source: draft-ietf-httpbis/latest/p1-messaging.xml @ 2651

Last change on this file since 2651 was 2651, checked in by julian.reschke@…, 9 years ago

"namespace" (#553)

  • Property svn:eol-style set to native
  • Property svn:mime-type set to text/xml
File size: 244.8 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 "May">
16  <!ENTITY ID-YEAR "2014">
17  <!ENTITY mdash "&#8212;">
18  <!ENTITY nbhy  "&#x2011;">
19  <!ENTITY Note "<x:h xmlns:x=''>Note:</x:h>">
20  <!ENTITY caching-overview       "<xref target='RFC7234' x:rel='#caching.overview' xmlns:x=''/>">
21  <!ENTITY cache-incomplete       "<xref target='RFC7234' x:rel='#response.cacheability' xmlns:x=''/>">
22  <!ENTITY cache-poisoning        "<xref target='RFC7234' x:rel='#security.considerations' xmlns:x=''/>">
23  <!ENTITY payload                "<xref target='RFC7231' x:rel='#payload' xmlns:x=''/>">
24  <!ENTITY media-type             "<xref target='RFC7231' x:rel='#media.type' xmlns:x=''/>">
25  <!ENTITY content-codings        "<xref target='RFC7231' x:rel='#content.codings' xmlns:x=''/>">
26  <!ENTITY CONNECT                "<xref target='RFC7231' x:rel='#CONNECT' xmlns:x=''/>">
27  <!ENTITY content.negotiation    "<xref target='RFC7231' x:rel='#content.negotiation' xmlns:x=''/>">
28  <!ENTITY diff-mime              "<xref target='RFC7231' x:rel='#differences.between.http.and.mime' xmlns:x=''/>">
29  <!ENTITY representation         "<xref target='RFC7231' x:rel='#representations' xmlns:x=''/>">
30  <!ENTITY GET                    "<xref target='RFC7231' x:rel='#GET' xmlns:x=''/>">
31  <!ENTITY HEAD                   "<xref target='RFC7231' x:rel='#HEAD' xmlns:x=''/>">
32  <!ENTITY header-allow           "<xref target='RFC7231' x:rel='#header.allow' xmlns:x=''/>">
33  <!ENTITY header-cache-control   "<xref target='RFC7234' x:rel='#header.cache-control' xmlns:x=''/>">
34  <!ENTITY header-content-encoding    "<xref target='RFC7231' x:rel='#header.content-encoding' xmlns:x=''/>">
35  <!ENTITY header-content-location    "<xref target='RFC7231' x:rel='#header.content-location' xmlns:x=''/>">
36  <!ENTITY header-content-range   "<xref target='RFC7233' x:rel='#header.content-range' xmlns:x=''/>">
37  <!ENTITY header-content-type    "<xref target='RFC7231' x:rel='#header.content-type' xmlns:x=''/>">
38  <!ENTITY header-date            "<xref target='RFC7231' x:rel='' xmlns:x=''/>">
39  <!ENTITY header-etag            "<xref target='RFC7232' x:rel='#header.etag' xmlns:x=''/>">
40  <!ENTITY header-expect          "<xref target='RFC7231' x:rel='#header.expect' xmlns:x=''/>">
41  <!ENTITY header-expires         "<xref target='RFC7234' x:rel='#header.expires' xmlns:x=''/>">
42  <!ENTITY header-last-modified   "<xref target='RFC7232' x:rel='#header.last-modified' xmlns:x=''/>">
43  <!ENTITY header-mime-version    "<xref target='RFC7231' x:rel='#mime-version' xmlns:x=''/>">
44  <!ENTITY header-pragma          "<xref target='RFC7234' x:rel='#header.pragma' xmlns:x=''/>">
45  <!ENTITY header-proxy-authenticate  "<xref target='RFC7235' x:rel='#header.proxy-authenticate' xmlns:x=''/>">
46  <!ENTITY header-proxy-authorization "<xref target='RFC7235' x:rel='#header.proxy-authorization' xmlns:x=''/>">
47  <!ENTITY header-server          "<xref target='RFC7231' x:rel='#header.server' xmlns:x=''/>">
48  <!ENTITY header-warning         "<xref target='RFC7234' x:rel='#header.warning' xmlns:x=''/>">
49  <!ENTITY idempotent-methods     "<xref target='RFC7231' x:rel='#idempotent.methods' xmlns:x=''/>">
50  <!ENTITY safe-methods           "<xref target='RFC7231' x:rel='#safe.methods' xmlns:x=''/>">
51  <!ENTITY methods                "<xref target='RFC7231' x:rel='#methods' xmlns:x=''/>">
52  <!ENTITY OPTIONS                "<xref target='RFC7231' x:rel='#OPTIONS' xmlns:x=''/>">
53  <!ENTITY qvalue                 "<xref target='RFC7231' x:rel='#quality.values' xmlns:x=''/>">
54  <!ENTITY request-header-fields  "<xref target='RFC7231' x:rel='#request.header.fields' xmlns:x=''/>">
55  <!ENTITY response-control-data  "<xref target='RFC7231' x:rel='' xmlns:x=''/>">
56  <!ENTITY resource               "<xref target='RFC7231' x:rel='#resources' xmlns:x=''/>">
57  <!ENTITY semantics              "<xref target='RFC7231' xmlns:x=''/>">
58  <!ENTITY status-codes           "<xref target='RFC7231' x:rel='' xmlns:x=''/>">
59  <!ENTITY status-1xx             "<xref target='RFC7231' x:rel='#status.1xx' xmlns:x=''/>">
60  <!ENTITY status-203             "<xref target='RFC7231' x:rel='#status.203' xmlns:x=''/>">
61  <!ENTITY status-3xx             "<xref target='RFC7231' x:rel='#status.3xx' xmlns:x=''/>">
62  <!ENTITY status-304             "<xref target='RFC7232' x:rel='#status.304' xmlns:x=''/>">
63  <!ENTITY status-4xx             "<xref target='RFC7231' x:rel='#status.4xx' xmlns:x=''/>">
64  <!ENTITY status-413             "<xref target='RFC7231' x:rel='#status.413' xmlns:x=''/>">
65  <!ENTITY status-414             "<xref target='RFC7231' x:rel='#status.414' xmlns:x=''/>">
66  <!ENTITY iana-header-registry   "<xref target='RFC7231' x:rel='#header.field.registry' xmlns:x=''/>">
68<?rfc toc="yes" ?>
69<?rfc symrefs="yes" ?>
70<?rfc sortrefs="yes" ?>
71<?rfc compact="yes"?>
72<?rfc subcompact="no" ?>
73<?rfc linkmailto="no" ?>
74<?rfc editing="no" ?>
75<?rfc comments="yes"?>
76<?rfc inline="yes"?>
77<?rfc rfcedstyle="yes"?>
78<?rfc-ext allow-markup-in-artwork="yes" ?>
79<?rfc-ext include-references-in-index="yes" ?>
80<rfc obsoletes="2145, 2616" updates="2817, 2818" category="std" x:maturity-level="proposed"
81     ipr="pre5378Trust200902" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
82     xmlns:x=''>
83<x:link rel="next" basename="p2-semantics"/>
84<x:feedback template="{docname},%20%22{section}%22&amp;body=&lt;{ref}&gt;:"/>
87  <title abbrev="HTTP/1.1 Message Syntax and Routing">Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing</title>
89  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
90    <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
91    <address>
92      <postal>
93        <street>345 Park Ave</street>
94        <city>San Jose</city>
95        <region>CA</region>
96        <code>95110</code>
97        <country>USA</country>
98      </postal>
99      <email></email>
100      <uri></uri>
101    </address>
102  </author>
104  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
105    <organization abbrev="greenbytes">greenbytes GmbH</organization>
106    <address>
107      <postal>
108        <street>Hafenweg 16</street>
109        <city>Muenster</city><region>NW</region><code>48155</code>
110        <country>Germany</country>
111      </postal>
112      <email></email>
113      <uri></uri>
114    </address>
115  </author>
117  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
119  <area>Applications</area>
120  <workgroup>HTTPbis</workgroup>
124   The Hypertext Transfer Protocol (HTTP) is a stateless application-level
125   protocol for distributed, collaborative, hypertext information systems.
126   This document provides an overview of HTTP architecture and its associated
127   terminology, defines the "http" and "https" Uniform Resource Identifier
128   (URI) schemes, defines the HTTP/1.1 message syntax and parsing
129   requirements, and describes related security concerns for implementations.
133<note title="Editorial Note (To be removed by RFC Editor)">
134  <t>
135    Discussion of this draft takes place on the HTTPBIS working group
136    mailing list (, which is archived at
137    <eref target=""/>.
138  </t>
139  <t>
140    The current issues list is at
141    <eref target=""/> and related
142    documents (including fancy diffs) can be found at
143    <eref target=""/>.
144  </t>
145  <t>
146    <spanx>This is a temporary document for the purpose of tracking the editorial changes made during the AUTH48 (RFC publication) phase.</spanx>
147  </t>
151<section title="Introduction" anchor="introduction">
153   The Hypertext Transfer Protocol (HTTP) is a stateless application-level
154   request/response protocol that uses extensible semantics and
155   self-descriptive message payloads for flexible interaction with
156   network-based hypertext information systems. This document is the first in
157   a series of documents that collectively form the HTTP/1.1 specification:
158   <list style="empty">
159    <t>RFC 7230: Message Syntax and Routing</t>
160    <t><xref target="RFC7231" x:fmt="none">RFC 7231</xref>: Semantics and Content</t>
161    <t><xref target="RFC7232" x:fmt="none">RFC 7232</xref>: Conditional Requests</t>
162    <t><xref target="RFC7233" x:fmt="none">RFC 7233</xref>: Range Requests</t>
163    <t><xref target="RFC7234" x:fmt="none">RFC 7234</xref>: Caching</t>
164    <t><xref target="RFC7234" x:fmt="none">RFC 7235</xref>: Authentication</t>
165   </list>
168   This HTTP/1.1 specification obsoletes
169   <xref target="RFC2616" x:fmt="none">RFC 2616</xref> and
170   <xref target="RFC2145" x:fmt="none">RFC 2145</xref> (on HTTP versioning).
171   This specification also updates the use of CONNECT to establish a tunnel,
172   previously defined in <xref target="RFC2817" x:fmt="none">RFC 2817</xref>,
173   and defines the "https" URI scheme that was described informally in
174   <xref target="RFC2818" x:fmt="none">RFC 2818</xref>.
177   HTTP is a generic interface protocol for information systems. It is
178   designed to hide the details of how a service is implemented by presenting
179   a uniform interface to clients that is independent of the types of
180   resources provided. Likewise, servers do not need to be aware of each
181   client's purpose: an HTTP request can be considered in isolation rather
182   than being associated with a specific type of client or a predetermined
183   sequence of application steps. The result is a protocol that can be used
184   effectively in many different contexts and for which implementations can
185   evolve independently over time.
188   HTTP is also designed for use as an intermediation protocol for translating
189   communication to and from non-HTTP information systems.
190   HTTP proxies and gateways can provide access to alternative information
191   services by translating their diverse protocols into a hypertext
192   format that can be viewed and manipulated by clients in the same way
193   as HTTP services.
196   One consequence of this flexibility is that the protocol cannot be
197   defined in terms of what occurs behind the interface. Instead, we
198   are limited to defining the syntax of communication, the intent
199   of received communication, and the expected behavior of recipients.
200   If the communication is considered in isolation, then successful
201   actions ought to be reflected in corresponding changes to the
202   observable interface provided by servers. However, since multiple
203   clients might act in parallel and perhaps at cross-purposes, we
204   cannot require that such changes be observable beyond the scope
205   of a single response.
208   This document describes the architectural elements that are used or
209   referred to in HTTP, defines the "http" and "https" URI schemes,
210   describes overall network operation and connection management,
211   and defines HTTP message framing and forwarding requirements.
212   Our goal is to define all of the mechanisms necessary for HTTP message
213   handling that are independent of message semantics, thereby defining the
214   complete set of requirements for message parsers and
215   message-forwarding intermediaries.
219<section title="Requirements Notation" anchor="intro.requirements">
221   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
222   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
223   document are to be interpreted as described in <xref target="RFC2119"/>.
226   Conformance criteria and considerations regarding error handling
227   are defined in <xref target="conformance"/>.
231<section title="Syntax Notation" anchor="notation">
232<iref primary="true" item="Grammar" subitem="ALPHA"/>
233<iref primary="true" item="Grammar" subitem="CR"/>
234<iref primary="true" item="Grammar" subitem="CRLF"/>
235<iref primary="true" item="Grammar" subitem="CTL"/>
236<iref primary="true" item="Grammar" subitem="DIGIT"/>
237<iref primary="true" item="Grammar" subitem="DQUOTE"/>
238<iref primary="true" item="Grammar" subitem="HEXDIG"/>
239<iref primary="true" item="Grammar" subitem="HTAB"/>
240<iref primary="true" item="Grammar" subitem="LF"/>
241<iref primary="true" item="Grammar" subitem="OCTET"/>
242<iref primary="true" item="Grammar" subitem="SP"/>
243<iref primary="true" item="Grammar" subitem="VCHAR"/>
245   This specification uses the Augmented Backus-Naur Form (ABNF) notation of
246   <xref target="RFC5234"/> with a list extension, defined in
247   <xref target="abnf.extension"/>, that allows for compact definition of
248   comma-separated lists using a '#' operator (similar to how the '*' operator
249   indicates repetition).
250   <xref target="collected.abnf"/> shows the collected grammar with all list
251   operators expanded to standard ABNF notation.
253<t anchor="core.rules">
254  <x:anchor-alias value="ALPHA"/>
255  <x:anchor-alias value="CTL"/>
256  <x:anchor-alias value="CR"/>
257  <x:anchor-alias value="CRLF"/>
258  <x:anchor-alias value="DIGIT"/>
259  <x:anchor-alias value="DQUOTE"/>
260  <x:anchor-alias value="HEXDIG"/>
261  <x:anchor-alias value="HTAB"/>
262  <x:anchor-alias value="LF"/>
263  <x:anchor-alias value="OCTET"/>
264  <x:anchor-alias value="SP"/>
265  <x:anchor-alias value="VCHAR"/>
266   The following core rules are included by
267   reference, as defined in <xref target="RFC5234" x:fmt="," x:sec="B.1"/>:
268   ALPHA (letters), CR (carriage return), CRLF (CR LF), CTL (controls),
269   DIGIT (decimal 0-9), DQUOTE (double quote),
270   HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line feed),
271   OCTET (any 8-bit sequence of data), SP (space), and
272   VCHAR (any visible <xref target="USASCII"/> character).
275   As a convention, ABNF rule names prefixed with "obs-" denote
276   "obsolete" grammar rules that appear for historical reasons.
281<section title="Architecture" anchor="architecture">
283   HTTP was created for the World Wide Web (WWW) architecture
284   and has evolved over time to support the scalability needs of a worldwide
285   hypertext system. Much of that architecture is reflected in the terminology
286   and syntax productions used to define HTTP.
289<section title="Client/Server Messaging" anchor="operation">
290<iref primary="true" item="client"/>
291<iref primary="true" item="server"/>
292<iref primary="true" item="connection"/>
294   HTTP is a stateless request/response protocol that operates by exchanging
295   <x:dfn>messages</x:dfn> (<xref target="http.message"/>) across a reliable
296   transport or session-layer
297   "<x:dfn>connection</x:dfn>" (<xref target=""/>).
298   An HTTP "<x:dfn>client</x:dfn>" is a program that establishes a connection
299   to a server for the purpose of sending one or more HTTP requests.
300   An HTTP "<x:dfn>server</x:dfn>" is a program that accepts connections
301   in order to service HTTP requests by sending HTTP responses.
303<iref primary="true" item="user agent"/>
304<iref primary="true" item="origin server"/>
305<iref primary="true" item="browser"/>
306<iref primary="true" item="spider"/>
307<iref primary="true" item="sender"/>
308<iref primary="true" item="recipient"/>
310   The terms "client" and "server" refer only to the roles that
311   these programs perform for a particular connection.  The same program
312   might act as a client on some connections and a server on others.
313   The term "<x:dfn>user agent</x:dfn>" refers to any of the various
314   client programs that initiate a request, including (but not limited to)
315   browsers, spiders (web-based robots), command-line tools, custom
316   applications, and mobile apps.
317   The term "<x:dfn>origin server</x:dfn>" refers to the program that can
318   originate authoritative responses for a given target resource.
319   The terms "<x:dfn>sender</x:dfn>" and "<x:dfn>recipient</x:dfn>" refer to
320   any implementation that sends or receives a given message, respectively.
323   HTTP relies upon the Uniform Resource Identifier (URI)
324   standard <xref target="RFC3986"/> to indicate the target resource
325   (<xref target="target-resource"/>) and relationships between resources.
326   Messages are passed in a format similar to that used by Internet mail
327   <xref target="RFC5322"/> and the Multipurpose Internet Mail Extensions
328   (MIME) <xref target="RFC2045"/> (see &diff-mime; for the differences
329   between HTTP and MIME messages).
332   Most HTTP communication consists of a retrieval request (GET) for
333   a representation of some resource identified by a URI.  In the
334   simplest case, this might be accomplished via a single bidirectional
335   connection (===) between the user agent (UA) and the origin server (O).
337<figure><artwork type="drawing">
338         request   &gt;
339    <x:highlight>UA</x:highlight> ======================================= <x:highlight>O</x:highlight>
340                                &lt;   response
342<iref primary="true" item="message"/>
343<iref primary="true" item="request"/>
344<iref primary="true" item="response"/>
346   A client sends an HTTP request to a server in the form of a <x:dfn>request</x:dfn>
347   message, beginning with a request-line that includes a method, URI, and
348   protocol version (<xref target="request.line"/>),
349   followed by header fields containing
350   request modifiers, client information, and representation metadata
351   (<xref target="header.fields"/>),
352   an empty line to indicate the end of the header section, and finally
353   a message body containing the payload body (if any,
354   <xref target="message.body"/>).
357   A server responds to a client's request by sending one or more HTTP
358   <x:dfn>response</x:dfn>
359   messages, each beginning with a status line that
360   includes the protocol version, a success or error code, and textual
361   reason phrase (<xref target="status.line"/>),
362   possibly followed by header fields containing server
363   information, resource metadata, and representation metadata
364   (<xref target="header.fields"/>),
365   an empty line to indicate the end of the header section, and finally
366   a message body containing the payload body (if any,
367   <xref target="message.body"/>).
370   A connection might be used for multiple request/response exchanges,
371   as defined in <xref target="persistent.connections"/>.
374   The following example illustrates a typical message exchange for a
375   GET request (&GET;) on the URI "":
378Client request:
379</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
380GET /hello.txt HTTP/1.1
381User-Agent: curl/7.16.3 libcurl/7.16.3 OpenSSL/0.9.7l zlib/1.2.3
383Accept-Language: en, mi
387Server response:
388</preamble><artwork type="message/http; msgtype=&#34;response&#34;" x:indent-with="  ">
389HTTP/1.1 200 OK
390Date: Mon, 27 Jul 2009 12:28:53 GMT
391Server: Apache
392Last-Modified: Wed, 22 Jul 2009 19:15:56 GMT
393ETag: "34aa387-d-1568eb00"
394Accept-Ranges: bytes
395Content-Length: <x:length-of target="exbody"/>
396Vary: Accept-Encoding
397Content-Type: text/plain
399<x:span anchor="exbody">Hello World! My payload includes a trailing CRLF.
404<section title="Implementation Diversity" anchor="implementation-diversity">
406   When considering the design of HTTP, it is easy to fall into a trap of
407   thinking that all user agents are general-purpose browsers and all origin
408   servers are large public websites. That is not the case in practice.
409   Common HTTP user agents include household appliances, stereos, scales,
410   firmware update scripts, command-line programs, mobile apps,
411   and communication devices in a multitude of shapes and sizes.  Likewise,
412   common HTTP origin servers include home automation units, configurable
413   networking components, office machines, autonomous robots, news feeds,
414   traffic cameras, ad selectors, and video-delivery platforms.
417   The term "user agent" does not imply that there is a human user directly
418   interacting with the software agent at the time of a request. In many
419   cases, a user agent is installed or configured to run in the background
420   and save its results for later inspection (or save only a subset of those
421   results that might be interesting or erroneous). Spiders, for example, are
422   typically given a start URI and configured to follow certain behavior while
423   crawling the Web as a hypertext graph.
426   The implementation diversity of HTTP means that not all user agents can
427   make interactive suggestions to their user or provide adequate warning for
428   security or privacy concerns. In the few cases where this
429   specification requires reporting of errors to the user, it is acceptable
430   for such reporting to only be observable in an error console or log file.
431   Likewise, requirements that an automated action be confirmed by the user
432   before proceeding might be met via advance configuration choices,
433   run-time options, or simple avoidance of the unsafe action; confirmation
434   does not imply any specific user interface or interruption of normal
435   processing if the user has already made that choice.
439<section title="Intermediaries" anchor="intermediaries">
440<iref primary="true" item="intermediary"/>
442   HTTP enables the use of intermediaries to satisfy requests through
443   a chain of connections.  There are three common forms of HTTP
444   <x:dfn>intermediary</x:dfn>: proxy, gateway, and tunnel.  In some cases,
445   a single intermediary might act as an origin server, proxy, gateway,
446   or tunnel, switching behavior based on the nature of each request.
448<figure><artwork type="drawing">
449         &gt;             &gt;             &gt;             &gt;
450    <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>
451               &lt;             &lt;             &lt;             &lt;
454   The figure above shows three intermediaries (A, B, and C) between the
455   user agent and origin server. A request or response message that
456   travels the whole chain will pass through four separate connections.
457   Some HTTP communication options
458   might apply only to the connection with the nearest, non-tunnel
459   neighbor, only to the endpoints of the chain, or to all connections
460   along the chain. Although the diagram is linear, each participant might
461   be engaged in multiple, simultaneous communications. For example, B
462   might be receiving requests from many clients other than A, and/or
463   forwarding requests to servers other than C, at the same time that it
464   is handling A's request. Likewise, later requests might be sent through a
465   different path of connections, often based on dynamic configuration for
466   load balancing.   
469<iref primary="true" item="upstream"/><iref primary="true" item="downstream"/>
470<iref primary="true" item="inbound"/><iref primary="true" item="outbound"/>
471   The terms "<x:dfn>upstream</x:dfn>" and "<x:dfn>downstream</x:dfn>" are
472   used to describe directional requirements in relation to the message flow:
473   all messages flow from upstream to downstream.
474   The terms "inbound" and "outbound" are used to describe directional
475   requirements in relation to the request route:
476   "<x:dfn>inbound</x:dfn>" means toward the origin server and
477   "<x:dfn>outbound</x:dfn>" means toward the user agent.
479<t><iref primary="true" item="proxy"/>
480   A "<x:dfn>proxy</x:dfn>" is a message-forwarding agent that is selected by the
481   client, usually via local configuration rules, to receive requests
482   for some type(s) of absolute URI and attempt to satisfy those
483   requests via translation through the HTTP interface.  Some translations
484   are minimal, such as for proxy requests for "http" URIs, whereas
485   other requests might require translation to and from entirely different
486   application-level protocols. Proxies are often used to group an
487   organization's HTTP requests through a common intermediary for the
488   sake of security, annotation services, or shared caching. Some proxies
489   are designed to apply transformations to selected messages or payloads
490   while they are being forwarded, as described in
491   <xref target="message.transformations"/>.
493<t><iref primary="true" item="gateway"/><iref primary="true" item="reverse proxy"/>
494<iref primary="true" item="accelerator"/>
495   A "<x:dfn>gateway</x:dfn>" (a.k.a. "<x:dfn>reverse proxy</x:dfn>") is an
496   intermediary that acts as an origin server for the outbound connection but
497   translates received requests and forwards them inbound to another server or
498   servers. Gateways are often used to encapsulate legacy or untrusted
499   information services, to improve server performance through
500   "<x:dfn>accelerator</x:dfn>" caching, and to enable partitioning or load
501   balancing of HTTP services across multiple machines.
504   All HTTP requirements applicable to an origin server
505   also apply to the outbound communication of a gateway.
506   A gateway communicates with inbound servers using any protocol that
507   it desires, including private extensions to HTTP that are outside
508   the scope of this specification.  However, an HTTP-to-HTTP gateway
509   that wishes to interoperate with third-party HTTP servers ought to conform
510   to user agent requirements on the gateway's inbound connection.
512<t><iref primary="true" item="tunnel"/>
513   A "<x:dfn>tunnel</x:dfn>" acts as a blind relay between two connections
514   without changing the messages. Once active, a tunnel is not
515   considered a party to the HTTP communication, though the tunnel might
516   have been initiated by an HTTP request. A tunnel ceases to exist when
517   both ends of the relayed connection are closed. Tunnels are used to
518   extend a virtual connection through an intermediary, such as when
519   Transport Layer Security (TLS, <xref target="RFC5246"/>) is used to
520   establish confidential communication through a shared firewall proxy.
523   The above categories for intermediary only consider those acting as
524   participants in the HTTP communication.  There are also intermediaries
525   that can act on lower layers of the network protocol stack, filtering or
526   redirecting HTTP traffic without the knowledge or permission of message
527   senders. Network intermediaries are indistinguishable (at a protocol level)
528   from a man-in-the-middle attack, often introducing security flaws or
529   interoperability problems due to mistakenly violating HTTP semantics.
531<t><iref primary="true" item="interception proxy"/>
532<iref primary="true" item="transparent proxy"/>
533<iref primary="true" item="captive portal"/>
534   For example, an
535   "<x:dfn>interception proxy</x:dfn>" <xref target="RFC3040"/> (also commonly
536   known as a "<x:dfn>transparent proxy</x:dfn>" <xref target="RFC1919"/> or
537   "<x:dfn>captive portal</x:dfn>")
538   differs from an HTTP proxy because it is not selected by the client.
539   Instead, an interception proxy filters or redirects outgoing TCP port 80
540   packets (and occasionally other common port traffic).
541   Interception proxies are commonly found on public network access points,
542   as a means of enforcing account subscription prior to allowing use of
543   non-local Internet services, and within corporate firewalls to enforce
544   network usage policies.
547   HTTP is defined as a stateless protocol, meaning that each request message
548   can be understood in isolation.  Many implementations depend on HTTP's
549   stateless design in order to reuse proxied connections or dynamically
550   load-balance requests across multiple servers.  Hence, a server &MUST-NOT;
551   assume that two requests on the same connection are from the same user
552   agent unless the connection is secured and specific to that agent.
553   Some non-standard HTTP extensions (e.g., <xref target="RFC4559"/>) have
554   been known to violate this requirement, resulting in security and
555   interoperability problems.
559<section title="Caches" anchor="caches">
560<iref primary="true" item="cache"/>
562   A "<x:dfn>cache</x:dfn>" is a local store of previous response messages and the
563   subsystem that controls its message storage, retrieval, and deletion.
564   A cache stores cacheable responses in order to reduce the response
565   time and network bandwidth consumption on future, equivalent
566   requests. Any client or server &MAY; employ a cache, though a cache
567   cannot be used by a server while it is acting as a tunnel.
570   The effect of a cache is that the request/response chain is shortened
571   if one of the participants along the chain has a cached response
572   applicable to that request. The following illustrates the resulting
573   chain if B has a cached copy of an earlier response from O (via C)
574   for a request that has not been cached by UA or A.
576<figure><artwork type="drawing">
577            &gt;             &gt;
578       <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>
579                  &lt;             &lt;
581<t><iref primary="true" item="cacheable"/>
582   A response is "<x:dfn>cacheable</x:dfn>" if a cache is allowed to store a copy of
583   the response message for use in answering subsequent requests.
584   Even when a response is cacheable, there might be additional
585   constraints placed by the client or by the origin server on when
586   that cached response can be used for a particular request. HTTP
587   requirements for cache behavior and cacheable responses are
588   defined in &caching-overview;. 
591   There is a wide variety of architectures and configurations
592   of caches deployed across the World Wide Web and
593   inside large organizations. These include national hierarchies
594   of proxy caches to save transoceanic bandwidth, collaborative systems that
595   broadcast or multicast cache entries, archives of pre-fetched cache
596   entries for use in off-line or high-latency environments, and so on.
600<section title="Conformance and Error Handling" anchor="conformance">
602   This specification targets conformance criteria according to the role of
603   a participant in HTTP communication.  Hence, HTTP requirements are placed
604   on senders, recipients, clients, servers, user agents, intermediaries,
605   origin servers, proxies, gateways, or caches, depending on what behavior
606   is being constrained by the requirement. Additional (social) requirements
607   are placed on implementations, resource owners, and protocol element
608   registrations when they apply beyond the scope of a single communication.
611   The verb "generate" is used instead of "send" where a requirement
612   differentiates between creating a protocol element and merely forwarding a
613   received element downstream.
616   An implementation is considered conformant if it complies with all of the
617   requirements associated with the roles it partakes in HTTP.
620   Conformance includes both the syntax and semantics of protocol
621   elements. A sender &MUST-NOT; generate protocol elements that convey a
622   meaning that is known by that sender to be false. A sender &MUST-NOT;
623   generate protocol elements that do not match the grammar defined by the
624   corresponding ABNF rules. Within a given message, a sender &MUST-NOT;
625   generate protocol elements or syntax alternatives that are only allowed to
626   be generated by participants in other roles (i.e., a role that the sender
627   does not have for that message).
630   When a received protocol element is parsed, the recipient &MUST; be able to
631   parse any value of reasonable length that is applicable to the recipient's
632   role and that matches the grammar defined by the corresponding ABNF rules.
633   Note, however, that some received protocol elements might not be parsed.
634   For example, an intermediary forwarding a message might parse a
635   header-field into generic field-name and field-value components, but then
636   forward the header field without further parsing inside the field-value.
639   HTTP does not have specific length limitations for many of its protocol
640   elements because the lengths that might be appropriate will vary widely,
641   depending on the deployment context and purpose of the implementation.
642   Hence, interoperability between senders and recipients depends on shared
643   expectations regarding what is a reasonable length for each protocol
644   element. Furthermore, what is commonly understood to be a reasonable length
645   for some protocol elements has changed over the course of the past two
646   decades of HTTP use and is expected to continue changing in the future.
649   At a minimum, a recipient &MUST; be able to parse and process protocol
650   element lengths that are at least as long as the values that it generates
651   for those same protocol elements in other messages. For example, an origin
652   server that publishes very long URI references to its own resources needs
653   to be able to parse and process those same references when received as a
654   request target.
657   A recipient &MUST; interpret a received protocol element according to the
658   semantics defined for it by this specification, including extensions to
659   this specification, unless the recipient has determined (through experience
660   or configuration) that the sender incorrectly implements what is implied by
661   those semantics.
662   For example, an origin server might disregard the contents of a received
663   <x:ref>Accept-Encoding</x:ref> header field if inspection of the
664   <x:ref>User-Agent</x:ref> header field indicates a specific implementation
665   version that is known to fail on receipt of certain content codings.
668   Unless noted otherwise, a recipient &MAY; attempt to recover a usable
669   protocol element from an invalid construct.  HTTP does not define
670   specific error handling mechanisms except when they have a direct impact
671   on security, since different applications of the protocol require
672   different error handling strategies.  For example, a Web browser might
673   wish to transparently recover from a response where the
674   <x:ref>Location</x:ref> header field doesn't parse according to the ABNF,
675   whereas a systems control client might consider any form of error recovery
676   to be dangerous.
680<section title="Protocol Versioning" anchor="http.version">
681  <x:anchor-alias value="HTTP-version"/>
682  <x:anchor-alias value="HTTP-name"/>
684   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate
685   versions of the protocol. This specification defines version "1.1".
686   The protocol version as a whole indicates the sender's conformance
687   with the set of requirements laid out in that version's corresponding
688   specification of HTTP.
691   The version of an HTTP message is indicated by an HTTP-version field
692   in the first line of the message. HTTP-version is case-sensitive.
694<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-version"/><iref primary="true" item="Grammar" subitem="HTTP-name"/>
695  <x:ref>HTTP-version</x:ref>  = <x:ref>HTTP-name</x:ref> "/" <x:ref>DIGIT</x:ref> "." <x:ref>DIGIT</x:ref>
696  <x:ref>HTTP-name</x:ref>     = <x:abnf-char-sequence>"HTTP"</x:abnf-char-sequence> ; "HTTP", case-sensitive
699   The HTTP version number consists of two decimal digits separated by a "."
700   (period or decimal point).  The first digit ("major version") indicates the
701   HTTP messaging syntax, whereas the second digit ("minor version") indicates
702   the highest minor version within that major version to which the sender is
703   conformant and able to understand for future communication.  The minor
704   version advertises the sender's communication capabilities even when the
705   sender is only using a backwards-compatible subset of the protocol,
706   thereby letting the recipient know that more advanced features can
707   be used in response (by servers) or in future requests (by clients).
710   When an HTTP/1.1 message is sent to an HTTP/1.0 recipient
711   <xref target="RFC1945"/> or a recipient whose version is unknown,
712   the HTTP/1.1 message is constructed such that it can be interpreted
713   as a valid HTTP/1.0 message if all of the newer features are ignored.
714   This specification places recipient-version requirements on some
715   new features so that a conformant sender will only use compatible
716   features until it has determined, through configuration or the
717   receipt of a message, that the recipient supports HTTP/1.1.
720   The interpretation of a header field does not change between minor
721   versions of the same major HTTP version, though the default
722   behavior of a recipient in the absence of such a field can change.
723   Unless specified otherwise, header fields defined in HTTP/1.1 are
724   defined for all versions of HTTP/1.x.  In particular, the <x:ref>Host</x:ref>
725   and <x:ref>Connection</x:ref> header fields ought to be implemented by all
726   HTTP/1.x implementations whether or not they advertise conformance with
727   HTTP/1.1.
730   New header fields can be introduced without changing the protocol version
731   if their defined semantics allow them to be safely ignored by recipients
732   that do not recognize them. Header field extensibility is discussed in
733   <xref target="field.extensibility"/>.
736   Intermediaries that process HTTP messages (i.e., all intermediaries
737   other than those acting as tunnels) &MUST; send their own HTTP-version
738   in forwarded messages.  In other words, they are not allowed to blindly
739   forward the first line of an HTTP message without ensuring that the
740   protocol version in that message matches a version to which that
741   intermediary is conformant for both the receiving and
742   sending of messages.  Forwarding an HTTP message without rewriting
743   the HTTP-version might result in communication errors when downstream
744   recipients use the message sender's version to determine what features
745   are safe to use for later communication with that sender.
748   A client &SHOULD; send a request version equal to the highest
749   version to which the client is conformant and
750   whose major version is no higher than the highest version supported
751   by the server, if this is known.  A client &MUST-NOT; send a
752   version to which it is not conformant.
755   A client &MAY; send a lower request version if it is known that
756   the server incorrectly implements the HTTP specification, but only
757   after the client has attempted at least one normal request and determined
758   from the response status code or header fields (e.g., <x:ref>Server</x:ref>) that
759   the server improperly handles higher request versions.
762   A server &SHOULD; send a response version equal to the highest version to
763   which the server is conformant that has a major version less than or equal
764   to the one received in the request.
765   A server &MUST-NOT; send a version to which it is not conformant.
766   A server can send a <x:ref>505 (HTTP Version Not Supported)</x:ref>
767   response if it wishes, for any reason, to refuse service of the client's
768   major protocol version.
771   A server &MAY; send an HTTP/1.0 response to a request
772   if it is known or suspected that the client incorrectly implements the
773   HTTP specification and is incapable of correctly processing later
774   version responses, such as when a client fails to parse the version
775   number correctly or when an intermediary is known to blindly forward
776   the HTTP-version even when it doesn't conform to the given minor
777   version of the protocol. Such protocol downgrades &SHOULD-NOT; be
778   performed unless triggered by specific client attributes, such as when
779   one or more of the request header fields (e.g., <x:ref>User-Agent</x:ref>)
780   uniquely match the values sent by a client known to be in error.
783   The intention of HTTP's versioning design is that the major number
784   will only be incremented if an incompatible message syntax is
785   introduced, and that the minor number will only be incremented when
786   changes made to the protocol have the effect of adding to the message
787   semantics or implying additional capabilities of the sender.  However,
788   the minor version was not incremented for the changes introduced between
789   <xref target="RFC2068"/> and <xref target="RFC2616"/>, and this revision
790   has specifically avoided any such changes to the protocol.
793   When an HTTP message is received with a major version number that the
794   recipient implements, but a higher minor version number than what the
795   recipient implements, the recipient &SHOULD; process the message as if it
796   were in the highest minor version within that major version to which the
797   recipient is conformant. A recipient can assume that a message with a
798   higher minor version, when sent to a recipient that has not yet indicated
799   support for that higher version, is sufficiently backwards-compatible to be
800   safely processed by any implementation of the same major version.
804<section title="Uniform Resource Identifiers" anchor="uri">
805<iref primary="true" item="resource"/>
807   Uniform Resource Identifiers (URIs) <xref target="RFC3986"/> are used
808   throughout HTTP as the means for identifying resources (&resource;).
809   URI references are used to target requests, indicate redirects, and define
810   relationships.
812  <x:anchor-alias value="URI-reference"/>
813  <x:anchor-alias value="absolute-URI"/>
814  <x:anchor-alias value="relative-part"/>
815  <x:anchor-alias value="scheme"/>
816  <x:anchor-alias value="authority"/>
817  <x:anchor-alias value="uri-host"/>
818  <x:anchor-alias value="port"/>
819  <x:anchor-alias value="path"/>
820  <x:anchor-alias value="path-abempty"/>
821  <x:anchor-alias value="segment"/>
822  <x:anchor-alias value="query"/>
823  <x:anchor-alias value="fragment"/>
824  <x:anchor-alias value="absolute-path"/>
825  <x:anchor-alias value="partial-URI"/>
827   The definitions of "URI-reference",
828   "absolute-URI", "relative-part", "scheme", "authority", "port", "host",
829   "path-abempty", "segment", "query", and "fragment" are adopted from the
830   URI generic syntax.
831   An "absolute-path" rule is defined for protocol elements that can contain a
832   non-empty path component. (This rule differs slightly from RFC 3986's
833   path-abempty rule, which allows for an empty path to be used in references,
834   and path-absolute rule, which does not allow paths that begin with "//".)
835   A "partial-URI" rule is defined for protocol elements
836   that can contain a relative URI but not a fragment component.
838<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="scheme"/><iref primary="true" item="Grammar" subitem="authority"/><iref primary="true" item="Grammar" subitem="absolute-path"/><iref primary="true" item="Grammar" subitem="port"/><iref primary="true" item="Grammar" subitem="query"/><iref primary="true" item="Grammar" subitem="fragment"/><iref primary="true" item="Grammar" subitem="segment"/><iref primary="true" item="Grammar" subitem="uri-host"/><iref primary="true" item="Grammar" subitem="partial-URI"><!--exported production--></iref>
839  <x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in <xref target="RFC3986" x:fmt="," x:sec="4.1"/>&gt;
840  <x:ref>absolute-URI</x:ref>  = &lt;absolute-URI, defined in <xref target="RFC3986" x:fmt="," x:sec="4.3"/>&gt;
841  <x:ref>relative-part</x:ref> = &lt;relative-part, defined in <xref target="RFC3986" x:fmt="," x:sec="4.2"/>&gt;
842  <x:ref>scheme</x:ref>        = &lt;scheme, defined in <xref target="RFC3986" x:fmt="," x:sec="3.1"/>&gt;
843  <x:ref>authority</x:ref>     = &lt;authority, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2"/>&gt;
844  <x:ref>uri-host</x:ref>      = &lt;host, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>&gt;
845  <x:ref>port</x:ref>          = &lt;port, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.3"/>&gt;
846  <x:ref>path-abempty</x:ref>  = &lt;path-abempty, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
847  <x:ref>segment</x:ref>       = &lt;segment, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
848  <x:ref>query</x:ref>         = &lt;query, defined in <xref target="RFC3986" x:fmt="," x:sec="3.4"/>&gt;
849  <x:ref>fragment</x:ref>      = &lt;fragment, defined in <xref target="RFC3986" x:fmt="," x:sec="3.5"/>&gt;
851  <x:ref>absolute-path</x:ref> = 1*( "/" segment )
852  <x:ref>partial-URI</x:ref>   = relative-part [ "?" query ]
855   Each protocol element in HTTP that allows a URI reference will indicate
856   in its ABNF production whether the element allows any form of reference
857   (URI-reference), only a URI in absolute form (absolute-URI), only the
858   path and optional query components, or some combination of the above.
859   Unless otherwise indicated, URI references are parsed
860   relative to the effective request URI
861   (<xref target="effective.request.uri"/>).
864<section title="http URI Scheme" anchor="http.uri">
865  <x:anchor-alias value="http-URI"/>
866  <iref item="http URI scheme" primary="true"/>
867  <iref item="URI scheme" subitem="http" primary="true"/>
869   The "http" URI scheme is hereby defined for the purpose of minting
870   identifiers according to their association with the hierarchical
871   namespace governed by a potential HTTP origin server listening for
872   TCP (<xref target="RFC0793"/>) connections on a given port.
874<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http-URI"><!--terminal production--></iref>
875  <x:ref>http-URI</x:ref> = "http:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
876             [ "#" <x:ref>fragment</x:ref> ]
879   The origin server for an "http" URI is identified by the
880   <x:ref>authority</x:ref> component, which includes a host identifier
881   and optional TCP port (<xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>).
882   The hierarchical path component and optional query component serve as an
883   identifier for a potential target resource within that origin server's name
884   space. The optional fragment component allows for indirect identification
885   of a secondary resource, independent of the URI scheme, as defined in
886   <xref target="RFC3986" x:fmt="of" x:sec="3.5"/>.
889   A sender &MUST-NOT; generate an "http" URI with an empty host identifier.
890   A recipient that processes such a URI reference &MUST; reject it as invalid.
893   If the host identifier is provided as an IP address, the origin server is
894   the listener (if any) on the indicated TCP port at that IP address.
895   If host is a registered name, the registered name is an indirect identifier
896   for use with a name resolution service, such as DNS, to find an address for
897   that origin server.
898   If the port subcomponent is empty or not given, TCP port 80 (the
899   reserved port for WWW services) is the default.
902   Note that the presence of a URI with a given authority component does not
903   imply that there is always an HTTP server listening for connections on
904   that host and port. Anyone can mint a URI. What the authority component
905   determines is who has the right to respond authoritatively to requests that
906   target the identified resource. The delegated nature of registered names
907   and IP addresses creates a federated namespace, based on control over the
908   indicated host and port, whether or not an HTTP server is present.
909   See <xref target="establishing.authority"/> for security considerations
910   related to establishing authority.
913   When an "http" URI is used within a context that calls for access to the
914   indicated resource, a client &MAY; attempt access by resolving
915   the host to an IP address, establishing a TCP connection to that address
916   on the indicated port, and sending an HTTP request message
917   (<xref target="http.message"/>) containing the URI's identifying data
918   (<xref target="message.routing"/>) to the server.
919   If the server responds to that request with a non-interim HTTP response
920   message, as described in &status-codes;, then that response
921   is considered an authoritative answer to the client's request.
924   Although HTTP is independent of the transport protocol, the "http"
925   scheme is specific to TCP-based services because the name delegation
926   process depends on TCP for establishing authority.
927   An HTTP service based on some other underlying connection protocol
928   would presumably be identified using a different URI scheme, just as
929   the "https" scheme (below) is used for resources that require an
930   end-to-end secured connection. Other protocols might also be used to
931   provide access to "http" identified resources &mdash; it is only the
932   authoritative interface that is specific to TCP.
935   The URI generic syntax for authority also includes a deprecated
936   userinfo subcomponent (<xref target="RFC3986" x:fmt="," x:sec="3.2.1"/>)
937   for including user authentication information in the URI.  Some
938   implementations make use of the userinfo component for internal
939   configuration of authentication information, such as within command
940   invocation options, configuration files, or bookmark lists, even
941   though such usage might expose a user identifier or password.
942   A sender &MUST-NOT; generate the userinfo subcomponent (and its "@"
943   delimiter) when an "http" URI reference is generated within a message as a
944   request target or header field value.
945   Before making use of an "http" URI reference received from an untrusted
946   source, a recipient &SHOULD; parse for userinfo and treat its presence as
947   an error; it is likely being used to obscure the authority for the sake of
948   phishing attacks.
952<section title="https URI Scheme" anchor="https.uri">
953   <x:anchor-alias value="https-URI"/>
954   <iref item="https URI scheme"/>
955   <iref item="URI scheme" subitem="https"/>
957   The "https" URI scheme is hereby defined for the purpose of minting
958   identifiers according to their association with the hierarchical
959   namespace governed by a potential HTTP origin server listening to a
960   given TCP port for TLS-secured connections (<xref target="RFC5246"/>).
963   All of the requirements listed above for the "http" scheme are also
964   requirements for the "https" scheme, except that TCP port 443 is the
965   default if the port subcomponent is empty or not given,
966   and the user agent &MUST; ensure that its connection to the origin
967   server is secured through the use of strong encryption, end-to-end,
968   prior to sending the first HTTP request.
970<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="https-URI"><!--terminal production--></iref>
971  <x:ref>https-URI</x:ref> = "https:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
972              [ "#" <x:ref>fragment</x:ref> ]
975   Note that the "https" URI scheme depends on both TLS and TCP for
976   establishing authority.
977   Resources made available via the "https" scheme have no shared
978   identity with the "http" scheme even if their resource identifiers
979   indicate the same authority (the same host listening to the same
980   TCP port).  They are distinct namespaces and are considered to be
981   distinct origin servers.  However, an extension to HTTP that is
982   defined to apply to entire host domains, such as the Cookie protocol
983   <xref target="RFC6265"/>, can allow information
984   set by one service to impact communication with other services
985   within a matching group of host domains.
988   The process for authoritative access to an "https" identified
989   resource is defined in <xref target="RFC2818"/>.
993<section title="http and https URI Normalization and Comparison" anchor="uri.comparison">
995   Since the "http" and "https" schemes conform to the URI generic syntax,
996   such URIs are normalized and compared according to the algorithm defined
997   in <xref target="RFC3986" x:fmt="of" x:sec="6"/>, using the defaults
998   described above for each scheme.
1001   If the port is equal to the default port for a scheme, the normal form is
1002   to omit the port subcomponent. When not being used in absolute form as the
1003   request target of an OPTIONS request, an empty path component is equivalent
1004   to an absolute path of "/", so the normal form is to provide a path of "/"
1005   instead. The scheme and host are case-insensitive and normally provided in
1006   lowercase; all other components are compared in a case-sensitive manner.
1007   Characters other than those in the "reserved" set are equivalent to their
1008   percent-encoded octets: the normal form is to not encode them
1009   (see Sections <xref target="RFC3986" x:fmt="number" x:sec="2.1"/> and
1010   <xref target="RFC3986" x:fmt="number" x:sec="2.2"/> of
1011   <xref target="RFC3986"/>).
1014   For example, the following three URIs are equivalent:
1016<figure><artwork type="example">
1025<section title="Message Format" anchor="http.message">
1026<x:anchor-alias value="generic-message"/>
1027<x:anchor-alias value="message.types"/>
1028<x:anchor-alias value="HTTP-message"/>
1029<x:anchor-alias value="start-line"/>
1030<iref item="header section"/>
1031<iref item="headers"/>
1032<iref item="header field"/>
1034   All HTTP/1.1 messages consist of a start-line followed by a sequence of
1035   octets in a format similar to the Internet Message Format
1036   <xref target="RFC5322"/>: zero or more header fields (collectively
1037   referred to as the "headers" or the "header section"), an empty line
1038   indicating the end of the header section, and an optional message body.
1040<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"><!--terminal production--></iref>
1041  <x:ref>HTTP-message</x:ref>   = <x:ref>start-line</x:ref>
1042                   *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
1043                   <x:ref>CRLF</x:ref>
1044                   [ <x:ref>message-body</x:ref> ]
1047   The normal procedure for parsing an HTTP message is to read the
1048   start-line into a structure, read each header field into a hash
1049   table by field name until the empty line, and then use the parsed
1050   data to determine if a message body is expected.  If a message body
1051   has been indicated, then it is read as a stream until an amount
1052   of octets equal to the message body length is read or the connection
1053   is closed.
1056   A recipient &MUST; parse an HTTP message as a sequence of octets in an
1057   encoding that is a superset of US-ASCII <xref target="USASCII"/>.
1058   Parsing an HTTP message as a stream of Unicode characters, without regard
1059   for the specific encoding, creates security vulnerabilities due to the
1060   varying ways that string processing libraries handle invalid multibyte
1061   character sequences that contain the octet LF (%x0A).  String-based
1062   parsers can only be safely used within protocol elements after the element
1063   has been extracted from the message, such as within a header field-value
1064   after message parsing has delineated the individual fields.
1067   An HTTP message can be parsed as a stream for incremental processing or
1068   forwarding downstream.  However, recipients cannot rely on incremental
1069   delivery of partial messages, since some implementations will buffer or
1070   delay message forwarding for the sake of network efficiency, security
1071   checks, or payload transformations.
1074   A sender &MUST-NOT; send whitespace between the start-line and
1075   the first header field.
1076   A recipient that receives whitespace between the start-line and
1077   the first header field &MUST; either reject the message as invalid or
1078   consume each whitespace-preceded line without further processing of it
1079   (i.e., ignore the entire line, along with any subsequent lines preceded
1080   by whitespace, until a properly formed header field is received or the
1081   header section is terminated).
1084   The presence of such whitespace in a request
1085   might be an attempt to trick a server into ignoring that field or
1086   processing the line after it as a new request, either of which might
1087   result in a security vulnerability if other implementations within
1088   the request chain interpret the same message differently.
1089   Likewise, the presence of such whitespace in a response might be
1090   ignored by some clients or cause others to cease parsing.
1093<section title="Start Line" anchor="start.line">
1094  <x:anchor-alias value="Start-Line"/>
1096   An HTTP message can be either a request from client to server or a
1097   response from server to client.  Syntactically, the two types of message
1098   differ only in the start-line, which is either a request-line (for requests)
1099   or a status-line (for responses), and in the algorithm for determining
1100   the length of the message body (<xref target="message.body"/>).
1103   In theory, a client could receive requests and a server could receive
1104   responses, distinguishing them by their different start-line formats,
1105   but, in practice, servers are implemented to only expect a request
1106   (a response is interpreted as an unknown or invalid request method)
1107   and clients are implemented to only expect a response.
1109<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="start-line"/>
1110  <x:ref>start-line</x:ref>     = <x:ref>request-line</x:ref> / <x:ref>status-line</x:ref>
1113<section title="Request Line" anchor="request.line">
1114  <x:anchor-alias value="Request"/>
1115  <x:anchor-alias value="request-line"/>
1117   A request-line begins with a method token, followed by a single
1118   space (SP), the request-target, another single space (SP), the
1119   protocol version, and ending with CRLF.
1121<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-line"/>
1122  <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>
1124<iref primary="true" item="method"/>
1125<t anchor="method">
1126   The method token indicates the request method to be performed on the
1127   target resource. The request method is case-sensitive.
1129<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="method"/>
1130  <x:ref>method</x:ref>         = <x:ref>token</x:ref>
1133   The request methods defined by this specification can be found in
1134   &methods;, along with information regarding the HTTP method registry
1135   and considerations for defining new methods.
1137<iref item="request-target"/>
1139   The request-target identifies the target resource upon which to apply
1140   the request, as defined in <xref target="request-target"/>.
1143   Recipients typically parse the request-line into its component parts by
1144   splitting on whitespace (see <xref target="message.robustness"/>), since
1145   no whitespace is allowed in the three components.
1146   Unfortunately, some user agents fail to properly encode or exclude
1147   whitespace found in hypertext references, resulting in those disallowed
1148   characters being sent in a request-target.
1151   Recipients of an invalid request-line &SHOULD; respond with either a
1152   <x:ref>400 (Bad Request)</x:ref> error or a <x:ref>301 (Moved Permanently)</x:ref>
1153   redirect with the request-target properly encoded.  A recipient &SHOULD-NOT;
1154   attempt to autocorrect and then process the request without a redirect,
1155   since the invalid request-line might be deliberately crafted to bypass
1156   security filters along the request chain.
1159   HTTP does not place a predefined limit on the length of a request-line,
1160   as described in <xref target="conformance"/>.
1161   A server that receives a method longer than any that it implements
1162   &SHOULD; respond with a <x:ref>501 (Not Implemented)</x:ref> status code.
1163   A server that receives a request-target longer than any URI it wishes to
1164   parse &MUST; respond with a
1165   <x:ref>414 (URI Too Long)</x:ref> status code (see &status-414;).
1168   Various ad hoc limitations on request-line length are found in practice.
1169   It is &RECOMMENDED; that all HTTP senders and recipients support, at a
1170   minimum, request-line lengths of 8000 octets.
1174<section title="Status Line" anchor="status.line">
1175  <x:anchor-alias value="response"/>
1176  <x:anchor-alias value="status-line"/>
1177  <x:anchor-alias value="status-code"/>
1178  <x:anchor-alias value="reason-phrase"/>
1180   The first line of a response message is the status-line, consisting
1181   of the protocol version, a space (SP), the status code, another space,
1182   a possibly empty textual phrase describing the status code, and
1183   ending with CRLF.
1185<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-line"/>
1186  <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>
1189   The status-code element is a 3-digit integer code describing the
1190   result of the server's attempt to understand and satisfy the client's
1191   corresponding request. The rest of the response message is to be
1192   interpreted in light of the semantics defined for that status code.
1193   See &status-codes; for information about the semantics of status codes,
1194   including the classes of status code (indicated by the first digit),
1195   the status codes defined by this specification, considerations for the
1196   definition of new status codes, and the IANA registry.
1198<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-code"/>
1199  <x:ref>status-code</x:ref>    = 3<x:ref>DIGIT</x:ref>
1202   The reason-phrase element exists for the sole purpose of providing a
1203   textual description associated with the numeric status code, mostly
1204   out of deference to earlier Internet application protocols that were more
1205   frequently used with interactive text clients. A client &SHOULD; ignore
1206   the reason-phrase content.
1208<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="reason-phrase"/>
1209  <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> )
1214<section title="Header Fields" anchor="header.fields">
1215  <x:anchor-alias value="header-field"/>
1216  <x:anchor-alias value="field-content"/>
1217  <x:anchor-alias value="field-name"/>
1218  <x:anchor-alias value="field-value"/>
1219  <x:anchor-alias value="field-vchar"/>
1220  <x:anchor-alias value="obs-fold"/>
1222   Each header field consists of a case-insensitive field name
1223   followed by a colon (":"), optional leading whitespace, the field value,
1224   and optional trailing whitespace.
1226<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-vchar"/><iref primary="true" item="Grammar" subitem="field-content"/><iref primary="true" item="Grammar" subitem="obs-fold"/>
1227  <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>OWS</x:ref>
1229  <x:ref>field-name</x:ref>     = <x:ref>token</x:ref>
1230  <x:ref>field-value</x:ref>    = *( <x:ref>field-content</x:ref> / <x:ref>obs-fold</x:ref> )
1231  <x:ref>field-content</x:ref>  = <x:ref>field-vchar</x:ref> [ 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> ) <x:ref>field-vchar</x:ref> ]
1232  <x:ref>field-vchar</x:ref>    = <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref>
1234  <x:ref>obs-fold</x:ref>       = <x:ref>CRLF</x:ref> 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1235                 ; obsolete line folding
1236                 ; see <xref target="field.parsing"/>
1239   The field-name token labels the corresponding field-value as having the
1240   semantics defined by that header field.  For example, the <x:ref>Date</x:ref>
1241   header field is defined in &header-date; as containing the origination
1242   timestamp for the message in which it appears.
1245<section title="Field Extensibility" anchor="field.extensibility">
1247   Header fields are fully extensible: there is no limit on the
1248   introduction of new field names, each presumably defining new semantics,
1249   nor on the number of header fields used in a given message.  Existing
1250   fields are defined in each part of this specification and in many other
1251   specifications outside this document set.
1254   New header fields can be defined such that, when they are understood by a
1255   recipient, they might override or enhance the interpretation of previously
1256   defined header fields, define preconditions on request evaluation, or
1257   refine the meaning of responses.
1260   A proxy &MUST; forward unrecognized header fields unless the
1261   field-name is listed in the <x:ref>Connection</x:ref> header field
1262   (<xref target="header.connection"/>) or the proxy is specifically
1263   configured to block, or otherwise transform, such fields.
1264   Other recipients &SHOULD; ignore unrecognized header fields.
1265   These requirements allow HTTP's functionality to be enhanced without
1266   requiring prior update of deployed intermediaries.
1269   All defined header fields ought to be registered with IANA in the
1270   Message Header Field Registry, as described in &iana-header-registry;.
1274<section title="Field Order" anchor="field.order">
1276   The order in which header fields with differing field names are
1277   received is not significant. However, it is good practice to send
1278   header fields that contain control data first, such as <x:ref>Host</x:ref>
1279   on requests and <x:ref>Date</x:ref> on responses, so that implementations
1280   can decide when not to handle a message as early as possible.
1281   A server &MUST-NOT; apply a request to the target resource until the entire
1282   request header section is received, since later header fields might include
1283   conditionals, authentication credentials, or deliberately misleading
1284   duplicate header fields that would impact request processing.
1287   A sender &MUST-NOT; generate multiple header fields with the same field
1288   name in a message unless either the entire field value for that
1289   header field is defined as a comma-separated list [i.e., #(values)]
1290   or the header field is a well-known exception (as noted below).
1293   A recipient &MAY; combine multiple header fields with the same field name
1294   into one "field-name: field-value" pair, without changing the semantics of
1295   the message, by appending each subsequent field value to the combined
1296   field value in order, separated by a comma. The order in which
1297   header fields with the same field name are received is therefore
1298   significant to the interpretation of the combined field value;
1299   a proxy &MUST-NOT; change the order of these field values when
1300   forwarding a message.
1303  <t>
1304   &Note; In practice, the "Set-Cookie" header field (<xref target="RFC6265"/>)
1305   often appears multiple times in a response message and does not use the
1306   list syntax, violating the above requirements on multiple header fields
1307   with the same name. Since it cannot be combined into a single field-value,
1308   recipients ought to handle "Set-Cookie" as a special case while processing
1309   header fields. (See Appendix A.2.3 of <xref target="Kri2001"/> for details.)
1310  </t>
1314<section title="Whitespace" anchor="whitespace">
1315<t anchor="rule.LWS">
1316   This specification uses three rules to denote the use of linear
1317   whitespace: OWS (optional whitespace), RWS (required whitespace), and
1318   BWS ("bad" whitespace).
1320<t anchor="rule.OWS">
1321   The OWS rule is used where zero or more linear whitespace octets might
1322   appear. For protocol elements where optional whitespace is preferred to
1323   improve readability, a sender &SHOULD; generate the optional whitespace
1324   as a single SP; otherwise, a sender &SHOULD-NOT; generate optional
1325   whitespace except as needed to white-out invalid or unwanted protocol
1326   elements during in-place message filtering.
1328<t anchor="rule.RWS">
1329   The RWS rule is used when at least one linear whitespace octet is required
1330   to separate field tokens. A sender &SHOULD; generate RWS as a single SP.
1332<t anchor="rule.BWS">
1333   The BWS rule is used where the grammar allows optional whitespace only for
1334   historical reasons. A sender &MUST-NOT; generate BWS in messages.
1335   A recipient &MUST; parse for such bad whitespace and remove it before
1336   interpreting the protocol element.
1338<t anchor="rule.whitespace">
1339  <x:anchor-alias value="BWS"/>
1340  <x:anchor-alias value="OWS"/>
1341  <x:anchor-alias value="RWS"/>
1343<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"/>
1344  <x:ref>OWS</x:ref>            = *( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1345                 ; optional whitespace
1346  <x:ref>RWS</x:ref>            = 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1347                 ; required whitespace
1348  <x:ref>BWS</x:ref>            = <x:ref>OWS</x:ref>
1349                 ; "bad" whitespace
1353<section title="Field Parsing" anchor="field.parsing">
1355   Messages are parsed using a generic algorithm, independent of the
1356   individual header field names. The contents within a given field value are
1357   not parsed until a later stage of message interpretation (usually after the
1358   message's entire header section has been processed).
1359   Consequently, this specification does not use ABNF rules to define each
1360   "Field-Name: Field Value" pair, as was done in previous editions.
1361   Instead, this specification uses ABNF rules that are named according to
1362   each registered field name, wherein the rule defines the valid grammar for
1363   that field's corresponding field values (i.e., after the field-value
1364   has been extracted from the header section by a generic field parser).
1367   No whitespace is allowed between the header field-name and colon.
1368   In the past, differences in the handling of such whitespace have led to
1369   security vulnerabilities in request routing and response handling.
1370   A server &MUST; reject any received request message that contains
1371   whitespace between a header field-name and colon with a response code of
1372   <x:ref>400 (Bad Request)</x:ref>. A proxy &MUST; remove any such whitespace
1373   from a response message before forwarding the message downstream.
1376   A field value might be preceded and/or followed by optional whitespace
1377   (OWS); a single SP preceding the field-value is preferred for consistent
1378   readability by humans.
1379   The field value does not include any leading or trailing whitespace: OWS
1380   occurring before the first non-whitespace octet of the field value or after
1381   the last non-whitespace octet of the field value ought to be excluded by
1382   parsers when extracting the field value from a header field.
1385   Historically, HTTP header field values could be extended over multiple
1386   lines by preceding each extra line with at least one space or horizontal
1387   tab (obs-fold). This specification deprecates such line folding except
1388   within the message/http media type
1389   (<xref target=""/>).
1390   A sender &MUST-NOT; generate a message that includes line folding
1391   (i.e., that has any field-value that contains a match to the
1392   <x:ref>obs-fold</x:ref> rule) unless the message is intended for packaging
1393   within the message/http media type.
1396   A server that receives an <x:ref>obs-fold</x:ref> in a request message that
1397   is not within a message/http container &MUST; either reject the message by
1398   sending a <x:ref>400 (Bad Request)</x:ref>, preferably with a
1399   representation explaining that obsolete line folding is unacceptable, or
1400   replace each received <x:ref>obs-fold</x:ref> with one or more
1401   <x:ref>SP</x:ref> octets prior to interpreting the field value or
1402   forwarding the message downstream.
1405   A proxy or gateway that receives an <x:ref>obs-fold</x:ref> in a response
1406   message that is not within a message/http container &MUST; either discard
1407   the message and replace it with a <x:ref>502 (Bad Gateway)</x:ref>
1408   response, preferably with a representation explaining that unacceptable
1409   line folding was received, or replace each received <x:ref>obs-fold</x:ref>
1410   with one or more <x:ref>SP</x:ref> octets prior to interpreting the field
1411   value or forwarding the message downstream.
1414   A user agent that receives an <x:ref>obs-fold</x:ref> in a response message
1415   that is not within a message/http container &MUST; replace each received
1416   <x:ref>obs-fold</x:ref> with one or more <x:ref>SP</x:ref> octets prior to
1417   interpreting the field value.
1420   Historically, HTTP has allowed field content with text in the ISO&nbhy;8859&nbhy;1
1421   charset <xref target="ISO-8859-1"/>, supporting other charsets only
1422   through use of <xref target="RFC2047"/> encoding.
1423   In practice, most HTTP header field values use only a subset of the
1424   US-ASCII charset <xref target="USASCII"/>. Newly defined
1425   header fields &SHOULD; limit their field values to US&nbhy;ASCII octets.
1426   A recipient &SHOULD; treat other octets in field content (obs&nbhy;text) as
1427   opaque data.
1431<section title="Field Limits" anchor="field.limits">
1433   HTTP does not place a predefined limit on the length of each header field
1434   or on the length of the header section as a whole, as described in
1435   <xref target="conformance"/>. Various ad hoc limitations on individual
1436   header field length are found in practice, often depending on the specific
1437   field semantics.
1440   A server that receives a request header field, or set of fields, larger
1441   than it wishes to process &MUST; respond with an appropriate
1442   <x:ref>4xx (Client Error)</x:ref> status code. Ignoring such header fields
1443   would increase the server's vulnerability to request smuggling attacks
1444   (<xref target="request.smuggling"/>).
1447   A client &MAY; discard or truncate received header fields that are larger
1448   than the client wishes to process if the field semantics are such that the
1449   dropped value(s) can be safely ignored without changing the
1450   message framing or response semantics.
1454<section title="Field Value Components" anchor="field.components">
1455<t anchor="rule.token.separators">
1456  <x:anchor-alias value="tchar"/>
1457  <x:anchor-alias value="token"/>
1458  <iref item="Delimiters"/>
1459   Most HTTP header field values are defined using common syntax components
1460   (token, quoted-string, and comment) separated by whitespace or specific
1461   delimiting characters. Delimiters are chosen from the set of US-ASCII
1462   visual characters not allowed in a <x:ref>token</x:ref>
1463   (DQUOTE and "(),/:;&lt;=>?@[\]{}").
1465<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="tchar"/>
1466  <x:ref>token</x:ref>          = 1*<x:ref>tchar</x:ref>
1468  NOTE: the definition of tchar and the prose above about special characters need to match!
1469 -->
1470  <x:ref>tchar</x:ref>          = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*"
1471                 / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~"
1472                 / <x:ref>DIGIT</x:ref> / <x:ref>ALPHA</x:ref>
1473                 ; any <x:ref>VCHAR</x:ref>, except delimiters
1475<t anchor="rule.quoted-string">
1476  <x:anchor-alias value="quoted-string"/>
1477  <x:anchor-alias value="qdtext"/>
1478  <x:anchor-alias value="obs-text"/>
1479   A string of text is parsed as a single value if it is quoted using
1480   double-quote marks.
1482<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"/>
1483  <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>
1484  <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>
1485  <x:ref>obs-text</x:ref>       = %x80-FF
1487<t anchor="rule.comment">
1488  <x:anchor-alias value="comment"/>
1489  <x:anchor-alias value="ctext"/>
1490   Comments can be included in some HTTP header fields by surrounding
1491   the comment text with parentheses. Comments are only allowed in
1492   fields containing "comment" as part of their field value definition.
1494<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
1495  <x:ref>comment</x:ref>        = "(" *( <x:ref>ctext</x:ref> / <x:ref>quoted-pair</x:ref> / <x:ref>comment</x:ref> ) ")"
1496  <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>
1498<t anchor="rule.quoted-pair">
1499  <x:anchor-alias value="quoted-pair"/>
1500   The backslash octet ("\") can be used as a single-octet
1501   quoting mechanism within quoted-string and comment constructs.
1502   Recipients that process the value of a quoted-string &MUST; handle a
1503   quoted-pair as if it were replaced by the octet following the backslash.
1505<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
1506  <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> )
1509   A sender &SHOULD-NOT; generate a quoted-pair in a quoted-string except
1510   where necessary to quote DQUOTE and backslash octets occurring within that
1511   string.
1512   A sender &SHOULD-NOT; generate a quoted-pair in a comment except
1513   where necessary to quote parentheses ["(" and ")"] and backslash octets
1514   occurring within that comment.
1520<section title="Message Body" anchor="message.body">
1521  <x:anchor-alias value="message-body"/>
1523   The message body (if any) of an HTTP message is used to carry the
1524   payload body of that request or response.  The message body is
1525   identical to the payload body unless a transfer coding has been
1526   applied, as described in <xref target="header.transfer-encoding"/>.
1528<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1529  <x:ref>message-body</x:ref> = *OCTET
1532   The rules for when a message body is allowed in a message differ for
1533   requests and responses.
1536   The presence of a message body in a request is signaled by a
1537   <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1538   field. Request message framing is independent of method semantics,
1539   even if the method does not define any use for a message body.
1542   The presence of a message body in a response depends on both
1543   the request method to which it is responding and the response
1544   status code (<xref target="status.line"/>).
1545   Responses to the HEAD request method (&HEAD;) never include a message body
1546   because the associated response header fields (e.g.,
1547   <x:ref>Transfer-Encoding</x:ref>, <x:ref>Content-Length</x:ref>, etc.),
1548   if present, indicate only what their values would have been if the request
1549   method had been GET (&GET;).
1550   <x:ref>2xx (Successful)</x:ref> responses to a CONNECT request method
1551   (&CONNECT;) switch to tunnel mode instead of having a message body.
1552   All <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, and
1553   <x:ref>304 (Not Modified)</x:ref> responses do not include a message body.
1554   All other responses do include a message body, although the body
1555   might be of zero length.
1558<section title="Transfer-Encoding" anchor="header.transfer-encoding">
1559  <iref primary="true" item="Transfer-Encoding header field" x:for-anchor=""/>
1560  <iref item="chunked (Coding Format)"/>
1561  <x:anchor-alias value="Transfer-Encoding"/>
1563   The Transfer-Encoding header field lists the transfer coding names
1564   corresponding to the sequence of transfer codings that have been
1565   (or will be) applied to the payload body in order to form the message body.
1566   Transfer codings are defined in <xref target="transfer.codings"/>.
1568<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
1569  <x:ref>Transfer-Encoding</x:ref> = 1#<x:ref>transfer-coding</x:ref>
1572   Transfer-Encoding is analogous to the Content-Transfer-Encoding field of
1573   MIME, which was designed to enable safe transport of binary data over a
1574   7-bit transport service (<xref target="RFC2045" x:fmt="," x:sec="6"/>).
1575   However, safe transport has a different focus for an 8bit-clean transfer
1576   protocol. In HTTP's case, Transfer-Encoding is primarily intended to
1577   accurately delimit a dynamically generated payload and to distinguish
1578   payload encodings that are only applied for transport efficiency or
1579   security from those that are characteristics of the selected resource.
1582   A recipient &MUST; be able to parse the chunked transfer coding
1583   (<xref target="chunked.encoding"/>) because it plays a crucial role in
1584   framing messages when the payload body size is not known in advance.
1585   A sender &MUST-NOT; apply chunked more than once to a message body
1586   (i.e., chunking an already chunked message is not allowed).
1587   If any transfer coding other than chunked is applied to a request payload
1588   body, the sender &MUST; apply chunked as the final transfer coding to
1589   ensure that the message is properly framed.
1590   If any transfer coding other than chunked is applied to a response payload
1591   body, the sender &MUST; either apply chunked as the final transfer coding
1592   or terminate the message by closing the connection.
1595   For example,
1596</preamble><artwork type="example">
1597  Transfer-Encoding: gzip, chunked
1599   indicates that the payload body has been compressed using the gzip
1600   coding and then chunked using the chunked coding while forming the
1601   message body.
1604   Unlike <x:ref>Content-Encoding</x:ref> (&content-codings;),
1605   Transfer-Encoding is a property of the message, not of the representation, and
1606   any recipient along the request/response chain &MAY; decode the received
1607   transfer coding(s) or apply additional transfer coding(s) to the message
1608   body, assuming that corresponding changes are made to the Transfer-Encoding
1609   field-value. Additional information about the encoding parameters can be
1610   provided by other header fields not defined by this specification.
1613   Transfer-Encoding &MAY; be sent in a response to a HEAD request or in a
1614   <x:ref>304 (Not Modified)</x:ref> response (&status-304;) to a GET request,
1615   neither of which includes a message body,
1616   to indicate that the origin server would have applied a transfer coding
1617   to the message body if the request had been an unconditional GET.
1618   This indication is not required, however, because any recipient on
1619   the response chain (including the origin server) can remove transfer
1620   codings when they are not needed.
1623   A server &MUST-NOT; send a Transfer-Encoding header field in any response
1624   with a status code of
1625   <x:ref>1xx (Informational)</x:ref> or <x:ref>204 (No Content)</x:ref>.
1626   A server &MUST-NOT; send a Transfer-Encoding header field in any
1627   <x:ref>2xx (Successful)</x:ref> response to a CONNECT request (&CONNECT;).
1630   Transfer-Encoding was added in HTTP/1.1.  It is generally assumed that
1631   implementations advertising only HTTP/1.0 support will not understand
1632   how to process a transfer-encoded payload.
1633   A client &MUST-NOT; send a request containing Transfer-Encoding unless it
1634   knows the server will handle HTTP/1.1 (or later) requests; such knowledge
1635   might be in the form of specific user configuration or by remembering the
1636   version of a prior received response.
1637   A server &MUST-NOT; send a response containing Transfer-Encoding unless
1638   the corresponding request indicates HTTP/1.1 (or later).
1641   A server that receives a request message with a transfer coding it does
1642   not understand &SHOULD; respond with <x:ref>501 (Not Implemented)</x:ref>.
1646<section title="Content-Length" anchor="header.content-length">
1647  <iref primary="true" item="Content-Length header field" x:for-anchor=""/>
1648  <x:anchor-alias value="Content-Length"/>
1650   When a message does not have a <x:ref>Transfer-Encoding</x:ref> header
1651   field, a Content-Length header field can provide the anticipated size,
1652   as a decimal number of octets, for a potential payload body.
1653   For messages that do include a payload body, the Content-Length field-value
1654   provides the framing information necessary for determining where the body
1655   (and message) ends.  For messages that do not include a payload body, the
1656   Content-Length indicates the size of the selected representation
1657   (&representation;).
1659<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
1660  <x:ref>Content-Length</x:ref> = 1*<x:ref>DIGIT</x:ref>
1663   An example is
1665<figure><artwork type="example">
1666  Content-Length: 3495
1669   A sender &MUST-NOT; send a Content-Length header field in any message that
1670   contains a <x:ref>Transfer-Encoding</x:ref> header field.
1673   A user agent &SHOULD; send a Content-Length in a request message when no
1674   <x:ref>Transfer-Encoding</x:ref> is sent and the request method defines
1675   a meaning for an enclosed payload body. For example, a Content-Length
1676   header field is normally sent in a POST request even when the value is
1677   0 (indicating an empty payload body).  A user agent &SHOULD-NOT; send a
1678   Content-Length header field when the request message does not contain a
1679   payload body and the method semantics do not anticipate such a body.
1682   A server &MAY; send a Content-Length header field in a response to a HEAD
1683   request (&HEAD;); a server &MUST-NOT; send Content-Length in such a
1684   response unless its field-value equals the decimal number of octets that
1685   would have been sent in the payload body of a response if the same
1686   request had used the GET method.
1689   A server &MAY; send a Content-Length header field in a
1690   <x:ref>304 (Not Modified)</x:ref> response to a conditional GET request
1691   (&status-304;); a server &MUST-NOT; send Content-Length in such a
1692   response unless its field-value equals the decimal number of octets that
1693   would have been sent in the payload body of a <x:ref>200 (OK)</x:ref>
1694   response to the same request.
1697   A server &MUST-NOT; send a Content-Length header field in any response
1698   with a status code of
1699   <x:ref>1xx (Informational)</x:ref> or <x:ref>204 (No Content)</x:ref>.
1700   A server &MUST-NOT; send a Content-Length header field in any
1701   <x:ref>2xx (Successful)</x:ref> response to a CONNECT request (&CONNECT;).
1704   Aside from the cases defined above, in the absence of Transfer-Encoding,
1705   an origin server &SHOULD; send a Content-Length header field when the
1706   payload body size is known prior to sending the complete header section.
1707   This will allow downstream recipients to measure transfer progress,
1708   know when a received message is complete, and potentially reuse the
1709   connection for additional requests.
1712   Any Content-Length field value greater than or equal to zero is valid.
1713   Since there is no predefined limit to the length of a payload, a
1714   recipient &MUST; anticipate potentially large decimal numerals and
1715   prevent parsing errors due to integer conversion overflows
1716   (<xref target="attack.protocol.element.length"/>).
1719   If a message is received that has multiple Content-Length header fields
1720   with field-values consisting of the same decimal value, or a single
1721   Content-Length header field with a field value containing a list of
1722   identical decimal values (e.g., "Content-Length: 42, 42"), indicating that
1723   duplicate Content-Length header fields have been generated or combined by an
1724   upstream message processor, then the recipient &MUST; either reject the
1725   message as invalid or replace the duplicated field-values with a single
1726   valid Content-Length field containing that decimal value prior to
1727   determining the message body length or forwarding the message.
1730  <t>
1731   &Note; HTTP's use of Content-Length for message framing differs
1732   significantly from the same field's use in MIME, where it is an optional
1733   field used only within the "message/external-body" media-type.
1734  </t>
1738<section title="Message Body Length" anchor="message.body.length">
1739  <iref item="chunked (Coding Format)"/>
1741   The length of a message body is determined by one of the following
1742   (in order of precedence):
1745  <list style="numbers">
1746    <x:lt><t>
1747     Any response to a HEAD request and any response with a
1748     <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, or
1749     <x:ref>304 (Not Modified)</x:ref> status code is always
1750     terminated by the first empty line after the header fields, regardless of
1751     the header fields present in the message, and thus cannot contain a
1752     message body.
1753    </t></x:lt>
1754    <x:lt><t>
1755     Any <x:ref>2xx (Successful)</x:ref> response to a CONNECT request implies that the
1756     connection will become a tunnel immediately after the empty line that
1757     concludes the header fields.  A client &MUST; ignore any
1758     <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1759     fields received in such a message.
1760    </t></x:lt>
1761    <x:lt><t>
1762     If a <x:ref>Transfer-Encoding</x:ref> header field is present
1763     and the chunked transfer coding (<xref target="chunked.encoding"/>)
1764     is the final encoding, the message body length is determined by reading
1765     and decoding the chunked data until the transfer coding indicates the
1766     data is complete.
1767    </t>
1768    <t>
1769     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a
1770     response and the chunked transfer coding is not the final encoding, the
1771     message body length is determined by reading the connection until it is
1772     closed by the server.
1773     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a request and the
1774     chunked transfer coding is not the final encoding, the message body
1775     length cannot be determined reliably; the server &MUST; respond with
1776     the <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1777    </t>
1778    <t>
1779     If a message is received with both a <x:ref>Transfer-Encoding</x:ref>
1780     and a <x:ref>Content-Length</x:ref> header field, the Transfer-Encoding
1781     overrides the Content-Length. Such a message might indicate an attempt to
1782     perform request smuggling (<xref target="request.smuggling"/>) or
1783     response splitting (<xref target="response.splitting"/>) and ought to be
1784     handled as an error. A sender &MUST; remove the received Content-Length
1785     field prior to forwarding such a message downstream.
1786    </t></x:lt>
1787    <x:lt><t>
1788     If a message is received without <x:ref>Transfer-Encoding</x:ref> and with
1789     either multiple <x:ref>Content-Length</x:ref> header fields having
1790     differing field-values or a single Content-Length header field having an
1791     invalid value, then the message framing is invalid and
1792     the recipient &MUST; treat it as an unrecoverable error.
1793     If this is a request message, the server &MUST; respond with
1794     a <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1795     If this is a response message received by a proxy,
1796     the proxy &MUST; close the connection to the server, discard the received
1797     response, and send a <x:ref>502 (Bad Gateway)</x:ref> response to the
1798     client.
1799     If this is a response message received by a user agent,
1800     the user agent &MUST; close the connection to the server and discard the
1801     received response.
1802    </t></x:lt>
1803    <x:lt><t>
1804     If a valid <x:ref>Content-Length</x:ref> header field is present without
1805     <x:ref>Transfer-Encoding</x:ref>, its decimal value defines the
1806     expected message body length in octets.
1807     If the sender closes the connection or the recipient times out before the
1808     indicated number of octets are received, the recipient &MUST; consider
1809     the message to be incomplete and close the connection.
1810    </t></x:lt>
1811    <x:lt><t>
1812     If this is a request message and none of the above are true, then the
1813     message body length is zero (no message body is present).
1814    </t></x:lt>
1815    <x:lt><t>
1816     Otherwise, this is a response message without a declared message body
1817     length, so the message body length is determined by the number of octets
1818     received prior to the server closing the connection.
1819    </t></x:lt>
1820  </list>
1823   Since there is no way to distinguish a successfully completed,
1824   close-delimited message from a partially-received message interrupted
1825   by network failure, a server &SHOULD; generate encoding or
1826   length-delimited messages whenever possible.  The close-delimiting
1827   feature exists primarily for backwards compatibility with HTTP/1.0.
1830   A server &MAY; reject a request that contains a message body but
1831   not a <x:ref>Content-Length</x:ref> by responding with
1832   <x:ref>411 (Length Required)</x:ref>.
1835   Unless a transfer coding other than chunked has been applied,
1836   a client that sends a request containing a message body &SHOULD;
1837   use a valid <x:ref>Content-Length</x:ref> header field if the message body
1838   length is known in advance, rather than the chunked transfer coding, since some
1839   existing services respond to chunked with a <x:ref>411 (Length Required)</x:ref>
1840   status code even though they understand the chunked transfer coding.  This
1841   is typically because such services are implemented via a gateway that
1842   requires a content-length in advance of being called and the server
1843   is unable or unwilling to buffer the entire request before processing.
1846   A user agent that sends a request containing a message body &MUST; send a
1847   valid <x:ref>Content-Length</x:ref> header field if it does not know the
1848   server will handle HTTP/1.1 (or later) requests; such knowledge can be in
1849   the form of specific user configuration or by remembering the version of a
1850   prior received response.
1853   If the final response to the last request on a connection has been
1854   completely received and there remains additional data to read, a user agent
1855   &MAY; discard the remaining data or attempt to determine if that data
1856   belongs as part of the prior response body, which might be the case if the
1857   prior message's Content-Length value is incorrect. A client &MUST-NOT;
1858   process, cache, or forward such extra data as a separate response, since
1859   such behavior would be vulnerable to cache poisoning.
1864<section anchor="incomplete.messages" title="Handling Incomplete Messages">
1866   A server that receives an incomplete request message, usually due to a
1867   canceled request or a triggered timeout exception, &MAY; send an error
1868   response prior to closing the connection.
1871   A client that receives an incomplete response message, which can occur
1872   when a connection is closed prematurely or when decoding a supposedly
1873   chunked transfer coding fails, &MUST; record the message as incomplete.
1874   Cache requirements for incomplete responses are defined in
1875   &cache-incomplete;.
1878   If a response terminates in the middle of the header section (before the
1879   empty line is received) and the status code might rely on header fields to
1880   convey the full meaning of the response, then the client cannot assume
1881   that meaning has been conveyed; the client might need to repeat the
1882   request in order to determine what action to take next.
1885   A message body that uses the chunked transfer coding is
1886   incomplete if the zero-sized chunk that terminates the encoding has not
1887   been received.  A message that uses a valid <x:ref>Content-Length</x:ref> is
1888   incomplete if the size of the message body received (in octets) is less than
1889   the value given by Content-Length.  A response that has neither chunked
1890   transfer coding nor Content-Length is terminated by closure of the
1891   connection and, thus, is considered complete regardless of the number of
1892   message body octets received, provided that the header section was received
1893   intact.
1897<section title="Message Parsing Robustness" anchor="message.robustness">
1899   Older HTTP/1.0 user agent implementations might send an extra CRLF
1900   after a POST request as a workaround for some early server
1901   applications that failed to read message body content that was
1902   not terminated by a line-ending. An HTTP/1.1 user agent &MUST-NOT;
1903   preface or follow a request with an extra CRLF.  If terminating
1904   the request message body with a line-ending is desired, then the
1905   user agent &MUST; count the terminating CRLF octets as part of the
1906   message body length.
1909   In the interest of robustness, a server that is expecting to receive and
1910   parse a request-line &SHOULD; ignore at least one empty line (CRLF)
1911   received prior to the request-line.
1914   Although the line terminator for the start-line and header
1915   fields is the sequence CRLF, a recipient &MAY; recognize a
1916   single LF as a line terminator and ignore any preceding CR.
1919   Although the request-line and status-line grammar rules require that each
1920   of the component elements be separated by a single SP octet, recipients
1921   &MAY; instead parse on whitespace-delimited word boundaries and, aside
1922   from the CRLF terminator, treat any form of whitespace as the SP separator
1923   while ignoring preceding or trailing whitespace;
1924   such whitespace includes one or more of the following octets:
1925   SP, HTAB, VT (%x0B), FF (%x0C), or bare CR.
1926   However, lenient parsing can result in security vulnerabilities if there
1927   are multiple recipients of the message and each has its own unique
1928   interpretation of robustness (see <xref target="request.smuggling"/>).
1931   When a server listening only for HTTP request messages, or processing
1932   what appears from the start-line to be an HTTP request message,
1933   receives a sequence of octets that does not match the HTTP-message
1934   grammar aside from the robustness exceptions listed above, the
1935   server &SHOULD; respond with a <x:ref>400 (Bad Request)</x:ref> response. 
1940<section title="Transfer Codings" anchor="transfer.codings">
1941  <x:anchor-alias value="transfer-coding"/>
1942  <x:anchor-alias value="transfer-extension"/>
1944   Transfer coding names are used to indicate an encoding
1945   transformation that has been, can be, or might need to be applied to a
1946   payload body in order to ensure "safe transport" through the network.
1947   This differs from a content coding in that the transfer coding is a
1948   property of the message rather than a property of the representation
1949   that is being transferred.
1951<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1952  <x:ref>transfer-coding</x:ref>    = "chunked" ; <xref target="chunked.encoding"/>
1953                     / "compress" ; <xref target="compress.coding"/>
1954                     / "deflate" ; <xref target="deflate.coding"/>
1955                     / "gzip" ; <xref target="gzip.coding"/>
1956                     / <x:ref>transfer-extension</x:ref>
1957  <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> )
1959<t anchor="rule.parameter">
1960  <x:anchor-alias value="transfer-parameter"/>
1961   Parameters are in the form of a name or name=value pair.
1963<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-parameter"/>
1964  <x:ref>transfer-parameter</x:ref> = <x:ref>token</x:ref> <x:ref>BWS</x:ref> "=" <x:ref>BWS</x:ref> ( <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref> )
1967   All transfer-coding names are case-insensitive and ought to be registered
1968   within the HTTP Transfer Coding registry, as defined in
1969   <xref target="transfer.coding.registry"/>.
1970   They are used in the <x:ref>TE</x:ref> (<xref target="header.te"/>) and
1971   <x:ref>Transfer-Encoding</x:ref> (<xref target="header.transfer-encoding"/>)
1972   header fields.
1975<section title="Chunked Transfer Coding" anchor="chunked.encoding">
1976  <iref primary="true" item="chunked (Coding Format)"/>
1977  <x:anchor-alias value="chunk"/>
1978  <x:anchor-alias value="chunked-body"/>
1979  <x:anchor-alias value="chunk-data"/>
1980  <x:anchor-alias value="chunk-size"/>
1981  <x:anchor-alias value="last-chunk"/>
1983   The chunked transfer coding wraps the payload body in order to transfer it
1984   as a series of chunks, each with its own size indicator, followed by an
1985   &OPTIONAL; trailer containing header fields. Chunked enables content
1986   streams of unknown size to be transferred as a sequence of length-delimited
1987   buffers, which enables the sender to retain connection persistence and the
1988   recipient to know when it has received the entire message.
1990<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="false" item="Grammar" subitem="trailer-part"/><iref primary="false" item="Grammar" subitem="chunk-ext"/><iref primary="true" item="Grammar" subitem="chunk-data"/>
1991  <x:ref>chunked-body</x:ref>   = *<x:ref>chunk</x:ref>
1992                   <x:ref>last-chunk</x:ref>
1993                   <x:ref>trailer-part</x:ref>
1994                   <x:ref>CRLF</x:ref>
1996  <x:ref>chunk</x:ref>          = <x:ref>chunk-size</x:ref> [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1997                   <x:ref>chunk-data</x:ref> <x:ref>CRLF</x:ref>
1998  <x:ref>chunk-size</x:ref>     = 1*<x:ref>HEXDIG</x:ref>
1999  <x:ref>last-chunk</x:ref>     = 1*("0") [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
2001  <x:ref>chunk-data</x:ref>     = 1*<x:ref>OCTET</x:ref> ; a sequence of chunk-size octets
2004   The chunk-size field is a string of hex digits indicating the size of
2005   the chunk-data in octets. The chunked transfer coding is complete when a
2006   chunk with a chunk-size of zero is received, possibly followed by a
2007   trailer, and finally terminated by an empty line.
2010   A recipient &MUST; be able to parse and decode the chunked transfer coding.
2013<section title="Chunk Extensions" anchor="chunked.extension">
2014  <x:anchor-alias value="chunk-ext"/>
2015  <x:anchor-alias value="chunk-ext-name"/>
2016  <x:anchor-alias value="chunk-ext-val"/>
2018   The chunked encoding allows each chunk to include zero or more chunk
2019   extensions, immediately following the <x:ref>chunk-size</x:ref>, for the
2020   sake of supplying per-chunk metadata (such as a signature or hash),
2021   mid-message control information, or randomization of message body size.
2023<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="chunked-body"><!--terminal production--></iref><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"/>
2024  <x:ref>chunk-ext</x:ref>      = *( ";" <x:ref>chunk-ext-name</x:ref> [ "=" <x:ref>chunk-ext-val</x:ref> ] )
2026  <x:ref>chunk-ext-name</x:ref> = <x:ref>token</x:ref>
2027  <x:ref>chunk-ext-val</x:ref>  = <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref>
2030   The chunked encoding is specific to each connection and is likely to be
2031   removed or recoded by each recipient (including intermediaries) before any
2032   higher-level application would have a chance to inspect the extensions.
2033   Hence, use of chunk extensions is generally limited to specialized HTTP
2034   services such as "long polling" (where client and server can have shared
2035   expectations regarding the use of chunk extensions) or for padding within
2036   an end-to-end secured connection.
2039   A recipient &MUST; ignore unrecognized chunk extensions.
2040   A server ought to limit the total length of chunk extensions received in a
2041   request to an amount reasonable for the services provided, in the same way
2042   that it applies length limitations and timeouts for other parts of a
2043   message, and generate an appropriate <x:ref>4xx (Client Error)</x:ref>
2044   response if that amount is exceeded.
2048<section title="Chunked Trailer Part" anchor="chunked.trailer.part">
2049  <x:anchor-alias value="trailer-part"/>
2051   A trailer allows the sender to include additional fields at the end of a
2052   chunked message in order to supply metadata that might be dynamically
2053   generated while the message body is sent, such as a message integrity
2054   check, digital signature, or post-processing status. The trailer fields are
2055   identical to header fields, except they are sent in a chunked trailer
2056   instead of the message's header section.
2058<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="trailer-part"/><iref primary="false" item="Grammar" subitem="header-field"/>
2059  <x:ref>trailer-part</x:ref>   = *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
2062   A sender &MUST-NOT; generate a trailer that contains a field necessary for
2063   message framing (e.g., <x:ref>Transfer-Encoding</x:ref> and
2064   <x:ref>Content-Length</x:ref>), routing (e.g., <x:ref>Host</x:ref>),
2065   request modifiers (e.g., controls and conditionals in
2066   &request-header-fields;), authentication (e.g., see <xref target="RFC7235"/>
2067   and <xref target="RFC6265"/>), response control data (e.g., see
2068   &response-control-data;), or determining how to process the payload
2069   (e.g., <x:ref>Content-Encoding</x:ref>, <x:ref>Content-Type</x:ref>,
2070   <x:ref>Content-Range</x:ref>, and <x:ref>Trailer</x:ref>).
2073   When a chunked message containing a non-empty trailer is received, the
2074   recipient &MAY; process the fields (aside from those forbidden above)
2075   as if they were appended to the message's header section.
2076   A recipient &MUST; ignore (or consider as an error) any fields that are
2077   forbidden to be sent in a trailer, since processing them as if they were
2078   present in the header section might bypass external security filters.
2081   Unless the request includes a <x:ref>TE</x:ref> header field indicating
2082   "trailers" is acceptable, as described in <xref target="header.te"/>, a
2083   server &SHOULD-NOT; generate trailer fields that it believes are necessary
2084   for the user agent to receive. Without a TE containing "trailers", the
2085   server ought to assume that the trailer fields might be silently discarded
2086   along the path to the user agent. This requirement allows intermediaries to
2087   forward a de-chunked message to an HTTP/1.0 recipient without buffering the
2088   entire response.
2092<section title="Decoding Chunked" anchor="decoding.chunked">
2094   A process for decoding the chunked transfer coding
2095   can be represented in pseudo-code as:
2097<figure><artwork type="code">
2098  length := 0
2099  read chunk-size, chunk-ext (if any), and CRLF
2100  while (chunk-size &gt; 0) {
2101     read chunk-data and CRLF
2102     append chunk-data to decoded-body
2103     length := length + chunk-size
2104     read chunk-size, chunk-ext (if any), and CRLF
2105  }
2106  read trailer field
2107  while (trailer field is not empty) {
2108     if (trailer field is allowed to be sent in a trailer) {
2109         append trailer field to existing header fields
2110     }
2111     read trailer-field
2112  }
2113  Content-Length := length
2114  Remove "chunked" from Transfer-Encoding
2115  Remove Trailer from existing header fields
2120<section title="Compression Codings" anchor="compression.codings">
2122   The codings defined below can be used to compress the payload of a
2123   message.
2126<section title="Compress Coding" anchor="compress.coding">
2127<iref item="compress (Coding Format)"/>
2129   The "compress" coding is an adaptive Lempel-Ziv-Welch (LZW) coding
2130   <xref target="Welch"/> that is commonly produced by the UNIX file
2131   compression program "compress".
2132   A recipient &SHOULD; consider "x-compress" to be equivalent to "compress".
2136<section title="Deflate Coding" anchor="deflate.coding">
2137<iref item="deflate (Coding Format)"/>
2139   The "deflate" coding is a "zlib" data format <xref target="RFC1950"/>
2140   containing a "deflate" compressed data stream <xref target="RFC1951"/>
2141   that uses a combination of the Lempel-Ziv (LZ77) compression algorithm and
2142   Huffman coding.
2145  <t>
2146    &Note; Some non-conformant implementations send the "deflate"
2147    compressed data without the zlib wrapper.
2148   </t>
2152<section title="Gzip Coding" anchor="gzip.coding">
2153<iref item="gzip (Coding Format)"/>
2155   The "gzip" coding is an LZ77 coding with a 32-bit Cyclic Redundancy Check
2156   (CRC) that is commonly
2157   produced by the gzip file compression program <xref target="RFC1952"/>.
2158   A recipient &SHOULD; consider "x-gzip" to be equivalent to "gzip".
2164<section title="TE" anchor="header.te">
2165  <iref primary="true" item="TE header field" x:for-anchor=""/>
2166  <x:anchor-alias value="TE"/>
2167  <x:anchor-alias value="t-codings"/>
2168  <x:anchor-alias value="t-ranking"/>
2169  <x:anchor-alias value="rank"/>
2171   The "TE" header field in a request indicates what transfer codings,
2172   besides chunked, the client is willing to accept in response, and
2173   whether or not the client is willing to accept trailer fields in a
2174   chunked transfer coding.
2177   The TE field-value consists of a comma-separated list of transfer coding
2178   names, each allowing for optional parameters (as described in
2179   <xref target="transfer.codings"/>), and/or the keyword "trailers".
2180   A client &MUST-NOT; send the chunked transfer coding name in TE;
2181   chunked is always acceptable for HTTP/1.1 recipients.
2183<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"/>
2184  <x:ref>TE</x:ref>        = #<x:ref>t-codings</x:ref>
2185  <x:ref>t-codings</x:ref> = "trailers" / ( <x:ref>transfer-coding</x:ref> [ <x:ref>t-ranking</x:ref> ] )
2186  <x:ref>t-ranking</x:ref> = <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> "q=" <x:ref>rank</x:ref>
2187  <x:ref>rank</x:ref>      = ( "0" [ "." 0*3<x:ref>DIGIT</x:ref> ] )
2188             / ( "1" [ "." 0*3("0") ] )
2191   Three examples of TE use are below.
2193<figure><artwork type="example">
2194  TE: deflate
2195  TE:
2196  TE: trailers, deflate;q=0.5
2199   The presence of the keyword "trailers" indicates that the client is willing
2200   to accept trailer fields in a chunked transfer coding, as defined in
2201   <xref target="chunked.trailer.part"/>, on behalf of itself and any downstream
2202   clients. For requests from an intermediary, this implies that either:
2203   (a) all downstream clients are willing to accept trailer fields in the
2204   forwarded response; or,
2205   (b) the intermediary will attempt to buffer the response on behalf of
2206   downstream recipients.
2207   Note that HTTP/1.1 does not define any means to limit the size of a
2208   chunked response such that an intermediary can be assured of buffering the
2209   entire response.
2212   When multiple transfer codings are acceptable, the client &MAY; rank the
2213   codings by preference using a case-insensitive "q" parameter (similar to
2214   the qvalues used in content negotiation fields, &qvalue;). The rank value
2215   is a real number in the range 0 through 1, where 0.001 is the least
2216   preferred and 1 is the most preferred; a value of 0 means "not acceptable".
2219   If the TE field-value is empty or if no TE field is present, the only
2220   acceptable transfer coding is chunked. A message with no transfer coding
2221   is always acceptable.
2224   Since the TE header field only applies to the immediate connection,
2225   a sender of TE &MUST; also send a "TE" connection option within the
2226   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
2227   in order to prevent the TE field from being forwarded by intermediaries
2228   that do not support its semantics.
2232<section title="Trailer" anchor="header.trailer">
2233  <iref primary="true" item="Trailer header field" x:for-anchor=""/>
2234  <x:anchor-alias value="Trailer"/>
2236   When a message includes a message body encoded with the chunked
2237   transfer coding and the sender desires to send metadata in the form of
2238   trailer fields at the end of the message, the sender &SHOULD; generate a
2239   <x:ref>Trailer</x:ref> header field before the message body to indicate
2240   which fields will be present in the trailers. This allows the recipient
2241   to prepare for receipt of that metadata before it starts processing the body,
2242   which is useful if the message is being streamed and the recipient wishes
2243   to confirm an integrity check on the fly.
2245<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/><iref primary="false" item="Grammar" subitem="field-name"/>
2246  <x:ref>Trailer</x:ref> = 1#<x:ref>field-name</x:ref>
2251<section title="Message Routing" anchor="message.routing">
2253   HTTP request message routing is determined by each client based on the
2254   target resource, the client's proxy configuration, and
2255   establishment or reuse of an inbound connection.  The corresponding
2256   response routing follows the same connection chain back to the client.
2259<section title="Identifying a Target Resource" anchor="target-resource">
2260  <iref primary="true" item="target resource"/>
2261  <iref primary="true" item="target URI"/>
2262  <x:anchor-alias value="target resource"/>
2263  <x:anchor-alias value="target URI"/>
2265   HTTP is used in a wide variety of applications, ranging from
2266   general-purpose computers to home appliances.  In some cases,
2267   communication options are hard-coded in a client's configuration.
2268   However, most HTTP clients rely on the same resource identification
2269   mechanism and configuration techniques as general-purpose Web browsers.
2272   HTTP communication is initiated by a user agent for some purpose.
2273   The purpose is a combination of request semantics, which are defined in
2274   <xref target="RFC7231"/>, and a target resource upon which to apply those
2275   semantics.  A URI reference (<xref target="uri"/>) is typically used as
2276   an identifier for the "<x:dfn>target resource</x:dfn>", which a user agent
2277   would resolve to its absolute form in order to obtain the
2278   "<x:dfn>target URI</x:dfn>".  The target URI
2279   excludes the reference's fragment component, if any,
2280   since fragment identifiers are reserved for client-side processing
2281   (<xref target="RFC3986" x:fmt="," x:sec="3.5"/>).
2285<section title="Connecting Inbound" anchor="connecting.inbound">
2287   Once the target URI is determined, a client needs to decide whether
2288   a network request is necessary to accomplish the desired semantics and,
2289   if so, where that request is to be directed.
2292   If the client has a cache <xref target="RFC7234"/> and the request can be
2293   satisfied by it, then the request is
2294   usually directed there first.
2297   If the request is not satisfied by a cache, then a typical client will
2298   check its configuration to determine whether a proxy is to be used to
2299   satisfy the request.  Proxy configuration is implementation-dependent,
2300   but is often based on URI prefix matching, selective authority matching,
2301   or both, and the proxy itself is usually identified by an "http" or
2302   "https" URI.  If a proxy is applicable, the client connects inbound by
2303   establishing (or reusing) a connection to that proxy.
2306   If no proxy is applicable, a typical client will invoke a handler routine,
2307   usually specific to the target URI's scheme, to connect directly
2308   to an authority for the target resource.  How that is accomplished is
2309   dependent on the target URI scheme and defined by its associated
2310   specification, similar to how this specification defines origin server
2311   access for resolution of the "http" (<xref target="http.uri"/>) and
2312   "https" (<xref target="https.uri"/>) schemes.
2315   HTTP requirements regarding connection management are defined in
2316   <xref target=""/>.
2320<section title="Request Target" anchor="request-target">
2322   Once an inbound connection is obtained,
2323   the client sends an HTTP request message (<xref target="http.message"/>)
2324   with a request-target derived from the target URI.
2325   There are four distinct formats for the request-target, depending on both
2326   the method being requested and whether the request is to a proxy.
2328<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-target"/><iref primary="false" item="Grammar" subitem="origin-form"/><iref primary="false" item="Grammar" subitem="absolute-form"/><iref primary="false" item="Grammar" subitem="authority-form"/><iref primary="false" item="Grammar" subitem="asterisk-form"/>
2329  <x:ref>request-target</x:ref> = <x:ref>origin-form</x:ref>
2330                 / <x:ref>absolute-form</x:ref>
2331                 / <x:ref>authority-form</x:ref>
2332                 / <x:ref>asterisk-form</x:ref>
2335<section title="origin-form" anchor="origin-form">
2336   <iref item="origin-form (of request-target)"/>
2338   The most common form of request-target is the <x:dfn>origin-form</x:dfn>.
2340<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="origin-form"/>
2341  <x:ref>origin-form</x:ref>    = <x:ref>absolute-path</x:ref> [ "?" <x:ref>query</x:ref> ]
2344   When making a request directly to an origin server, other than a CONNECT
2345   or server-wide OPTIONS request (as detailed below),
2346   a client &MUST; send only the absolute path and query components of
2347   the target URI as the request-target.
2348   If the target URI's path component is empty, the client &MUST; send
2349   "/" as the path within the origin-form of request-target.
2350   A <x:ref>Host</x:ref> header field is also sent, as defined in
2351   <xref target=""/>.
2354   For example, a client wishing to retrieve a representation of the resource
2355   identified as
2357<figure><artwork x:indent-with="  " type="example">
2361   directly from the origin server would open (or reuse) a TCP connection
2362   to port 80 of the host "" and send the lines:
2364<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2365GET /where?q=now HTTP/1.1
2369   followed by the remainder of the request message.
2373<section title="absolute-form" anchor="absolute-form">
2374   <iref item="absolute-form (of request-target)"/>
2376   When making a request to a proxy, other than a CONNECT or server-wide
2377   OPTIONS request (as detailed below), a client &MUST; send the target URI
2378   in <x:dfn>absolute-form</x:dfn> as the request-target.
2380<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="absolute-form"/>
2381  <x:ref>absolute-form</x:ref>  = <x:ref>absolute-URI</x:ref>
2384   The proxy is requested to either service that request from a valid cache,
2385   if possible, or make the same request on the client's behalf to either
2386   the next inbound proxy server or directly to the origin server indicated
2387   by the request-target.  Requirements on such "forwarding" of messages are
2388   defined in <xref target="message.forwarding"/>.
2391   An example absolute-form of request-line would be:
2393<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2394GET HTTP/1.1
2397   To allow for transition to the absolute-form for all requests in some
2398   future version of HTTP, a server &MUST; accept the absolute-form
2399   in requests, even though HTTP/1.1 clients will only send them in requests
2400   to proxies.
2404<section title="authority-form" anchor="authority-form">
2405   <iref item="authority-form (of request-target)"/>
2407   The <x:dfn>authority-form</x:dfn> of request-target is only used for
2408   CONNECT requests (&CONNECT;).
2410<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="authority-form"/>
2411  <x:ref>authority-form</x:ref> = <x:ref>authority</x:ref>
2414   When making a CONNECT request to establish a
2415   tunnel through one or more proxies, a client &MUST; send only the target
2416   URI's authority component (excluding any userinfo and its "@" delimiter) as
2417   the request-target. For example,
2419<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2424<section title="asterisk-form" anchor="asterisk-form">
2425   <iref item="asterisk-form (of request-target)"/>
2427   The <x:dfn>asterisk-form</x:dfn> of request-target is only used for a server-wide
2428   OPTIONS request (&OPTIONS;).
2430<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="asterisk-form"/>
2431  <x:ref>asterisk-form</x:ref>  = "*"
2434   When a client wishes to request OPTIONS
2435   for the server as a whole, as opposed to a specific named resource of
2436   that server, the client &MUST; send only "*" (%x2A) as the request-target.
2437   For example,
2439<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2440OPTIONS * HTTP/1.1
2443   If a proxy receives an OPTIONS request with an absolute-form of
2444   request-target in which the URI has an empty path and no query component,
2445   then the last proxy on the request chain &MUST; send a request-target
2446   of "*" when it forwards the request to the indicated origin server.
2449   For example, the request
2450</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2454  would be forwarded by the final proxy as
2455</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2456OPTIONS * HTTP/1.1
2460   after connecting to port 8001 of host "".
2466<section title="Host" anchor="">
2467  <iref primary="true" item="Host header field" x:for-anchor=""/>
2468  <x:anchor-alias value="Host"/>
2470   The "Host" header field in a request provides the host and port
2471   information from the target URI, enabling the origin
2472   server to distinguish among resources while servicing requests
2473   for multiple host names on a single IP address.
2475<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2476  <x:ref>Host</x:ref> = <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ; <xref target="http.uri"/>
2479   A client &MUST; send a Host header field in all HTTP/1.1 request messages.
2480   If the target URI includes an authority component, then a client &MUST;
2481   send a field-value for Host that is identical to that authority
2482   component, excluding any userinfo subcomponent and its "@" delimiter
2483   (<xref target="http.uri"/>).
2484   If the authority component is missing or undefined for the target URI,
2485   then a client &MUST; send a Host header field with an empty field-value.
2488   Since the Host field-value is critical information for handling a request,
2489   a user agent &SHOULD; generate Host as the first header field following the
2490   request-line.
2493   For example, a GET request to the origin server for
2494   &lt;; would begin with:
2496<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2497GET /pub/WWW/ HTTP/1.1
2501   A client &MUST; send a Host header field in an HTTP/1.1 request even
2502   if the request-target is in the absolute-form, since this
2503   allows the Host information to be forwarded through ancient HTTP/1.0
2504   proxies that might not have implemented Host.
2507   When a proxy receives a request with an absolute-form of
2508   request-target, the proxy &MUST; ignore the received
2509   Host header field (if any) and instead replace it with the host
2510   information of the request-target.  A proxy that forwards such a request
2511   &MUST; generate a new Host field-value based on the received
2512   request-target rather than forward the received Host field-value.
2515   Since the Host header field acts as an application-level routing
2516   mechanism, it is a frequent target for malware seeking to poison
2517   a shared cache or redirect a request to an unintended server.
2518   An interception proxy is particularly vulnerable if it relies on
2519   the Host field-value for redirecting requests to internal
2520   servers, or for use as a cache key in a shared cache, without
2521   first verifying that the intercepted connection is targeting a
2522   valid IP address for that host.
2525   A server &MUST; respond with a <x:ref>400 (Bad Request)</x:ref> status code
2526   to any HTTP/1.1 request message that lacks a Host header field and
2527   to any request message that contains more than one Host header field
2528   or a Host header field with an invalid field-value.
2532<section title="Effective Request URI" anchor="effective.request.uri">
2533  <iref primary="true" item="effective request URI"/>
2534  <x:anchor-alias value="effective request URI"/>
2536   Since the request-target often contains only part of the user agent's
2537   target URI, a server reconstructs the intended target as an
2538   "<x:dfn>effective request URI</x:dfn>" to properly service the request.
2539   This reconstruction involves both the server's local configuration and
2540   information communicated in the <x:ref>request-target</x:ref>,
2541   <x:ref>Host</x:ref> header field, and connection context.
2544   For a user agent, the effective request URI is the target URI.
2547   If the <x:ref>request-target</x:ref> is in <x:ref>absolute-form</x:ref>,
2548   the effective request URI is the same as the request-target. Otherwise, the
2549   effective request URI is constructed as follows:
2550<list style="empty">
2552   If the server's configuration (or outbound gateway) provides a fixed URI
2553   <x:ref>scheme</x:ref>, that scheme is used for the effective request URI.
2554   Otherwise, if the request is received over a TLS-secured TCP connection,
2555   the effective request URI's scheme is "https"; if not, the scheme is "http".
2558   If the server's configuration (or outbound gateway) provides a fixed URI
2559   <x:ref>authority</x:ref> component, that authority is used for the
2560   effective request URI. If not, then if the request-target is in
2561   <x:ref>authority-form</x:ref>, the effective request URI's authority
2562   component is the same as the request-target.
2563   If not, then if a <x:ref>Host</x:ref> header field is supplied with a
2564   non-empty field-value, the authority component is the same as the
2565   Host field-value. Otherwise, the authority component is assigned
2566   the default name configured for the server and, if the connection's
2567   incoming TCP port number differs from the default port for the effective
2568   request URI's scheme, then a colon (":") and the incoming port number (in
2569   decimal form) are appended to the authority component.
2572   If the request-target is in <x:ref>authority-form</x:ref> or
2573   <x:ref>asterisk-form</x:ref>, the effective request URI's combined
2574   <x:ref>path</x:ref> and <x:ref>query</x:ref> component is empty. Otherwise,
2575   the combined <x:ref>path</x:ref> and <x:ref>query</x:ref> component is the
2576   same as the request-target.
2579   The components of the effective request URI, once determined as above, can
2580   be combined into <x:ref>absolute-URI</x:ref> form by concatenating the
2581   scheme, "://", authority, and combined path and query component.
2587   Example 1: the following message received over an insecure TCP connection
2589<artwork type="example" x:indent-with="  ">
2590GET /pub/WWW/TheProject.html HTTP/1.1
2596  has an effective request URI of
2598<artwork type="example" x:indent-with="  ">
2604   Example 2: the following message received over a TLS-secured TCP connection
2606<artwork type="example" x:indent-with="  ">
2607OPTIONS * HTTP/1.1
2613  has an effective request URI of
2615<artwork type="example" x:indent-with="  ">
2620   Recipients of an HTTP/1.0 request that lacks a <x:ref>Host</x:ref> header
2621   field might need to use heuristics (e.g., examination of the URI path for
2622   something unique to a particular host) in order to guess the
2623   effective request URI's authority component.
2626   Once the effective request URI has been constructed, an origin server needs
2627   to decide whether or not to provide service for that URI via the connection
2628   in which the request was received. For example, the request might have been
2629   misdirected, deliberately or accidentally, such that the information within
2630   a received <x:ref>request-target</x:ref> or <x:ref>Host</x:ref> header
2631   field differs from the host or port upon which the connection has been
2632   made. If the connection is from a trusted gateway, that inconsistency might
2633   be expected; otherwise, it might indicate an attempt to bypass security
2634   filters, trick the server into delivering non-public content, or poison a
2635   cache. See <xref target="security.considerations"/> for security
2636   considerations regarding message routing.
2640<section title="Associating a Response to a Request" anchor="">
2642   HTTP does not include a request identifier for associating a given
2643   request message with its corresponding one or more response messages.
2644   Hence, it relies on the order of response arrival to correspond exactly
2645   to the order in which requests are made on the same connection.
2646   More than one response message per request only occurs when one or more
2647   informational responses (<x:ref>1xx</x:ref>, see &status-1xx;) precede a
2648   final response to the same request.
2651   A client that has more than one outstanding request on a connection &MUST;
2652   maintain a list of outstanding requests in the order sent and &MUST;
2653   associate each received response message on that connection to the highest
2654   ordered request that has not yet received a final (non-<x:ref>1xx</x:ref>)
2655   response.
2659<section title="Message Forwarding" anchor="message.forwarding">
2661   As described in <xref target="intermediaries"/>, intermediaries can serve
2662   a variety of roles in the processing of HTTP requests and responses.
2663   Some intermediaries are used to improve performance or availability.
2664   Others are used for access control or to filter content.
2665   Since an HTTP stream has characteristics similar to a pipe-and-filter
2666   architecture, there are no inherent limits to the extent an intermediary
2667   can enhance (or interfere) with either direction of the stream.
2670   An intermediary not acting as a tunnel &MUST; implement the
2671   <x:ref>Connection</x:ref> header field, as specified in
2672   <xref target="header.connection"/>, and exclude fields from being forwarded
2673   that are only intended for the incoming connection.
2676   An intermediary &MUST-NOT; forward a message to itself unless it is
2677   protected from an infinite request loop. In general, an intermediary ought
2678   to recognize its own server names, including any aliases, local variations,
2679   or literal IP addresses, and respond to such requests directly.
2682<section title="Via" anchor="header.via">
2683  <iref primary="true" item="Via header field" x:for-anchor=""/>
2684  <x:anchor-alias value="pseudonym"/>
2685  <x:anchor-alias value="received-by"/>
2686  <x:anchor-alias value="received-protocol"/>
2687  <x:anchor-alias value="Via"/>
2689   The "Via" header field indicates the presence of intermediate protocols and
2690   recipients between the user agent and the server (on requests) or between
2691   the origin server and the client (on responses), similar to the
2692   "Received" header field in email
2693   (<xref target="RFC5322" x:fmt="of" x:sec="3.6.7"/>).
2694   Via can be used for tracking message forwards,
2695   avoiding request loops, and identifying the protocol capabilities of
2696   senders along the request/response chain.
2698<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"/>
2699  <x:ref>Via</x:ref> = 1#( <x:ref>received-protocol</x:ref> <x:ref>RWS</x:ref> <x:ref>received-by</x:ref> [ <x:ref>RWS</x:ref> <x:ref>comment</x:ref> ] )
2701  <x:ref>received-protocol</x:ref> = [ <x:ref>protocol-name</x:ref> "/" ] <x:ref>protocol-version</x:ref>
2702                      ; see <xref target="header.upgrade"/>
2703  <x:ref>received-by</x:ref>       = ( <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ) / <x:ref>pseudonym</x:ref>
2704  <x:ref>pseudonym</x:ref>         = <x:ref>token</x:ref>
2707   Multiple Via field values represent each proxy or gateway that has
2708   forwarded the message. Each intermediary appends its own information
2709   about how the message was received, such that the end result is ordered
2710   according to the sequence of forwarding recipients.
2713   A proxy &MUST; send an appropriate Via header field, as described below, in
2714   each message that it forwards.
2715   An HTTP-to-HTTP gateway &MUST; send an appropriate Via header field in
2716   each inbound request message and &MAY; send a Via header field in
2717   forwarded response messages.
2720   For each intermediary, the received-protocol indicates the protocol and
2721   protocol version used by the upstream sender of the message. Hence, the
2722   Via field value records the advertised protocol capabilities of the
2723   request/response chain such that they remain visible to downstream
2724   recipients; this can be useful for determining what backwards-incompatible
2725   features might be safe to use in response, or within a later request, as
2726   described in <xref target="http.version"/>. For brevity, the protocol-name
2727   is omitted when the received protocol is HTTP.
2730   The received-by portion of the field value is normally the host and optional
2731   port number of a recipient server or client that subsequently forwarded the
2732   message.
2733   However, if the real host is considered to be sensitive information, a
2734   sender &MAY; replace it with a pseudonym. If a port is not provided,
2735   a recipient &MAY; interpret that as meaning it was received on the default
2736   TCP port, if any, for the received-protocol.
2739   A sender &MAY; generate comments in the Via header field to identify the
2740   software of each recipient, analogous to the <x:ref>User-Agent</x:ref> and
2741   <x:ref>Server</x:ref> header fields. However, all comments in the Via field
2742   are optional, and a recipient &MAY; remove them prior to forwarding the
2743   message.
2746   For example, a request message could be sent from an HTTP/1.0 user
2747   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2748   forward the request to a public proxy at, which completes
2749   the request by forwarding it to the origin server at
2750   The request received by would then have the following
2751   Via header field:
2753<figure><artwork type="example">
2754  Via: 1.0 fred, 1.1
2757   An intermediary used as a portal through a network firewall
2758   &SHOULD-NOT; forward the names and ports of hosts within the firewall
2759   region unless it is explicitly enabled to do so. If not enabled, such an
2760   intermediary &SHOULD; replace each received-by host of any host behind the
2761   firewall by an appropriate pseudonym for that host.
2764   An intermediary &MAY; combine an ordered subsequence of Via header
2765   field entries into a single such entry if the entries have identical
2766   received-protocol values. For example,
2768<figure><artwork type="example">
2769  Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2772  could be collapsed to
2774<figure><artwork type="example">
2775  Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2778   A sender &SHOULD-NOT; combine multiple entries unless they are all
2779   under the same organizational control and the hosts have already been
2780   replaced by pseudonyms. A sender &MUST-NOT; combine entries that
2781   have different received-protocol values.
2785<section title="Transformations" anchor="message.transformations">
2786   <iref primary="true" item="transforming proxy"/>
2787   <iref primary="true" item="non-transforming proxy"/>
2789   Some intermediaries include features for transforming messages and their
2790   payloads. A proxy might, for example, convert between image formats in
2791   order to save cache space or to reduce the amount of traffic on a slow
2792   link. However, operational problems might occur when these transformations
2793   are applied to payloads intended for critical applications, such as medical
2794   imaging or scientific data analysis, particularly when integrity checks or
2795   digital signatures are used to ensure that the payload received is
2796   identical to the original.
2799   An HTTP-to-HTTP proxy is called a "<x:dfn>transforming proxy</x:dfn>"
2800   if it is designed or configured to modify messages in a semantically
2801   meaningful way (i.e., modifications, beyond those required by normal
2802   HTTP processing, that change the message in a way that would be
2803   significant to the original sender or potentially significant to
2804   downstream recipients).  For example, a transforming proxy might be
2805   acting as a shared annotation server (modifying responses to include
2806   references to a local annotation database), a malware filter, a
2807   format transcoder, or a privacy filter. Such transformations are presumed
2808   to be desired by whichever client (or client organization) selected the
2809   proxy.
2812   If a proxy receives a request-target with a host name that is not a
2813   fully qualified domain name, it &MAY; add its own domain to the host name
2814   it received when forwarding the request.  A proxy &MUST-NOT; change the
2815   host name if the request-target contains a fully qualified domain name.
2818   A proxy &MUST-NOT; modify the "absolute-path" and "query" parts of the
2819   received request-target when forwarding it to the next inbound server,
2820   except as noted above to replace an empty path with "/" or "*".
2823   A proxy &MAY; modify the message body through application
2824   or removal of a transfer coding (<xref target="transfer.codings"/>).
2827   A proxy &MUST-NOT; transform the payload (&payload;) of a message that
2828   contains a no-transform cache-control directive (&header-cache-control;).
2831   A proxy &MAY; transform the payload of a message
2832   that does not contain a no-transform cache-control directive.
2833   A proxy that transforms a payload &MUST; add a <x:ref>Warning</x:ref>
2834   header field with the warn-code of 214 ("Transformation Applied")
2835   if one is not already in the message (see &header-warning;).
2836   A proxy that transforms the payload of a <x:ref>200 (OK)</x:ref> response
2837   can further inform downstream recipients that a transformation has been
2838   applied by changing the response status code to
2839   <x:ref>203 (Non-Authoritative Information)</x:ref> (&status-203;).
2842   A proxy &SHOULD-NOT; modify header fields that provide information about
2843   the end points of the communication chain, the resource state, or the
2844   selected representation (other than the payload) unless the field's
2845   definition specifically allows such modification or the modification is
2846   deemed necessary for privacy or security.
2852<section title="Connection Management" anchor="">
2854   HTTP messaging is independent of the underlying transport or
2855   session-layer connection protocol(s).  HTTP only presumes a reliable
2856   transport with in-order delivery of requests and the corresponding
2857   in-order delivery of responses.  The mapping of HTTP request and
2858   response structures onto the data units of an underlying transport
2859   protocol is outside the scope of this specification.
2862   As described in <xref target="connecting.inbound"/>, the specific
2863   connection protocols to be used for an HTTP interaction are determined by
2864   client configuration and the <x:ref>target URI</x:ref>.
2865   For example, the "http" URI scheme
2866   (<xref target="http.uri"/>) indicates a default connection of TCP
2867   over IP, with a default TCP port of 80, but the client might be
2868   configured to use a proxy via some other connection, port, or protocol.
2871   HTTP implementations are expected to engage in connection management,
2872   which includes maintaining the state of current connections,
2873   establishing a new connection or reusing an existing connection,
2874   processing messages received on a connection, detecting connection
2875   failures, and closing each connection.
2876   Most clients maintain multiple connections in parallel, including
2877   more than one connection per server endpoint.
2878   Most servers are designed to maintain thousands of concurrent connections,
2879   while controlling request queues to enable fair use and detect
2880   denial-of-service attacks.
2883<section title="Connection" anchor="header.connection">
2884  <iref primary="true" item="Connection header field" x:for-anchor=""/>
2885  <iref primary="true" item="close" x:for-anchor=""/>
2886  <x:anchor-alias value="Connection"/>
2887  <x:anchor-alias value="connection-option"/>
2888  <x:anchor-alias value="close"/>
2890   The "Connection" header field allows the sender to indicate desired
2891   control options for the current connection.  In order to avoid confusing
2892   downstream recipients, a proxy or gateway &MUST; remove or replace any
2893   received connection options before forwarding the message.
2896   When a header field aside from Connection is used to supply control
2897   information for or about the current connection, the sender &MUST; list
2898   the corresponding field-name within the "Connection" header field.
2899   A proxy or gateway &MUST; parse a received Connection
2900   header field before a message is forwarded and, for each
2901   connection-option in this field, remove any header field(s) from
2902   the message with the same name as the connection-option, and then
2903   remove the Connection header field itself (or replace it with the
2904   intermediary's own connection options for the forwarded message).
2907   Hence, the Connection header field provides a declarative way of
2908   distinguishing header fields that are only intended for the
2909   immediate recipient ("hop-by-hop") from those fields that are
2910   intended for all recipients on the chain ("end-to-end"), enabling the
2911   message to be self-descriptive and allowing future connection-specific
2912   extensions to be deployed without fear that they will be blindly
2913   forwarded by older intermediaries.
2916   The Connection header field's value has the following grammar:
2918<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-option"/>
2919  <x:ref>Connection</x:ref>        = 1#<x:ref>connection-option</x:ref>
2920  <x:ref>connection-option</x:ref> = <x:ref>token</x:ref>
2923   Connection options are case-insensitive.
2926   A sender &MUST-NOT; send a connection option corresponding to a header
2927   field that is intended for all recipients of the payload.
2928   For example, <x:ref>Cache-Control</x:ref> is never appropriate as a
2929   connection option (&header-cache-control;).
2932   The connection options do not always correspond to a header field
2933   present in the message, since a connection-specific header field
2934   might not be needed if there are no parameters associated with a
2935   connection option. In contrast, a connection-specific header field that
2936   is received without a corresponding connection option usually indicates
2937   that the field has been improperly forwarded by an intermediary and
2938   ought to be ignored by the recipient.
2941   When defining new connection options, specification authors ought to survey
2942   existing header field names and ensure that the new connection option does
2943   not share the same name as an already deployed header field.
2944   Defining a new connection option essentially reserves that potential
2945   field-name for carrying additional information related to the
2946   connection option, since it would be unwise for senders to use
2947   that field-name for anything else.
2950   The "<x:dfn>close</x:dfn>" connection option is defined for a
2951   sender to signal that this connection will be closed after completion of
2952   the response. For example,
2954<figure><artwork type="example">
2955  Connection: close
2958   in either the request or the response header fields indicates that the
2959   sender is going to close the connection after the current request/response
2960   is complete (<xref target="persistent.tear-down"/>).
2963   A client that does not support <x:ref>persistent connections</x:ref> &MUST;
2964   send the "close" connection option in every request message.
2967   A server that does not support <x:ref>persistent connections</x:ref> &MUST;
2968   send the "close" connection option in every response message that
2969   does not have a <x:ref>1xx (Informational)</x:ref> status code.
2973<section title="Establishment" anchor="persistent.establishment">
2975   It is beyond the scope of this specification to describe how connections
2976   are established via various transport or session-layer protocols.
2977   Each connection applies to only one transport link.
2981<section title="Persistence" anchor="persistent.connections">
2982   <x:anchor-alias value="persistent connections"/>
2984   HTTP/1.1 defaults to the use of "<x:dfn>persistent connections</x:dfn>",
2985   allowing multiple requests and responses to be carried over a single
2986   connection. The "<x:ref>close</x:ref>" connection-option is used to signal
2987   that a connection will not persist after the current request/response.
2988   HTTP implementations &SHOULD; support persistent connections.
2991   A recipient determines whether a connection is persistent or not based on
2992   the most recently received message's protocol version and
2993   <x:ref>Connection</x:ref> header field (if any):
2994   <list style="symbols">
2995     <t>If the <x:ref>close</x:ref> connection option is present, the
2996        connection will not persist after the current response; else,</t>
2997     <t>If the received protocol is HTTP/1.1 (or later), the connection will
2998        persist after the current response; else,</t>
2999     <t>If the received protocol is HTTP/1.0, the "keep-alive"
3000        connection option is present, the recipient is not a proxy, and
3001        the recipient wishes to honor the HTTP/1.0 "keep-alive" mechanism,
3002        the connection will persist after the current response; otherwise,</t>
3003     <t>The connection will close after the current response.</t>
3004   </list>
3007   A client &MAY; send additional requests on a persistent connection until it
3008   sends or receives a <x:ref>close</x:ref> connection option or receives an
3009   HTTP/1.0 response without a "keep-alive" connection option.
3012   In order to remain persistent, all messages on a connection need to
3013   have a self-defined message length (i.e., one not defined by closure
3014   of the connection), as described in <xref target="message.body"/>.
3015   A server &MUST; read the entire request message body or close
3016   the connection after sending its response, since otherwise the
3017   remaining data on a persistent connection would be misinterpreted
3018   as the next request.  Likewise,
3019   a client &MUST; read the entire response message body if it intends
3020   to reuse the same connection for a subsequent request.
3023   A proxy server &MUST-NOT; maintain a persistent connection with an
3024   HTTP/1.0 client (see <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/> for
3025   information and discussion of the problems with the Keep-Alive header field
3026   implemented by many HTTP/1.0 clients).
3029   See <xref target="compatibility.with.http.1.0.persistent.connections"/>
3030   for more information on backwards compatibility with HTTP/1.0 clients.
3033<section title="Retrying Requests" anchor="persistent.retrying.requests">
3035   Connections can be closed at any time, with or without intention.
3036   Implementations ought to anticipate the need to recover
3037   from asynchronous close events.
3040   When an inbound connection is closed prematurely, a client &MAY; open a new
3041   connection and automatically retransmit an aborted sequence of requests if
3042   all of those requests have idempotent methods (&idempotent-methods;).
3043   A proxy &MUST-NOT; automatically retry non-idempotent requests.
3046   A user agent &MUST-NOT; automatically retry a request with a non-idempotent
3047   method unless it has some means to know that the request semantics are
3048   actually idempotent, regardless of the method, or some means to detect that
3049   the original request was never applied. For example, a user agent that
3050   knows (through design or configuration) that a POST request to a given
3051   resource is safe can repeat that request automatically.
3052   Likewise, a user agent designed specifically to operate on a version
3053   control repository might be able to recover from partial failure conditions
3054   by checking the target resource revision(s) after a failed connection,
3055   reverting or fixing any changes that were partially applied, and then
3056   automatically retrying the requests that failed.
3059   A client &SHOULD-NOT; automatically retry a failed automatic retry.
3063<section title="Pipelining" anchor="pipelining">
3064   <x:anchor-alias value="pipeline"/>
3066   A client that supports persistent connections &MAY; "<x:dfn>pipeline</x:dfn>"
3067   its requests (i.e., send multiple requests without waiting for each
3068   response). A server &MAY; process a sequence of pipelined requests in
3069   parallel if they all have safe methods (&safe-methods;), but it &MUST; send
3070   the corresponding responses in the same order that the requests were
3071   received.
3074   A client that pipelines requests &SHOULD; retry unanswered requests if the
3075   connection closes before it receives all of the corresponding responses.
3076   When retrying pipelined requests after a failed connection (a connection
3077   not explicitly closed by the server in its last complete response), a
3078   client &MUST-NOT; pipeline immediately after connection establishment,
3079   since the first remaining request in the prior pipeline might have caused
3080   an error response that can be lost again if multiple requests are sent on a
3081   prematurely closed connection (see the TCP reset problem described in
3082   <xref target="persistent.tear-down"/>).
3085   Idempotent methods (&idempotent-methods;) are significant to pipelining
3086   because they can be automatically retried after a connection failure.
3087   A user agent &SHOULD-NOT; pipeline requests after a non-idempotent method,
3088   until the final response status code for that method has been received,
3089   unless the user agent has a means to detect and recover from partial
3090   failure conditions involving the pipelined sequence.
3093   An intermediary that receives pipelined requests &MAY; pipeline those
3094   requests when forwarding them inbound, since it can rely on the outbound
3095   user agent(s) to determine what requests can be safely pipelined. If the
3096   inbound connection fails before receiving a response, the pipelining
3097   intermediary &MAY; attempt to retry a sequence of requests that have yet
3098   to receive a response if the requests all have idempotent methods;
3099   otherwise, the pipelining intermediary &SHOULD; forward any received
3100   responses and then close the corresponding outbound connection(s) so that
3101   the outbound user agent(s) can recover accordingly.
3106<section title="Concurrency" anchor="persistent.concurrency">
3108   A client ought to limit the number of simultaneous open
3109   connections that it maintains to a given server.
3112   Previous revisions of HTTP gave a specific number of connections as a
3113   ceiling, but this was found to be impractical for many applications. As a
3114   result, this specification does not mandate a particular maximum number of
3115   connections but, instead, encourages clients to be conservative when opening
3116   multiple connections.
3119   Multiple connections are typically used to avoid the "head-of-line
3120   blocking" problem, wherein a request that takes significant server-side
3121   processing and/or has a large payload blocks subsequent requests on the
3122   same connection. However, each connection consumes server resources.
3123   Furthermore, using multiple connections can cause undesirable side effects
3124   in congested networks.
3127   Note that a server might reject traffic that it deems abusive or
3128   characteristic of a denial-of-service attack, such as an excessive number
3129   of open connections from a single client.
3133<section title="Failures and Timeouts" anchor="persistent.failures">
3135   Servers will usually have some timeout value beyond which they will
3136   no longer maintain an inactive connection. Proxy servers might make
3137   this a higher value since it is likely that the client will be making
3138   more connections through the same proxy server. The use of persistent
3139   connections places no requirements on the length (or existence) of
3140   this timeout for either the client or the server.
3143   A client or server that wishes to time out &SHOULD; issue a graceful close
3144   on the connection. Implementations &SHOULD; constantly monitor open
3145   connections for a received closure signal and respond to it as appropriate,
3146   since prompt closure of both sides of a connection enables allocated system
3147   resources to be reclaimed.
3150   A client, server, or proxy &MAY; close the transport connection at any
3151   time. For example, a client might have started to send a new request
3152   at the same time that the server has decided to close the "idle"
3153   connection. From the server's point of view, the connection is being
3154   closed while it was idle, but from the client's point of view, a
3155   request is in progress.
3158   A server &SHOULD; sustain persistent connections, when possible, and allow
3159   the underlying
3160   transport's flow control mechanisms to resolve temporary overloads, rather
3161   than terminate connections with the expectation that clients will retry.
3162   The latter technique can exacerbate network congestion.
3165   A client sending a message body &SHOULD; monitor
3166   the network connection for an error response while it is transmitting
3167   the request. If the client sees a response that indicates the server does
3168   not wish to receive the message body and is closing the connection, the
3169   client &SHOULD; immediately cease transmitting the body and close its side
3170   of the connection.
3174<section title="Tear-down" anchor="persistent.tear-down">
3175  <iref primary="false" item="Connection header field" x:for-anchor=""/>
3176  <iref primary="false" item="close" x:for-anchor=""/>
3178   The <x:ref>Connection</x:ref> header field
3179   (<xref target="header.connection"/>) provides a "<x:ref>close</x:ref>"
3180   connection option that a sender &SHOULD; send when it wishes to close
3181   the connection after the current request/response pair.
3184   A client that sends a <x:ref>close</x:ref> connection option &MUST-NOT;
3185   send further requests on that connection (after the one containing
3186   <x:ref>close</x:ref>) and &MUST; close the connection after reading the
3187   final response message corresponding to this request.
3190   A server that receives a <x:ref>close</x:ref> connection option &MUST;
3191   initiate a close of the connection (see below) after it sends the
3192   final response to the request that contained <x:ref>close</x:ref>.
3193   The server &SHOULD; send a <x:ref>close</x:ref> connection option
3194   in its final response on that connection. The server &MUST-NOT; process
3195   any further requests received on that connection.
3198   A server that sends a <x:ref>close</x:ref> connection option &MUST;
3199   initiate a close of the connection (see below) after it sends the
3200   response containing <x:ref>close</x:ref>. The server &MUST-NOT; process
3201   any further requests received on that connection.
3204   A client that receives a <x:ref>close</x:ref> connection option &MUST;
3205   cease sending requests on that connection and close the connection
3206   after reading the response message containing the close; if additional
3207   pipelined requests had been sent on the connection, the client &SHOULD-NOT;
3208   assume that they will be processed by the server.
3211   If a server performs an immediate close of a TCP connection, there is a
3212   significant risk that the client will not be able to read the last HTTP
3213   response.  If the server receives additional data from the client on a
3214   fully-closed connection, such as another request that was sent by the
3215   client before receiving the server's response, the server's TCP stack will
3216   send a reset packet to the client; unfortunately, the reset packet might
3217   erase the client's unacknowledged input buffers before they can be read
3218   and interpreted by the client's HTTP parser.
3221   To avoid the TCP reset problem, servers typically close a connection in
3222   stages. First, the server performs a half-close by closing only the write
3223   side of the read/write connection. The server then continues to read from
3224   the connection until it receives a corresponding close by the client, or
3225   until the server is reasonably certain that its own TCP stack has received
3226   the client's acknowledgement of the packet(s) containing the server's last
3227   response. Finally, the server fully closes the connection.
3230   It is unknown whether the reset problem is exclusive to TCP or might also
3231   be found in other transport connection protocols.
3235<section title="Upgrade" anchor="header.upgrade">
3236  <iref primary="true" item="Upgrade header field" x:for-anchor=""/>
3237  <x:anchor-alias value="Upgrade"/>
3238  <x:anchor-alias value="protocol"/>
3239  <x:anchor-alias value="protocol-name"/>
3240  <x:anchor-alias value="protocol-version"/>
3242   The "Upgrade" header field is intended to provide a simple mechanism
3243   for transitioning from HTTP/1.1 to some other protocol on the same
3244   connection.  A client &MAY; send a list of protocols in the Upgrade
3245   header field of a request to invite the server to switch to one or
3246   more of those protocols, in order of descending preference, before sending
3247   the final response. A server &MAY; ignore a received Upgrade header field
3248   if it wishes to continue using the current protocol on that connection.
3249   Upgrade cannot be used to insist on a protocol change.
3251<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
3252  <x:ref>Upgrade</x:ref>          = 1#<x:ref>protocol</x:ref>
3254  <x:ref>protocol</x:ref>         = <x:ref>protocol-name</x:ref> ["/" <x:ref>protocol-version</x:ref>]
3255  <x:ref>protocol-name</x:ref>    = <x:ref>token</x:ref>
3256  <x:ref>protocol-version</x:ref> = <x:ref>token</x:ref>
3259   A server that sends a <x:ref>101 (Switching Protocols)</x:ref> response
3260   &MUST; send an Upgrade header field to indicate the new protocol(s) to
3261   which the connection is being switched; if multiple protocol layers are
3262   being switched, the sender &MUST; list the protocols in layer-ascending
3263   order. A server &MUST-NOT; switch to a protocol that was not indicated by
3264   the client in the corresponding request's Upgrade header field.
3265   A server &MAY; choose to ignore the order of preference indicated by the
3266   client and select the new protocol(s) based on other factors, such as the
3267   nature of the request or the current load on the server.
3270   A server that sends a <x:ref>426 (Upgrade Required)</x:ref> response
3271   &MUST; send an Upgrade header field to indicate the acceptable protocols,
3272   in order of descending preference.
3275   A server &MAY; send an Upgrade header field in any other response to
3276   advertise that it implements support for upgrading to the listed protocols,
3277   in order of descending preference, when appropriate for a future request.
3280   The following is a hypothetical example sent by a client:
3281</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
3282GET /hello.txt HTTP/1.1
3284Connection: upgrade
3285Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
3289   The capabilities and nature of the
3290   application-level communication after the protocol change is entirely
3291   dependent upon the new protocol(s) chosen. However, immediately after
3292   sending the 101 response, the server is expected to continue responding to
3293   the original request as if it had received its equivalent within the new
3294   protocol (i.e., the server still has an outstanding request to satisfy
3295   after the protocol has been changed, and is expected to do so without
3296   requiring the request to be repeated).
3299   For example, if the Upgrade header field is received in a GET request
3300   and the server decides to switch protocols, it first responds
3301   with a <x:ref>101 (Switching Protocols)</x:ref> message in HTTP/1.1 and
3302   then immediately follows that with the new protocol's equivalent of a
3303   response to a GET on the target resource.  This allows a connection to be
3304   upgraded to protocols with the same semantics as HTTP without the
3305   latency cost of an additional round trip.  A server &MUST-NOT; switch
3306   protocols unless the received message semantics can be honored by the new
3307   protocol; an OPTIONS request can be honored by any protocol.
3310   The following is an example response to the above hypothetical request:
3311</preamble><artwork type="message/http; msgtype=&#34;response&#34;" x:indent-with="  ">
3312HTTP/1.1 101 Switching Protocols
3313Connection: upgrade
3314Upgrade: HTTP/2.0
3316[... data stream switches to HTTP/2.0 with an appropriate response
3317(as defined by new protocol) to the "GET /hello.txt" request ...]
3320   When Upgrade is sent, the sender &MUST; also send a
3321   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
3322   that contains an "upgrade" connection option, in order to prevent Upgrade
3323   from being accidentally forwarded by intermediaries that might not implement
3324   the listed protocols.  A server &MUST; ignore an Upgrade header field that
3325   is received in an HTTP/1.0 request.
3328   A client cannot begin using an upgraded protocol on the connection until
3329   it has completely sent the request message (i.e., the client can't change
3330   the protocol it is sending in the middle of a message).
3331   If a server receives both an Upgrade and an <x:ref>Expect</x:ref> header field
3332   with the "100-continue" expectation (&header-expect;), the
3333   server &MUST; send a <x:ref>100 (Continue)</x:ref> response before sending
3334   a <x:ref>101 (Switching Protocols)</x:ref> response.
3337   The Upgrade header field only applies to switching protocols on top of the
3338   existing connection; it cannot be used to switch the underlying connection
3339   (transport) protocol, nor to switch the existing communication to a
3340   different connection. For those purposes, it is more appropriate to use a
3341   <x:ref>3xx (Redirection)</x:ref> response (&status-3xx;).
3344   This specification only defines the protocol name "HTTP" for use by
3345   the family of Hypertext Transfer Protocols, as defined by the HTTP
3346   version rules of <xref target="http.version"/> and future updates to this
3347   specification. Additional tokens ought to be registered with IANA using the
3348   registration procedure defined in <xref target="upgrade.token.registry"/>.
3353<section title="ABNF List Extension: #rule" anchor="abnf.extension">
3355   A #rule extension to the ABNF rules of <xref target="RFC5234"/> is used to
3356   improve readability in the definitions of some header field values.
3359   A construct "#" is defined, similar to "*", for defining comma-delimited
3360   lists of elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating
3361   at least &lt;n&gt; and at most &lt;m&gt; elements, each separated by a single
3362   comma (",") and optional whitespace (OWS).   
3365   In any production that uses the list construct, a sender &MUST-NOT;
3366   generate empty list elements. In other words, a sender &MUST; generate
3367   lists that satisfy the following syntax:
3368</preamble><artwork type="example">
3369  1#element =&gt; element *( OWS "," OWS element )
3372   and:
3373</preamble><artwork type="example">
3374  #element =&gt; [ 1#element ]
3377   and for n &gt;= 1 and m &gt; 1:
3378</preamble><artwork type="example">
3379  &lt;n&gt;#&lt;m&gt;element =&gt; element &lt;n-1&gt;*&lt;m-1&gt;( OWS "," OWS element )
3382   For compatibility with legacy list rules, a recipient &MUST; parse and ignore
3383   a reasonable number of empty list elements: enough to handle common mistakes
3384   by senders that merge values, but not so much that they could be used as a
3385   denial-of-service mechanism. In other words, a recipient &MUST; accept lists
3386   that satisfy the following syntax:
3388<figure><artwork type="example">
3389  #element =&gt; [ ( "," / element ) *( OWS "," [ OWS element ] ) ]
3391  1#element =&gt; *( "," OWS ) element *( OWS "," [ OWS element ] )
3394   Empty elements do not contribute to the count of elements present.
3395   For example, given these ABNF productions:
3397<figure><artwork type="example">
3398  example-list      = 1#example-list-elmt
3399  example-list-elmt = token ; see <xref target="field.components"/>
3402   Then the following are valid values for example-list (not including the
3403   double quotes, which are present for delimitation only):
3405<figure><artwork type="example">
3406  "foo,bar"
3407  "foo ,bar,"
3408  "foo , ,bar,charlie   "
3411   In contrast, the following values would be invalid, since at least one
3412   non-empty element is required by the example-list production:
3414<figure><artwork type="example">
3415  ""
3416  ","
3417  ",   ,"
3420   <xref target="collected.abnf"/> shows the collected ABNF for recipients
3421   after the list constructs have been expanded.
3425<section title="IANA Considerations" anchor="IANA.considerations">
3427<section title="Header Field Registration" anchor="header.field.registration">
3429   HTTP header fields are registered within the Message Header Field Registry
3430   maintained at
3431   <eref target=""/>.
3434   This document defines the following HTTP header fields, so their
3435   associated registry entries shall be updated according to the permanent
3436   registrations below (see <xref target="BCP90"/>):
3438<?BEGININC p1-messaging.iana-headers ?>
3439<!--AUTOGENERATED FROM extract-header-defs.xslt, do not edit manually-->
3440<texttable align="left" suppress-title="true" anchor="iana.header.registration.table">
3441   <ttcol>Header Field Name</ttcol>
3442   <ttcol>Protocol</ttcol>
3443   <ttcol>Status</ttcol>
3444   <ttcol>Reference</ttcol>
3446   <c>Connection</c>
3447   <c>http</c>
3448   <c>standard</c>
3449   <c>
3450      <xref target="header.connection"/>
3451   </c>
3452   <c>Content-Length</c>
3453   <c>http</c>
3454   <c>standard</c>
3455   <c>
3456      <xref target="header.content-length"/>
3457   </c>
3458   <c>Host</c>
3459   <c>http</c>
3460   <c>standard</c>
3461   <c>
3462      <xref target=""/>
3463   </c>
3464   <c>TE</c>
3465   <c>http</c>
3466   <c>standard</c>
3467   <c>
3468      <xref target="header.te"/>
3469   </c>
3470   <c>Trailer</c>
3471   <c>http</c>
3472   <c>standard</c>
3473   <c>
3474      <xref target="header.trailer"/>
3475   </c>
3476   <c>Transfer-Encoding</c>
3477   <c>http</c>
3478   <c>standard</c>
3479   <c>
3480      <xref target="header.transfer-encoding"/>
3481   </c>
3482   <c>Upgrade</c>
3483   <c>http</c>
3484   <c>standard</c>
3485   <c>
3486      <xref target="header.upgrade"/>
3487   </c>
3488   <c>Via</c>
3489   <c>http</c>
3490   <c>standard</c>
3491   <c>
3492      <xref target="header.via"/>
3493   </c>
3496<?ENDINC p1-messaging.iana-headers ?>
3498   Furthermore, the header field-name "Close" shall be registered as
3499   "reserved", since using that name as an HTTP header field might
3500   conflict with the "close" connection option of the "<x:ref>Connection</x:ref>"
3501   header field (<xref target="header.connection"/>).
3503<texttable align="left" suppress-title="true">
3504   <ttcol>Header Field Name</ttcol>
3505   <ttcol>Protocol</ttcol>
3506   <ttcol>Status</ttcol>
3507   <ttcol>Reference</ttcol>
3509   <c>Close</c>
3510   <c>http</c>
3511   <c>reserved</c>
3512   <c>
3513      <xref target="header.field.registration"/>
3514   </c>
3517   The change controller is: "IETF ( - Internet Engineering Task Force".
3521<section title="URI Scheme Registration" anchor="uri.scheme.registration">
3523   IANA maintains the registry of URI Schemes <xref target="BCP115"/> at
3524   <eref target=""/>.
3527   This document defines the following URI schemes, so their
3528   associated registry entries shall be updated according to the permanent
3529   registrations below:
3531<texttable align="left" suppress-title="true">
3532   <ttcol>URI Scheme</ttcol>
3533   <ttcol>Description</ttcol>
3534   <ttcol>Reference</ttcol>
3536   <c>http</c>
3537   <c>Hypertext Transfer Protocol</c>
3538   <c><xref target="http.uri"/></c>
3540   <c>https</c>
3541   <c>Hypertext Transfer Protocol Secure</c>
3542   <c><xref target="https.uri"/></c>
3546<section title="Internet Media Type Registration" anchor="">
3548   IANA maintains the registry of Internet media types <xref target="BCP13"/> at
3549   <eref target=""/>.
3552   This document serves as the specification for the Internet media types
3553   "message/http" and "application/http". The following is to be registered with
3554   IANA.
3556<section title="Internet Media Type message/http" anchor="">
3557<iref item="Media Type" subitem="message/http" primary="true"/>
3558<iref item="message/http Media Type" primary="true"/>
3560   The message/http type can be used to enclose a single HTTP request or
3561   response message, provided that it obeys the MIME restrictions for all
3562   "message" types regarding line length and encodings.
3565  <list style="hanging" x:indent="12em">
3566    <t hangText="Type name:">
3567      message
3568    </t>
3569    <t hangText="Subtype name:">
3570      http
3571    </t>
3572    <t hangText="Required parameters:">
3573      N/A
3574    </t>
3575    <t hangText="Optional parameters:">
3576      version, msgtype
3577      <list style="hanging">
3578        <t hangText="version:">
3579          The HTTP-version number of the enclosed message
3580          (e.g., "1.1"). If not present, the version can be
3581          determined from the first line of the body.
3582        </t>
3583        <t hangText="msgtype:">
3584          The message type &mdash; "request" or "response". If not
3585          present, the type can be determined from the first
3586          line of the body.
3587        </t>
3588      </list>
3589    </t>
3590    <t hangText="Encoding considerations:">
3591      only "7bit", "8bit", or "binary" are permitted
3592    </t>
3593    <t hangText="Security considerations:">
3594      see <xref target="security.considerations"/>
3595    </t>
3596    <t hangText="Interoperability considerations:">
3597      N/A
3598    </t>
3599    <t hangText="Published specification:">
3600      This specification (see <xref target=""/>).
3601    </t>
3602    <t hangText="Applications that use this media type:">
3603      N/A
3604    </t>
3605    <t hangText="Fragment identifier considerations:">
3606      N/A
3607    </t>
3608    <t hangText="Additional information:">
3609      <list style="hanging">
3610        <t hangText="Magic number(s):">N/A</t>
3611        <t hangText="Deprecated alias names for this type:">N/A</t>
3612        <t hangText="File extension(s):">N/A</t>
3613        <t hangText="Macintosh file type code(s):">N/A</t>
3614      </list>
3615    </t>
3616    <t hangText="Person and email address to contact for further information:">
3617      See Authors' Addresses Section.
3618    </t>
3619    <t hangText="Intended usage:">
3620      COMMON
3621    </t>
3622    <t hangText="Restrictions on usage:">
3623      N/A
3624    </t>
3625    <t hangText="Author:">
3626      See Authors' Addresses Section.
3627    </t>
3628    <t hangText="Change controller:">
3629      IESG
3630    </t>
3631  </list>
3634<section title="Internet Media Type application/http" anchor="">
3635<iref item="Media Type" subitem="application/http" primary="true"/>
3636<iref item="application/http Media Type" primary="true"/>
3638   The application/http type can be used to enclose a pipeline of one or more
3639   HTTP request or response messages (not intermixed).
3642  <list style="hanging" x:indent="12em">
3643    <t hangText="Type name:">
3644      application
3645    </t>
3646    <t hangText="Subtype name:">
3647      http
3648    </t>
3649    <t hangText="Required parameters:">
3650      N/A
3651    </t>
3652    <t hangText="Optional parameters:">
3653      version, msgtype
3654      <list style="hanging">
3655        <t hangText="version:">
3656          The HTTP-version number of the enclosed messages
3657          (e.g., "1.1"). If not present, the version can be
3658          determined from the first line of the body.
3659        </t>
3660        <t hangText="msgtype:">
3661          The message type &mdash; "request" or "response". If not
3662          present, the type can be determined from the first
3663          line of the body.
3664        </t>
3665      </list>
3666    </t>
3667    <t hangText="Encoding considerations:">
3668      HTTP messages enclosed by this type
3669      are in "binary" format; use of an appropriate
3670      Content-Transfer-Encoding is required when
3671      transmitted via E-mail.
3672    </t>
3673    <t hangText="Security considerations:">
3674      see <xref target="security.considerations"/>
3675    </t>
3676    <t hangText="Interoperability considerations:">
3677      N/A
3678    </t>
3679    <t hangText="Published specification:">
3680      This specification (see <xref target=""/>).
3681    </t>
3682    <t hangText="Applications that use this media type:">
3683      N/A
3684    </t>
3685    <t hangText="Fragment identifier considerations:">
3686      N/A
3687    </t>
3688    <t hangText="Additional information:">
3689      <list style="hanging">
3690        <t hangText="Deprecated alias names for this type:">N/A</t>
3691        <t hangText="Magic number(s):">N/A</t>
3692        <t hangText="File extension(s):">N/A</t>
3693        <t hangText="Macintosh file type code(s):">N/A</t>
3694      </list>
3695    </t>
3696    <t hangText="Person and email address to contact for further information:">
3697      See Authors' Addresses Section.
3698    </t>
3699    <t hangText="Intended usage:">
3700      COMMON
3701    </t>
3702    <t hangText="Restrictions on usage:">
3703      N/A
3704    </t>
3705    <t hangText="Author:">
3706      See Authors' Addresses Section.
3707    </t>
3708    <t hangText="Change controller:">
3709      IESG
3710    </t>
3711  </list>
3716<section title="Transfer Coding Registry" anchor="transfer.coding.registry">
3718   The HTTP Transfer Coding Registry defines the namespace for transfer
3719   coding names. It is maintained at <eref target=""/>.
3722<section title="Procedure" anchor="transfer.coding.registry.procedure">
3724   Registrations &MUST; include the following fields:
3725   <list style="symbols">
3726     <t>Name</t>
3727     <t>Description</t>
3728     <t>Pointer to specification text</t>
3729   </list>
3732   Names of transfer codings &MUST-NOT; overlap with names of content codings
3733   (&content-codings;) unless the encoding transformation is identical, as
3734   is the case for the compression codings defined in
3735   <xref target="compression.codings"/>.
3738   Values to be added to this namespace require IETF Review (see
3739   <xref target="RFC5226" x:fmt="of" x:sec="4.1"/>), and &MUST;
3740   conform to the purpose of transfer coding defined in this specification.
3743   Use of program names for the identification of encoding formats
3744   is not desirable and is discouraged for future encodings.
3748<section title="Registration" anchor="transfer.coding.registration">
3750   The HTTP Transfer Coding Registry shall be updated with the registrations
3751   below:
3753<texttable align="left" suppress-title="true" anchor="iana.transfer.coding.registration.table">
3754   <ttcol>Name</ttcol>
3755   <ttcol>Description</ttcol>
3756   <ttcol>Reference</ttcol>
3757   <c>chunked</c>
3758   <c>Transfer in a series of chunks</c>
3759   <c>
3760      <xref target="chunked.encoding"/>
3761   </c>
3762   <c>compress</c>
3763   <c>UNIX "compress" data format <xref target="Welch"/></c>
3764   <c>
3765      <xref target="compress.coding"/>
3766   </c>
3767   <c>deflate</c>
3768   <c>"deflate" compressed data (<xref target="RFC1951"/>) inside
3769   the "zlib" data format (<xref target="RFC1950"/>)
3770   </c>
3771   <c>
3772      <xref target="deflate.coding"/>
3773   </c>
3774   <c>gzip</c>
3775   <c>GZIP file format <xref target="RFC1952"/></c>
3776   <c>
3777      <xref target="gzip.coding"/>
3778   </c>
3779   <c>x-compress</c>
3780   <c>Deprecated (alias for compress)</c>
3781   <c>
3782      <xref target="compress.coding"/>
3783   </c>
3784   <c>x-gzip</c>
3785   <c>Deprecated (alias for gzip)</c>
3786   <c>
3787      <xref target="gzip.coding"/>
3788   </c>
3793<section title="Content Coding Registration" anchor="content.coding.registration">
3795   IANA maintains the registry of HTTP Content Codings at
3796   <eref target=""/>.
3799   The HTTP Content Codings Registry shall be updated with the registrations
3800   below:
3802<texttable align="left" suppress-title="true" anchor="iana.content.coding.registration.table">
3803   <ttcol>Name</ttcol>
3804   <ttcol>Description</ttcol>
3805   <ttcol>Reference</ttcol>
3806   <c>compress</c>
3807   <c>UNIX "compress" data format <xref target="Welch"/></c>
3808   <c>
3809      <xref target="compress.coding"/>
3810   </c>
3811   <c>deflate</c>
3812   <c>"deflate" compressed data (<xref target="RFC1951"/>) inside
3813   the "zlib" data format (<xref target="RFC1950"/>)</c>
3814   <c>
3815      <xref target="deflate.coding"/>
3816   </c>
3817   <c>gzip</c>
3818   <c>GZIP file format <xref target="RFC1952"/></c>
3819   <c>
3820      <xref target="gzip.coding"/>
3821   </c>
3822   <c>x-compress</c>
3823   <c>Deprecated (alias for compress)</c>
3824   <c>
3825      <xref target="compress.coding"/>
3826   </c>
3827   <c>x-gzip</c>
3828   <c>Deprecated (alias for gzip)</c>
3829   <c>
3830      <xref target="gzip.coding"/>
3831   </c>
3835<section title="Upgrade Token Registry" anchor="upgrade.token.registry">
3837   The HTTP Upgrade Token Registry defines the namespace for protocol-name
3838   tokens used to identify protocols in the <x:ref>Upgrade</x:ref> header
3839   field. The registry is maintained at <eref target=""/>.
3842<section title="Procedure" anchor="upgrade.token.registry.procedure">  
3844   Each registered protocol name is associated with contact information
3845   and an optional set of specifications that details how the connection
3846   will be processed after it has been upgraded.
3849   Registrations happen on a "First Come First Served" basis (see
3850   <xref target="RFC5226" x:sec="4.1" x:fmt="of"/>) and are subject to the
3851   following rules:
3852  <list style="numbers">
3853    <t>A protocol-name token, once registered, stays registered forever.</t>
3854    <t>The registration &MUST; name a responsible party for the
3855       registration.</t>
3856    <t>The registration &MUST; name a point of contact.</t>
3857    <t>The registration &MAY; name a set of specifications associated with
3858       that token. Such specifications need not be publicly available.</t>
3859    <t>The registration &SHOULD; name a set of expected "protocol-version"
3860       tokens associated with that token at the time of registration.</t>
3861    <t>The responsible party &MAY; change the registration at any time.
3862       The IANA will keep a record of all such changes, and make them
3863       available upon request.</t>
3864    <t>The IESG &MAY; reassign responsibility for a protocol token.
3865       This will normally only be used in the case when a
3866       responsible party cannot be contacted.</t>
3867  </list>
3870   This registration procedure for HTTP Upgrade Tokens replaces that
3871   previously defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
3875<section title="Upgrade Token Registration" anchor="upgrade.token.registration">
3877   The "HTTP" entry in the HTTP Upgrade Token Registry shall be updated with
3878   the registration below:
3880<texttable align="left" suppress-title="true">
3881   <ttcol>Value</ttcol>
3882   <ttcol>Description</ttcol>
3883   <ttcol>Expected Version Tokens</ttcol>
3884   <ttcol>Reference</ttcol>
3886   <c>HTTP</c>
3887   <c>Hypertext Transfer Protocol</c>
3888   <c>any DIGIT.DIGIT (e.g, "2.0")</c>
3889   <c><xref target="http.version"/></c>
3892   The responsible party is: "IETF ( - Internet Engineering Task Force".
3899<section title="Security Considerations" anchor="security.considerations">
3901   This section is meant to inform developers, information providers, and
3902   users of known security considerations relevant to HTTP message syntax,
3903   parsing, and routing. Security considerations about HTTP semantics and
3904   payloads are addressed in &semantics;.
3907<section title="Establishing Authority" anchor="establishing.authority">
3908  <iref item="authoritative response" primary="true"/>
3909  <iref item="phishing" primary="true"/>
3911   HTTP relies on the notion of an <x:dfn>authoritative response</x:dfn>: a
3912   response that has been determined by (or at the direction of) the authority
3913   identified within the target URI to be the most appropriate response for
3914   that request given the state of the target resource at the time of
3915   response message origination. Providing a response from a non-authoritative
3916   source, such as a shared cache, is often useful to improve performance and
3917   availability, but only to the extent that the source can be trusted or
3918   the distrusted response can be safely used.
3921   Unfortunately, establishing authority can be difficult.
3922   For example, <x:dfn>phishing</x:dfn> is an attack on the user's perception
3923   of authority, where that perception can be misled by presenting similar
3924   branding in hypertext, possibly aided by userinfo obfuscating the authority
3925   component (see <xref target="http.uri"/>).
3926   User agents can reduce the impact of phishing attacks by enabling users to
3927   easily inspect a target URI prior to making an action, by prominently
3928   distinguishing (or rejecting) userinfo when present, and by not sending
3929   stored credentials and cookies when the referring document is from an
3930   unknown or untrusted source.
3933   When a registered name is used in the authority component, the "http" URI
3934   scheme (<xref target="http.uri"/>) relies on the user's local name
3935   resolution service to determine where it can find authoritative responses.
3936   This means that any attack on a user's network host table, cached names, or
3937   name resolution libraries becomes an avenue for attack on establishing
3938   authority. Likewise, the user's choice of server for Domain Name Service
3939   (DNS), and the hierarchy of servers from which it obtains resolution
3940   results, could impact the authenticity of address mappings;
3941   DNS Security Extensions (DNSSEC, <xref target="RFC4033"/>) are one way to
3942   improve authenticity.
3945   Furthermore, after an IP address is obtained, establishing authority for
3946   an "http" URI is vulnerable to attacks on Internet Protocol routing.
3949   The "https" scheme (<xref target="https.uri"/>) is intended to prevent
3950   (or at least reveal) many of these potential attacks on establishing
3951   authority, provided that the negotiated TLS connection is secured and
3952   the client properly verifies that the communicating server's identity
3953   matches the target URI's authority component
3954   (see <xref target="RFC2818"/>). Correctly implementing such verification
3955   can be difficult (see <xref target="Georgiev"/>).
3959<section title="Risks of Intermediaries" anchor="risks.intermediaries">
3961   By their very nature, HTTP intermediaries are men-in-the-middle, and thus
3962   represent an opportunity for man-in-the-middle attacks. Compromise of
3963   the systems on which the intermediaries run can result in serious security
3964   and privacy problems. Intermediaries might have access to security-related
3965   information, personal information about individual users and
3966   organizations, and proprietary information belonging to users and
3967   content providers. A compromised intermediary, or an intermediary
3968   implemented or configured without regard to security and privacy
3969   considerations, might be used in the commission of a wide range of
3970   potential attacks.
3973   Intermediaries that contain a shared cache are especially vulnerable
3974   to cache poisoning attacks, as described in &cache-poisoning;.
3977   Implementers need to consider the privacy and security
3978   implications of their design and coding decisions, and of the
3979   configuration options they provide to operators (especially the
3980   default configuration).
3983   Users need to be aware that intermediaries are no more trustworthy than
3984   the people who run them; HTTP itself cannot solve this problem.
3988<section title="Attacks via Protocol Element Length" anchor="attack.protocol.element.length">
3990   Because HTTP uses mostly textual, character-delimited fields, parsers are
3991   often vulnerable to attacks based on sending very long (or very slow)
3992   streams of data, particularly where an implementation is expecting a
3993   protocol element with no predefined length.
3996   To promote interoperability, specific recommendations are made for minimum
3997   size limits on request-line (<xref target="request.line"/>)
3998   and header fields (<xref target="header.fields"/>). These are
3999   minimum recommendations, chosen to be supportable even by implementations
4000   with limited resources; it is expected that most implementations will
4001   choose substantially higher limits.
4004   A server can reject a message that
4005   has a request-target that is too long (&status-414;) or a request payload
4006   that is too large (&status-413;). Additional status codes related to
4007   capacity limits have been defined by extensions to HTTP
4008   <xref target="RFC6585"/>.
4011   Recipients ought to carefully limit the extent to which they process other
4012   protocol elements, including (but not limited to) request methods, response
4013   status phrases, header field-names, numeric values, and body chunks.
4014   Failure to limit such processing can result in buffer overflows, arithmetic
4015   overflows, or increased vulnerability to denial-of-service attacks.
4019<section title="Response Splitting" anchor="response.splitting">
4021   Response splitting (a.k.a, CRLF injection) is a common technique, used in
4022   various attacks on Web usage, that exploits the line-based nature of HTTP
4023   message framing and the ordered association of requests to responses on
4024   persistent connections <xref target="Klein"/>. This technique can be
4025   particularly damaging when the requests pass through a shared cache.
4028   Response splitting exploits a vulnerability in servers (usually within an
4029   application server) where an attacker can send encoded data within some
4030   parameter of the request that is later decoded and echoed within any of the
4031   response header fields of the response. If the decoded data is crafted to
4032   look like the response has ended and a subsequent response has begun, the
4033   response has been split and the content within the apparent second response
4034   is controlled by the attacker. The attacker can then make any other request
4035   on the same persistent connection and trick the recipients (including
4036   intermediaries) into believing that the second half of the split is an
4037   authoritative answer to the second request.
4040   For example, a parameter within the request-target might be read by an
4041   application server and reused within a redirect, resulting in the same
4042   parameter being echoed in the <x:ref>Location</x:ref> header field of the
4043   response. If the parameter is decoded by the application and not properly
4044   encoded when placed in the response field, the attacker can send encoded
4045   CRLF octets and other content that will make the application's single
4046   response look like two or more responses.
4049   A common defense against response splitting is to filter requests for data
4050   that looks like encoded CR and LF (e.g., "%0D" and "%0A"). However, that
4051   assumes the application server is only performing URI decoding, rather
4052   than more obscure data transformations like charset transcoding, XML entity
4053   translation, base64 decoding, sprintf reformatting, etc.  A more effective
4054   mitigation is to prevent anything other than the server's core protocol
4055   libraries from sending a CR or LF within the header section, which means
4056   restricting the output of header fields to APIs that filter for bad octets
4057   and not allowing application servers to write directly to the protocol
4058   stream.
4062<section title="Request Smuggling" anchor="request.smuggling">
4064   Request smuggling (<xref target="Linhart"/>) is a technique that exploits
4065   differences in protocol parsing among various recipients to hide additional
4066   requests (which might otherwise be blocked or disabled by policy) within an
4067   apparently harmless request.  Like response splitting, request smuggling
4068   can lead to a variety of attacks on HTTP usage.
4071   This specification has introduced new requirements on request parsing,
4072   particularly with regard to message framing in
4073   <xref target="message.body.length"/>, to reduce the effectiveness of
4074   request smuggling.
4078<section title="Message Integrity" anchor="message.integrity">
4080   HTTP does not define a specific mechanism for ensuring message integrity,
4081   instead relying on the error-detection ability of underlying transport
4082   protocols and the use of length or chunk-delimited framing to detect
4083   completeness. Additional integrity mechanisms, such as hash functions or
4084   digital signatures applied to the content, can be selectively added to
4085   messages via extensible metadata header fields. Historically, the lack of
4086   a single integrity mechanism has been justified by the informal nature of
4087   most HTTP communication.  However, the prevalence of HTTP as an information
4088   access mechanism has resulted in its increasing use within environments
4089   where verification of message integrity is crucial.
4092   User agents are encouraged to implement configurable means for detecting
4093   and reporting failures of message integrity such that those means can be
4094   enabled within environments for which integrity is necessary. For example,
4095   a browser being used to view medical history or drug interaction
4096   information needs to indicate to the user when such information is detected
4097   by the protocol to be incomplete, expired, or corrupted during transfer.
4098   Such mechanisms might be selectively enabled via user agent extensions or
4099   the presence of message integrity metadata in a response.
4100   At a minimum, user agents ought to provide some indication that allows a
4101   user to distinguish between a complete and incomplete response message
4102   (<xref target="incomplete.messages"/>) when such verification is desired.
4106<section title="Message Confidentiality" anchor="message.confidentiality">
4108   HTTP relies on underlying transport protocols to provide message
4109   confidentiality when that is desired. HTTP has been specifically designed
4110   to be independent of the transport protocol, such that it can be used
4111   over many different forms of encrypted connection, with the selection of
4112   such transports being identified by the choice of URI scheme or within
4113   user agent configuration.
4116   The "https" scheme can be used to identify resources that require a
4117   confidential connection, as described in <xref target="https.uri"/>.
4121<section title="Privacy of Server Log Information" anchor="privacy.of.server.log.information">
4123   A server is in the position to save personal data about a user's requests
4124   over time, which might identify their reading patterns or subjects of
4125   interest.  In particular, log information gathered at an intermediary
4126   often contains a history of user agent interaction, across a multitude
4127   of sites, that can be traced to individual users.
4130   HTTP log information is confidential in nature; its handling is often
4131   constrained by laws and regulations.  Log information needs to be securely
4132   stored and appropriate guidelines followed for its analysis.
4133   Anonymization of personal information within individual entries helps,
4134   but it is generally not sufficient to prevent real log traces from being
4135   re-identified based on correlation with other access characteristics.
4136   As such, access traces that are keyed to a specific client are unsafe to
4137   publish even if the key is pseudonymous.
4140   To minimize the risk of theft or accidental publication, log information
4141   ought to be purged of personally identifiable information, including
4142   user identifiers, IP addresses, and user-provided query parameters,
4143   as soon as that information is no longer necessary to support operational
4144   needs for security, auditing, or fraud control.
4149<section title="Acknowledgments" anchor="acks">
4151   This edition of HTTP/1.1 builds on the many contributions that went into
4152   <xref target="RFC1945" format="none">RFC 1945</xref>,
4153   <xref target="RFC2068" format="none">RFC 2068</xref>,
4154   <xref target="RFC2145" format="none">RFC 2145</xref>, and
4155   <xref target="RFC2616" format="none">RFC 2616</xref>, including
4156   substantial contributions made by the previous authors, editors, and
4157   Working Group Chairs: Tim Berners-Lee, Ari Luotonen, Roy T. Fielding,
4158   Henrik Frystyk Nielsen, Jim Gettys, Jeffrey C. Mogul, Larry Masinter,
4159   and Paul J. Leach. Mark Nottingham oversaw this effort as Working Group Chair.
4162   Since 1999, the following contributors have helped improve the HTTP
4163   specification by reporting bugs, asking smart questions, drafting or
4164   reviewing text, and evaluating open issues:
4166<?BEGININC acks ?>
4167<t>Adam Barth,
4168Adam Roach,
4169Addison Phillips,
4170Adrian Chadd,
4171Adrian Cole,
4172Adrien W. de Croy,
4173Alan Ford,
4174Alan Ruttenberg,
4175Albert Lunde,
4176Alek Storm,
4177Alex Rousskov,
4178Alexandre Morgaut,
4179Alexey Melnikov,
4180Alisha Smith,
4181Amichai Rothman,
4182Amit Klein,
4183Amos Jeffries,
4184Andreas Maier,
4185Andreas Petersson,
4186Andrei Popov,
4187Anil Sharma,
4188Anne van Kesteren,
4189Anthony Bryan,
4190Asbjorn Ulsberg,
4191Ashok Kumar,
4192Balachander Krishnamurthy,
4193Barry Leiba,
4194Ben Laurie,
4195Benjamin Carlyle,
4196Benjamin Niven-Jenkins,
4197Benoit Claise,
4198Bil Corry,
4199Bill Burke,
4200Bjoern Hoehrmann,
4201Bob Scheifler,
4202Boris Zbarsky,
4203Brett Slatkin,
4204Brian Kell,
4205Brian McBarron,
4206Brian Pane,
4207Brian Raymor,
4208Brian Smith,
4209Bruce Perens,
4210Bryce Nesbitt,
4211Cameron Heavon-Jones,
4212Carl Kugler,
4213Carsten Bormann,
4214Charles Fry,
4215Chris Burdess,
4216Chris Newman,
4217Christian Huitema,
4218Cyrus Daboo,
4219Dale Robert Anderson,
4220Dan Wing,
4221Dan Winship,
4222Daniel Stenberg,
4223Darrel Miller,
4224Dave Cridland,
4225Dave Crocker,
4226Dave Kristol,
4227Dave Thaler,
4228David Booth,
4229David Singer,
4230David W. Morris,
4231Diwakar Shetty,
4232Dmitry Kurochkin,
4233Drummond Reed,
4234Duane Wessels,
4235Edward Lee,
4236Eitan Adler,
4237Eliot Lear,
4238Emile Stephan,
4239Eran Hammer-Lahav,
4240Eric D. Williams,
4241Eric J. Bowman,
4242Eric Lawrence,
4243Eric Rescorla,
4244Erik Aronesty,
4245EungJun Yi,
4246Evan Prodromou,
4247Felix Geisendoerfer,
4248Florian Weimer,
4249Frank Ellermann,
4250Fred Akalin,
4251Fred Bohle,
4252Frederic Kayser,
4253Gabor Molnar,
4254Gabriel Montenegro,
4255Geoffrey Sneddon,
4256Gervase Markham,
4257Gili Tzabari,
4258Grahame Grieve,
4259Greg Slepak,
4260Greg Wilkins,
4261Grzegorz Calkowski,
4262Harald Tveit Alvestrand,
4263Harry Halpin,
4264Helge Hess,
4265Henrik Nordstrom,
4266Henry S. Thompson,
4267Henry Story,
4268Herbert van de Sompel,
4269Herve Ruellan,
4270Howard Melman,
4271Hugo Haas,
4272Ian Fette,
4273Ian Hickson,
4274Ido Safruti,
4275Ilari Liusvaara,
4276Ilya Grigorik,
4277Ingo Struck,
4278J. Ross Nicoll,
4279James Cloos,
4280James H. Manger,
4281James Lacey,
4282James M. Snell,
4283Jamie Lokier,
4284Jan Algermissen,
4285Jari Arkko,
4286Jeff Hodges (who came up with the term 'effective Request-URI'),
4287Jeff Pinner,
4288Jeff Walden,
4289Jim Luther,
4290Jitu Padhye,
4291Joe D. Williams,
4292Joe Gregorio,
4293Joe Orton,
4294Joel Jaeggli,
4295John C. Klensin,
4296John C. Mallery,
4297John Cowan,
4298John Kemp,
4299John Panzer,
4300John Schneider,
4301John Stracke,
4302John Sullivan,
4303Jonas Sicking,
4304Jonathan A. Rees,
4305Jonathan Billington,
4306Jonathan Moore,
4307Jonathan Silvera,
4308Jordi Ros,
4309Joris Dobbelsteen,
4310Josh Cohen,
4311Julien Pierre,
4312Jungshik Shin,
4313Justin Chapweske,
4314Justin Erenkrantz,
4315Justin James,
4316Kalvinder Singh,
4317Karl Dubost,
4318Kathleen Moriarty,
4319Keith Hoffman,
4320Keith Moore,
4321Ken Murchison,
4322Koen Holtman,
4323Konstantin Voronkov,
4324Kris Zyp,
4325Leif Hedstrom,
4326Lionel Morand,
4327Lisa Dusseault,
4328Maciej Stachowiak,
4329Manu Sporny,
4330Marc Schneider,
4331Marc Slemko,
4332Mark Baker,
4333Mark Pauley,
4334Mark Watson,
4335Markus Isomaki,
4336Markus Lanthaler,
4337Martin J. Duerst,
4338Martin Musatov,
4339Martin Nilsson,
4340Martin Thomson,
4341Matt Lynch,
4342Matthew Cox,
4343Matthew Kerwin,
4344Max Clark,
4345Menachem Dodge,
4346Meral Shirazipour,
4347Michael Burrows,
4348Michael Hausenblas,
4349Michael Scharf,
4350Michael Sweet,
4351Michael Tuexen,
4352Michael Welzl,
4353Mike Amundsen,
4354Mike Belshe,
4355Mike Bishop,
4356Mike Kelly,
4357Mike Schinkel,
4358Miles Sabin,
4359Murray S. Kucherawy,
4360Mykyta Yevstifeyev,
4361Nathan Rixham,
4362Nicholas Shanks,
4363Nico Williams,
4364Nicolas Alvarez,
4365Nicolas Mailhot,
4366Noah Slater,
4367Osama Mazahir,
4368Pablo Castro,
4369Pat Hayes,
4370Patrick R. McManus,
4371Paul E. Jones,
4372Paul Hoffman,
4373Paul Marquess,
4374Pete Resnick,
4375Peter Lepeska,
4376Peter Occil,
4377Peter Saint-Andre,
4378Peter Watkins,
4379Phil Archer,
4380Phil Hunt,
4381Philippe Mougin,
4382Phillip Hallam-Baker,
4383Piotr Dobrogost,
4384Poul-Henning Kamp,
4385Preethi Natarajan,
4386Rajeev Bector,
4387Ray Polk,
4388Reto Bachmann-Gmuer,
4389Richard Barnes,
4390Richard Cyganiak,
4391Rob Trace,
4392Robby Simpson,
4393Robert Brewer,
4394Robert Collins,
4395Robert Mattson,
4396Robert O'Callahan,
4397Robert Olofsson,
4398Robert Sayre,
4399Robert Siemer,
4400Robert de Wilde,
4401Roberto Javier Godoy,
4402Roberto Peon,
4403Roland Zink,
4404Ronny Widjaja,
4405Ryan Hamilton,
4406S. Mike Dierken,
4407Salvatore Loreto,
4408Sam Johnston,
4409Sam Pullara,
4410Sam Ruby,
4411Saurabh Kulkarni,
4412Scott Lawrence (who maintained the original issues list),
4413Sean B. Palmer,
4414Sean Turner,
4415Sebastien Barnoud,
4416Shane McCarron,
4417Shigeki Ohtsu,
4418Simon Yarde,
4419Stefan Eissing,
4420Stefan Tilkov,
4421Stefanos Harhalakis,
4422Stephane Bortzmeyer,
4423Stephen Farrell,
4424Stephen Kent,
4425Stephen Ludin,
4426Stuart Williams,
4427Subbu Allamaraju,
4428Subramanian Moonesamy,
4429Susan Hares,
4430Sylvain Hellegouarch,
4431Tapan Divekar,
4432Tatsuhiro Tsujikawa,
4433Tatsuya Hayashi,
4434Ted Hardie,
4435Ted Lemon,
4436Thomas Broyer,
4437Thomas Fossati,
4438Thomas Maslen,
4439Thomas Nadeau,
4440Thomas Nordin,
4441Thomas Roessler,
4442Tim Bray,
4443Tim Morgan,
4444Tim Olsen,
4445Tom Zhou,
4446Travis Snoozy,
4447Tyler Close,
4448Vincent Murphy,
4449Wenbo Zhu,
4450Werner Baumann,
4451Wilbur Streett,
4452Wilfredo Sanchez Vega,
4453William A. Rowe Jr.,
4454William Chan,
4455Willy Tarreau,
4456Xiaoshu Wang,
4457Yaron Goland,
4458Yngve Nysaeter Pettersen,
4459Yoav Nir,
4460Yogesh Bang,
4461Yuchung Cheng,
4462Yutaka Oiwa,
4463Yves Lafon (long-time member of the editor team),
4464Zed A. Shaw, and
4465Zhong Yu.
4467<?ENDINC acks ?>
4469   See <xref target="RFC2616" x:fmt="of" x:sec="16"/> for additional
4470   acknowledgements from prior revisions.
4477<references title="Normative References">
4479<reference anchor="RFC7231">
4480  <front>
4481    <title>Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content</title>
4482    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4483      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4484      <address><email></email></address>
4485    </author>
4486    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4487      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4488      <address><email></email></address>
4489    </author>
4490    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4491  </front>
4492  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-&ID-VERSION;"/>
4493  <x:source href="p2-semantics.xml" basename="p2-semantics">
4494    <x:defines>1xx (Informational)</x:defines>
4495    <x:defines>1xx</x:defines>
4496    <x:defines>100 (Continue)</x:defines>
4497    <x:defines>101 (Switching Protocols)</x:defines>
4498    <x:defines>2xx (Successful)</x:defines>
4499    <x:defines>2xx</x:defines>
4500    <x:defines>200 (OK)</x:defines>
4501    <x:defines>203 (Non-Authoritative Information)</x:defines>
4502    <x:defines>204 (No Content)</x:defines>
4503    <x:defines>3xx (Redirection)</x:defines>
4504    <x:defines>3xx</x:defines>
4505    <x:defines>301 (Moved Permanently)</x:defines>
4506    <x:defines>4xx (Client Error)</x:defines>
4507    <x:defines>4xx</x:defines>
4508    <x:defines>400 (Bad Request)</x:defines>
4509    <x:defines>411 (Length Required)</x:defines>
4510    <x:defines>414 (URI Too Long)</x:defines>
4511    <x:defines>417 (Expectation Failed)</x:defines>
4512    <x:defines>426 (Upgrade Required)</x:defines>
4513    <x:defines>501 (Not Implemented)</x:defines>
4514    <x:defines>502 (Bad Gateway)</x:defines>
4515    <x:defines>505 (HTTP Version Not Supported)</x:defines>
4516    <x:defines>Accept-Encoding</x:defines>
4517    <x:defines>Allow</x:defines>
4518    <x:defines>Content-Encoding</x:defines>
4519    <x:defines>Content-Location</x:defines>
4520    <x:defines>Content-Type</x:defines>
4521    <x:defines>Date</x:defines>
4522    <x:defines>Expect</x:defines>
4523    <x:defines>Location</x:defines>
4524    <x:defines>Server</x:defines>
4525    <x:defines>User-Agent</x:defines>
4526  </x:source>
4529<reference anchor="RFC7232">
4530  <front>
4531    <title>Hypertext Transfer Protocol (HTTP/1.1): Conditional Requests</title>
4532    <author fullname="Roy T. Fielding" initials="R." role="editor" surname="Fielding">
4533      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4534      <address><email></email></address>
4535    </author>
4536    <author fullname="Julian F. Reschke" initials="J. F." role="editor" surname="Reschke">
4537      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4538      <address><email></email></address>
4539    </author>
4540    <date month="&ID-MONTH;" year="&ID-YEAR;" />
4541  </front>
4542  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p4-conditional-&ID-VERSION;" />
4543  <x:source basename="p4-conditional" href="p4-conditional.xml">
4544    <x:defines>304 (Not Modified)</x:defines>
4545    <x:defines>ETag</x:defines>
4546    <x:defines>Last-Modified</x:defines>
4547  </x:source>
4550<reference anchor="RFC7233">
4551  <front>
4552    <title>Hypertext Transfer Protocol (HTTP/1.1): Range Requests</title>
4553    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4554      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4555      <address><email></email></address>
4556    </author>
4557    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
4558      <organization abbrev="W3C">World Wide Web Consortium</organization>
4559      <address><email></email></address>
4560    </author>
4561    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4562      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4563      <address><email></email></address>
4564    </author>
4565    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4566  </front>
4567  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-&ID-VERSION;"/>
4568  <x:source href="p5-range.xml" basename="p5-range">
4569    <x:defines>Content-Range</x:defines>
4570  </x:source>
4573<reference anchor="RFC7234">
4574  <front>
4575    <title>Hypertext Transfer Protocol (HTTP/1.1): Caching</title>
4576    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4577      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4578      <address><email></email></address>
4579    </author>
4580    <author initials="M." surname="Nottingham" fullname="Mark Nottingham" role="editor">
4581      <organization>Akamai</organization>
4582      <address><email></email></address>
4583    </author>
4584    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4585      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4586      <address><email></email></address>
4587    </author>
4588    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4589  </front>
4590  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-&ID-VERSION;"/>
4591  <x:source href="p6-cache.xml" basename="p6-cache">
4592    <x:defines>Cache-Control</x:defines>
4593    <x:defines>Expires</x:defines>
4594    <x:defines>Warning</x:defines>
4595  </x:source>
4598<reference anchor="RFC7235">
4599  <front>
4600    <title>Hypertext Transfer Protocol (HTTP/1.1): Authentication</title>
4601    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4602      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4603      <address><email></email></address>
4604    </author>
4605    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4606      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4607      <address><email></email></address>
4608    </author>
4609    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4610  </front>
4611  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p7-auth-&ID-VERSION;"/>
4612  <x:source href="p7-auth.xml" basename="p7-auth">
4613    <x:defines>Proxy-Authenticate</x:defines>
4614    <x:defines>Proxy-Authorization</x:defines>
4615  </x:source>
4618<reference anchor="RFC5234">
4619  <front>
4620    <title abbrev="ABNF for Syntax Specifications">Augmented BNF for Syntax Specifications: ABNF</title>
4621    <author initials="D." surname="Crocker" fullname="Dave Crocker" role="editor">
4622      <organization>Brandenburg InternetWorking</organization>
4623      <address>
4624        <email></email>
4625      </address> 
4626    </author>
4627    <author initials="P." surname="Overell" fullname="Paul Overell">
4628      <organization>THUS plc.</organization>
4629      <address>
4630        <email></email>
4631      </address>
4632    </author>
4633    <date month="January" year="2008"/>
4634  </front>
4635  <seriesInfo name="STD" value="68"/>
4636  <seriesInfo name="RFC" value="5234"/>
4639<reference anchor="RFC2119">
4640  <front>
4641    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
4642    <author initials="S." surname="Bradner" fullname="Scott Bradner">
4643      <organization>Harvard University</organization>
4644      <address><email></email></address>
4645    </author>
4646    <date month="March" year="1997"/>
4647  </front>
4648  <seriesInfo name="BCP" value="14"/>
4649  <seriesInfo name="RFC" value="2119"/>
4652<reference anchor="RFC3986">
4653 <front>
4654  <title abbrev='URI Generic Syntax'>Uniform Resource Identifier (URI): Generic Syntax</title>
4655  <author initials='T.' surname='Berners-Lee' fullname='Tim Berners-Lee'>
4656    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
4657    <address>
4658       <email></email>
4659       <uri></uri>
4660    </address>
4661  </author>
4662  <author initials='R.' surname='Fielding' fullname='Roy T. Fielding'>
4663    <organization abbrev="Day Software">Day Software</organization>
4664    <address>
4665      <email></email>
4666      <uri></uri>
4667    </address>
4668  </author>
4669  <author initials='L.' surname='Masinter' fullname='Larry Masinter'>
4670    <organization abbrev="Adobe Systems">Adobe Systems Incorporated</organization>
4671    <address>
4672      <email></email>
4673      <uri></uri>
4674    </address>
4675  </author>
4676  <date month='January' year='2005'></date>
4677 </front>
4678 <seriesInfo name="STD" value="66"/>
4679 <seriesInfo name="RFC" value="3986"/>
4682<reference anchor="RFC0793">
4683  <front>
4684    <title>Transmission Control Protocol</title>
4685    <author initials='J.' surname='Postel' fullname='Jon Postel'>
4686      <organization>University of Southern California (USC)/Information Sciences Institute</organization>
4687    </author>
4688    <date year='1981' month='September' />
4689  </front>
4690  <seriesInfo name='STD' value='7' />
4691  <seriesInfo name='RFC' value='793' />
4694<reference anchor="USASCII">
4695  <front>
4696    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
4697    <author>
4698      <organization>American National Standards Institute</organization>
4699    </author>
4700    <date year="1986"/>
4701  </front>
4702  <seriesInfo name="ANSI" value="X3.4"/>
4705<reference anchor="RFC1950">
4706  <front>
4707    <title>ZLIB Compressed Data Format Specification version 3.3</title>
4708    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4709      <organization>Aladdin Enterprises</organization>
4710      <address><email></email></address>
4711    </author>
4712    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly"/>
4713    <date month="May" year="1996"/>
4714  </front>
4715  <seriesInfo name="RFC" value="1950"/>
4716  <!--<annotation>
4717    RFC 1950 is an Informational RFC, thus it might be less stable than
4718    this specification. On the other hand, this downward reference was
4719    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4720    therefore it is unlikely to cause problems in practice. See also
4721    <xref target="BCP97"/>.
4722  </annotation>-->
4725<reference anchor="RFC1951">
4726  <front>
4727    <title>DEFLATE Compressed Data Format Specification version 1.3</title>
4728    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4729      <organization>Aladdin Enterprises</organization>
4730      <address><email></email></address>
4731    </author>
4732    <date month="May" year="1996"/>
4733  </front>
4734  <seriesInfo name="RFC" value="1951"/>
4735  <!--<annotation>
4736    RFC 1951 is an Informational RFC, thus it might be less stable than
4737    this specification. On the other hand, this downward reference was
4738    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4739    therefore it is unlikely to cause problems in practice. See also
4740    <xref target="BCP97"/>.
4741  </annotation>-->
4744<reference anchor="RFC1952">
4745  <front>
4746    <title>GZIP file format specification version 4.3</title>
4747    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4748      <organization>Aladdin Enterprises</organization>
4749      <address><email></email></address>
4750    </author>
4751    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly">
4752      <address><email></email></address>
4753    </author>
4754    <author initials="M." surname="Adler" fullname="Mark Adler">
4755      <address><email></email></address>
4756    </author>
4757    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4758      <address><email></email></address>
4759    </author>
4760    <author initials="G." surname="Randers-Pehrson" fullname="Glenn Randers-Pehrson">
4761      <address><email></email></address>
4762    </author>
4763    <date month="May" year="1996"/>
4764  </front>
4765  <seriesInfo name="RFC" value="1952"/>
4766  <!--<annotation>
4767    RFC 1952 is an Informational RFC, thus it might be less stable than
4768    this specification. On the other hand, this downward reference was
4769    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4770    therefore it is unlikely to cause problems in practice. See also
4771    <xref target="BCP97"/>.
4772  </annotation>-->
4775<reference anchor="Welch">
4776  <front>
4777    <title>A Technique for High-Performance Data Compression</title>
4778    <author initials="T. A." surname="Welch" fullname="Terry A. Welch"/>
4779    <date month="June" year="1984"/>
4780  </front>
4781  <seriesInfo name="IEEE Computer" value="17(6)"/>
4786<references title="Informative References">
4788<reference anchor="ISO-8859-1">
4789  <front>
4790    <title>
4791     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
4792    </title>
4793    <author>
4794      <organization>International Organization for Standardization</organization>
4795    </author>
4796    <date year="1998"/>
4797  </front>
4798  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
4801<reference anchor='RFC1919'>
4802  <front>
4803    <title>Classical versus Transparent IP Proxies</title>
4804    <author initials='M.' surname='Chatel' fullname='Marc Chatel'>
4805      <address><email></email></address>
4806    </author>
4807    <date year='1996' month='March' />
4808  </front>
4809  <seriesInfo name='RFC' value='1919' />
4812<reference anchor="RFC1945">
4813  <front>
4814    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
4815    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4816      <organization>MIT, Laboratory for Computer Science</organization>
4817      <address><email></email></address>
4818    </author>
4819    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4820      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4821      <address><email></email></address>
4822    </author>
4823    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4824      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
4825      <address><email></email></address>
4826    </author>
4827    <date month="May" year="1996"/>
4828  </front>
4829  <seriesInfo name="RFC" value="1945"/>
4832<reference anchor="RFC2045">
4833  <front>
4834    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
4835    <author initials="N." surname="Freed" fullname="Ned Freed">
4836      <organization>Innosoft International, Inc.</organization>
4837      <address><email></email></address>
4838    </author>
4839    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
4840      <organization>First Virtual Holdings</organization>
4841      <address><email></email></address>
4842    </author>
4843    <date month="November" year="1996"/>
4844  </front>
4845  <seriesInfo name="RFC" value="2045"/>
4848<reference anchor="RFC2047">
4849  <front>
4850    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
4851    <author initials="K." surname="Moore" fullname="Keith Moore">
4852      <organization>University of Tennessee</organization>
4853      <address><email></email></address>
4854    </author>
4855    <date month="November" year="1996"/>
4856  </front>
4857  <seriesInfo name="RFC" value="2047"/>
4860<reference anchor="RFC2068">
4861  <front>
4862    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4863    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
4864      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4865      <address><email></email></address>
4866    </author>
4867    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4868      <organization>MIT Laboratory for Computer Science</organization>
4869      <address><email></email></address>
4870    </author>
4871    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
4872      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
4873      <address><email></email></address>
4874    </author>
4875    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4876      <organization>MIT Laboratory for Computer Science</organization>
4877      <address><email></email></address>
4878    </author>
4879    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4880      <organization>MIT Laboratory for Computer Science</organization>
4881      <address><email></email></address>
4882    </author>
4883    <date month="January" year="1997"/>
4884  </front>
4885  <seriesInfo name="RFC" value="2068"/>
4888<reference anchor="RFC2145">
4889  <front>
4890    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
4891    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
4892      <organization>Western Research Laboratory</organization>
4893      <address><email></email></address>
4894    </author>
4895    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4896      <organization>Department of Information and Computer Science</organization>
4897      <address><email></email></address>
4898    </author>
4899    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4900      <organization>MIT Laboratory for Computer Science</organization>
4901      <address><email></email></address>
4902    </author>
4903    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4904      <organization>W3 Consortium</organization>
4905      <address><email></email></address>
4906    </author>
4907    <date month="May" year="1997"/>
4908  </front>
4909  <seriesInfo name="RFC" value="2145"/>
4912<reference anchor="RFC2616">
4913  <front>
4914    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4915    <author initials="R." surname="Fielding" fullname="R. Fielding">
4916      <organization>University of California, Irvine</organization>
4917      <address><email></email></address>
4918    </author>
4919    <author initials="J." surname="Gettys" fullname="J. Gettys">
4920      <organization>W3C</organization>
4921      <address><email></email></address>
4922    </author>
4923    <author initials="J." surname="Mogul" fullname="J. Mogul">
4924      <organization>Compaq Computer Corporation</organization>
4925      <address><email></email></address>
4926    </author>
4927    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
4928      <organization>MIT Laboratory for Computer Science</organization>
4929      <address><email></email></address>
4930    </author>
4931    <author initials="L." surname="Masinter" fullname="L. Masinter">
4932      <organization>Xerox Corporation</organization>
4933      <address><email></email></address>
4934    </author>
4935    <author initials="P." surname="Leach" fullname="P. Leach">
4936      <organization>Microsoft Corporation</organization>
4937      <address><email></email></address>
4938    </author>
4939    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
4940      <organization>W3C</organization>
4941      <address><email></email></address>
4942    </author>
4943    <date month="June" year="1999"/>
4944  </front>
4945  <seriesInfo name="RFC" value="2616"/>
4948<reference anchor='RFC2817'>
4949  <front>
4950    <title>Upgrading to TLS Within HTTP/1.1</title>
4951    <author initials='R.' surname='Khare' fullname='R. Khare'>
4952      <organization>4K Associates / UC Irvine</organization>
4953      <address><email></email></address>
4954    </author>
4955    <author initials='S.' surname='Lawrence' fullname='S. Lawrence'>
4956      <organization>Agranat Systems, Inc.</organization>
4957      <address><email></email></address>
4958    </author>
4959    <date year='2000' month='May' />
4960  </front>
4961  <seriesInfo name='RFC' value='2817' />
4964<reference anchor='RFC2818'>
4965  <front>
4966    <title>HTTP Over TLS</title>
4967    <author initials='E.' surname='Rescorla' fullname='Eric Rescorla'>
4968      <organization>RTFM, Inc.</organization>
4969      <address><email></email></address>
4970    </author>
4971    <date year='2000' month='May' />
4972  </front>
4973  <seriesInfo name='RFC' value='2818' />
4976<reference anchor='RFC3040'>
4977  <front>
4978    <title>Internet Web Replication and Caching Taxonomy</title>
4979    <author initials='I.' surname='Cooper' fullname='I. Cooper'>
4980      <organization>Equinix, Inc.</organization>
4981    </author>
4982    <author initials='I.' surname='Melve' fullname='I. Melve'>
4983      <organization>UNINETT</organization>
4984    </author>
4985    <author initials='G.' surname='Tomlinson' fullname='G. Tomlinson'>
4986      <organization>CacheFlow Inc.</organization>
4987    </author>
4988    <date year='2001' month='January' />
4989  </front>
4990  <seriesInfo name='RFC' value='3040' />
4993<reference anchor='BCP90'>
4994  <front>
4995    <title>Registration Procedures for Message Header Fields</title>
4996    <author initials='G.' surname='Klyne' fullname='G. Klyne'>
4997      <organization>Nine by Nine</organization>
4998      <address><email></email></address>
4999    </author>
5000    <author initials='M.' surname='Nottingham' fullname='M. Nottingham'>
5001      <organization>BEA Systems</organization>
5002      <address><email></email></address>
5003    </author>
5004    <author initials='J.' surname='Mogul' fullname='J. Mogul'>
5005      <organization>HP Labs</organization>
5006      <address><email></email></address>
5007    </author>
5008    <date year='2004' month='September' />
5009  </front>
5010  <seriesInfo name='BCP' value='90' />
5011  <seriesInfo name='RFC' value='3864' />
5014<reference anchor='RFC4033'>
5015  <front>
5016    <title>DNS Security Introduction and Requirements</title>
5017    <author initials='R.' surname='Arends' fullname='R. Arends'/>
5018    <author initials='R.' surname='Austein' fullname='R. Austein'/>
5019    <author initials='M.' surname='Larson' fullname='M. Larson'/>
5020    <author initials='D.' surname='Massey' fullname='D. Massey'/>
5021    <author initials='S.' surname='Rose' fullname='S. Rose'/>
5022    <date year='2005' month='March' />
5023  </front>
5024  <seriesInfo name='RFC' value='4033' />
5027<reference anchor="BCP13">
5028  <front>
5029    <title>Media Type Specifications and Registration Procedures</title>
5030    <author initials="N." surname="Freed" fullname="Ned Freed">
5031      <organization>Oracle</organization>
5032      <address>
5033        <email></email>
5034      </address>
5035    </author>
5036    <author initials="J." surname="Klensin" fullname="John C. Klensin">
5037      <address>
5038        <email></email>
5039      </address>
5040    </author>
5041    <author initials="T." surname="Hansen" fullname="Tony Hansen">
5042      <organization>AT&amp;T Laboratories</organization>
5043      <address>
5044        <email></email>
5045      </address>
5046    </author>
5047    <date year="2013" month="January"/>
5048  </front>
5049  <seriesInfo name="BCP" value="13"/>
5050  <seriesInfo name="RFC" value="6838"/>
5053<reference anchor='BCP115'>
5054  <front>
5055    <title>Guidelines and Registration Procedures for New URI Schemes</title>
5056    <author initials='T.' surname='Hansen' fullname='T. Hansen'>
5057      <organization>AT&amp;T Laboratories</organization>
5058      <address>
5059        <email></email>
5060      </address>
5061    </author>
5062    <author initials='T.' surname='Hardie' fullname='T. Hardie'>
5063      <organization>Qualcomm, Inc.</organization>
5064      <address>
5065        <email></email>
5066      </address>
5067    </author>
5068    <author initials='L.' surname='Masinter' fullname='L. Masinter'>
5069      <organization>Adobe Systems</organization>
5070      <address>
5071        <email></email>
5072      </address>
5073    </author>
5074    <date year='2006' month='February' />
5075  </front>
5076  <seriesInfo name='BCP' value='115' />
5077  <seriesInfo name='RFC' value='4395' />
5080<reference anchor='RFC4559'>
5081  <front>
5082    <title>SPNEGO-based Kerberos and NTLM HTTP Authentication in Microsoft Windows</title>
5083    <author initials='K.' surname='Jaganathan' fullname='K. Jaganathan'/>
5084    <author initials='L.' surname='Zhu' fullname='L. Zhu'/>
5085    <author initials='J.' surname='Brezak' fullname='J. Brezak'/>
5086    <date year='2006' month='June' />
5087  </front>
5088  <seriesInfo name='RFC' value='4559' />
5091<reference anchor='RFC5226'>
5092  <front>
5093    <title>Guidelines for Writing an IANA Considerations Section in RFCs</title>
5094    <author initials='T.' surname='Narten' fullname='T. Narten'>
5095      <organization>IBM</organization>
5096      <address><email></email></address>
5097    </author>
5098    <author initials='H.' surname='Alvestrand' fullname='H. Alvestrand'>
5099      <organization>Google</organization>
5100      <address><email></email></address>
5101    </author>
5102    <date year='2008' month='May' />
5103  </front>
5104  <seriesInfo name='BCP' value='26' />
5105  <seriesInfo name='RFC' value='5226' />
5108<reference anchor='RFC5246'>
5109   <front>
5110      <title>The Transport Layer Security (TLS) Protocol Version 1.2</title>
5111      <author initials='T.' surname='Dierks' fullname='T. Dierks'/>
5112      <author initials='E.' surname='Rescorla' fullname='E. Rescorla'>
5113         <organization>RTFM, Inc.</organization>
5114      </author>
5115      <date year='2008' month='August' />
5116   </front>
5117   <seriesInfo name='RFC' value='5246' />
5120<reference anchor="RFC5322">
5121  <front>
5122    <title>Internet Message Format</title>
5123    <author initials="P." surname="Resnick" fullname="P. Resnick">
5124      <organization>Qualcomm Incorporated</organization>
5125    </author>
5126    <date year="2008" month="October"/>
5127  </front>
5128  <seriesInfo name="RFC" value="5322"/>
5131<reference anchor="RFC6265">
5132  <front>
5133    <title>HTTP State Management Mechanism</title>
5134    <author initials="A." surname="Barth" fullname="Adam Barth">
5135      <organization abbrev="U.C. Berkeley">
5136        University of California, Berkeley
5137      </organization>
5138      <address><email></email></address>
5139    </author>
5140    <date year="2011" month="April" />
5141  </front>
5142  <seriesInfo name="RFC" value="6265"/>
5145<reference anchor='RFC6585'>
5146  <front>
5147    <title>Additional HTTP Status Codes</title>
5148    <author initials='M.' surname='Nottingham' fullname='M. Nottingham'>
5149      <organization>Rackspace</organization>
5150    </author>
5151    <author initials='R.' surname='Fielding' fullname='R. Fielding'>
5152      <organization>Adobe</organization>
5153    </author>
5154    <date year='2012' month='April' />
5155   </front>
5156   <seriesInfo name='RFC' value='6585' />
5159<!--<reference anchor='BCP97'>
5160  <front>
5161    <title>Handling Normative References to Standards-Track Documents</title>
5162    <author initials='J.' surname='Klensin' fullname='J. Klensin'>
5163      <address>
5164        <email></email>
5165      </address>
5166    </author>
5167    <author initials='S.' surname='Hartman' fullname='S. Hartman'>
5168      <organization>MIT</organization>
5169      <address>
5170        <email></email>
5171      </address>
5172    </author>
5173    <date year='2007' month='June' />
5174  </front>
5175  <seriesInfo name='BCP' value='97' />
5176  <seriesInfo name='RFC' value='4897' />
5179<reference anchor="Kri2001" target="">
5180  <front>
5181    <title>HTTP Cookies: Standards, Privacy, and Politics</title>
5182    <author initials="D." surname="Kristol" fullname="David M. Kristol"/>
5183    <date year="2001" month="November"/>
5184  </front>
5185  <seriesInfo name="ACM Transactions on Internet Technology" value="1(2)"/>
5188<reference anchor="Klein" target="">
5189  <front>
5190    <title>Divide and Conquer - HTTP Response Splitting, Web Cache Poisoning Attacks, and Related Topics</title>
5191    <author initials="A." surname="Klein" fullname="Amit Klein">
5192      <organization>Sanctum, Inc.</organization>
5193    </author>
5194    <date year="2004" month="March"/>
5195  </front>
5198<reference anchor="Georgiev" target="">
5199  <front>
5200    <title>The Most Dangerous Code in the World: Validating SSL Certificates in Non-browser Software</title>
5201    <author initials="M." surname="Georgiev" fullname="Martin Georgiev"/>
5202    <author initials="S." surname="Iyengar" fullname="Subodh Iyengar"/>
5203    <author initials="S." surname="Jana" fullname="Suman Jana"/>
5204    <author initials="R." surname="Anubhai" fullname="Rishita Anubhai"/>
5205    <author initials="D." surname="Boneh" fullname="Dan Boneh"/>
5206    <author initials="V." surname="Shmatikov" fullname="Vitaly Shmatikov"/>
5207    <date year="2012" month="October"/>
5208  </front>
5209  <x:prose>In Proceedings of the 2012 ACM Conference on Computer and Communications Security (CCS '12), pp. 38-49</x:prose>
5212<reference anchor="Linhart" target="">
5213  <front>
5214    <title>HTTP Request Smuggling</title>
5215    <author initials="C." surname="Linhart" fullname="Chaim Linhart"/>
5216    <author initials="A." surname="Klein" fullname="Amit Klein"/>
5217    <author initials="R." surname="Heled" fullname="Ronen Heled"/>
5218    <author initials="S." surname="Orrin" fullname="Steve Orrin"/>
5219    <date year="2005" month="June"/>
5220  </front>
5226<section title="HTTP Version History" anchor="compatibility">
5228   HTTP has been in use since 1990. The first version, later referred to as
5229   HTTP/0.9, was a simple protocol for hypertext data transfer across the
5230   Internet, using only a single request method (GET) and no metadata.
5231   HTTP/1.0, as defined by <xref target="RFC1945"/>, added a range of request
5232   methods and MIME-like messaging, allowing for metadata to be transferred
5233   and modifiers placed on the request/response semantics. However,
5234   HTTP/1.0 did not sufficiently take into consideration the effects of
5235   hierarchical proxies, caching, the need for persistent connections, or
5236   name-based virtual hosts. The proliferation of incompletely-implemented
5237   applications calling themselves "HTTP/1.0" further necessitated a
5238   protocol version change in order for two communicating applications
5239   to determine each other's true capabilities.
5242   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
5243   requirements that enable reliable implementations, adding only
5244   those features that can either be safely ignored by an HTTP/1.0
5245   recipient or only be sent when communicating with a party advertising
5246   conformance with HTTP/1.1.
5249   HTTP/1.1 has been designed to make supporting previous versions easy.
5250   A general-purpose HTTP/1.1 server ought to be able to understand any valid
5251   request in the format of HTTP/1.0, responding appropriately with an
5252   HTTP/1.1 message that only uses features understood (or safely ignored) by
5253   HTTP/1.0 clients. Likewise, an HTTP/1.1 client can be expected to
5254   understand any valid HTTP/1.0 response.
5257   Since HTTP/0.9 did not support header fields in a request, there is no
5258   mechanism for it to support name-based virtual hosts (selection of resource
5259   by inspection of the <x:ref>Host</x:ref> header field).
5260   Any server that implements name-based virtual hosts ought to disable
5261   support for HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in
5262   fact, badly constructed HTTP/1.x requests caused by a client failing to
5263   properly encode the request-target.
5266<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
5268   This section summarizes major differences between versions HTTP/1.0
5269   and HTTP/1.1.
5272<section title="Multi-homed Web Servers" anchor="">
5274   The requirements that clients and servers support the <x:ref>Host</x:ref>
5275   header field (<xref target=""/>), report an error if it is
5276   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-target"/>)
5277   are among the most important changes defined by HTTP/1.1.
5280   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
5281   addresses and servers; there was no other established mechanism for
5282   distinguishing the intended server of a request than the IP address
5283   to which that request was directed. The <x:ref>Host</x:ref> header field was
5284   introduced during the development of HTTP/1.1 and, though it was
5285   quickly implemented by most HTTP/1.0 browsers, additional requirements
5286   were placed on all HTTP/1.1 requests in order to ensure complete
5287   adoption.  At the time of this writing, most HTTP-based services
5288   are dependent upon the Host header field for targeting requests.
5292<section title="Keep-Alive Connections" anchor="compatibility.with.http.1.0.persistent.connections">
5294   In HTTP/1.0, each connection is established by the client prior to the
5295   request and closed by the server after sending the response. However, some
5296   implementations implement the explicitly negotiated ("Keep-Alive") version
5297   of persistent connections described in <xref x:sec="19.7.1" x:fmt="of"
5298   target="RFC2068"/>.
5301   Some clients and servers might wish to be compatible with these previous
5302   approaches to persistent connections, by explicitly negotiating for them
5303   with a "Connection: keep-alive" request header field. However, some
5304   experimental implementations of HTTP/1.0 persistent connections are faulty;
5305   for example, if an HTTP/1.0 proxy server doesn't understand
5306   <x:ref>Connection</x:ref>, it will erroneously forward that header field
5307   to the next inbound server, which would result in a hung connection.
5310   One attempted solution was the introduction of a Proxy-Connection header
5311   field, targeted specifically at proxies. In practice, this was also
5312   unworkable, because proxies are often deployed in multiple layers, bringing
5313   about the same problem discussed above.
5316   As a result, clients are encouraged not to send the Proxy-Connection header
5317   field in any requests.
5320   Clients are also encouraged to consider the use of Connection: keep-alive
5321   in requests carefully; while they can enable persistent connections with
5322   HTTP/1.0 servers, clients using them will need to monitor the
5323   connection for "hung" requests (which indicate that the client ought stop
5324   sending the header field), and this mechanism ought not be used by clients
5325   at all when a proxy is being used.
5329<section title="Introduction of Transfer-Encoding" anchor="introduction.of.transfer-encoding">
5331   HTTP/1.1 introduces the <x:ref>Transfer-Encoding</x:ref> header field
5332   (<xref target="header.transfer-encoding"/>).
5333   Transfer codings need to be decoded prior to forwarding an HTTP message
5334   over a MIME-compliant protocol.
5340<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
5342  HTTP's approach to error handling has been explained.
5343  (<xref target="conformance" />)
5346  The HTTP-version ABNF production has been clarified to be case-sensitive.
5347  Additionally, version numbers have been restricted to single digits, due
5348  to the fact that implementations are known to handle multi-digit version
5349  numbers incorrectly.
5350  (<xref target="http.version"/>)
5353  Userinfo (i.e., username and password) are now disallowed in HTTP and
5354  HTTPS URIs, because of security issues related to their transmission on the
5355  wire.
5356  (<xref target="http.uri" />)
5359  The HTTPS URI scheme is now defined by this specification; previously,
5360  it was done in  <xref target="RFC2818" x:fmt="of" x:sec="2.4"/>.
5361  Furthermore, it implies end-to-end security.
5362  (<xref target="https.uri"/>)
5365  HTTP messages can be (and often are) buffered by implementations; despite
5366  it sometimes being available as a stream, HTTP is fundamentally a
5367  message-oriented protocol.
5368  Minimum supported sizes for various protocol elements have been
5369  suggested, to improve interoperability.
5370  (<xref target="http.message" />)
5373  Invalid whitespace around field-names is now required to be rejected,
5374  because accepting it represents a security vulnerability.
5375  The ABNF productions defining header fields now only list the field value.
5376  (<xref target="header.fields"/>)
5379  Rules about implicit linear whitespace between certain grammar productions
5380  have been removed; now whitespace is only allowed where specifically
5381  defined in the ABNF.
5382  (<xref target="whitespace"/>)
5385  Header fields that span multiple lines ("line folding") are deprecated.
5386  (<xref target="field.parsing" />)
5389  The NUL octet is no longer allowed in comment and quoted-string text, and
5390  handling of backslash-escaping in them has been clarified.
5391  The quoted-pair rule no longer allows escaping control characters other than
5392  HTAB.
5393  Non-ASCII content in header fields and the reason phrase has been obsoleted
5394  and made opaque (the TEXT rule was removed).
5395  (<xref target="field.components"/>)
5398  Bogus "<x:ref>Content-Length</x:ref>" header fields are now required to be
5399  handled as errors by recipients.
5400  (<xref target="header.content-length"/>)
5403  The algorithm for determining the message body length has been clarified
5404  to indicate all of the special cases (e.g., driven by methods or status
5405  codes) that affect it, and that new protocol elements cannot define such
5406  special cases.
5407  CONNECT is a new, special case in determining message body length.
5408  "multipart/byteranges" is no longer a way of determining message body length
5409  detection.
5410  (<xref target="message.body.length"/>)
5413  The "identity" transfer coding token has been removed.
5414  (Sections <xref format="counter" target="message.body"/> and
5415  <xref format="counter" target="transfer.codings"/>)
5418  Chunk length does not include the count of the octets in the
5419  chunk header and trailer.
5420  Line folding in chunk extensions is  disallowed.
5421  (<xref target="chunked.encoding"/>)
5424  The meaning of the "deflate" content coding has been clarified.
5425  (<xref target="deflate.coding" />)
5428  The segment + query components of RFC 3986 have been used to define the
5429  request-target, instead of abs_path from RFC 1808.
5430  The asterisk-form of the request-target is only allowed with the OPTIONS
5431  method.
5432  (<xref target="request-target"/>)
5435  The term "Effective Request URI" has been introduced.
5436  (<xref target="effective.request.uri" />)
5439  Gateways do not need to generate <x:ref>Via</x:ref> header fields anymore.
5440  (<xref target="header.via"/>)
5443  Exactly when "close" connection options have to be sent has been clarified.
5444  Also, "hop-by-hop" header fields are required to appear in the Connection header
5445  field; just because they're defined as hop-by-hop in this specification
5446  doesn't exempt them.
5447  (<xref target="header.connection"/>)
5450  The limit of two connections per server has been removed.
5451  An idempotent sequence of requests is no longer required to be retried.
5452  The requirement to retry requests under certain circumstances when the
5453  server prematurely closes the connection has been removed.
5454  Also, some extraneous requirements about when servers are allowed to close
5455  connections prematurely have been removed.
5456  (<xref target="persistent.connections"/>)
5459  The semantics of the <x:ref>Upgrade</x:ref> header field is now defined in
5460  responses other than 101 (this was incorporated from <xref
5461  target="RFC2817"/>). Furthermore, the ordering in the field value is now
5462  significant.
5463  (<xref target="header.upgrade"/>)
5466  Empty list elements in list productions (e.g., a list header field containing
5467  ", ,") have been deprecated.
5468  (<xref target="abnf.extension"/>)
5471  Registration of Transfer Codings now requires IETF Review
5472  (<xref target="transfer.coding.registry"/>)
5475  This specification now defines the Upgrade Token Registry, previously
5476  defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
5477  (<xref target="upgrade.token.registry"/>)
5480  The expectation to support HTTP/0.9 requests has been removed.
5481  (<xref target="compatibility"/>)
5484  Issues with the Keep-Alive and Proxy-Connection header fields in requests
5485  are pointed out, with use of the latter being discouraged altogether.
5486  (<xref target="compatibility.with.http.1.0.persistent.connections" />)
5491<?BEGININC p1-messaging.abnf-appendix ?>
5492<section xmlns:x="" title="Collected ABNF" anchor="collected.abnf">
5494<artwork type="abnf" name="p1-messaging.parsed-abnf">
5495<x:ref>BWS</x:ref> = OWS
5497<x:ref>Connection</x:ref> = *( "," OWS ) connection-option *( OWS "," [ OWS
5498 connection-option ] )
5499<x:ref>Content-Length</x:ref> = 1*DIGIT
5501<x:ref>HTTP-message</x:ref> = start-line *( header-field CRLF ) CRLF [ message-body
5502 ]
5503<x:ref>HTTP-name</x:ref> = %x48.54.54.50 ; HTTP
5504<x:ref>HTTP-version</x:ref> = HTTP-name "/" DIGIT "." DIGIT
5505<x:ref>Host</x:ref> = uri-host [ ":" port ]
5507<x:ref>OWS</x:ref> = *( SP / HTAB )
5509<x:ref>RWS</x:ref> = 1*( SP / HTAB )
5511<x:ref>TE</x:ref> = [ ( "," / t-codings ) *( OWS "," [ OWS t-codings ] ) ]
5512<x:ref>Trailer</x:ref> = *( "," OWS ) field-name *( OWS "," [ OWS field-name ] )
5513<x:ref>Transfer-Encoding</x:ref> = *( "," OWS ) transfer-coding *( OWS "," [ OWS
5514 transfer-coding ] )
5516<x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in [RFC3986], Section 4.1&gt;
5517<x:ref>Upgrade</x:ref> = *( "," OWS ) protocol *( OWS "," [ OWS protocol ] )
5519<x:ref>Via</x:ref> = *( "," OWS ) ( received-protocol RWS received-by [ RWS comment
5520 ] ) *( OWS "," [ OWS ( received-protocol RWS received-by [ RWS
5521 comment ] ) ] )
5523<x:ref>absolute-URI</x:ref> = &lt;absolute-URI, defined in [RFC3986], Section 4.3&gt;
5524<x:ref>absolute-form</x:ref> = absolute-URI
5525<x:ref>absolute-path</x:ref> = 1*( "/" segment )
5526<x:ref>asterisk-form</x:ref> = "*"
5527<x:ref>authority</x:ref> = &lt;authority, defined in [RFC3986], Section 3.2&gt;
5528<x:ref>authority-form</x:ref> = authority
5530<x:ref>chunk</x:ref> = chunk-size [ chunk-ext ] CRLF chunk-data CRLF
5531<x:ref>chunk-data</x:ref> = 1*OCTET
5532<x:ref>chunk-ext</x:ref> = *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
5533<x:ref>chunk-ext-name</x:ref> = token
5534<x:ref>chunk-ext-val</x:ref> = token / quoted-string
5535<x:ref>chunk-size</x:ref> = 1*HEXDIG
5536<x:ref>chunked-body</x:ref> = *chunk last-chunk trailer-part CRLF
5537<x:ref>comment</x:ref> = "(" *( ctext / quoted-pair / comment ) ")"
5538<x:ref>connection-option</x:ref> = token
5539<x:ref>ctext</x:ref> = HTAB / SP / %x21-27 ; '!'-'''
5540 / %x2A-5B ; '*'-'['
5541 / %x5D-7E ; ']'-'~'
5542 / obs-text
5544<x:ref>field-content</x:ref> = field-vchar [ 1*( SP / HTAB ) field-vchar ]
5545<x:ref>field-name</x:ref> = token
5546<x:ref>field-value</x:ref> = *( field-content / obs-fold )
5547<x:ref>field-vchar</x:ref> = VCHAR / obs-text
5548<x:ref>fragment</x:ref> = &lt;fragment, defined in [RFC3986], Section 3.5&gt;
5550<x:ref>header-field</x:ref> = field-name ":" OWS field-value OWS
5551<x:ref>http-URI</x:ref> = "http://" authority path-abempty [ "?" query ] [ "#"
5552 fragment ]
5553<x:ref>https-URI</x:ref> = "https://" authority path-abempty [ "?" query ] [ "#"
5554 fragment ]
5556<x:ref>last-chunk</x:ref> = 1*"0" [ chunk-ext ] CRLF
5558<x:ref>message-body</x:ref> = *OCTET
5559<x:ref>method</x:ref> = token
5561<x:ref>obs-fold</x:ref> = CRLF 1*( SP / HTAB )
5562<x:ref>obs-text</x:ref> = %x80-FF
5563<x:ref>origin-form</x:ref> = absolute-path [ "?" query ]
5565<x:ref>partial-URI</x:ref> = relative-part [ "?" query ]
5566<x:ref>path-abempty</x:ref> = &lt;path-abempty, defined in [RFC3986], Section 3.3&gt;
5567<x:ref>port</x:ref> = &lt;port, defined in [RFC3986], Section 3.2.3&gt;
5568<x:ref>protocol</x:ref> = protocol-name [ "/" protocol-version ]
5569<x:ref>protocol-name</x:ref> = token
5570<x:ref>protocol-version</x:ref> = token
5571<x:ref>pseudonym</x:ref> = token
5573<x:ref>qdtext</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5574 / %x5D-7E ; ']'-'~'
5575 / obs-text
5576<x:ref>query</x:ref> = &lt;query, defined in [RFC3986], Section 3.4&gt;
5577<x:ref>quoted-pair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5578<x:ref>quoted-string</x:ref> = DQUOTE *( qdtext / quoted-pair ) DQUOTE
5580<x:ref>rank</x:ref> = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] )
5581<x:ref>reason-phrase</x:ref> = *( HTAB / SP / VCHAR / obs-text )
5582<x:ref>received-by</x:ref> = ( uri-host [ ":" port ] ) / pseudonym
5583<x:ref>received-protocol</x:ref> = [ protocol-name "/" ] protocol-version
5584<x:ref>relative-part</x:ref> = &lt;relative-part, defined in [RFC3986], Section 4.2&gt;
5585<x:ref>request-line</x:ref> = method SP request-target SP HTTP-version CRLF
5586<x:ref>request-target</x:ref> = origin-form / absolute-form / authority-form /
5587 asterisk-form
5589<x:ref>scheme</x:ref> = &lt;scheme, defined in [RFC3986], Section 3.1&gt;
5590<x:ref>segment</x:ref> = &lt;segment, defined in [RFC3986], Section 3.3&gt;
5591<x:ref>start-line</x:ref> = request-line / status-line
5592<x:ref>status-code</x:ref> = 3DIGIT
5593<x:ref>status-line</x:ref> = HTTP-version SP status-code SP reason-phrase CRLF
5595<x:ref>t-codings</x:ref> = "trailers" / ( transfer-coding [ t-ranking ] )
5596<x:ref>t-ranking</x:ref> = OWS ";" OWS "q=" rank
5597<x:ref>tchar</x:ref> = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*" / "+" / "-" / "." /
5598 "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
5599<x:ref>token</x:ref> = 1*tchar
5600<x:ref>trailer-part</x:ref> = *( header-field CRLF )
5601<x:ref>transfer-coding</x:ref> = "chunked" / "compress" / "deflate" / "gzip" /
5602 transfer-extension
5603<x:ref>transfer-extension</x:ref> = token *( OWS ";" OWS transfer-parameter )
5604<x:ref>transfer-parameter</x:ref> = token BWS "=" BWS ( token / quoted-string )
5606<x:ref>uri-host</x:ref> = &lt;host, defined in [RFC3986], Section 3.2.2&gt;
5610<?ENDINC p1-messaging.abnf-appendix ?>
Note: See TracBrowser for help on using the repository browser.