partial
2893   failure conditions involving the pipelined sequence.
2896   An intermediary that receives pipelined requests &MAY; pipeline those
2897   requests when forwarding them inbound, since it can rely on the outbound
2898   user agent(s) to determine what requests can be safely pipelined. If the
2899   inbound connection fails before receiving a response, the pipelining
2900   intermediary &MAY; attempt to retry a sequence of requests that have yet
2901   to receive a response if the requests all have idempotent methods;
2902   otherwise, the pipelining intermediary &SHOULD; forward any received
2903   responses and then close the corresponding outbound connection(s) so that
2904   the outbound user agent(s) can recover accordingly.
2909<section title="Concurrency" anchor="persistent.concurrency">
2911   Clients &SHOULD; limit the number of simultaneous
2912   connections that they maintain to a given server.
2915   Previous revisions of HTTP gave a specific number of connections as a
2916   ceiling, but this was found to be impractical for many applications. As a
2917   result, this specification does not mandate a particular maximum number of
2918   connections, but instead encourages clients to be conservative when opening
2919   multiple connections.
2922   Multiple connections are typically used to avoid the "head-of-line
2923   blocking" problem, wherein a request that takes significant server-side
2924   processing and/or has a large payload blocks subsequent requests on the
2925   same connection. However, each connection consumes server resources.
2926   Furthermore, using multiple connections can cause undesirable side effects
2927   in congested networks.
2930   Note that servers might reject traffic that they deem abusive, including an
2931   excessive number of connections from a client.
2935<section title="Failures and Time-outs" anchor="persistent.failures">
2937   Servers will usually have some time-out value beyond which they will
2938   no longer maintain an inactive connection. Proxy servers might make
2939   this a higher value since it is likely that the client will be making
2940   more connections through the same server. The use of persistent
2941   connections places no requirements on the length (or existence) of
2942   this time-out for either the client or the server.
2945   When a client or server wishes to time-out it &SHOULD; issue a graceful
2946   close on the transport connection. Clients and servers &SHOULD; both
2947   constantly watch for the other side of the transport close, and
2948   respond to it as appropriate. If a client or server does not detect
2949   the other side's close promptly it could cause unnecessary resource
2950   drain on the network.
2953   A client, server, or proxy &MAY; close the transport connection at any
2954   time. For example, a client might have started to send a new request
2955   at the same time that the server has decided to close the "idle"
2956   connection. From the server's point of view, the connection is being
2957   closed while it was idle, but from the client's point of view, a
2958   request is in progress.
2961   Servers &SHOULD; maintain persistent connections and allow the underlying
2962   transport's flow control mechanisms to resolve temporary overloads, rather
2963   than terminate connections with the expectation that clients will retry.
2964   The latter technique can exacerbate network congestion.
2967   A client sending a message body &SHOULD; monitor
2968   the network connection for an error status code while it is transmitting
2969   the request. If the client sees an error status code, it &SHOULD;
2970   immediately cease transmitting the body and close the connection.
2974<section title="Tear-down" anchor="persistent.tear-down">
2975  <iref primary="false" item="Connection header field" x:for-anchor=""/>
2976  <iref primary="false" item="close" x:for-anchor=""/>
2978   The <x:ref>Connection</x:ref> header field
2979   (<xref target="header.connection"/>) provides a "<x:ref>close</x:ref>"
2980   connection option that a sender &SHOULD; send when it wishes to close
2981   the connection after the current request/response pair.
2984   A client that sends a <x:ref>close</x:ref> connection option &MUST-NOT;
2985   send further requests on that connection (after the one containing
2986   <x:ref>close</x:ref>) and &MUST; close the connection after reading the
2987   final response message corresponding to this request.
2990   A server that receives a <x:ref>close</x:ref> connection option &MUST;
2991   initiate a lingering close (see below) of the connection after it sends the
2992   final response to the request that contained <x:ref>close</x:ref>.
2993   The server &SHOULD; send a <x:ref>close</x:ref> connection option
2994   in its final response on that connection. The server &MUST-NOT; process
2995   any further requests received on that connection.
2998   A server that sends a <x:ref>close</x:ref> connection option &MUST;
2999   initiate a lingering close of the connection after it sends the
3000   response containing <x:ref>close</x:ref>. The server &MUST-NOT; process
3001   any further requests received on that connection.
3004   A client that receives a <x:ref>close</x:ref> connection option &MUST;
3005   cease sending requests on that connection and close the connection
3006   after reading the response message containing the close; if additional
3007   pipelined requests had been sent on the connection, the client &SHOULD;
3008   assume that they will not be processed by the server.
3011   If a server performs an immediate close of a TCP connection, there is a
3012   significant risk that the client will not be able to read the last HTTP
3013   response.  If the server receives additional data from the client on a
3014   fully-closed connection, such as another request that was sent by the
3015   client before receiving the server's response, the server's TCP stack will
3016   send a reset packet to the client; unfortunately, the reset packet might
3017   erase the client's unacknowledged input buffers before they can be read
3018   and interpreted by the client's HTTP parser.
3021   To avoid the TCP reset problem, a server can perform a lingering close on a
3022   connection by closing only the write side of the read/write connection
3023   (a half-close) and continuing to read from the connection until the
3024   connection is closed by the client or the server is reasonably certain
3025   that its own TCP stack has received the client's acknowledgement of the
3026   packet(s) containing the server's last response. It is then safe for the
3027   server to fully close the connection.
3030   It is unknown whether the reset problem is exclusive to TCP or might also
3031   be found in other transport connection protocols.
3035<section title="Upgrade" anchor="header.upgrade">
3036  <iref primary="true" item="Upgrade header field" x:for-anchor=""/>
3037  <x:anchor-alias value="Upgrade"/>
3038  <x:anchor-alias value="protocol"/>
3039  <x:anchor-alias value="protocol-name"/>
3040  <x:anchor-alias value="protocol-version"/>
3042   The "Upgrade" header field is intended to provide a simple mechanism
3043   for transitioning from HTTP/1.1 to some other protocol on the same
3044   connection.  A client &MAY; send a list of protocols in the Upgrade
3045   header field of a request to invite the server to switch to one or
3046   more of those protocols before sending the final response.
3047   A server &MUST; send an Upgrade header field in <x:ref>101 (Switching
3048   Protocols)</x:ref> responses to indicate which protocol(s) are being
3049   switched to, and &MUST; send it in <x:ref>426 (Upgrade Required)</x:ref>
3050   responses to indicate acceptable protocols.
3051   A server &MAY; send an Upgrade header field in any other response to
3052   indicate that they might be willing to upgrade to one of the
3053   specified protocols for a future request.
3055<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
3056  <x:ref>Upgrade</x:ref>          = 1#<x:ref>protocol</x:ref>
3058  <x:ref>protocol</x:ref>         = <x:ref>protocol-name</x:ref> ["/" <x:ref>protocol-version</x:ref>]
3059  <x:ref>protocol-name</x:ref>    = <x:ref>token</x:ref>
3060  <x:ref>protocol-version</x:ref> = <x:ref>token</x:ref>
3063   For example,
3065<figure><artwork type="example">
3066  Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
3069   Upgrade eases the difficult transition between incompatible protocols by
3070   allowing the client to initiate a request in the more commonly
3071   supported protocol while indicating to the server that it would like
3072   to use a "better" protocol if available (where "better" is determined
3073   by the server, possibly according to the nature of the request method
3074   or target resource).
3077   Upgrade cannot be used to insist on a protocol change; its acceptance and
3078   use by the server is optional. The capabilities and nature of the
3079   application-level communication after the protocol change is entirely
3080   dependent upon the new protocol chosen, although the first action
3081   after changing the protocol &MUST; be a response to the initial HTTP
3082   request that contained the Upgrade header field.
3085   For example, if the Upgrade header field is received in a GET request
3086   and the server decides to switch protocols, then it first responds
3087   with a <x:ref>101 (Switching Protocols)</x:ref> message in HTTP/1.1 and
3088   then immediately follows that with the new protocol's equivalent of a
3089   response to a GET on the target resource.  This allows a connection to be
3090   upgraded to protocols with the same semantics as HTTP without the
3091   latency cost of an additional round-trip.  A server &MUST-NOT; switch
3092   protocols unless the received message semantics can be honored by the new
3093   protocol; an OPTIONS request can be honored by any protocol.
3096   When Upgrade is sent, a sender &MUST; also send a
3097   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
3098   that contains the "upgrade" connection option, in order to prevent Upgrade
3099   from being accidentally forwarded by intermediaries that might not implement
3100   the listed protocols.  A server &MUST; ignore an Upgrade header field that
3101   is received in an HTTP/1.0 request.
3104   The Upgrade header field only applies to switching application-level
3105   protocols on the existing connection; it cannot be used
3106   to switch to a protocol on a different connection. For that purpose, it is
3107   more appropriate to use a <x:ref>3xx (Redirection)</x:ref> response
3108   (&status-3xx;).
3111   This specification only defines the protocol name "HTTP" for use by
3112   the family of Hypertext Transfer Protocols, as defined by the HTTP
3113   version rules of <xref target="http.version"/> and future updates to this
3114   specification. Additional tokens ought to be registered with IANA using the
3115   registration procedure defined in <xref target="upgrade.token.registry"/>.
3120<section title="IANA Considerations" anchor="IANA.considerations">
3122<section title="Header Field Registration" anchor="header.field.registration">
3124   HTTP header fields are registered within the Message Header Field Registry
3125   <xref target="BCP90"/> maintained by IANA at
3126   <eref target=""/>.
3129   This document defines the following HTTP header fields, so their
3130   associated registry entries shall be updated according to the permanent
3131   registrations below:
3133<?BEGININC p1-messaging.iana-headers ?>
3134<!--AUTOGENERATED FROM extract-header-defs.xslt, do not edit manually-->
3135<texttable align="left" suppress-title="true" anchor="iana.header.registration.table">
3136   <ttcol>Header Field Name</ttcol>
3137   <ttcol>Protocol</ttcol>
3138   <ttcol>Status</ttcol>
3139   <ttcol>Reference</ttcol>
3141   <c>Connection</c>
3142   <c>http</c>
3143   <c>standard</c>
3144   <c>
3145      <xref target="header.connection"/>
3146   </c>
3147   <c>Content-Length</c>
3148   <c>http</c>
3149   <c>standard</c>
3150   <c>
3151      <xref target="header.content-length"/>
3152   </c>
3153   <c>Host</c>
3154   <c>http</c>
3155   <c>standard</c>
3156   <c>
3157      <xref target=""/>
3158   </c>
3159   <c>TE</c>
3160   <c>http</c>
3161   <c>standard</c>
3162   <c>
3163      <xref target="header.te"/>
3164   </c>
3165   <c>Trailer</c>
3166   <c>http</c>
3167   <c>standard</c>
3168   <c>
3169      <xref target="header.trailer"/>
3170   </c>
3171   <c>Transfer-Encoding</c>
3172   <c>http</c>
3173   <c>standard</c>
3174   <c>
3175      <xref target="header.transfer-encoding"/>
3176   </c>
3177   <c>Upgrade</c>
3178   <c>http</c>
3179   <c>standard</c>
3180   <c>
3181      <xref target="header.upgrade"/>
3182   </c>
3183   <c>Via</c>
3184   <c>http</c>
3185   <c>standard</c>
3186   <c>
3187      <xref target="header.via"/>
3188   </c>
3191<?ENDINC p1-messaging.iana-headers ?>
3193   Furthermore, the header field-name "Close" shall be registered as
3194   "reserved", since using that name as an HTTP header field might
3195   conflict with the "close" connection option of the "<x:ref>Connection</x:ref>"
3196   header field (<xref target="header.connection"/>).
3198<texttable align="left" suppress-title="true">
3199   <ttcol>Header Field Name</ttcol>
3200   <ttcol>Protocol</ttcol>
3201   <ttcol>Status</ttcol>
3202   <ttcol>Reference</ttcol>
3204   <c>Close</c>
3205   <c>http</c>
3206   <c>reserved</c>
3207   <c>
3208      <xref target="header.field.registration"/>
3209   </c>
3212   The change controller is: "IETF ( - Internet Engineering Task Force".
3216<section title="URI Scheme Registration" anchor="uri.scheme.registration">
3218   IANA maintains the registry of URI Schemes <xref target="BCP115"/> at
3219   <eref target=""/>.
3222   This document defines the following URI schemes, so their
3223   associated registry entries shall be updated according to the permanent
3224   registrations below:
3226<texttable align="left" suppress-title="true">
3227   <ttcol>URI Scheme</ttcol>
3228   <ttcol>Description</ttcol>
3229   <ttcol>Reference</ttcol>
3231   <c>http</c>
3232   <c>Hypertext Transfer Protocol</c>
3233   <c><xref target="http.uri"/></c>
3235   <c>https</c>
3236   <c>Hypertext Transfer Protocol Secure</c>
3237   <c><xref target="https.uri"/></c>
3241<section title="Internet Media Type Registration" anchor="">
3243   This document serves as the specification for the Internet media types
3244   "message/http" and "application/http". The following is to be registered with
3245   IANA (see <xref target="BCP13"/>).
3247<section title="Internet Media Type message/http" anchor="">
3248<iref item="Media Type" subitem="message/http" primary="true"/>
3249<iref item="message/http Media Type" primary="true"/>
3251   The message/http type can be used to enclose a single HTTP request or
3252   response message, provided that it obeys the MIME restrictions for all
3253   "message" types regarding line length and encodings.
3256  <list style="hanging" x:indent="12em">
3257    <t hangText="Type name:">
3258      message
3259    </t>
3260    <t hangText="Subtype name:">
3261      http
3262    </t>
3263    <t hangText="Required parameters:">
3264      none
3265    </t>
3266    <t hangText="Optional parameters:">
3267      version, msgtype
3268      <list style="hanging">
3269        <t hangText="version:">
3270          The HTTP-version number of the enclosed message
3271          (e.g., "1.1"). If not present, the version can be
3272          determined from the first line of the body.
3273        </t>
3274        <t hangText="msgtype:">
3275          The message type &mdash; "request" or "response". If not
3276          present, the type can be determined from the first
3277          line of the body.
3278        </t>
3279      </list>
3280    </t>
3281    <t hangText="Encoding considerations:">
3282      only "7bit", "8bit", or "binary" are permitted
3283    </t>
3284    <t hangText="Security considerations:">
3285      none
3286    </t>
3287    <t hangText="Interoperability considerations:">
3288      none
3289    </t>
3290    <t hangText="Published specification:">
3291      This specification (see <xref target=""/>).
3292    </t>
3293    <t hangText="Applications that use this media type:">
3294    </t>
3295    <t hangText="Additional information:">
3296      <list style="hanging">
3297        <t hangText="Magic number(s):">none</t>
3298        <t hangText="File extension(s):">none</t>
3299        <t hangText="Macintosh file type code(s):">none</t>
3300      </list>
3301    </t>
3302    <t hangText="Person and email address to contact for further information:">
3303      See Authors Section.
3304    </t>
3305    <t hangText="Intended usage:">
3306      COMMON
3307    </t>
3308    <t hangText="Restrictions on usage:">
3309      none
3310    </t>
3311    <t hangText="Author/Change controller:">
3312      IESG
3313    </t>
3314  </list>
3317<section title="Internet Media Type application/http" anchor="">
3318<iref item="Media Type" subitem="application/http" primary="true"/>
3319<iref item="application/http Media Type" primary="true"/>
3321   The application/http type can be used to enclose a pipeline of one or more
3322   HTTP request or response messages (not intermixed).
3325  <list style="hanging" x:indent="12em">
3326    <t hangText="Type name:">
3327      application
3328    </t>
3329    <t hangText="Subtype name:">
3330      http
3331    </t>
3332    <t hangText="Required parameters:">
3333      none
3334    </t>
3335    <t hangText="Optional parameters:">
3336      version, msgtype
3337      <list style="hanging">
3338        <t hangText="version:">
3339          The HTTP-version number of the enclosed messages
3340          (e.g., "1.1"). If not present, the version can be
3341          determined from the first line of the body.
3342        </t>
3343        <t hangText="msgtype:">
3344          The message type &mdash; "request" or "response". If not
3345          present, the type can be determined from the first
3346          line of the body.
3347        </t>
3348      </list>
3349    </t>
3350    <t hangText="Encoding considerations:">
3351      HTTP messages enclosed by this type
3352      are in "binary" format; use of an appropriate
3353      Content-Transfer-Encoding is required when
3354      transmitted via E-mail.
3355    </t>
3356    <t hangText="Security considerations:">
3357      none
3358    </t>
3359    <t hangText="Interoperability considerations:">
3360      none
3361    </t>
3362    <t hangText="Published specification:">
3363      This specification (see <xref target=""/>).
3364    </t>
3365    <t hangText="Applications that use this media type:">
3366    </t>
3367    <t hangText="Additional information:">
3368      <list style="hanging">
3369        <t hangText="Magic number(s):">none</t>
3370        <t hangText="File extension(s):">none</t>
3371        <t hangText="Macintosh file type code(s):">none</t>
3372      </list>
3373    </t>
3374    <t hangText="Person and email address to contact for further information:">
3375      See Authors Section.
3376    </t>
3377    <t hangText="Intended usage:">
3378      COMMON
3379    </t>
3380    <t hangText="Restrictions on usage:">
3381      none
3382    </t>
3383    <t hangText="Author/Change controller:">
3384      IESG
3385    </t>
3386  </list>
3391<section title="Transfer Coding Registry" anchor="transfer.coding.registry">
3393   The HTTP Transfer Coding Registry defines the name space for transfer
3394   coding names.
3397   Registrations &MUST; include the following fields:
3398   <list style="symbols">
3399     <t>Name</t>
3400     <t>Description</t>
3401     <t>Pointer to specification text</t>
3402   </list>
3405   Names of transfer codings &MUST-NOT; overlap with names of content codings
3406   (&content-codings;) unless the encoding transformation is identical, as
3407   is the case for the compression codings defined in
3408   <xref target="compression.codings"/>.
3411   Values to be added to this name space require IETF Review (see
3412   <xref target="RFC5226" x:fmt="of" x:sec="4.1"/>), and &MUST;
3413   conform to the purpose of transfer coding defined in this section.
3414   Use of program names for the identification of encoding formats
3415   is not desirable and is discouraged for future encodings.
3418   The registry itself is maintained at
3419   <eref target=""/>.
3423<section title="Transfer Coding Registration" anchor="transfer.coding.registration">
3425   The HTTP Transfer Coding Registry shall be updated with the registrations
3426   below:
3428<texttable align="left" suppress-title="true" anchor="iana.transfer.coding.registration.table">
3429   <ttcol>Name</ttcol>
3430   <ttcol>Description</ttcol>
3431   <ttcol>Reference</ttcol>
3432   <c>chunked</c>
3433   <c>Transfer in a series of chunks</c>
3434   <c>
3435      <xref target="chunked.encoding"/>
3436   </c>
3437   <c>compress</c>
3438   <c>UNIX "compress" program method</c>
3439   <c>
3440      <xref target="compress.coding"/>
3441   </c>
3442   <c>deflate</c>
3443   <c>"deflate" compression mechanism (<xref target="RFC1951"/>) used inside
3444   the "zlib" data format (<xref target="RFC1950"/>)
3445   </c>
3446   <c>
3447      <xref target="deflate.coding"/>
3448   </c>
3449   <c>gzip</c>
3450   <c>Same as GNU zip <xref target="RFC1952"/></c>
3451   <c>
3452      <xref target="gzip.coding"/>
3453   </c>
3457<section title="Upgrade Token Registry" anchor="upgrade.token.registry">
3459   The HTTP Upgrade Token Registry defines the name space for protocol-name
3460   tokens used to identify protocols in the <x:ref>Upgrade</x:ref> header
3461   field. Each registered protocol name is associated with contact information
3462   and an optional set of specifications that details how the connection
3463   will be processed after it has been upgraded.
3466   Registrations happen on a "First Come First Served" basis (see
3467   <xref target="RFC5226" x:sec="4.1" x:fmt="of"/>) and are subject to the
3468   following rules:
3469  <list style="numbers">
3470    <t>A protocol-name token, once registered, stays registered forever.</t>
3471    <t>The registration &MUST; name a responsible party for the
3472       registration.</t>
3473    <t>The registration &MUST; name a point of contact.</t>
3474    <t>The registration &MAY; name a set of specifications associated with
3475       that token. Such specifications need not be publicly available.</t>
3476    <t>The registration &SHOULD; name a set of expected "protocol-version"
3477       tokens associated with that token at the time of registration.</t>
3478    <t>The responsible party &MAY; change the registration at any time.
3479       The IANA will keep a record of all such changes, and make them
3480       available upon request.</t>
3481    <t>The IESG &MAY; reassign responsibility for a protocol token.
3482       This will normally only be used in the case when a
3483       responsible party cannot be contacted.</t>
3484  </list>
3487   This registration procedure for HTTP Upgrade Tokens replaces that
3488   previously defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
3492<section title="Upgrade Token Registration" anchor="upgrade.token.registration">
3494   The HTTP Upgrade Token Registry shall be updated with the registration
3495   below:
3497<texttable align="left" suppress-title="true">
3498   <ttcol>Value</ttcol>
3499   <ttcol>Description</ttcol>
3500   <ttcol>Expected Version Tokens</ttcol>
3501   <ttcol>Reference</ttcol>
3503   <c>HTTP</c>
3504   <c>Hypertext Transfer Protocol</c>
3505   <c>any DIGIT.DIGIT (e.g, "2.0")</c>
3506   <c><xref target="http.version"/></c>
3509   The responsible party is: "IETF ( - Internet Engineering Task Force".
3515<section title="Security Considerations" anchor="security.considerations">
3517   This section is meant to inform developers, information providers, and
3518   users of known security concerns relevant to HTTP/1.1 message syntax,
3519   parsing, and routing.
3522<section title="DNS-related Attacks" anchor="dns.related.attacks">
3524   HTTP clients rely heavily on the Domain Name Service (DNS), and are thus
3525   generally prone to security attacks based on the deliberate misassociation
3526   of IP addresses and DNS names not protected by DNSSEC. Clients need to be
3527   cautious in assuming the validity of an IP number/DNS name association unless
3528   the response is protected by DNSSEC (<xref target="RFC4033"/>).
3532<section title="Intermediaries and Caching" anchor="attack.intermediaries">
3534   By their very nature, HTTP intermediaries are men-in-the-middle, and
3535   represent an opportunity for man-in-the-middle attacks. Compromise of
3536   the systems on which the intermediaries run can result in serious security
3537   and privacy problems. Intermediaries have access to security-related
3538   information, personal information about individual users and
3539   organizations, and proprietary information belonging to users and
3540   content providers. A compromised intermediary, or an intermediary
3541   implemented or configured without regard to security and privacy
3542   considerations, might be used in the commission of a wide range of
3543   potential attacks.
3546   Intermediaries that contain a shared cache are especially vulnerable
3547   to cache poisoning attacks.
3550   Implementers need to consider the privacy and security
3551   implications of their design and coding decisions, and of the
3552   configuration options they provide to operators (especially the
3553   default configuration).
3556   Users need to be aware that intermediaries are no more trustworthy than
3557   the people who run them; HTTP itself cannot solve this problem.
3561<section title="Buffer Overflows" anchor="attack.protocol.element.size.overflows">
3563   Because HTTP uses mostly textual, character-delimited fields, attackers can
3564   overflow buffers in implementations, and/or perform a Denial of Service
3565   against implementations that accept fields with unlimited lengths.
3568   To promote interoperability, this specification makes specific
3569   recommendations for minimum size limits on request-line
3570   (<xref target="request.line"/>)
3571   and blocks of header fields (<xref target="header.fields"/>). These are
3572   minimum recommendations, chosen to be supportable even by implementations
3573   with limited resources; it is expected that most implementations will
3574   choose substantially higher limits.
3577   This specification also provides a way for servers to reject messages that
3578   have request-targets that are too long (&status-414;) or request entities
3579   that are too large (&status-4xx;).
3582   Recipients &SHOULD; carefully limit the extent to which they read other
3583   fields, including (but not limited to) request methods, response status
3584   phrases, header field-names, and body chunks, so as to avoid denial of
3585   service attacks without impeding interoperability.
3589<section title="Message Integrity" anchor="message.integrity">
3591   HTTP does not define a specific mechanism for ensuring message integrity,
3592   instead relying on the error-detection ability of underlying transport
3593   protocols and the use of length or chunk-delimited framing to detect
3594   completeness. Additional integrity mechanisms, such as hash functions or
3595   digital signatures applied to the content, can be selectively added to
3596   messages via extensible metadata header fields. Historically, the lack of
3597   a single integrity mechanism has been justified by the informal nature of
3598   most HTTP communication.  However, the prevalence of HTTP as an information
3599   access mechanism has resulted in its increasing use within environments
3600   where verification of message integrity is crucial.
3603   User agents are encouraged to implement configurable means for detecting
3604   and reporting failures of message integrity such that those means can be
3605   enabled within environments for which integrity is necessary. For example,
3606   a browser being used to view medical history or drug interaction
3607   information needs to indicate to the user when such information is detected
3608   by the protocol to be incomplete, expired, or corrupted during transfer.
3609   Such mechanisms might be selectively enabled via user agent extensions or
3610   the presence of message integrity metadata in a response.
3611   At a minimum, user agents ought to provide some indication that allows a
3612   user to distinguish between a complete and incomplete response message
3613   (<xref target="incomplete.messages"/>) when such verification is desired.
3617<section title="Server Log Information" anchor="abuse.of.server.log.information">
3619   A server is in the position to save personal data about a user's requests
3620   over time, which might identify their reading patterns or subjects of
3621   interest.  In particular, log information gathered at an intermediary
3622   often contains a history of user agent interaction, across a multitude
3623   of sites, that can be traced to individual users.
3626   HTTP log information is confidential in nature; its handling is often
3627   constrained by laws and regulations.  Log information needs to be securely
3628   stored and appropriate guidelines followed for its analysis.
3629   Anonymization of personal information within individual entries helps,
3630   but is generally not sufficient to prevent real log traces from being
3631   re-identified based on correlation with other access characteristics.
3632   As such, access traces that are keyed to a specific client should not
3633   be published even if the key is pseudonymous.
3636   To minimize the risk of theft or accidental publication, log information
3637   should be purged of personally identifiable information, including
3638   user identifiers, IP addresses, and user-provided query parameters,
3639   as soon as that information is no longer necessary to support operational
3640   needs for security, auditing, or fraud control.
3645<section title="Acknowledgments" anchor="acks">
3647   This edition of HTTP/1.1 builds on the many contributions that went into
3648   <xref target="RFC1945" format="none">RFC 1945</xref>,
3649   <xref target="RFC2068" format="none">RFC 2068</xref>,
3650   <xref target="RFC2145" format="none">RFC 2145</xref>, and
3651   <xref target="RFC2616" format="none">RFC 2616</xref>, including
3652   substantial contributions made by the previous authors, editors, and
3653   working group chairs: Tim Berners-Lee, Ari Luotonen, Roy T. Fielding,
3654   Henrik Frystyk Nielsen, Jim Gettys, Jeffrey C. Mogul, Larry Masinter,
3655   and Paul J. Leach. Mark Nottingham oversaw this effort as working group chair.
3658   Since 1999, the following contributors have helped improve the HTTP
3659   specification by reporting bugs, asking smart questions, drafting or
3660   reviewing text, and evaluating open issues:
3662<?BEGININC acks ?>
3663<t>Adam Barth,
3664Adam Roach,
3665Addison Phillips,
3666Adrian Chadd,
3667Adrien W. de Croy,
3668Alan Ford,
3669Alan Ruttenberg,
3670Albert Lunde,
3671Alek Storm,
3672Alex Rousskov,
3673Alexandre Morgaut,
3674Alexey Melnikov,
3675Alisha Smith,
3676Amichai Rothman,
3677Amit Klein,
3678Amos Jeffries,
3679Andreas Maier,
3680Andreas Petersson,
3681Anil Sharma,
3682Anne van Kesteren,
3683Anthony Bryan,
3684Asbjorn Ulsberg,
3685Ashok Kumar,
3686Balachander Krishnamurthy,
3687Barry Leiba,
3688Ben Laurie,
3689Benjamin Niven-Jenkins,
3690Bil Corry,
3691Bill Burke,
3692Bjoern Hoehrmann,
3693Bob Scheifler,
3694Boris Zbarsky,
3695Brett Slatkin,
3696Brian Kell,
3697Brian McBarron,
3698Brian Pane,
3699Brian Smith,
3700Bryce Nesbitt,
3701Cameron Heavon-Jones,
3702Carl Kugler,
3703Carsten Bormann,
3704Charles Fry,
3705Chris Newman,
3706Chris Weber,
3707Cyrus Daboo,
3708Dale Robert Anderson,
3709Dan Wing,
3710Dan Winship,
3711Daniel Stenberg,
3712Darrel Miller,
3713Dave Cridland,
3714Dave Crocker,
3715Dave Kristol,
3716David Booth,
3717David Singer,
3718David W. Morris,
3719Diwakar Shetty,
3720Dmitry Kurochkin,
3721Drummond Reed,
3722Duane Wessels,
3723Duncan Cragg,
3724Edward Lee,
3725Eliot Lear,
3726Eran Hammer-Lahav,
3727Eric D. Williams,
3728Eric J. Bowman,
3729Eric Lawrence,
3730Eric Rescorla,
3731Erik Aronesty,
3732Evan Prodromou,
3733Florian Weimer,
3734Frank Ellermann,
3735Fred Bohle,
3736Gabriel Montenegro,
3737Geoffrey Sneddon,
3738Gervase Markham,
3739Grahame Grieve,
3740Greg Wilkins,
3741Harald Tveit Alvestrand,
3742Harry Halpin,
3743Helge Hess,
3744Henrik Nordstrom,
3745Henry S. Thompson,
3746Henry Story,
3747Herbert van de Sompel,
3748Howard Melman,
3749Hugo Haas,
3750Ian Fette,
3751Ian Hickson,
3752Ido Safruti,
3753Ilya Grigorik,
3754Ingo Struck,
3755J. Ross Nicoll,
3756James H. Manger,
3757James Lacey,
3758James M. Snell,
3759Jamie Lokier,
3760Jan Algermissen,
3761Jeff Hodges (who came up with the term 'effective Request-URI'),
3762Jeff Walden,
3763Jeroen de Borst,
3764Jim Luther,
3765Joe D. Williams,
3766Joe Gregorio,
3767Joe Orton,
3768John C. Klensin,
3769John C. Mallery,
3770John Cowan,
3771John Kemp,
3772John Panzer,
3773John Schneider,
3774John Stracke,
3775John Sullivan,
3776Jonas Sicking,
3777Jonathan A. Rees,
3778Jonathan Billington,
3779Jonathan Moore,
3780Jonathan Rees,
3781Jonathan Silvera,
3782Jordi Ros,
3783Joris Dobbelsteen,
3784Josh Cohen,
3785Julien Pierre,
3786Jungshik Shin,
3787Justin Chapweske,
3788Justin Erenkrantz,
3789Justin James,
3790Kalvinder Singh,
3791Karl Dubost,
3792Keith Hoffman,
3793Keith Moore,
3794Ken Murchison,
3795Koen Holtman,
3796Konstantin Voronkov,
3797Kris Zyp,
3798Lisa Dusseault,
3799Maciej Stachowiak,
3800Marc Schneider,
3801Marc Slemko,
3802Mark Baker,
3803Mark Pauley,
3804Mark Watson,
3805Markus Isomaki,
3806Markus Lanthaler,
3807Martin J. Duerst,
3808Martin Musatov,
3809Martin Nilsson,
3810Martin Thomson,
3811Matt Lynch,
3812Matthew Cox,
3813Max Clark,
3814Michael Burrows,
3815Michael Hausenblas,
3816Mike Amundsen,
3817Mike Belshe,
3818Mike Kelly,
3819Mike Schinkel,
3820Miles Sabin,
3821Murray S. Kucherawy,
3822Mykyta Yevstifeyev,
3823Nathan Rixham,
3824Nicholas Shanks,
3825Nico Williams,
3826Nicolas Alvarez,
3827Nicolas Mailhot,
3828Noah Slater,
3829Pablo Castro,
3830Pat Hayes,
3831Patrick R. McManus,
3832Patrik Faltstrom,
3833Paul E. Jones,
3834Paul Hoffman,
3835Paul Marquess,
3836Peter Lepeska,
3837Peter Saint-Andre,
3838Peter Watkins,
3839Phil Archer,
3840Philippe Mougin,
3841Phillip Hallam-Baker,
3842Poul-Henning Kamp,
3843Preethi Natarajan,
3844Rajeev Bector,
3845Ray Polk,
3846Reto Bachmann-Gmuer,
3847Richard Cyganiak,
3848Robert Brewer,
3849Robert Collins,
3850Robert O'Callahan,
3851Robert Olofsson,
3852Robert Sayre,
3853Robert Siemer,
3854Robert de Wilde,
3855Roberto Javier Godoy,
3856Roberto Peon,
3857Roland Zink,
3858Ronny Widjaja,
3859S. Mike Dierken,
3860Salvatore Loreto,
3861Sam Johnston,
3862Sam Ruby,
3863Scott Lawrence (who maintained the original issues list),
3864Sean B. Palmer,
3865Shane McCarron,
3866Stefan Eissing,
3867Stefan Tilkov,
3868Stefanos Harhalakis,
3869Stephane Bortzmeyer,
3870Stephen Farrell,
3871Stephen Ludin,
3872Stuart Williams,
3873Subbu Allamaraju,
3874Subramanian Moonesamy,
3875Sylvain Hellegouarch,
3876Tapan Divekar,
3877Tatsuya Hayashi,
3878Ted Hardie,
3879Thomas Broyer,
3880Thomas Fossati,
3881Thomas Nordin,
3882Thomas Roessler,
3883Tim Bray,
3884Tim Morgan,
3885Tim Olsen,
3886Tobias Oberstein,
3887Tom Zhou,
3888Travis Snoozy,
3889Tyler Close,
3890Vincent Murphy,
3891Wenbo Zhu,
3892Werner Baumann,
3893Wilbur Streett,
3894Wilfredo Sanchez Vega,
3895William A. Rowe Jr.,
3896William Chan,
3897Willy Tarreau,
3898Xiaoshu Wang,
3899Yaron Goland,
3900Yngve Nysaeter Pettersen,
3901Yoav Nir,
3902Yogesh Bang,
3903Yutaka Oiwa,
3904Yves Lafon (long-time member of the editor team),
3905Zed A. Shaw, and
3906Zhong Yu.
3908<?ENDINC acks ?>
3910   See <xref target="RFC2616" x:fmt="of" x:sec="16"/> for additional
3911   acknowledgements from prior revisions.
3918<references title="Normative References">
3920<reference anchor="Part2">
3921  <front>
3922    <title>Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content</title>
3923    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3924      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3925      <address><email></email></address>
3926    </author>
3927    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3928      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3929      <address><email></email></address>
3930    </author>
3931    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3932  </front>
3933  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-&ID-VERSION;"/>
3934  <x:source href="p2-semantics.xml" basename="p2-semantics">
3935    <x:defines>1xx (Informational)</x:defines>
3936    <x:defines>1xx</x:defines>
3937    <x:defines>100 (Continue)</x:defines>
3938    <x:defines>101 (Switching Protocols)</x:defines>
3939    <x:defines>2xx (Successful)</x:defines>
3940    <x:defines>2xx</x:defines>
3941    <x:defines>200 (OK)</x:defines>
3942    <x:defines>204 (No Content)</x:defines>
3943    <x:defines>3xx (Redirection)</x:defines>
3944    <x:defines>3xx</x:defines>
3945    <x:defines>301 (Moved Permanently)</x:defines>
3946    <x:defines>4xx (Client Error)</x:defines>
3947    <x:defines>4xx</x:defines>
3948    <x:defines>400 (Bad Request)</x:defines>
3949    <x:defines>411 (Length Required)</x:defines>
3950    <x:defines>414 (URI Too Long)</x:defines>
3951    <x:defines>417 (Expectation Failed)</x:defines>
3952    <x:defines>426 (Upgrade Required)</x:defines>
3953    <x:defines>501 (Not Implemented)</x:defines>
3954    <x:defines>502 (Bad Gateway)</x:defines>
3955    <x:defines>505 (HTTP Version Not Supported)</x:defines>
3956    <x:defines>Allow</x:defines>
3957    <x:defines>Content-Encoding</x:defines>
3958    <x:defines>Content-Location</x:defines>
3959    <x:defines>Content-Type</x:defines>
3960    <x:defines>Date</x:defines>
3961    <x:defines>Expect</x:defines>
3962    <x:defines>Location</x:defines>
3963    <x:defines>Server</x:defines>
3964    <x:defines>User-Agent</x:defines>
3965  </x:source>
3968<reference anchor="Part4">
3969  <front>
3970    <title>Hypertext Transfer Protocol (HTTP/1.1): Conditional Requests</title>
3971    <author fullname="Roy T. Fielding" initials="R." role="editor" surname="Fielding">
3972      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3973      <address><email></email></address>
3974    </author>
3975    <author fullname="Julian F. Reschke" initials="J. F." role="editor" surname="Reschke">
3976      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3977      <address><email></email></address>
3978    </author>
3979    <date month="&ID-MONTH;" year="&ID-YEAR;" />
3980  </front>
3981  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p4-conditional-&ID-VERSION;" />
3982  <x:source basename="p4-conditional" href="p4-conditional.xml">
3983    <x:defines>304 (Not Modified)</x:defines>
3984    <x:defines>ETag</x:defines>
3985    <x:defines>Last-Modified</x:defines>
3986  </x:source>
3989<reference anchor="Part5">
3990  <front>
3991    <title>Hypertext Transfer Protocol (HTTP/1.1): Range Requests</title>
3992    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3993      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3994      <address><email></email></address>
3995    </author>
3996    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3997      <organization abbrev="W3C">World Wide Web Consortium</organization>
3998      <address><email></email></address>
3999    </author>
4000    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4001      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4002      <address><email></email></address>
4003    </author>
4004    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4005  </front>
4006  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-&ID-VERSION;"/>
4007  <x:source href="p5-range.xml" basename="p5-range">
4008    <x:defines>Content-Range</x:defines>
4009  </x:source>
4012<reference anchor="Part6">
4013  <front>
4014    <title>Hypertext Transfer Protocol (HTTP/1.1): Caching</title>
4015    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4016      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4017      <address><email></email></address>
4018    </author>
4019    <author initials="M." surname="Nottingham" fullname="Mark Nottingham" role="editor">
4020      <organization>Akamai</organization>
4021      <address><email></email></address>
4022    </author>
4023    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4024      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4025      <address><email></email></address>
4026    </author>
4027    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4028  </front>
4029  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-&ID-VERSION;"/>
4030  <x:source href="p6-cache.xml" basename="p6-cache">
4031    <x:defines>Cache-Control</x:defines>
4032    <x:defines>Expires</x:defines>
4033  </x:source>
4036<reference anchor="Part7">
4037  <front>
4038    <title>Hypertext Transfer Protocol (HTTP/1.1): Authentication</title>
4039    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4040      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4041      <address><email></email></address>
4042    </author>
4043    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4044      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4045      <address><email></email></address>
4046    </author>
4047    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4048  </front>
4049  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p7-auth-&ID-VERSION;"/>
4050  <x:source href="p7-auth.xml" basename="p7-auth">
4051    <x:defines>Proxy-Authenticate</x:defines>
4052    <x:defines>Proxy-Authorization</x:defines>
4053  </x:source>
4056<reference anchor="RFC5234">
4057  <front>
4058    <title abbrev="ABNF for Syntax Specifications">Augmented BNF for Syntax Specifications: ABNF</title>
4059    <author initials="D." surname="Crocker" fullname="Dave Crocker" role="editor">
4060      <organization>Brandenburg InternetWorking</organization>
4061      <address>
4062        <email></email>
4063      </address> 
4064    </author>
4065    <author initials="P." surname="Overell" fullname="Paul Overell">
4066      <organization>THUS plc.</organization>
4067      <address>
4068        <email></email>
4069      </address>
4070    </author>
4071    <date month="January" year="2008"/>
4072  </front>
4073  <seriesInfo name="STD" value="68"/>
4074  <seriesInfo name="RFC" value="5234"/>
4077<reference anchor="RFC2119">
4078  <front>
4079    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
4080    <author initials="S." surname="Bradner" fullname="Scott Bradner">
4081      <organization>Harvard University</organization>
4082      <address><email></email></address>
4083    </author>
4084    <date month="March" year="1997"/>
4085  </front>
4086  <seriesInfo name="BCP" value="14"/>
4087  <seriesInfo name="RFC" value="2119"/>
4090<reference anchor="RFC3986">
4091 <front>
4092  <title abbrev='URI Generic Syntax'>Uniform Resource Identifier (URI): Generic Syntax</title>
4093  <author initials='T.' surname='Berners-Lee' fullname='Tim Berners-Lee'>
4094    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
4095    <address>
4096       <email></email>
4097       <uri></uri>
4098    </address>
4099  </author>
4100  <author initials='R.' surname='Fielding' fullname='Roy T. Fielding'>
4101    <organization abbrev="Day Software">Day Software</organization>
4102    <address>
4103      <email></email>
4104      <uri></uri>
4105    </address>
4106  </author>
4107  <author initials='L.' surname='Masinter' fullname='Larry Masinter'>
4108    <organization abbrev="Adobe Systems">Adobe Systems Incorporated</organization>
4109    <address>
4110      <email></email>
4111      <uri></uri>
4112    </address>
4113  </author>
4114  <date month='January' year='2005'></date>
4115 </front>
4116 <seriesInfo name="STD" value="66"/>
4117 <seriesInfo name="RFC" value="3986"/>
4120<reference anchor="USASCII">
4121  <front>
4122    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
4123    <author>
4124      <organization>American National Standards Institute</organization>
4125    </author>
4126    <date year="1986"/>
4127  </front>
4128  <seriesInfo name="ANSI" value="X3.4"/>
4131<reference anchor="RFC1950">
4132  <front>
4133    <title>ZLIB Compressed Data Format Specification version 3.3</title>
4134    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4135      <organization>Aladdin Enterprises</organization>
4136      <address><email></email></address>
4137    </author>
4138    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly"/>
4139    <date month="May" year="1996"/>
4140  </front>
4141  <seriesInfo name="RFC" value="1950"/>
4142  <!--<annotation>
4143    RFC 1950 is an Informational RFC, thus it might be less stable than
4144    this specification. On the other hand, this downward reference was
4145    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4146    therefore it is unlikely to cause problems in practice. See also
4147    <xref target="BCP97"/>.
4148  </annotation>-->
4151<reference anchor="RFC1951">
4152  <front>
4153    <title>DEFLATE Compressed Data Format Specification version 1.3</title>
4154    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4155      <organization>Aladdin Enterprises</organization>
4156      <address><email></email></address>
4157    </author>
4158    <date month="May" year="1996"/>
4159  </front>
4160  <seriesInfo name="RFC" value="1951"/>
4161  <!--<annotation>
4162    RFC 1951 is an Informational RFC, thus it might be less stable than
4163    this specification. On the other hand, this downward reference was
4164    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4165    therefore it is unlikely to cause problems in practice. See also
4166    <xref target="BCP97"/>.
4167  </annotation>-->
4170<reference anchor="RFC1952">
4171  <front>
4172    <title>GZIP file format specification version 4.3</title>
4173    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4174      <organization>Aladdin Enterprises</organization>
4175      <address><email></email></address>
4176    </author>
4177    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly">
4178      <address><email></email></address>
4179    </author>
4180    <author initials="M." surname="Adler" fullname="Mark Adler">
4181      <address><email></email></address>
4182    </author>
4183    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4184      <address><email></email></address>
4185    </author>
4186    <author initials="G." surname="Randers-Pehrson" fullname="Glenn Randers-Pehrson">
4187      <address><email></email></address>
4188    </author>
4189    <date month="May" year="1996"/>
4190  </front>
4191  <seriesInfo name="RFC" value="1952"/>
4192  <!--<annotation>
4193    RFC 1952 is an Informational RFC, thus it might be less stable than
4194    this specification. On the other hand, this downward reference was
4195    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4196    therefore it is unlikely to cause problems in practice. See also
4197    <xref target="BCP97"/>.
4198  </annotation>-->
4203<references title="Informative References">
4205<reference anchor="ISO-8859-1">
4206  <front>
4207    <title>
4208     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
4209    </title>
4210    <author>
4211      <organization>International Organization for Standardization</organization>
4212    </author>
4213    <date year="1998"/>
4214  </front>
4215  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
4218<reference anchor='RFC1919'>
4219  <front>
4220    <title>Classical versus Transparent IP Proxies</title>
4221    <author initials='M.' surname='Chatel' fullname='Marc Chatel'>
4222      <address><email></email></address>
4223    </author>
4224    <date year='1996' month='March' />
4225  </front>
4226  <seriesInfo name='RFC' value='1919' />
4229<reference anchor="RFC1945">
4230  <front>
4231    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
4232    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4233      <organization>MIT, Laboratory for Computer Science</organization>
4234      <address><email></email></address>
4235    </author>
4236    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4237      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4238      <address><email></email></address>
4239    </author>
4240    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4241      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
4242      <address><email></email></address>
4243    </author>
4244    <date month="May" year="1996"/>
4245  </front>
4246  <seriesInfo name="RFC" value="1945"/>
4249<reference anchor="RFC2045">
4250  <front>
4251    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
4252    <author initials="N." surname="Freed" fullname="Ned Freed">
4253      <organization>Innosoft International, Inc.</organization>
4254      <address><email></email></address>
4255    </author>
4256    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
4257      <organization>First Virtual Holdings</organization>
4258      <address><email></email></address>
4259    </author>
4260    <date month="November" year="1996"/>
4261  </front>
4262  <seriesInfo name="RFC" value="2045"/>
4265<reference anchor="RFC2047">
4266  <front>
4267    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
4268    <author initials="K." surname="Moore" fullname="Keith Moore">
4269      <organization>University of Tennessee</organization>
4270      <address><email></email></address>
4271    </author>
4272    <date month="November" year="1996"/>
4273  </front>
4274  <seriesInfo name="RFC" value="2047"/>
4277<reference anchor="RFC2068">
4278  <front>
4279    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4280    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
4281      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4282      <address><email></email></address>
4283    </author>
4284    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4285      <organization>MIT Laboratory for Computer Science</organization>
4286      <address><email></email></address>
4287    </author>
4288    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
4289      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
4290      <address><email></email></address>
4291    </author>
4292    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4293      <organization>MIT Laboratory for Computer Science</organization>
4294      <address><email></email></address>
4295    </author>
4296    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4297      <organization>MIT Laboratory for Computer Science</organization>
4298      <address><email></email></address>
4299    </author>
4300    <date month="January" year="1997"/>
4301  </front>
4302  <seriesInfo name="RFC" value="2068"/>
4305<reference anchor="RFC2145">
4306  <front>
4307    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
4308    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
4309      <organization>Western Research Laboratory</organization>
4310      <address><email></email></address>
4311    </author>
4312    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4313      <organization>Department of Information and Computer Science</organization>
4314      <address><email></email></address>
4315    </author>
4316    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4317      <organization>MIT Laboratory for Computer Science</organization>
4318      <address><email></email></address>
4319    </author>
4320    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4321      <organization>W3 Consortium</organization>
4322      <address><email></email></address>
4323    </author>
4324    <date month="May" year="1997"/>
4325  </front>
4326  <seriesInfo name="RFC" value="2145"/>
4329<reference anchor="RFC2616">
4330  <front>
4331    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4332    <author initials="R." surname="Fielding" fullname="R. Fielding">
4333      <organization>University of California, Irvine</organization>
4334      <address><email></email></address>
4335    </author>
4336    <author initials="J." surname="Gettys" fullname="J. Gettys">
4337      <organization>W3C</organization>
4338      <address><email></email></address>
4339    </author>
4340    <author initials="J." surname="Mogul" fullname="J. Mogul">
4341      <organization>Compaq Computer Corporation</organization>
4342      <address><email></email></address>
4343    </author>
4344    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
4345      <organization>MIT Laboratory for Computer Science</organization>
4346      <address><email></email></address>
4347    </author>
4348    <author initials="L." surname="Masinter" fullname="L. Masinter">
4349      <organization>Xerox Corporation</organization>
4350      <address><email></email></address>
4351    </author>
4352    <author initials="P." surname="Leach" fullname="P. Leach">
4353      <organization>Microsoft Corporation</organization>
4354      <address><email></email></address>
4355    </author>
4356    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
4357      <organization>W3C</organization>
4358      <address><email></email></address>
4359    </author>
4360    <date month="June" year="1999"/>
4361  </front>
4362  <seriesInfo name="RFC" value="2616"/>
4365<reference anchor='RFC2817'>
4366  <front>
4367    <title>Upgrading to TLS Within HTTP/1.1</title>
4368    <author initials='R.' surname='Khare' fullname='R. Khare'>
4369      <organization>4K Associates / UC Irvine</organization>
4370      <address><email></email></address>
4371    </author>
4372    <author initials='S.' surname='Lawrence' fullname='S. Lawrence'>
4373      <organization>Agranat Systems, Inc.</organization>
4374      <address><email></email></address>
4375    </author>
4376    <date year='2000' month='May' />
4377  </front>
4378  <seriesInfo name='RFC' value='2817' />
4381<reference anchor='RFC2818'>
4382  <front>
4383    <title>HTTP Over TLS</title>
4384    <author initials='E.' surname='Rescorla' fullname='Eric Rescorla'>
4385      <organization>RTFM, Inc.</organization>
4386      <address><email></email></address>
4387    </author>
4388    <date year='2000' month='May' />
4389  </front>
4390  <seriesInfo name='RFC' value='2818' />
4393<reference anchor='RFC3040'>
4394  <front>
4395    <title>Internet Web Replication and Caching Taxonomy</title>
4396    <author initials='I.' surname='Cooper' fullname='I. Cooper'>
4397      <organization>Equinix, Inc.</organization>
4398    </author>
4399    <author initials='I.' surname='Melve' fullname='I. Melve'>
4400      <organization>UNINETT</organization>
4401    </author>
4402    <author initials='G.' surname='Tomlinson' fullname='G. Tomlinson'>
4403      <organization>CacheFlow Inc.</organization>
4404    </author>
4405    <date year='2001' month='January' />
4406  </front>
4407  <seriesInfo name='RFC' value='3040' />
4410<reference anchor='BCP90'>
4411  <front>
4412    <title>Registration Procedures for Message Header Fields</title>
4413    <author initials='G.' surname='Klyne' fullname='G. Klyne'>
4414      <organization>Nine by Nine</organization>
4415      <address><email></email></address>
4416    </author>
4417    <author initials='M.' surname='Nottingham' fullname='M. Nottingham'>
4418      <organization>BEA Systems</organization>
4419      <address><email></email></address>
4420    </author>
4421    <author initials='J.' surname='Mogul' fullname='J. Mogul'>
4422      <organization>HP Labs</organization>
4423      <address><email></email></address>
4424    </author>
4425    <date year='2004' month='September' />
4426  </front>
4427  <seriesInfo name='BCP' value='90' />
4428  <seriesInfo name='RFC' value='3864' />
4431<reference anchor='RFC4033'>
4432  <front>
4433    <title>DNS Security Introduction and Requirements</title>
4434    <author initials='R.' surname='Arends' fullname='R. Arends'/>
4435    <author initials='R.' surname='Austein' fullname='R. Austein'/>
4436    <author initials='M.' surname='Larson' fullname='M. Larson'/>
4437    <author initials='D.' surname='Massey' fullname='D. Massey'/>
4438    <author initials='S.' surname='Rose' fullname='S. Rose'/>
4439    <date year='2005' month='March' />
4440  </front>
4441  <seriesInfo name='RFC' value='4033' />
4444<reference anchor="BCP13">
4445  <front>
4446    <title>Media Type Specifications and Registration Procedures</title>
4447    <author initials="N." surname="Freed" fullname="N. Freed">
4448      <organization>Sun Microsystems</organization>
4449      <address>
4450        <email></email>
4451      </address>
4452    </author>
4453    <author initials="J." surname="Klensin" fullname="J. Klensin">
4454      <address>
4455        <email></email>
4456      </address>
4457    </author>
4458    <date year="2005" month="December"/>
4459  </front>
4460  <seriesInfo name="BCP" value="13"/>
4461  <seriesInfo name="RFC" value="4288"/>
4464<reference anchor='BCP115'>
4465  <front>
4466    <title>Guidelines and Registration Procedures for New URI Schemes</title>
4467    <author initials='T.' surname='Hansen' fullname='T. Hansen'>
4468      <organization>AT&amp;T Laboratories</organization>
4469      <address>
4470        <email></email>
4471      </address>
4472    </author>
4473    <author initials='T.' surname='Hardie' fullname='T. Hardie'>
4474      <organization>Qualcomm, Inc.</organization>
4475      <address>
4476        <email></email>
4477      </address>
4478    </author>
4479    <author initials='L.' surname='Masinter' fullname='L. Masinter'>
4480      <organization>Adobe Systems</organization>
4481      <address>
4482        <email></email>
4483      </address>
4484    </author>
4485    <date year='2006' month='February' />
4486  </front>
4487  <seriesInfo name='BCP' value='115' />
4488  <seriesInfo name='RFC' value='4395' />
4491<reference anchor='RFC4559'>
4492  <front>
4493    <title>SPNEGO-based Kerberos and NTLM HTTP Authentication in Microsoft Windows</title>
4494    <author initials='K.' surname='Jaganathan' fullname='K. Jaganathan'/>
4495    <author initials='L.' surname='Zhu' fullname='L. Zhu'/>
4496    <author initials='J.' surname='Brezak' fullname='J. Brezak'/>
4497    <date year='2006' month='June' />
4498  </front>
4499  <seriesInfo name='RFC' value='4559' />
4502<reference anchor='RFC5226'>
4503  <front>
4504    <title>Guidelines for Writing an IANA Considerations Section in RFCs</title>
4505    <author initials='T.' surname='Narten' fullname='T. Narten'>
4506      <organization>IBM</organization>
4507      <address><email></email></address>
4508    </author>
4509    <author initials='H.' surname='Alvestrand' fullname='H. Alvestrand'>
4510      <organization>Google</organization>
4511      <address><email></email></address>
4512    </author>
4513    <date year='2008' month='May' />
4514  </front>
4515  <seriesInfo name='BCP' value='26' />
4516  <seriesInfo name='RFC' value='5226' />
4519<reference anchor='RFC5246'>
4520   <front>
4521      <title>The Transport Layer Security (TLS) Protocol Version 1.2</title>
4522      <author initials='T.' surname='Dierks' fullname='T. Dierks'>
4523         <organization />
4524      </author>
4525      <author initials='E.' surname='Rescorla' fullname='E. Rescorla'>
4526         <organization>RTFM, Inc.</organization>
4527      </author>
4528      <date year='2008' month='August' />
4529   </front>
4530   <seriesInfo name='RFC' value='5246' />
4533<reference anchor="RFC5322">
4534  <front>
4535    <title>Internet Message Format</title>
4536    <author initials="P." surname="Resnick" fullname="P. Resnick">
4537      <organization>Qualcomm Incorporated</organization>
4538    </author>
4539    <date year="2008" month="October"/>
4540  </front>
4541  <seriesInfo name="RFC" value="5322"/>
4544<reference anchor="RFC6265">
4545  <front>
4546    <title>HTTP State Management Mechanism</title>
4547    <author initials="A." surname="Barth" fullname="Adam Barth">
4548      <organization abbrev="U.C. Berkeley">
4549        University of California, Berkeley
4550      </organization>
4551      <address><email></email></address>
4552    </author>
4553    <date year="2011" month="April" />
4554  </front>
4555  <seriesInfo name="RFC" value="6265"/>
4558<!--<reference anchor='BCP97'>
4559  <front>
4560    <title>Handling Normative References to Standards-Track Documents</title>
4561    <author initials='J.' surname='Klensin' fullname='J. Klensin'>
4562      <address>
4563        <email></email>
4564      </address>
4565    </author>
4566    <author initials='S.' surname='Hartman' fullname='S. Hartman'>
4567      <organization>MIT</organization>
4568      <address>
4569        <email></email>
4570      </address>
4571    </author>
4572    <date year='2007' month='June' />
4573  </front>
4574  <seriesInfo name='BCP' value='97' />
4575  <seriesInfo name='RFC' value='4897' />
4578<reference anchor="Kri2001" target="">
4579  <front>
4580    <title>HTTP Cookies: Standards, Privacy, and Politics</title>
4581    <author initials="D." surname="Kristol" fullname="David M. Kristol"/>
4582    <date year="2001" month="November"/>
4583  </front>
4584  <seriesInfo name="ACM Transactions on Internet Technology" value="Vol. 1, #2"/>
4590<section title="HTTP Version History" anchor="compatibility">
4592   HTTP has been in use by the World-Wide Web global information initiative
4593   since 1990. The first version of HTTP, later referred to as HTTP/0.9,
4594   was a simple protocol for hypertext data transfer across the Internet
4595   with only a single request method (GET) and no metadata.
4596   HTTP/1.0, as defined by <xref target="RFC1945"/>, added a range of request
4597   methods and MIME-like messaging that could include metadata about the data
4598   transferred and modifiers on the request/response semantics. However,
4599   HTTP/1.0 did not sufficiently take into consideration the effects of
4600   hierarchical proxies, caching, the need for persistent connections, or
4601   name-based virtual hosts. The proliferation of incompletely-implemented
4602   applications calling themselves "HTTP/1.0" further necessitated a
4603   protocol version change in order for two communicating applications
4604   to determine each other's true capabilities.
4607   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
4608   requirements that enable reliable implementations, adding only
4609   those new features that will either be safely ignored by an HTTP/1.0
4610   recipient or only sent when communicating with a party advertising
4611   conformance with HTTP/1.1.
4614   It is beyond the scope of a protocol specification to mandate
4615   conformance with previous versions. HTTP/1.1 was deliberately
4616   designed, however, to make supporting previous versions easy.
4617   We would expect a general-purpose HTTP/1.1 server to understand
4618   any valid request in the format of HTTP/1.0 and respond appropriately
4619   with an HTTP/1.1 message that only uses features understood (or
4620   safely ignored) by HTTP/1.0 clients.  Likewise, we would expect
4621   an HTTP/1.1 client to understand any valid HTTP/1.0 response.
4624   Since HTTP/0.9 did not support header fields in a request,
4625   there is no mechanism for it to support name-based virtual
4626   hosts (selection of resource by inspection of the <x:ref>Host</x:ref> header
4627   field).  Any server that implements name-based virtual hosts
4628   ought to disable support for HTTP/0.9.  Most requests that
4629   appear to be HTTP/0.9 are, in fact, badly constructed HTTP/1.x
4630   requests wherein a buggy client failed to properly encode
4631   linear whitespace found in a URI reference and placed in
4632   the request-target.
4635<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
4637   This section summarizes major differences between versions HTTP/1.0
4638   and HTTP/1.1.
4641<section title="Multi-homed Web Servers" anchor="">
4643   The requirements that clients and servers support the <x:ref>Host</x:ref>
4644   header field (<xref target=""/>), report an error if it is
4645   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-target"/>)
4646   are among the most important changes defined by HTTP/1.1.
4649   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
4650   addresses and servers; there was no other established mechanism for
4651   distinguishing the intended server of a request than the IP address
4652   to which that request was directed. The <x:ref>Host</x:ref> header field was
4653   introduced during the development of HTTP/1.1 and, though it was
4654   quickly implemented by most HTTP/1.0 browsers, additional requirements
4655   were placed on all HTTP/1.1 requests in order to ensure complete
4656   adoption.  At the time of this writing, most HTTP-based services
4657   are dependent upon the Host header field for targeting requests.
4661<section title="Keep-Alive Connections" anchor="compatibility.with.http.1.0.persistent.connections">
4663   In HTTP/1.0, each connection is established by the client prior to the
4664   request and closed by the server after sending the response. However, some
4665   implementations implement the explicitly negotiated ("Keep-Alive") version
4666   of persistent connections described in <xref x:sec="19.7.1" x:fmt="of"
4667   target="RFC2068"/>.
4670   Some clients and servers might wish to be compatible with these previous
4671   approaches to persistent connections, by explicitly negotiating for them
4672   with a "Connection: keep-alive" request header field. However, some
4673   experimental implementations of HTTP/1.0 persistent connections are faulty;
4674   for example, if an HTTP/1.0 proxy server doesn't understand
4675   <x:ref>Connection</x:ref>, it will erroneously forward that header field
4676   to the next inbound server, which would result in a hung connection.
4679   One attempted solution was the introduction of a Proxy-Connection header
4680   field, targeted specifically at proxies. In practice, this was also
4681   unworkable, because proxies are often deployed in multiple layers, bringing
4682   about the same problem discussed above.
4685   As a result, clients are encouraged not to send the Proxy-Connection header
4686   field in any requests.
4689   Clients are also encouraged to consider the use of Connection: keep-alive
4690   in requests carefully; while they can enable persistent connections with
4691   HTTP/1.0 servers, clients using them need will need to monitor the
4692   connection for "hung" requests (which indicate that the client ought stop
4693   sending the header field), and this mechanism ought not be used by clients
4694   at all when a proxy is being used.
4698<section title="Introduction of Transfer-Encoding" anchor="introduction.of.transfer-encoding">
4700   HTTP/1.1 introduces the <x:ref>Transfer-Encoding</x:ref> header field
4701   (<xref target="header.transfer-encoding"/>).
4702   Transfer codings need to be decoded prior to forwarding an HTTP message
4703   over a MIME-compliant protocol.
4709<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
4711  HTTP's approach to error handling has been explained.
4712  (<xref target="conformance"/>)
4715  The expectation to support HTTP/0.9 requests has been removed.
4718  The term "Effective Request URI" has been introduced.
4719  (<xref target="effective.request.uri" />)
4722  HTTP messages can be (and often are) buffered by implementations; despite
4723  it sometimes being available as a stream, HTTP is fundamentally a
4724  message-oriented protocol.
4725  (<xref target="http.message" />)
4728  Minimum supported sizes for various protocol elements have been
4729  suggested, to improve interoperability.
4732  Header fields that span multiple lines ("line folding") are deprecated.
4733  (<xref target="field.parsing" />)
4736  The HTTP-version ABNF production has been clarified to be case-sensitive.
4737  Additionally, version numbers has been restricted to single digits, due
4738  to the fact that implementations are known to handle multi-digit version
4739  numbers incorrectly.
4740  (<xref target="http.version"/>)
4743  The HTTPS URI scheme is now defined by this specification; previously,
4744  it was done in  <xref target="RFC2818" x:fmt="of" x:sec="2.4"/>.
4745  (<xref target="https.uri"/>)
4748  The HTTPS URI scheme implies end-to-end security.
4749  (<xref target="https.uri"/>)
4752  Userinfo (i.e., username and password) are now disallowed in HTTP and
4753  HTTPS URIs, because of security issues related to their transmission on the
4754  wire.
4755  (<xref target="http.uri" />)
4758  Invalid whitespace around field-names is now required to be rejected,
4759  because accepting it represents a security vulnerability.
4760  (<xref target="header.fields"/>)
4763  The ABNF productions defining header fields now only list the field value.
4764  (<xref target="header.fields"/>)
4767  Rules about implicit linear whitespace between certain grammar productions
4768  have been removed; now whitespace is only allowed where specifically
4769  defined in the ABNF.
4770  (<xref target="whitespace"/>)
4773  The NUL octet is no longer allowed in comment and quoted-string text, and
4774  handling of backslash-escaping in them has been clarified.
4775  (<xref target="field.components"/>)
4778  The quoted-pair rule no longer allows escaping control characters other than
4779  HTAB.
4780  (<xref target="field.components"/>)
4783  Non-ASCII content in header fields and the reason phrase has been obsoleted
4784  and made opaque (the TEXT rule was removed).
4785  (<xref target="field.components"/>)
4788  Bogus "<x:ref>Content-Length</x:ref>" header fields are now required to be
4789  handled as errors by recipients.
4790  (<xref target="header.content-length"/>)
4793  The "identity" transfer coding token has been removed.
4794  (Sections <xref format="counter" target="message.body"/> and
4795  <xref format="counter" target="transfer.codings"/>)
4798  The algorithm for determining the message body length has been clarified
4799  to indicate all of the special cases (e.g., driven by methods or status
4800  codes) that affect it, and that new protocol elements cannot define such
4801  special cases.
4802  (<xref target="message.body.length"/>)
4805  "multipart/byteranges" is no longer a way of determining message body length
4806  detection.
4807  (<xref target="message.body.length"/>)
4810  CONNECT is a new, special case in determining message body length.
4811  (<xref target="message.body.length"/>)
4814  Chunk length does not include the count of the octets in the
4815  chunk header and trailer.
4816  (<xref target="chunked.encoding"/>)
4819  Use of chunk extensions is deprecated, and line folding in them is
4820  disallowed.
4821  (<xref target="chunked.encoding"/>)
4824  The path-absolute + query components of RFC3986 have been used to define the
4825  request-target, instead of abs_path from RFC 1808.
4826  (<xref target="request-target"/>)
4829  The asterisk form of the request-target is only allowed in the OPTIONS
4830  method.
4831  (<xref target="request-target"/>)
4834  Exactly when "close" connection options have to be sent has been clarified.
4835  (<xref target="header.connection"/>)
4838  "hop-by-hop" header fields are required to appear in the Connection header
4839  field; just because they're defined as hop-by-hop in this specification
4840  doesn't exempt them.
4841  (<xref target="header.connection"/>)
4844  The limit of two connections per server has been removed.
4845  (<xref target="persistent.connections"/>)
4848  An idempotent sequence of requests is no longer required to be retried.
4849  (<xref target="persistent.connections"/>)
4852  The requirement to retry requests under certain circumstances when the
4853  server prematurely closes the connection has been removed.
4854  (<xref target="persistent.connections"/>)
4857  Some extraneous requirements about when servers are allowed to close
4858  connections prematurely have been removed.
4859  (<xref target="persistent.connections"/>)
4862  The semantics of the <x:ref>Upgrade</x:ref> header field is now defined in
4863  responses other than 101 (this was incorporated from <xref
4864  target="RFC2817"/>).
4865  (<xref target="header.upgrade"/>)
4868  Registration of Transfer Codings now requires IETF Review
4869  (<xref target="transfer.coding.registry"/>)
4872  The meaning of the "deflate" content coding has been clarified.
4873  (<xref target="deflate.coding" />)
4876  This specification now defines the Upgrade Token Registry, previously
4877  defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
4878  (<xref target="upgrade.token.registry"/>)
4881  Empty list elements in list productions (e.g., a list header containing
4882  ", ,") have been deprecated.
4883  (<xref target="abnf.extension"/>)
4886  Issues with the Keep-Alive and Proxy-Connection headers in requests
4887  are pointed out, with use of the latter being discouraged altogether.
4888  (<xref target="compatibility.with.http.1.0.persistent.connections" />)
4893<section title="ABNF list extension: #rule" anchor="abnf.extension">
4895  A #rule extension to the ABNF rules of <xref target="RFC5234"/> is used to
4896  improve readability in the definitions of some header field values.
4899  A construct "#" is defined, similar to "*", for defining comma-delimited
4900  lists of elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating
4901  at least &lt;n&gt; and at most &lt;m&gt; elements, each separated by a single
4902  comma (",") and optional whitespace (OWS).   
4905  Thus,
4906</preamble><artwork type="example">
4907  1#element =&gt; element *( OWS "," OWS element )
4910  and:
4911</preamble><artwork type="example">
4912  #element =&gt; [ 1#element ]
4915  and for n &gt;= 1 and m &gt; 1:
4916</preamble><artwork type="example">
4917  &lt;n&gt;#&lt;m&gt;element =&gt; element &lt;n-1&gt;*&lt;m-1&gt;( OWS "," OWS element )
4920  For compatibility with legacy list rules, recipients &SHOULD; accept empty
4921  list elements. In other words, consumers would follow the list productions:
4923<figure><artwork type="example">
4924  #element =&gt; [ ( "," / element ) *( OWS "," [ OWS element ] ) ]
4926  1#element =&gt; *( "," OWS ) element *( OWS "," [ OWS element ] )
4929  Note that empty elements do not contribute to the count of elements present,
4930  though.
4933  For example, given these ABNF productions:
4935<figure><artwork type="example">
4936  example-list      = 1#example-list-elmt
4937  example-list-elmt = token ; see <xref target="field.components"/>
4940  Then these are valid values for example-list (not including the double
4941  quotes, which are present for delimitation only):
4943<figure><artwork type="example">
4944  "foo,bar"
4945  "foo ,bar,"
4946  "foo , ,bar,charlie   "
4949  But these values would be invalid, as at least one non-empty element is
4950  required:
4952<figure><artwork type="example">
4953  ""
4954  ","
4955  ",   ,"
4958  <xref target="collected.abnf"/> shows the collected ABNF, with the list rules
4959  expanded as explained above.
4963<?BEGININC p1-messaging.abnf-appendix ?>
4964<section xmlns:x="" title="Collected ABNF" anchor="collected.abnf">
4966<artwork type="abnf" name="p1-messaging.parsed-abnf">
4967<x:ref>BWS</x:ref> = OWS
4969<x:ref>Connection</x:ref> = *( "," OWS ) connection-option *( OWS "," [ OWS
4970 connection-option ] )
4971<x:ref>Content-Length</x:ref> = 1*DIGIT
4973<x:ref>HTTP-message</x:ref> = start-line *( header-field CRLF ) CRLF [ message-body
4974 ]
4975<x:ref>HTTP-name</x:ref> = %x48.54.54.50 ; HTTP
4976<x:ref>HTTP-version</x:ref> = HTTP-name "/" DIGIT "." DIGIT
4977<x:ref>Host</x:ref> = uri-host [ ":" port ]
4979<x:ref>OWS</x:ref> = *( SP / HTAB )
4981<x:ref>RWS</x:ref> = 1*( SP / HTAB )
4983<x:ref>TE</x:ref> = [ ( "," / t-codings ) *( OWS "," [ OWS t-codings ] ) ]
4984<x:ref>Trailer</x:ref> = *( "," OWS ) field-name *( OWS "," [ OWS field-name ] )
4985<x:ref>Transfer-Encoding</x:ref> = *( "," OWS ) transfer-coding *( OWS "," [ OWS
4986 transfer-coding ] )
4988<x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in [RFC3986], Section 4.1&gt;
4989<x:ref>Upgrade</x:ref> = *( "," OWS ) protocol *( OWS "," [ OWS protocol ] )
4991<x:ref>Via</x:ref> = *( "," OWS ) ( received-protocol RWS received-by [ RWS comment
4992 ] ) *( OWS "," [ OWS ( received-protocol RWS received-by [ RWS
4993 comment ] ) ] )
4995<x:ref>absolute-URI</x:ref> = &lt;absolute-URI, defined in [RFC3986], Section 4.3&gt;
4996<x:ref>absolute-form</x:ref> = absolute-URI
4997<x:ref>asterisk-form</x:ref> = "*"
4998<x:ref>attribute</x:ref> = token
4999<x:ref>authority</x:ref> = &lt;authority, defined in [RFC3986], Section 3.2&gt;
5000<x:ref>authority-form</x:ref> = authority
5002<x:ref>chunk</x:ref> = chunk-size [ chunk-ext ] CRLF chunk-data CRLF
5003<x:ref>chunk-data</x:ref> = 1*OCTET
5004<x:ref>chunk-ext</x:ref> = *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
5005<x:ref>chunk-ext-name</x:ref> = token
5006<x:ref>chunk-ext-val</x:ref> = token / quoted-str-nf
5007<x:ref>chunk-size</x:ref> = 1*HEXDIG
5008<x:ref>chunked-body</x:ref> = *chunk last-chunk trailer-part CRLF
5009<x:ref>comment</x:ref> = "(" *( ctext / quoted-cpair / comment ) ")"
5010<x:ref>connection-option</x:ref> = token
5011<x:ref>ctext</x:ref> = HTAB / SP / %x21-27 ; '!'-'''
5012 / %x2A-5B ; '*'-'['
5013 / %x5D-7E ; ']'-'~'
5014 / obs-text
5016<x:ref>field-content</x:ref> = *( HTAB / SP / VCHAR / obs-text )
5017<x:ref>field-name</x:ref> = token
5018<x:ref>field-value</x:ref> = *( field-content / obs-fold )
5020<x:ref>header-field</x:ref> = field-name ":" OWS field-value BWS
5021<x:ref>http-URI</x:ref> = "http://" authority path-abempty [ "?" query ]
5022<x:ref>https-URI</x:ref> = "https://" authority path-abempty [ "?" query ]
5024<x:ref>last-chunk</x:ref> = 1*"0" [ chunk-ext ] CRLF
5026<x:ref>message-body</x:ref> = *OCTET
5027<x:ref>method</x:ref> = token
5029<x:ref>obs-fold</x:ref> = CRLF ( SP / HTAB )
5030<x:ref>obs-text</x:ref> = %x80-FF
5031<x:ref>origin-form</x:ref> = path-absolute [ "?" query ]
5033<x:ref>partial-URI</x:ref> = relative-part [ "?" query ]
5034<x:ref>path-abempty</x:ref> = &lt;path-abempty, defined in [RFC3986], Section 3.3&gt;
5035<x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in [RFC3986], Section 3.3&gt;
5036<x:ref>port</x:ref> = &lt;port, defined in [RFC3986], Section 3.2.3&gt;
5037<x:ref>protocol</x:ref> = protocol-name [ "/" protocol-version ]
5038<x:ref>protocol-name</x:ref> = token
5039<x:ref>protocol-version</x:ref> = token
5040<x:ref>pseudonym</x:ref> = token
5042<x:ref>qdtext</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5043 / %x5D-7E ; ']'-'~'
5044 / obs-text
5045<x:ref>qdtext-nf</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5046 / %x5D-7E ; ']'-'~'
5047 / obs-text
5048<x:ref>query</x:ref> = &lt;query, defined in [RFC3986], Section 3.4&gt;
5049<x:ref>quoted-cpair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5050<x:ref>quoted-pair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5051<x:ref>quoted-str-nf</x:ref> = DQUOTE *( qdtext-nf / quoted-pair ) DQUOTE
5052<x:ref>quoted-string</x:ref> = DQUOTE *( qdtext / quoted-pair ) DQUOTE
5054<x:ref>rank</x:ref> = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] )
5055<x:ref>reason-phrase</x:ref> = *( HTAB / SP / VCHAR / obs-text )
5056<x:ref>received-by</x:ref> = ( uri-host [ ":" port ] ) / pseudonym
5057<x:ref>received-protocol</x:ref> = [ protocol-name "/" ] protocol-version
5058<x:ref>relative-part</x:ref> = &lt;relative-part, defined in [RFC3986], Section 4.2&gt;
5059<x:ref>request-line</x:ref> = method SP request-target SP HTTP-version CRLF
5060<x:ref>request-target</x:ref> = origin-form / absolute-form / authority-form /
5061 asterisk-form
5063<x:ref>special</x:ref> = "(" / ")" / "&lt;" / "&gt;" / "@" / "," / ";" / ":" / "\" /
5064 DQUOTE / "/" / "[" / "]" / "?" / "=" / "{" / "}"
5065<x:ref>start-line</x:ref> = request-line / status-line
5066<x:ref>status-code</x:ref> = 3DIGIT
5067<x:ref>status-line</x:ref> = HTTP-version SP status-code SP reason-phrase CRLF
5069<x:ref>t-codings</x:ref> = "trailers" / ( transfer-coding [ t-ranking ] )
5070<x:ref>t-ranking</x:ref> = OWS ";" OWS "q=" rank
5071<x:ref>tchar</x:ref> = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*" / "+" / "-" / "." /
5072 "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
5073<x:ref>token</x:ref> = 1*tchar
5074<x:ref>trailer-part</x:ref> = *( header-field CRLF )
5075<x:ref>transfer-coding</x:ref> = "chunked" / "compress" / "deflate" / "gzip" /
5076 transfer-extension
5077<x:ref>transfer-extension</x:ref> = token *( OWS ";" OWS transfer-parameter )
5078<x:ref>transfer-parameter</x:ref> = attribute BWS "=" BWS value
5080<x:ref>uri-host</x:ref> = &lt;host, defined in [RFC3986], Section 3.2.2&gt;
5082<x:ref>value</x:ref> = word
5084<x:ref>word</x:ref> = token / quoted-string
5088<?ENDINC p1-messaging.abnf-appendix ?>
5090<section title="Change Log (to be removed by RFC Editor before publication)" anchor="change.log">
5092<section title="Since RFC 2616">
5094  Changes up to the first Working Group Last Call draft are summarized
5095  in <eref target=""/>.
5099<section title="Since draft-ietf-httpbis-p1-messaging-21" anchor="changes.since.21">
5101  Closed issues:
5102  <list style="symbols">
5103    <t>
5104      <eref target=""/>:
5105      "Cite HTTPS URI scheme definition" (the spec now includes the HTTPs
5106      scheme definition and thus updates RFC 2818)
5107    </t>
5108    <t>
5109      <eref target=""/>:
5110      "mention of 'proxies' in section about caches"
5111    </t>
5112    <t>
5113      <eref target=""/>:
5114      "use of ABNF terms from RFC 3986"
5115    </t>
5116    <t>
5117      <eref target=""/>:
5118      "editorial improvements to message length definition"
5119    </t>
5120    <t>
5121      <eref target=""/>:
5122      "Connection header field MUST vs SHOULD"
5123    </t>
5124    <t>
5125      <eref target=""/>:
5126      "editorial improvements to persistent connections section"
5127    </t>
5128    <t>
5129      <eref target=""/>:
5130      "URI normalization vs empty path"
5131    </t>
5132    <t>
5133      <eref target=""/>:
5134      "p1 feedback"
5135    </t>
5136    <t>
5137      <eref target=""/>:
5138      "is parsing OBS-FOLD mandatory?"
5139    </t>
5140    <t>
5141      <eref target=""/>:
5142      "HTTPS and Shared Caching"
5143    </t>
5144    <t>
5145      <eref target=""/>:
5146      "Requirements for recipients of ws between start-line and first header field"
5147    </t>
5148    <t>
5149      <eref target=""/>:
5150      "SP and HT when being tolerant"
5151    </t>
5152    <t>
5153      <eref target=""/>:
5154      "Message Parsing Strictness"
5155    </t>
5156    <t>
5157      <eref target=""/>:
5158      "'Render'"
5159    </t>
5160    <t>
5161      <eref target=""/>:
5162      "No-Transform"
5163    </t>
5164    <t>
5165      <eref target=""/>:
5166      "p2 editorial feedback"
5167    </t>
5168    <t>
5169      <eref target=""/>:
5170      "Content-Length SHOULD be sent"
5171    </t>
5172  </list>
Note: See TracBrowser for help on using the repository browser.