source: draft-ietf-httpbis/latest/p1-messaging.xml @ 2110

Last change on this file since 2110 was 2110, checked in by fielding@…, 10 years ago

(editorial) move definition of selected representtation to top and change name of selected representation header fields to validator header fields for consistency with p4

  • Property svn:eol-style set to native
  • Property svn:mime-type set to text/xml
File size: 222.1 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 methods                "<xref target='Part2' x:rel='#methods' xmlns:x=''/>">
47  <!ENTITY OPTIONS                "<xref target='Part2' x:rel='#OPTIONS' xmlns:x=''/>">
48  <!ENTITY qvalue                 "<xref target='Part2' x:rel='#quality.values' xmlns:x=''/>">
49  <!ENTITY resource               "<xref target='Part2' x:rel='#resources' xmlns:x=''/>">
50  <!ENTITY status-codes           "<xref target='Part2' x:rel='' xmlns:x=''/>">
51  <!ENTITY status-1xx             "<xref target='Part2' x:rel='#status.1xx' xmlns:x=''/>">
52  <!ENTITY status-203             "<xref target='Part2' x:rel='#status.203' xmlns:x=''/>">
53  <!ENTITY status-3xx             "<xref target='Part2' x:rel='#status.3xx' xmlns:x=''/>">
54  <!ENTITY status-304             "<xref target='Part4' x:rel='#status.304' xmlns:x=''/>">
55  <!ENTITY status-4xx             "<xref target='Part2' x:rel='#status.4xx' xmlns:x=''/>">
56  <!ENTITY status-414             "<xref target='Part2' x:rel='#status.414' xmlns:x=''/>">
57  <!ENTITY iana-header-registry   "<xref target='Part2' x:rel='#header.field.registry' xmlns:x=''/>">
59<?rfc toc="yes" ?>
60<?rfc symrefs="yes" ?>
61<?rfc sortrefs="yes" ?>
62<?rfc compact="yes"?>
63<?rfc subcompact="no" ?>
64<?rfc linkmailto="no" ?>
65<?rfc editing="no" ?>
66<?rfc comments="yes"?>
67<?rfc inline="yes"?>
68<?rfc rfcedstyle="yes"?>
69<?rfc-ext allow-markup-in-artwork="yes" ?>
70<?rfc-ext include-references-in-index="yes" ?>
71<rfc obsoletes="2145,2616" updates="2817,2818" category="std" x:maturity-level="proposed"
72     ipr="pre5378Trust200902" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
73     xmlns:x=''>
74<x:link rel="next" basename="p2-semantics"/>
75<x:feedback template="{docname},%20%22{section}%22&amp;body=&lt;{ref}&gt;:"/>
78  <title abbrev="HTTP/1.1 Message Syntax and Routing">Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing</title>
80  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
81    <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
82    <address>
83      <postal>
84        <street>345 Park Ave</street>
85        <city>San Jose</city>
86        <region>CA</region>
87        <code>95110</code>
88        <country>USA</country>
89      </postal>
90      <email></email>
91      <uri></uri>
92    </address>
93  </author>
95  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
96    <organization abbrev="greenbytes">greenbytes GmbH</organization>
97    <address>
98      <postal>
99        <street>Hafenweg 16</street>
100        <city>Muenster</city><region>NW</region><code>48155</code>
101        <country>Germany</country>
102      </postal>
103      <email></email>
104      <uri></uri>
105    </address>
106  </author>
108  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
109  <workgroup>HTTPbis Working Group</workgroup>
113   The Hypertext Transfer Protocol (HTTP) is an application-level protocol for
114   distributed, collaborative, hypertext information systems. HTTP has been in
115   use by the World Wide Web global information initiative since 1990.
116   This document provides an overview of HTTP architecture and its associated
117   terminology, defines the "http" and "https" Uniform Resource Identifier
118   (URI) schemes, defines the HTTP/1.1 message syntax and parsing requirements,
119   and describes general security concerns for implementations.
123<note title="Editorial Note (To be removed by RFC Editor)">
124  <t>
125    Discussion of this draft takes place on the HTTPBIS working group
126    mailing list (, which is archived at
127    <eref target=""/>.
128  </t>
129  <t>
130    The current issues list is at
131    <eref target=""/> and related
132    documents (including fancy diffs) can be found at
133    <eref target=""/>.
134  </t>
135  <t>
136    The changes in this draft are summarized in <xref target="changes.since.21"/>.
137  </t>
141<section title="Introduction" anchor="introduction">
143   The Hypertext Transfer Protocol (HTTP) is an application-level
144   request/response protocol that uses extensible semantics and self-descriptive
145   message payloads for flexible interaction with network-based hypertext
146   information systems. This document is the first in a series of documents
147   that collectively form the HTTP/1.1 specification:
148   <list style="empty">
149    <t>RFC xxx1: Message Syntax and Routing</t>
150    <t><xref target="Part2" x:fmt="none">RFC xxx2</xref>: Semantics and Content</t>
151    <t><xref target="Part4" x:fmt="none">RFC xxx3</xref>: Conditional Requests</t>
152    <t><xref target="Part5" x:fmt="none">RFC xxx4</xref>: Range Requests</t>
153    <t><xref target="Part6" x:fmt="none">RFC xxx5</xref>: Caching</t>
154    <t><xref target="Part7" x:fmt="none">RFC xxx6</xref>: Authentication</t>
155   </list>
158   This HTTP/1.1 specification obsoletes and moves to historic status
159   <xref target="RFC2616" x:fmt="none">RFC 2616</xref>, its predecessor
160   <xref target="RFC2068" x:fmt="none">RFC 2068</xref>, and
161   <xref target="RFC2145" x:fmt="none">RFC 2145</xref> (on HTTP versioning).
162   This specification also updates the use of CONNECT to establish a tunnel,
163   previously defined in <xref target="RFC2817" x:fmt="none">RFC 2817</xref>,
164   and defines the "https" URI scheme that was described informally in
165   <xref target="RFC2818" x:fmt="none">RFC 2818</xref>.
168   HTTP is a generic interface protocol for information systems. It is
169   designed to hide the details of how a service is implemented by presenting
170   a uniform interface to clients that is independent of the types of
171   resources provided. Likewise, servers do not need to be aware of each
172   client's purpose: an HTTP request can be considered in isolation rather
173   than being associated with a specific type of client or a predetermined
174   sequence of application steps. The result is a protocol that can be used
175   effectively in many different contexts and for which implementations can
176   evolve independently over time.
179   HTTP is also designed for use as an intermediation protocol for translating
180   communication to and from non-HTTP information systems.
181   HTTP proxies and gateways can provide access to alternative information
182   services by translating their diverse protocols into a hypertext
183   format that can be viewed and manipulated by clients in the same way
184   as HTTP services.
187   One consequence of this flexibility is that the protocol cannot be
188   defined in terms of what occurs behind the interface. Instead, we
189   are limited to defining the syntax of communication, the intent
190   of received communication, and the expected behavior of recipients.
191   If the communication is considered in isolation, then successful
192   actions ought to be reflected in corresponding changes to the
193   observable interface provided by servers. However, since multiple
194   clients might act in parallel and perhaps at cross-purposes, we
195   cannot require that such changes be observable beyond the scope
196   of a single response.
199   This document describes the architectural elements that are used or
200   referred to in HTTP, defines the "http" and "https" URI schemes,
201   describes overall network operation and connection management,
202   and defines HTTP message framing and forwarding requirements.
203   Our goal is to define all of the mechanisms necessary for HTTP message
204   handling that are independent of message semantics, thereby defining the
205   complete set of requirements for message parsers and
206   message-forwarding intermediaries.
210<section title="Requirement Notation" anchor="intro.requirements">
212   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
213   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
214   document are to be interpreted as described in <xref target="RFC2119"/>.
217   Conformance criteria and considerations regarding error handling
218   are defined in <xref target="conformance"/>.
222<section title="Syntax Notation" anchor="notation">
223<iref primary="true" item="Grammar" subitem="ALPHA"/>
224<iref primary="true" item="Grammar" subitem="CR"/>
225<iref primary="true" item="Grammar" subitem="CRLF"/>
226<iref primary="true" item="Grammar" subitem="CTL"/>
227<iref primary="true" item="Grammar" subitem="DIGIT"/>
228<iref primary="true" item="Grammar" subitem="DQUOTE"/>
229<iref primary="true" item="Grammar" subitem="HEXDIG"/>
230<iref primary="true" item="Grammar" subitem="HTAB"/>
231<iref primary="true" item="Grammar" subitem="LF"/>
232<iref primary="true" item="Grammar" subitem="OCTET"/>
233<iref primary="true" item="Grammar" subitem="SP"/>
234<iref primary="true" item="Grammar" subitem="VCHAR"/>
236   This specification uses the Augmented Backus-Naur Form (ABNF) notation
237   of <xref target="RFC5234"/> with the list rule extension defined in
238   <xref target="abnf.extension"/>.  <xref target="collected.abnf"/> shows
239   the collected ABNF with the list rule expanded.
241<t anchor="core.rules">
242  <x:anchor-alias value="ALPHA"/>
243  <x:anchor-alias value="CTL"/>
244  <x:anchor-alias value="CR"/>
245  <x:anchor-alias value="CRLF"/>
246  <x:anchor-alias value="DIGIT"/>
247  <x:anchor-alias value="DQUOTE"/>
248  <x:anchor-alias value="HEXDIG"/>
249  <x:anchor-alias value="HTAB"/>
250  <x:anchor-alias value="LF"/>
251  <x:anchor-alias value="OCTET"/>
252  <x:anchor-alias value="SP"/>
253  <x:anchor-alias value="VCHAR"/>
254   The following core rules are included by
255   reference, as defined in <xref target="RFC5234" x:fmt="," x:sec="B.1"/>:
256   ALPHA (letters), CR (carriage return), CRLF (CR LF), CTL (controls),
257   DIGIT (decimal 0-9), DQUOTE (double quote),
258   HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line feed),
259   OCTET (any 8-bit sequence of data), SP (space), and
260   VCHAR (any visible <xref target="USASCII"/> character).
263   As a convention, ABNF rule names prefixed with "obs-" denote
264   "obsolete" grammar rules that appear for historical reasons.
269<section title="Architecture" anchor="architecture">
271   HTTP was created for the World Wide Web architecture
272   and has evolved over time to support the scalability needs of a worldwide
273   hypertext system. Much of that architecture is reflected in the terminology
274   and syntax productions used to define HTTP.
277<section title="Client/Server Messaging" anchor="operation">
278<iref primary="true" item="client"/>
279<iref primary="true" item="server"/>
280<iref primary="true" item="connection"/>
282   HTTP is a stateless request/response protocol that operates by exchanging
283   <x:dfn>messages</x:dfn> (<xref target="http.message"/>) across a reliable
284   transport or session-layer
285   "<x:dfn>connection</x:dfn>" (<xref target=""/>).
286   An HTTP "<x:dfn>client</x:dfn>" is a program that establishes a connection
287   to a server for the purpose of sending one or more HTTP requests.
288   An HTTP "<x:dfn>server</x:dfn>" is a program that accepts connections
289   in order to service HTTP requests by sending HTTP responses.
291<iref primary="true" item="user agent"/>
292<iref primary="true" item="origin server"/>
293<iref primary="true" item="browser"/>
294<iref primary="true" item="spider"/>
295<iref primary="true" item="sender"/>
296<iref primary="true" item="recipient"/>
298   The terms client and server refer only to the roles that
299   these programs perform for a particular connection.  The same program
300   might act as a client on some connections and a server on others.
301   We use the term "<x:dfn>user agent</x:dfn>" to refer to any of the various
302   client programs that initiate a request, including (but not limited to)
303   browsers, spiders (web-based robots), command-line tools, native
304   applications, and mobile apps.  The term "<x:dfn>origin server</x:dfn>" is
305   used to refer to the program that can originate authoritative responses to
306   a request. For general requirements, we use the terms
307   "<x:dfn>sender</x:dfn>" and "<x:dfn>recipient</x:dfn>" to refer to any
308   component that sends or receives, respectively, a given message.
311   HTTP relies upon the Uniform Resource Identifier (URI)
312   standard <xref target="RFC3986"/> to indicate the target resource
313   (<xref target="target-resource"/>) and relationships between resources.
314   Messages are passed in a format similar to that used by Internet mail
315   <xref target="RFC5322"/> and the Multipurpose Internet Mail Extensions
316   (MIME) <xref target="RFC2045"/> (see &diff-mime; for the differences
317   between HTTP and MIME messages).
320   Most HTTP communication consists of a retrieval request (GET) for
321   a representation of some resource identified by a URI.  In the
322   simplest case, this might be accomplished via a single bidirectional
323   connection (===) between the user agent (UA) and the origin server (O).
325<figure><artwork type="drawing">
326         request   &gt;
327    <x:highlight>UA</x:highlight> ======================================= <x:highlight>O</x:highlight>
328                                &lt;   response
330<iref primary="true" item="message"/>
331<iref primary="true" item="request"/>
332<iref primary="true" item="response"/>
334   A client sends an HTTP request to a server in the form of a <x:dfn>request</x:dfn>
335   message, beginning with a request-line that includes a method, URI, and
336   protocol version (<xref target="request.line"/>),
337   followed by header fields containing
338   request modifiers, client information, and representation metadata
339   (<xref target="header.fields"/>),
340   an empty line to indicate the end of the header section, and finally
341   a message body containing the payload body (if any,
342   <xref target="message.body"/>).
345   A server responds to a client's request by sending one or more HTTP
346   <x:dfn>response</x:dfn>
347   messages, each beginning with a status line that
348   includes the protocol version, a success or error code, and textual
349   reason phrase (<xref target="status.line"/>),
350   possibly followed by header fields containing server
351   information, resource metadata, and representation metadata
352   (<xref target="header.fields"/>),
353   an empty line to indicate the end of the header section, and finally
354   a message body containing the payload body (if any,
355   <xref target="message.body"/>).
358   A connection might be used for multiple request/response exchanges,
359   as defined in <xref target="persistent.connections"/>.
362   The following example illustrates a typical message exchange for a
363   GET request on the URI "":
366client request:
367</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
368GET /hello.txt HTTP/1.1
369User-Agent: curl/7.16.3 libcurl/7.16.3 OpenSSL/0.9.7l zlib/1.2.3
371Accept-Language: en, mi
375server response:
376</preamble><artwork type="message/http; msgtype=&#34;response&#34;" x:indent-with="  ">
377HTTP/1.1 200 OK
378Date: Mon, 27 Jul 2009 12:28:53 GMT
379Server: Apache
380Last-Modified: Wed, 22 Jul 2009 19:15:56 GMT
381ETag: "34aa387-d-1568eb00"
382Accept-Ranges: bytes
383Content-Length: <x:length-of target="exbody"/>
384Vary: Accept-Encoding
385Content-Type: text/plain
387<x:span anchor="exbody">Hello World!
391<section title="Implementation Diversity" anchor="implementation-diversity">
393   When considering the design of HTTP, it is easy to fall into a trap of
394   thinking that all user agents are general-purpose browsers and all origin
395   servers are large public websites. That is not the case in practice.
396   Common HTTP user agents include household appliances, stereos, scales,
397   firmware update scripts, command-line programs, mobile apps,
398   and communication devices in a multitude of shapes and sizes.  Likewise,
399   common HTTP origin servers include home automation units, configurable
400   networking components, office machines, autonomous robots, news feeds,
401   traffic cameras, ad selectors, and video delivery platforms.
404   The term "user agent" does not imply that there is a human user directly
405   interacting with the software agent at the time of a request. In many
406   cases, a user agent is installed or configured to run in the background
407   and save its results for later inspection (or save only a subset of those
408   results that might be interesting or erroneous). Spiders, for example, are
409   typically given a start URI and configured to follow certain behavior while
410   crawling the Web as a hypertext graph.
413   The implementation diversity of HTTP means that we cannot assume the
414   user agent can make interactive suggestions to a user or provide adequate
415   warning for security or privacy options.  In the few cases where this
416   specification requires reporting of errors to the user, it is acceptable
417   for such reporting to only be observable in an error console or log file.
418   Likewise, requirements that an automated action be confirmed by the user
419   before proceeding can be met via advance configuration choices,
420   run-time options, or simply not proceeding with the unsafe action.
424<section title="Intermediaries" anchor="intermediaries">
425<iref primary="true" item="intermediary"/>
427   HTTP enables the use of intermediaries to satisfy requests through
428   a chain of connections.  There are three common forms of HTTP
429   <x:dfn>intermediary</x:dfn>: proxy, gateway, and tunnel.  In some cases,
430   a single intermediary might act as an origin server, proxy, gateway,
431   or tunnel, switching behavior based on the nature of each request.
433<figure><artwork type="drawing">
434         &gt;             &gt;             &gt;             &gt;
435    <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>
436               &lt;             &lt;             &lt;             &lt;
439   The figure above shows three intermediaries (A, B, and C) between the
440   user agent and origin server. A request or response message that
441   travels the whole chain will pass through four separate connections.
442   Some HTTP communication options
443   might apply only to the connection with the nearest, non-tunnel
444   neighbor, only to the end-points of the chain, or to all connections
445   along the chain. Although the diagram is linear, each participant might
446   be engaged in multiple, simultaneous communications. For example, B
447   might be receiving requests from many clients other than A, and/or
448   forwarding requests to servers other than C, at the same time that it
449   is handling A's request.
452<iref primary="true" item="upstream"/><iref primary="true" item="downstream"/>
453<iref primary="true" item="inbound"/><iref primary="true" item="outbound"/>
454   We use the terms "<x:dfn>upstream</x:dfn>" and "<x:dfn>downstream</x:dfn>"
455   to describe various requirements in relation to the directional flow of a
456   message: all messages flow from upstream to downstream.
457   Likewise, we use the terms inbound and outbound to refer to
458   directions in relation to the request path:
459   "<x:dfn>inbound</x:dfn>" means toward the origin server and
460   "<x:dfn>outbound</x:dfn>" means toward the user agent.
462<t><iref primary="true" item="proxy"/>
463   A "<x:dfn>proxy</x:dfn>" is a message forwarding agent that is selected by the
464   client, usually via local configuration rules, to receive requests
465   for some type(s) of absolute URI and attempt to satisfy those
466   requests via translation through the HTTP interface.  Some translations
467   are minimal, such as for proxy requests for "http" URIs, whereas
468   other requests might require translation to and from entirely different
469   application-level protocols. Proxies are often used to group an
470   organization's HTTP requests through a common intermediary for the
471   sake of security, annotation services, or shared caching.
474<iref primary="true" item="transforming proxy"/>
475<iref primary="true" item="non-transforming proxy"/>
476   An HTTP-to-HTTP proxy is called a "<x:dfn>transforming proxy</x:dfn>" if it is designed
477   or configured to modify request or response messages in a semantically
478   meaningful way (i.e., modifications, beyond those required by normal
479   HTTP processing, that change the message in a way that would be
480   significant to the original sender or potentially significant to
481   downstream recipients).  For example, a transforming proxy might be
482   acting as a shared annotation server (modifying responses to include
483   references to a local annotation database), a malware filter, a
484   format transcoder, or an intranet-to-Internet privacy filter.  Such
485   transformations are presumed to be desired by the client (or client
486   organization) that selected the proxy and are beyond the scope of
487   this specification.  However, when a proxy is not intended to transform
488   a given message, we use the term "<x:dfn>non-transforming proxy</x:dfn>" to target
489   requirements that preserve HTTP message semantics. See &status-203; and
490   &header-warning; for status and warning codes related to transformations.
492<t><iref primary="true" item="gateway"/><iref primary="true" item="reverse proxy"/>
493<iref primary="true" item="accelerator"/>
494   A "<x:dfn>gateway</x:dfn>" (a.k.a., "<x:dfn>reverse proxy</x:dfn>")
495   is a receiving agent that acts
496   as a layer above some other server(s) and translates the received
497   requests to the underlying server's protocol.  Gateways are often
498   used to encapsulate legacy or untrusted information services, to
499   improve server performance through "<x:dfn>accelerator</x:dfn>" caching, and to
500   enable partitioning or load-balancing of HTTP services across
501   multiple machines.
504   A gateway behaves as an origin server on its outbound connection and
505   as a user agent on its inbound connection.
506   All HTTP requirements applicable to an origin server
507   also apply to the outbound communication of a gateway.
508   A gateway communicates with inbound servers using any protocol that
509   it desires, including private extensions to HTTP that are outside
510   the scope of this specification.  However, an HTTP-to-HTTP gateway
511   that wishes to interoperate with third-party HTTP servers &MUST;
512   conform to HTTP user agent requirements on the gateway's inbound
513   connection and &MUST; implement the <x:ref>Connection</x:ref>
514   (<xref target="header.connection"/>) and <x:ref>Via</x:ref>
515   (<xref target="header.via"/>) header fields for both connections.
517<t><iref primary="true" item="tunnel"/>
518   A "<x:dfn>tunnel</x:dfn>" acts as a blind relay between two connections
519   without changing the messages. Once active, a tunnel is not
520   considered a party to the HTTP communication, though the tunnel might
521   have been initiated by an HTTP request. A tunnel ceases to exist when
522   both ends of the relayed connection are closed. Tunnels are used to
523   extend a virtual connection through an intermediary, such as when
524   Transport Layer Security (TLS, <xref target="RFC5246"/>) is used to
525   establish confidential communication through a shared firewall proxy.
527<t><iref primary="true" item="interception proxy"/>
528<iref primary="true" item="transparent proxy"/>
529<iref primary="true" item="captive portal"/>
530   The above categories for intermediary only consider those acting as
531   participants in the HTTP communication.  There are also intermediaries
532   that can act on lower layers of the network protocol stack, filtering or
533   redirecting HTTP traffic without the knowledge or permission of message
534   senders. Network intermediaries often introduce security flaws or
535   interoperability problems by violating HTTP semantics.  For example, an
536   "<x:dfn>interception proxy</x:dfn>" <xref target="RFC3040"/> (also commonly
537   known as a "<x:dfn>transparent proxy</x:dfn>" <xref target="RFC1919"/> or
538   "<x:dfn>captive portal</x:dfn>")
539   differs from an HTTP proxy because it is not selected by the client.
540   Instead, an interception proxy filters or redirects outgoing TCP port 80
541   packets (and occasionally other common port traffic).
542   Interception proxies are commonly found on public network access points,
543   as a means of enforcing account subscription prior to allowing use of
544   non-local Internet services, and within corporate firewalls to enforce
545   network usage policies.
546   They are indistinguishable from a man-in-the-middle attack.
549   HTTP is defined as a stateless protocol, meaning that each request message
550   can be understood in isolation.  Many implementations depend on HTTP's
551   stateless design in order to reuse proxied connections or dynamically
552   load-balance requests across multiple servers.  Hence, servers &MUST-NOT;
553   assume that two requests on the same connection are from the same user
554   agent unless the connection is secured and specific to that agent.
555   Some non-standard HTTP extensions (e.g., <xref target="RFC4559"/>) have
556   been known to violate this requirement, resulting in security and
557   interoperability problems.
561<section title="Caches" anchor="caches">
562<iref primary="true" item="cache"/>
564   A "<x:dfn>cache</x:dfn>" is a local store of previous response messages and the
565   subsystem that controls its message storage, retrieval, and deletion.
566   A cache stores cacheable responses in order to reduce the response
567   time and network bandwidth consumption on future, equivalent
568   requests. Any client or server &MAY; employ a cache, though a cache
569   cannot be used by a server while it is acting as a tunnel.
572   The effect of a cache is that the request/response chain is shortened
573   if one of the participants along the chain has a cached response
574   applicable to that request. The following illustrates the resulting
575   chain if B has a cached copy of an earlier response from O (via C)
576   for a request that has not been cached by UA or A.
578<figure><artwork type="drawing">
579            &gt;             &gt;
580       <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>
581                  &lt;             &lt;
583<t><iref primary="true" item="cacheable"/>
584   A response is "<x:dfn>cacheable</x:dfn>" if a cache is allowed to store a copy of
585   the response message for use in answering subsequent requests.
586   Even when a response is cacheable, there might be additional
587   constraints placed by the client or by the origin server on when
588   that cached response can be used for a particular request. HTTP
589   requirements for cache behavior and cacheable responses are
590   defined in &caching-overview;. 
593   There are a wide variety of architectures and configurations
594   of caches deployed across the World Wide Web and
595   inside large organizations. These include national hierarchies
596   of proxy caches to save transoceanic bandwidth, collaborative systems that
597   broadcast or multicast cache entries, archives of pre-fetched cache
598   entries for use in off-line or high-latency environments, and so on.
602<section title="Conformance and Error Handling" anchor="conformance">
604   This specification targets conformance criteria according to the role of
605   a participant in HTTP communication.  Hence, HTTP requirements are placed
606   on senders, recipients, clients, servers, user agents, intermediaries,
607   origin servers, proxies, gateways, or caches, depending on what behavior
608   is being constrained by the requirement. Additional (social) requirements
609   are placed on implementations, resource owners, and protocol element
610   registrations when they apply beyond the scope of a single communication.
613   The verb "generate" is used instead of "send" where a requirement
614   differentiates between creating a protocol element and merely forwarding a
615   received element downstream.
618   An implementation is considered conformant if it complies with all of the
619   requirements associated with the roles it partakes in HTTP. Note that
620   SHOULD-level requirements are relevant here, unless one of the documented
621   exceptions is applicable.
624   Conformance applies to both the syntax and semantics of HTTP protocol
625   elements. A sender &MUST-NOT; generate protocol elements that convey a
626   meaning that is known by that sender to be false. A sender &MUST-NOT;
627   generate protocol elements that do not match the grammar defined by the
628   ABNF rules for those protocol elements that are applicable to the sender's
629   role. If a received protocol element is processed, the recipient &MUST; be
630   able to parse any value that would match the ABNF rules for that protocol
631   element, excluding only those rules not applicable to the recipient's role.
634   Unless noted otherwise, a recipient &MAY; attempt to recover a usable
635   protocol element from an invalid construct.  HTTP does not define
636   specific error handling mechanisms except when they have a direct impact
637   on security, since different applications of the protocol require
638   different error handling strategies.  For example, a Web browser might
639   wish to transparently recover from a response where the
640   <x:ref>Location</x:ref> header field doesn't parse according to the ABNF,
641   whereas a systems control client might consider any form of error recovery
642   to be dangerous.
646<section title="Protocol Versioning" anchor="http.version">
647  <x:anchor-alias value="HTTP-version"/>
648  <x:anchor-alias value="HTTP-name"/>
650   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate
651   versions of the protocol. This specification defines version "1.1".
652   The protocol version as a whole indicates the sender's conformance
653   with the set of requirements laid out in that version's corresponding
654   specification of HTTP.
657   The version of an HTTP message is indicated by an HTTP-version field
658   in the first line of the message. HTTP-version is case-sensitive.
660<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-version"/><iref primary="true" item="Grammar" subitem="HTTP-name"/>
661  <x:ref>HTTP-version</x:ref>  = <x:ref>HTTP-name</x:ref> "/" <x:ref>DIGIT</x:ref> "." <x:ref>DIGIT</x:ref>
662  <x:ref>HTTP-name</x:ref>     = <x:abnf-char-sequence>"HTTP"</x:abnf-char-sequence> ; "HTTP", case-sensitive
665   The HTTP version number consists of two decimal digits separated by a "."
666   (period or decimal point).  The first digit ("major version") indicates the
667   HTTP messaging syntax, whereas the second digit ("minor version") indicates
668   the highest minor version to which the sender is
669   conformant and able to understand for future communication.  The minor
670   version advertises the sender's communication capabilities even when the
671   sender is only using a backwards-compatible subset of the protocol,
672   thereby letting the recipient know that more advanced features can
673   be used in response (by servers) or in future requests (by clients).
676   When an HTTP/1.1 message is sent to an HTTP/1.0 recipient
677   <xref target="RFC1945"/> or a recipient whose version is unknown,
678   the HTTP/1.1 message is constructed such that it can be interpreted
679   as a valid HTTP/1.0 message if all of the newer features are ignored.
680   This specification places recipient-version requirements on some
681   new features so that a conformant sender will only use compatible
682   features until it has determined, through configuration or the
683   receipt of a message, that the recipient supports HTTP/1.1.
686   The interpretation of a header field does not change between minor
687   versions of the same major HTTP version, though the default
688   behavior of a recipient in the absence of such a field can change.
689   Unless specified otherwise, header fields defined in HTTP/1.1 are
690   defined for all versions of HTTP/1.x.  In particular, the <x:ref>Host</x:ref>
691   and <x:ref>Connection</x:ref> header fields ought to be implemented by all
692   HTTP/1.x implementations whether or not they advertise conformance with
693   HTTP/1.1.
696   New header fields can be defined such that, when they are
697   understood by a recipient, they might override or enhance the
698   interpretation of previously defined header fields.  When an
699   implementation receives an unrecognized header field, the recipient
700   &MUST; ignore that header field for local processing regardless of
701   the message's HTTP version.  An unrecognized header field received
702   by a proxy &MUST; be forwarded downstream unless the header field's
703   field-name is listed in the message's <x:ref>Connection</x:ref> header field
704   (see <xref target="header.connection"/>).
705   These requirements allow HTTP's functionality to be enhanced without
706   requiring prior update of deployed intermediaries.
709   Intermediaries that process HTTP messages (i.e., all intermediaries
710   other than those acting as tunnels) &MUST; send their own HTTP-version
711   in forwarded messages.  In other words, they &MUST-NOT; blindly
712   forward the first line of an HTTP message without ensuring that the
713   protocol version in that message matches a version to which that
714   intermediary is conformant for both the receiving and
715   sending of messages.  Forwarding an HTTP message without rewriting
716   the HTTP-version might result in communication errors when downstream
717   recipients use the message sender's version to determine what features
718   are safe to use for later communication with that sender.
721   An HTTP client &SHOULD; send a request version equal to the highest
722   version to which the client is conformant and
723   whose major version is no higher than the highest version supported
724   by the server, if this is known.  An HTTP client &MUST-NOT; send a
725   version to which it is not conformant.
728   An HTTP client &MAY; send a lower request version if it is known that
729   the server incorrectly implements the HTTP specification, but only
730   after the client has attempted at least one normal request and determined
731   from the response status or header fields (e.g., <x:ref>Server</x:ref>) that
732   the server improperly handles higher request versions.
735   An HTTP server &SHOULD; send a response version equal to the highest
736   version to which the server is conformant and
737   whose major version is less than or equal to the one received in the
738   request.  An HTTP server &MUST-NOT; send a version to which it is not
739   conformant.  A server &MAY; send a <x:ref>505 (HTTP Version Not
740   Supported)</x:ref> response if it cannot send a response using the
741   major version used in the client's request.
744   An HTTP server &MAY; send an HTTP/1.0 response to an HTTP/1.0 request
745   if it is known or suspected that the client incorrectly implements the
746   HTTP specification and is incapable of correctly processing later
747   version responses, such as when a client fails to parse the version
748   number correctly or when an intermediary is known to blindly forward
749   the HTTP-version even when it doesn't conform to the given minor
750   version of the protocol. Such protocol downgrades &SHOULD-NOT; be
751   performed unless triggered by specific client attributes, such as when
752   one or more of the request header fields (e.g., <x:ref>User-Agent</x:ref>)
753   uniquely match the values sent by a client known to be in error.
756   The intention of HTTP's versioning design is that the major number
757   will only be incremented if an incompatible message syntax is
758   introduced, and that the minor number will only be incremented when
759   changes made to the protocol have the effect of adding to the message
760   semantics or implying additional capabilities of the sender.  However,
761   the minor version was not incremented for the changes introduced between
762   <xref target="RFC2068"/> and <xref target="RFC2616"/>, and this revision
763   has specifically avoiding any such changes to the protocol.
767<section title="Uniform Resource Identifiers" anchor="uri">
768<iref primary="true" item="resource"/>
770   Uniform Resource Identifiers (URIs) <xref target="RFC3986"/> are used
771   throughout HTTP as the means for identifying resources (&resource;).
772   URI references are used to target requests, indicate redirects, and define
773   relationships.
775  <x:anchor-alias value="URI-reference"/>
776  <x:anchor-alias value="absolute-URI"/>
777  <x:anchor-alias value="relative-part"/>
778  <x:anchor-alias value="authority"/>
779  <x:anchor-alias value="path-abempty"/>
780  <x:anchor-alias value="path-absolute"/>
781  <x:anchor-alias value="port"/>
782  <x:anchor-alias value="query"/>
783  <x:anchor-alias value="uri-host"/>
784  <x:anchor-alias value="partial-URI"/>
786   This specification adopts the definitions of "URI-reference",
787   "absolute-URI", "relative-part", "port", "host",
788   "path-abempty", "path-absolute", "query", and "authority" from the
789   URI generic syntax.
790   In addition, we define a partial-URI rule for protocol elements
791   that allow a relative URI but not a fragment.
793<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>
794  <x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in <xref target="RFC3986" x:fmt="," x:sec="4.1"/>&gt;
795  <x:ref>absolute-URI</x:ref>  = &lt;absolute-URI, defined in <xref target="RFC3986" x:fmt="," x:sec="4.3"/>&gt;
796  <x:ref>relative-part</x:ref> = &lt;relative-part, defined in <xref target="RFC3986" x:fmt="," x:sec="4.2"/>&gt;
797  <x:ref>authority</x:ref>     = &lt;authority, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2"/>&gt;
798  <x:ref>path-abempty</x:ref>  = &lt;path-abempty, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
799  <x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
800  <x:ref>port</x:ref>          = &lt;port, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.3"/>&gt;
801  <x:ref>query</x:ref>         = &lt;query, defined in <xref target="RFC3986" x:fmt="," x:sec="3.4"/>&gt;
802  <x:ref>uri-host</x:ref>      = &lt;host, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>&gt;
804  <x:ref>partial-URI</x:ref>   = relative-part [ "?" query ]
807   Each protocol element in HTTP that allows a URI reference will indicate
808   in its ABNF production whether the element allows any form of reference
809   (URI-reference), only a URI in absolute form (absolute-URI), only the
810   path and optional query components, or some combination of the above.
811   Unless otherwise indicated, URI references are parsed
812   relative to the effective request URI
813   (<xref target="effective.request.uri"/>).
816<section title="http URI scheme" anchor="http.uri">
817  <x:anchor-alias value="http-URI"/>
818  <iref item="http URI scheme" primary="true"/>
819  <iref item="URI scheme" subitem="http" primary="true"/>
821   The "http" URI scheme is hereby defined for the purpose of minting
822   identifiers according to their association with the hierarchical
823   namespace governed by a potential HTTP origin server listening for
824   TCP connections on a given port.
826<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http-URI"><!--terminal production--></iref>
827  <x:ref>http-URI</x:ref> = "http:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
830   The HTTP origin server is identified by the generic syntax's
831   <x:ref>authority</x:ref> component, which includes a host identifier
832   and optional TCP port (<xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>).
833   The remainder of the URI, consisting of both the hierarchical path
834   component and optional query component, serves as an identifier for
835   a potential resource within that origin server's name space.
838   If the host identifier is provided as an IP address,
839   then the origin server is any listener on the indicated TCP port at
840   that IP address. If host is a registered name, then that name is
841   considered an indirect identifier and the recipient might use a name
842   resolution service, such as DNS, to find the address of a listener
843   for that host.
844   The host &MUST-NOT; be empty; if an "http" URI is received with an
845   empty host, then it &MUST; be rejected as invalid.
846   If the port subcomponent is empty or not given, then TCP port 80 is
847   assumed (the default reserved port for WWW services).
850   Regardless of the form of host identifier, access to that host is not
851   implied by the mere presence of its name or address. The host might or might
852   not exist and, even when it does exist, might or might not be running an
853   HTTP server or listening to the indicated port. The "http" URI scheme
854   makes use of the delegated nature of Internet names and addresses to
855   establish a naming authority (whatever entity has the ability to place
856   an HTTP server at that Internet name or address) and allows that
857   authority to determine which names are valid and how they might be used.
860   When an "http" URI is used within a context that calls for access to the
861   indicated resource, a client &MAY; attempt access by resolving
862   the host to an IP address, establishing a TCP connection to that address
863   on the indicated port, and sending an HTTP request message
864   (<xref target="http.message"/>) containing the URI's identifying data
865   (<xref target="message.routing"/>) to the server.
866   If the server responds to that request with a non-interim HTTP response
867   message, as described in &status-codes;, then that response
868   is considered an authoritative answer to the client's request.
871   Although HTTP is independent of the transport protocol, the "http"
872   scheme is specific to TCP-based services because the name delegation
873   process depends on TCP for establishing authority.
874   An HTTP service based on some other underlying connection protocol
875   would presumably be identified using a different URI scheme, just as
876   the "https" scheme (below) is used for resources that require an
877   end-to-end secured connection. Other protocols might also be used to
878   provide access to "http" identified resources &mdash; it is only the
879   authoritative interface used for mapping the namespace that is
880   specific to TCP.
883   The URI generic syntax for authority also includes a deprecated
884   userinfo subcomponent (<xref target="RFC3986" x:fmt="," x:sec="3.2.1"/>)
885   for including user authentication information in the URI.  Some
886   implementations make use of the userinfo component for internal
887   configuration of authentication information, such as within command
888   invocation options, configuration files, or bookmark lists, even
889   though such usage might expose a user identifier or password.
890   Senders &MUST; exclude the userinfo subcomponent (and its "@"
891   delimiter) when an "http" URI is transmitted within a message as a
892   request target or header field value.
893   Recipients of an "http" URI reference &SHOULD; parse for userinfo and
894   treat its presence as an error, since it is likely being used to obscure
895   the authority for the sake of phishing attacks.
899<section title="https URI scheme" anchor="https.uri">
900   <x:anchor-alias value="https-URI"/>
901   <iref item="https URI scheme"/>
902   <iref item="URI scheme" subitem="https"/>
904   The "https" URI scheme is hereby defined for the purpose of minting
905   identifiers according to their association with the hierarchical
906   namespace governed by a potential HTTP origin server listening to a
907   given TCP port for TLS-secured connections <xref target="RFC5246"/>.
910   All of the requirements listed above for the "http" scheme are also
911   requirements for the "https" scheme, except that a default TCP port
912   of 443 is assumed if the port subcomponent is empty or not given,
913   and the TCP connection &MUST; be secured, end-to-end, through the
914   use of strong encryption prior to sending the first HTTP request.
916<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="https-URI"><!--terminal production--></iref>
917  <x:ref>https-URI</x:ref> = "https:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
920   Resources made available via the "https" scheme have no shared
921   identity with the "http" scheme even if their resource identifiers
922   indicate the same authority (the same host listening to the same
923   TCP port).  They are distinct name spaces and are considered to be
924   distinct origin servers.  However, an extension to HTTP that is
925   defined to apply to entire host domains, such as the Cookie protocol
926   <xref target="RFC6265"/>, can allow information
927   set by one service to impact communication with other services
928   within a matching group of host domains.
931   The process for authoritative access to an "https" identified
932   resource is defined in <xref target="RFC2818"/>.
936<section title="http and https URI Normalization and Comparison" anchor="uri.comparison">
938   Since the "http" and "https" schemes conform to the URI generic syntax,
939   such URIs are normalized and compared according to the algorithm defined
940   in <xref target="RFC3986" x:fmt="," x:sec="6"/>, using the defaults
941   described above for each scheme.
944   If the port is equal to the default port for a scheme, the normal form is
945   to elide the port subcomponent. When not being used in absolute form as the
946   request target of an OPTIONS request, an empty path component is equivalent
947   to an absolute path of "/", so the normal form is to provide a path of "/"
948   instead. The scheme and host are case-insensitive and normally provided in
949   lowercase; all other components are compared in a case-sensitive manner.
950   Characters other than those in the "reserved" set are equivalent to their
951   percent-encoded octets (see <xref target="RFC3986" x:fmt=","
952   x:sec="2.1"/>): the normal form is to not encode them.
955   For example, the following three URIs are equivalent:
957<figure><artwork type="example">
966<section title="Message Format" anchor="http.message">
967<x:anchor-alias value="generic-message"/>
968<x:anchor-alias value="message.types"/>
969<x:anchor-alias value="HTTP-message"/>
970<x:anchor-alias value="start-line"/>
971<iref item="header section"/>
972<iref item="headers"/>
973<iref item="header field"/>
975   All HTTP/1.1 messages consist of a start-line followed by a sequence of
976   octets in a format similar to the Internet Message Format
977   <xref target="RFC5322"/>: zero or more header fields (collectively
978   referred to as the "headers" or the "header section"), an empty line
979   indicating the end of the header section, and an optional message body.
981<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"><!--terminal production--></iref>
982  <x:ref>HTTP-message</x:ref>   = <x:ref>start-line</x:ref>
983                   *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
984                   <x:ref>CRLF</x:ref>
985                   [ <x:ref>message-body</x:ref> ]
988   The normal procedure for parsing an HTTP message is to read the
989   start-line into a structure, read each header field into a hash
990   table by field name until the empty line, and then use the parsed
991   data to determine if a message body is expected.  If a message body
992   has been indicated, then it is read as a stream until an amount
993   of octets equal to the message body length is read or the connection
994   is closed.
997   Recipients &MUST; parse an HTTP message as a sequence of octets in an
998   encoding that is a superset of US-ASCII <xref target="USASCII"/>.
999   Parsing an HTTP message as a stream of Unicode characters, without regard
1000   for the specific encoding, creates security vulnerabilities due to the
1001   varying ways that string processing libraries handle invalid multibyte
1002   character sequences that contain the octet LF (%x0A).  String-based
1003   parsers can only be safely used within protocol elements after the element
1004   has been extracted from the message, such as within a header field-value
1005   after message parsing has delineated the individual fields.
1008   An HTTP message can be parsed as a stream for incremental processing or
1009   forwarding downstream.  However, recipients cannot rely on incremental
1010   delivery of partial messages, since some implementations will buffer or
1011   delay message forwarding for the sake of network efficiency, security
1012   checks, or payload transformations.
1015<section title="Start Line" anchor="start.line">
1016  <x:anchor-alias value="Start-Line"/>
1018   An HTTP message can either be a request from client to server or a
1019   response from server to client.  Syntactically, the two types of message
1020   differ only in the start-line, which is either a request-line (for requests)
1021   or a status-line (for responses), and in the algorithm for determining
1022   the length of the message body (<xref target="message.body"/>).
1025   In theory, a client could receive requests and a server could receive
1026   responses, distinguishing them by their different start-line formats,
1027   but in practice servers are implemented to only expect a request
1028   (a response is interpreted as an unknown or invalid request method)
1029   and clients are implemented to only expect a response.
1031<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="start-line"/>
1032  <x:ref>start-line</x:ref>     = <x:ref>request-line</x:ref> / <x:ref>status-line</x:ref>
1035   A sender &MUST-NOT; send whitespace between the start-line and
1036   the first header field. The presence of such whitespace in a request
1037   might be an attempt to trick a server into ignoring that field or
1038   processing the line after it as a new request, either of which might
1039   result in a security vulnerability if other implementations within
1040   the request chain interpret the same message differently.
1041   Likewise, the presence of such whitespace in a response might be
1042   ignored by some clients or cause others to cease parsing.
1045   A recipient that receives whitespace between the start-line and
1046   the first header field &MUST; either reject the message as invalid or
1047   consume each whitespace-preceded line without further processing of it
1048   (i.e., ignore the entire line, along with any subsequent lines preceded
1049   by whitespace, until a properly formed header field is received or the
1050   header block is terminated).
1053<section title="Request Line" anchor="request.line">
1054  <x:anchor-alias value="Request"/>
1055  <x:anchor-alias value="request-line"/>
1057   A request-line begins with a method token, followed by a single
1058   space (SP), the request-target, another single space (SP), the
1059   protocol version, and ending with CRLF.
1061<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-line"/>
1062  <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>
1064<iref primary="true" item="method"/>
1065<t anchor="method">
1066   The method token indicates the request method to be performed on the
1067   target resource. The request method is case-sensitive.
1069<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="method"/>
1070  <x:ref>method</x:ref>         = <x:ref>token</x:ref>
1073   The methods defined by this specification can be found in
1074   &methods;, along with information regarding the HTTP method registry
1075   and considerations for defining new methods.
1077<iref item="request-target"/>
1079   The request-target identifies the target resource upon which to apply
1080   the request, as defined in <xref target="request-target"/>.
1083   No whitespace is allowed inside the method, request-target, and
1084   protocol version.  Hence, recipients typically parse the request-line
1085   into its component parts by splitting on whitespace
1086   (see <xref target="message.robustness"/>).
1089   Unfortunately, some user agents fail to properly encode hypertext
1090   references that have embedded whitespace, sending the characters directly
1091   instead of properly encoding or excluding the disallowed characters.
1092   Recipients of an invalid request-line &SHOULD; respond with either a
1093   <x:ref>400 (Bad Request)</x:ref> error or a <x:ref>301 (Moved Permanently)</x:ref>
1094   redirect with the request-target properly encoded.  Recipients &SHOULD-NOT;
1095   attempt to autocorrect and then process the request without a redirect,
1096   since the invalid request-line might be deliberately crafted to bypass
1097   security filters along the request chain.
1100   HTTP does not place a pre-defined limit on the length of a request-line.
1101   A server that receives a method longer than any that it implements
1102   &SHOULD; respond with a <x:ref>501 (Not Implemented)</x:ref> status code.
1103   A server &MUST; be prepared to receive URIs of unbounded length and
1104   respond with the <x:ref>414 (URI Too Long)</x:ref> status code if the received
1105   request-target would be longer than the server wishes to handle
1106   (see &status-414;).
1109   Various ad-hoc limitations on request-line length are found in practice.
1110   It is &RECOMMENDED; that all HTTP senders and recipients support, at a
1111   minimum, request-line lengths of 8000 octets.
1115<section title="Status Line" anchor="status.line">
1116  <x:anchor-alias value="response"/>
1117  <x:anchor-alias value="status-line"/>
1118  <x:anchor-alias value="status-code"/>
1119  <x:anchor-alias value="reason-phrase"/>
1121   The first line of a response message is the status-line, consisting
1122   of the protocol version, a space (SP), the status code, another space,
1123   a possibly-empty textual phrase describing the status code, and
1124   ending with CRLF.
1126<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-line"/>
1127  <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>
1130   The status-code element is a 3-digit integer code describing the
1131   result of the server's attempt to understand and satisfy the client's
1132   corresponding request. The rest of the response message is to be
1133   interpreted in light of the semantics defined for that status code.
1134   See &status-codes; for information about the semantics of status codes,
1135   including the classes of status code (indicated by the first digit),
1136   the status codes defined by this specification, considerations for the
1137   definition of new status codes, and the IANA registry.
1139<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-code"/>
1140  <x:ref>status-code</x:ref>    = 3<x:ref>DIGIT</x:ref>
1143   The reason-phrase element exists for the sole purpose of providing a
1144   textual description associated with the numeric status code, mostly
1145   out of deference to earlier Internet application protocols that were more
1146   frequently used with interactive text clients. A client &SHOULD; ignore
1147   the reason-phrase content.
1149<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="reason-phrase"/>
1150  <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> )
1155<section title="Header Fields" anchor="header.fields">
1156  <x:anchor-alias value="header-field"/>
1157  <x:anchor-alias value="field-content"/>
1158  <x:anchor-alias value="field-name"/>
1159  <x:anchor-alias value="field-value"/>
1160  <x:anchor-alias value="obs-fold"/>
1162   Each HTTP header field consists of a case-insensitive field name
1163   followed by a colon (":"), optional whitespace, and the field value.
1165<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"/>
1166  <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>
1167  <x:ref>field-name</x:ref>     = <x:ref>token</x:ref>
1168  <x:ref>field-value</x:ref>    = *( <x:ref>field-content</x:ref> / <x:ref>obs-fold</x:ref> )
1169  <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> )
1170  <x:ref>obs-fold</x:ref>       = <x:ref>CRLF</x:ref> ( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1171                 ; obsolete line folding
1172                 ; see <xref target="field.parsing"/>
1175   The field-name token labels the corresponding field-value as having the
1176   semantics defined by that header field.  For example, the <x:ref>Date</x:ref>
1177   header field is defined in &header-date; as containing the origination
1178   timestamp for the message in which it appears.
1181<section title="Field Extensibility" anchor="field.extensibility">
1183   HTTP header fields are fully extensible: there is no limit on the
1184   introduction of new field names, each presumably defining new semantics,
1185   nor on the number of header fields used in a given message.  Existing
1186   fields are defined in each part of this specification and in many other
1187   specifications outside the core standard.
1188   New header fields can be introduced without changing the protocol version
1189   if their defined semantics allow them to be safely ignored by recipients
1190   that do not recognize them.
1193   New HTTP header fields ought to be be registered with IANA in the
1194   Message Header Field Registry, as described in &iana-header-registry;.
1195   A proxy &MUST; forward unrecognized header fields unless the
1196   field-name is listed in the <x:ref>Connection</x:ref> header field
1197   (<xref target="header.connection"/>) or the proxy is specifically
1198   configured to block, or otherwise transform, such fields.
1199   Other recipients &SHOULD; ignore unrecognized header fields.
1203<section title="Field Order" anchor="field.order">
1205   The order in which header fields with differing field names are
1206   received is not significant. However, it is "good practice" to send
1207   header fields that contain control data first, such as <x:ref>Host</x:ref>
1208   on requests and <x:ref>Date</x:ref> on responses, so that implementations
1209   can decide when not to handle a message as early as possible.  A server
1210   &MUST; wait until the entire header section is received before interpreting
1211   a request message, since later header fields might include conditionals,
1212   authentication credentials, or deliberately misleading duplicate
1213   header fields that would impact request processing.
1216   A sender &MUST-NOT; generate multiple header fields with the same field
1217   name in a message unless either the entire field value for that
1218   header field is defined as a comma-separated list [i.e., #(values)]
1219   or the header field is a well-known exception (as noted below).
1222   Multiple header fields with the same field name can be combined into
1223   one "field-name: field-value" pair, without changing the semantics of the
1224   message, by appending each subsequent field value to the combined
1225   field value in order, separated by a comma. The order in which
1226   header fields with the same field name are received is therefore
1227   significant to the interpretation of the combined field value;
1228   a proxy &MUST-NOT; change the order of these field values when
1229   forwarding a message.
1232  <t>
1233   &Note; In practice, the "Set-Cookie" header field (<xref target="RFC6265"/>)
1234   often appears multiple times in a response message and does not use the
1235   list syntax, violating the above requirements on multiple header fields
1236   with the same name. Since it cannot be combined into a single field-value,
1237   recipients ought to handle "Set-Cookie" as a special case while processing
1238   header fields. (See Appendix A.2.3 of <xref target="Kri2001"/> for details.)
1239  </t>
1243<section title="Whitespace" anchor="whitespace">
1244<t anchor="rule.LWS">
1245   This specification uses three rules to denote the use of linear
1246   whitespace: OWS (optional whitespace), RWS (required whitespace), and
1247   BWS ("bad" whitespace).
1249<t anchor="rule.OWS">
1250   The OWS rule is used where zero or more linear whitespace octets might
1251   appear. OWS &SHOULD; either not be generated or be generated as a single
1252   SP. Multiple OWS octets that occur within field-content &SHOULD; either
1253   be replaced with a single SP or transformed to all SP octets (each
1254   octet other than SP replaced with SP) before interpreting the field value
1255   or forwarding the message downstream.
1257<t anchor="rule.RWS">
1258   RWS is used when at least one linear whitespace octet is required to
1259   separate field tokens. RWS &SHOULD; be generated as a single SP.
1260   Multiple RWS octets that occur within field-content &SHOULD; either
1261   be replaced with a single SP or transformed to all SP octets before
1262   interpreting the field value or forwarding the message downstream.
1264<t anchor="rule.BWS">
1265   BWS is used where the grammar allows optional whitespace, for historical
1266   reasons, but senders &SHOULD-NOT; generate it in messages;
1267   recipients &MUST; accept such bad optional whitespace and remove it before
1268   interpreting the field value or forwarding the message downstream.
1270<t anchor="rule.whitespace">
1271  <x:anchor-alias value="BWS"/>
1272  <x:anchor-alias value="OWS"/>
1273  <x:anchor-alias value="RWS"/>
1275<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"/>
1276  <x:ref>OWS</x:ref>            = *( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1277                 ; optional whitespace
1278  <x:ref>RWS</x:ref>            = 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1279                 ; required whitespace
1280  <x:ref>BWS</x:ref>            = <x:ref>OWS</x:ref>
1281                 ; "bad" whitespace
1285<section title="Field Parsing" anchor="field.parsing">
1287   No whitespace is allowed between the header field-name and colon.
1288   In the past, differences in the handling of such whitespace have led to
1289   security vulnerabilities in request routing and response handling.
1290   A server &MUST; reject any received request message that contains
1291   whitespace between a header field-name and colon with a response code of
1292   <x:ref>400 (Bad Request)</x:ref>. A proxy &MUST; remove any such whitespace
1293   from a response message before forwarding the message downstream.
1296   A field value is preceded by optional whitespace (OWS); a single SP is
1297   preferred. The field value does not include any leading or trailing white
1298   space: OWS occurring before the first non-whitespace octet of the
1299   field value or after the last non-whitespace octet of the field value
1300   is ignored and &SHOULD; be removed before further processing (as this does
1301   not change the meaning of the header field).
1304   Historically, HTTP header field values could be extended over multiple
1305   lines by preceding each extra line with at least one space or horizontal
1306   tab (obs-fold). This specification deprecates such line
1307   folding except within the message/http media type
1308   (<xref target=""/>).
1309   Senders &MUST-NOT; generate messages that include line folding
1310   (i.e., that contain any field-value that matches the obs-fold rule) unless
1311   the message is intended for packaging within the message/http media type.
1312   Recipients &MUST; accept line folding and replace any embedded
1313   obs-fold whitespace with either a single SP or a matching number of SP
1314   octets (to avoid buffer copying) prior to interpreting the field value or
1315   forwarding the message downstream.
1318   Historically, HTTP has allowed field content with text in the ISO-8859-1
1319   <xref target="ISO-8859-1"/> charset, supporting other charsets only
1320   through use of <xref target="RFC2047"/> encoding.
1321   In practice, most HTTP header field values use only a subset of the
1322   US-ASCII charset <xref target="USASCII"/>. Newly defined
1323   header fields &SHOULD; limit their field values to US-ASCII octets.
1324   Recipients &SHOULD; treat other octets in field content (obs-text) as
1325   opaque data.
1329<section title="Field Limits" anchor="field.limits">
1331   HTTP does not place a pre-defined limit on the length of each header field
1332   or on the length of the header block as a whole.  Various ad-hoc
1333   limitations on individual header field length are found in practice,
1334   often depending on the specific field semantics.
1337   A server &MUST; be prepared to receive request header fields of unbounded
1338   length and respond with an appropriate <x:ref>4xx (Client Error)</x:ref>
1339   status code if the received header field(s) are larger than the server
1340   wishes to process.
1343   A client &MUST; be prepared to receive response header fields of unbounded
1344   length. A client &MAY; discard or truncate received header fields that are
1345   larger than the client wishes to process if the field semantics are such
1346   that the dropped value(s) can be safely ignored without changing the
1347   response semantics.
1351<section title="Field value components" anchor="field.components">
1352<t anchor="rule.token.separators">
1353  <x:anchor-alias value="tchar"/>
1354  <x:anchor-alias value="token"/>
1355  <x:anchor-alias value="special"/>
1356  <x:anchor-alias value="word"/>
1357   Many HTTP header field values consist of words (token or quoted-string)
1358   separated by whitespace or special characters. These special characters
1359   &MUST; be in a quoted string to be used within a parameter value (as defined
1360   in <xref target="transfer.codings"/>).
1362<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>
1363  <x:ref>word</x:ref>           = <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref>
1365  <x:ref>token</x:ref>          = 1*<x:ref>tchar</x:ref>
1367  IMPORTANT: when editing "tchar" make sure that "special" is updated accordingly!!!
1368 -->
1369  <x:ref>tchar</x:ref>          = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*"
1370                 / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~"
1371                 / <x:ref>DIGIT</x:ref> / <x:ref>ALPHA</x:ref>
1372                 ; any <x:ref>VCHAR</x:ref>, except <x:ref>special</x:ref>
1374  <x:ref>special</x:ref>        = "(" / ")" / "&lt;" / ">" / "@" / ","
1375                 / ";" / ":" / "\" / DQUOTE / "/" / "["
1376                 / "]" / "?" / "=" / "{" / "}"
1378<t anchor="rule.quoted-string">
1379  <x:anchor-alias value="quoted-string"/>
1380  <x:anchor-alias value="qdtext"/>
1381  <x:anchor-alias value="obs-text"/>
1382   A string of text is parsed as a single word if it is quoted using
1383   double-quote marks.
1385<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"/>
1386  <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>
1387  <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>
1388  <x:ref>obs-text</x:ref>       = %x80-FF
1390<t anchor="rule.quoted-pair">
1391  <x:anchor-alias value="quoted-pair"/>
1392   The backslash octet ("\") can be used as a single-octet
1393   quoting mechanism within quoted-string constructs:
1395<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
1396  <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> )
1399   Recipients that process the value of a quoted-string &MUST; handle a
1400   quoted-pair as if it were replaced by the octet following the backslash.
1403   Senders &SHOULD-NOT; generate a quoted-pair in a quoted-string except where
1404   necessary to quote DQUOTE and backslash octets occurring within that string.
1406<t anchor="rule.comment">
1407  <x:anchor-alias value="comment"/>
1408  <x:anchor-alias value="ctext"/>
1409   Comments can be included in some HTTP header fields by surrounding
1410   the comment text with parentheses. Comments are only allowed in
1411   fields containing "comment" as part of their field value definition.
1413<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
1414  <x:ref>comment</x:ref>        = "(" *( <x:ref>ctext</x:ref> / <x:ref>quoted-cpair</x:ref> / <x:ref>comment</x:ref> ) ")"
1415  <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>
1417<t anchor="rule.quoted-cpair">
1418  <x:anchor-alias value="quoted-cpair"/>
1419   The backslash octet ("\") can be used as a single-octet
1420   quoting mechanism within comment constructs:
1422<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-cpair"/>
1423  <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> )
1426   Senders &SHOULD-NOT; escape octets in comments that do not require escaping
1427   (i.e., other than the backslash octet "\" and the parentheses "(" and ")").
1433<section title="Message Body" anchor="message.body">
1434  <x:anchor-alias value="message-body"/>
1436   The message body (if any) of an HTTP message is used to carry the
1437   payload body of that request or response.  The message body is
1438   identical to the payload body unless a transfer coding has been
1439   applied, as described in <xref target="header.transfer-encoding"/>.
1441<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1442  <x:ref>message-body</x:ref> = *OCTET
1445   The rules for when a message body is allowed in a message differ for
1446   requests and responses.
1449   The presence of a message body in a request is signaled by a
1450   <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1451   field. Request message framing is independent of method semantics,
1452   even if the method does not define any use for a message body.
1455   The presence of a message body in a response depends on both
1456   the request method to which it is responding and the response
1457   status code (<xref target="status.line"/>).
1458   Responses to the HEAD request method never include a message body
1459   because the associated response header fields (e.g.,
1460   <x:ref>Transfer-Encoding</x:ref>, <x:ref>Content-Length</x:ref>, etc.),
1461   if present, indicate only what their values would have been if the request
1462   method had been GET (&HEAD;).
1463   <x:ref>2xx (Successful)</x:ref> responses to CONNECT switch to tunnel
1464   mode instead of having a message body (&CONNECT;).
1465   All <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, and
1466   <x:ref>304 (Not Modified)</x:ref> responses do not include a message body.
1467   All other responses do include a message body, although the body
1468   might be of zero length.
1471<section title="Transfer-Encoding" anchor="header.transfer-encoding">
1472  <iref primary="true" item="Transfer-Encoding header field" x:for-anchor=""/>
1473  <iref item="chunked (Coding Format)"/>
1474  <x:anchor-alias value="Transfer-Encoding"/>
1476   The Transfer-Encoding header field lists the transfer coding names
1477   corresponding to the sequence of transfer codings that have been
1478   (or will be) applied to the payload body in order to form the message body.
1479   Transfer codings are defined in <xref target="transfer.codings"/>.
1481<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
1482  <x:ref>Transfer-Encoding</x:ref> = 1#<x:ref>transfer-coding</x:ref>
1485   Transfer-Encoding is analogous to the Content-Transfer-Encoding field of
1486   MIME, which was designed to enable safe transport of binary data over a
1487   7-bit transport service (<xref target="RFC2045" x:fmt="," x:sec="6"/>).
1488   However, safe transport has a different focus for an 8bit-clean transfer
1489   protocol. In HTTP's case, Transfer-Encoding is primarily intended to
1490   accurately delimit a dynamically generated payload and to distinguish
1491   payload encodings that are only applied for transport efficiency or
1492   security from those that are characteristics of the selected resource.
1495   All HTTP/1.1 recipients &MUST; implement the chunked transfer coding
1496   (<xref target="chunked.encoding"/>) because it plays a crucial role in
1497   framing messages when the payload body size is not known in advance.
1498   If chunked is applied to a payload body, the sender &MUST-NOT; apply
1499   chunked more than once (i.e., chunking an already chunked message is not
1500   allowed).
1501   If any transfer coding is applied to a request payload body, the
1502   sender &MUST; apply chunked as the final transfer coding to ensure that
1503   the message is properly framed.
1504   If any transfer coding is applied to a response payload body, the
1505   sender &MUST; either apply chunked as the final transfer coding or
1506   terminate the message by closing the connection.
1509   For example,
1510</preamble><artwork type="example">
1511  Transfer-Encoding: gzip, chunked
1513   indicates that the payload body has been compressed using the gzip
1514   coding and then chunked using the chunked coding while forming the
1515   message body.
1518   Unlike <x:ref>Content-Encoding</x:ref> (&content-codings;),
1519   Transfer-Encoding is a property of the message, not of the payload, and
1520   any recipient along the request/response chain &MAY; decode the received
1521   transfer coding(s) or apply additional transfer coding(s) to the message
1522   body, assuming that corresponding changes are made to the Transfer-Encoding
1523   field-value. Additional information about the encoding parameters &MAY; be
1524   provided by other header fields not defined by this specification.
1527   Transfer-Encoding &MAY; be sent in a response to a HEAD request or in a
1528   <x:ref>304 (Not Modified)</x:ref> response (&status-304;) to a GET request,
1529   neither of which includes a message body,
1530   to indicate that the origin server would have applied a transfer coding
1531   to the message body if the request had been an unconditional GET.
1532   This indication is not required, however, because any recipient on
1533   the response chain (including the origin server) can remove transfer
1534   codings when they are not needed.
1537   Transfer-Encoding was added in HTTP/1.1.  It is generally assumed that
1538   implementations advertising only HTTP/1.0 support will not understand
1539   how to process a transfer-encoded payload.
1540   A client &MUST-NOT; send a request containing Transfer-Encoding unless it
1541   knows the server will handle HTTP/1.1 (or later) requests; such knowledge
1542   might be in the form of specific user configuration or by remembering the
1543   version of a prior received response.
1544   A server &MUST-NOT; send a response containing Transfer-Encoding unless
1545   the corresponding request indicates HTTP/1.1 (or later).
1548   A server that receives a request message with a transfer coding it does
1549   not understand &SHOULD; respond with <x:ref>501 (Not Implemented)</x:ref>.
1553<section title="Content-Length" anchor="header.content-length">
1554  <iref primary="true" item="Content-Length header field" x:for-anchor=""/>
1555  <x:anchor-alias value="Content-Length"/>
1557   When a message does not have a <x:ref>Transfer-Encoding</x:ref> header
1558   field, a Content-Length header field can provide the anticipated size,
1559   as a decimal number of octets, for a potential payload body.
1560   For messages that do include a payload body, the Content-Length field-value
1561   provides the framing information necessary for determining where the body
1562   (and message) ends.  For messages that do not include a payload body, the
1563   Content-Length indicates the size of the selected representation
1564   (&representation;).
1566<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
1567  <x:ref>Content-Length</x:ref> = 1*<x:ref>DIGIT</x:ref>
1570   An example is
1572<figure><artwork type="example">
1573  Content-Length: 3495
1576   A sender &MUST-NOT; send a Content-Length header field in any message that
1577   contains a <x:ref>Transfer-Encoding</x:ref> header field.
1580   A user agent &SHOULD; send a Content-Length in a request message when no
1581   <x:ref>Transfer-Encoding</x:ref> is sent and the request method defines
1582   a meaning for an enclosed payload body. For example, a Content-Length
1583   header field is normally sent in a POST request even when the value is
1584   0 (indicating an empty payload body).  A user agent &SHOULD-NOT; send a
1585   Content-Length header field when the request message does not contain a
1586   payload body and the method semantics do not anticipate such a body.
1589   A server &MAY; send a Content-Length header field in a response to a HEAD
1590   request (&HEAD;); a server &MUST-NOT; send Content-Length in such a
1591   response unless its field-value equals the decimal number of octets that
1592   would have been sent in the payload body of a response if the same
1593   request had used the GET method.
1596   A server &MAY; send a Content-Length header field in a
1597   <x:ref>304 (Not Modified)</x:ref> response to a conditional GET request
1598   (&status-304;); a server &MUST-NOT; send Content-Length in such a
1599   response unless its field-value equals the decimal number of octets that
1600   would have been sent in the payload body of a <x:ref>200 (OK)</x:ref>
1601   response to the same request.
1604   A server &MUST-NOT; send a Content-Length header field in any response
1605   with a status code of
1606   <x:ref>1xx (Informational)</x:ref> or <x:ref>204 (No Content)</x:ref>.
1607   A server &SHOULD-NOT; send a Content-Length header field in any
1608   <x:ref>2xx (Successful)</x:ref> response to a CONNECT request (&CONNECT;).
1611   Aside from the cases defined above, in the absence of Transfer-Encoding,
1612   an origin server &SHOULD; send a Content-Length header field when the
1613   payload body size is known prior to sending the complete header block.
1614   This will allow downstream recipients to measure transfer progress,
1615   know when a received message is complete, and potentially reuse the
1616   connection for additional requests.
1619   Any Content-Length field value greater than or equal to zero is valid.
1620   Since there is no predefined limit to the length of an HTTP payload,
1621   recipients &SHOULD; anticipate potentially large decimal numerals and
1622   prevent parsing errors due to integer conversion overflows
1623   (<xref target="attack.protocol.element.size.overflows"/>).
1626   If a message is received that has multiple Content-Length header fields
1627   with field-values consisting of the same decimal value, or a single
1628   Content-Length header field with a field value containing a list of
1629   identical decimal values (e.g., "Content-Length: 42, 42"), indicating that
1630   duplicate Content-Length header fields have been generated or combined by an
1631   upstream message processor, then the recipient &MUST; either reject the
1632   message as invalid or replace the duplicated field-values with a single
1633   valid Content-Length field containing that decimal value prior to
1634   determining the message body length.
1637  <t>
1638   &Note; HTTP's use of Content-Length for message framing differs
1639   significantly from the same field's use in MIME, where it is an optional
1640   field used only within the "message/external-body" media-type.
1641  </t>
1645<section title="Message Body Length" anchor="message.body.length">
1646  <iref item="chunked (Coding Format)"/>
1648   The length of a message body is determined by one of the following
1649   (in order of precedence):
1652  <list style="numbers">
1653    <x:lt><t>
1654     Any response to a HEAD request and any response with a
1655     <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, or
1656     <x:ref>304 (Not Modified)</x:ref> status code is always
1657     terminated by the first empty line after the header fields, regardless of
1658     the header fields present in the message, and thus cannot contain a
1659     message body.
1660    </t></x:lt>
1661    <x:lt><t>
1662     Any <x:ref>2xx (Successful)</x:ref> response to a CONNECT request implies that the
1663     connection will become a tunnel immediately after the empty line that
1664     concludes the header fields.  A client &MUST; ignore any
1665     <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1666     fields received in such a message.
1667    </t></x:lt>
1668    <x:lt><t>
1669     If a <x:ref>Transfer-Encoding</x:ref> header field is present
1670     and the chunked transfer coding (<xref target="chunked.encoding"/>)
1671     is the final encoding, the message body length is determined by reading
1672     and decoding the chunked data until the transfer coding indicates the
1673     data is complete.
1674    </t>
1675    <t>
1676     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a
1677     response and the chunked transfer coding is not the final encoding, the
1678     message body length is determined by reading the connection until it is
1679     closed by the server.
1680     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a request and the
1681     chunked transfer coding is not the final encoding, the message body
1682     length cannot be determined reliably; the server &MUST; respond with
1683     the <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1684    </t>
1685    <t>
1686     If a message is received with both a <x:ref>Transfer-Encoding</x:ref>
1687     and a <x:ref>Content-Length</x:ref> header field, the Transfer-Encoding
1688     overrides the Content-Length. Such a message might indicate an attempt
1689     to perform request or response smuggling (bypass of security-related
1690     checks on message routing or content) and thus ought to be handled as
1691     an error.  A sender &MUST; remove the received Content-Length field
1692     prior to forwarding such a message downstream.
1693    </t></x:lt>
1694    <x:lt><t>
1695     If a message is received without <x:ref>Transfer-Encoding</x:ref> and with
1696     either multiple <x:ref>Content-Length</x:ref> header fields having
1697     differing field-values or a single Content-Length header field having an
1698     invalid value, then the message framing is invalid and &MUST; be treated
1699     as an error to prevent request or response smuggling.
1700     If this is a request message, the server &MUST; respond with
1701     a <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1702     If this is a response message received by a proxy, the proxy
1703     &MUST; discard the received response, send a <x:ref>502 (Bad Gateway)</x:ref>
1704     status code as its downstream response, and then close the connection.
1705     If this is a response message received by a user agent, it &MUST; be
1706     treated as an error by discarding the message and closing the connection.
1707    </t></x:lt>
1708    <x:lt><t>
1709     If a valid <x:ref>Content-Length</x:ref> header field is present without
1710     <x:ref>Transfer-Encoding</x:ref>, its decimal value defines the
1711     expected message body length in octets.
1712     If the sender closes the connection or the recipient times out before the
1713     indicated number of octets are received, the recipient &MUST; consider
1714     the message to be incomplete and close the connection.
1715    </t></x:lt>
1716    <x:lt><t>
1717     If this is a request message and none of the above are true, then the
1718     message body length is zero (no message body is present).
1719    </t></x:lt>
1720    <x:lt><t>
1721     Otherwise, this is a response message without a declared message body
1722     length, so the message body length is determined by the number of octets
1723     received prior to the server closing the connection.
1724    </t></x:lt>
1725  </list>
1728   Since there is no way to distinguish a successfully completed,
1729   close-delimited message from a partially-received message interrupted
1730   by network failure, a server &SHOULD; use encoding or
1731   length-delimited messages whenever possible.  The close-delimiting
1732   feature exists primarily for backwards compatibility with HTTP/1.0.
1735   A server &MAY; reject a request that contains a message body but
1736   not a <x:ref>Content-Length</x:ref> by responding with
1737   <x:ref>411 (Length Required)</x:ref>.
1740   Unless a transfer coding other than chunked has been applied,
1741   a client that sends a request containing a message body &SHOULD;
1742   use a valid <x:ref>Content-Length</x:ref> header field if the message body
1743   length is known in advance, rather than the chunked transfer coding, since some
1744   existing services respond to chunked with a <x:ref>411 (Length Required)</x:ref>
1745   status code even though they understand the chunked transfer coding.  This
1746   is typically because such services are implemented via a gateway that
1747   requires a content-length in advance of being called and the server
1748   is unable or unwilling to buffer the entire request before processing.
1751   A user agent that sends a request containing a message body &MUST; send a
1752   valid <x:ref>Content-Length</x:ref> header field if it does not know the
1753   server will handle HTTP/1.1 (or later) requests; such knowledge can be in
1754   the form of specific user configuration or by remembering the version of a
1755   prior received response.
1758   If the final response to the last request on a connection has been
1759   completely received and there remains additional data to read, a user agent
1760   &MAY; discard the remaining data or attempt to determine if that data
1761   belongs as part of the prior response body, which might be the case if the
1762   prior message's Content-Length value is incorrect. A client &MUST-NOT;
1763   process, cache, or forward such extra data as a separate response, since
1764   such behavior would be vulnerable to cache poisoning.
1769<section anchor="incomplete.messages" title="Handling Incomplete Messages">
1771   A server that receives an incomplete request message, usually due to a
1772   canceled request or a triggered time-out exception, &MAY; send an error
1773   response prior to closing the connection.
1776   A client that receives an incomplete response message, which can occur
1777   when a connection is closed prematurely or when decoding a supposedly
1778   chunked transfer coding fails, &MUST; record the message as incomplete.
1779   Cache requirements for incomplete responses are defined in
1780   &cache-incomplete;.
1783   If a response terminates in the middle of the header block (before the
1784   empty line is received) and the status code might rely on header fields to
1785   convey the full meaning of the response, then the client cannot assume
1786   that meaning has been conveyed; the client might need to repeat the
1787   request in order to determine what action to take next.
1790   A message body that uses the chunked transfer coding is
1791   incomplete if the zero-sized chunk that terminates the encoding has not
1792   been received.  A message that uses a valid <x:ref>Content-Length</x:ref> is
1793   incomplete if the size of the message body received (in octets) is less than
1794   the value given by Content-Length.  A response that has neither chunked
1795   transfer coding nor Content-Length is terminated by closure of the
1796   connection, and thus is considered complete regardless of the number of
1797   message body octets received, provided that the header block was received
1798   intact.
1802<section title="Message Parsing Robustness" anchor="message.robustness">
1804   Older HTTP/1.0 user agent implementations might send an extra CRLF
1805   after a POST request as a lame workaround for some early server
1806   applications that failed to read message body content that was
1807   not terminated by a line-ending. An HTTP/1.1 user agent &MUST-NOT;
1808   preface or follow a request with an extra CRLF.  If terminating
1809   the request message body with a line-ending is desired, then the
1810   user agent &MUST; count the terminating CRLF octets as part of the
1811   message body length.
1814   In the interest of robustness, servers &SHOULD; ignore at least one
1815   empty line received where a request-line is expected. In other words, if
1816   a server is reading the protocol stream at the beginning of a
1817   message and receives a CRLF first, the server &SHOULD; ignore the CRLF.
1820   Although the line terminator for the start-line and header
1821   fields is the sequence CRLF, recipients &MAY; recognize a
1822   single LF as a line terminator and ignore any preceding CR.
1825   Although the request-line and status-line grammar rules require that each
1826   of the component elements be separated by a single SP octet, recipients
1827   &MAY; instead parse on whitespace-delimited word boundaries and, aside
1828   from the CRLF terminator, treat any form of whitespace as the SP separator
1829   while ignoring preceding or trailing whitespace;
1830   such whitespace includes one or more of the following octets:
1831   SP, HTAB, VT (%x0B), FF (%x0C), or bare CR.
1834   When a server listening only for HTTP request messages, or processing
1835   what appears from the start-line to be an HTTP request message,
1836   receives a sequence of octets that does not match the HTTP-message
1837   grammar aside from the robustness exceptions listed above, the
1838   server &SHOULD; respond with a <x:ref>400 (Bad Request)</x:ref> response. 
1843<section title="Transfer Codings" anchor="transfer.codings">
1844  <x:anchor-alias value="transfer-coding"/>
1845  <x:anchor-alias value="transfer-extension"/>
1847   Transfer coding names are used to indicate an encoding
1848   transformation that has been, can be, or might need to be applied to a
1849   payload body in order to ensure "safe transport" through the network.
1850   This differs from a content coding in that the transfer coding is a
1851   property of the message rather than a property of the representation
1852   that is being transferred.
1854<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1855  <x:ref>transfer-coding</x:ref>    = "chunked" ; <xref target="chunked.encoding"/>
1856                     / "compress" ; <xref target="compress.coding"/>
1857                     / "deflate" ; <xref target="deflate.coding"/>
1858                     / "gzip" ; <xref target="gzip.coding"/>
1859                     / <x:ref>transfer-extension</x:ref>
1860  <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> )
1862<t anchor="rule.parameter">
1863  <x:anchor-alias value="attribute"/>
1864  <x:anchor-alias value="transfer-parameter"/>
1865  <x:anchor-alias value="value"/>
1866   Parameters are in the form of attribute/value pairs.
1868<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"/>
1869  <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>
1870  <x:ref>attribute</x:ref>          = <x:ref>token</x:ref>
1871  <x:ref>value</x:ref>              = <x:ref>word</x:ref>
1874   All transfer-coding names are case-insensitive and ought to be registered
1875   within the HTTP Transfer Coding registry, as defined in
1876   <xref target="transfer.coding.registry"/>.
1877   They are used in the <x:ref>TE</x:ref> (<xref target="header.te"/>) and
1878   <x:ref>Transfer-Encoding</x:ref> (<xref target="header.transfer-encoding"/>)
1879   header fields.
1882<section title="Chunked Transfer Coding" anchor="chunked.encoding">
1883  <iref primary="true" item="chunked (Coding Format)"/>
1884  <x:anchor-alias value="chunk"/>
1885  <x:anchor-alias value="chunked-body"/>
1886  <x:anchor-alias value="chunk-data"/>
1887  <x:anchor-alias value="chunk-ext"/>
1888  <x:anchor-alias value="chunk-ext-name"/>
1889  <x:anchor-alias value="chunk-ext-val"/>
1890  <x:anchor-alias value="chunk-size"/>
1891  <x:anchor-alias value="last-chunk"/>
1892  <x:anchor-alias value="trailer-part"/>
1893  <x:anchor-alias value="quoted-str-nf"/>
1894  <x:anchor-alias value="qdtext-nf"/>
1896   The chunked transfer coding modifies the body of a message in order to
1897   transfer it as a series of chunks, each with its own size indicator,
1898   followed by an &OPTIONAL; trailer containing header fields. This
1899   allows dynamically generated content to be transferred along with the
1900   information necessary for the recipient to verify that it has
1901   received the full message.
1903<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"/>
1904  <x:ref>chunked-body</x:ref>   = *<x:ref>chunk</x:ref>
1905                   <x:ref>last-chunk</x:ref>
1906                   <x:ref>trailer-part</x:ref>
1907                   <x:ref>CRLF</x:ref>
1909  <x:ref>chunk</x:ref>          = <x:ref>chunk-size</x:ref> [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1910                   <x:ref>chunk-data</x:ref> <x:ref>CRLF</x:ref>
1911  <x:ref>chunk-size</x:ref>     = 1*<x:ref>HEXDIG</x:ref>
1912  <x:ref>last-chunk</x:ref>     = 1*("0") [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1914  <x:ref>chunk-ext</x:ref>      = *( ";" <x:ref>chunk-ext-name</x:ref> [ "=" <x:ref>chunk-ext-val</x:ref> ] )
1915  <x:ref>chunk-ext-name</x:ref> = <x:ref>token</x:ref>
1916  <x:ref>chunk-ext-val</x:ref>  = <x:ref>token</x:ref> / <x:ref>quoted-str-nf</x:ref>
1917  <x:ref>chunk-data</x:ref>     = 1*<x:ref>OCTET</x:ref> ; a sequence of chunk-size octets
1918  <x:ref>trailer-part</x:ref>   = *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
1920  <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>
1921                 ; like <x:ref>quoted-string</x:ref>, but disallowing line folding
1922  <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>
1925   Chunk extensions within the chunked transfer coding are deprecated.
1926   Senders &SHOULD-NOT; send chunk-ext.
1927   Definition of new chunk extensions is discouraged.
1930   The chunk-size field is a string of hex digits indicating the size of
1931   the chunk-data in octets. The chunked transfer coding is complete when a
1932   chunk with a chunk-size of zero is received, possibly followed by a
1933   trailer, and finally terminated by an empty line.
1936<section title="Trailer" anchor="header.trailer">
1937  <iref primary="true" item="Trailer header field" x:for-anchor=""/>
1938  <x:anchor-alias value="Trailer"/>
1940   A trailer allows the sender to include additional fields at the end of a
1941   chunked message in order to supply metadata that might be dynamically
1942   generated while the message body is sent, such as a message integrity
1943   check, digital signature, or post-processing status.
1944   The trailer &MUST-NOT; contain fields that need to be known before a
1945   recipient processes the body, such as <x:ref>Transfer-Encoding</x:ref>,
1946   <x:ref>Content-Length</x:ref>, and <x:ref>Trailer</x:ref>.
1949   When a message includes a message body encoded with the chunked
1950   transfer coding and the sender desires to send metadata in the form of
1951   trailer fields at the end of the message, the sender &SHOULD; send a
1952   <x:ref>Trailer</x:ref> header field before the message body to indicate
1953   which fields will be present in the trailers. This allows the recipient
1954   to prepare for receipt of that metadata before it starts processing the body,
1955   which is useful if the message is being streamed and the recipient wishes
1956   to confirm an integrity check on the fly.
1958<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
1959  <x:ref>Trailer</x:ref> = 1#<x:ref>field-name</x:ref>
1962   If no <x:ref>Trailer</x:ref> header field is present, the sender of a
1963   chunked message body &SHOULD; send an empty trailer.
1966   A server &MUST; send an empty trailer with the chunked transfer coding
1967   unless at least one of the following is true:
1968  <list style="numbers">
1969    <t>the request included a <x:ref>TE</x:ref> header field that indicates
1970    "trailers" is acceptable in the transfer coding of the response, as
1971    described in <xref target="header.te"/>; or,</t>
1973    <t>the trailer fields consist entirely of optional metadata and the
1974    recipient could use the message (in a manner acceptable to the server where
1975    the field originated) without receiving that metadata. In other words,
1976    the server that generated the header field is willing to accept the
1977    possibility that the trailer fields might be silently discarded along
1978    the path to the client.</t>
1979  </list>
1982   The above requirement prevents the need for an infinite buffer when a
1983   message is being received by an HTTP/1.1 (or later) proxy and forwarded to
1984   an HTTP/1.0 recipient.
1988<section title="Decoding chunked" anchor="decoding.chunked">
1990   A process for decoding the chunked transfer coding
1991   can be represented in pseudo-code as:
1993<figure><artwork type="code">
1994  length := 0
1995  read chunk-size, chunk-ext (if any) and CRLF
1996  while (chunk-size &gt; 0) {
1997     read chunk-data and CRLF
1998     append chunk-data to decoded-body
1999     length := length + chunk-size
2000     read chunk-size and CRLF
2001  }
2002  read header-field
2003  while (header-field not empty) {
2004     append header-field to existing header fields
2005     read header-field
2006  }
2007  Content-Length := length
2008  Remove "chunked" from Transfer-Encoding
2009  Remove Trailer from existing header fields
2012   All recipients &MUST; be able to receive and decode the
2013   chunked transfer coding and &MUST; ignore chunk-ext extensions
2014   they do not understand.
2019<section title="Compression Codings" anchor="compression.codings">
2021   The codings defined below can be used to compress the payload of a
2022   message.
2025<section title="Compress Coding" anchor="compress.coding">
2026<iref item="compress (Coding Format)"/>
2028   The "compress" format is produced by the common UNIX file compression
2029   program "compress". This format is an adaptive Lempel-Ziv-Welch
2030   coding (LZW). Recipients &SHOULD; consider "x-compress" to be
2031   equivalent to "compress".
2035<section title="Deflate Coding" anchor="deflate.coding">
2036<iref item="deflate (Coding Format)"/>
2038   The "deflate" format is defined as the "deflate" compression mechanism
2039   (described in <xref target="RFC1951"/>) used inside the "zlib"
2040   data format (<xref target="RFC1950"/>).
2043  <t>
2044    &Note; Some incorrect implementations send the "deflate"
2045    compressed data without the zlib wrapper.
2046   </t>
2050<section title="Gzip Coding" anchor="gzip.coding">
2051<iref item="gzip (Coding Format)"/>
2053   The "gzip" format is produced by the file compression program
2054   "gzip" (GNU zip), as described in <xref target="RFC1952"/>. This format is a
2055   Lempel-Ziv coding (LZ77) with a 32 bit CRC.
2056   Recipients &SHOULD; consider "x-gzip" to be equivalent to "gzip".
2062<section title="TE" anchor="header.te">
2063  <iref primary="true" item="TE header field" x:for-anchor=""/>
2064  <x:anchor-alias value="TE"/>
2065  <x:anchor-alias value="t-codings"/>
2066  <x:anchor-alias value="t-ranking"/>
2067  <x:anchor-alias value="rank"/>
2069   The "TE" header field in a request indicates what transfer codings,
2070   besides chunked, the client is willing to accept in response, and
2071   whether or not the client is willing to accept trailer fields in a
2072   chunked transfer coding.
2075   The TE field-value consists of a comma-separated list of transfer coding
2076   names, each allowing for optional parameters (as described in
2077   <xref target="transfer.codings"/>), and/or the keyword "trailers".
2078   Clients &MUST-NOT; send the chunked transfer coding name in TE;
2079   chunked is always acceptable for HTTP/1.1 recipients.
2081<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"/>
2082  <x:ref>TE</x:ref>        = #<x:ref>t-codings</x:ref>
2083  <x:ref>t-codings</x:ref> = "trailers" / ( <x:ref>transfer-coding</x:ref> [ <x:ref>t-ranking</x:ref> ] )
2084  <x:ref>t-ranking</x:ref> = <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> "q=" <x:ref>rank</x:ref>
2085  <x:ref>rank</x:ref>      = ( "0" [ "." 0*3<x:ref>DIGIT</x:ref> ] )
2086             / ( "1" [ "." 0*3("0") ] )
2089   Three examples of TE use are below.
2091<figure><artwork type="example">
2092  TE: deflate
2093  TE:
2094  TE: trailers, deflate;q=0.5
2097   The presence of the keyword "trailers" indicates that the client is
2098   willing to accept trailer fields in a chunked transfer coding,
2099   as defined in <xref target="chunked.encoding"/>, on behalf of itself and
2100   any downstream clients. For chained requests, this implies that either:
2101   (a) all downstream clients are willing to accept trailer fields in the
2102   forwarded response; or,
2103   (b) the client will attempt to buffer the response on behalf of downstream
2104   recipients.
2105   Note that HTTP/1.1 does not define any means to limit the size of a
2106   chunked response such that a client can be assured of buffering the
2107   entire response.
2110   When multiple transfer codings are acceptable, the client &MAY; rank the
2111   codings by preference using a case-insensitive "q" parameter (similar to
2112   the qvalues used in content negotiation fields, &qvalue;). The rank value
2113   is a real number in the range 0 through 1, where 0.001 is the least
2114   preferred and 1 is the most preferred; a value of 0 means "not acceptable".
2117   If the TE field-value is empty or if no TE field is present, the only
2118   acceptable transfer coding is chunked. A message with no transfer coding
2119   is always acceptable.
2122   Since the TE header field only applies to the immediate connection,
2123   a sender of TE &MUST; also send a "TE" connection option within the
2124   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
2125   in order to prevent the TE field from being forwarded by intermediaries
2126   that do not support its semantics.
2131<section title="Message Routing" anchor="message.routing">
2133   HTTP request message routing is determined by each client based on the
2134   target resource, the client's proxy configuration, and
2135   establishment or reuse of an inbound connection.  The corresponding
2136   response routing follows the same connection chain back to the client.
2139<section title="Identifying a Target Resource" anchor="target-resource">
2140  <iref primary="true" item="target resource"/>
2141  <iref primary="true" item="target URI"/>
2142  <x:anchor-alias value="target resource"/>
2143  <x:anchor-alias value="target URI"/>
2145   HTTP is used in a wide variety of applications, ranging from
2146   general-purpose computers to home appliances.  In some cases,
2147   communication options are hard-coded in a client's configuration.
2148   However, most HTTP clients rely on the same resource identification
2149   mechanism and configuration techniques as general-purpose Web browsers.
2152   HTTP communication is initiated by a user agent for some purpose.
2153   The purpose is a combination of request semantics, which are defined in
2154   <xref target="Part2"/>, and a target resource upon which to apply those
2155   semantics.  A URI reference (<xref target="uri"/>) is typically used as
2156   an identifier for the "<x:dfn>target resource</x:dfn>", which a user agent
2157   would resolve to its absolute form in order to obtain the
2158   "<x:dfn>target URI</x:dfn>".  The target URI
2159   excludes the reference's fragment identifier component, if any,
2160   since fragment identifiers are reserved for client-side processing
2161   (<xref target="RFC3986" x:fmt="," x:sec="3.5"/>).
2165<section title="Connecting Inbound" anchor="connecting.inbound">
2167   Once the target URI is determined, a client needs to decide whether
2168   a network request is necessary to accomplish the desired semantics and,
2169   if so, where that request is to be directed.
2172   If the client has a response cache and the request semantics can be
2173   satisfied by a cache (<xref target="Part6"/>), then the request is
2174   usually directed to the cache first.
2177   If the request is not satisfied by a cache, then a typical client will
2178   check its configuration to determine whether a proxy is to be used to
2179   satisfy the request.  Proxy configuration is implementation-dependent,
2180   but is often based on URI prefix matching, selective authority matching,
2181   or both, and the proxy itself is usually identified by an "http" or
2182   "https" URI.  If a proxy is applicable, the client connects inbound by
2183   establishing (or reusing) a connection to that proxy.
2186   If no proxy is applicable, a typical client will invoke a handler routine,
2187   usually specific to the target URI's scheme, to connect directly
2188   to an authority for the target resource.  How that is accomplished is
2189   dependent on the target URI scheme and defined by its associated
2190   specification, similar to how this specification defines origin server
2191   access for resolution of the "http" (<xref target="http.uri"/>) and
2192   "https" (<xref target="https.uri"/>) schemes.
2195   HTTP requirements regarding connection management are defined in
2196   <xref target=""/>.
2200<section title="Request Target" anchor="request-target">
2202   Once an inbound connection is obtained,
2203   the client sends an HTTP request message (<xref target="http.message"/>)
2204   with a request-target derived from the target URI.
2205   There are four distinct formats for the request-target, depending on both
2206   the method being requested and whether the request is to a proxy.
2208<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"/>
2209  <x:ref>request-target</x:ref> = <x:ref>origin-form</x:ref>
2210                 / <x:ref>absolute-form</x:ref>
2211                 / <x:ref>authority-form</x:ref>
2212                 / <x:ref>asterisk-form</x:ref>
2214  <x:ref>origin-form</x:ref>    = <x:ref>path-absolute</x:ref> [ "?" <x:ref>query</x:ref> ]
2215  <x:ref>absolute-form</x:ref>  = <x:ref>absolute-URI</x:ref>
2216  <x:ref>authority-form</x:ref> = <x:ref>authority</x:ref>
2217  <x:ref>asterisk-form</x:ref>  = "*"
2219<t anchor="origin-form"><iref item="origin-form (of request-target)"/>
2220   The most common form of request-target is the origin-form.
2221   When making a request directly to an origin server, other than a CONNECT
2222   or server-wide OPTIONS request (as detailed below),
2223   a client &MUST; send only the absolute path and query components of
2224   the target URI as the request-target.
2225   If the target URI's path component is empty, then the client &MUST; send
2226   "/" as the path within the origin-form of request-target.
2227   A <x:ref>Host</x:ref> header field is also sent, as defined in
2228   <xref target=""/>, containing the target URI's
2229   authority component (excluding any userinfo).
2232   For example, a client wishing to retrieve a representation of the resource
2233   identified as
2235<figure><artwork x:indent-with="  " type="example">
2239   directly from the origin server would open (or reuse) a TCP connection
2240   to port 80 of the host "" and send the lines:
2242<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2243GET /where?q=now HTTP/1.1
2247   followed by the remainder of the request message.
2249<t anchor="absolute-form"><iref item="absolute-form (of request-target)"/>
2250   When making a request to a proxy, other than a CONNECT or server-wide
2251   OPTIONS request (as detailed below), a client &MUST; send the target URI
2252   in absolute-form as the request-target.
2253   The proxy is requested to either service that request from a valid cache,
2254   if possible, or make the same request on the client's behalf to either
2255   the next inbound proxy server or directly to the origin server indicated
2256   by the request-target.  Requirements on such "forwarding" of messages are
2257   defined in <xref target="message.forwarding"/>.
2260   An example absolute-form of request-line would be:
2262<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2263GET HTTP/1.1
2266   To allow for transition to the absolute-form for all requests in some
2267   future version of HTTP, HTTP/1.1 servers &MUST; accept the absolute-form
2268   in requests, even though HTTP/1.1 clients will only send them in requests
2269   to proxies.
2271<t anchor="authority-form"><iref item="authority-form (of request-target)"/>
2272   The authority-form of request-target is only used for CONNECT requests
2273   (&CONNECT;).  When making a CONNECT request to establish a tunnel through
2274   one or more proxies, a client &MUST; send only the target URI's
2275   authority component (excluding any userinfo) as the request-target.
2276   For example,
2278<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2281<t anchor="asterisk-form"><iref item="asterisk-form (of request-target)"/>
2282   The asterisk-form of request-target is only used for a server-wide
2283   OPTIONS request (&OPTIONS;).  When a client wishes to request OPTIONS
2284   for the server as a whole, as opposed to a specific named resource of
2285   that server, the client &MUST; send only "*" (%x2A) as the request-target.
2286   For example,
2288<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2289OPTIONS * HTTP/1.1
2292   If a proxy receives an OPTIONS request with an absolute-form of
2293   request-target in which the URI has an empty path and no query component,
2294   then the last proxy on the request chain &MUST; send a request-target
2295   of "*" when it forwards the request to the indicated origin server.
2298   For example, the request
2299</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2303  would be forwarded by the final proxy as
2304</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2305OPTIONS * HTTP/1.1
2309   after connecting to port 8001 of host "".
2314<section title="Host" anchor="">
2315  <iref primary="true" item="Host header field" x:for-anchor=""/>
2316  <x:anchor-alias value="Host"/>
2318   The "Host" header field in a request provides the host and port
2319   information from the target URI, enabling the origin
2320   server to distinguish among resources while servicing requests
2321   for multiple host names on a single IP address.  Since the Host
2322   field-value is critical information for handling a request, it
2323   &SHOULD; be sent as the first header field following the request-line.
2325<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2326  <x:ref>Host</x:ref> = <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ; <xref target="http.uri"/>
2329   A client &MUST; send a Host header field in all HTTP/1.1 request
2330   messages.  If the target URI includes an authority component, then
2331   the Host field-value &MUST; be identical to that authority component
2332   after excluding any userinfo (<xref target="http.uri"/>).
2333   If the authority component is missing or undefined for the target URI,
2334   then the Host header field &MUST; be sent with an empty field-value.
2337   For example, a GET request to the origin server for
2338   &lt;; would begin with:
2340<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2341GET /pub/WWW/ HTTP/1.1
2345   The Host header field &MUST; be sent in an HTTP/1.1 request even
2346   if the request-target is in the absolute-form, since this
2347   allows the Host information to be forwarded through ancient HTTP/1.0
2348   proxies that might not have implemented Host.
2351   When a proxy receives a request with an absolute-form of
2352   request-target, the proxy &MUST; ignore the received
2353   Host header field (if any) and instead replace it with the host
2354   information of the request-target.  If the proxy forwards the request,
2355   it &MUST; generate a new Host field-value based on the received
2356   request-target rather than forward the received Host field-value.
2359   Since the Host header field acts as an application-level routing
2360   mechanism, it is a frequent target for malware seeking to poison
2361   a shared cache or redirect a request to an unintended server.
2362   An interception proxy is particularly vulnerable if it relies on
2363   the Host field-value for redirecting requests to internal
2364   servers, or for use as a cache key in a shared cache, without
2365   first verifying that the intercepted connection is targeting a
2366   valid IP address for that host.
2369   A server &MUST; respond with a <x:ref>400 (Bad Request)</x:ref> status code
2370   to any HTTP/1.1 request message that lacks a Host header field and
2371   to any request message that contains more than one Host header field
2372   or a Host header field with an invalid field-value.
2376<section title="Effective Request URI" anchor="effective.request.uri">
2377  <iref primary="true" item="effective request URI"/>
2378  <x:anchor-alias value="effective request URI"/>
2380   A server that receives an HTTP request message &MUST; reconstruct
2381   the user agent's original target URI, based on the pieces of information
2382   learned from the request-target, <x:ref>Host</x:ref> header field, and
2383   connection context, in order to identify the intended target resource and
2384   properly service the request. The URI derived from this reconstruction
2385   process is referred to as the "<x:dfn>effective request URI</x:dfn>".
2388   For a user agent, the effective request URI is the target URI.
2391   If the request-target is in absolute-form, then the effective request URI
2392   is the same as the request-target.  Otherwise, the effective request URI
2393   is constructed as follows.
2396   If the request is received over a TLS-secured TCP connection,
2397   then the effective request URI's scheme is "https"; otherwise, the
2398   scheme is "http".
2401   If the request-target is in authority-form, then the effective
2402   request URI's authority component is the same as the request-target.
2403   Otherwise, if a <x:ref>Host</x:ref> header field is supplied with a
2404   non-empty field-value, then the authority component is the same as the
2405   Host field-value. Otherwise, the authority component is the concatenation of
2406   the default host name configured for the server, a colon (":"), and the
2407   connection's incoming TCP port number in decimal form.
2410   If the request-target is in authority-form or asterisk-form, then the
2411   effective request URI's combined path and query component is empty.
2412   Otherwise, the combined path and query component is the same as the
2413   request-target.
2416   The components of the effective request URI, once determined as above,
2417   can be combined into absolute-URI form by concatenating the scheme,
2418   "://", authority, and combined path and query component.
2422   Example 1: the following message received over an insecure TCP connection
2424<artwork type="example" x:indent-with="  ">
2425GET /pub/WWW/TheProject.html HTTP/1.1
2431  has an effective request URI of
2433<artwork type="example" x:indent-with="  ">
2439   Example 2: the following message received over a TLS-secured TCP connection
2441<artwork type="example" x:indent-with="  ">
2442OPTIONS * HTTP/1.1
2448  has an effective request URI of
2450<artwork type="example" x:indent-with="  ">
2455   An origin server that does not allow resources to differ by requested
2456   host &MAY; ignore the <x:ref>Host</x:ref> field-value and instead replace it
2457   with a configured server name when constructing the effective request URI.
2460   Recipients of an HTTP/1.0 request that lacks a <x:ref>Host</x:ref> header
2461   field &MAY; attempt to use heuristics (e.g., examination of the URI path for
2462   something unique to a particular host) in order to guess the
2463   effective request URI's authority component.
2467<section title="Associating a Response to a Request" anchor="">
2469   HTTP does not include a request identifier for associating a given
2470   request message with its corresponding one or more response messages.
2471   Hence, it relies on the order of response arrival to correspond exactly
2472   to the order in which requests are made on the same connection.
2473   More than one response message per request only occurs when one or more
2474   informational responses (<x:ref>1xx</x:ref>, see &status-1xx;) precede a
2475   final response to the same request.
2478   A client that has more than one outstanding request on a connection &MUST;
2479   maintain a list of outstanding requests in the order sent and &MUST;
2480   associate each received response message on that connection to the highest
2481   ordered request that has not yet received a final (non-<x:ref>1xx</x:ref>)
2482   response.
2486<section title="Message Forwarding" anchor="message.forwarding">
2488   As described in <xref target="intermediaries"/>, intermediaries can serve
2489   a variety of roles in the processing of HTTP requests and responses.
2490   Some intermediaries are used to improve performance or availability.
2491   Others are used for access control or to filter content.
2492   Since an HTTP stream has characteristics similar to a pipe-and-filter
2493   architecture, there are no inherent limits to the extent an intermediary
2494   can enhance (or interfere) with either direction of the stream.
2497   Intermediaries that forward a message &MUST; implement the
2498   <x:ref>Connection</x:ref> header field, as specified in
2499   <xref target="header.connection"/>, to exclude fields that are only
2500   intended for the incoming connection.
2503   In order to avoid request loops, a proxy that forwards requests to other
2504   proxies &MUST; be able to recognize and exclude all of its own server
2505   names, including any aliases, local variations, or literal IP addresses.
2508<section title="Via" anchor="header.via">
2509  <iref primary="true" item="Via header field" x:for-anchor=""/>
2510  <x:anchor-alias value="pseudonym"/>
2511  <x:anchor-alias value="received-by"/>
2512  <x:anchor-alias value="received-protocol"/>
2513  <x:anchor-alias value="Via"/>
2515   The "Via" header field &MUST; be sent by a proxy or gateway in forwarded
2516   messages to indicate the intermediate protocols and recipients between the
2517   user agent and the server on requests, and between the origin server and
2518   the client on responses. It is analogous to the "Received" field
2519   used by email systems (<xref target="RFC5322" x:fmt="of" x:sec="3.6.7"/>).
2520   Via is used in HTTP for tracking message forwards,
2521   avoiding request loops, and identifying the protocol capabilities of
2522   all senders along the request/response chain.
2524<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"/>
2525  <x:ref>Via</x:ref>               = 1#( <x:ref>received-protocol</x:ref> <x:ref>RWS</x:ref> <x:ref>received-by</x:ref>
2526                          [ <x:ref>RWS</x:ref> <x:ref>comment</x:ref> ] )
2527  <x:ref>received-protocol</x:ref> = [ <x:ref>protocol-name</x:ref> "/" ] <x:ref>protocol-version</x:ref>
2528                      ; see <xref target="header.upgrade"/>
2529  <x:ref>received-by</x:ref>       = ( <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ) / <x:ref>pseudonym</x:ref>
2530  <x:ref>pseudonym</x:ref>         = <x:ref>token</x:ref>
2533   The received-protocol indicates the protocol version of the message
2534   received by the server or client along each segment of the
2535   request/response chain. The received-protocol version is appended to
2536   the Via field value when the message is forwarded so that information
2537   about the protocol capabilities of upstream applications remains
2538   visible to all recipients.
2541   The protocol-name is excluded if and only if it would be "HTTP". The
2542   received-by field is normally the host and optional port number of a
2543   recipient server or client that subsequently forwarded the message.
2544   However, if the real host is considered to be sensitive information,
2545   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
2546   be assumed to be the default port of the received-protocol.
2549   Multiple Via field values represent each proxy or gateway that has
2550   forwarded the message. Each recipient &MUST; append its information
2551   such that the end result is ordered according to the sequence of
2552   forwarding applications.
2555   Comments &MAY; be used in the Via header field to identify the software
2556   of each recipient, analogous to the <x:ref>User-Agent</x:ref> and
2557   <x:ref>Server</x:ref> header fields. However, all comments in the Via field
2558   are optional and &MAY; be removed by any recipient prior to forwarding the
2559   message.
2562   For example, a request message could be sent from an HTTP/1.0 user
2563   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2564   forward the request to a public proxy at, which completes
2565   the request by forwarding it to the origin server at
2566   The request received by would then have the following
2567   Via header field:
2569<figure><artwork type="example">
2570  Via: 1.0 fred, 1.1 (Apache/1.1)
2573   A proxy or gateway used as a portal through a network firewall
2574   &SHOULD-NOT; forward the names and ports of hosts within the firewall
2575   region unless it is explicitly enabled to do so. If not enabled, the
2576   received-by host of any host behind the firewall &SHOULD; be replaced
2577   by an appropriate pseudonym for that host.
2580   A proxy or gateway &MAY; combine an ordered subsequence of Via header
2581   field entries into a single such entry if the entries have identical
2582   received-protocol values. For example,
2584<figure><artwork type="example">
2585  Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2588  could be collapsed to
2590<figure><artwork type="example">
2591  Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2594   Senders &SHOULD-NOT; combine multiple entries unless they are all
2595   under the same organizational control and the hosts have already been
2596   replaced by pseudonyms. Senders &MUST-NOT; combine entries that
2597   have different received-protocol values.
2601<section title="Transformations" anchor="message.transformations">
2603   Some intermediaries include features for transforming messages and their
2604   payloads.  A transforming proxy might, for example, convert between image
2605   formats in order to save cache space or to reduce the amount of traffic on
2606   a slow link. However, operational problems might occur when these
2607   transformations are applied to payloads intended for critical applications,
2608   such as medical imaging or scientific data analysis, particularly when
2609   integrity checks or digital signatures are used to ensure that the payload
2610   received is identical to the original.
2613   If a proxy receives a request-target with a host name that is not a
2614   fully qualified domain name, it &MAY; add its own domain to the host name
2615   it received when forwarding the request.  A proxy &MUST-NOT; change the
2616   host name if it is a fully qualified domain name.
2619   A proxy &MUST-NOT; modify the "path-absolute" and "query" parts of the
2620   received request-target when forwarding it to the next inbound server,
2621   except as noted above to replace an empty path with "/" or "*".
2624   A proxy &MUST-NOT; modify header fields that provide information about the
2625   end points of the communication chain, the resource state, or the selected
2626   representation. A proxy &MAY; change the message body through application
2627   or removal of a transfer coding (<xref target="transfer.codings"/>).
2630   A non-transforming proxy &MUST; preserve the message payload (&payload;).
2631   A transforming proxy &MUST; preserve the payload of a message that
2632   contains the no-transform cache-control directive.
2635   A transforming proxy &MAY; transform the payload of a message
2636   that does not contain the no-transform cache-control directive;
2637   if the payload is transformed, the transforming proxy &MUST; add a
2638   Warning 214 (Transformation applied) header field if one does not
2639   already appear in the message (see &header-warning;).
2645<section title="Connection Management" anchor="">
2647   HTTP messaging is independent of the underlying transport or
2648   session-layer connection protocol(s).  HTTP only presumes a reliable
2649   transport with in-order delivery of requests and the corresponding
2650   in-order delivery of responses.  The mapping of HTTP request and
2651   response structures onto the data units of an underlying transport
2652   protocol is outside the scope of this specification.
2655   As described in <xref target="connecting.inbound"/>, the specific
2656   connection protocols to be used for an HTTP interaction are determined by
2657   client configuration and the <x:ref>target URI</x:ref>.
2658   For example, the "http" URI scheme
2659   (<xref target="http.uri"/>) indicates a default connection of TCP
2660   over IP, with a default TCP port of 80, but the client might be
2661   configured to use a proxy via some other connection, port, or protocol.
2664   HTTP implementations are expected to engage in connection management,
2665   which includes maintaining the state of current connections,
2666   establishing a new connection or reusing an existing connection,
2667   processing messages received on a connection, detecting connection
2668   failures, and closing each connection.
2669   Most clients maintain multiple connections in parallel, including
2670   more than one connection per server endpoint.
2671   Most servers are designed to maintain thousands of concurrent connections,
2672   while controlling request queues to enable fair use and detect
2673   denial of service attacks.
2676<section title="Connection" anchor="header.connection">
2677  <iref primary="true" item="Connection header field" x:for-anchor=""/>
2678  <iref primary="true" item="close" x:for-anchor=""/>
2679  <x:anchor-alias value="Connection"/>
2680  <x:anchor-alias value="connection-option"/>
2681  <x:anchor-alias value="close"/>
2683   The "Connection" header field allows the sender to indicate desired
2684   control options for the current connection.  In order to avoid confusing
2685   downstream recipients, a proxy or gateway &MUST; remove or replace any
2686   received connection options before forwarding the message.
2689   When a header field aside from Connection is used to supply control
2690   information for or about the current connection, the sender &MUST; list
2691   the corresponding field-name within the "Connection" header field.
2692   A proxy or gateway &MUST; parse a received Connection
2693   header field before a message is forwarded and, for each
2694   connection-option in this field, remove any header field(s) from
2695   the message with the same name as the connection-option, and then
2696   remove the Connection header field itself (or replace it with the
2697   intermediary's own connection options for the forwarded message).
2700   Hence, the Connection header field provides a declarative way of
2701   distinguishing header fields that are only intended for the
2702   immediate recipient ("hop-by-hop") from those fields that are
2703   intended for all recipients on the chain ("end-to-end"), enabling the
2704   message to be self-descriptive and allowing future connection-specific
2705   extensions to be deployed without fear that they will be blindly
2706   forwarded by older intermediaries.
2709   The Connection header field's value has the following grammar:
2711<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-option"/>
2712  <x:ref>Connection</x:ref>        = 1#<x:ref>connection-option</x:ref>
2713  <x:ref>connection-option</x:ref> = <x:ref>token</x:ref>
2716   Connection options are case-insensitive.
2719   A sender &MUST-NOT; send a connection option corresponding to a header
2720   field that is intended for all recipients of the payload.
2721   For example, <x:ref>Cache-Control</x:ref> is never appropriate as a
2722   connection option (&header-cache-control;).
2725   The connection options do not have to correspond to a header field
2726   present in the message, since a connection-specific header field
2727   might not be needed if there are no parameters associated with that
2728   connection option.  Recipients that trigger certain connection
2729   behavior based on the presence of connection options &MUST; do so
2730   based on the presence of the connection-option rather than only the
2731   presence of the optional header field.  In other words, if the
2732   connection option is received as a header field but not indicated
2733   within the Connection field-value, then the recipient &MUST; ignore
2734   the connection-specific header field because it has likely been
2735   forwarded by an intermediary that is only partially conformant.
2738   When defining new connection options, specifications ought to
2739   carefully consider existing deployed header fields and ensure
2740   that the new connection option does not share the same name as
2741   an unrelated header field that might already be deployed.
2742   Defining a new connection option essentially reserves that potential
2743   field-name for carrying additional information related to the
2744   connection option, since it would be unwise for senders to use
2745   that field-name for anything else.
2748   The "<x:dfn>close</x:dfn>" connection option is defined for a
2749   sender to signal that this connection will be closed after completion of
2750   the response. For example,
2752<figure><artwork type="example">
2753  Connection: close
2756   in either the request or the response header fields indicates that
2757   the connection &MUST; be closed after the current request/response
2758   is complete (<xref target="persistent.tear-down"/>).
2761   A client that does not support <x:ref>persistent connections</x:ref> &MUST;
2762   send the "close" connection option in every request message.
2765   A server that does not support <x:ref>persistent connections</x:ref> &MUST;
2766   send the "close" connection option in every response message that
2767   does not have a <x:ref>1xx (Informational)</x:ref> status code.
2771<section title="Establishment" anchor="persistent.establishment">
2773   It is beyond the scope of this specification to describe how connections
2774   are established via various transport or session-layer protocols.
2775   Each connection applies to only one transport link.
2779<section title="Persistence" anchor="persistent.connections">
2780   <x:anchor-alias value="persistent connections"/>
2782   HTTP/1.1 defaults to the use of "<x:dfn>persistent connections</x:dfn>",
2783   allowing multiple requests and responses to be carried over a single
2784   connection. The "<x:ref>close</x:ref>" connection-option is used to signal
2785   that a connection will not persist after the current request/response.
2786   HTTP implementations &SHOULD; support persistent connections.
2789   A recipient determines whether a connection is persistent or not based on
2790   the most recently received message's protocol version and
2791   <x:ref>Connection</x:ref> header field (if any):
2792   <list style="symbols">
2793     <t>If the <x:ref>close</x:ref> connection option is present, the
2794        connection will not persist after the current response; else,</t>
2795     <t>If the received protocol is HTTP/1.1 (or later), the connection will
2796        persist after the current response; else,</t>
2797     <t>If the received protocol is HTTP/1.0, the "keep-alive"
2798        connection option is present, the recipient is not a proxy, and
2799        the recipient wishes to honor the HTTP/1.0 "keep-alive" mechanism,
2800        the connection will persist after the current response; otherwise,</t>
2801     <t>The connection will close after the current response.</t>
2802   </list>
2805   A server &MAY; assume that an HTTP/1.1 client intends to maintain a
2806   persistent connection until a <x:ref>close</x:ref> connection option
2807   is received in a request.
2810   A client &MAY; reuse a persistent connection until it sends or receives
2811   a <x:ref>close</x:ref> connection option or receives an HTTP/1.0 response
2812   without a "keep-alive" connection option.
2815   In order to remain persistent, all messages on a connection &MUST;
2816   have a self-defined message length (i.e., one not defined by closure
2817   of the connection), as described in <xref target="message.body"/>.
2818   A server &MUST; read the entire request message body or close
2819   the connection after sending its response, since otherwise the
2820   remaining data on a persistent connection would be misinterpreted
2821   as the next request.  Likewise,
2822   a client &MUST; read the entire response message body if it intends
2823   to reuse the same connection for a subsequent request.
2826   A proxy server &MUST-NOT; maintain a persistent connection with an
2827   HTTP/1.0 client (see <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/> for
2828   information and discussion of the problems with the Keep-Alive header field
2829   implemented by many HTTP/1.0 clients).
2832   Clients and servers &SHOULD-NOT; assume that a persistent connection is
2833   maintained for HTTP versions less than 1.1 unless it is explicitly
2834   signaled.
2835   See <xref target="compatibility.with.http.1.0.persistent.connections"/>
2836   for more information on backward compatibility with HTTP/1.0 clients.
2839<section title="Pipelining" anchor="pipelining">
2841   A client that supports persistent connections &MAY; "pipeline" its
2842   requests (i.e., send multiple requests without waiting for each
2843   response). A server &MUST; send its responses to those requests in the
2844   same order that the requests were received.
2847   Clients that assume persistent connections and pipeline immediately
2848   after connection establishment &SHOULD; be prepared to retry their
2849   connection if the first pipelined attempt fails. If a client does
2850   such a retry, it &MUST-NOT; pipeline before it knows the connection is
2851   persistent. Clients &MUST; also be prepared to resend their requests if
2852   the server closes the connection before sending all of the
2853   corresponding responses.
2856   Clients &SHOULD-NOT; pipeline requests using non-idempotent request methods
2857   or non-idempotent sequences of request methods (see &idempotent-methods;).
2858   Otherwise, a premature termination of the transport connection could lead
2859   to indeterminate results. A client wishing to send a non-idempotent
2860   request &SHOULD; wait to send that request until it has received the
2861   response status line for the previous request.
2865<section title="Retrying Requests" anchor="persistent.retrying.requests">
2867   Connections can be closed at any time, with or without intention.
2868   Implementations ought to anticipate the need to recover
2869   from asynchronous close events.
2870   A client &MAY; open a new connection and retransmit an aborted sequence
2871   of requests without user interaction so long as the request sequence is
2872   idempotent (see &idempotent-methods;).
2873   A client &MUST-NOT; automatically retry non-idempotent request sequences,
2874   although user agents &MAY; offer a human operator the choice of retrying
2875   the request(s). Confirmation by
2876   user agent software with semantic understanding of the application
2877   &MAY; substitute for user confirmation. An automatic retry &SHOULD-NOT;
2878   be repeated if a second sequence of requests fails.
2883<section title="Concurrency" anchor="persistent.concurrency">
2885   Clients &SHOULD; limit the number of simultaneous
2886   connections that they maintain to a given server.
2889   Previous revisions of HTTP gave a specific number of connections as a
2890   ceiling, but this was found to be impractical for many applications. As a
2891   result, this specification does not mandate a particular maximum number of
2892   connections, but instead encourages clients to be conservative when opening
2893   multiple connections.
2896   Multiple connections are typically used to avoid the "head-of-line
2897   blocking" problem, wherein a request that takes significant server-side
2898   processing and/or has a large payload blocks subsequent requests on the
2899   same connection. However, each connection consumes server resources.
2900   Furthermore, using multiple connections can cause undesirable side effects
2901   in congested networks.
2904   Note that servers might reject traffic that they deem abusive, including an
2905   excessive number of connections from a client.
2909<section title="Failures and Time-outs" anchor="persistent.failures">
2911   Servers will usually have some time-out value beyond which they will
2912   no longer maintain an inactive connection. Proxy servers might make
2913   this a higher value since it is likely that the client will be making
2914   more connections through the same server. The use of persistent
2915   connections places no requirements on the length (or existence) of
2916   this time-out for either the client or the server.
2919   When a client or server wishes to time-out it &SHOULD; issue a graceful
2920   close on the transport connection. Clients and servers &SHOULD; both
2921   constantly watch for the other side of the transport close, and
2922   respond to it as appropriate. If a client or server does not detect
2923   the other side's close promptly it could cause unnecessary resource
2924   drain on the network.
2927   A client, server, or proxy &MAY; close the transport connection at any
2928   time. For example, a client might have started to send a new request
2929   at the same time that the server has decided to close the "idle"
2930   connection. From the server's point of view, the connection is being
2931   closed while it was idle, but from the client's point of view, a
2932   request is in progress.
2935   Servers &SHOULD; maintain persistent connections and allow the underlying
2936   transport's flow control mechanisms to resolve temporary overloads, rather
2937   than terminate connections with the expectation that clients will retry.
2938   The latter technique can exacerbate network congestion.
2941   A client sending a message body &SHOULD; monitor
2942   the network connection for an error status code while it is transmitting
2943   the request. If the client sees an error status code, it &SHOULD;
2944   immediately cease transmitting the body and close the connection.
2948<section title="Tear-down" anchor="persistent.tear-down">
2949  <iref primary="false" item="Connection header field" x:for-anchor=""/>
2950  <iref primary="false" item="close" x:for-anchor=""/>
2952   The <x:ref>Connection</x:ref> header field
2953   (<xref target="header.connection"/>) provides a "<x:ref>close</x:ref>"
2954   connection option that a sender &SHOULD; send when it wishes to close
2955   the connection after the current request/response pair.
2958   A client that sends a <x:ref>close</x:ref> connection option &MUST-NOT;
2959   send further requests on that connection (after the one containing
2960   <x:ref>close</x:ref>) and &MUST; close the connection after reading the
2961   final response message corresponding to this request.
2964   A server that receives a <x:ref>close</x:ref> connection option &MUST;
2965   initiate a lingering close (see below) of the connection after it sends the
2966   final response to the request that contained <x:ref>close</x:ref>.
2967   The server &SHOULD; send a <x:ref>close</x:ref> connection option
2968   in its final response on that connection. The server &MUST-NOT; process
2969   any further requests received on that connection.
2972   A server that sends a <x:ref>close</x:ref> connection option &MUST;
2973   initiate a lingering close of the connection after it sends the
2974   response containing <x:ref>close</x:ref>. The server &MUST-NOT; process
2975   any further requests received on that connection.
2978   A client that receives a <x:ref>close</x:ref> connection option &MUST;
2979   cease sending requests on that connection and close the connection
2980   after reading the response message containing the close; if additional
2981   pipelined requests had been sent on the connection, the client &SHOULD;
2982   assume that they will not be processed by the server.
2985   If a server performs an immediate close of a TCP connection, there is a
2986   significant risk that the client will not be able to read the last HTTP
2987   response.  If the server receives additional data from the client on a
2988   fully-closed connection, such as another request that was sent by the
2989   client before receiving the server's response, the server's TCP stack will
2990   send a reset packet to the client; unfortunately, the reset packet might
2991   erase the client's unacknowledged input buffers before they can be read
2992   and interpreted by the client's HTTP parser.
2995   To avoid the TCP reset problem, a server can perform a lingering close on a
2996   connection by closing only the write side of the read/write connection
2997   (a half-close) and continuing to read from the connection until the
2998   connection is closed by the client or the server is reasonably certain
2999   that its own TCP stack has received the client's acknowledgement of the
3000   packet(s) containing the server's last response. It is then safe for the
3001   server to fully close the connection.
3004   It is unknown whether the reset problem is exclusive to TCP or might also
3005   be found in other transport connection protocols.
3009<section title="Upgrade" anchor="header.upgrade">
3010  <iref primary="true" item="Upgrade header field" x:for-anchor=""/>
3011  <x:anchor-alias value="Upgrade"/>
3012  <x:anchor-alias value="protocol"/>
3013  <x:anchor-alias value="protocol-name"/>
3014  <x:anchor-alias value="protocol-version"/>
3016   The "Upgrade" header field is intended to provide a simple mechanism
3017   for transitioning from HTTP/1.1 to some other protocol on the same
3018   connection.  A client &MAY; send a list of protocols in the Upgrade
3019   header field of a request to invite the server to switch to one or
3020   more of those protocols before sending the final response.
3021   A server &MUST; send an Upgrade header field in <x:ref>101 (Switching
3022   Protocols)</x:ref> responses to indicate which protocol(s) are being
3023   switched to, and &MUST; send it in <x:ref>426 (Upgrade Required)</x:ref>
3024   responses to indicate acceptable protocols.
3025   A server &MAY; send an Upgrade header field in any other response to
3026   indicate that they might be willing to upgrade to one of the
3027   specified protocols for a future request.
3029<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
3030  <x:ref>Upgrade</x:ref>          = 1#<x:ref>protocol</x:ref>
3032  <x:ref>protocol</x:ref>         = <x:ref>protocol-name</x:ref> ["/" <x:ref>protocol-version</x:ref>]
3033  <x:ref>protocol-name</x:ref>    = <x:ref>token</x:ref>
3034  <x:ref>protocol-version</x:ref> = <x:ref>token</x:ref>
3037   For example,
3039<figure><artwork type="example">
3040  Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
3043   Upgrade eases the difficult transition between incompatible protocols by
3044   allowing the client to initiate a request in the more commonly
3045   supported protocol while indicating to the server that it would like
3046   to use a "better" protocol if available (where "better" is determined
3047   by the server, possibly according to the nature of the request method
3048   or target resource).
3051   Upgrade cannot be used to insist on a protocol change; its acceptance and
3052   use by the server is optional. The capabilities and nature of the
3053   application-level communication after the protocol change is entirely
3054   dependent upon the new protocol chosen, although the first action
3055   after changing the protocol &MUST; be a response to the initial HTTP
3056   request that contained the Upgrade header field.
3059   For example, if the Upgrade header field is received in a GET request
3060   and the server decides to switch protocols, then it first responds
3061   with a <x:ref>101 (Switching Protocols)</x:ref> message in HTTP/1.1 and
3062   then immediately follows that with the new protocol's equivalent of a
3063   response to a GET on the target resource.  This allows a connection to be
3064   upgraded to protocols with the same semantics as HTTP without the
3065   latency cost of an additional round-trip.  A server &MUST-NOT; switch
3066   protocols unless the received message semantics can be honored by the new
3067   protocol; an OPTIONS request can be honored by any protocol.
3070   When Upgrade is sent, a sender &MUST; also send a
3071   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
3072   that contains the "upgrade" connection option, in order to prevent Upgrade
3073   from being accidentally forwarded by intermediaries that might not implement
3074   the listed protocols.  A server &MUST; ignore an Upgrade header field that
3075   is received in an HTTP/1.0 request.
3078   The Upgrade header field only applies to switching application-level
3079   protocols on the existing connection; it cannot be used
3080   to switch to a protocol on a different connection. For that purpose, it is
3081   more appropriate to use a <x:ref>3xx (Redirection)</x:ref> response
3082   (&status-3xx;).
3085   This specification only defines the protocol name "HTTP" for use by
3086   the family of Hypertext Transfer Protocols, as defined by the HTTP
3087   version rules of <xref target="http.version"/> and future updates to this
3088   specification. Additional tokens ought to be registered with IANA using the
3089   registration procedure defined in <xref target="upgrade.token.registry"/>.
3094<section title="IANA Considerations" anchor="IANA.considerations">
3096<section title="Header Field Registration" anchor="header.field.registration">
3098   HTTP header fields are registered within the Message Header Field Registry
3099   <xref target="BCP90"/> maintained by IANA at
3100   <eref target=""/>.
3103   This document defines the following HTTP header fields, so their
3104   associated registry entries shall be updated according to the permanent
3105   registrations below:
3107<?BEGININC p1-messaging.iana-headers ?>
3108<!--AUTOGENERATED FROM extract-header-defs.xslt, do not edit manually-->
3109<texttable align="left" suppress-title="true" anchor="iana.header.registration.table">
3110   <ttcol>Header Field Name</ttcol>
3111   <ttcol>Protocol</ttcol>
3112   <ttcol>Status</ttcol>
3113   <ttcol>Reference</ttcol>
3115   <c>Connection</c>
3116   <c>http</c>
3117   <c>standard</c>
3118   <c>
3119      <xref target="header.connection"/>
3120   </c>
3121   <c>Content-Length</c>
3122   <c>http</c>
3123   <c>standard</c>
3124   <c>
3125      <xref target="header.content-length"/>
3126   </c>
3127   <c>Host</c>
3128   <c>http</c>
3129   <c>standard</c>
3130   <c>
3131      <xref target=""/>
3132   </c>
3133   <c>TE</c>
3134   <c>http</c>
3135   <c>standard</c>
3136   <c>
3137      <xref target="header.te"/>
3138   </c>
3139   <c>Trailer</c>
3140   <c>http</c>
3141   <c>standard</c>
3142   <c>
3143      <xref target="header.trailer"/>
3144   </c>
3145   <c>Transfer-Encoding</c>
3146   <c>http</c>
3147   <c>standard</c>
3148   <c>
3149      <xref target="header.transfer-encoding"/>
3150   </c>
3151   <c>Upgrade</c>
3152   <c>http</c>
3153   <c>standard</c>
3154   <c>
3155      <xref target="header.upgrade"/>
3156   </c>
3157   <c>Via</c>
3158   <c>http</c>
3159   <c>standard</c>
3160   <c>
3161      <xref target="header.via"/>
3162   </c>
3165<?ENDINC p1-messaging.iana-headers ?>
3167   Furthermore, the header field-name "Close" shall be registered as
3168   "reserved", since using that name as an HTTP header field might
3169   conflict with the "close" connection option of the "<x:ref>Connection</x:ref>"
3170   header field (<xref target="header.connection"/>).
3172<texttable align="left" suppress-title="true">
3173   <ttcol>Header Field Name</ttcol>
3174   <ttcol>Protocol</ttcol>
3175   <ttcol>Status</ttcol>
3176   <ttcol>Reference</ttcol>
3178   <c>Close</c>
3179   <c>http</c>
3180   <c>reserved</c>
3181   <c>
3182      <xref target="header.field.registration"/>
3183   </c>
3186   The change controller is: "IETF ( - Internet Engineering Task Force".
3190<section title="URI Scheme Registration" anchor="uri.scheme.registration">
3192   IANA maintains the registry of URI Schemes <xref target="BCP115"/> at
3193   <eref target=""/>.
3196   This document defines the following URI schemes, so their
3197   associated registry entries shall be updated according to the permanent
3198   registrations below:
3200<texttable align="left" suppress-title="true">
3201   <ttcol>URI Scheme</ttcol>
3202   <ttcol>Description</ttcol>
3203   <ttcol>Reference</ttcol>
3205   <c>http</c>
3206   <c>Hypertext Transfer Protocol</c>
3207   <c><xref target="http.uri"/></c>
3209   <c>https</c>
3210   <c>Hypertext Transfer Protocol Secure</c>
3211   <c><xref target="https.uri"/></c>
3215<section title="Internet Media Type Registration" anchor="">
3217   This document serves as the specification for the Internet media types
3218   "message/http" and "application/http". The following is to be registered with
3219   IANA (see <xref target="BCP13"/>).
3221<section title="Internet Media Type message/http" anchor="">
3222<iref item="Media Type" subitem="message/http" primary="true"/>
3223<iref item="message/http Media Type" primary="true"/>
3225   The message/http type can be used to enclose a single HTTP request or
3226   response message, provided that it obeys the MIME restrictions for all
3227   "message" types regarding line length and encodings.
3230  <list style="hanging" x:indent="12em">
3231    <t hangText="Type name:">
3232      message
3233    </t>
3234    <t hangText="Subtype name:">
3235      http
3236    </t>
3237    <t hangText="Required parameters:">
3238      none
3239    </t>
3240    <t hangText="Optional parameters:">
3241      version, msgtype
3242      <list style="hanging">
3243        <t hangText="version:">
3244          The HTTP-version number of the enclosed message
3245          (e.g., "1.1"). If not present, the version can be
3246          determined from the first line of the body.
3247        </t>
3248        <t hangText="msgtype:">
3249          The message type &mdash; "request" or "response". If not
3250          present, the type can be determined from the first
3251          line of the body.
3252        </t>
3253      </list>
3254    </t>
3255    <t hangText="Encoding considerations:">
3256      only "7bit", "8bit", or "binary" are permitted
3257    </t>
3258    <t hangText="Security considerations:">
3259      none
3260    </t>
3261    <t hangText="Interoperability considerations:">
3262      none
3263    </t>
3264    <t hangText="Published specification:">
3265      This specification (see <xref target=""/>).
3266    </t>
3267    <t hangText="Applications that use this media type:">
3268    </t>
3269    <t hangText="Additional information:">
3270      <list style="hanging">
3271        <t hangText="Magic number(s):">none</t>
3272        <t hangText="File extension(s):">none</t>
3273        <t hangText="Macintosh file type code(s):">none</t>
3274      </list>
3275    </t>
3276    <t hangText="Person and email address to contact for further information:">
3277      See Authors Section.
3278    </t>
3279    <t hangText="Intended usage:">
3280      COMMON
3281    </t>
3282    <t hangText="Restrictions on usage:">
3283      none
3284    </t>
3285    <t hangText="Author/Change controller:">
3286      IESG
3287    </t>
3288  </list>
3291<section title="Internet Media Type application/http" anchor="">
3292<iref item="Media Type" subitem="application/http" primary="true"/>
3293<iref item="application/http Media Type" primary="true"/>
3295   The application/http type can be used to enclose a pipeline of one or more
3296   HTTP request or response messages (not intermixed).
3299  <list style="hanging" x:indent="12em">
3300    <t hangText="Type name:">
3301      application
3302    </t>
3303    <t hangText="Subtype name:">
3304      http
3305    </t>
3306    <t hangText="Required parameters:">
3307      none
3308    </t>
3309    <t hangText="Optional parameters:">
3310      version, msgtype
3311      <list style="hanging">
3312        <t hangText="version:">
3313          The HTTP-version number of the enclosed messages
3314          (e.g., "1.1"). If not present, the version can be
3315          determined from the first line of the body.
3316        </t>
3317        <t hangText="msgtype:">
3318          The message type &mdash; "request" or "response". If not
3319          present, the type can be determined from the first
3320          line of the body.
3321        </t>
3322      </list>
3323    </t>
3324    <t hangText="Encoding considerations:">
3325      HTTP messages enclosed by this type
3326      are in "binary" format; use of an appropriate
3327      Content-Transfer-Encoding is required when
3328      transmitted via E-mail.
3329    </t>
3330    <t hangText="Security considerations:">
3331      none
3332    </t>
3333    <t hangText="Interoperability considerations:">
3334      none
3335    </t>
3336    <t hangText="Published specification:">
3337      This specification (see <xref target=""/>).
3338    </t>
3339    <t hangText="Applications that use this media type:">
3340    </t>
3341    <t hangText="Additional information:">
3342      <list style="hanging">
3343        <t hangText="Magic number(s):">none</t>
3344        <t hangText="File extension(s):">none</t>
3345        <t hangText="Macintosh file type code(s):">none</t>
3346      </list>
3347    </t>
3348    <t hangText="Person and email address to contact for further information:">
3349      See Authors Section.
3350    </t>
3351    <t hangText="Intended usage:">
3352      COMMON
3353    </t>
3354    <t hangText="Restrictions on usage:">
3355      none
3356    </t>
3357    <t hangText="Author/Change controller:">
3358      IESG
3359    </t>
3360  </list>
3365<section title="Transfer Coding Registry" anchor="transfer.coding.registry">
3367   The HTTP Transfer Coding Registry defines the name space for transfer
3368   coding names.
3371   Registrations &MUST; include the following fields:
3372   <list style="symbols">
3373     <t>Name</t>
3374     <t>Description</t>
3375     <t>Pointer to specification text</t>
3376   </list>
3379   Names of transfer codings &MUST-NOT; overlap with names of content codings
3380   (&content-codings;) unless the encoding transformation is identical, as
3381   is the case for the compression codings defined in
3382   <xref target="compression.codings"/>.
3385   Values to be added to this name space require IETF Review (see
3386   <xref target="RFC5226" x:fmt="of" x:sec="4.1"/>), and &MUST;
3387   conform to the purpose of transfer coding defined in this section.
3388   Use of program names for the identification of encoding formats
3389   is not desirable and is discouraged for future encodings.
3392   The registry itself is maintained at
3393   <eref target=""/>.
3397<section title="Transfer Coding Registration" anchor="transfer.coding.registration">
3399   The HTTP Transfer Coding Registry shall be updated with the registrations
3400   below:
3402<texttable align="left" suppress-title="true" anchor="iana.transfer.coding.registration.table">
3403   <ttcol>Name</ttcol>
3404   <ttcol>Description</ttcol>
3405   <ttcol>Reference</ttcol>
3406   <c>chunked</c>
3407   <c>Transfer in a series of chunks</c>
3408   <c>
3409      <xref target="chunked.encoding"/>
3410   </c>
3411   <c>compress</c>
3412   <c>UNIX "compress" program method</c>
3413   <c>
3414      <xref target="compress.coding"/>
3415   </c>
3416   <c>deflate</c>
3417   <c>"deflate" compression mechanism (<xref target="RFC1951"/>) used inside
3418   the "zlib" data format (<xref target="RFC1950"/>)
3419   </c>
3420   <c>
3421      <xref target="deflate.coding"/>
3422   </c>
3423   <c>gzip</c>
3424   <c>Same as GNU zip <xref target="RFC1952"/></c>
3425   <c>
3426      <xref target="gzip.coding"/>
3427   </c>
3431<section title="Upgrade Token Registry" anchor="upgrade.token.registry">
3433   The HTTP Upgrade Token Registry defines the name space for protocol-name
3434   tokens used to identify protocols in the <x:ref>Upgrade</x:ref> header
3435   field. Each registered protocol name is associated with contact information
3436   and an optional set of specifications that details how the connection
3437   will be processed after it has been upgraded.
3440   Registrations happen on a "First Come First Served" basis (see
3441   <xref target="RFC5226" x:sec="4.1" x:fmt="of"/>) and are subject to the
3442   following rules:
3443  <list style="numbers">
3444    <t>A protocol-name token, once registered, stays registered forever.</t>
3445    <t>The registration &MUST; name a responsible party for the
3446       registration.</t>
3447    <t>The registration &MUST; name a point of contact.</t>
3448    <t>The registration &MAY; name a set of specifications associated with
3449       that token. Such specifications need not be publicly available.</t>
3450    <t>The registration &SHOULD; name a set of expected "protocol-version"
3451       tokens associated with that token at the time of registration.</t>
3452    <t>The responsible party &MAY; change the registration at any time.
3453       The IANA will keep a record of all such changes, and make them
3454       available upon request.</t>
3455    <t>The IESG &MAY; reassign responsibility for a protocol token.
3456       This will normally only be used in the case when a
3457       responsible party cannot be contacted.</t>
3458  </list>
3461   This registration procedure for HTTP Upgrade Tokens replaces that
3462   previously defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
3466<section title="Upgrade Token Registration" anchor="upgrade.token.registration">
3468   The HTTP Upgrade Token Registry shall be updated with the registration
3469   below:
3471<texttable align="left" suppress-title="true">
3472   <ttcol>Value</ttcol>
3473   <ttcol>Description</ttcol>
3474   <ttcol>Expected Version Tokens</ttcol>
3475   <ttcol>Reference</ttcol>
3477   <c>HTTP</c>
3478   <c>Hypertext Transfer Protocol</c>
3479   <c>any DIGIT.DIGIT (e.g, "2.0")</c>
3480   <c><xref target="http.version"/></c>
3483   The responsible party is: "IETF ( - Internet Engineering Task Force".
3489<section title="Security Considerations" anchor="security.considerations">
3491   This section is meant to inform developers, information providers, and
3492   users of known security concerns relevant to HTTP/1.1 message syntax,
3493   parsing, and routing.
3496<section title="DNS-related Attacks" anchor="dns.related.attacks">
3498   HTTP clients rely heavily on the Domain Name Service (DNS), and are thus
3499   generally prone to security attacks based on the deliberate misassociation
3500   of IP addresses and DNS names not protected by DNSSEC. Clients need to be
3501   cautious in assuming the validity of an IP number/DNS name association unless
3502   the response is protected by DNSSEC (<xref target="RFC4033"/>).
3506<section title="Intermediaries and Caching" anchor="attack.intermediaries">
3508   By their very nature, HTTP intermediaries are men-in-the-middle, and
3509   represent an opportunity for man-in-the-middle attacks. Compromise of
3510   the systems on which the intermediaries run can result in serious security
3511   and privacy problems. Intermediaries have access to security-related
3512   information, personal information about individual users and
3513   organizations, and proprietary information belonging to users and
3514   content providers. A compromised intermediary, or an intermediary
3515   implemented or configured without regard to security and privacy
3516   considerations, might be used in the commission of a wide range of
3517   potential attacks.
3520   Intermediaries that contain a shared cache are especially vulnerable
3521   to cache poisoning attacks.
3524   Implementers need to consider the privacy and security
3525   implications of their design and coding decisions, and of the
3526   configuration options they provide to operators (especially the
3527   default configuration).
3530   Users need to be aware that intermediaries are no more trustworthy than
3531   the people who run them; HTTP itself cannot solve this problem.
3535<section title="Buffer Overflows" anchor="attack.protocol.element.size.overflows">
3537   Because HTTP uses mostly textual, character-delimited fields, attackers can
3538   overflow buffers in implementations, and/or perform a Denial of Service
3539   against implementations that accept fields with unlimited lengths.
3542   To promote interoperability, this specification makes specific
3543   recommendations for minimum size limits on request-line
3544   (<xref target="request.line"/>)
3545   and blocks of header fields (<xref target="header.fields"/>). These are
3546   minimum recommendations, chosen to be supportable even by implementations
3547   with limited resources; it is expected that most implementations will
3548   choose substantially higher limits.
3551   This specification also provides a way for servers to reject messages that
3552   have request-targets that are too long (&status-414;) or request entities
3553   that are too large (&status-4xx;).
3556   Recipients &SHOULD; carefully limit the extent to which they read other
3557   fields, including (but not limited to) request methods, response status
3558   phrases, header field-names, and body chunks, so as to avoid denial of
3559   service attacks without impeding interoperability.
3563<section title="Message Integrity" anchor="message.integrity">
3565   HTTP does not define a specific mechanism for ensuring message integrity,
3566   instead relying on the error-detection ability of underlying transport
3567   protocols and the use of length or chunk-delimited framing to detect
3568   completeness. Additional integrity mechanisms, such as hash functions or
3569   digital signatures applied to the content, can be selectively added to
3570   messages via extensible metadata header fields. Historically, the lack of
3571   a single integrity mechanism has been justified by the informal nature of
3572   most HTTP communication.  However, the prevalence of HTTP as an information
3573   access mechanism has resulted in its increasing use within environments
3574   where verification of message integrity is crucial.
3577   User agents are encouraged to implement configurable means for detecting
3578   and reporting failures of message integrity such that those means can be
3579   enabled within environments for which integrity is necessary. For example,
3580   a browser being used to view medical history or drug interaction
3581   information needs to indicate to the user when such information is detected
3582   by the protocol to be incomplete, expired, or corrupted during transfer.
3583   Such mechanisms might be selectively enabled via user agent extensions or
3584   the presence of message integrity metadata in a response.
3585   At a minimum, user agents ought to provide some indication that allows a
3586   user to distinguish between a complete and incomplete response message
3587   (<xref target="incomplete.messages"/>) when such verification is desired.
3591<section title="Server Log Information" anchor="abuse.of.server.log.information">
3593   A server is in the position to save personal data about a user's requests
3594   over time, which might identify their reading patterns or subjects of
3595   interest.  In particular, log information gathered at an intermediary
3596   often contains a history of user agent interaction, across a multitude
3597   of sites, that can be traced to individual users.
3600   HTTP log information is confidential in nature; its handling is often
3601   constrained by laws and regulations.  Log information needs to be securely
3602   stored and appropriate guidelines followed for its analysis.
3603   Anonymization of personal information within individual entries helps,
3604   but is generally not sufficient to prevent real log traces from being
3605   re-identified based on correlation with other access characteristics.
3606   As such, access traces that are keyed to a specific client should not
3607   be published even if the key is pseudonymous.
3610   To minimize the risk of theft or accidental publication, log information
3611   should be purged of personally identifiable information, including
3612   user identifiers, IP addresses, and user-provided query parameters,
3613   as soon as that information is no longer necessary to support operational
3614   needs for security, auditing, or fraud control.
3619<section title="Acknowledgments" anchor="acks">
3621   This edition of HTTP/1.1 builds on the many contributions that went into
3622   <xref target="RFC1945" format="none">RFC 1945</xref>,
3623   <xref target="RFC2068" format="none">RFC 2068</xref>,
3624   <xref target="RFC2145" format="none">RFC 2145</xref>, and
3625   <xref target="RFC2616" format="none">RFC 2616</xref>, including
3626   substantial contributions made by the previous authors, editors, and
3627   working group chairs: Tim Berners-Lee, Ari Luotonen, Roy T. Fielding,
3628   Henrik Frystyk Nielsen, Jim Gettys, Jeffrey C. Mogul, Larry Masinter,
3629   and Paul J. Leach. Mark Nottingham oversaw this effort as working group chair.
3632   Since 1999, the following contributors have helped improve the HTTP
3633   specification by reporting bugs, asking smart questions, drafting or
3634   reviewing text, and evaluating open issues:
3636<?BEGININC acks ?>
3637<t>Adam Barth,
3638Adam Roach,
3639Addison Phillips,
3640Adrian Chadd,
3641Adrien W. de Croy,
3642Alan Ford,
3643Alan Ruttenberg,
3644Albert Lunde,
3645Alek Storm,
3646Alex Rousskov,
3647Alexandre Morgaut,
3648Alexey Melnikov,
3649Alisha Smith,
3650Amichai Rothman,
3651Amit Klein,
3652Amos Jeffries,
3653Andreas Maier,
3654Andreas Petersson,
3655Anil Sharma,
3656Anne van Kesteren,
3657Anthony Bryan,
3658Asbjorn Ulsberg,
3659Ashok Kumar,
3660Balachander Krishnamurthy,
3661Barry Leiba,
3662Ben Laurie,
3663Benjamin Niven-Jenkins,
3664Bil Corry,
3665Bill Burke,
3666Bjoern Hoehrmann,
3667Bob Scheifler,
3668Boris Zbarsky,
3669Brett Slatkin,
3670Brian Kell,
3671Brian McBarron,
3672Brian Pane,
3673Brian Smith,
3674Bryce Nesbitt,
3675Cameron Heavon-Jones,
3676Carl Kugler,
3677Carsten Bormann,
3678Charles Fry,
3679Chris Newman,
3680Chris Weber,
3681Cyrus Daboo,
3682Dale Robert Anderson,
3683Dan Wing,
3684Dan Winship,
3685Daniel Stenberg,
3686Darrel Miller,
3687Dave Cridland,
3688Dave Crocker,
3689Dave Kristol,
3690David Booth,
3691David Singer,
3692David W. Morris,
3693Diwakar Shetty,
3694Dmitry Kurochkin,
3695Drummond Reed,
3696Duane Wessels,
3697Duncan Cragg,
3698Edward Lee,
3699Eliot Lear,
3700Eran Hammer-Lahav,
3701Eric D. Williams,
3702Eric J. Bowman,
3703Eric Lawrence,
3704Eric Rescorla,
3705Erik Aronesty,
3706Evan Prodromou,
3707Florian Weimer,
3708Frank Ellermann,
3709Fred Bohle,
3710Gabriel Montenegro,
3711Geoffrey Sneddon,
3712Gervase Markham,
3713Grahame Grieve,
3714Greg Wilkins,
3715Harald Tveit Alvestrand,
3716Harry Halpin,
3717Helge Hess,
3718Henrik Nordstrom,
3719Henry S. Thompson,
3720Henry Story,
3721Herbert van de Sompel,
3722Howard Melman,
3723Hugo Haas,
3724Ian Fette,
3725Ian Hickson,
3726Ido Safruti,
3727Ilya Grigorik,
3728Ingo Struck,
3729J. Ross Nicoll,
3730James H. Manger,
3731James Lacey,
3732James M. Snell,
3733Jamie Lokier,
3734Jan Algermissen,
3735Jeff Hodges (who came up with the term 'effective Request-URI'),
3736Jeff Walden,
3737Jeroen de Borst,
3738Jim Luther,
3739Joe D. Williams,
3740Joe Gregorio,
3741Joe Orton,
3742John C. Klensin,
3743John C. Mallery,
3744John Cowan,
3745John Kemp,
3746John Panzer,
3747John Schneider,
3748John Stracke,
3749John Sullivan,
3750Jonas Sicking,
3751Jonathan A. Rees,
3752Jonathan Billington,
3753Jonathan Moore,
3754Jonathan Rees,
3755Jonathan Silvera,
3756Jordi Ros,
3757Joris Dobbelsteen,
3758Josh Cohen,
3759Julien Pierre,
3760Jungshik Shin,
3761Justin Chapweske,
3762Justin Erenkrantz,
3763Justin James,
3764Kalvinder Singh,
3765Karl Dubost,
3766Keith Hoffman,
3767Keith Moore,
3768Ken Murchison,
3769Koen Holtman,
3770Konstantin Voronkov,
3771Kris Zyp,
3772Lisa Dusseault,
3773Maciej Stachowiak,
3774Marc Schneider,
3775Marc Slemko,
3776Mark Baker,
3777Mark Pauley,
3778Mark Watson,
3779Markus Isomaki,
3780Markus Lanthaler,
3781Martin J. Duerst,
3782Martin Musatov,
3783Martin Nilsson,
3784Martin Thomson,
3785Matt Lynch,
3786Matthew Cox,
3787Max Clark,
3788Michael Burrows,
3789Michael Hausenblas,
3790Mike Amundsen,
3791Mike Belshe,
3792Mike Kelly,
3793Mike Schinkel,
3794Miles Sabin,
3795Murray S. Kucherawy,
3796Mykyta Yevstifeyev,
3797Nathan Rixham,
3798Nicholas Shanks,
3799Nico Williams,
3800Nicolas Alvarez,
3801Nicolas Mailhot,
3802Noah Slater,
3803Pablo Castro,
3804Pat Hayes,
3805Patrick R. McManus,
3806Patrik Faltstrom,
3807Paul E. Jones,
3808Paul Hoffman,
3809Paul Marquess,
3810Peter Lepeska,
3811Peter Saint-Andre,
3812Peter Watkins,
3813Phil Archer,
3814Philippe Mougin,
3815Phillip Hallam-Baker,
3816Poul-Henning Kamp,
3817Preethi Natarajan,
3818Rajeev Bector,
3819Ray Polk,
3820Reto Bachmann-Gmuer,
3821Richard Cyganiak,
3822Robert Brewer,
3823Robert Collins,
3824Robert O'Callahan,
3825Robert Olofsson,
3826Robert Sayre,
3827Robert Siemer,
3828Robert de Wilde,
3829Roberto Javier Godoy,
3830Roberto Peon,
3831Roland Zink,
3832Ronny Widjaja,
3833S. Mike Dierken,
3834Salvatore Loreto,
3835Sam Johnston,
3836Sam Ruby,
3837Scott Lawrence (who maintained the original issues list),
3838Sean B. Palmer,
3839Shane McCarron,
3840Stefan Eissing,
3841Stefan Tilkov,
3842Stefanos Harhalakis,
3843Stephane Bortzmeyer,
3844Stephen Farrell,
3845Stephen Ludin,
3846Stuart Williams,
3847Subbu Allamaraju,
3848Subramanian Moonesamy,
3849Sylvain Hellegouarch,
3850Tapan Divekar,
3851Tatsuya Hayashi,
3852Ted Hardie,
3853Thomas Broyer,
3854Thomas Fossati,
3855Thomas Nordin,
3856Thomas Roessler,
3857Tim Bray,
3858Tim Morgan,
3859Tim Olsen,
3860Tobias Oberstein,
3861Tom Zhou,
3862Travis Snoozy,
3863Tyler Close,
3864Vincent Murphy,
3865Wenbo Zhu,
3866Werner Baumann,
3867Wilbur Streett,
3868Wilfredo Sanchez Vega,
3869William A. Rowe Jr.,
3870William Chan,
3871Willy Tarreau,
3872Xiaoshu Wang,
3873Yaron Goland,
3874Yngve Nysaeter Pettersen,
3875Yoav Nir,
3876Yogesh Bang,
3877Yutaka Oiwa,
3878Yves Lafon (long-time member of the editor team),
3879Zed A. Shaw, and
3880Zhong Yu.
3882<?ENDINC acks ?>
3884   See <xref target="RFC2616" x:fmt="of" x:sec="16"/> for additional
3885   acknowledgements from prior revisions.
3892<references title="Normative References">
3894<reference anchor="Part2">
3895  <front>
3896    <title>Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content</title>
3897    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3898      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3899      <address><email></email></address>
3900    </author>
3901    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3902      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3903      <address><email></email></address>
3904    </author>
3905    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3906  </front>
3907  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-&ID-VERSION;"/>
3908  <x:source href="p2-semantics.xml" basename="p2-semantics">
3909    <x:defines>1xx (Informational)</x:defines>
3910    <x:defines>1xx</x:defines>
3911    <x:defines>100 (Continue)</x:defines>
3912    <x:defines>101 (Switching Protocols)</x:defines>
3913    <x:defines>2xx (Successful)</x:defines>
3914    <x:defines>2xx</x:defines>
3915    <x:defines>200 (OK)</x:defines>
3916    <x:defines>204 (No Content)</x:defines>
3917    <x:defines>3xx (Redirection)</x:defines>
3918    <x:defines>3xx</x:defines>
3919    <x:defines>301 (Moved Permanently)</x:defines>
3920    <x:defines>4xx (Client Error)</x:defines>
3921    <x:defines>4xx</x:defines>
3922    <x:defines>400 (Bad Request)</x:defines>
3923    <x:defines>411 (Length Required)</x:defines>
3924    <x:defines>414 (URI Too Long)</x:defines>
3925    <x:defines>417 (Expectation Failed)</x:defines>
3926    <x:defines>426 (Upgrade Required)</x:defines>
3927    <x:defines>501 (Not Implemented)</x:defines>
3928    <x:defines>502 (Bad Gateway)</x:defines>
3929    <x:defines>505 (HTTP Version Not Supported)</x:defines>
3930    <x:defines>Allow</x:defines>
3931    <x:defines>Content-Encoding</x:defines>
3932    <x:defines>Content-Location</x:defines>
3933    <x:defines>Content-Type</x:defines>
3934    <x:defines>Date</x:defines>
3935    <x:defines>Expect</x:defines>
3936    <x:defines>Location</x:defines>
3937    <x:defines>Server</x:defines>
3938    <x:defines>User-Agent</x:defines>
3939  </x:source>
3942<reference anchor="Part4">
3943  <front>
3944    <title>Hypertext Transfer Protocol (HTTP/1.1): Conditional Requests</title>
3945    <author fullname="Roy T. Fielding" initials="R." role="editor" surname="Fielding">
3946      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3947      <address><email></email></address>
3948    </author>
3949    <author fullname="Julian F. Reschke" initials="J. F." role="editor" surname="Reschke">
3950      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3951      <address><email></email></address>
3952    </author>
3953    <date month="&ID-MONTH;" year="&ID-YEAR;" />
3954  </front>
3955  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p4-conditional-&ID-VERSION;" />
3956  <x:source basename="p4-conditional" href="p4-conditional.xml">
3957    <x:defines>304 (Not Modified)</x:defines>
3958    <x:defines>ETag</x:defines>
3959    <x:defines>Last-Modified</x:defines>
3960  </x:source>
3963<reference anchor="Part5">
3964  <front>
3965    <title>Hypertext Transfer Protocol (HTTP/1.1): Range Requests</title>
3966    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3967      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3968      <address><email></email></address>
3969    </author>
3970    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3971      <organization abbrev="W3C">World Wide Web Consortium</organization>
3972      <address><email></email></address>
3973    </author>
3974    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3975      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3976      <address><email></email></address>
3977    </author>
3978    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3979  </front>
3980  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-&ID-VERSION;"/>
3981  <x:source href="p5-range.xml" basename="p5-range">
3982    <x:defines>Content-Range</x:defines>
3983  </x:source>
3986<reference anchor="Part6">
3987  <front>
3988    <title>Hypertext Transfer Protocol (HTTP/1.1): Caching</title>
3989    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3990      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3991      <address><email></email></address>
3992    </author>
3993    <author initials="M." surname="Nottingham" fullname="Mark Nottingham" role="editor">
3994      <organization>Akamai</organization>
3995      <address><email></email></address>
3996    </author>
3997    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3998      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3999      <address><email></email></address>
4000    </author>
4001    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4002  </front>
4003  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-&ID-VERSION;"/>
4004  <x:source href="p6-cache.xml" basename="p6-cache">
4005    <x:defines>Cache-Control</x:defines>
4006    <x:defines>Expires</x:defines>
4007  </x:source>
4010<reference anchor="Part7">
4011  <front>
4012    <title>Hypertext Transfer Protocol (HTTP/1.1): Authentication</title>
4013    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4014      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4015      <address><email></email></address>
4016    </author>
4017    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4018      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4019      <address><email></email></address>
4020    </author>
4021    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4022  </front>
4023  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p7-auth-&ID-VERSION;"/>
4024  <x:source href="p7-auth.xml" basename="p7-auth">
4025    <x:defines>Proxy-Authenticate</x:defines>
4026    <x:defines>Proxy-Authorization</x:defines>
4027  </x:source>
4030<reference anchor="RFC5234">
4031  <front>
4032    <title abbrev="ABNF for Syntax Specifications">Augmented BNF for Syntax Specifications: ABNF</title>
4033    <author initials="D." surname="Crocker" fullname="Dave Crocker" role="editor">
4034      <organization>Brandenburg InternetWorking</organization>
4035      <address>
4036        <email></email>
4037      </address> 
4038    </author>
4039    <author initials="P." surname="Overell" fullname="Paul Overell">
4040      <organization>THUS plc.</organization>
4041      <address>
4042        <email></email>
4043      </address>
4044    </author>
4045    <date month="January" year="2008"/>
4046  </front>
4047  <seriesInfo name="STD" value="68"/>
4048  <seriesInfo name="RFC" value="5234"/>
4051<reference anchor="RFC2119">
4052  <front>
4053    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
4054    <author initials="S." surname="Bradner" fullname="Scott Bradner">
4055      <organization>Harvard University</organization>
4056      <address><email></email></address>
4057    </author>
4058    <date month="March" year="1997"/>
4059  </front>
4060  <seriesInfo name="BCP" value="14"/>
4061  <seriesInfo name="RFC" value="2119"/>
4064<reference anchor="RFC3986">
4065 <front>
4066  <title abbrev='URI Generic Syntax'>Uniform Resource Identifier (URI): Generic Syntax</title>
4067  <author initials='T.' surname='Berners-Lee' fullname='Tim Berners-Lee'>
4068    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
4069    <address>
4070       <email></email>
4071       <uri></uri>
4072    </address>
4073  </author>
4074  <author initials='R.' surname='Fielding' fullname='Roy T. Fielding'>
4075    <organization abbrev="Day Software">Day Software</organization>
4076    <address>
4077      <email></email>
4078      <uri></uri>
4079    </address>
4080  </author>
4081  <author initials='L.' surname='Masinter' fullname='Larry Masinter'>
4082    <organization abbrev="Adobe Systems">Adobe Systems Incorporated</organization>
4083    <address>
4084      <email></email>
4085      <uri></uri>
4086    </address>
4087  </author>
4088  <date month='January' year='2005'></date>
4089 </front>
4090 <seriesInfo name="STD" value="66"/>
4091 <seriesInfo name="RFC" value="3986"/>
4094<reference anchor="USASCII">
4095  <front>
4096    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
4097    <author>
4098      <organization>American National Standards Institute</organization>
4099    </author>
4100    <date year="1986"/>
4101  </front>
4102  <seriesInfo name="ANSI" value="X3.4"/>
4105<reference anchor="RFC1950">
4106  <front>
4107    <title>ZLIB Compressed Data Format Specification version 3.3</title>
4108    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4109      <organization>Aladdin Enterprises</organization>
4110      <address><email></email></address>
4111    </author>
4112    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly"/>
4113    <date month="May" year="1996"/>
4114  </front>
4115  <seriesInfo name="RFC" value="1950"/>
4116  <!--<annotation>
4117    RFC 1950 is an Informational RFC, thus it might be less stable than
4118    this specification. On the other hand, this downward reference was
4119    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4120    therefore it is unlikely to cause problems in practice. See also
4121    <xref target="BCP97"/>.
4122  </annotation>-->
4125<reference anchor="RFC1951">
4126  <front>
4127    <title>DEFLATE Compressed Data Format Specification version 1.3</title>
4128    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4129      <organization>Aladdin Enterprises</organization>
4130      <address><email></email></address>
4131    </author>
4132    <date month="May" year="1996"/>
4133  </front>
4134  <seriesInfo name="RFC" value="1951"/>
4135  <!--<annotation>
4136    RFC 1951 is an Informational RFC, thus it might be less stable than
4137    this specification. On the other hand, this downward reference was
4138    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4139    therefore it is unlikely to cause problems in practice. See also
4140    <xref target="BCP97"/>.
4141  </annotation>-->
4144<reference anchor="RFC1952">
4145  <front>
4146    <title>GZIP file format specification version 4.3</title>
4147    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4148      <organization>Aladdin Enterprises</organization>
4149      <address><email></email></address>
4150    </author>
4151    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly">
4152      <address><email></email></address>
4153    </author>
4154    <author initials="M." surname="Adler" fullname="Mark Adler">
4155      <address><email></email></address>
4156    </author>
4157    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4158      <address><email></email></address>
4159    </author>
4160    <author initials="G." surname="Randers-Pehrson" fullname="Glenn Randers-Pehrson">
4161      <address><email></email></address>
4162    </author>
4163    <date month="May" year="1996"/>
4164  </front>
4165  <seriesInfo name="RFC" value="1952"/>
4166  <!--<annotation>
4167    RFC 1952 is an Informational RFC, thus it might be less stable than
4168    this specification. On the other hand, this downward reference was
4169    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4170    therefore it is unlikely to cause problems in practice. See also
4171    <xref target="BCP97"/>.
4172  </annotation>-->
4177<references title="Informative References">
4179<reference anchor="ISO-8859-1">
4180  <front>
4181    <title>
4182     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
4183    </title>
4184    <author>
4185      <organization>International Organization for Standardization</organization>
4186    </author>
4187    <date year="1998"/>
4188  </front>
4189  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
4192<reference anchor='RFC1919'>
4193  <front>
4194    <title>Classical versus Transparent IP Proxies</title>
4195    <author initials='M.' surname='Chatel' fullname='Marc Chatel'>
4196      <address><email></email></address>
4197    </author>
4198    <date year='1996' month='March' />
4199  </front>
4200  <seriesInfo name='RFC' value='1919' />
4203<reference anchor="RFC1945">
4204  <front>
4205    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
4206    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4207      <organization>MIT, Laboratory for Computer Science</organization>
4208      <address><email></email></address>
4209    </author>
4210    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4211      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4212      <address><email></email></address>
4213    </author>
4214    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4215      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
4216      <address><email></email></address>
4217    </author>
4218    <date month="May" year="1996"/>
4219  </front>
4220  <seriesInfo name="RFC" value="1945"/>
4223<reference anchor="RFC2045">
4224  <front>
4225    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
4226    <author initials="N." surname="Freed" fullname="Ned Freed">
4227      <organization>Innosoft International, Inc.</organization>
4228      <address><email></email></address>
4229    </author>
4230    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
4231      <organization>First Virtual Holdings</organization>
4232      <address><email></email></address>
4233    </author>
4234    <date month="November" year="1996"/>
4235  </front>
4236  <seriesInfo name="RFC" value="2045"/>
4239<reference anchor="RFC2047">
4240  <front>
4241    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
4242    <author initials="K." surname="Moore" fullname="Keith Moore">
4243      <organization>University of Tennessee</organization>
4244      <address><email></email></address>
4245    </author>
4246    <date month="November" year="1996"/>
4247  </front>
4248  <seriesInfo name="RFC" value="2047"/>
4251<reference anchor="RFC2068">
4252  <front>
4253    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4254    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
4255      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4256      <address><email></email></address>
4257    </author>
4258    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4259      <organization>MIT Laboratory for Computer Science</organization>
4260      <address><email></email></address>
4261    </author>
4262    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
4263      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
4264      <address><email></email></address>
4265    </author>
4266    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4267      <organization>MIT Laboratory for Computer Science</organization>
4268      <address><email></email></address>
4269    </author>
4270    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4271      <organization>MIT Laboratory for Computer Science</organization>
4272      <address><email></email></address>
4273    </author>
4274    <date month="January" year="1997"/>
4275  </front>
4276  <seriesInfo name="RFC" value="2068"/>
4279<reference anchor="RFC2145">
4280  <front>
4281    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
4282    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
4283      <organization>Western Research Laboratory</organization>
4284      <address><email></email></address>
4285    </author>
4286    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4287      <organization>Department of Information and Computer Science</organization>
4288      <address><email></email></address>
4289    </author>
4290    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4291      <organization>MIT Laboratory for Computer Science</organization>
4292      <address><email></email></address>
4293    </author>
4294    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4295      <organization>W3 Consortium</organization>
4296      <address><email></email></address>
4297    </author>
4298    <date month="May" year="1997"/>
4299  </front>
4300  <seriesInfo name="RFC" value="2145"/>
4303<reference anchor="RFC2616">
4304  <front>
4305    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4306    <author initials="R." surname="Fielding" fullname="R. Fielding">
4307      <organization>University of California, Irvine</organization>
4308      <address><email></email></address>
4309    </author>
4310    <author initials="J." surname="Gettys" fullname="J. Gettys">
4311      <organization>W3C</organization>
4312      <address><email></email></address>
4313    </author>
4314    <author initials="J." surname="Mogul" fullname="J. Mogul">
4315      <organization>Compaq Computer Corporation</organization>
4316      <address><email></email></address>
4317    </author>
4318    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
4319      <organization>MIT Laboratory for Computer Science</organization>
4320      <address><email></email></address>
4321    </author>
4322    <author initials="L." surname="Masinter" fullname="L. Masinter">
4323      <organization>Xerox Corporation</organization>
4324      <address><email></email></address>
4325    </author>
4326    <author initials="P." surname="Leach" fullname="P. Leach">
4327      <organization>Microsoft Corporation</organization>
4328      <address><email></email></address>
4329    </author>
4330    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
4331      <organization>W3C</organization>
4332      <address><email></email></address>
4333    </author>
4334    <date month="June" year="1999"/>
4335  </front>
4336  <seriesInfo name="RFC" value="2616"/>
4339<reference anchor='RFC2817'>
4340  <front>
4341    <title>Upgrading to TLS Within HTTP/1.1</title>
4342    <author initials='R.' surname='Khare' fullname='R. Khare'>
4343      <organization>4K Associates / UC Irvine</organization>
4344      <address><email></email></address>
4345    </author>
4346    <author initials='S.' surname='Lawrence' fullname='S. Lawrence'>
4347      <organization>Agranat Systems, Inc.</organization>
4348      <address><email></email></address>
4349    </author>
4350    <date year='2000' month='May' />
4351  </front>
4352  <seriesInfo name='RFC' value='2817' />
4355<reference anchor='RFC2818'>
4356  <front>
4357    <title>HTTP Over TLS</title>
4358    <author initials='E.' surname='Rescorla' fullname='Eric Rescorla'>
4359      <organization>RTFM, Inc.</organization>
4360      <address><email></email></address>
4361    </author>
4362    <date year='2000' month='May' />
4363  </front>
4364  <seriesInfo name='RFC' value='2818' />
4367<reference anchor='RFC3040'>
4368  <front>
4369    <title>Internet Web Replication and Caching Taxonomy</title>
4370    <author initials='I.' surname='Cooper' fullname='I. Cooper'>
4371      <organization>Equinix, Inc.</organization>
4372    </author>
4373    <author initials='I.' surname='Melve' fullname='I. Melve'>
4374      <organization>UNINETT</organization>
4375    </author>
4376    <author initials='G.' surname='Tomlinson' fullname='G. Tomlinson'>
4377      <organization>CacheFlow Inc.</organization>
4378    </author>
4379    <date year='2001' month='January' />
4380  </front>
4381  <seriesInfo name='RFC' value='3040' />
4384<reference anchor='BCP90'>
4385  <front>
4386    <title>Registration Procedures for Message Header Fields</title>
4387    <author initials='G.' surname='Klyne' fullname='G. Klyne'>
4388      <organization>Nine by Nine</organization>
4389      <address><email></email></address>
4390    </author>
4391    <author initials='M.' surname='Nottingham' fullname='M. Nottingham'>
4392      <organization>BEA Systems</organization>
4393      <address><email></email></address>
4394    </author>
4395    <author initials='J.' surname='Mogul' fullname='J. Mogul'>
4396      <organization>HP Labs</organization>
4397      <address><email></email></address>
4398    </author>
4399    <date year='2004' month='September' />
4400  </front>
4401  <seriesInfo name='BCP' value='90' />
4402  <seriesInfo name='RFC' value='3864' />
4405<reference anchor='RFC4033'>
4406  <front>
4407    <title>DNS Security Introduction and Requirements</title>
4408    <author initials='R.' surname='Arends' fullname='R. Arends'/>
4409    <author initials='R.' surname='Austein' fullname='R. Austein'/>
4410    <author initials='M.' surname='Larson' fullname='M. Larson'/>
4411    <author initials='D.' surname='Massey' fullname='D. Massey'/>
4412    <author initials='S.' surname='Rose' fullname='S. Rose'/>
4413    <date year='2005' month='March' />
4414  </front>
4415  <seriesInfo name='RFC' value='4033' />
4418<reference anchor="BCP13">
4419  <front>
4420    <title>Media Type Specifications and Registration Procedures</title>
4421    <author initials="N." surname="Freed" fullname="N. Freed">
4422      <organization>Sun Microsystems</organization>
4423      <address>
4424        <email></email>
4425      </address>
4426    </author>
4427    <author initials="J." surname="Klensin" fullname="J. Klensin">
4428      <address>
4429        <email></email>
4430      </address>
4431    </author>
4432    <date year="2005" month="December"/>
4433  </front>
4434  <seriesInfo name="BCP" value="13"/>
4435  <seriesInfo name="RFC" value="4288"/>
4438<reference anchor='BCP115'>
4439  <front>
4440    <title>Guidelines and Registration Procedures for New URI Schemes</title>
4441    <author initials='T.' surname='Hansen' fullname='T. Hansen'>
4442      <organization>AT&amp;T Laboratories</organization>
4443      <address>
4444        <email></email>
4445      </address>
4446    </author>
4447    <author initials='T.' surname='Hardie' fullname='T. Hardie'>
4448      <organization>Qualcomm, Inc.</organization>
4449      <address>
4450        <email></email>
4451      </address>
4452    </author>
4453    <author initials='L.' surname='Masinter' fullname='L. Masinter'>
4454      <organization>Adobe Systems</organization>
4455      <address>
4456        <email></email>
4457      </address>
4458    </author>
4459    <date year='2006' month='February' />
4460  </front>
4461  <seriesInfo name='BCP' value='115' />
4462  <seriesInfo name='RFC' value='4395' />
4465<reference anchor='RFC4559'>
4466  <front>
4467    <title>SPNEGO-based Kerberos and NTLM HTTP Authentication in Microsoft Windows</title>
4468    <author initials='K.' surname='Jaganathan' fullname='K. Jaganathan'/>
4469    <author initials='L.' surname='Zhu' fullname='L. Zhu'/>
4470    <author initials='J.' surname='Brezak' fullname='J. Brezak'/>
4471    <date year='2006' month='June' />
4472  </front>
4473  <seriesInfo name='RFC' value='4559' />
4476<reference anchor='RFC5226'>
4477  <front>
4478    <title>Guidelines for Writing an IANA Considerations Section in RFCs</title>
4479    <author initials='T.' surname='Narten' fullname='T. Narten'>
4480      <organization>IBM</organization>
4481      <address><email></email></address>
4482    </author>
4483    <author initials='H.' surname='Alvestrand' fullname='H. Alvestrand'>
4484      <organization>Google</organization>
4485      <address><email></email></address>
4486    </author>
4487    <date year='2008' month='May' />
4488  </front>
4489  <seriesInfo name='BCP' value='26' />
4490  <seriesInfo name='RFC' value='5226' />
4493<reference anchor='RFC5246'>
4494   <front>
4495      <title>The Transport Layer Security (TLS) Protocol Version 1.2</title>
4496      <author initials='T.' surname='Dierks' fullname='T. Dierks'>
4497         <organization />
4498      </author>
4499      <author initials='E.' surname='Rescorla' fullname='E. Rescorla'>
4500         <organization>RTFM, Inc.</organization>
4501      </author>
4502      <date year='2008' month='August' />
4503   </front>
4504   <seriesInfo name='RFC' value='5246' />
4507<reference anchor="RFC5322">
4508  <front>
4509    <title>Internet Message Format</title>
4510    <author initials="P." surname="Resnick" fullname="P. Resnick">
4511      <organization>Qualcomm Incorporated</organization>
4512    </author>
4513    <date year="2008" month="October"/>
4514  </front>
4515  <seriesInfo name="RFC" value="5322"/>
4518<reference anchor="RFC6265">
4519  <front>
4520    <title>HTTP State Management Mechanism</title>
4521    <author initials="A." surname="Barth" fullname="Adam Barth">
4522      <organization abbrev="U.C. Berkeley">
4523        University of California, Berkeley
4524      </organization>
4525      <address><email></email></address>
4526    </author>
4527    <date year="2011" month="April" />
4528  </front>
4529  <seriesInfo name="RFC" value="6265"/>
4532<!--<reference anchor='BCP97'>
4533  <front>
4534    <title>Handling Normative References to Standards-Track Documents</title>
4535    <author initials='J.' surname='Klensin' fullname='J. Klensin'>
4536      <address>
4537        <email></email>
4538      </address>
4539    </author>
4540    <author initials='S.' surname='Hartman' fullname='S. Hartman'>
4541      <organization>MIT</organization>
4542      <address>
4543        <email></email>
4544      </address>
4545    </author>
4546    <date year='2007' month='June' />
4547  </front>
4548  <seriesInfo name='BCP' value='97' />
4549  <seriesInfo name='RFC' value='4897' />
4552<reference anchor="Kri2001" target="">
4553  <front>
4554    <title>HTTP Cookies: Standards, Privacy, and Politics</title>
4555    <author initials="D." surname="Kristol" fullname="David M. Kristol"/>
4556    <date year="2001" month="November"/>
4557  </front>
4558  <seriesInfo name="ACM Transactions on Internet Technology" value="Vol. 1, #2"/>
4564<section title="HTTP Version History" anchor="compatibility">
4566   HTTP has been in use by the World-Wide Web global information initiative
4567   since 1990. The first version of HTTP, later referred to as HTTP/0.9,
4568   was a simple protocol for hypertext data transfer across the Internet
4569   with only a single request method (GET) and no metadata.
4570   HTTP/1.0, as defined by <xref target="RFC1945"/>, added a range of request
4571   methods and MIME-like messaging that could include metadata about the data
4572   transferred and modifiers on the request/response semantics. However,
4573   HTTP/1.0 did not sufficiently take into consideration the effects of
4574   hierarchical proxies, caching, the need for persistent connections, or
4575   name-based virtual hosts. The proliferation of incompletely-implemented
4576   applications calling themselves "HTTP/1.0" further necessitated a
4577   protocol version change in order for two communicating applications
4578   to determine each other's true capabilities.
4581   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
4582   requirements that enable reliable implementations, adding only
4583   those new features that will either be safely ignored by an HTTP/1.0
4584   recipient or only sent when communicating with a party advertising
4585   conformance with HTTP/1.1.
4588   It is beyond the scope of a protocol specification to mandate
4589   conformance with previous versions. HTTP/1.1 was deliberately
4590   designed, however, to make supporting previous versions easy.
4591   We would expect a general-purpose HTTP/1.1 server to understand
4592   any valid request in the format of HTTP/1.0 and respond appropriately
4593   with an HTTP/1.1 message that only uses features understood (or
4594   safely ignored) by HTTP/1.0 clients.  Likewise, we would expect
4595   an HTTP/1.1 client to understand any valid HTTP/1.0 response.
4598   Since HTTP/0.9 did not support header fields in a request,
4599   there is no mechanism for it to support name-based virtual
4600   hosts (selection of resource by inspection of the <x:ref>Host</x:ref> header
4601   field).  Any server that implements name-based virtual hosts
4602   ought to disable support for HTTP/0.9.  Most requests that
4603   appear to be HTTP/0.9 are, in fact, badly constructed HTTP/1.x
4604   requests wherein a buggy client failed to properly encode
4605   linear whitespace found in a URI reference and placed in
4606   the request-target.
4609<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
4611   This section summarizes major differences between versions HTTP/1.0
4612   and HTTP/1.1.
4615<section title="Multi-homed Web Servers" anchor="">
4617   The requirements that clients and servers support the <x:ref>Host</x:ref>
4618   header field (<xref target=""/>), report an error if it is
4619   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-target"/>)
4620   are among the most important changes defined by HTTP/1.1.
4623   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
4624   addresses and servers; there was no other established mechanism for
4625   distinguishing the intended server of a request than the IP address
4626   to which that request was directed. The <x:ref>Host</x:ref> header field was
4627   introduced during the development of HTTP/1.1 and, though it was
4628   quickly implemented by most HTTP/1.0 browsers, additional requirements
4629   were placed on all HTTP/1.1 requests in order to ensure complete
4630   adoption.  At the time of this writing, most HTTP-based services
4631   are dependent upon the Host header field for targeting requests.
4635<section title="Keep-Alive Connections" anchor="compatibility.with.http.1.0.persistent.connections">
4637   In HTTP/1.0, each connection is established by the client prior to the
4638   request and closed by the server after sending the response. However, some
4639   implementations implement the explicitly negotiated ("Keep-Alive") version
4640   of persistent connections described in <xref x:sec="19.7.1" x:fmt="of"
4641   target="RFC2068"/>.
4644   Some clients and servers might wish to be compatible with these previous
4645   approaches to persistent connections, by explicitly negotiating for them
4646   with a "Connection: keep-alive" request header field. However, some
4647   experimental implementations of HTTP/1.0 persistent connections are faulty;
4648   for example, if an HTTP/1.0 proxy server doesn't understand
4649   <x:ref>Connection</x:ref>, it will erroneously forward that header field
4650   to the next inbound server, which would result in a hung connection.
4653   One attempted solution was the introduction of a Proxy-Connection header
4654   field, targeted specifically at proxies. In practice, this was also
4655   unworkable, because proxies are often deployed in multiple layers, bringing
4656   about the same problem discussed above.
4659   As a result, clients are encouraged not to send the Proxy-Connection header
4660   field in any requests.
4663   Clients are also encouraged to consider the use of Connection: keep-alive
4664   in requests carefully; while they can enable persistent connections with
4665   HTTP/1.0 servers, clients using them need will need to monitor the
4666   connection for "hung" requests (which indicate that the client ought stop
4667   sending the header field), and this mechanism ought not be used by clients
4668   at all when a proxy is being used.
4672<section title="Introduction of Transfer-Encoding" anchor="introduction.of.transfer-encoding">
4674   HTTP/1.1 introduces the <x:ref>Transfer-Encoding</x:ref> header field
4675   (<xref target="header.transfer-encoding"/>).
4676   Transfer codings need to be decoded prior to forwarding an HTTP message
4677   over a MIME-compliant protocol.
4683<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
4685  HTTP's approach to error handling has been explained.
4686  (<xref target="conformance"/>)
4689  The expectation to support HTTP/0.9 requests has been removed.
4692  The term "Effective Request URI" has been introduced.
4693  (<xref target="effective.request.uri" />)
4696  HTTP messages can be (and often are) buffered by implementations; despite
4697  it sometimes being available as a stream, HTTP is fundamentally a
4698  message-oriented protocol.
4699  (<xref target="http.message" />)
4702  Minimum supported sizes for various protocol elements have been
4703  suggested, to improve interoperability.
4706  Header fields that span multiple lines ("line folding") are deprecated.
4707  (<xref target="field.parsing" />)
4710  The HTTP-version ABNF production has been clarified to be case-sensitive.
4711  Additionally, version numbers has been restricted to single digits, due
4712  to the fact that implementations are known to handle multi-digit version
4713  numbers incorrectly.
4714  (<xref target="http.version"/>)
4717  The HTTPS URI scheme is now defined by this specification; previously,
4718  it was done in  <xref target="RFC2818" x:fmt="of" x:sec="2.4"/>.
4719  (<xref target="https.uri"/>)
4722  The HTTPS URI scheme implies end-to-end security.
4723  (<xref target="https.uri"/>)
4726  Userinfo (i.e., username and password) are now disallowed in HTTP and
4727  HTTPS URIs, because of security issues related to their transmission on the
4728  wire.
4729  (<xref target="http.uri" />)
4732  Invalid whitespace around field-names is now required to be rejected,
4733  because accepting it represents a security vulnerability.
4734  (<xref target="header.fields"/>)
4737  The ABNF productions defining header fields now only list the field value.
4738  (<xref target="header.fields"/>)
4741  Rules about implicit linear whitespace between certain grammar productions
4742  have been removed; now whitespace is only allowed where specifically
4743  defined in the ABNF.
4744  (<xref target="whitespace"/>)
4747  The NUL octet is no longer allowed in comment and quoted-string text, and
4748  handling of backslash-escaping in them has been clarified.
4749  (<xref target="field.components"/>)
4752  The quoted-pair rule no longer allows escaping control characters other than
4753  HTAB.
4754  (<xref target="field.components"/>)
4757  Non-ASCII content in header fields and the reason phrase has been obsoleted
4758  and made opaque (the TEXT rule was removed).
4759  (<xref target="field.components"/>)
4762  Bogus "<x:ref>Content-Length</x:ref>" header fields are now required to be
4763  handled as errors by recipients.
4764  (<xref target="header.content-length"/>)
4767  The "identity" transfer coding token has been removed.
4768  (Sections <xref format="counter" target="message.body"/> and
4769  <xref format="counter" target="transfer.codings"/>)
4772  The algorithm for determining the message body length has been clarified
4773  to indicate all of the special cases (e.g., driven by methods or status
4774  codes) that affect it, and that new protocol elements cannot define such
4775  special cases.
4776  (<xref target="message.body.length"/>)
4779  "multipart/byteranges" is no longer a way of determining message body length
4780  detection.
4781  (<xref target="message.body.length"/>)
4784  CONNECT is a new, special case in determining message body length.
4785  (<xref target="message.body.length"/>)
4788  Chunk length does not include the count of the octets in the
4789  chunk header and trailer.
4790  (<xref target="chunked.encoding"/>)
4793  Use of chunk extensions is deprecated, and line folding in them is
4794  disallowed.
4795  (<xref target="chunked.encoding"/>)
4798  The path-absolute + query components of RFC3986 have been used to define the
4799  request-target, instead of abs_path from RFC 1808.
4800  (<xref target="request-target"/>)
4803  The asterisk form of the request-target is only allowed in the OPTIONS
4804  method.
4805  (<xref target="request-target"/>)
4808  Exactly when "close" connection options have to be sent has been clarified.
4809  (<xref target="header.connection"/>)
4812  "hop-by-hop" header fields are required to appear in the Connection header
4813  field; just because they're defined as hop-by-hop in this specification
4814  doesn't exempt them.
4815  (<xref target="header.connection"/>)
4818  The limit of two connections per server has been removed.
4819  (<xref target="persistent.connections"/>)
4822  An idempotent sequence of requests is no longer required to be retried.
4823  (<xref target="persistent.connections"/>)
4826  The requirement to retry requests under certain circumstances when the
4827  server prematurely closes the connection has been removed.
4828  (<xref target="persistent.connections"/>)
4831  Some extraneous requirements about when servers are allowed to close
4832  connections prematurely have been removed.
4833  (<xref target="persistent.connections"/>)
4836  The semantics of the <x:ref>Upgrade</x:ref> header field is now defined in
4837  responses other than 101 (this was incorporated from <xref
4838  target="RFC2817"/>).
4839  (<xref target="header.upgrade"/>)
4842  Registration of Transfer Codings now requires IETF Review
4843  (<xref target="transfer.coding.registry"/>)
4846  The meaning of the "deflate" content coding has been clarified.
4847  (<xref target="deflate.coding" />)
4850  This specification now defines the Upgrade Token Registry, previously
4851  defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
4852  (<xref target="upgrade.token.registry"/>)
4855  Empty list elements in list productions (e.g., a list header containing
4856  ", ,") have been deprecated.
4857  (<xref target="abnf.extension"/>)
4860  Issues with the Keep-Alive and Proxy-Connection headers in requests
4861  are pointed out, with use of the latter being discouraged altogether.
4862  (<xref target="compatibility.with.http.1.0.persistent.connections" />)
4867<section title="ABNF list extension: #rule" anchor="abnf.extension">
4869  A #rule extension to the ABNF rules of <xref target="RFC5234"/> is used to
4870  improve readability in the definitions of some header field values.
4873  A construct "#" is defined, similar to "*", for defining comma-delimited
4874  lists of elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating
4875  at least &lt;n&gt; and at most &lt;m&gt; elements, each separated by a single
4876  comma (",") and optional whitespace (OWS).   
4879  Thus,
4880</preamble><artwork type="example">
4881  1#element =&gt; element *( OWS "," OWS element )
4884  and:
4885</preamble><artwork type="example">
4886  #element =&gt; [ 1#element ]
4889  and for n &gt;= 1 and m &gt; 1:
4890</preamble><artwork type="example">
4891  &lt;n&gt;#&lt;m&gt;element =&gt; element &lt;n-1&gt;*&lt;m-1&gt;( OWS "," OWS element )
4894  For compatibility with legacy list rules, recipients &SHOULD; accept empty
4895  list elements. In other words, consumers would follow the list productions:
4897<figure><artwork type="example">
4898  #element =&gt; [ ( "," / element ) *( OWS "," [ OWS element ] ) ]
4900  1#element =&gt; *( "," OWS ) element *( OWS "," [ OWS element ] )
4903  Note that empty elements do not contribute to the count of elements present,
4904  though.
4907  For example, given these ABNF productions:
4909<figure><artwork type="example">
4910  example-list      = 1#example-list-elmt
4911  example-list-elmt = token ; see <xref target="field.components"/>
4914  Then these are valid values for example-list (not including the double
4915  quotes, which are present for delimitation only):
4917<figure><artwork type="example">
4918  "foo,bar"
4919  "foo ,bar,"
4920  "foo , ,bar,charlie   "
4923  But these values would be invalid, as at least one non-empty element is
4924  required:
4926<figure><artwork type="example">
4927  ""
4928  ","
4929  ",   ,"
4932  <xref target="collected.abnf"/> shows the collected ABNF, with the list rules
4933  expanded as explained above.
4937<?BEGININC p1-messaging.abnf-appendix ?>
4938<section xmlns:x="" title="Collected ABNF" anchor="collected.abnf">
4940<artwork type="abnf" name="p1-messaging.parsed-abnf">
4941<x:ref>BWS</x:ref> = OWS
4943<x:ref>Connection</x:ref> = *( "," OWS ) connection-option *( OWS "," [ OWS
4944 connection-option ] )
4945<x:ref>Content-Length</x:ref> = 1*DIGIT
4947<x:ref>HTTP-message</x:ref> = start-line *( header-field CRLF ) CRLF [ message-body
4948 ]
4949<x:ref>HTTP-name</x:ref> = %x48.54.54.50 ; HTTP
4950<x:ref>HTTP-version</x:ref> = HTTP-name "/" DIGIT "." DIGIT
4951<x:ref>Host</x:ref> = uri-host [ ":" port ]
4953<x:ref>OWS</x:ref> = *( SP / HTAB )
4955<x:ref>RWS</x:ref> = 1*( SP / HTAB )
4957<x:ref>TE</x:ref> = [ ( "," / t-codings ) *( OWS "," [ OWS t-codings ] ) ]
4958<x:ref>Trailer</x:ref> = *( "," OWS ) field-name *( OWS "," [ OWS field-name ] )
4959<x:ref>Transfer-Encoding</x:ref> = *( "," OWS ) transfer-coding *( OWS "," [ OWS
4960 transfer-coding ] )
4962<x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in [RFC3986], Section 4.1&gt;
4963<x:ref>Upgrade</x:ref> = *( "," OWS ) protocol *( OWS "," [ OWS protocol ] )
4965<x:ref>Via</x:ref> = *( "," OWS ) ( received-protocol RWS received-by [ RWS comment
4966 ] ) *( OWS "," [ OWS ( received-protocol RWS received-by [ RWS
4967 comment ] ) ] )
4969<x:ref>absolute-URI</x:ref> = &lt;absolute-URI, defined in [RFC3986], Section 4.3&gt;
4970<x:ref>absolute-form</x:ref> = absolute-URI
4971<x:ref>asterisk-form</x:ref> = "*"
4972<x:ref>attribute</x:ref> = token
4973<x:ref>authority</x:ref> = &lt;authority, defined in [RFC3986], Section 3.2&gt;
4974<x:ref>authority-form</x:ref> = authority
4976<x:ref>chunk</x:ref> = chunk-size [ chunk-ext ] CRLF chunk-data CRLF
4977<x:ref>chunk-data</x:ref> = 1*OCTET
4978<x:ref>chunk-ext</x:ref> = *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
4979<x:ref>chunk-ext-name</x:ref> = token
4980<x:ref>chunk-ext-val</x:ref> = token / quoted-str-nf
4981<x:ref>chunk-size</x:ref> = 1*HEXDIG
4982<x:ref>chunked-body</x:ref> = *chunk last-chunk trailer-part CRLF
4983<x:ref>comment</x:ref> = "(" *( ctext / quoted-cpair / comment ) ")"
4984<x:ref>connection-option</x:ref> = token
4985<x:ref>ctext</x:ref> = HTAB / SP / %x21-27 ; '!'-'''
4986 / %x2A-5B ; '*'-'['
4987 / %x5D-7E ; ']'-'~'
4988 / obs-text
4990<x:ref>field-content</x:ref> = *( HTAB / SP / VCHAR / obs-text )
4991<x:ref>field-name</x:ref> = token
4992<x:ref>field-value</x:ref> = *( field-content / obs-fold )
4994<x:ref>header-field</x:ref> = field-name ":" OWS field-value BWS
4995<x:ref>http-URI</x:ref> = "http://" authority path-abempty [ "?" query ]
4996<x:ref>https-URI</x:ref> = "https://" authority path-abempty [ "?" query ]
4998<x:ref>last-chunk</x:ref> = 1*"0" [ chunk-ext ] CRLF
5000<x:ref>message-body</x:ref> = *OCTET
5001<x:ref>method</x:ref> = token
5003<x:ref>obs-fold</x:ref> = CRLF ( SP / HTAB )
5004<x:ref>obs-text</x:ref> = %x80-FF
5005<x:ref>origin-form</x:ref> = path-absolute [ "?" query ]
5007<x:ref>partial-URI</x:ref> = relative-part [ "?" query ]
5008<x:ref>path-abempty</x:ref> = &lt;path-abempty, defined in [RFC3986], Section 3.3&gt;
5009<x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in [RFC3986], Section 3.3&gt;
5010<x:ref>port</x:ref> = &lt;port, defined in [RFC3986], Section 3.2.3&gt;
5011<x:ref>protocol</x:ref> = protocol-name [ "/" protocol-version ]
5012<x:ref>protocol-name</x:ref> = token
5013<x:ref>protocol-version</x:ref> = token
5014<x:ref>pseudonym</x:ref> = token
5016<x:ref>qdtext</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5017 / %x5D-7E ; ']'-'~'
5018 / obs-text
5019<x:ref>qdtext-nf</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5020 / %x5D-7E ; ']'-'~'
5021 / obs-text
5022<x:ref>query</x:ref> = &lt;query, defined in [RFC3986], Section 3.4&gt;
5023<x:ref>quoted-cpair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5024<x:ref>quoted-pair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5025<x:ref>quoted-str-nf</x:ref> = DQUOTE *( qdtext-nf / quoted-pair ) DQUOTE
5026<x:ref>quoted-string</x:ref> = DQUOTE *( qdtext / quoted-pair ) DQUOTE
5028<x:ref>rank</x:ref> = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] )
5029<x:ref>reason-phrase</x:ref> = *( HTAB / SP / VCHAR / obs-text )
5030<x:ref>received-by</x:ref> = ( uri-host [ ":" port ] ) / pseudonym
5031<x:ref>received-protocol</x:ref> = [ protocol-name "/" ] protocol-version
5032<x:ref>relative-part</x:ref> = &lt;relative-part, defined in [RFC3986], Section 4.2&gt;
5033<x:ref>request-line</x:ref> = method SP request-target SP HTTP-version CRLF
5034<x:ref>request-target</x:ref> = origin-form / absolute-form / authority-form /
5035 asterisk-form
5037<x:ref>special</x:ref> = "(" / ")" / "&lt;" / "&gt;" / "@" / "," / ";" / ":" / "\" /
5038 DQUOTE / "/" / "[" / "]" / "?" / "=" / "{" / "}"
5039<x:ref>start-line</x:ref> = request-line / status-line
5040<x:ref>status-code</x:ref> = 3DIGIT
5041<x:ref>status-line</x:ref> = HTTP-version SP status-code SP reason-phrase CRLF
5043<x:ref>t-codings</x:ref> = "trailers" / ( transfer-coding [ t-ranking ] )
5044<x:ref>t-ranking</x:ref> = OWS ";" OWS "q=" rank
5045<x:ref>tchar</x:ref> = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*" / "+" / "-" / "." /
5046 "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
5047<x:ref>token</x:ref> = 1*tchar
5048<x:ref>trailer-part</x:ref> = *( header-field CRLF )
5049<x:ref>transfer-coding</x:ref> = "chunked" / "compress" / "deflate" / "gzip" /
5050 transfer-extension
5051<x:ref>transfer-extension</x:ref> = token *( OWS ";" OWS transfer-parameter )
5052<x:ref>transfer-parameter</x:ref> = attribute BWS "=" BWS value
5054<x:ref>uri-host</x:ref> = &lt;host, defined in [RFC3986], Section 3.2.2&gt;
5056<x:ref>value</x:ref> = word
5058<x:ref>word</x:ref> = token / quoted-string
5062<?ENDINC p1-messaging.abnf-appendix ?>
5064<section title="Change Log (to be removed by RFC Editor before publication)" anchor="change.log">
5066<section title="Since RFC 2616">
5068  Changes up to the first Working Group Last Call draft are summarized
5069  in <eref target=""/>.
5073<section title="Since draft-ietf-httpbis-p1-messaging-21" anchor="changes.since.21">
5075  Closed issues:
5076  <list style="symbols">
5077    <t>
5078      <eref target=""/>:
5079      "Cite HTTPS URI scheme definition" (the spec now includes the HTTPs
5080      scheme definition and thus updates RFC 2818)
5081    </t>
5082    <t>
5083      <eref target=""/>:
5084      "mention of 'proxies' in section about caches"
5085    </t>
5086    <t>
5087      <eref target=""/>:
5088      "use of ABNF terms from RFC 3986"
5089    </t>
5090    <t>
5091      <eref target=""/>:
5092      "editorial improvements to message length definition"
5093    </t>
5094    <t>
5095      <eref target=""/>:
5096      "Connection header field MUST vs SHOULD"
5097    </t>
5098    <t>
5099      <eref target=""/>:
5100      "editorial improvements to persistent connections section"
5101    </t>
5102    <t>
5103      <eref target=""/>:
5104      "URI normalization vs empty path"
5105    </t>
5106    <t>
5107      <eref target=""/>:
5108      "p1 feedback"
5109    </t>
5110    <t>
5111      <eref target=""/>:
5112      "is parsing OBS-FOLD mandatory?"
5113    </t>
5114    <t>
5115      <eref target=""/>:
5116      "HTTPS and Shared Caching"
5117    </t>
5118    <t>
5119      <eref target=""/>:
5120      "Requirements for recipients of ws between start-line and first header field"
5121    </t>
5122    <t>
5123      <eref target=""/>:
5124      "SP and HT when being tolerant"
5125    </t>
5126    <t>
5127      <eref target=""/>:
5128      "Message Parsing Strictness"
5129    </t>
5130    <t>
5131      <eref target=""/>:
5132      "'Render'"
5133    </t>
5134    <t>
5135      <eref target=""/>:
5136      "No-Transform"
5137    </t>
5138    <t>
5139      <eref target=""/>:
5140      "p2 editorial feedback"
5141    </t>
5142    <t>
5143      <eref target=""/>:
5144      "Content-Length SHOULD be sent"
5145    </t>
5146  </list>
Note: See TracBrowser for help on using the repository browser.