source: draft-ietf-httpbis/latest/p1-messaging.xml @ 203

Last change on this file since 203 was 203, checked in by julian.reschke@…, 15 years ago

Add processing instructions for inline comments throughout, and use them for the IANA TBDs.

  • Property svn:eol-style set to native
File size: 180.9 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 "February">
16  <!ENTITY ID-YEAR "2008">
17  <!ENTITY caching                "<xref target='Part6' x:rel='#caching' xmlns:x=''/>">
18  <!ENTITY payload                "<xref target='Part3' xmlns:x=''/>">
19  <!ENTITY media-types            "<xref target='Part3' x:rel='#media.types' xmlns:x=''/>">
20  <!ENTITY content-codings        "<xref target='Part3' x:rel='#content.codings' xmlns:x=''/>">
21  <!ENTITY CONNECT                "<xref target='Part2' x:rel='#CONNECT' xmlns:x=''/>">
22  <!ENTITY content.negotiation    "<xref target='Part3' x:rel='#content.negotiation' xmlns:x=''/>">
23  <!ENTITY diff2045entity         "<xref target='Part3' x:rel='#differences.between.http.entities.and.rfc.2045.entities' xmlns:x=''/>">
24  <!ENTITY entity                 "<xref target='Part3' x:rel='#entity' xmlns:x=''/>">
25  <!ENTITY entity-header-fields   "<xref target='Part3' x:rel='#entity.header.fields' xmlns:x=''/>">
26  <!ENTITY header-cache-control   "<xref target='Part6' x:rel='#header.cache-control' xmlns:x=''/>">
27  <!ENTITY header-expect          "<xref target='Part2' x:rel='#header.expect' xmlns:x=''/>">
28  <!ENTITY header-pragma          "<xref target='Part6' x:rel='#header.pragma' xmlns:x=''/>">
29  <!ENTITY header-warning         "<xref target='Part6' x:rel='#header.warning' xmlns:x=''/>">
30  <!ENTITY idempotent-methods     "<xref target='Part2' x:rel='#idempotent.methods' xmlns:x=''/>">
31  <!ENTITY qvalue                 "<xref target='Part3' x:rel='#quality.values' xmlns:x=''/>">
32  <!ENTITY request-header-fields  "<xref target='Part2' x:rel='#request.header.fields' xmlns:x=''/>">
33  <!ENTITY response-header-fields "<xref target='Part2' x:rel='#response.header.fields' xmlns:x=''/>">
34  <!ENTITY method                 "<xref target='Part2' x:rel='#method' xmlns:x=''/>">
35  <!ENTITY status-codes           "<xref target='Part2' x:rel='' xmlns:x=''/>">
36  <!ENTITY status-100             "<xref target='Part2' x:rel='#status.100' xmlns:x=''/>">
37  <!ENTITY status-1xx             "<xref target='Part2' x:rel='#status.1xx' xmlns:x=''/>">
38  <!ENTITY status-414             "<xref target='Part2' x:rel='#status.414' xmlns:x=''/>">
40<?rfc toc="yes" ?>
41<?rfc symrefs="yes" ?>
42<?rfc sortrefs="yes" ?>
43<?rfc compact="yes"?>
44<?rfc subcompact="no" ?>
45<?rfc linkmailto="no" ?>
46<?rfc editing="no" ?>
47<?rfc comments="yes"?>
48<?rfc inline="yes"?>
49<?rfc-ext allow-markup-in-artwork="yes" ?>
50<?rfc-ext include-references-in-index="yes" ?>
51<rfc obsoletes="2616" category="std"
52     ipr="full3978" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
53     xmlns:x=''>
56  <title abbrev="HTTP/1.1, Part 1">HTTP/1.1, part 1: URIs, Connections, and Message Parsing</title>
58  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
59    <organization abbrev="Day Software">Day Software</organization>
60    <address>
61      <postal>
62        <street>23 Corporate Plaza DR, Suite 280</street>
63        <city>Newport Beach</city>
64        <region>CA</region>
65        <code>92660</code>
66        <country>USA</country>
67      </postal>
68      <phone>+1-949-706-5300</phone>
69      <facsimile>+1-949-706-5305</facsimile>
70      <email></email>
71      <uri></uri>
72    </address>
73  </author>
75  <author initials="J." surname="Gettys" fullname="Jim Gettys">
76    <organization>One Laptop per Child</organization>
77    <address>
78      <postal>
79        <street>21 Oak Knoll Road</street>
80        <city>Carlisle</city>
81        <region>MA</region>
82        <code>01741</code>
83        <country>USA</country>
84      </postal>
85      <email></email>
86      <uri></uri>
87    </address>
88  </author>
90  <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
91    <organization abbrev="HP">Hewlett-Packard Company</organization>
92    <address>
93      <postal>
94        <street>HP Labs, Large Scale Systems Group</street>
95        <street>1501 Page Mill Road, MS 1177</street>
96        <city>Palo Alto</city>
97        <region>CA</region>
98        <code>94304</code>
99        <country>USA</country>
100      </postal>
101      <email></email>
102    </address>
103  </author>
105  <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
106    <organization abbrev="Microsoft">Microsoft Corporation</organization>
107    <address>
108      <postal>
109        <street>1 Microsoft Way</street>
110        <city>Redmond</city>
111        <region>WA</region>
112        <code>98052</code>
113        <country>USA</country>
114      </postal>
115      <email></email>
116    </address>
117  </author>
119  <author initials="L." surname="Masinter" fullname="Larry Masinter">
120    <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
121    <address>
122      <postal>
123        <street>345 Park Ave</street>
124        <city>San Jose</city>
125        <region>CA</region>
126        <code>95110</code>
127        <country>USA</country>
128      </postal>
129      <email></email>
130      <uri></uri>
131    </address>
132  </author>
134  <author initials="P." surname="Leach" fullname="Paul J. Leach">
135    <organization abbrev="Microsoft">Microsoft Corporation</organization>
136    <address>
137      <postal>
138        <street>1 Microsoft Way</street>
139        <city>Redmond</city>
140        <region>WA</region>
141        <code>98052</code>
142      </postal>
143      <email></email>
144    </address>
145  </author>
147  <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
148    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
149    <address>
150      <postal>
151        <street>MIT Computer Science and Artificial Intelligence Laboratory</street>
152        <street>The Stata Center, Building 32</street>
153        <street>32 Vassar Street</street>
154        <city>Cambridge</city>
155        <region>MA</region>
156        <code>02139</code>
157        <country>USA</country>
158      </postal>
159      <email></email>
160      <uri></uri>
161    </address>
162  </author>
164  <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
165    <organization abbrev="W3C">World Wide Web Consortium</organization>
166    <address>
167      <postal>
168        <street>W3C / ERCIM</street>
169        <street>2004, rte des Lucioles</street>
170        <city>Sophia-Antipolis</city>
171        <region>AM</region>
172        <code>06902</code>
173        <country>France</country>
174      </postal>
175      <email></email>
176      <uri></uri>
177    </address>
178  </author>
180  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
181    <organization abbrev="greenbytes">greenbytes GmbH</organization>
182    <address>
183      <postal>
184        <street>Hafenweg 16</street>
185        <city>Muenster</city><region>NW</region><code>48155</code>
186        <country>Germany</country>
187      </postal>
188      <phone>+49 251 2807760</phone>   
189      <facsimile>+49 251 2807761</facsimile>   
190      <email></email>       
191      <uri></uri>     
192    </address>
193  </author>
195  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
199   The Hypertext Transfer Protocol (HTTP) is an application-level
200   protocol for distributed, collaborative, hypermedia information
201   systems. HTTP has been in use by the World Wide Web global information
202   initiative since 1990. This document is Part 1 of the seven-part specification
203   that defines the protocol referred to as "HTTP/1.1" and, taken together,
204   obsoletes RFC 2616.  Part 1 provides an overview of HTTP and
205   its associated terminology, defines the "http" and "https" Uniform
206   Resource Identifier (URI) schemes, defines the generic message syntax
207   and parsing requirements for HTTP message frames, and describes
208   general security concerns for implementations.
212<note title="Editorial Note (To be removed by RFC Editor)">
213  <t>
214    Discussion of this draft should take place on the HTTPBIS working group
215    mailing list ( The current issues list is
216    at <eref target=""/>
217    and related documents (including fancy diffs) can be found at
218    <eref target=""/>.
219  </t>
220  <t>
221    This draft incorporates those issue resolutions that were either
222    collected in the original RFC2616 errata list (<eref target=""/>),
223    or which were agreed upon on the mailing list between October 2006 and
224    November 2007 (as published in "draft-lafon-rfc2616bis-03").
225  </t>
229<section title="Introduction" anchor="introduction">
231   The Hypertext Transfer Protocol (HTTP) is an application-level
232   protocol for distributed, collaborative, hypermedia information
233   systems. HTTP has been in use by the World-Wide Web global
234   information initiative since 1990. The first version of HTTP, commonly
235   referred to as HTTP/0.9, was a simple protocol for raw data transfer
236   across the Internet with only a single method and no metadata.
237   HTTP/1.0, as defined by <xref target="RFC1945"/>, improved
238   the protocol by allowing messages to be in the format of MIME-like
239   messages, containing metadata about the data transferred and
240   modifiers on the request/response semantics. However, HTTP/1.0 did
241   not sufficiently take into consideration the effects of hierarchical
242   proxies, caching, the need for persistent connections, or name-based
243   virtual hosts. In addition, the proliferation of incompletely-implemented
244   applications calling themselves "HTTP/1.0" necessitated a
245   protocol version change in order for two communicating applications
246   to determine each other's true capabilities.
249   This document is Part 1 of the seven-part specification that defines
250   the protocol referred to as "HTTP/1.1", obsoleting <xref target="RFC2616"/>.
251   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
252   requirements that enable reliable implementations and adding only
253   those new features that will either be safely ignored by an HTTP/1.0
254   recipient or only sent when communicating with a party advertising
255   compliance with HTTP/1.1.
256   Part 1 defines those aspects of HTTP/1.1 related to overall network
257   operation, message framing, interaction with transport protocols, and
258   URI schemes.
261   This document is currently disorganized in order to minimize the changes
262   between drafts and enable reviewers to see the smaller errata changes.
263   The next draft will reorganize the sections to better reflect the content.
264   In particular, the sections will be organized according to the typical
265   process of deciding when to use HTTP (URI schemes), overall network operation,
266   connection management, message framing, and generic message parsing.
267   The current mess reflects how widely dispersed these topics and associated
268   requirements had become in <xref target="RFC2616"/>.
271<section title="Purpose" anchor="intro.purpose">
273   Practical information systems require more functionality than simple
274   retrieval, including search, front-end update, and annotation. HTTP
275   allows an open-ended set of methods and headers that indicate the
276   purpose of a request <xref target="RFC2324"/>. It builds on the discipline of reference
277   provided by the Uniform Resource Identifier (URI) <xref target="RFC1630"/>, as a location
278   (URL) <xref target="RFC1738"/> or name (URN) <xref target="RFC1737"/>, for indicating the resource to which a
279   method is to be applied. Messages are passed in a format similar to
280   that used by Internet mail <xref target="RFC2822"/> as defined by the Multipurpose
281   Internet Mail Extensions (MIME) <xref target="RFC2045"/>.
284   HTTP is also used as a generic protocol for communication between
285   user agents and proxies/gateways to other Internet systems, including
286   those supported by the SMTP <xref target="RFC2821"/>, NNTP <xref target="RFC3977"/>, FTP <xref target="RFC959"/>, Gopher <xref target="RFC1436"/>,
287   and WAIS <xref target="WAIS"/> protocols. In this way, HTTP allows basic hypermedia
288   access to resources available from diverse applications.
292<section title="Requirements" anchor="intro.requirements">
294   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
295   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
296   document are to be interpreted as described in <xref target="RFC2119"/>.
299   An implementation is not compliant if it fails to satisfy one or more
300   of the &MUST; or &REQUIRED; level requirements for the protocols it
301   implements. An implementation that satisfies all the &MUST; or &REQUIRED;
302   level and all the &SHOULD; level requirements for its protocols is said
303   to be "unconditionally compliant"; one that satisfies all the &MUST;
304   level requirements but not all the &SHOULD; level requirements for its
305   protocols is said to be "conditionally compliant."
309<section title="Terminology" anchor="intro.terminology">
311   This specification uses a number of terms to refer to the roles
312   played by participants in, and objects of, the HTTP communication.
315  <iref item="connection"/>
316  <x:dfn>connection</x:dfn>
317  <list>
318    <t>
319      A transport layer virtual circuit established between two programs
320      for the purpose of communication.
321    </t>
322  </list>
325  <iref item="message"/>
326  <x:dfn>message</x:dfn>
327  <list>
328    <t>
329      The basic unit of HTTP communication, consisting of a structured
330      sequence of octets matching the syntax defined in <xref target="http.message"/> and
331      transmitted via the connection.
332    </t>
333  </list>
336  <iref item="request"/>
337  <x:dfn>request</x:dfn>
338  <list>
339    <t>
340      An HTTP request message, as defined in <xref target="request"/>.
341    </t>
342  </list>
345  <iref item="response"/>
346  <x:dfn>response</x:dfn>
347  <list>
348    <t>
349      An HTTP response message, as defined in <xref target="response"/>.
350    </t>
351  </list>
354  <iref item="resource"/>
355  <x:dfn>resource</x:dfn>
356  <list>
357    <t>
358      A network data object or service that can be identified by a URI,
359      as defined in <xref target="uri"/>. Resources may be available in multiple
360      representations (e.g. multiple languages, data formats, size, and
361      resolutions) or vary in other ways.
362    </t>
363  </list>
366  <iref item="entity"/>
367  <x:dfn>entity</x:dfn>
368  <list>
369    <t>
370      The information transferred as the payload of a request or
371      response. An entity consists of metainformation in the form of
372      entity-header fields and content in the form of an entity-body, as
373      described in &entity;.
374    </t>
375  </list>
378  <iref item="representation"/>
379  <x:dfn>representation</x:dfn>
380  <list>
381    <t>
382      An entity included with a response that is subject to content
383      negotiation, as described in &content.negotiation;. There may exist multiple
384      representations associated with a particular response status.
385    </t>
386  </list>
389  <iref item="content negotiation"/>
390  <x:dfn>content negotiation</x:dfn>
391  <list>
392    <t>
393      The mechanism for selecting the appropriate representation when
394      servicing a request, as described in &content.negotiation;. The
395      representation of entities in any response can be negotiated
396      (including error responses).
397    </t>
398  </list>
401  <iref item="variant"/>
402  <x:dfn>variant</x:dfn>
403  <list>
404    <t>
405      A resource may have one, or more than one, representation(s)
406      associated with it at any given instant. Each of these
407      representations is termed a `variant'.  Use of the term `variant'
408      does not necessarily imply that the resource is subject to content
409      negotiation.
410    </t>
411  </list>
414  <iref item="client"/>
415  <x:dfn>client</x:dfn>
416  <list>
417    <t>
418      A program that establishes connections for the purpose of sending
419      requests.
420    </t>
421  </list>
424  <iref item="user agent"/>
425  <x:dfn>user agent</x:dfn>
426  <list>
427    <t>
428      The client which initiates a request. These are often browsers,
429      editors, spiders (web-traversing robots), or other end user tools.
430    </t>
431  </list>
434  <iref item="server"/>
435  <x:dfn>server</x:dfn>
436  <list>
437    <t>
438      An application program that accepts connections in order to
439      service requests by sending back responses. Any given program may
440      be capable of being both a client and a server; our use of these
441      terms refers only to the role being performed by the program for a
442      particular connection, rather than to the program's capabilities
443      in general. Likewise, any server may act as an origin server,
444      proxy, gateway, or tunnel, switching behavior based on the nature
445      of each request.
446    </t>
447  </list>
450  <iref item="origin server"/>
451  <x:dfn>origin server</x:dfn>
452  <list>
453    <t>
454      The server on which a given resource resides or is to be created.
455    </t>
456  </list>
459  <iref item="proxy"/>
460  <x:dfn>proxy</x:dfn>
461  <list>
462    <t>
463      An intermediary program which acts as both a server and a client
464      for the purpose of making requests on behalf of other clients.
465      Requests are serviced internally or by passing them on, with
466      possible translation, to other servers. A proxy &MUST; implement
467      both the client and server requirements of this specification. A
468      "transparent proxy" is a proxy that does not modify the request or
469      response beyond what is required for proxy authentication and
470      identification. A "non-transparent proxy" is a proxy that modifies
471      the request or response in order to provide some added service to
472      the user agent, such as group annotation services, media type
473      transformation, protocol reduction, or anonymity filtering. Except
474      where either transparent or non-transparent behavior is explicitly
475      stated, the HTTP proxy requirements apply to both types of
476      proxies.
477    </t>
478  </list>
481  <iref item="gateway"/>
482  <x:dfn>gateway</x:dfn>
483  <list>
484    <t>
485      A server which acts as an intermediary for some other server.
486      Unlike a proxy, a gateway receives requests as if it were the
487      origin server for the requested resource; the requesting client
488      may not be aware that it is communicating with a gateway.
489    </t>
490  </list>
493  <iref item="tunnel"/>
494  <x:dfn>tunnel</x:dfn>
495  <list>
496    <t>
497      An intermediary program which is acting as a blind relay between
498      two connections. Once active, a tunnel is not considered a party
499      to the HTTP communication, though the tunnel may have been
500      initiated by an HTTP request. The tunnel ceases to exist when both
501      ends of the relayed connections are closed.
502    </t>
503  </list>
506  <iref item="cache"/>
507  <x:dfn>cache</x:dfn>
508  <list>
509    <t>
510      A program's local store of response messages and the subsystem
511      that controls its message storage, retrieval, and deletion. A
512      cache stores cacheable responses in order to reduce the response
513      time and network bandwidth consumption on future, equivalent
514      requests. Any client or server may include a cache, though a cache
515      cannot be used by a server that is acting as a tunnel.
516    </t>
517  </list>
520  <iref item="cacheable"/>
521  <x:dfn>cacheable</x:dfn>
522  <list>
523    <t>
524      A response is cacheable if a cache is allowed to store a copy of
525      the response message for use in answering subsequent requests. The
526      rules for determining the cacheability of HTTP responses are
527      defined in &caching;. Even if a resource is cacheable, there may
528      be additional constraints on whether a cache can use the cached
529      copy for a particular request.
530    </t>
531  </list>
534  <iref item="upstream"/>
535  <iref item="downstream"/>
536  <x:dfn>upstream</x:dfn>/<x:dfn>downstream</x:dfn>
537  <list>
538    <t>
539      Upstream and downstream describe the flow of a message: all
540      messages flow from upstream to downstream.
541    </t>
542  </list>
545  <iref item="inbound"/>
546  <iref item="outbound"/>
547  <x:dfn>inbound</x:dfn>/<x:dfn>outbound</x:dfn>
548  <list>
549    <t>
550      Inbound and outbound refer to the request and response paths for
551      messages: "inbound" means "traveling toward the origin server",
552      and "outbound" means "traveling toward the user agent"
553    </t>
554  </list>
558<section title="Overall Operation" anchor="intro.overall.operation">
560   HTTP is a request/response protocol. A client sends a
561   request to the server in the form of a request method, URI, and
562   protocol version, followed by a MIME-like message containing request
563   modifiers, client information, and possible body content over a
564   connection with a server. The server responds with a status line,
565   including the message's protocol version and a success or error code,
566   followed by a MIME-like message containing server information, entity
567   metainformation, and possible entity-body content. The relationship
568   between HTTP and MIME is described in &diff2045entity;.
571   Most HTTP communication is initiated by a user agent and consists of
572   a request to be applied to a resource on some origin server. In the
573   simplest case, this may be accomplished via a single connection (v)
574   between the user agent (UA) and the origin server (O).
576<figure><artwork type="drawing">
577       request chain ------------------------&gt;
578    UA -------------------v------------------- O
579       &lt;----------------------- response chain
582   A more complicated situation occurs when one or more intermediaries
583   are present in the request/response chain. There are three common
584   forms of intermediary: proxy, gateway, and tunnel. A proxy is a
585   forwarding agent, receiving requests for a URI in its absolute form,
586   rewriting all or part of the message, and forwarding the reformatted
587   request toward the server identified by the URI. A gateway is a
588   receiving agent, acting as a layer above some other server(s) and, if
589   necessary, translating the requests to the underlying server's
590   protocol. A tunnel acts as a relay point between two connections
591   without changing the messages; tunnels are used when the
592   communication needs to pass through an intermediary (such as a
593   firewall) even when the intermediary cannot understand the contents
594   of the messages.
596<figure><artwork type="drawing">
597       request chain --------------------------------------&gt;
598    UA -----v----- A -----v----- B -----v----- C -----v----- O
599       &lt;------------------------------------- response chain
602   The figure above shows three intermediaries (A, B, and C) between the
603   user agent and origin server. A request or response message that
604   travels the whole chain will pass through four separate connections.
605   This distinction is important because some HTTP communication options
606   may apply only to the connection with the nearest, non-tunnel
607   neighbor, only to the end-points of the chain, or to all connections
608   along the chain. Although the diagram is linear, each participant may
609   be engaged in multiple, simultaneous communications. For example, B
610   may be receiving requests from many clients other than A, and/or
611   forwarding requests to servers other than C, at the same time that it
612   is handling A's request.
615   Any party to the communication which is not acting as a tunnel may
616   employ an internal cache for handling requests. The effect of a cache
617   is that the request/response chain is shortened if one of the
618   participants along the chain has a cached response applicable to that
619   request. The following illustrates the resulting chain if B has a
620   cached copy of an earlier response from O (via C) for a request which
621   has not been cached by UA or A.
623<figure><artwork type="drawing">
624          request chain ----------&gt;
625       UA -----v----- A -----v----- B - - - - - - C - - - - - - O
626          &lt;--------- response chain
629   Not all responses are usefully cacheable, and some requests may
630   contain modifiers which place special requirements on cache behavior.
631   HTTP requirements for cache behavior and cacheable responses are
632   defined in &caching;.
635   In fact, there are a wide variety of architectures and configurations
636   of caches and proxies currently being experimented with or deployed
637   across the World Wide Web. These systems include national hierarchies
638   of proxy caches to save transoceanic bandwidth, systems that
639   broadcast or multicast cache entries, organizations that distribute
640   subsets of cached data via CD-ROM, and so on. HTTP systems are used
641   in corporate intranets over high-bandwidth links, and for access via
642   PDAs with low-power radio links and intermittent connectivity. The
643   goal of HTTP/1.1 is to support the wide diversity of configurations
644   already deployed while introducing protocol constructs that meet the
645   needs of those who build web applications that require high
646   reliability and, failing that, at least reliable indications of
647   failure.
650   HTTP communication usually takes place over TCP/IP connections. The
651   default port is TCP 80 (<eref target=""/>), but other ports can be used. This does
652   not preclude HTTP from being implemented on top of any other protocol
653   on the Internet, or on other networks. HTTP only presumes a reliable
654   transport; any protocol that provides such guarantees can be used;
655   the mapping of the HTTP/1.1 request and response structures onto the
656   transport data units of the protocol in question is outside the scope
657   of this specification.
660   In HTTP/1.0, most implementations used a new connection for each
661   request/response exchange. In HTTP/1.1, a connection may be used for
662   one or more request/response exchanges, although connections may be
663   closed for a variety of reasons (see <xref target="persistent.connections"/>).
668<section title="Notational Conventions and Generic Grammar" anchor="notation">
670<section title="Augmented BNF" anchor="notation.abnf">
672   All of the mechanisms specified in this document are described in
673   both prose and an augmented Backus-Naur Form (BNF) similar to that
674   used by <xref target="RFC822ABNF"/>. Implementors will need to be familiar with the
675   notation in order to understand this specification. The augmented BNF
676   includes the following constructs:
679   name = definition
680  <list>
681    <t>
682      The name of a rule is simply the name itself (without any
683      enclosing "&lt;" and "&gt;") and is separated from its definition by the
684      equal "=" character. White space is only significant in that
685      indentation of continuation lines is used to indicate a rule
686      definition that spans more than one line. Certain basic rules are
687      in uppercase, such as SP, LWS, HTAB, CRLF, DIGIT, ALPHA, etc. Angle
688      brackets are used within definitions whenever their presence will
689      facilitate discerning the use of rule names.
690    </t>
691  </list>
694   "literal"
695  <list>
696    <t>
697      Quotation marks surround literal text. Unless stated otherwise,
698      the text is case-insensitive.
699    </t>
700  </list>
703   rule1 | rule2
704  <list>
705    <t>
706      Elements separated by a bar ("|") are alternatives, e.g., "yes |
707      no" will accept yes or no.
708    </t>
709  </list>
712   (rule1 rule2)
713  <list>
714    <t>
715      Elements enclosed in parentheses are treated as a single element.
716      Thus, "(elem (foo | bar) elem)" allows the token sequences "elem
717      foo elem" and "elem bar elem".
718    </t>
719  </list>
722   *rule
723  <list>
724    <t>
725      The character "*" preceding an element indicates repetition. The
726      full form is "&lt;n&gt;*&lt;m&gt;element" indicating at least &lt;n&gt; and at most
727      &lt;m&gt; occurrences of element. Default values are 0 and infinity so
728      that "*(element)" allows any number, including zero; "1*element"
729      requires at least one; and "1*2element" allows one or two.
730    </t>
731  </list>
734   [rule]
735  <list>
736    <t>
737      Square brackets enclose optional elements; "[foo bar]" is
738      equivalent to "*1(foo bar)".
739    </t>
740  </list>
743   N rule
744  <list>
745    <t>
746      Specific repetition: "&lt;n&gt;(element)" is equivalent to
747      "&lt;n&gt;*&lt;n&gt;(element)"; that is, exactly &lt;n&gt; occurrences of (element).
748      Thus 2DIGIT is a 2-digit number, and 3ALPHA is a string of three
749      alphabetic characters.
750    </t>
751  </list>
754   #rule
755  <list>
756    <t>
757      A construct "#" is defined, similar to "*", for defining lists of
758      elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating at least
759      &lt;n&gt; and at most &lt;m&gt; elements, each separated by one or more commas
760      (",") and &OPTIONAL; linear white space (LWS). This makes the usual
761      form of lists very easy; a rule such as
762    </t>
763    <t>
764         ( *LWS element *( *LWS "," *LWS element ))
765    </t>
766    <t>
767      can be shown as
768    </t>
769    <t>
770         1#element
771    </t>
772    <t>
773      Wherever this construct is used, null elements are allowed, but do
774      not contribute to the count of elements present. That is,
775      "(element), , (element) " is permitted, but counts as only two
776      elements. Therefore, where at least one element is required, at
777      least one non-null element &MUST; be present. Default values are 0
778      and infinity so that "#element" allows any number, including zero;
779      "1#element" requires at least one; and "1#2element" allows one or
780      two.
781    </t>
782  </list>
785   ; comment
786  <list>
787    <t>
788      A semi-colon, set off some distance to the right of rule text,
789      starts a comment that continues to the end of line. This is a
790      simple way of including useful notes in parallel with the
791      specifications.
792    </t>
793  </list>
796   implied *LWS
797  <list>
798    <t>
799      The grammar described by this specification is word-based. Except
800      where noted otherwise, linear white space (LWS) can be included
801      between any two adjacent words (token or quoted-string), and
802      between adjacent words and separators, without changing the
803      interpretation of a field. At least one delimiter (LWS and/or
804      separators) &MUST; exist between any two tokens (for the definition
805      of "token" below), since they would otherwise be interpreted as a
806      single token.
807    </t>
808  </list>
812<section title="Basic Rules" anchor="basic.rules">
813<x:anchor-alias value="OCTET"/>
814<x:anchor-alias value="CHAR"/>
815<x:anchor-alias value="ALPHA"/>
816<x:anchor-alias value="DIGIT"/>
817<x:anchor-alias value="CTL"/>
818<x:anchor-alias value="CR"/>
819<x:anchor-alias value="LF"/>
820<x:anchor-alias value="SP"/>
821<x:anchor-alias value="HTAB"/>
822<x:anchor-alias value="CRLF"/>
823<x:anchor-alias value="LWS"/>
824<x:anchor-alias value="TEXT"/>
825<x:anchor-alias value="HEX"/>
826<x:anchor-alias value="token"/>
827<x:anchor-alias value="separators"/>
828<x:anchor-alias value="comment"/>
829<x:anchor-alias value="ctext"/>
830<x:anchor-alias value="quoted-string"/>
831<x:anchor-alias value="qdtext"/>
832<x:anchor-alias value="quoted-pair"/>
834   The following rules are used throughout this specification to
835   describe basic parsing constructs. The US-ASCII coded character set
836   is defined by ANSI X3.4-1986 <xref target="USASCII"/>.
838<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="OCTET"/><iref primary="true" item="Grammar" subitem="CHAR"/><iref primary="true" item="Grammar" subitem="ALPHA"/><iref primary="true" item="Grammar" subitem="DIGIT"/><iref primary="true" item="Grammar" subitem="CTL"/><iref primary="true" item="Grammar" subitem="CR"/><iref primary="true" item="Grammar" subitem="LF"/><iref primary="true" item="Grammar" subitem="SP"/><iref primary="true" item="Grammar" subitem="HTAB"/><iref primary="true" item="Grammar" subitem="DQUOTE"/>
839  OCTET          = %x00-FF
840                   ; any 8-bit sequence of data
841  CHAR           = %x01-7F
842                   ; any US-ASCII character, excluding NUL
843  ALPHA          = %x41-5A | %x61-7A
844                   ; A-Z | a-z
845  DIGIT          = %x30-39
846                   ; any US-ASCII digit "0".."9"
847  CTL            = %x00-1F | %x7F
848                   ; (octets 0 - 31) and DEL (127)
849  CR             = %x0D
850                   ; US-ASCII CR, carriage return (13)
851  LF             = %x0A
852                   ; US-ASCII LF, linefeed (10)
853  SP             = %x20
854                   ; US-ASCII SP, space (32)
855  HTAB           = %x09
856                   ; US-ASCII HT, horizontal-tab (9)
857  DQUOTE         = %x22
858                   ; US-ASCII double-quote mark (34)
861   HTTP/1.1 defines the sequence CR LF as the end-of-line marker for all
862   protocol elements except the entity-body (see <xref target="tolerant.applications"/> for
863   tolerant applications). The end-of-line marker within an entity-body
864   is defined by its associated media type, as described in &media-types;.
866<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="CRLF"/>
867  CRLF           = CR LF
870   HTTP/1.1 header field values can be folded onto multiple lines if the
871   continuation line begins with a space or horizontal tab. All linear
872   white space, including folding, has the same semantics as SP. A
873   recipient &MAY; replace any linear white space with a single SP before
874   interpreting the field value or forwarding the message downstream.
876<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="LWS"/>
877  LWS            = [CRLF] 1*( SP | HTAB )
880   The TEXT rule is only used for descriptive field contents and values
881   that are not intended to be interpreted by the message parser. Words
882   of *TEXT &MAY; contain characters from character sets other than ISO-8859-1
883   <xref target="ISO-8859-1"/> only when encoded according to the rules of
884   <xref target="RFC2047"/>.
886<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TEXT"/>
887  TEXT           = &lt;any OCTET except CTLs, but including LWS&gt;
890   A CRLF is allowed in the definition of TEXT only as part of a header
891   field continuation. It is expected that the folding LWS will be
892   replaced with a single SP before interpretation of the TEXT value.
895   Hexadecimal numeric characters are used in several protocol elements.
897<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HEX"/>
898  HEX            = "A" | "B" | "C" | "D" | "E" | "F"
899                 | "a" | "b" | "c" | "d" | "e" | "f" | DIGIT
902   Many HTTP/1.1 header field values consist of words separated by LWS
903   or special characters. These special characters &MUST; be in a quoted
904   string to be used within a parameter value (as defined in
905   <xref target="transfer.codings"/>).
907<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="separators"/>
908  token          = 1*&lt;any CHAR except CTLs or separators&gt;
909  separators     = "(" | ")" | "&lt;" | "&gt;" | "@"
910                 | "," | ";" | ":" | "\" | DQUOTE
911                 | "/" | "[" | "]" | "?" | "="
912                 | "{" | "}" | SP | HTAB
915   Comments can be included in some HTTP header fields by surrounding
916   the comment text with parentheses. Comments are only allowed in
917   fields containing "comment" as part of their field value definition.
918   In all other fields, parentheses are considered part of the field
919   value.
921<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
922  comment        = "(" *( ctext | quoted-pair | comment ) ")"
923  ctext          = &lt;any TEXT excluding "(" and ")"&gt;
926   A string of text is parsed as a single word if it is quoted using
927   double-quote marks.
929<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-string"/><iref primary="true" item="Grammar" subitem="qdtext"/>
930  quoted-string  = ( DQUOTE *(qdtext | quoted-pair ) DQUOTE )
931  qdtext         = &lt;any TEXT excluding DQUOTE and "\">
934   The backslash character ("\") &MAY; be used as a single-character
935   quoting mechanism only within quoted-string and comment constructs.
937<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
938  quoted-pair    = "\" CHAR
943<section title="Protocol Parameters" anchor="protocol.parameters">
945<section title="HTTP Version" anchor="http.version">
947   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate versions
948   of the protocol. The protocol versioning policy is intended to allow
949   the sender to indicate the format of a message and its capacity for
950   understanding further HTTP communication, rather than the features
951   obtained via that communication. No change is made to the version
952   number for the addition of message components which do not affect
953   communication behavior or which only add to extensible field values.
954   The &lt;minor&gt; number is incremented when the changes made to the
955   protocol add features which do not change the general message parsing
956   algorithm, but which may add to the message semantics and imply
957   additional capabilities of the sender. The &lt;major&gt; number is
958   incremented when the format of a message within the protocol is
959   changed. See <xref target="RFC2145"/> for a fuller explanation.
962   The version of an HTTP message is indicated by an HTTP-Version field
963   in the first line of the message. HTTP-Version is case-sensitive.
965<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-Version"/>
966  HTTP-Version   = "HTTP" "/" 1*DIGIT "." 1*DIGIT
969   Note that the major and minor numbers &MUST; be treated as separate
970   integers and that each &MAY; be incremented higher than a single digit.
971   Thus, HTTP/2.4 is a lower version than HTTP/2.13, which in turn is
972   lower than HTTP/12.3. Leading zeros &MUST; be ignored by recipients and
973   &MUST-NOT; be sent.
976   An application that sends a request or response message that includes
977   HTTP-Version of "HTTP/1.1" &MUST; be at least conditionally compliant
978   with this specification. Applications that are at least conditionally
979   compliant with this specification &SHOULD; use an HTTP-Version of
980   "HTTP/1.1" in their messages, and &MUST; do so for any message that is
981   not compatible with HTTP/1.0. For more details on when to send
982   specific HTTP-Version values, see <xref target="RFC2145"/>.
985   The HTTP version of an application is the highest HTTP version for
986   which the application is at least conditionally compliant.
989   Proxy and gateway applications need to be careful when forwarding
990   messages in protocol versions different from that of the application.
991   Since the protocol version indicates the protocol capability of the
992   sender, a proxy/gateway &MUST-NOT; send a message with a version
993   indicator which is greater than its actual version. If a higher
994   version request is received, the proxy/gateway &MUST; either downgrade
995   the request version, or respond with an error, or switch to tunnel
996   behavior.
999   Due to interoperability problems with HTTP/1.0 proxies discovered
1000   since the publication of <xref target="RFC2068"/>, caching proxies &MUST;, gateways
1001   &MAY;, and tunnels &MUST-NOT; upgrade the request to the highest version
1002   they support. The proxy/gateway's response to that request &MUST; be in
1003   the same major version as the request.
1006  <list>
1007    <t>
1008      <x:h>Note:</x:h> Converting between versions of HTTP may involve modification
1009      of header fields required or forbidden by the versions involved.
1010    </t>
1011  </list>
1015<section title="Uniform Resource Identifiers" anchor="uri">
1017   URIs have been known by many names: WWW addresses, Universal Document
1018   Identifiers, Universal Resource Identifiers <xref target="RFC1630"/>, and finally the
1019   combination of Uniform Resource Locators (URL) <xref target="RFC1738"/> and Names (URN)
1020   <xref target="RFC1737"/>. As far as HTTP is concerned, Uniform Resource Identifiers are
1021   simply formatted strings which identify--via name, location, or any
1022   other characteristic--a resource.
1025<section title="General Syntax" anchor="general.syntax">
1027   URIs in HTTP can be represented in absolute form or relative to some
1028   known base URI <xref target="RFC1808"/>, depending upon the context of their use. The two
1029   forms are differentiated by the fact that absolute URIs always begin
1030   with a scheme name followed by a colon. For definitive information on
1031   URL syntax and semantics, see "Uniform Resource Identifiers (URI):
1032   Generic Syntax and Semantics," <xref target="RFC2396"/> (which replaces <xref target="RFC1738"/>
1033   and <xref target="RFC1808"/>). This specification adopts the
1034   definitions of "URI-reference", "absoluteURI", "relativeURI", "port",
1035   "host", "abs_path", "query", and "authority" from that specification:
1037<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="absoluteURI"/><iref primary="true" item="Grammar" subitem="authority"/><iref primary="true" item="Grammar" subitem="path-absolute"/><iref primary="true" item="Grammar" subitem="port"/><iref primary="true" item="Grammar" subitem="query"/><iref primary="true" item="Grammar" subitem="relativeURI"/><iref primary="true" item="Grammar" subitem="uri-host"/>
1038  absoluteURI   = &lt;absoluteURI, defined in <xref target="RFC2396" x:fmt="," x:sec="3"/>>
1039  authority     = &lt;authority, defined in <xref target="RFC2396" x:fmt="," x:sec="3.2"/>>
1040  path-absolute = &lt;abs_path, defined in <xref target="RFC2396" x:fmt="," x:sec="3"/>>
1041  port          = &lt;port, defined in <xref target="RFC2396" x:fmt="," x:sec="3.2.2"/>>
1042  query         = &lt;query, defined in <xref target="RFC2396" x:fmt="," x:sec="3.4"/>>
1043  relativeURI   = &lt;relativeURI, defined in <xref target="RFC2396" x:fmt="," x:sec="5"/>>
1044  uri-host      = &lt;host, defined in <xref target="RFC2396" x:fmt="," x:sec="3.2.2"/>>
1047   HTTP does not place any a priori limit on the length of
1048   a URI. Servers &MUST; be able to handle the URI of any resource they
1049   serve, and &SHOULD; be able to handle URIs of unbounded length if they
1050   provide GET-based forms that could generate such URIs. A server
1051   &SHOULD; return 414 (Request-URI Too Long) status if a URI is longer
1052   than the server can handle (see &status-414;).
1055  <list>
1056    <t>
1057      <x:h>Note:</x:h> Servers ought to be cautious about depending on URI lengths
1058      above 255 bytes, because some older client or proxy
1059      implementations might not properly support these lengths.
1060    </t>
1061  </list>
1065<section title="http URL" anchor="http.url">
1067   The "http" scheme is used to locate network resources via the HTTP
1068   protocol. This section defines the scheme-specific syntax and
1069   semantics for http URLs.
1071<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http-URL"/>
1072  http-URL = "http:" "//" uri-host [ ":" port ]
1073             [ path-absolute [ "?" query ]]
1076   If the port is empty or not given, port 80 is assumed. The semantics
1077   are that the identified resource is located at the server listening
1078   for TCP connections on that port of that host, and the Request-URI
1079   for the resource is path-absolute (<xref target="request-uri"/>). The use of IP addresses
1080   in URLs &SHOULD; be avoided whenever possible (see <xref target="RFC1900"/>). If
1081   the path-absolute is not present in the URL, it &MUST; be given as "/" when
1082   used as a Request-URI for a resource (<xref target="request-uri"/>). If a proxy
1083   receives a host name which is not a fully qualified domain name, it
1084   &MAY; add its domain to the host name it received. If a proxy receives
1085   a fully qualified domain name, the proxy &MUST-NOT; change the host
1086   name.
1090<section title="URI Comparison" anchor="uri.comparison">
1092   When comparing two URIs to decide if they match or not, a client
1093   &SHOULD; use a case-sensitive octet-by-octet comparison of the entire
1094   URIs, with these exceptions:
1095  <list style="symbols">
1096    <t>A port that is empty or not given is equivalent to the default
1097        port for that URI-reference;</t>
1098    <t>Comparisons of host names &MUST; be case-insensitive;</t>
1099    <t>Comparisons of scheme names &MUST; be case-insensitive;</t>
1100    <t>An empty path-absolute is equivalent to an path-absolute of "/".</t>
1101  </list>
1104   Characters other than those in the "reserved" set (see
1105   <xref target="RFC2396"/>) are equivalent to their ""%" HEX HEX" encoding.
1108   For example, the following three URIs are equivalent:
1110<figure><artwork type="example">
1118<section title="Date/Time Formats" anchor="date.time.formats">
1119<section title="Full Date" anchor="">
1121   HTTP applications have historically allowed three different formats
1122   for the representation of date/time stamps:
1124<figure><artwork type="example">
1125   Sun, 06 Nov 1994 08:49:37 GMT  ; RFC 822, updated by RFC 1123
1126   Sunday, 06-Nov-94 08:49:37 GMT ; obsolete RFC 850 format
1127   Sun Nov  6 08:49:37 1994       ; ANSI C's asctime() format
1130   The first format is preferred as an Internet standard and represents
1131   a fixed-length subset of that defined by <xref target="RFC1123"/> (an update to
1132   <xref target="RFC822"/>). The other formats are described here only for
1133   compatibility with obsolete implementations.
1134   HTTP/1.1 clients and servers that parse the date value &MUST; accept
1135   all three formats (for compatibility with HTTP/1.0), though they &MUST;
1136   only generate the RFC 1123 format for representing HTTP-date values
1137   in header fields. See <xref target="tolerant.applications"/> for further information.
1140      <x:h>Note:</x:h> Recipients of date values are encouraged to be robust in
1141      accepting date values that may have been sent by non-HTTP
1142      applications, as is sometimes the case when retrieving or posting
1143      messages via proxies/gateways to SMTP or NNTP.
1146   All HTTP date/time stamps &MUST; be represented in Greenwich Mean Time
1147   (GMT), without exception. For the purposes of HTTP, GMT is exactly
1148   equal to UTC (Coordinated Universal Time). This is indicated in the
1149   first two formats by the inclusion of "GMT" as the three-letter
1150   abbreviation for time zone, and &MUST; be assumed when reading the
1151   asctime format. HTTP-date is case sensitive and &MUST-NOT; include
1152   additional LWS beyond that specifically included as SP in the
1153   grammar.
1155<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-date"/><iref primary="true" item="Grammar" subitem="rfc1123-date"/><iref primary="true" item="Grammar" subitem="rfc850-date"/><iref primary="true" item="Grammar" subitem="asctime-date"/><iref primary="true" item="Grammar" subitem="date1"/><iref primary="true" item="Grammar" subitem="date2"/><iref primary="true" item="Grammar" subitem="date3"/><iref primary="true" item="Grammar" subitem="time"/><iref primary="true" item="Grammar" subitem="wkday"/><iref primary="true" item="Grammar" subitem="weekday"/><iref primary="true" item="Grammar" subitem="month"/>
1156  HTTP-date    = rfc1123-date | rfc850-date | asctime-date
1157  rfc1123-date = wkday "," SP date1 SP time SP "GMT"
1158  rfc850-date  = weekday "," SP date2 SP time SP "GMT"
1159  asctime-date = wkday SP date3 SP time SP 4DIGIT
1160  date1        = 2DIGIT SP month SP 4DIGIT
1161                 ; day month year (e.g., 02 Jun 1982)
1162  date2        = 2DIGIT "-" month "-" 2DIGIT
1163                 ; day-month-year (e.g., 02-Jun-82)
1164  date3        = month SP ( 2DIGIT | ( SP 1DIGIT ))
1165                 ; month day (e.g., Jun  2)
1166  time         = 2DIGIT ":" 2DIGIT ":" 2DIGIT
1167                 ; 00:00:00 - 23:59:59
1168  wkday        = "Mon" | "Tue" | "Wed"
1169               | "Thu" | "Fri" | "Sat" | "Sun"
1170  weekday      = "Monday" | "Tuesday" | "Wednesday"
1171               | "Thursday" | "Friday" | "Saturday" | "Sunday"
1172  month        = "Jan" | "Feb" | "Mar" | "Apr"
1173               | "May" | "Jun" | "Jul" | "Aug"
1174               | "Sep" | "Oct" | "Nov" | "Dec"
1177      <x:h>Note:</x:h> HTTP requirements for the date/time stamp format apply only
1178      to their usage within the protocol stream. Clients and servers are
1179      not required to use these formats for user presentation, request
1180      logging, etc.
1185<section title="Transfer Codings" anchor="transfer.codings">
1187   Transfer-coding values are used to indicate an encoding
1188   transformation that has been, can be, or may need to be applied to an
1189   entity-body in order to ensure "safe transport" through the network.
1190   This differs from a content coding in that the transfer-coding is a
1191   property of the message, not of the original entity.
1193<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1194  transfer-coding         = "chunked" | transfer-extension
1195  transfer-extension      = token *( ";" parameter )
1198   Parameters are in  the form of attribute/value pairs.
1200<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="parameter"/><iref primary="true" item="Grammar" subitem="attribute"/><iref primary="true" item="Grammar" subitem="value"/>
1201  parameter               = attribute "=" value
1202  attribute               = token
1203  value                   = token | quoted-string
1206   All transfer-coding values are case-insensitive. HTTP/1.1 uses
1207   transfer-coding values in the TE header field (<xref target="header.te"/>) and in
1208   the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1211   Whenever a transfer-coding is applied to a message-body, the set of
1212   transfer-codings &MUST; include "chunked", unless the message is
1213   terminated by closing the connection. When the "chunked" transfer-coding
1214   is used, it &MUST; be the last transfer-coding applied to the
1215   message-body. The "chunked" transfer-coding &MUST-NOT; be applied more
1216   than once to a message-body. These rules allow the recipient to
1217   determine the transfer-length of the message (<xref target="message.length"/>).
1220   Transfer-codings are analogous to the Content-Transfer-Encoding
1221   values of MIME <xref target="RFC2045"/>, which were designed to enable safe transport of
1222   binary data over a 7-bit transport service. However, safe transport
1223   has a different focus for an 8bit-clean transfer protocol. In HTTP,
1224   the only unsafe characteristic of message-bodies is the difficulty in
1225   determining the exact body length (<xref target="message.length"/>), or the desire to
1226   encrypt data over a shared transport.
1229   The Internet Assigned Numbers Authority (IANA) acts as a registry for
1230   transfer-coding value tokens. Initially, the registry contains the
1231   following tokens: "chunked" (<xref target="chunked.transfer.encoding"/>),
1232   "gzip", "compress", and "deflate" (&content-codings;).
1235   New transfer-coding value tokens &SHOULD; be registered in the same way
1236   as new content-coding value tokens (&content-codings;).
1239   A server which receives an entity-body with a transfer-coding it does
1240   not understand &SHOULD; return 501 (Not Implemented), and close the
1241   connection. A server &MUST-NOT; send transfer-codings to an HTTP/1.0
1242   client.
1245<section title="Chunked Transfer Coding" anchor="chunked.transfer.encoding">
1247   The chunked encoding modifies the body of a message in order to
1248   transfer it as a series of chunks, each with its own size indicator,
1249   followed by an &OPTIONAL; trailer containing entity-header fields. This
1250   allows dynamically produced content to be transferred along with the
1251   information necessary for the recipient to verify that it has
1252   received the full message.
1254<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Chunked-Body"/><iref primary="true" item="Grammar" subitem="chunk"/><iref primary="true" item="Grammar" subitem="chunk-size"/><iref primary="true" item="Grammar" subitem="last-chunk"/><iref primary="true" item="Grammar" subitem="chunk-extension"/><iref primary="true" item="Grammar" subitem="chunk-ext-name"/><iref primary="true" item="Grammar" subitem="chunk-ext-val"/><iref primary="true" item="Grammar" subitem="chunk-data"/><iref primary="true" item="Grammar" subitem="trailer-part"/>
1255  Chunked-Body   = *chunk
1256                   last-chunk
1257                   trailer-part
1258                   CRLF
1260  chunk          = chunk-size [ chunk-extension ] CRLF
1261                   chunk-data CRLF
1262  chunk-size     = 1*HEX
1263  last-chunk     = 1*("0") [ chunk-extension ] CRLF
1265  chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
1266  chunk-ext-name = token
1267  chunk-ext-val  = token | quoted-string
1268  chunk-data     = 1*OCTET ; a sequence of chunk-size octets
1269  trailer-part   = *(entity-header CRLF)
1272   The chunk-size field is a string of hex digits indicating the size of
1273   the chunk-data in octets. The chunked encoding is ended by any chunk whose size is
1274   zero, followed by the trailer, which is terminated by an empty line.
1277   The trailer allows the sender to include additional HTTP header
1278   fields at the end of the message. The Trailer header field can be
1279   used to indicate which header fields are included in a trailer (see
1280   <xref target="header.trailer"/>).
1283   A server using chunked transfer-coding in a response &MUST-NOT; use the
1284   trailer for any header fields unless at least one of the following is
1285   true:
1286  <list style="numbers">
1287    <t>the request included a TE header field that indicates "trailers" is
1288     acceptable in the transfer-coding of the  response, as described in
1289     <xref target="header.te"/>; or,</t>
1291    <t>the server is the origin server for the response, the trailer
1292     fields consist entirely of optional metadata, and the recipient
1293     could use the message (in a manner acceptable to the origin server)
1294     without receiving this metadata.  In other words, the origin server
1295     is willing to accept the possibility that the trailer fields might
1296     be silently discarded along the path to the client.</t>
1297  </list>
1300   This requirement prevents an interoperability failure when the
1301   message is being received by an HTTP/1.1 (or later) proxy and
1302   forwarded to an HTTP/1.0 recipient. It avoids a situation where
1303   compliance with the protocol would have necessitated a possibly
1304   infinite buffer on the proxy.
1307   A process for decoding the "chunked" transfer-coding
1308   can be represented in pseudo-code as:
1310<figure><artwork type="code">
1311    length := 0
1312    read chunk-size, chunk-extension (if any) and CRLF
1313    while (chunk-size &gt; 0) {
1314       read chunk-data and CRLF
1315       append chunk-data to entity-body
1316       length := length + chunk-size
1317       read chunk-size and CRLF
1318    }
1319    read entity-header
1320    while (entity-header not empty) {
1321       append entity-header to existing header fields
1322       read entity-header
1323    }
1324    Content-Length := length
1325    Remove "chunked" from Transfer-Encoding
1328   All HTTP/1.1 applications &MUST; be able to receive and decode the
1329   "chunked" transfer-coding, and &MUST; ignore chunk-extension extensions
1330   they do not understand.
1335<section title="Product Tokens" anchor="product.tokens">
1337   Product tokens are used to allow communicating applications to
1338   identify themselves by software name and version. Most fields using
1339   product tokens also allow sub-products which form a significant part
1340   of the application to be listed, separated by white space. By
1341   convention, the products are listed in order of their significance
1342   for identifying the application.
1344<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="product"/><iref primary="true" item="Grammar" subitem="product-version"/>
1345  product         = token ["/" product-version]
1346  product-version = token
1349   Examples:
1351<figure><artwork type="example">
1352    User-Agent: CERN-LineMode/2.15 libwww/2.17b3
1353    Server: Apache/0.8.4
1356   Product tokens &SHOULD; be short and to the point. They &MUST-NOT; be
1357   used for advertising or other non-essential information. Although any
1358   token character &MAY; appear in a product-version, this token &SHOULD;
1359   only be used for a version identifier (i.e., successive versions of
1360   the same product &SHOULD; only differ in the product-version portion of
1361   the product value).
1367<section title="HTTP Message" anchor="http.message">
1369<section title="Message Types" anchor="message.types">
1371   HTTP messages consist of requests from client to server and responses
1372   from server to client.
1374<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"/>
1375  HTTP-message   = Request | Response     ; HTTP/1.1 messages
1378   Request (<xref target="request"/>) and Response (<xref target="response"/>) messages use the generic
1379   message format of <xref target="RFC2822"/> for transferring entities (the payload
1380   of the message). Both types of message consist of a start-line, zero
1381   or more header fields (also known as "headers"), an empty line (i.e.,
1382   a line with nothing preceding the CRLF) indicating the end of the
1383   header fields, and possibly a message-body.
1385<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="generic-message"/><iref primary="true" item="Grammar" subitem="start-line"/>
1386  generic-message = start-line
1387                    *(message-header CRLF)
1388                    CRLF
1389                    [ message-body ]
1390  start-line      = Request-Line | Status-Line
1393   In the interest of robustness, servers &SHOULD; ignore any empty
1394   line(s) received where a Request-Line is expected. In other words, if
1395   the server is reading the protocol stream at the beginning of a
1396   message and receives a CRLF first, it should ignore the CRLF.
1399   Certain buggy HTTP/1.0 client implementations generate extra CRLF's
1400   after a POST request. To restate what is explicitly forbidden by the
1401   BNF, an HTTP/1.1 client &MUST-NOT; preface or follow a request with an
1402   extra CRLF.
1406<section title="Message Headers" anchor="message.headers">
1408   HTTP header fields, which include general-header (<xref target="general.header.fields"/>),
1409   request-header (&request-header-fields;), response-header (&response-header-fields;), and
1410   entity-header (&entity-header-fields;) fields, follow the same generic format as
1411   that given in <xref target="RFC2822" x:fmt="of" x:sec="2.1"/>. Each header field consists
1412   of a name followed by a colon (":") and the field value. Field names
1413   are case-insensitive. The field value &MAY; be preceded by any amount
1414   of LWS, though a single SP is preferred. Header fields can be
1415   extended over multiple lines by preceding each extra line with at
1416   least one SP or HTAB. Applications ought to follow "common form", where
1417   one is known or indicated, when generating HTTP constructs, since
1418   there might exist some implementations that fail to accept anything
1419   beyond the common forms.
1421<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-header"/><iref primary="true" item="Grammar" subitem="field-name"/><iref primary="true" item="Grammar" subitem="field-value"/><iref primary="true" item="Grammar" subitem="field-content"/>
1422  message-header = field-name ":" [ field-value ]
1423  field-name     = token
1424  field-value    = *( field-content | LWS )
1425  field-content  = &lt;field content&gt;
1426                   ; the OCTETs making up the field-value
1427                   ; and consisting of either *TEXT or combinations
1428                   ; of token, separators, and quoted-string
1431   The field-content does not include any leading or trailing LWS:
1432   linear white space occurring before the first non-whitespace
1433   character of the field-value or after the last non-whitespace
1434   character of the field-value. Such leading or trailing LWS &MAY; be
1435   removed without changing the semantics of the field value. Any LWS
1436   that occurs between field-content &MAY; be replaced with a single SP
1437   before interpreting the field value or forwarding the message
1438   downstream.
1441   The order in which header fields with differing field names are
1442   received is not significant. However, it is "good practice" to send
1443   general-header fields first, followed by request-header or response-header
1444   fields, and ending with the entity-header fields.
1447   Multiple message-header fields with the same field-name &MAY; be
1448   present in a message if and only if the entire field-value for that
1449   header field is defined as a comma-separated list [i.e., #(values)].
1450   It &MUST; be possible to combine the multiple header fields into one
1451   "field-name: field-value" pair, without changing the semantics of the
1452   message, by appending each subsequent field-value to the first, each
1453   separated by a comma. The order in which header fields with the same
1454   field-name are received is therefore significant to the
1455   interpretation of the combined field value, and thus a proxy &MUST-NOT;
1456   change the order of these field values when a message is forwarded.
1460<section title="Message Body" anchor="message.body">
1462   The message-body (if any) of an HTTP message is used to carry the
1463   entity-body associated with the request or response. The message-body
1464   differs from the entity-body only when a transfer-coding has been
1465   applied, as indicated by the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1467<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1468  message-body = entity-body
1469               | &lt;entity-body encoded as per Transfer-Encoding&gt;
1472   Transfer-Encoding &MUST; be used to indicate any transfer-codings
1473   applied by an application to ensure safe and proper transfer of the
1474   message. Transfer-Encoding is a property of the message, not of the
1475   entity, and thus &MAY; be added or removed by any application along the
1476   request/response chain. (However, <xref target="transfer.codings"/> places restrictions on
1477   when certain transfer-codings may be used.)
1480   The rules for when a message-body is allowed in a message differ for
1481   requests and responses.
1484   The presence of a message-body in a request is signaled by the
1485   inclusion of a Content-Length or Transfer-Encoding header field in
1486   the request's message-headers. A message-body &MUST-NOT; be included in
1487   a request if the specification of the request method (&method;)
1488   explicitly disallows an entity-body in requests.
1489   When a request message contains both a message-body of non-zero
1490   length and a method that does not define any semantics for that
1491   request message-body, then an origin server &SHOULD; either ignore
1492   the message-body or respond with an appropriate error message
1493   (e.g., 413).  A proxy or gateway, when presented the same request,
1494   &SHOULD; either forward the request inbound with the message-body or
1495   ignore the message-body when determining a response.
1498   For response messages, whether or not a message-body is included with
1499   a message is dependent on both the request method and the response
1500   status code (<xref target="status.code.and.reason.phrase"/>). All responses to the HEAD request method
1501   &MUST-NOT; include a message-body, even though the presence of entity-header
1502   fields might lead one to believe they do. All 1xx
1503   (informational), 204 (No Content), and 304 (Not Modified) responses
1504   &MUST-NOT; include a message-body. All other responses do include a
1505   message-body, although it &MAY; be of zero length.
1509<section title="Message Length" anchor="message.length">
1511   The transfer-length of a message is the length of the message-body as
1512   it appears in the message; that is, after any transfer-codings have
1513   been applied. When a message-body is included with a message, the
1514   transfer-length of that body is determined by one of the following
1515   (in order of precedence):
1518  <list style="numbers">
1519    <x:lt><t>
1520     Any response message which "&MUST-NOT;" include a message-body (such
1521     as the 1xx, 204, and 304 responses and any response to a HEAD
1522     request) is always terminated by the first empty line after the
1523     header fields, regardless of the entity-header fields present in
1524     the message.
1525    </t></x:lt>
1526    <x:lt><t>
1527     If a Transfer-Encoding header field (<xref target="header.transfer-encoding"/>)
1528     is present, then the transfer-length is
1529     defined by use of the "chunked" transfer-coding (<xref target="transfer.codings"/>),
1530     unless the message is terminated by closing the connection.
1531    </t></x:lt>
1532    <x:lt><t>
1533     If a Content-Length header field (<xref target="header.content-length"/>) is present, its
1534     decimal value in OCTETs represents both the entity-length and the
1535     transfer-length. The Content-Length header field &MUST-NOT; be sent
1536     if these two lengths are different (i.e., if a Transfer-Encoding
1537     header field is present). If a message is received with both a
1538     Transfer-Encoding header field and a Content-Length header field,
1539     the latter &MUST; be ignored.
1540    </t></x:lt>
1541    <x:lt><t>
1542     If the message uses the media type "multipart/byteranges", and the
1543     transfer-length is not otherwise specified, then this self-delimiting
1544     media type defines the transfer-length. This media type
1545     &MUST-NOT; be used unless the sender knows that the recipient can parse
1546     it; the presence in a request of a Range header with multiple byte-range
1547     specifiers from a 1.1 client implies that the client can parse
1548     multipart/byteranges responses.
1549    <list style="empty"><t>
1550       A range header might be forwarded by a 1.0 proxy that does not
1551       understand multipart/byteranges; in this case the server &MUST;
1552       delimit the message using methods defined in items 1, 3 or 5 of
1553       this section.
1554    </t></list>
1555    </t></x:lt>
1556    <x:lt><t>
1557     By the server closing the connection. (Closing the connection
1558     cannot be used to indicate the end of a request body, since that
1559     would leave no possibility for the server to send back a response.)
1560    </t></x:lt>
1561  </list>
1564   For compatibility with HTTP/1.0 applications, HTTP/1.1 requests
1565   containing a message-body &MUST; include a valid Content-Length header
1566   field unless the server is known to be HTTP/1.1 compliant. If a
1567   request contains a message-body and a Content-Length is not given,
1568   the server &SHOULD; respond with 400 (Bad Request) if it cannot
1569   determine the length of the message, or with 411 (Length Required) if
1570   it wishes to insist on receiving a valid Content-Length.
1573   All HTTP/1.1 applications that receive entities &MUST; accept the
1574   "chunked" transfer-coding (<xref target="transfer.codings"/>), thus allowing this mechanism
1575   to be used for messages when the message length cannot be determined
1576   in advance.
1579   Messages &MUST-NOT; include both a Content-Length header field and a
1580   transfer-coding. If the message does include a
1581   transfer-coding, the Content-Length &MUST; be ignored.
1584   When a Content-Length is given in a message where a message-body is
1585   allowed, its field value &MUST; exactly match the number of OCTETs in
1586   the message-body. HTTP/1.1 user agents &MUST; notify the user when an
1587   invalid length is received and detected.
1591<section title="General Header Fields" anchor="general.header.fields">
1593   There are a few header fields which have general applicability for
1594   both request and response messages, but which do not apply to the
1595   entity being transferred. These header fields apply only to the
1596   message being transmitted.
1598<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="general-header"/>
1599  general-header = Cache-Control            ; &header-cache-control;
1600                 | Connection               ; <xref target="header.connection"/>
1601                 | Date                     ; <xref target=""/>
1602                 | Pragma                   ; &header-pragma;
1603                 | Trailer                  ; <xref target="header.trailer"/>
1604                 | Transfer-Encoding        ; <xref target="header.transfer-encoding"/>
1605                 | Upgrade                  ; <xref target="header.upgrade"/>
1606                 | Via                      ; <xref target="header.via"/>
1607                 | Warning                  ; &header-warning;
1610   General-header field names can be extended reliably only in
1611   combination with a change in the protocol version. However, new or
1612   experimental header fields may be given the semantics of general
1613   header fields if all parties in the communication recognize them to
1614   be general-header fields. Unrecognized header fields are treated as
1615   entity-header fields.
1620<section title="Request" anchor="request">
1622   A request message from a client to a server includes, within the
1623   first line of that message, the method to be applied to the resource,
1624   the identifier of the resource, and the protocol version in use.
1626<!--                 Host                      ; should be moved here eventually -->
1627<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request"/>
1628  Request       = Request-Line              ; <xref target="request-line"/>
1629                  *(( general-header        ; <xref target="general.header.fields"/>
1630                   | request-header         ; &request-header-fields;
1631                   | entity-header ) CRLF)  ; &entity-header-fields;
1632                  CRLF
1633                  [ message-body ]          ; <xref target="message.body"/>
1636<section title="Request-Line" anchor="request-line">
1638   The Request-Line begins with a method token, followed by the
1639   Request-URI and the protocol version, and ending with CRLF. The
1640   elements are separated by SP characters. No CR or LF is allowed
1641   except in the final CRLF sequence.
1643<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-Line"/>
1644  Request-Line   = Method SP Request-URI SP HTTP-Version CRLF
1647<section title="Method" anchor="method">
1649   The Method  token indicates the method to be performed on the
1650   resource identified by the Request-URI. The method is case-sensitive.
1652<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Method"/><iref primary="true" item="Grammar" subitem="extension-method"/>
1653  Method         = token
1657<section title="Request-URI" anchor="request-uri">
1659   The Request-URI is a Uniform Resource Identifier (<xref target="uri"/>) and
1660   identifies the resource upon which to apply the request.
1662<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-URI"/>
1663  Request-URI    = "*"
1664                 | absoluteURI
1665                 | ( path-absolute [ "?" query ] )
1666                 | authority
1669   The four options for Request-URI are dependent on the nature of the
1670   request. The asterisk "*" means that the request does not apply to a
1671   particular resource, but to the server itself, and is only allowed
1672   when the method used does not necessarily apply to a resource. One
1673   example would be
1675<figure><artwork type="example">
1676    OPTIONS * HTTP/1.1
1679   The absoluteURI form is &REQUIRED; when the request is being made to a
1680   proxy. The proxy is requested to forward the request or service it
1681   from a valid cache, and return the response. Note that the proxy &MAY;
1682   forward the request on to another proxy or directly to the server
1683   specified by the absoluteURI. In order to avoid request loops, a
1684   proxy &MUST; be able to recognize all of its server names, including
1685   any aliases, local variations, and the numeric IP address. An example
1686   Request-Line would be:
1688<figure><artwork type="example">
1689    GET HTTP/1.1
1692   To allow for transition to absoluteURIs in all requests in future
1693   versions of HTTP, all HTTP/1.1 servers &MUST; accept the absoluteURI
1694   form in requests, even though HTTP/1.1 clients will only generate
1695   them in requests to proxies.
1698   The authority form is only used by the CONNECT method (&CONNECT;).
1701   The most common form of Request-URI is that used to identify a
1702   resource on an origin server or gateway. In this case the absolute
1703   path of the URI &MUST; be transmitted (see <xref target="general.syntax"/>, path-absolute) as
1704   the Request-URI, and the network location of the URI (authority) &MUST;
1705   be transmitted in a Host header field. For example, a client wishing
1706   to retrieve the resource above directly from the origin server would
1707   create a TCP connection to port 80 of the host "" and send
1708   the lines:
1710<figure><artwork type="example">
1711    GET /pub/WWW/TheProject.html HTTP/1.1
1712    Host:
1715   followed by the remainder of the Request. Note that the absolute path
1716   cannot be empty; if none is present in the original URI, it &MUST; be
1717   given as "/" (the server root).
1720   The Request-URI is transmitted in the format specified in
1721   <xref target="general.syntax"/>. If the Request-URI is encoded using the "% HEX HEX" encoding
1722   <xref target="RFC2396"/>, the origin server &MUST; decode the Request-URI in order to
1723   properly interpret the request. Servers &SHOULD; respond to invalid
1724   Request-URIs with an appropriate status code.
1727   A transparent proxy &MUST-NOT; rewrite the "path-absolute" part of the
1728   received Request-URI when forwarding it to the next inbound server,
1729   except as noted above to replace a null path-absolute with "/".
1732  <list><t>
1733      <x:h>Note:</x:h> The "no rewrite" rule prevents the proxy from changing the
1734      meaning of the request when the origin server is improperly using
1735      a non-reserved URI character for a reserved purpose.  Implementors
1736      should be aware that some pre-HTTP/1.1 proxies have been known to
1737      rewrite the Request-URI.
1738  </t></list>
1743<section title="The Resource Identified by a Request" anchor="">
1745   The exact resource identified by an Internet request is determined by
1746   examining both the Request-URI and the Host header field.
1749   An origin server that does not allow resources to differ by the
1750   requested host &MAY; ignore the Host header field value when
1751   determining the resource identified by an HTTP/1.1 request. (But see
1752   <xref target=""/>
1753   for other requirements on Host support in HTTP/1.1.)
1756   An origin server that does differentiate resources based on the host
1757   requested (sometimes referred to as virtual hosts or vanity host
1758   names) &MUST; use the following rules for determining the requested
1759   resource on an HTTP/1.1 request:
1760  <list style="numbers">
1761    <t>If Request-URI is an absoluteURI, the host is part of the
1762     Request-URI. Any Host header field value in the request &MUST; be
1763     ignored.</t>
1764    <t>If the Request-URI is not an absoluteURI, and the request includes
1765     a Host header field, the host is determined by the Host header
1766     field value.</t>
1767    <t>If the host as determined by rule 1 or 2 is not a valid host on
1768     the server, the response &MUST; be a 400 (Bad Request) error message.</t>
1769  </list>
1772   Recipients of an HTTP/1.0 request that lacks a Host header field &MAY;
1773   attempt to use heuristics (e.g., examination of the URI path for
1774   something unique to a particular host) in order to determine what
1775   exact resource is being requested.
1782<section title="Response" anchor="response">
1784   After receiving and interpreting a request message, a server responds
1785   with an HTTP response message.
1787<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Response"/>
1788  Response      = Status-Line               ; <xref target="status-line"/>
1789                  *(( general-header        ; <xref target="general.header.fields"/>
1790                   | response-header        ; &response-header-fields;
1791                   | entity-header ) CRLF)  ; &entity-header-fields;
1792                  CRLF
1793                  [ message-body ]          ; <xref target="message.body"/>
1796<section title="Status-Line" anchor="status-line">
1798   The first line of a Response message is the Status-Line, consisting
1799   of the protocol version followed by a numeric status code and its
1800   associated textual phrase, with each element separated by SP
1801   characters. No CR or LF is allowed except in the final CRLF sequence.
1803<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Line"/>
1804  Status-Line = HTTP-Version SP Status-Code SP Reason-Phrase CRLF
1807<section title="Status Code and Reason Phrase" anchor="status.code.and.reason.phrase">
1809   The Status-Code element is a 3-digit integer result code of the
1810   attempt to understand and satisfy the request. These codes are fully
1811   defined in &status-codes;.  The Reason Phrase exists for the sole
1812   purpose of providing a textual description associated with the numeric
1813   status code, out of deference to earlier Internet application protocols
1814   that were more frequently used with interactive text clients.
1815   A client &SHOULD; ignore the content of the Reason Phrase.
1818   The first digit of the Status-Code defines the class of response. The
1819   last two digits do not have any categorization role. There are 5
1820   values for the first digit:
1821  <list style="symbols">
1822    <t>
1823      1xx: Informational - Request received, continuing process
1824    </t>
1825    <t>
1826      2xx: Success - The action was successfully received,
1827        understood, and accepted
1828    </t>
1829    <t>
1830      3xx: Redirection - Further action must be taken in order to
1831        complete the request
1832    </t>
1833    <t>
1834      4xx: Client Error - The request contains bad syntax or cannot
1835        be fulfilled
1836    </t>
1837    <t>
1838      5xx: Server Error - The server failed to fulfill an apparently
1839        valid request
1840    </t>
1841  </list>
1843<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Code"/><iref primary="true" item="Grammar" subitem="extension-code"/><iref primary="true" item="Grammar" subitem="Reason-Phrase"/>
1844  Status-Code    = 3DIGIT
1845  Reason-Phrase  = *&lt;TEXT, excluding CR, LF&gt;
1853<section title="Connections" anchor="connections">
1855<section title="Persistent Connections" anchor="persistent.connections">
1857<section title="Purpose" anchor="persistent.purpose">
1859   Prior to persistent connections, a separate TCP connection was
1860   established to fetch each URL, increasing the load on HTTP servers
1861   and causing congestion on the Internet. The use of inline images and
1862   other associated data often require a client to make multiple
1863   requests of the same server in a short amount of time. Analysis of
1864   these performance problems and results from a prototype
1865   implementation are available <xref target="Pad1995"/> <xref target="Spe"/>. Implementation experience and
1866   measurements of actual HTTP/1.1 (<xref target="RFC2068" x:fmt="none">RFC 2068</xref>) implementations show good
1867   results <xref target="Nie1997"/>. Alternatives have also been explored, for example,
1868   T/TCP <xref target="Tou1998"/>.
1871   Persistent HTTP connections have a number of advantages:
1872  <list style="symbols">
1873      <t>
1874        By opening and closing fewer TCP connections, CPU time is saved
1875        in routers and hosts (clients, servers, proxies, gateways,
1876        tunnels, or caches), and memory used for TCP protocol control
1877        blocks can be saved in hosts.
1878      </t>
1879      <t>
1880        HTTP requests and responses can be pipelined on a connection.
1881        Pipelining allows a client to make multiple requests without
1882        waiting for each response, allowing a single TCP connection to
1883        be used much more efficiently, with much lower elapsed time.
1884      </t>
1885      <t>
1886        Network congestion is reduced by reducing the number of packets
1887        caused by TCP opens, and by allowing TCP sufficient time to
1888        determine the congestion state of the network.
1889      </t>
1890      <t>
1891        Latency on subsequent requests is reduced since there is no time
1892        spent in TCP's connection opening handshake.
1893      </t>
1894      <t>
1895        HTTP can evolve more gracefully, since errors can be reported
1896        without the penalty of closing the TCP connection. Clients using
1897        future versions of HTTP might optimistically try a new feature,
1898        but if communicating with an older server, retry with old
1899        semantics after an error is reported.
1900      </t>
1901    </list>
1904   HTTP implementations &SHOULD; implement persistent connections.
1908<section title="Overall Operation" anchor="persistent.overall">
1910   A significant difference between HTTP/1.1 and earlier versions of
1911   HTTP is that persistent connections are the default behavior of any
1912   HTTP connection. That is, unless otherwise indicated, the client
1913   &SHOULD; assume that the server will maintain a persistent connection,
1914   even after error responses from the server.
1917   Persistent connections provide a mechanism by which a client and a
1918   server can signal the close of a TCP connection. This signaling takes
1919   place using the Connection header field (<xref target="header.connection"/>). Once a close
1920   has been signaled, the client &MUST-NOT; send any more requests on that
1921   connection.
1924<section title="Negotiation" anchor="persistent.negotiation">
1926   An HTTP/1.1 server &MAY; assume that a HTTP/1.1 client intends to
1927   maintain a persistent connection unless a Connection header including
1928   the connection-token "close" was sent in the request. If the server
1929   chooses to close the connection immediately after sending the
1930   response, it &SHOULD; send a Connection header including the
1931   connection-token close.
1934   An HTTP/1.1 client &MAY; expect a connection to remain open, but would
1935   decide to keep it open based on whether the response from a server
1936   contains a Connection header with the connection-token close. In case
1937   the client does not want to maintain a connection for more than that
1938   request, it &SHOULD; send a Connection header including the
1939   connection-token close.
1942   If either the client or the server sends the close token in the
1943   Connection header, that request becomes the last one for the
1944   connection.
1947   Clients and servers &SHOULD-NOT;  assume that a persistent connection is
1948   maintained for HTTP versions less than 1.1 unless it is explicitly
1949   signaled. See <xref target="compatibility.with.http.1.0.persistent.connections"/> for more information on backward
1950   compatibility with HTTP/1.0 clients.
1953   In order to remain persistent, all messages on the connection &MUST;
1954   have a self-defined message length (i.e., one not defined by closure
1955   of the connection), as described in <xref target="message.length"/>.
1959<section title="Pipelining" anchor="pipelining">
1961   A client that supports persistent connections &MAY; "pipeline" its
1962   requests (i.e., send multiple requests without waiting for each
1963   response). A server &MUST; send its responses to those requests in the
1964   same order that the requests were received.
1967   Clients which assume persistent connections and pipeline immediately
1968   after connection establishment &SHOULD; be prepared to retry their
1969   connection if the first pipelined attempt fails. If a client does
1970   such a retry, it &MUST-NOT; pipeline before it knows the connection is
1971   persistent. Clients &MUST; also be prepared to resend their requests if
1972   the server closes the connection before sending all of the
1973   corresponding responses.
1976   Clients &SHOULD-NOT;  pipeline requests using non-idempotent methods or
1977   non-idempotent sequences of methods (see &idempotent-methods;). Otherwise, a
1978   premature termination of the transport connection could lead to
1979   indeterminate results. A client wishing to send a non-idempotent
1980   request &SHOULD; wait to send that request until it has received the
1981   response status for the previous request.
1986<section title="Proxy Servers" anchor="persistent.proxy">
1988   It is especially important that proxies correctly implement the
1989   properties of the Connection header field as specified in <xref target="header.connection"/>.
1992   The proxy server &MUST; signal persistent connections separately with
1993   its clients and the origin servers (or other proxy servers) that it
1994   connects to. Each persistent connection applies to only one transport
1995   link.
1998   A proxy server &MUST-NOT; establish a HTTP/1.1 persistent connection
1999   with an HTTP/1.0 client (but see <xref target="RFC2068"/> for information and
2000   discussion of the problems with the Keep-Alive header implemented by
2001   many HTTP/1.0 clients).
2005<section title="Practical Considerations" anchor="persistent.practical">
2007   Servers will usually have some time-out value beyond which they will
2008   no longer maintain an inactive connection. Proxy servers might make
2009   this a higher value since it is likely that the client will be making
2010   more connections through the same server. The use of persistent
2011   connections places no requirements on the length (or existence) of
2012   this time-out for either the client or the server.
2015   When a client or server wishes to time-out it &SHOULD; issue a graceful
2016   close on the transport connection. Clients and servers &SHOULD; both
2017   constantly watch for the other side of the transport close, and
2018   respond to it as appropriate. If a client or server does not detect
2019   the other side's close promptly it could cause unnecessary resource
2020   drain on the network.
2023   A client, server, or proxy &MAY; close the transport connection at any
2024   time. For example, a client might have started to send a new request
2025   at the same time that the server has decided to close the "idle"
2026   connection. From the server's point of view, the connection is being
2027   closed while it was idle, but from the client's point of view, a
2028   request is in progress.
2031   This means that clients, servers, and proxies &MUST; be able to recover
2032   from asynchronous close events. Client software &SHOULD; reopen the
2033   transport connection and retransmit the aborted sequence of requests
2034   without user interaction so long as the request sequence is
2035   idempotent (see &idempotent-methods;). Non-idempotent methods or sequences
2036   &MUST-NOT; be automatically retried, although user agents &MAY; offer a
2037   human operator the choice of retrying the request(s). Confirmation by
2038   user-agent software with semantic understanding of the application
2039   &MAY; substitute for user confirmation. The automatic retry &SHOULD-NOT;
2040   be repeated if the second sequence of requests fails.
2043   Servers &SHOULD; always respond to at least one request per connection,
2044   if at all possible. Servers &SHOULD-NOT;  close a connection in the
2045   middle of transmitting a response, unless a network or client failure
2046   is suspected.
2049   Clients that use persistent connections &SHOULD; limit the number of
2050   simultaneous connections that they maintain to a given server. A
2051   single-user client &SHOULD-NOT; maintain more than 2 connections with
2052   any server or proxy. A proxy &SHOULD; use up to 2*N connections to
2053   another server or proxy, where N is the number of simultaneously
2054   active users. These guidelines are intended to improve HTTP response
2055   times and avoid congestion.
2060<section title="Message Transmission Requirements" anchor="message.transmission.requirements">
2062<section title="Persistent Connections and Flow Control" anchor="persistent.flow">
2064   HTTP/1.1 servers &SHOULD; maintain persistent connections and use TCP's
2065   flow control mechanisms to resolve temporary overloads, rather than
2066   terminating connections with the expectation that clients will retry.
2067   The latter technique can exacerbate network congestion.
2071<section title="Monitoring Connections for Error Status Messages" anchor="persistent.monitor">
2073   An HTTP/1.1 (or later) client sending a message-body &SHOULD; monitor
2074   the network connection for an error status while it is transmitting
2075   the request. If the client sees an error status, it &SHOULD;
2076   immediately cease transmitting the body. If the body is being sent
2077   using a "chunked" encoding (<xref target="transfer.codings"/>), a zero length chunk and
2078   empty trailer &MAY; be used to prematurely mark the end of the message.
2079   If the body was preceded by a Content-Length header, the client &MUST;
2080   close the connection.
2084<section title="Use of the 100 (Continue) Status" anchor="use.of.the.100.status">
2086   The purpose of the 100 (Continue) status (see &status-100;) is to
2087   allow a client that is sending a request message with a request body
2088   to determine if the origin server is willing to accept the request
2089   (based on the request headers) before the client sends the request
2090   body. In some cases, it might either be inappropriate or highly
2091   inefficient for the client to send the body if the server will reject
2092   the message without looking at the body.
2095   Requirements for HTTP/1.1 clients:
2096  <list style="symbols">
2097    <t>
2098        If a client will wait for a 100 (Continue) response before
2099        sending the request body, it &MUST; send an Expect request-header
2100        field (&header-expect;) with the "100-continue" expectation.
2101    </t>
2102    <t>
2103        A client &MUST-NOT; send an Expect request-header field (&header-expect;)
2104        with the "100-continue" expectation if it does not intend
2105        to send a request body.
2106    </t>
2107  </list>
2110   Because of the presence of older implementations, the protocol allows
2111   ambiguous situations in which a client may send "Expect: 100-continue"
2112   without receiving either a 417 (Expectation Failed) status
2113   or a 100 (Continue) status. Therefore, when a client sends this
2114   header field to an origin server (possibly via a proxy) from which it
2115   has never seen a 100 (Continue) status, the client &SHOULD-NOT;  wait
2116   for an indefinite period before sending the request body.
2119   Requirements for HTTP/1.1 origin servers:
2120  <list style="symbols">
2121    <t> Upon receiving a request which includes an Expect request-header
2122        field with the "100-continue" expectation, an origin server &MUST;
2123        either respond with 100 (Continue) status and continue to read
2124        from the input stream, or respond with a final status code. The
2125        origin server &MUST-NOT; wait for the request body before sending
2126        the 100 (Continue) response. If it responds with a final status
2127        code, it &MAY; close the transport connection or it &MAY; continue
2128        to read and discard the rest of the request.  It &MUST-NOT;
2129        perform the requested method if it returns a final status code.
2130    </t>
2131    <t> An origin server &SHOULD-NOT;  send a 100 (Continue) response if
2132        the request message does not include an Expect request-header
2133        field with the "100-continue" expectation, and &MUST-NOT; send a
2134        100 (Continue) response if such a request comes from an HTTP/1.0
2135        (or earlier) client. There is an exception to this rule: for
2136        compatibility with <xref target="RFC2068"/>, a server &MAY; send a 100 (Continue)
2137        status in response to an HTTP/1.1 PUT or POST request that does
2138        not include an Expect request-header field with the "100-continue"
2139        expectation. This exception, the purpose of which is
2140        to minimize any client processing delays associated with an
2141        undeclared wait for 100 (Continue) status, applies only to
2142        HTTP/1.1 requests, and not to requests with any other HTTP-version
2143        value.
2144    </t>
2145    <t> An origin server &MAY; omit a 100 (Continue) response if it has
2146        already received some or all of the request body for the
2147        corresponding request.
2148    </t>
2149    <t> An origin server that sends a 100 (Continue) response &MUST;
2150    ultimately send a final status code, once the request body is
2151        received and processed, unless it terminates the transport
2152        connection prematurely.
2153    </t>
2154    <t> If an origin server receives a request that does not include an
2155        Expect request-header field with the "100-continue" expectation,
2156        the request includes a request body, and the server responds
2157        with a final status code before reading the entire request body
2158        from the transport connection, then the server &SHOULD-NOT;  close
2159        the transport connection until it has read the entire request,
2160        or until the client closes the connection. Otherwise, the client
2161        might not reliably receive the response message. However, this
2162        requirement is not be construed as preventing a server from
2163        defending itself against denial-of-service attacks, or from
2164        badly broken client implementations.
2165      </t>
2166    </list>
2169   Requirements for HTTP/1.1 proxies:
2170  <list style="symbols">
2171    <t> If a proxy receives a request that includes an Expect request-header
2172        field with the "100-continue" expectation, and the proxy
2173        either knows that the next-hop server complies with HTTP/1.1 or
2174        higher, or does not know the HTTP version of the next-hop
2175        server, it &MUST; forward the request, including the Expect header
2176        field.
2177    </t>
2178    <t> If the proxy knows that the version of the next-hop server is
2179        HTTP/1.0 or lower, it &MUST-NOT; forward the request, and it &MUST;
2180        respond with a 417 (Expectation Failed) status.
2181    </t>
2182    <t> Proxies &SHOULD; maintain a cache recording the HTTP version
2183        numbers received from recently-referenced next-hop servers.
2184    </t>
2185    <t> A proxy &MUST-NOT; forward a 100 (Continue) response if the
2186        request message was received from an HTTP/1.0 (or earlier)
2187        client and did not include an Expect request-header field with
2188        the "100-continue" expectation. This requirement overrides the
2189        general rule for forwarding of 1xx responses (see &status-1xx;).
2190    </t>
2191  </list>
2195<section title="Client Behavior if Server Prematurely Closes Connection" anchor="connection.premature">
2197   If an HTTP/1.1 client sends a request which includes a request body,
2198   but which does not include an Expect request-header field with the
2199   "100-continue" expectation, and if the client is not directly
2200   connected to an HTTP/1.1 origin server, and if the client sees the
2201   connection close before receiving any status from the server, the
2202   client &SHOULD; retry the request.  If the client does retry this
2203   request, it &MAY; use the following "binary exponential backoff"
2204   algorithm to be assured of obtaining a reliable response:
2205  <list style="numbers">
2206    <t>
2207      Initiate a new connection to the server
2208    </t>
2209    <t>
2210      Transmit the request-headers
2211    </t>
2212    <t>
2213      Initialize a variable R to the estimated round-trip time to the
2214         server (e.g., based on the time it took to establish the
2215         connection), or to a constant value of 5 seconds if the round-trip
2216         time is not available.
2217    </t>
2218    <t>
2219       Compute T = R * (2**N), where N is the number of previous
2220         retries of this request.
2221    </t>
2222    <t>
2223       Wait either for an error response from the server, or for T
2224         seconds (whichever comes first)
2225    </t>
2226    <t>
2227       If no error response is received, after T seconds transmit the
2228         body of the request.
2229    </t>
2230    <t>
2231       If client sees that the connection is closed prematurely,
2232         repeat from step 1 until the request is accepted, an error
2233         response is received, or the user becomes impatient and
2234         terminates the retry process.
2235    </t>
2236  </list>
2239   If at any point an error status is received, the client
2240  <list style="symbols">
2241      <t>&SHOULD-NOT;  continue and</t>
2243      <t>&SHOULD; close the connection if it has not completed sending the
2244        request message.</t>
2245    </list>
2252<section title="Header Field Definitions" anchor="header.fields">
2254   This section defines the syntax and semantics of HTTP/1.1 header fields
2255   related to message framing and transport protocols.
2258   For entity-header fields, both sender and recipient refer to either the
2259   client or the server, depending on who sends and who receives the entity.
2262<section title="Connection" anchor="header.connection">
2263  <iref primary="true" item="Connection header" x:for-anchor=""/>
2264  <iref primary="true" item="Headers" subitem="Connection" x:for-anchor=""/>
2266   The Connection general-header field allows the sender to specify
2267   options that are desired for that particular connection and &MUST-NOT;
2268   be communicated by proxies over further connections.
2271   The Connection header has the following grammar:
2273<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-token"/>
2274  Connection = "Connection" ":" 1#(connection-token)
2275  connection-token  = token
2278   HTTP/1.1 proxies &MUST; parse the Connection header field before a
2279   message is forwarded and, for each connection-token in this field,
2280   remove any header field(s) from the message with the same name as the
2281   connection-token. Connection options are signaled by the presence of
2282   a connection-token in the Connection header field, not by any
2283   corresponding additional header field(s), since the additional header
2284   field may not be sent if there are no parameters associated with that
2285   connection option.
2288   Message headers listed in the Connection header &MUST-NOT; include
2289   end-to-end headers, such as Cache-Control.
2292   HTTP/1.1 defines the "close" connection option for the sender to
2293   signal that the connection will be closed after completion of the
2294   response. For example,
2296<figure><artwork type="example">
2297    Connection: close
2300   in either the request or the response header fields indicates that
2301   the connection &SHOULD-NOT;  be considered `persistent' (<xref target="persistent.connections"/>)
2302   after the current request/response is complete.
2305   An HTTP/1.1 client that does not support persistent connections &MUST;
2306   include the "close" connection option in every request message.
2309   An HTTP/1.1 server that does not support persistent connections &MUST;
2310   include the "close" connection option in every response message that
2311   does not have a 1xx (informational) status code.
2314   A system receiving an HTTP/1.0 (or lower-version) message that
2315   includes a Connection header &MUST;, for each connection-token in this
2316   field, remove and ignore any header field(s) from the message with
2317   the same name as the connection-token. This protects against mistaken
2318   forwarding of such header fields by pre-HTTP/1.1 proxies. See <xref target="compatibility.with.http.1.0.persistent.connections"/>.
2322<section title="Content-Length" anchor="header.content-length">
2323  <iref primary="true" item="Content-Length header" x:for-anchor=""/>
2324  <iref primary="true" item="Headers" subitem="Content-Length" x:for-anchor=""/>
2326   The Content-Length entity-header field indicates the size of the
2327   entity-body, in decimal number of OCTETs, sent to the recipient or,
2328   in the case of the HEAD method, the size of the entity-body that
2329   would have been sent had the request been a GET.
2331<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
2332  Content-Length    = "Content-Length" ":" 1*DIGIT
2335   An example is
2337<figure><artwork type="example">
2338    Content-Length: 3495
2341   Applications &SHOULD; use this field to indicate the transfer-length of
2342   the message-body, unless this is prohibited by the rules in <xref target="message.length"/>.
2345   Any Content-Length greater than or equal to zero is a valid value.
2346   <xref target="message.length"/> describes how to determine the length of a message-body
2347   if a Content-Length is not given.
2350   Note that the meaning of this field is significantly different from
2351   the corresponding definition in MIME, where it is an optional field
2352   used within the "message/external-body" content-type. In HTTP, it
2353   &SHOULD; be sent whenever the message's length can be determined prior
2354   to being transferred, unless this is prohibited by the rules in
2355   <xref target="message.length"/>.
2359<section title="Date" anchor="">
2360  <iref primary="true" item="Date header" x:for-anchor=""/>
2361  <iref primary="true" item="Headers" subitem="Date" x:for-anchor=""/>
2363   The Date general-header field represents the date and time at which
2364   the message was originated, having the same semantics as orig-date in
2365   <xref target="RFC2822" x:fmt="of" x:sec="3.6.1"/>. The field value is an HTTP-date, as described in <xref target=""/>;
2366   it &MUST; be sent in rfc1123-date format.
2368<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Date"/>
2369  Date  = "Date" ":" HTTP-date
2372   An example is
2374<figure><artwork type="example">
2375    Date: Tue, 15 Nov 1994 08:12:31 GMT
2378   Origin servers &MUST; include a Date header field in all responses,
2379   except in these cases:
2380  <list style="numbers">
2381      <t>If the response status code is 100 (Continue) or 101 (Switching
2382         Protocols), the response &MAY; include a Date header field, at
2383         the server's option.</t>
2385      <t>If the response status code conveys a server error, e.g. 500
2386         (Internal Server Error) or 503 (Service Unavailable), and it is
2387         inconvenient or impossible to generate a valid Date.</t>
2389      <t>If the server does not have a clock that can provide a
2390         reasonable approximation of the current time, its responses
2391         &MUST-NOT; include a Date header field. In this case, the rules
2392         in <xref target="clockless.origin.server.operation"/> &MUST; be followed.</t>
2393  </list>
2396   A received message that does not have a Date header field &MUST; be
2397   assigned one by the recipient if the message will be cached by that
2398   recipient or gatewayed via a protocol which requires a Date. An HTTP
2399   implementation without a clock &MUST-NOT; cache responses without
2400   revalidating them on every use. An HTTP cache, especially a shared
2401   cache, &SHOULD; use a mechanism, such as NTP <xref target="RFC1305"/>, to synchronize its
2402   clock with a reliable external standard.
2405   Clients &SHOULD; only send a Date header field in messages that include
2406   an entity-body, as in the case of the PUT and POST requests, and even
2407   then it is optional. A client without a clock &MUST-NOT; send a Date
2408   header field in a request.
2411   The HTTP-date sent in a Date header &SHOULD-NOT;  represent a date and
2412   time subsequent to the generation of the message. It &SHOULD; represent
2413   the best available approximation of the date and time of message
2414   generation, unless the implementation has no means of generating a
2415   reasonably accurate date and time. In theory, the date ought to
2416   represent the moment just before the entity is generated. In
2417   practice, the date can be generated at any time during the message
2418   origination without affecting its semantic value.
2421<section title="Clockless Origin Server Operation" anchor="clockless.origin.server.operation">
2423   Some origin server implementations might not have a clock available.
2424   An origin server without a clock &MUST-NOT; assign Expires or Last-Modified
2425   values to a response, unless these values were associated
2426   with the resource by a system or user with a reliable clock. It &MAY;
2427   assign an Expires value that is known, at or before server
2428   configuration time, to be in the past (this allows "pre-expiration"
2429   of responses without storing separate Expires values for each
2430   resource).
2435<section title="Host" anchor="">
2436  <iref primary="true" item="Host header" x:for-anchor=""/>
2437  <iref primary="true" item="Headers" subitem="Host" x:for-anchor=""/>
2439   The Host request-header field specifies the Internet host and port
2440   number of the resource being requested, as obtained from the original
2441   URI given by the user or referring resource (generally an HTTP URL,
2442   as described in <xref target="http.url"/>). The Host field value &MUST; represent
2443   the naming authority of the origin server or gateway given by the
2444   original URL. This allows the origin server or gateway to
2445   differentiate between internally-ambiguous URLs, such as the root "/"
2446   URL of a server for multiple host names on a single IP address.
2448<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2449  Host = "Host" ":" uri-host [ ":" port ] ; <xref target="http.url"/>
2452   A "host" without any trailing port information implies the default
2453   port for the service requested (e.g., "80" for an HTTP URL). For
2454   example, a request on the origin server for
2455   &lt;; would properly include:
2457<figure><artwork type="example">
2458    GET /pub/WWW/ HTTP/1.1
2459    Host:
2462   A client &MUST; include a Host header field in all HTTP/1.1 request
2463   messages. If the requested URI does not include an Internet host
2464   name for the service being requested, then the Host header field &MUST;
2465   be given with an empty value. An HTTP/1.1 proxy &MUST; ensure that any
2466   request message it forwards does contain an appropriate Host header
2467   field that identifies the service being requested by the proxy. All
2468   Internet-based HTTP/1.1 servers &MUST; respond with a 400 (Bad Request)
2469   status code to any HTTP/1.1 request message which lacks a Host header
2470   field.
2473   See Sections <xref target="" format="counter"/>
2474   and <xref target="" format="counter"/>
2475   for other requirements relating to Host.
2479<section title="TE" anchor="header.te">
2480  <iref primary="true" item="TE header" x:for-anchor=""/>
2481  <iref primary="true" item="Headers" subitem="TE" x:for-anchor=""/>
2483   The TE request-header field indicates what extension transfer-codings
2484   it is willing to accept in the response and whether or not it is
2485   willing to accept trailer fields in a chunked transfer-coding. Its
2486   value may consist of the keyword "trailers" and/or a comma-separated
2487   list of extension transfer-coding names with optional accept
2488   parameters (as described in <xref target="transfer.codings"/>).
2490<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TE"/><iref primary="true" item="Grammar" subitem="t-codings"/>
2491  TE        = "TE" ":" #( t-codings )
2492  t-codings = "trailers" | ( transfer-extension [ accept-params ] )
2495   The presence of the keyword "trailers" indicates that the client is
2496   willing to accept trailer fields in a chunked transfer-coding, as
2497   defined in <xref target="chunked.transfer.encoding"/>. This keyword is reserved for use with
2498   transfer-coding values even though it does not itself represent a
2499   transfer-coding.
2502   Examples of its use are:
2504<figure><artwork type="example">
2505    TE: deflate
2506    TE:
2507    TE: trailers, deflate;q=0.5
2510   The TE header field only applies to the immediate connection.
2511   Therefore, the keyword &MUST; be supplied within a Connection header
2512   field (<xref target="header.connection"/>) whenever TE is present in an HTTP/1.1 message.
2515   A server tests whether a transfer-coding is acceptable, according to
2516   a TE field, using these rules:
2517  <list style="numbers">
2518    <x:lt>
2519      <t>The "chunked" transfer-coding is always acceptable. If the
2520         keyword "trailers" is listed, the client indicates that it is
2521         willing to accept trailer fields in the chunked response on
2522         behalf of itself and any downstream clients. The implication is
2523         that, if given, the client is stating that either all
2524         downstream clients are willing to accept trailer fields in the
2525         forwarded response, or that it will attempt to buffer the
2526         response on behalf of downstream recipients.
2527      </t><t>
2528         <x:h>Note:</x:h> HTTP/1.1 does not define any means to limit the size of a
2529         chunked response such that a client can be assured of buffering
2530         the entire response.</t>
2531    </x:lt>
2532    <x:lt>
2533      <t>If the transfer-coding being tested is one of the transfer-codings
2534         listed in the TE field, then it is acceptable unless it
2535         is accompanied by a qvalue of 0. (As defined in &qvalue;, a
2536         qvalue of 0 means "not acceptable.")</t>
2537    </x:lt>
2538    <x:lt>
2539      <t>If multiple transfer-codings are acceptable, then the
2540         acceptable transfer-coding with the highest non-zero qvalue is
2541         preferred.  The "chunked" transfer-coding always has a qvalue
2542         of 1.</t>
2543    </x:lt>
2544  </list>
2547   If the TE field-value is empty or if no TE field is present, the only
2548   transfer-coding  is "chunked". A message with no transfer-coding is
2549   always acceptable.
2553<section title="Trailer" anchor="header.trailer">
2554  <iref primary="true" item="Trailer header" x:for-anchor=""/>
2555  <iref primary="true" item="Headers" subitem="Trailer" x:for-anchor=""/>
2557   The Trailer general field value indicates that the given set of
2558   header fields is present in the trailer of a message encoded with
2559   chunked transfer-coding.
2561<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
2562  Trailer  = "Trailer" ":" 1#field-name
2565   An HTTP/1.1 message &SHOULD; include a Trailer header field in a
2566   message using chunked transfer-coding with a non-empty trailer. Doing
2567   so allows the recipient to know which header fields to expect in the
2568   trailer.
2571   If no Trailer header field is present, the trailer &SHOULD-NOT;  include
2572   any header fields. See <xref target="chunked.transfer.encoding"/> for restrictions on the use of
2573   trailer fields in a "chunked" transfer-coding.
2576   Message header fields listed in the Trailer header field &MUST-NOT;
2577   include the following header fields:
2578  <list style="symbols">
2579    <t>Transfer-Encoding</t>
2580    <t>Content-Length</t>
2581    <t>Trailer</t>
2582  </list>
2586<section title="Transfer-Encoding" anchor="header.transfer-encoding">
2587  <iref primary="true" item="Transfer-Encoding header" x:for-anchor=""/>
2588  <iref primary="true" item="Headers" subitem="Transfer-Encoding" x:for-anchor=""/>
2590   The Transfer-Encoding general-header field indicates what (if any)
2591   type of transformation has been applied to the message body in order
2592   to safely transfer it between the sender and the recipient. This
2593   differs from the content-coding in that the transfer-coding is a
2594   property of the message, not of the entity.
2596<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
2597  Transfer-Encoding       = "Transfer-Encoding" ":" 1#transfer-coding
2600   Transfer-codings are defined in <xref target="transfer.codings"/>. An example is:
2602<figure><artwork type="example">
2603  Transfer-Encoding: chunked
2606   If multiple encodings have been applied to an entity, the transfer-codings
2607   &MUST; be listed in the order in which they were applied.
2608   Additional information about the encoding parameters &MAY; be provided
2609   by other entity-header fields not defined by this specification.
2612   Many older HTTP/1.0 applications do not understand the Transfer-Encoding
2613   header.
2617<section title="Upgrade" anchor="header.upgrade">
2618  <iref primary="true" item="Upgrade header" x:for-anchor=""/>
2619  <iref primary="true" item="Headers" subitem="Upgrade" x:for-anchor=""/>
2621   The Upgrade general-header allows the client to specify what
2622   additional communication protocols it supports and would like to use
2623   if the server finds it appropriate to switch protocols. The server
2624   &MUST; use the Upgrade header field within a 101 (Switching Protocols)
2625   response to indicate which protocol(s) are being switched.
2627<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
2628  Upgrade        = "Upgrade" ":" 1#product
2631   For example,
2633<figure><artwork type="example">
2634    Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
2637   The Upgrade header field is intended to provide a simple mechanism
2638   for transition from HTTP/1.1 to some other, incompatible protocol. It
2639   does so by allowing the client to advertise its desire to use another
2640   protocol, such as a later version of HTTP with a higher major version
2641   number, even though the current request has been made using HTTP/1.1.
2642   This eases the difficult transition between incompatible protocols by
2643   allowing the client to initiate a request in the more commonly
2644   supported protocol while indicating to the server that it would like
2645   to use a "better" protocol if available (where "better" is determined
2646   by the server, possibly according to the nature of the method and/or
2647   resource being requested).
2650   The Upgrade header field only applies to switching application-layer
2651   protocols upon the existing transport-layer connection. Upgrade
2652   cannot be used to insist on a protocol change; its acceptance and use
2653   by the server is optional. The capabilities and nature of the
2654   application-layer communication after the protocol change is entirely
2655   dependent upon the new protocol chosen, although the first action
2656   after changing the protocol &MUST; be a response to the initial HTTP
2657   request containing the Upgrade header field.
2660   The Upgrade header field only applies to the immediate connection.
2661   Therefore, the upgrade keyword &MUST; be supplied within a Connection
2662   header field (<xref target="header.connection"/>) whenever Upgrade is present in an
2663   HTTP/1.1 message.
2666   The Upgrade header field cannot be used to indicate a switch to a
2667   protocol on a different connection. For that purpose, it is more
2668   appropriate to use a 301, 302, 303, or 305 redirection response.
2671   This specification only defines the protocol name "HTTP" for use by
2672   the family of Hypertext Transfer Protocols, as defined by the HTTP
2673   version rules of <xref target="http.version"/> and future updates to this
2674   specification. Any token can be used as a protocol name; however, it
2675   will only be useful if both the client and server associate the name
2676   with the same protocol.
2680<section title="Via" anchor="header.via">
2681  <iref primary="true" item="Via header" x:for-anchor=""/>
2682  <iref primary="true" item="Headers" subitem="Via" x:for-anchor=""/>
2684   The Via general-header field &MUST; be used by gateways and proxies to
2685   indicate the intermediate protocols and recipients between the user
2686   agent and the server on requests, and between the origin server and
2687   the client on responses. It is analogous to the "Received" field of
2688   <xref target="RFC2822"/> and is intended to be used for tracking message forwards,
2689   avoiding request loops, and identifying the protocol capabilities of
2690   all senders along the request/response chain.
2692<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"/>
2693  Via =  "Via" ":" 1#( received-protocol received-by [ comment ] )
2694  received-protocol = [ protocol-name "/" ] protocol-version
2695  protocol-name     = token
2696  protocol-version  = token
2697  received-by       = ( uri-host [ ":" port ] ) | pseudonym
2698  pseudonym         = token
2701   The received-protocol indicates the protocol version of the message
2702   received by the server or client along each segment of the
2703   request/response chain. The received-protocol version is appended to
2704   the Via field value when the message is forwarded so that information
2705   about the protocol capabilities of upstream applications remains
2706   visible to all recipients.
2709   The protocol-name is optional if and only if it would be "HTTP". The
2710   received-by field is normally the host and optional port number of a
2711   recipient server or client that subsequently forwarded the message.
2712   However, if the real host is considered to be sensitive information,
2713   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
2714   be assumed to be the default port of the received-protocol.
2717   Multiple Via field values represents each proxy or gateway that has
2718   forwarded the message. Each recipient &MUST; append its information
2719   such that the end result is ordered according to the sequence of
2720   forwarding applications.
2723   Comments &MAY; be used in the Via header field to identify the software
2724   of the recipient proxy or gateway, analogous to the User-Agent and
2725   Server header fields. However, all comments in the Via field are
2726   optional and &MAY; be removed by any recipient prior to forwarding the
2727   message.
2730   For example, a request message could be sent from an HTTP/1.0 user
2731   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2732   forward the request to a public proxy at, which completes
2733   the request by forwarding it to the origin server at
2734   The request received by would then have the following
2735   Via header field:
2737<figure><artwork type="example">
2738    Via: 1.0 fred, 1.1 (Apache/1.1)
2741   Proxies and gateways used as a portal through a network firewall
2742   &SHOULD-NOT;, by default, forward the names and ports of hosts within
2743   the firewall region. This information &SHOULD; only be propagated if
2744   explicitly enabled. If not enabled, the received-by host of any host
2745   behind the firewall &SHOULD; be replaced by an appropriate pseudonym
2746   for that host.
2749   For organizations that have strong privacy requirements for hiding
2750   internal structures, a proxy &MAY; combine an ordered subsequence of
2751   Via header field entries with identical received-protocol values into
2752   a single such entry. For example,
2754<figure><artwork type="example">
2755    Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2758        could be collapsed to
2760<figure><artwork type="example">
2761    Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2764   Applications &SHOULD-NOT;  combine multiple entries unless they are all
2765   under the same organizational control and the hosts have already been
2766   replaced by pseudonyms. Applications &MUST-NOT; combine entries which
2767   have different received-protocol values.
2773<section title="IANA Considerations" anchor="IANA.considerations">
2775   <cref>TBD.</cref>
2779<section title="Security Considerations" anchor="security.considerations">
2781   This section is meant to inform application developers, information
2782   providers, and users of the security limitations in HTTP/1.1 as
2783   described by this document. The discussion does not include
2784   definitive solutions to the problems revealed, though it does make
2785   some suggestions for reducing security risks.
2788<section title="Personal Information" anchor="personal.information">
2790   HTTP clients are often privy to large amounts of personal information
2791   (e.g. the user's name, location, mail address, passwords, encryption
2792   keys, etc.), and &SHOULD; be very careful to prevent unintentional
2793   leakage of this information.
2794   We very strongly recommend that a convenient interface be provided
2795   for the user to control dissemination of such information, and that
2796   designers and implementors be particularly careful in this area.
2797   History shows that errors in this area often create serious security
2798   and/or privacy problems and generate highly adverse publicity for the
2799   implementor's company.
2803<section title="Abuse of Server Log Information" anchor="abuse.of.server.log.information">
2805   A server is in the position to save personal data about a user's
2806   requests which might identify their reading patterns or subjects of
2807   interest. This information is clearly confidential in nature and its
2808   handling can be constrained by law in certain countries. People using
2809   HTTP to provide data are responsible for ensuring that
2810   such material is not distributed without the permission of any
2811   individuals that are identifiable by the published results.
2815<section title="Attacks Based On File and Path Names" anchor="attack.pathname">
2817   Implementations of HTTP origin servers &SHOULD; be careful to restrict
2818   the documents returned by HTTP requests to be only those that were
2819   intended by the server administrators. If an HTTP server translates
2820   HTTP URIs directly into file system calls, the server &MUST; take
2821   special care not to serve files that were not intended to be
2822   delivered to HTTP clients. For example, UNIX, Microsoft Windows, and
2823   other operating systems use ".." as a path component to indicate a
2824   directory level above the current one. On such a system, an HTTP
2825   server &MUST; disallow any such construct in the Request-URI if it
2826   would otherwise allow access to a resource outside those intended to
2827   be accessible via the HTTP server. Similarly, files intended for
2828   reference only internally to the server (such as access control
2829   files, configuration files, and script code) &MUST; be protected from
2830   inappropriate retrieval, since they might contain sensitive
2831   information. Experience has shown that minor bugs in such HTTP server
2832   implementations have turned into security risks.
2836<section title="DNS Spoofing" anchor="dns.spoofing">
2838   Clients using HTTP rely heavily on the Domain Name Service, and are
2839   thus generally prone to security attacks based on the deliberate
2840   mis-association of IP addresses and DNS names. Clients need to be
2841   cautious in assuming the continuing validity of an IP number/DNS name
2842   association.
2845   In particular, HTTP clients &SHOULD; rely on their name resolver for
2846   confirmation of an IP number/DNS name association, rather than
2847   caching the result of previous host name lookups. Many platforms
2848   already can cache host name lookups locally when appropriate, and
2849   they &SHOULD; be configured to do so. It is proper for these lookups to
2850   be cached, however, only when the TTL (Time To Live) information
2851   reported by the name server makes it likely that the cached
2852   information will remain useful.
2855   If HTTP clients cache the results of host name lookups in order to
2856   achieve a performance improvement, they &MUST; observe the TTL
2857   information reported by DNS.
2860   If HTTP clients do not observe this rule, they could be spoofed when
2861   a previously-accessed server's IP address changes. As network
2862   renumbering is expected to become increasingly common <xref target="RFC1900"/>, the
2863   possibility of this form of attack will grow. Observing this
2864   requirement thus reduces this potential security vulnerability.
2867   This requirement also improves the load-balancing behavior of clients
2868   for replicated servers using the same DNS name and reduces the
2869   likelihood of a user's experiencing failure in accessing sites which
2870   use that strategy.
2874<section title="Proxies and Caching" anchor="attack.proxies">
2876   By their very nature, HTTP proxies are men-in-the-middle, and
2877   represent an opportunity for man-in-the-middle attacks. Compromise of
2878   the systems on which the proxies run can result in serious security
2879   and privacy problems. Proxies have access to security-related
2880   information, personal information about individual users and
2881   organizations, and proprietary information belonging to users and
2882   content providers. A compromised proxy, or a proxy implemented or
2883   configured without regard to security and privacy considerations,
2884   might be used in the commission of a wide range of potential attacks.
2887   Proxy operators should protect the systems on which proxies run as
2888   they would protect any system that contains or transports sensitive
2889   information. In particular, log information gathered at proxies often
2890   contains highly sensitive personal information, and/or information
2891   about organizations. Log information should be carefully guarded, and
2892   appropriate guidelines for use developed and followed. (<xref target="abuse.of.server.log.information"/>).
2895   Proxy implementors should consider the privacy and security
2896   implications of their design and coding decisions, and of the
2897   configuration options they provide to proxy operators (especially the
2898   default configuration).
2901   Users of a proxy need to be aware that they are no trustworthier than
2902   the people who run the proxy; HTTP itself cannot solve this problem.
2905   The judicious use of cryptography, when appropriate, may suffice to
2906   protect against a broad range of security and privacy attacks. Such
2907   cryptography is beyond the scope of the HTTP/1.1 specification.
2911<section title="Denial of Service Attacks on Proxies" anchor="attack.DoS">
2913   They exist. They are hard to defend against. Research continues.
2914   Beware.
2919<section title="Acknowledgments" anchor="ack">
2921   This specification makes heavy use of the augmented BNF and generic
2922   constructs defined by David H. Crocker for <xref target="RFC822ABNF"/>. Similarly, it
2923   reuses many of the definitions provided by Nathaniel Borenstein and
2924   Ned Freed for MIME <xref target="RFC2045"/>. We hope that their inclusion in this
2925   specification will help reduce past confusion over the relationship
2926   between HTTP and Internet mail message formats.
2929   HTTP has evolved considerably over the years. It has
2930   benefited from a large and active developer community--the many
2931   people who have participated on the www-talk mailing list--and it is
2932   that community which has been most responsible for the success of
2933   HTTP and of the World-Wide Web in general. Marc Andreessen, Robert
2934   Cailliau, Daniel W. Connolly, Bob Denny, John Franks, Jean-Francois
2935   Groff, Phillip M. Hallam-Baker, Hakon W. Lie, Ari Luotonen, Rob
2936   McCool, Lou Montulli, Dave Raggett, Tony Sanders, and Marc
2937   VanHeyningen deserve special recognition for their efforts in
2938   defining early aspects of the protocol.
2941   This document has benefited greatly from the comments of all those
2942   participating in the HTTP-WG. In addition to those already mentioned,
2943   the following individuals have contributed to this specification:
2946   Gary Adams, Harald Tveit Alvestrand, Keith Ball, Brian Behlendorf,
2947   Paul Burchard, Maurizio Codogno, Mike Cowlishaw, Roman Czyborra,
2948   Michael A. Dolan, Daniel DuBois, David J. Fiander, Alan Freier, Marc Hedlund, Greg Herlihy,
2949   Koen Holtman, Alex Hopmann, Bob Jernigan, Shel Kaphan, Rohit Khare,
2950   John Klensin, Martijn Koster, Alexei Kosut, David M. Kristol,
2951   Daniel LaLiberte, Ben Laurie, Paul J. Leach, Albert Lunde,
2952   John C. Mallery, Jean-Philippe Martin-Flatin, Mitra, David Morris,
2953   Gavin Nicol, Ross Patterson, Bill Perry, Jeffrey Perry, Scott Powers, Owen Rees,
2954   Luigi Rizzo, David Robinson, Marc Salomon, Rich Salz,
2955   Allan M. Schiffman, Jim Seidman, Chuck Shotton, Eric W. Sink,
2956   Simon E. Spero, Richard N. Taylor, Robert S. Thau,
2957   Bill (BearHeart) Weinman, Francois Yergeau, Mary Ellen Zurko,
2958   Josh Cohen.
2961   Thanks to the "cave men" of Palo Alto. You know who you are.
2964   Jim Gettys (the editor of <xref target="RFC2616"/>) wishes particularly
2965   to thank Roy Fielding, the editor of <xref target="RFC2068"/>, along
2966   with John Klensin, Jeff Mogul, Paul Leach, Dave Kristol, Koen
2967   Holtman, John Franks, Josh Cohen, Alex Hopmann, Scott Lawrence, and
2968   Larry Masinter for their help. And thanks go particularly to Jeff
2969   Mogul and Scott Lawrence for performing the "MUST/MAY/SHOULD" audit.
2972   The Apache Group, Anselm Baird-Smith, author of Jigsaw, and Henrik
2973   Frystyk implemented RFC 2068 early, and we wish to thank them for the
2974   discovery of many of the problems that this document attempts to
2975   rectify.
2982<references title="Normative References">
2984<reference anchor="ISO-8859-1">
2985  <front>
2986    <title>
2987     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
2988    </title>
2989    <author>
2990      <organization>International Organization for Standardization</organization>
2991    </author>
2992    <date year="1998"/>
2993  </front>
2994  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
2997<reference anchor="Part2">
2998  <front>
2999    <title abbrev="HTTP/1.1">HTTP/1.1, part 2: Message Semantics</title>
3000    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3001      <organization abbrev="Day Software">Day Software</organization>
3002      <address><email></email></address>
3003    </author>
3004    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3005      <organization>One Laptop per Child</organization>
3006      <address><email></email></address>
3007    </author>
3008    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3009      <organization abbrev="HP">Hewlett-Packard Company</organization>
3010      <address><email></email></address>
3011    </author>
3012    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3013      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3014      <address><email></email></address>
3015    </author>
3016    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3017      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3018      <address><email></email></address>
3019    </author>
3020    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3021      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3022      <address><email></email></address>
3023    </author>
3024    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3025      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3026      <address><email></email></address>
3027    </author>
3028    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3029      <organization abbrev="W3C">World Wide Web Consortium</organization>
3030      <address><email></email></address>
3031    </author>
3032    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3033      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3034      <address><email></email></address>
3035    </author>
3036    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3037  </front>
3038  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-&ID-VERSION;"/>
3039  <x:source href="p2-semantics.xml" basename="p2-semantics"/>
3042<reference anchor="Part3">
3043  <front>
3044    <title abbrev="HTTP/1.1">HTTP/1.1, part 3: Message Payload and Content Negotiation</title>
3045    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3046      <organization abbrev="Day Software">Day Software</organization>
3047      <address><email></email></address>
3048    </author>
3049    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3050      <organization>One Laptop per Child</organization>
3051      <address><email></email></address>
3052    </author>
3053    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3054      <organization abbrev="HP">Hewlett-Packard Company</organization>
3055      <address><email></email></address>
3056    </author>
3057    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3058      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3059      <address><email></email></address>
3060    </author>
3061    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3062      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3063      <address><email></email></address>
3064    </author>
3065    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3066      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3067      <address><email></email></address>
3068    </author>
3069    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3070      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3071      <address><email></email></address>
3072    </author>
3073    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3074      <organization abbrev="W3C">World Wide Web Consortium</organization>
3075      <address><email></email></address>
3076    </author>
3077    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3078      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3079      <address><email></email></address>
3080    </author>
3081    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3082  </front>
3083  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p3-payload-&ID-VERSION;"/>
3084  <x:source href="p3-payload.xml" basename="p3-payload"/>
3087<reference anchor="Part5">
3088  <front>
3089    <title abbrev="HTTP/1.1">HTTP/1.1, part 5: Range Requests and Partial Responses</title>
3090    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3091      <organization abbrev="Day Software">Day Software</organization>
3092      <address><email></email></address>
3093    </author>
3094    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3095      <organization>One Laptop per Child</organization>
3096      <address><email></email></address>
3097    </author>
3098    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3099      <organization abbrev="HP">Hewlett-Packard Company</organization>
3100      <address><email></email></address>
3101    </author>
3102    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3103      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3104      <address><email></email></address>
3105    </author>
3106    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3107      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3108      <address><email></email></address>
3109    </author>
3110    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3111      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3112      <address><email></email></address>
3113    </author>
3114    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3115      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3116      <address><email></email></address>
3117    </author>
3118    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3119      <organization abbrev="W3C">World Wide Web Consortium</organization>
3120      <address><email></email></address>
3121    </author>
3122    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3123      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3124      <address><email></email></address>
3125    </author>
3126    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3127  </front>
3128  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-&ID-VERSION;"/>
3129  <x:source href="p5-range.xml" basename="p5-range"/>
3132<reference anchor="Part6">
3133  <front>
3134    <title abbrev="HTTP/1.1">HTTP/1.1, part 6: Caching</title>
3135    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3136      <organization abbrev="Day Software">Day Software</organization>
3137      <address><email></email></address>
3138    </author>
3139    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3140      <organization>One Laptop per Child</organization>
3141      <address><email></email></address>
3142    </author>
3143    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3144      <organization abbrev="HP">Hewlett-Packard Company</organization>
3145      <address><email></email></address>
3146    </author>
3147    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3148      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3149      <address><email></email></address>
3150    </author>
3151    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3152      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3153      <address><email></email></address>
3154    </author>
3155    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3156      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3157      <address><email></email></address>
3158    </author>
3159    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3160      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3161      <address><email></email></address>
3162    </author>
3163    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3164      <organization abbrev="W3C">World Wide Web Consortium</organization>
3165      <address><email></email></address>
3166    </author>
3167    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3168      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3169      <address><email></email></address>
3170    </author>
3171    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3172  </front>
3173  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-&ID-VERSION;"/>
3174  <x:source href="p6-cache.xml" basename="p6-cache"/>
3177<reference anchor="RFC822ABNF">
3178  <front>
3179    <title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3180    <author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3181      <organization>University of Delaware, Dept. of Electrical Engineering</organization>
3182      <address><email>DCrocker@UDel-Relay</email></address>
3183    </author>
3184    <date month="August" day="13" year="1982"/>
3185  </front>
3186  <seriesInfo name="STD" value="11"/>
3187  <seriesInfo name="RFC" value="822"/>
3190<reference anchor="RFC2045">
3191  <front>
3192    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
3193    <author initials="N." surname="Freed" fullname="Ned Freed">
3194      <organization>Innosoft International, Inc.</organization>
3195      <address><email></email></address>
3196    </author>
3197    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
3198      <organization>First Virtual Holdings</organization>
3199      <address><email></email></address>
3200    </author>
3201    <date month="November" year="1996"/>
3202  </front>
3203  <seriesInfo name="RFC" value="2045"/>
3206<reference anchor="RFC2047">
3207  <front>
3208    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
3209    <author initials="K." surname="Moore" fullname="Keith Moore">
3210      <organization>University of Tennessee</organization>
3211      <address><email></email></address>
3212    </author>
3213    <date month="November" year="1996"/>
3214  </front>
3215  <seriesInfo name="RFC" value="2047"/>
3218<reference anchor="RFC2119">
3219  <front>
3220    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
3221    <author initials="S." surname="Bradner" fullname="Scott Bradner">
3222      <organization>Harvard University</organization>
3223      <address><email></email></address>
3224    </author>
3225    <date month="March" year="1997"/>
3226  </front>
3227  <seriesInfo name="BCP" value="14"/>
3228  <seriesInfo name="RFC" value="2119"/>
3231<reference anchor="RFC2396">
3232  <front>
3233    <title abbrev="URI Generic Syntax">Uniform Resource Identifiers (URI): Generic Syntax</title>
3234    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3235      <organization abbrev="MIT/LCS">World Wide Web Consortium</organization>
3236      <address><email></email></address>
3237    </author>
3238    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3239      <organization abbrev="U.C. Irvine">Department of Information and Computer Science</organization>
3240      <address><email></email></address>
3241    </author>
3242    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3243      <organization abbrev="Xerox Corporation">Xerox PARC</organization>
3244      <address><email></email></address>
3245    </author>
3246    <date month="August" year="1998"/>
3247  </front>
3248  <seriesInfo name="RFC" value="2396"/>
3251<reference anchor="USASCII">
3252  <front>
3253    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
3254    <author>
3255      <organization>American National Standards Institute</organization>
3256    </author>
3257    <date year="1986"/>
3258  </front>
3259  <seriesInfo name="ANSI" value="X3.4"/>
3264<references title="Informative References">
3266<reference anchor="Nie1997" target="">
3267  <front>
3268    <title>Network Performance Effects of HTTP/1.1, CSS1, and PNG</title>
3269    <author initials="H.F.." surname="Nielsen" fullname="H.F. Nielsen">
3270      <organization/>
3271    </author>
3272    <author initials="J." surname="Gettys" fullname="J. Gettys">
3273      <organization/>
3274    </author>
3275    <author initials="E." surname="Prud'hommeaux" fullname="E. Prud'hommeaux">
3276      <organization/>
3277    </author>
3278    <author initials="H." surname="Lie" fullname="H. Lie">
3279      <organization/>
3280    </author>
3281    <author initials="C." surname="Lilley" fullname="C. Lilley">
3282      <organization/>
3283    </author>
3284    <date year="1997" month="September"/>
3285  </front>
3286  <seriesInfo name="ACM" value="Proceedings of the ACM SIGCOMM '97 conference on Applications, technologies, architectures, and protocols for computer communication SIGCOMM '97"/>
3289<reference anchor="Pad1995">
3290  <front>
3291    <title>Improving HTTP Latency</title>
3292    <author initials="V.N." surname="Padmanabhan" fullname="Venkata N. Padmanabhan">
3293      <organization/>
3294    </author>
3295    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3296      <organization/>
3297    </author>
3298    <date year="1995" month="December"/>
3299  </front>
3300  <seriesInfo name="Computer Networks and ISDN Systems" value="v. 28, pp. 25-35"/>
3301  <annotation>
3302    Slightly revised version of paper in Proc. 2nd International WWW Conference '94: Mosaic and the Web, Oct. 1994,
3303    which is available at <eref target=""/>.
3304  </annotation>
3307<reference anchor="RFC822">
3308  <front>
3309    <title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3310    <author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3311      <organization>University of Delaware, Dept. of Electrical Engineering</organization>
3312      <address><email>DCrocker@UDel-Relay</email></address>
3313    </author>
3314    <date month="August" day="13" year="1982"/>
3315  </front>
3316  <seriesInfo name="STD" value="11"/>
3317  <seriesInfo name="RFC" value="822"/>
3320<reference anchor="RFC959">
3321  <front>
3322    <title abbrev="File Transfer Protocol">File Transfer Protocol</title>
3323    <author initials="J." surname="Postel" fullname="J. Postel">
3324      <organization>Information Sciences Institute (ISI)</organization>
3325    </author>
3326    <author initials="J." surname="Reynolds" fullname="J. Reynolds">
3327      <organization/>
3328    </author>
3329    <date month="October" year="1985"/>
3330  </front>
3331  <seriesInfo name="STD" value="9"/>
3332  <seriesInfo name="RFC" value="959"/>
3335<reference anchor="RFC1123">
3336  <front>
3337    <title>Requirements for Internet Hosts - Application and Support</title>
3338    <author initials="R." surname="Braden" fullname="Robert Braden">
3339      <organization>University of Southern California (USC), Information Sciences Institute</organization>
3340      <address><email>Braden@ISI.EDU</email></address>
3341    </author>
3342    <date month="October" year="1989"/>
3343  </front>
3344  <seriesInfo name="STD" value="3"/>
3345  <seriesInfo name="RFC" value="1123"/>
3348<reference anchor="RFC1305">
3349  <front>
3350    <title>Network Time Protocol (Version 3) Specification, Implementation</title>
3351    <author initials="D." surname="Mills" fullname="David L. Mills">
3352      <organization>University of Delaware, Electrical Engineering Department</organization>
3353      <address><email></email></address>
3354    </author>
3355    <date month="March" year="1992"/>
3356  </front>
3357  <seriesInfo name="RFC" value="1305"/>
3360<reference anchor="RFC1436">
3361  <front>
3362    <title abbrev="Gopher">The Internet Gopher Protocol (a distributed document search and retrieval protocol)</title>
3363    <author initials="F." surname="Anklesaria" fullname="Farhad Anklesaria">
3364      <organization>University of Minnesota, Computer and Information Services</organization>
3365      <address><email></email></address>
3366    </author>
3367    <author initials="M." surname="McCahill" fullname="Mark McCahill">
3368      <organization>University of Minnesota, Computer and Information Services</organization>
3369      <address><email></email></address>
3370    </author>
3371    <author initials="P." surname="Lindner" fullname="Paul Lindner">
3372      <organization>University of Minnesota, Computer and Information Services</organization>
3373      <address><email></email></address>
3374    </author>
3375    <author initials="D." surname="Johnson" fullname="David Johnson">
3376      <organization>University of Minnesota, Computer and Information Services</organization>
3377      <address><email></email></address>
3378    </author>
3379    <author initials="D." surname="Torrey" fullname="Daniel Torrey">
3380      <organization>University of Minnesota, Computer and Information Services</organization>
3381      <address><email></email></address>
3382    </author>
3383    <author initials="B." surname="Alberti" fullname="Bob Alberti">
3384      <organization>University of Minnesota, Computer and Information Services</organization>
3385      <address><email></email></address>
3386    </author>
3387    <date month="March" year="1993"/>
3388  </front>
3389  <seriesInfo name="RFC" value="1436"/>
3392<reference anchor="RFC1630">
3393  <front>
3394    <title abbrev="URIs in WWW">Universal Resource Identifiers in WWW: A Unifying Syntax for the Expression of Names and Addresses of Objects on the Network as used in the World-Wide Web</title>
3395    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3396      <organization>CERN, World-Wide Web project</organization>
3397      <address><email></email></address>
3398    </author>
3399    <date month="June" year="1994"/>
3400  </front>
3401  <seriesInfo name="RFC" value="1630"/>
3404<reference anchor="RFC1737">
3405  <front>
3406    <title abbrev="Requirements for Uniform Resource Names">Functional Requirements for Uniform Resource Names</title>
3407    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3408      <organization>Xerox Palo Alto Research Center</organization>
3409      <address><email></email></address>
3410    </author>
3411    <author initials="K." surname="Sollins" fullname="Karen Sollins">
3412      <organization>MIT Laboratory for Computer Science</organization>
3413      <address><email></email></address>
3414    </author>
3415    <date month="December" year="1994"/>
3416  </front>
3417  <seriesInfo name="RFC" value="1737"/>
3420<reference anchor="RFC1738">
3421  <front>
3422    <title>Uniform Resource Locators (URL)</title>
3423    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3424      <organization>CERN, World-Wide Web project</organization>
3425      <address><email></email></address>
3426    </author>
3427    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3428      <organization>Xerox PARC</organization>
3429      <address><email></email></address>
3430    </author>
3431    <author initials="M." surname="McCahill" fullname="Mark McCahill">
3432      <organization>University of Minnesota, Computer and Information Services</organization>
3433      <address><email></email></address>
3434    </author>
3435    <date month="December" year="1994"/>
3436  </front>
3437  <seriesInfo name="RFC" value="1738"/>
3440<reference anchor="RFC1808">
3441  <front>
3442    <title>Relative Uniform Resource Locators</title>
3443    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3444      <organization>University of California Irvine, Department of Information and Computer Science</organization>
3445      <address><email></email></address>
3446    </author>
3447    <date month="June" year="1995"/>
3448  </front>
3449  <seriesInfo name="RFC" value="1808"/>
3452<reference anchor="RFC1900">
3453  <front>
3454    <title>Renumbering Needs Work</title>
3455    <author initials="B." surname="Carpenter" fullname="Brian E. Carpenter">
3456      <organization>CERN, Computing and Networks Division</organization>
3457      <address><email></email></address>
3458    </author>
3459    <author initials="Y." surname="Rekhter" fullname="Yakov Rekhter">
3460      <organization>cisco Systems</organization>
3461      <address><email></email></address>
3462    </author>
3463    <date month="February" year="1996"/>
3464  </front>
3465  <seriesInfo name="RFC" value="1900"/>
3468<reference anchor="RFC1945">
3469  <front>
3470    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
3471    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3472      <organization>MIT, Laboratory for Computer Science</organization>
3473      <address><email></email></address>
3474    </author>
3475    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3476      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
3477      <address><email></email></address>
3478    </author>
3479    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3480      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
3481      <address><email></email></address>
3482    </author>
3483    <date month="May" year="1996"/>
3484  </front>
3485  <seriesInfo name="RFC" value="1945"/>
3488<reference anchor="RFC2068">
3489  <front>
3490    <title abbrev="HTTP/1.1">Hypertext Transfer Protocol -- HTTP/1.1</title>
3491    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3492      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
3493      <address><email></email></address>
3494    </author>
3495    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3496      <organization>MIT Laboratory for Computer Science</organization>
3497      <address><email></email></address>
3498    </author>
3499    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3500      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
3501      <address><email></email></address>
3502    </author>
3503    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3504      <organization>MIT Laboratory for Computer Science</organization>
3505      <address><email></email></address>
3506    </author>
3507    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3508      <organization>MIT Laboratory for Computer Science</organization>
3509      <address><email></email></address>
3510    </author>
3511    <date month="January" year="1997"/>
3512  </front>
3513  <seriesInfo name="RFC" value="2068"/>
3516<reference anchor="RFC2145">
3517  <front>
3518    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
3519    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3520      <organization>Western Research Laboratory</organization>
3521      <address><email></email></address>
3522    </author>
3523    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3524      <organization>Department of Information and Computer Science</organization>
3525      <address><email></email></address>
3526    </author>
3527    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3528      <organization>MIT Laboratory for Computer Science</organization>
3529      <address><email></email></address>
3530    </author>
3531    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3532      <organization>W3 Consortium</organization>
3533      <address><email></email></address>
3534    </author>
3535    <date month="May" year="1997"/>
3536  </front>
3537  <seriesInfo name="RFC" value="2145"/>
3540<reference anchor="RFC2324">
3541  <front>
3542    <title abbrev="HTCPCP/1.0">Hyper Text Coffee Pot Control Protocol (HTCPCP/1.0)</title>
3543    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3544      <organization>Xerox Palo Alto Research Center</organization>
3545      <address><email></email></address>
3546    </author>
3547    <date month="April" day="1" year="1998"/>
3548  </front>
3549  <seriesInfo name="RFC" value="2324"/>
3552<reference anchor="RFC2616">
3553  <front>
3554    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
3555    <author initials="R." surname="Fielding" fullname="R. Fielding">
3556      <organization>University of California, Irvine</organization>
3557      <address><email></email></address>
3558    </author>
3559    <author initials="J." surname="Gettys" fullname="J. Gettys">
3560      <organization>W3C</organization>
3561      <address><email></email></address>
3562    </author>
3563    <author initials="J." surname="Mogul" fullname="J. Mogul">
3564      <organization>Compaq Computer Corporation</organization>
3565      <address><email></email></address>
3566    </author>
3567    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
3568      <organization>MIT Laboratory for Computer Science</organization>
3569      <address><email></email></address>
3570    </author>
3571    <author initials="L." surname="Masinter" fullname="L. Masinter">
3572      <organization>Xerox Corporation</organization>
3573      <address><email></email></address>
3574    </author>
3575    <author initials="P." surname="Leach" fullname="P. Leach">
3576      <organization>Microsoft Corporation</organization>
3577      <address><email></email></address>
3578    </author>
3579    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
3580      <organization>W3C</organization>
3581      <address><email></email></address>
3582    </author>
3583    <date month="June" year="1999"/>
3584  </front>
3585  <seriesInfo name="RFC" value="2616"/>
3588<reference anchor="RFC2821">
3589  <front>
3590    <title>Simple Mail Transfer Protocol</title>
3591    <author initials="J." surname="Klensin" fullname="J. Klensin">
3592      <organization>AT&amp;T Laboratories</organization>
3593      <address><email></email></address>
3594    </author>
3595    <date year="2001" month="April"/>
3596  </front>
3597  <seriesInfo name="RFC" value="2821"/>
3600<reference anchor="RFC2822">
3601  <front>
3602    <title>Internet Message Format</title>
3603    <author initials="P." surname="Resnick" fullname="P. Resnick">
3604      <organization>QUALCOMM Incorporated</organization>
3605    </author>
3606    <date year="2001" month="April"/>
3607  </front>
3608  <seriesInfo name="RFC" value="2822"/>
3611<reference anchor='RFC3977'>
3612  <front>
3613    <title>Network News Transfer Protocol (NNTP)</title>
3614    <author initials='C.' surname='Feather' fullname='C. Feather'>
3615      <organization>THUS plc</organization>
3616      <address><email></email></address>
3617    </author>
3618    <date year='2006' month='October' />
3619  </front>
3620  <seriesInfo name="RFC" value="3977"/>
3623<reference anchor="RFC4288">
3624  <front>
3625    <title>Media Type Specifications and Registration Procedures</title>
3626    <author initials="N." surname="Freed" fullname="N. Freed">
3627      <organization>Sun Microsystems</organization>
3628      <address>
3629        <email></email>
3630      </address>
3631    </author>
3632    <author initials="J." surname="Klensin" fullname="J. Klensin">
3633      <organization/>
3634      <address>
3635        <email></email>
3636      </address>
3637    </author>
3638    <date year="2005" month="December"/>
3639  </front>
3640  <seriesInfo name="BCP" value="13"/>
3641  <seriesInfo name="RFC" value="4288"/>
3644<reference anchor="Spe" target="">
3645  <front>
3646  <title>Analysis of HTTP Performance Problems</title>
3647  <author initials="S." surname="Spero" fullname="Simon E. Spero">
3648    <organization/>
3649  </author>
3650  <date/>
3651  </front>
3654<reference anchor="Tou1998" target="">
3655  <front>
3656  <title>Analysis of HTTP Performance</title>
3657  <author initials="J." surname="Touch" fullname="Joe Touch">
3658    <organization>USC/Information Sciences Institute</organization>
3659    <address><email></email></address>
3660  </author>
3661  <author initials="J." surname="Heidemann" fullname="John Heidemann">
3662    <organization>USC/Information Sciences Institute</organization>
3663    <address><email></email></address>
3664  </author>
3665  <author initials="K." surname="Obraczka" fullname="Katia Obraczka">
3666    <organization>USC/Information Sciences Institute</organization>
3667    <address><email></email></address>
3668  </author>
3669  <date year="1998" month="Aug"/>
3670  </front>
3671  <seriesInfo name="ISI Research Report" value="ISI/RR-98-463"/>
3672  <annotation>(original report dated Aug. 1996)</annotation>
3675<reference anchor="WAIS">
3676  <front>
3677    <title>WAIS Interface Protocol Prototype Functional Specification (v1.5)</title>
3678    <author initials="F." surname="Davis" fullname="F. Davis">
3679      <organization>Thinking Machines Corporation</organization>
3680    </author>
3681    <author initials="B." surname="Kahle" fullname="B. Kahle">
3682      <organization>Thinking Machines Corporation</organization>
3683    </author>
3684    <author initials="H." surname="Morris" fullname="H. Morris">
3685      <organization>Thinking Machines Corporation</organization>
3686    </author>
3687    <author initials="J." surname="Salem" fullname="J. Salem">
3688      <organization>Thinking Machines Corporation</organization>
3689    </author>
3690    <author initials="T." surname="Shen" fullname="T. Shen">
3691      <organization>Thinking Machines Corporation</organization>
3692    </author>
3693    <author initials="R." surname="Wang" fullname="R. Wang">
3694      <organization>Thinking Machines Corporation</organization>
3695    </author>
3696    <author initials="J." surname="Sui" fullname="J. Sui">
3697      <organization>Thinking Machines Corporation</organization>
3698    </author>
3699    <author initials="M." surname="Grinbaum" fullname="M. Grinbaum">
3700      <organization>Thinking Machines Corporation</organization>
3701    </author>
3702    <date month="April" year="1990"/>
3703  </front>
3704  <seriesInfo name="Thinking Machines Corporation" value=""/>
3710<section title="Internet Media Types" anchor="">
3712   In addition to defining HTTP/1.1, this document serves
3713   as the specification for the Internet media type "message/http" and
3714   "application/http". The following is to be registered with IANA <xref target="RFC4288"/>.
3716<section title="Internet Media Type message/http" anchor="">
3717<iref item="Media Type" subitem="message/http" primary="true"/>
3718<iref item="message/http Media Type" primary="true"/>
3720   The message/http type can be used to enclose a single HTTP request or
3721   response message, provided that it obeys the MIME restrictions for all
3722   "message" types regarding line length and encodings.
3725  <list style="hanging" x:indent="12em">
3726    <t hangText="Type name:">
3727      message
3728    </t>
3729    <t hangText="Subtype name:">
3730      http
3731    </t>
3732    <t hangText="Required parameters:">
3733      none
3734    </t>
3735    <t hangText="Optional parameters:">
3736      version, msgtype
3737      <list style="hanging">
3738        <t hangText="version:">
3739          The HTTP-Version number of the enclosed message
3740          (e.g., "1.1"). If not present, the version can be
3741          determined from the first line of the body.
3742        </t>
3743        <t hangText="msgtype:">
3744          The message type -- "request" or "response". If not
3745          present, the type can be determined from the first
3746          line of the body.
3747        </t>
3748      </list>
3749    </t>
3750    <t hangText="Encoding considerations:">
3751      only "7bit", "8bit", or "binary" are permitted
3752    </t>
3753    <t hangText="Security considerations:">
3754      none
3755    </t>
3756    <t hangText="Interoperability considerations:">
3757      none
3758    </t>
3759    <t hangText="Published specification:">
3760      This specification (see <xref target=""/>).
3761    </t>
3762    <t hangText="Applications that use this media type:">
3763    </t>
3764    <t hangText="Additional information:">
3765      <list style="hanging">
3766        <t hangText="Magic number(s):">none</t>
3767        <t hangText="File extension(s):">none</t>
3768        <t hangText="Macintosh file type code(s):">none</t>
3769      </list>
3770    </t>
3771    <t hangText="Person and email address to contact for further information:">
3772      See Authors Section.
3773    </t>
3774                <t hangText="Intended usage:">
3775                  COMMON
3776    </t>
3777                <t hangText="Restrictions on usage:">
3778                  none
3779    </t>
3780    <t hangText="Author/Change controller:">
3781      IESG
3782    </t>
3783  </list>
3786<section title="Internet Media Type application/http" anchor="">
3787<iref item="Media Type" subitem="application/http" primary="true"/>
3788<iref item="application/http Media Type" primary="true"/>
3790   The application/http type can be used to enclose a pipeline of one or more
3791   HTTP request or response messages (not intermixed).
3794  <list style="hanging" x:indent="12em">
3795    <t hangText="Type name:">
3796      application
3797    </t>
3798    <t hangText="Subtype name:">
3799      http
3800    </t>
3801    <t hangText="Required parameters:">
3802      none
3803    </t>
3804    <t hangText="Optional parameters:">
3805      version, msgtype
3806      <list style="hanging">
3807        <t hangText="version:">
3808          The HTTP-Version number of the enclosed messages
3809          (e.g., "1.1"). If not present, the version can be
3810          determined from the first line of the body.
3811        </t>
3812        <t hangText="msgtype:">
3813          The message type -- "request" or "response". If not
3814          present, the type can be determined from the first
3815          line of the body.
3816        </t>
3817      </list>
3818    </t>
3819    <t hangText="Encoding considerations:">
3820      HTTP messages enclosed by this type
3821      are in "binary" format; use of an appropriate
3822      Content-Transfer-Encoding is required when
3823      transmitted via E-mail.
3824    </t>
3825    <t hangText="Security considerations:">
3826      none
3827    </t>
3828    <t hangText="Interoperability considerations:">
3829      none
3830    </t>
3831    <t hangText="Published specification:">
3832      This specification (see <xref target=""/>).
3833    </t>
3834    <t hangText="Applications that use this media type:">
3835    </t>
3836    <t hangText="Additional information:">
3837      <list style="hanging">
3838        <t hangText="Magic number(s):">none</t>
3839        <t hangText="File extension(s):">none</t>
3840        <t hangText="Macintosh file type code(s):">none</t>
3841      </list>
3842    </t>
3843    <t hangText="Person and email address to contact for further information:">
3844      See Authors Section.
3845    </t>
3846                <t hangText="Intended usage:">
3847                  COMMON
3848    </t>
3849                <t hangText="Restrictions on usage:">
3850                  none
3851    </t>
3852    <t hangText="Author/Change controller:">
3853      IESG
3854    </t>
3855  </list>
3860<section title="Tolerant Applications" anchor="tolerant.applications">
3862   Although this document specifies the requirements for the generation
3863   of HTTP/1.1 messages, not all applications will be correct in their
3864   implementation. We therefore recommend that operational applications
3865   be tolerant of deviations whenever those deviations can be
3866   interpreted unambiguously.
3869   Clients &SHOULD; be tolerant in parsing the Status-Line and servers
3870   tolerant when parsing the Request-Line. In particular, they &SHOULD;
3871   accept any amount of SP or HTAB characters between fields, even though
3872   only a single SP is required.
3875   The line terminator for message-header fields is the sequence CRLF.
3876   However, we recommend that applications, when parsing such headers,
3877   recognize a single LF as a line terminator and ignore the leading CR.
3880   The character set of an entity-body &SHOULD; be labeled as the lowest
3881   common denominator of the character codes used within that body, with
3882   the exception that not labeling the entity is preferred over labeling
3883   the entity with the labels US-ASCII or ISO-8859-1. See &payload;.
3886   Additional rules for requirements on parsing and encoding of dates
3887   and other potential problems with date encodings include:
3890  <list style="symbols">
3891     <t>HTTP/1.1 clients and caches &SHOULD; assume that an RFC-850 date
3892        which appears to be more than 50 years in the future is in fact
3893        in the past (this helps solve the "year 2000" problem).</t>
3895     <t>An HTTP/1.1 implementation &MAY; internally represent a parsed
3896        Expires date as earlier than the proper value, but &MUST-NOT;
3897        internally represent a parsed Expires date as later than the
3898        proper value.</t>
3900     <t>All expiration-related calculations &MUST; be done in GMT. The
3901        local time zone &MUST-NOT; influence the calculation or comparison
3902        of an age or expiration time.</t>
3904     <t>If an HTTP header incorrectly carries a date value with a time
3905        zone other than GMT, it &MUST; be converted into GMT using the
3906        most conservative possible conversion.</t>
3907  </list>
3911<section title="Conversion of Date Formats" anchor="">
3913   HTTP/1.1 uses a restricted set of date formats (<xref target=""/>) to
3914   simplify the process of date comparison. Proxies and gateways from
3915   other protocols &SHOULD; ensure that any Date header field present in a
3916   message conforms to one of the HTTP/1.1 formats and rewrite the date
3917   if necessary.
3921<section title="Compatibility with Previous Versions" anchor="compatibility">
3923   It is beyond the scope of a protocol specification to mandate
3924   compliance with previous versions. HTTP/1.1 was deliberately
3925   designed, however, to make supporting previous versions easy. It is
3926   worth noting that, at the time of composing this specification
3927   (1996), we would expect commercial HTTP/1.1 servers to:
3928  <list style="symbols">
3929     <t>recognize the format of the Request-Line for HTTP/0.9, 1.0, and
3930        1.1 requests;</t>
3932     <t>understand any valid request in the format of HTTP/0.9, 1.0, or
3933        1.1;</t>
3935     <t>respond appropriately with a message in the same major version
3936        used by the client.</t>
3937  </list>
3940   And we would expect HTTP/1.1 clients to:
3941  <list style="symbols">
3942     <t>recognize the format of the Status-Line for HTTP/1.0 and 1.1
3943        responses;</t>
3945     <t>understand any valid response in the format of HTTP/0.9, 1.0, or
3946        1.1.</t>
3947  </list>
3950   For most implementations of HTTP/1.0, each connection is established
3951   by the client prior to the request and closed by the server after
3952   sending the response. Some implementations implement the Keep-Alive
3953   version of persistent connections described in <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/>.
3956<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
3958   This section summarizes major differences between versions HTTP/1.0
3959   and HTTP/1.1.
3962<section title="Changes to Simplify Multi-homed Web Servers and Conserve IP Addresses" anchor="">
3964   The requirements that clients and servers support the Host request-header,
3965   report an error if the Host request-header (<xref target=""/>) is
3966   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-uri"/>)
3967   are among the most important changes defined by this
3968   specification.
3971   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
3972   addresses and servers; there was no other established mechanism for
3973   distinguishing the intended server of a request than the IP address
3974   to which that request was directed. The changes outlined above will
3975   allow the Internet, once older HTTP clients are no longer common, to
3976   support multiple Web sites from a single IP address, greatly
3977   simplifying large operational Web servers, where allocation of many
3978   IP addresses to a single host has created serious problems. The
3979   Internet will also be able to recover the IP addresses that have been
3980   allocated for the sole purpose of allowing special-purpose domain
3981   names to be used in root-level HTTP URLs. Given the rate of growth of
3982   the Web, and the number of servers already deployed, it is extremely
3983   important that all implementations of HTTP (including updates to
3984   existing HTTP/1.0 applications) correctly implement these
3985   requirements:
3986  <list style="symbols">
3987     <t>Both clients and servers &MUST; support the Host request-header.</t>
3989     <t>A client that sends an HTTP/1.1 request &MUST; send a Host header.</t>
3991     <t>Servers &MUST; report a 400 (Bad Request) error if an HTTP/1.1
3992        request does not include a Host request-header.</t>
3994     <t>Servers &MUST; accept absolute URIs.</t>
3995  </list>
4000<section title="Compatibility with HTTP/1.0 Persistent Connections" anchor="compatibility.with.http.1.0.persistent.connections">
4002   Some clients and servers might wish to be compatible with some
4003   previous implementations of persistent connections in HTTP/1.0
4004   clients and servers. Persistent connections in HTTP/1.0 are
4005   explicitly negotiated as they are not the default behavior. HTTP/1.0
4006   experimental implementations of persistent connections are faulty,
4007   and the new facilities in HTTP/1.1 are designed to rectify these
4008   problems. The problem was that some existing 1.0 clients may be
4009   sending Keep-Alive to a proxy server that doesn't understand
4010   Connection, which would then erroneously forward it to the next
4011   inbound server, which would establish the Keep-Alive connection and
4012   result in a hung HTTP/1.0 proxy waiting for the close on the
4013   response. The result is that HTTP/1.0 clients must be prevented from
4014   using Keep-Alive when talking to proxies.
4017   However, talking to proxies is the most important use of persistent
4018   connections, so that prohibition is clearly unacceptable. Therefore,
4019   we need some other mechanism for indicating a persistent connection
4020   is desired, which is safe to use even when talking to an old proxy
4021   that ignores Connection. Persistent connections are the default for
4022   HTTP/1.1 messages; we introduce a new keyword (Connection: close) for
4023   declaring non-persistence. See <xref target="header.connection"/>.
4026   The original HTTP/1.0 form of persistent connections (the Connection:
4027   Keep-Alive and Keep-Alive header) is documented in <xref target="RFC2068"/>.
4031<section title="Changes from RFC 2068" anchor="changes.from.rfc.2068">
4033   This specification has been carefully audited to correct and
4034   disambiguate key word usage; RFC 2068 had many problems in respect to
4035   the conventions laid out in <xref target="RFC2119"/>.
4038   Transfer-coding and message lengths all interact in ways that
4039   required fixing exactly when chunked encoding is used (to allow for
4040   transfer encoding that may not be self delimiting); it was important
4041   to straighten out exactly how message lengths are computed. (Sections
4042   <xref target="transfer.codings" format="counter"/>, <xref target="message.length" format="counter"/>,
4043   <xref target="header.content-length" format="counter"/>,
4044   see also <xref target="Part3"/>, <xref target="Part5"/> and <xref target="Part6"/>)
4047   The use and interpretation of HTTP version numbers has been clarified
4048   by <xref target="RFC2145"/>. Require proxies to upgrade requests to highest protocol
4049   version they support to deal with problems discovered in HTTP/1.0
4050   implementations (<xref target="http.version"/>)
4053   Transfer-coding had significant problems, particularly with
4054   interactions with chunked encoding. The solution is that transfer-codings
4055   become as full fledged as content-codings. This involves
4056   adding an IANA registry for transfer-codings (separate from content
4057   codings), a new header field (TE) and enabling trailer headers in the
4058   future. Transfer encoding is a major performance benefit, so it was
4059   worth fixing <xref target="Nie1997"/>. TE also solves another, obscure, downward
4060   interoperability problem that could have occurred due to interactions
4061   between authentication trailers, chunked encoding and HTTP/1.0
4062   clients.(Section <xref target="transfer.codings" format="counter"/>, <xref target="chunked.transfer.encoding" format="counter"/>,
4063   and <xref target="header.te" format="counter"/>)
4067<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
4069  The CHAR rule does not allow the NUL character anymore (this affects
4070  the comment and quoted-string rules).
4071  (<xref target="basic.rules"/>)
4074  Clarify that HTTP-Version is case sensitive.
4075  (<xref target="http.version"/>)
4078  Remove reference to non-existant identity transfer-coding value tokens.
4079  (Sections <xref format="counter" target="transfer.codings"/> and
4080  <xref format="counter" target="message.length"/>)
4083  Clarification that the chunk length does not include
4084  the count of the octets in the chunk header and trailer.
4085  (<xref target="chunked.transfer.encoding"/>)
4088  Fix BNF to add query, as the abs_path production in
4089  <xref x:sec="3" x:fmt="of" target="RFC2396"/> doesn't define it.
4090  (<xref target="request-uri"/>)
4093  Clarify exactly when close connection options must be sent.
4094  (<xref target="header.connection"/>)
4099<section title="Change Log (to be removed by RFC Editor before publication)">
4101<section title="Since RFC2616">
4103  Extracted relevant partitions from <xref target="RFC2616"/>.
4107<section title="Since draft-ietf-httpbis-p1-messaging-00">
4109  Closed issues:
4110  <list style="symbols">
4111    <t>
4112      <eref target=""/>:
4113      "HTTP Version should be case sensitive"
4114      (<eref target=""/>)
4115    </t>
4116    <t>
4117      <eref target=""/>:
4118      "'unsafe' characters"
4119      (<eref target=""/>)
4120    </t>
4121    <t>
4122      <eref target=""/>:
4123      "Chunk Size Definition"
4124      (<eref target=""/>)
4125    </t>
4126    <t>
4127      <eref target=""/>:
4128      "Message Length"
4129      (<eref target=""/>)
4130    </t>
4131    <t>
4132      <eref target=""/>:
4133      "Media Type Registrations"
4134      (<eref target=""/>)
4135    </t>
4136    <t>
4137      <eref target=""/>:
4138      "URI includes query"
4139      (<eref target=""/>)
4140    </t>
4141    <t>
4142      <eref target=""/>:
4143      "No close on 1xx responses"
4144      (<eref target=""/>)
4145    </t>
4146    <t>
4147      <eref target=""/>:
4148      "Remove 'identity' token references"
4149      (<eref target=""/>)
4150    </t>
4151    <t>
4152      <eref target=""/>:
4153      "Import query BNF"
4154    </t>
4155    <t>
4156      <eref target=""/>:
4157      "qdtext BNF"
4158    </t>
4159    <t>
4160      <eref target=""/>:
4161      "Normative and Informative references"
4162    </t>
4163    <t>
4164      <eref target=""/>:
4165      "RFC2606 Compliance"
4166    </t>
4167    <t>
4168      <eref target=""/>:
4169      "RFC977 reference"
4170    </t>
4171    <t>
4172      <eref target=""/>:
4173      "RFC1700 references"
4174    </t>
4175    <t>
4176      <eref target=""/>:
4177      "inconsistency in date format explanation"
4178    </t>
4179    <t>
4180      <eref target=""/>:
4181      "Date reference typo"
4182    </t>
4183    <t>
4184      <eref target=""/>:
4185      "Informative references"
4186    </t>
4187    <t>
4188      <eref target=""/>:
4189      "ISO-8859-1 Reference"
4190    </t>
4191    <t>
4192      <eref target=""/>:
4193      "Normative up-to-date references"
4194    </t>
4195  </list>
4198  Other changes:
4199  <list style="symbols">
4200    <t>
4201      Update media type registrations to use RFC4288 template.
4202    </t>
4203    <t>
4204      Use names of RFC4234 core rules DQUOTE and HTAB,
4205      fix broken ABNF for chunk-data
4206      (work in progress on <eref target=""/>)
4207    </t>
4208  </list>
4212<section title="Since draft-ietf-httpbis-p1-messaging-01">
4214  Closed issues:
4215  <list style="symbols">
4216    <t>
4217      <eref target=""/>:
4218      "Bodies on GET (and other) requests"
4219    </t>
4220    <t>
4221      <eref target=""/>:
4222      "Updating to RFC4288"
4223    </t>
4224    <t>
4225      <eref target=""/>:
4226      "Status Code and Reason Phrase"
4227    </t>
4228    <t>
4229      <eref target=""/>:
4230      "rel_path not used"
4231    </t>
4232  </list>
4235  Ongoing work on ABNF conversion (<eref target=""/>):
4236  <list style="symbols">
4237    <t>
4238      Get rid of duplicate BNF rule names ("host" -> "uri-host", "trailer" ->
4239      "trailer-part").
4240    </t>
4241    <t>
4242      Avoid underscore character in rule names ("http_URL" ->
4243      "http-URL", "abs_path" -> "path-absolute").
4244    </t>
4245    <t>
4246      Add rules for terms imported from URI spec ("absoluteURI", "authority",
4247      "path-absolute", "port", "query", "relativeURI", "host) -- these will
4248      have to be updated when switching over to RFC3986.
4249    </t>
4250    <t>
4251      Synchronize core rules with RFC5234 (this includes a change to CHAR
4252      which now excludes NUL).
4253    </t>
4254    <t>
4255      Get rid of prose rules that span multiple lines.
4256    </t>
4257    <t>
4258      Get rid of unused rules LOALPHA and UPALPHA.
4259    </t>
4260    <t>
4261      Move "Product Tokens" section (back) into Part 1, as "token" is used
4262      in the definition of the Upgrade header.
4263    </t>
4264  </list>
Note: See TracBrowser for help on using the repository browser.