source: draft-ietf-httpbis/latest/p1-messaging.xml @ 173

Last change on this file since 173 was 173, checked in by julian.reschke@…, 14 years ago

Resolve #82: do not use rel_path in prose anymore, it hasn't been used inside the ABNF for http_URL since 2616 (closes #82).

  • Property svn:eol-style set to native
File size: 177.2 KB
1<?xml version="1.0" encoding="utf-8"?>
2<?xml-stylesheet type='text/xsl' href='../myxml2rfc.xslt'?>
3<!DOCTYPE rfc [
4  <!ENTITY MAY "<bcp14 xmlns=''>MAY</bcp14>">
5  <!ENTITY MUST "<bcp14 xmlns=''>MUST</bcp14>">
6  <!ENTITY MUST-NOT "<bcp14 xmlns=''>MUST NOT</bcp14>">
7  <!ENTITY OPTIONAL "<bcp14 xmlns=''>OPTIONAL</bcp14>">
8  <!ENTITY RECOMMENDED "<bcp14 xmlns=''>RECOMMENDED</bcp14>">
9  <!ENTITY REQUIRED "<bcp14 xmlns=''>REQUIRED</bcp14>">
10  <!ENTITY SHALL "<bcp14 xmlns=''>SHALL</bcp14>">
11  <!ENTITY SHALL-NOT "<bcp14 xmlns=''>SHALL NOT</bcp14>">
12  <!ENTITY SHOULD "<bcp14 xmlns=''>SHOULD</bcp14>">
13  <!ENTITY SHOULD-NOT "<bcp14 xmlns=''>SHOULD NOT</bcp14>">
14  <!ENTITY ID-VERSION "latest">
15  <!ENTITY ID-MONTH "January">
16  <!ENTITY ID-YEAR "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-ext allow-markup-in-artwork="yes" ?>
48<?rfc-ext include-references-in-index="yes" ?>
49<rfc obsoletes="2616" category="std"
50     ipr="full3978" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
51     xmlns:x=''>
54  <title abbrev="HTTP/1.1, Part 1">HTTP/1.1, part 1: URIs, Connections, and Message Parsing</title>
56  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
57    <organization abbrev="Day Software">Day Software</organization>
58    <address>
59      <postal>
60        <street>23 Corporate Plaza DR, Suite 280</street>
61        <city>Newport Beach</city>
62        <region>CA</region>
63        <code>92660</code>
64        <country>USA</country>
65      </postal>
66      <phone>+1-949-706-5300</phone>
67      <facsimile>+1-949-706-5305</facsimile>
68      <email></email>
69      <uri></uri>
70    </address>
71  </author>
73  <author initials="J." surname="Gettys" fullname="Jim Gettys">
74    <organization>One Laptop per Child</organization>
75    <address>
76      <postal>
77        <street>21 Oak Knoll Road</street>
78        <city>Carlisle</city>
79        <region>MA</region>
80        <code>01741</code>
81        <country>USA</country>
82      </postal>
83      <email></email>
84      <uri></uri>
85    </address>
86  </author>
88  <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
89    <organization abbrev="HP">Hewlett-Packard Company</organization>
90    <address>
91      <postal>
92        <street>HP Labs, Large Scale Systems Group</street>
93        <street>1501 Page Mill Road, MS 1177</street>
94        <city>Palo Alto</city>
95        <region>CA</region>
96        <code>94304</code>
97        <country>USA</country>
98      </postal>
99      <email></email>
100    </address>
101  </author>
103  <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
104    <organization abbrev="Microsoft">Microsoft Corporation</organization>
105    <address>
106      <postal>
107        <street>1 Microsoft Way</street>
108        <city>Redmond</city>
109        <region>WA</region>
110        <code>98052</code>
111        <country>USA</country>
112      </postal>
113      <email></email>
114    </address>
115  </author>
117  <author initials="L." surname="Masinter" fullname="Larry Masinter">
118    <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
119    <address>
120      <postal>
121        <street>345 Park Ave</street>
122        <city>San Jose</city>
123        <region>CA</region>
124        <code>95110</code>
125        <country>USA</country>
126      </postal>
127      <email></email>
128      <uri></uri>
129    </address>
130  </author>
132  <author initials="P." surname="Leach" fullname="Paul J. Leach">
133    <organization abbrev="Microsoft">Microsoft Corporation</organization>
134    <address>
135      <postal>
136        <street>1 Microsoft Way</street>
137        <city>Redmond</city>
138        <region>WA</region>
139        <code>98052</code>
140      </postal>
141      <email></email>
142    </address>
143  </author>
145  <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
146    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
147    <address>
148      <postal>
149        <street>MIT Computer Science and Artificial Intelligence Laboratory</street>
150        <street>The Stata Center, Building 32</street>
151        <street>32 Vassar Street</street>
152        <city>Cambridge</city>
153        <region>MA</region>
154        <code>02139</code>
155        <country>USA</country>
156      </postal>
157      <email></email>
158      <uri></uri>
159    </address>
160  </author>
162  <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
163    <organization abbrev="W3C">World Wide Web Consortium</organization>
164    <address>
165      <postal>
166        <street>W3C / ERCIM</street>
167        <street>2004, rte des Lucioles</street>
168        <city>Sophia-Antipolis</city>
169        <region>AM</region>
170        <code>06902</code>
171        <country>France</country>
172      </postal>
173      <email></email>
174      <uri></uri>
175    </address>
176  </author>
178  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
179    <organization abbrev="greenbytes">greenbytes GmbH</organization>
180    <address>
181      <postal>
182        <street>Hafenweg 16</street>
183        <city>Muenster</city><region>NW</region><code>48155</code>
184        <country>Germany</country>
185      </postal>
186      <phone>+49 251 2807760</phone>   
187      <facsimile>+49 251 2807761</facsimile>   
188      <email></email>       
189      <uri></uri>     
190    </address>
191  </author>
193  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
197   The Hypertext Transfer Protocol (HTTP) is an application-level
198   protocol for distributed, collaborative, hypermedia information
199   systems. HTTP has been in use by the World Wide Web global information
200   initiative since 1990. This document is Part 1 of the seven-part specification
201   that defines the protocol referred to as "HTTP/1.1" and, taken together,
202   obsoletes RFC 2616.  Part 1 provides an overview of HTTP and
203   its associated terminology, defines the "http" and "https" Uniform
204   Resource Identifier (URI) schemes, defines the generic message syntax
205   and parsing requirements for HTTP message frames, and describes
206   general security concerns for implementations.
210<note title="Editorial Note (To be removed by RFC Editor)">
211  <t>
212    Discussion of this draft should take place on the HTTPBIS working group
213    mailing list ( The current issues list is
214    at <eref target=""/>
215    and related documents (including fancy diffs) can be found at
216    <eref target=""/>.
217  </t>
218  <t>
219    This draft incorporates those issue resolutions that were either
220    collected in the original RFC2616 errata list (<eref target=""/>),
221    or which were agreed upon on the mailing list between October 2006 and
222    November 2007 (as published in "draft-lafon-rfc2616bis-03").
223  </t>
227<section title="Introduction" anchor="introduction">
229   The Hypertext Transfer Protocol (HTTP) is an application-level
230   protocol for distributed, collaborative, hypermedia information
231   systems. HTTP has been in use by the World-Wide Web global
232   information initiative since 1990. The first version of HTTP, commonly
233   referred to as HTTP/0.9, was a simple protocol for raw data transfer
234   across the Internet with only a single method and no metadata.
235   HTTP/1.0, as defined by <xref target="RFC1945"/>, improved
236   the protocol by allowing messages to be in the format of MIME-like
237   messages, containing metadata about the data transferred and
238   modifiers on the request/response semantics. However, HTTP/1.0 did
239   not sufficiently take into consideration the effects of hierarchical
240   proxies, caching, the need for persistent connections, or name-based
241   virtual hosts. In addition, the proliferation of incompletely-implemented
242   applications calling themselves "HTTP/1.0" necessitated a
243   protocol version change in order for two communicating applications
244   to determine each other's true capabilities.
247   This document is Part 1 of the seven-part specification that defines
248   the protocol referred to as "HTTP/1.1", obsoleting <xref target="RFC2616"/>.
249   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
250   requirements that enable reliable implementations and adding only
251   those new features that will either be safely ignored by an HTTP/1.0
252   recipient or only sent when communicating with a party advertising
253   compliance with HTTP/1.1.
254   Part 1 defines those aspects of HTTP/1.1 related to overall network
255   operation, message framing, interaction with transport protocols, and
256   URI schemes.
259   This document is currently disorganized in order to minimize the changes
260   between drafts and enable reviewers to see the smaller errata changes.
261   The next draft will reorganize the sections to better reflect the content.
262   In particular, the sections will be organized according to the typical
263   process of deciding when to use HTTP (URI schemes), overall network operation,
264   connection management, message framing, and generic message parsing.
265   The current mess reflects how widely dispersed these topics and associated
266   requirements had become in <xref target="RFC2616"/>.
269<section title="Purpose" anchor="intro.purpose">
271   Practical information systems require more functionality than simple
272   retrieval, including search, front-end update, and annotation. HTTP
273   allows an open-ended set of methods and headers that indicate the
274   purpose of a request <xref target="RFC2324"/>. It builds on the discipline of reference
275   provided by the Uniform Resource Identifier (URI) <xref target="RFC1630"/>, as a location
276   (URL) <xref target="RFC1738"/> or name (URN) <xref target="RFC1737"/>, for indicating the resource to which a
277   method is to be applied. Messages are passed in a format similar to
278   that used by Internet mail <xref target="RFC2822"/> as defined by the Multipurpose
279   Internet Mail Extensions (MIME) <xref target="RFC2045"/>.
282   HTTP is also used as a generic protocol for communication between
283   user agents and proxies/gateways to other Internet systems, including
284   those supported by the SMTP <xref target="RFC2821"/>, NNTP <xref target="RFC3977"/>, FTP <xref target="RFC959"/>, Gopher <xref target="RFC1436"/>,
285   and WAIS <xref target="WAIS"/> protocols. In this way, HTTP allows basic hypermedia
286   access to resources available from diverse applications.
290<section title="Requirements" anchor="intro.requirements">
292   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
293   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
294   document are to be interpreted as described in <xref target="RFC2119"/>.
297   An implementation is not compliant if it fails to satisfy one or more
298   of the &MUST; or &REQUIRED; level requirements for the protocols it
299   implements. An implementation that satisfies all the &MUST; or &REQUIRED;
300   level and all the &SHOULD; level requirements for its protocols is said
301   to be "unconditionally compliant"; one that satisfies all the &MUST;
302   level requirements but not all the &SHOULD; level requirements for its
303   protocols is said to be "conditionally compliant."
307<section title="Terminology" anchor="intro.terminology">
309   This specification uses a number of terms to refer to the roles
310   played by participants in, and objects of, the HTTP communication.
313  <iref item="connection"/>
314  <x:dfn>connection</x:dfn>
315  <list>
316    <t>
317      A transport layer virtual circuit established between two programs
318      for the purpose of communication.
319    </t>
320  </list>
323  <iref item="message"/>
324  <x:dfn>message</x:dfn>
325  <list>
326    <t>
327      The basic unit of HTTP communication, consisting of a structured
328      sequence of octets matching the syntax defined in <xref target="http.message"/> and
329      transmitted via the connection.
330    </t>
331  </list>
334  <iref item="request"/>
335  <x:dfn>request</x:dfn>
336  <list>
337    <t>
338      An HTTP request message, as defined in <xref target="request"/>.
339    </t>
340  </list>
343  <iref item="response"/>
344  <x:dfn>response</x:dfn>
345  <list>
346    <t>
347      An HTTP response message, as defined in <xref target="response"/>.
348    </t>
349  </list>
352  <iref item="resource"/>
353  <x:dfn>resource</x:dfn>
354  <list>
355    <t>
356      A network data object or service that can be identified by a URI,
357      as defined in <xref target="uri"/>. Resources may be available in multiple
358      representations (e.g. multiple languages, data formats, size, and
359      resolutions) or vary in other ways.
360    </t>
361  </list>
364  <iref item="entity"/>
365  <x:dfn>entity</x:dfn>
366  <list>
367    <t>
368      The information transferred as the payload of a request or
369      response. An entity consists of metainformation in the form of
370      entity-header fields and content in the form of an entity-body, as
371      described in &entity;.
372    </t>
373  </list>
376  <iref item="representation"/>
377  <x:dfn>representation</x:dfn>
378  <list>
379    <t>
380      An entity included with a response that is subject to content
381      negotiation, as described in &content.negotiation;. There may exist multiple
382      representations associated with a particular response status.
383    </t>
384  </list>
387  <iref item="content negotiation"/>
388  <x:dfn>content negotiation</x:dfn>
389  <list>
390    <t>
391      The mechanism for selecting the appropriate representation when
392      servicing a request, as described in &content.negotiation;. The
393      representation of entities in any response can be negotiated
394      (including error responses).
395    </t>
396  </list>
399  <iref item="variant"/>
400  <x:dfn>variant</x:dfn>
401  <list>
402    <t>
403      A resource may have one, or more than one, representation(s)
404      associated with it at any given instant. Each of these
405      representations is termed a `variant'.  Use of the term `variant'
406      does not necessarily imply that the resource is subject to content
407      negotiation.
408    </t>
409  </list>
412  <iref item="client"/>
413  <x:dfn>client</x:dfn>
414  <list>
415    <t>
416      A program that establishes connections for the purpose of sending
417      requests.
418    </t>
419  </list>
422  <iref item="user agent"/>
423  <x:dfn>user agent</x:dfn>
424  <list>
425    <t>
426      The client which initiates a request. These are often browsers,
427      editors, spiders (web-traversing robots), or other end user tools.
428    </t>
429  </list>
432  <iref item="server"/>
433  <x:dfn>server</x:dfn>
434  <list>
435    <t>
436      An application program that accepts connections in order to
437      service requests by sending back responses. Any given program may
438      be capable of being both a client and a server; our use of these
439      terms refers only to the role being performed by the program for a
440      particular connection, rather than to the program's capabilities
441      in general. Likewise, any server may act as an origin server,
442      proxy, gateway, or tunnel, switching behavior based on the nature
443      of each request.
444    </t>
445  </list>
448  <iref item="origin server"/>
449  <x:dfn>origin server</x:dfn>
450  <list>
451    <t>
452      The server on which a given resource resides or is to be created.
453    </t>
454  </list>
457  <iref item="proxy"/>
458  <x:dfn>proxy</x:dfn>
459  <list>
460    <t>
461      An intermediary program which acts as both a server and a client
462      for the purpose of making requests on behalf of other clients.
463      Requests are serviced internally or by passing them on, with
464      possible translation, to other servers. A proxy &MUST; implement
465      both the client and server requirements of this specification. A
466      "transparent proxy" is a proxy that does not modify the request or
467      response beyond what is required for proxy authentication and
468      identification. A "non-transparent proxy" is a proxy that modifies
469      the request or response in order to provide some added service to
470      the user agent, such as group annotation services, media type
471      transformation, protocol reduction, or anonymity filtering. Except
472      where either transparent or non-transparent behavior is explicitly
473      stated, the HTTP proxy requirements apply to both types of
474      proxies.
475    </t>
476  </list>
479  <iref item="gateway"/>
480  <x:dfn>gateway</x:dfn>
481  <list>
482    <t>
483      A server which acts as an intermediary for some other server.
484      Unlike a proxy, a gateway receives requests as if it were the
485      origin server for the requested resource; the requesting client
486      may not be aware that it is communicating with a gateway.
487    </t>
488  </list>
491  <iref item="tunnel"/>
492  <x:dfn>tunnel</x:dfn>
493  <list>
494    <t>
495      An intermediary program which is acting as a blind relay between
496      two connections. Once active, a tunnel is not considered a party
497      to the HTTP communication, though the tunnel may have been
498      initiated by an HTTP request. The tunnel ceases to exist when both
499      ends of the relayed connections are closed.
500    </t>
501  </list>
504  <iref item="cache"/>
505  <x:dfn>cache</x:dfn>
506  <list>
507    <t>
508      A program's local store of response messages and the subsystem
509      that controls its message storage, retrieval, and deletion. A
510      cache stores cacheable responses in order to reduce the response
511      time and network bandwidth consumption on future, equivalent
512      requests. Any client or server may include a cache, though a cache
513      cannot be used by a server that is acting as a tunnel.
514    </t>
515  </list>
518  <iref item="cacheable"/>
519  <x:dfn>cacheable</x:dfn>
520  <list>
521    <t>
522      A response is cacheable if a cache is allowed to store a copy of
523      the response message for use in answering subsequent requests. The
524      rules for determining the cacheability of HTTP responses are
525      defined in &caching;. Even if a resource is cacheable, there may
526      be additional constraints on whether a cache can use the cached
527      copy for a particular request.
528    </t>
529  </list>
532  <iref item="upstream"/>
533  <iref item="downstream"/>
534  <x:dfn>upstream</x:dfn>/<x:dfn>downstream</x:dfn>
535  <list>
536    <t>
537      Upstream and downstream describe the flow of a message: all
538      messages flow from upstream to downstream.
539    </t>
540  </list>
543  <iref item="inbound"/>
544  <iref item="outbound"/>
545  <x:dfn>inbound</x:dfn>/<x:dfn>outbound</x:dfn>
546  <list>
547    <t>
548      Inbound and outbound refer to the request and response paths for
549      messages: "inbound" means "traveling toward the origin server",
550      and "outbound" means "traveling toward the user agent"
551    </t>
552  </list>
556<section title="Overall Operation" anchor="intro.overall.operation">
558   HTTP is a request/response protocol. A client sends a
559   request to the server in the form of a request method, URI, and
560   protocol version, followed by a MIME-like message containing request
561   modifiers, client information, and possible body content over a
562   connection with a server. The server responds with a status line,
563   including the message's protocol version and a success or error code,
564   followed by a MIME-like message containing server information, entity
565   metainformation, and possible entity-body content. The relationship
566   between HTTP and MIME is described in &diff2045entity;.
569   Most HTTP communication is initiated by a user agent and consists of
570   a request to be applied to a resource on some origin server. In the
571   simplest case, this may be accomplished via a single connection (v)
572   between the user agent (UA) and the origin server (O).
574<figure><artwork type="drawing">
575       request chain ------------------------&gt;
576    UA -------------------v------------------- O
577       &lt;----------------------- response chain
580   A more complicated situation occurs when one or more intermediaries
581   are present in the request/response chain. There are three common
582   forms of intermediary: proxy, gateway, and tunnel. A proxy is a
583   forwarding agent, receiving requests for a URI in its absolute form,
584   rewriting all or part of the message, and forwarding the reformatted
585   request toward the server identified by the URI. A gateway is a
586   receiving agent, acting as a layer above some other server(s) and, if
587   necessary, translating the requests to the underlying server's
588   protocol. A tunnel acts as a relay point between two connections
589   without changing the messages; tunnels are used when the
590   communication needs to pass through an intermediary (such as a
591   firewall) even when the intermediary cannot understand the contents
592   of the messages.
594<figure><artwork type="drawing">
595       request chain --------------------------------------&gt;
596    UA -----v----- A -----v----- B -----v----- C -----v----- O
597       &lt;------------------------------------- response chain
600   The figure above shows three intermediaries (A, B, and C) between the
601   user agent and origin server. A request or response message that
602   travels the whole chain will pass through four separate connections.
603   This distinction is important because some HTTP communication options
604   may apply only to the connection with the nearest, non-tunnel
605   neighbor, only to the end-points of the chain, or to all connections
606   along the chain. Although the diagram is linear, each participant may
607   be engaged in multiple, simultaneous communications. For example, B
608   may be receiving requests from many clients other than A, and/or
609   forwarding requests to servers other than C, at the same time that it
610   is handling A's request.
613   Any party to the communication which is not acting as a tunnel may
614   employ an internal cache for handling requests. The effect of a cache
615   is that the request/response chain is shortened if one of the
616   participants along the chain has a cached response applicable to that
617   request. The following illustrates the resulting chain if B has a
618   cached copy of an earlier response from O (via C) for a request which
619   has not been cached by UA or A.
621<figure><artwork type="drawing">
622          request chain ----------&gt;
623       UA -----v----- A -----v----- B - - - - - - C - - - - - - O
624          &lt;--------- response chain
627   Not all responses are usefully cacheable, and some requests may
628   contain modifiers which place special requirements on cache behavior.
629   HTTP requirements for cache behavior and cacheable responses are
630   defined in &caching;.
633   In fact, there are a wide variety of architectures and configurations
634   of caches and proxies currently being experimented with or deployed
635   across the World Wide Web. These systems include national hierarchies
636   of proxy caches to save transoceanic bandwidth, systems that
637   broadcast or multicast cache entries, organizations that distribute
638   subsets of cached data via CD-ROM, and so on. HTTP systems are used
639   in corporate intranets over high-bandwidth links, and for access via
640   PDAs with low-power radio links and intermittent connectivity. The
641   goal of HTTP/1.1 is to support the wide diversity of configurations
642   already deployed while introducing protocol constructs that meet the
643   needs of those who build web applications that require high
644   reliability and, failing that, at least reliable indications of
645   failure.
648   HTTP communication usually takes place over TCP/IP connections. The
649   default port is TCP 80 (<eref target=""/>), but other ports can be used. This does
650   not preclude HTTP from being implemented on top of any other protocol
651   on the Internet, or on other networks. HTTP only presumes a reliable
652   transport; any protocol that provides such guarantees can be used;
653   the mapping of the HTTP/1.1 request and response structures onto the
654   transport data units of the protocol in question is outside the scope
655   of this specification.
658   In HTTP/1.0, most implementations used a new connection for each
659   request/response exchange. In HTTP/1.1, a connection may be used for
660   one or more request/response exchanges, although connections may be
661   closed for a variety of reasons (see <xref target="persistent.connections"/>).
666<section title="Notational Conventions and Generic Grammar" anchor="notation">
668<section title="Augmented BNF" anchor="notation.abnf">
670   All of the mechanisms specified in this document are described in
671   both prose and an augmented Backus-Naur Form (BNF) similar to that
672   used by <xref target="RFC822ABNF"/>. Implementors will need to be familiar with the
673   notation in order to understand this specification. The augmented BNF
674   includes the following constructs:
677   name = definition
678  <list>
679    <t>
680      The name of a rule is simply the name itself (without any
681      enclosing "&lt;" and "&gt;") and is separated from its definition by the
682      equal "=" character. White space is only significant in that
683      indentation of continuation lines is used to indicate a rule
684      definition that spans more than one line. Certain basic rules are
685      in uppercase, such as SP, LWS, HTAB, CRLF, DIGIT, ALPHA, etc. Angle
686      brackets are used within definitions whenever their presence will
687      facilitate discerning the use of rule names.
688    </t>
689  </list>
692   "literal"
693  <list>
694    <t>
695      Quotation marks surround literal text. Unless stated otherwise,
696      the text is case-insensitive.
697    </t>
698  </list>
701   rule1 | rule2
702  <list>
703    <t>
704      Elements separated by a bar ("|") are alternatives, e.g., "yes |
705      no" will accept yes or no.
706    </t>
707  </list>
710   (rule1 rule2)
711  <list>
712    <t>
713      Elements enclosed in parentheses are treated as a single element.
714      Thus, "(elem (foo | bar) elem)" allows the token sequences "elem
715      foo elem" and "elem bar elem".
716    </t>
717  </list>
720   *rule
721  <list>
722    <t>
723      The character "*" preceding an element indicates repetition. The
724      full form is "&lt;n&gt;*&lt;m&gt;element" indicating at least &lt;n&gt; and at most
725      &lt;m&gt; occurrences of element. Default values are 0 and infinity so
726      that "*(element)" allows any number, including zero; "1*element"
727      requires at least one; and "1*2element" allows one or two.
728    </t>
729  </list>
732   [rule]
733  <list>
734    <t>
735      Square brackets enclose optional elements; "[foo bar]" is
736      equivalent to "*1(foo bar)".
737    </t>
738  </list>
741   N rule
742  <list>
743    <t>
744      Specific repetition: "&lt;n&gt;(element)" is equivalent to
745      "&lt;n&gt;*&lt;n&gt;(element)"; that is, exactly &lt;n&gt; occurrences of (element).
746      Thus 2DIGIT is a 2-digit number, and 3ALPHA is a string of three
747      alphabetic characters.
748    </t>
749  </list>
752   #rule
753  <list>
754    <t>
755      A construct "#" is defined, similar to "*", for defining lists of
756      elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating at least
757      &lt;n&gt; and at most &lt;m&gt; elements, each separated by one or more commas
758      (",") and &OPTIONAL; linear white space (LWS). This makes the usual
759      form of lists very easy; a rule such as
760    </t>
761    <t>
762         ( *LWS element *( *LWS "," *LWS element ))
763    </t>
764    <t>
765      can be shown as
766    </t>
767    <t>
768         1#element
769    </t>
770    <t>
771      Wherever this construct is used, null elements are allowed, but do
772      not contribute to the count of elements present. That is,
773      "(element), , (element) " is permitted, but counts as only two
774      elements. Therefore, where at least one element is required, at
775      least one non-null element &MUST; be present. Default values are 0
776      and infinity so that "#element" allows any number, including zero;
777      "1#element" requires at least one; and "1#2element" allows one or
778      two.
779    </t>
780  </list>
783   ; comment
784  <list>
785    <t>
786      A semi-colon, set off some distance to the right of rule text,
787      starts a comment that continues to the end of line. This is a
788      simple way of including useful notes in parallel with the
789      specifications.
790    </t>
791  </list>
794   implied *LWS
795  <list>
796    <t>
797      The grammar described by this specification is word-based. Except
798      where noted otherwise, linear white space (LWS) can be included
799      between any two adjacent words (token or quoted-string), and
800      between adjacent words and separators, without changing the
801      interpretation of a field. At least one delimiter (LWS and/or
802      separators) &MUST; exist between any two tokens (for the definition
803      of "token" below), since they would otherwise be interpreted as a
804      single token.
805    </t>
806  </list>
810<section title="Basic Rules" anchor="basic.rules">
811<x:anchor-alias value="OCTET"/>
812<x:anchor-alias value="CHAR"/>
813<x:anchor-alias value="UPALPHA"/>
814<x:anchor-alias value="LOALPHA"/>
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="UPALPHA"/><iref primary="true" item="Grammar" subitem="LOALPHA"/><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          = &lt;any 8-bit sequence of data&gt;
840  CHAR           = &lt;any US-ASCII character (octets 0 - 127)&gt;
841  UPALPHA        = &lt;any US-ASCII uppercase letter "A".."Z"&gt;
842  LOALPHA        = &lt;any US-ASCII lowercase letter "a".."z"&gt;
843  ALPHA          = UPALPHA | LOALPHA
844  DIGIT          = &lt;any US-ASCII digit "0".."9"&gt;
845  CTL            = &lt;any US-ASCII control character
846                   (octets 0 - 31) and DEL (127)&gt;
847  CR             = &lt;US-ASCII CR, carriage return (13)&gt;
848  LF             = &lt;US-ASCII LF, linefeed (10)&gt;
849  SP             = &lt;US-ASCII SP, space (32)&gt;
850  HTAB           = &lt;US-ASCII HT, horizontal-tab (9)&gt;
851  DQUOTE         = &lt;US-ASCII double-quote mark (34)&gt;
854   HTTP/1.1 defines the sequence CR LF as the end-of-line marker for all
855   protocol elements except the entity-body (see <xref target="tolerant.applications"/> for
856   tolerant applications). The end-of-line marker within an entity-body
857   is defined by its associated media type, as described in &media-types;.
859<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="CRLF"/>
860  CRLF           = CR LF
863   HTTP/1.1 header field values can be folded onto multiple lines if the
864   continuation line begins with a space or horizontal tab. All linear
865   white space, including folding, has the same semantics as SP. A
866   recipient &MAY; replace any linear white space with a single SP before
867   interpreting the field value or forwarding the message downstream.
869<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="LWS"/>
870  LWS            = [CRLF] 1*( SP | HTAB )
873   The TEXT rule is only used for descriptive field contents and values
874   that are not intended to be interpreted by the message parser. Words
875   of *TEXT &MAY; contain characters from character sets other than ISO-8859-1
876   <xref target="ISO-8859-1"/> only when encoded according to the rules of
877   <xref target="RFC2047"/>.
879<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TEXT"/>
880  TEXT           = &lt;any OCTET except CTLs,
881                   but including LWS&gt;
884   A CRLF is allowed in the definition of TEXT only as part of a header
885   field continuation. It is expected that the folding LWS will be
886   replaced with a single SP before interpretation of the TEXT value.
889   Hexadecimal numeric characters are used in several protocol elements.
891<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HEX"/>
892  HEX            = "A" | "B" | "C" | "D" | "E" | "F"
893                 | "a" | "b" | "c" | "d" | "e" | "f" | DIGIT
896   Many HTTP/1.1 header field values consist of words separated by LWS
897   or special characters. These special characters &MUST; be in a quoted
898   string to be used within a parameter value (as defined in
899   <xref target="transfer.codings"/>).
901<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="separators"/>
902  token          = 1*&lt;any CHAR except CTLs or separators&gt;
903  separators     = "(" | ")" | "&lt;" | "&gt;" | "@"
904                 | "," | ";" | ":" | "\" | DQUOTE
905                 | "/" | "[" | "]" | "?" | "="
906                 | "{" | "}" | SP | HTAB
909   Comments can be included in some HTTP header fields by surrounding
910   the comment text with parentheses. Comments are only allowed in
911   fields containing "comment" as part of their field value definition.
912   In all other fields, parentheses are considered part of the field
913   value.
915<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
916  comment        = "(" *( ctext | quoted-pair | comment ) ")"
917  ctext          = &lt;any TEXT excluding "(" and ")"&gt;
920   A string of text is parsed as a single word if it is quoted using
921   double-quote marks.
923<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-string"/><iref primary="true" item="Grammar" subitem="qdtext"/>
924  quoted-string  = ( DQUOTE *(qdtext | quoted-pair ) DQUOTE )
925  qdtext         = &lt;any TEXT excluding DQUOTE and "\">
928   The backslash character ("\") &MAY; be used as a single-character
929   quoting mechanism only within quoted-string and comment constructs.
931<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
932  quoted-pair    = "\" CHAR
937<section title="Protocol Parameters" anchor="protocol.parameters">
939<section title="HTTP Version" anchor="http.version">
941   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate versions
942   of the protocol. The protocol versioning policy is intended to allow
943   the sender to indicate the format of a message and its capacity for
944   understanding further HTTP communication, rather than the features
945   obtained via that communication. No change is made to the version
946   number for the addition of message components which do not affect
947   communication behavior or which only add to extensible field values.
948   The &lt;minor&gt; number is incremented when the changes made to the
949   protocol add features which do not change the general message parsing
950   algorithm, but which may add to the message semantics and imply
951   additional capabilities of the sender. The &lt;major&gt; number is
952   incremented when the format of a message within the protocol is
953   changed. See <xref target="RFC2145"/> for a fuller explanation.
956   The version of an HTTP message is indicated by an HTTP-Version field
957   in the first line of the message. HTTP-Version is case-sensitive.
959<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-Version"/>
960  HTTP-Version   = "HTTP" "/" 1*DIGIT "." 1*DIGIT
963   Note that the major and minor numbers &MUST; be treated as separate
964   integers and that each &MAY; be incremented higher than a single digit.
965   Thus, HTTP/2.4 is a lower version than HTTP/2.13, which in turn is
966   lower than HTTP/12.3. Leading zeros &MUST; be ignored by recipients and
967   &MUST-NOT; be sent.
970   An application that sends a request or response message that includes
971   HTTP-Version of "HTTP/1.1" &MUST; be at least conditionally compliant
972   with this specification. Applications that are at least conditionally
973   compliant with this specification &SHOULD; use an HTTP-Version of
974   "HTTP/1.1" in their messages, and &MUST; do so for any message that is
975   not compatible with HTTP/1.0. For more details on when to send
976   specific HTTP-Version values, see <xref target="RFC2145"/>.
979   The HTTP version of an application is the highest HTTP version for
980   which the application is at least conditionally compliant.
983   Proxy and gateway applications need to be careful when forwarding
984   messages in protocol versions different from that of the application.
985   Since the protocol version indicates the protocol capability of the
986   sender, a proxy/gateway &MUST-NOT; send a message with a version
987   indicator which is greater than its actual version. If a higher
988   version request is received, the proxy/gateway &MUST; either downgrade
989   the request version, or respond with an error, or switch to tunnel
990   behavior.
993   Due to interoperability problems with HTTP/1.0 proxies discovered
994   since the publication of <xref target="RFC2068"/>, caching proxies &MUST;, gateways
995   &MAY;, and tunnels &MUST-NOT; upgrade the request to the highest version
996   they support. The proxy/gateway's response to that request &MUST; be in
997   the same major version as the request.
1000  <list>
1001    <t>
1002      <x:h>Note:</x:h> Converting between versions of HTTP may involve modification
1003      of header fields required or forbidden by the versions involved.
1004    </t>
1005  </list>
1009<section title="Uniform Resource Identifiers" anchor="uri">
1011   URIs have been known by many names: WWW addresses, Universal Document
1012   Identifiers, Universal Resource Identifiers <xref target="RFC1630"/>, and finally the
1013   combination of Uniform Resource Locators (URL) <xref target="RFC1738"/> and Names (URN)
1014   <xref target="RFC1737"/>. As far as HTTP is concerned, Uniform Resource Identifiers are
1015   simply formatted strings which identify--via name, location, or any
1016   other characteristic--a resource.
1019<section title="General Syntax" anchor="general.syntax">
1021   URIs in HTTP can be represented in absolute form or relative to some
1022   known base URI <xref target="RFC1808"/>, depending upon the context of their use. The two
1023   forms are differentiated by the fact that absolute URIs always begin
1024   with a scheme name followed by a colon. For definitive information on
1025   URL syntax and semantics, see "Uniform Resource Identifiers (URI):
1026   Generic Syntax and Semantics," <xref target="RFC2396"/> (which replaces <xref target="RFC1738"/>
1027   and <xref target="RFC1808"/>). This specification adopts the
1028   definitions of "URI-reference", "absoluteURI", "relativeURI", "port",
1029   "host", "abs_path", "query", and "authority" from that specification.
1032   HTTP does not place any a priori limit on the length of
1033   a URI. Servers &MUST; be able to handle the URI of any resource they
1034   serve, and &SHOULD; be able to handle URIs of unbounded length if they
1035   provide GET-based forms that could generate such URIs. A server
1036   &SHOULD; return 414 (Request-URI Too Long) status if a URI is longer
1037   than the server can handle (see &status-414;).
1040  <list>
1041    <t>
1042      <x:h>Note:</x:h> Servers ought to be cautious about depending on URI lengths
1043      above 255 bytes, because some older client or proxy
1044      implementations might not properly support these lengths.
1045    </t>
1046  </list>
1050<section title="http URL" anchor="http.url">
1052   The "http" scheme is used to locate network resources via the HTTP
1053   protocol. This section defines the scheme-specific syntax and
1054   semantics for http URLs.
1056<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http_URL"/>
1057  http_URL = "http:" "//" host [ ":" port ] [ abs_path [ "?" query ]]
1060   If the port is empty or not given, port 80 is assumed. The semantics
1061   are that the identified resource is located at the server listening
1062   for TCP connections on that port of that host, and the Request-URI
1063   for the resource is abs_path (<xref target="request-uri"/>). The use of IP addresses
1064   in URLs &SHOULD; be avoided whenever possible (see <xref target="RFC1900"/>). If
1065   the abs_path is not present in the URL, it &MUST; be given as "/" when
1066   used as a Request-URI for a resource (<xref target="request-uri"/>). If a proxy
1068   receives a host name which is not a fully qualified domain name, it
1069   &MAY; add its domain to the host name it received. If a proxy receives
1070   a fully qualified domain name, the proxy &MUST-NOT; change the host
1071   name.
1075<section title="URI Comparison" anchor="uri.comparison">
1077   When comparing two URIs to decide if they match or not, a client
1078   &SHOULD; use a case-sensitive octet-by-octet comparison of the entire
1079   URIs, with these exceptions:
1080  <list style="symbols">
1081    <t>A port that is empty or not given is equivalent to the default
1082        port for that URI-reference;</t>
1083    <t>Comparisons of host names &MUST; be case-insensitive;</t>
1084    <t>Comparisons of scheme names &MUST; be case-insensitive;</t>
1085    <t>An empty abs_path is equivalent to an abs_path of "/".</t>
1086  </list>
1089   Characters other than those in the "reserved" set (see
1090   <xref target="RFC2396"/>) are equivalent to their ""%" HEX HEX" encoding.
1093   For example, the following three URIs are equivalent:
1095<figure><artwork type="example">
1103<section title="Date/Time Formats" anchor="date.time.formats">
1104<section title="Full Date" anchor="">
1106   HTTP applications have historically allowed three different formats
1107   for the representation of date/time stamps:
1109<figure><artwork type="example">
1110   Sun, 06 Nov 1994 08:49:37 GMT  ; RFC 822, updated by RFC 1123
1111   Sunday, 06-Nov-94 08:49:37 GMT ; obsolete RFC 850 format
1112   Sun Nov  6 08:49:37 1994       ; ANSI C's asctime() format
1115   The first format is preferred as an Internet standard and represents
1116   a fixed-length subset of that defined by <xref target="RFC1123"/> (an update to
1117   <xref target="RFC822"/>). The other formats are described here only for
1118   compatibility with obsolete implementations.
1119   HTTP/1.1 clients and servers that parse the date value &MUST; accept
1120   all three formats (for compatibility with HTTP/1.0), though they &MUST;
1121   only generate the RFC 1123 format for representing HTTP-date values
1122   in header fields. See <xref target="tolerant.applications"/> for further information.
1125      <x:h>Note:</x:h> Recipients of date values are encouraged to be robust in
1126      accepting date values that may have been sent by non-HTTP
1127      applications, as is sometimes the case when retrieving or posting
1128      messages via proxies/gateways to SMTP or NNTP.
1131   All HTTP date/time stamps &MUST; be represented in Greenwich Mean Time
1132   (GMT), without exception. For the purposes of HTTP, GMT is exactly
1133   equal to UTC (Coordinated Universal Time). This is indicated in the
1134   first two formats by the inclusion of "GMT" as the three-letter
1135   abbreviation for time zone, and &MUST; be assumed when reading the
1136   asctime format. HTTP-date is case sensitive and &MUST-NOT; include
1137   additional LWS beyond that specifically included as SP in the
1138   grammar.
1140<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"/>
1141  HTTP-date    = rfc1123-date | rfc850-date | asctime-date
1142  rfc1123-date = wkday "," SP date1 SP time SP "GMT"
1143  rfc850-date  = weekday "," SP date2 SP time SP "GMT"
1144  asctime-date = wkday SP date3 SP time SP 4DIGIT
1145  date1        = 2DIGIT SP month SP 4DIGIT
1146                 ; day month year (e.g., 02 Jun 1982)
1147  date2        = 2DIGIT "-" month "-" 2DIGIT
1148                 ; day-month-year (e.g., 02-Jun-82)
1149  date3        = month SP ( 2DIGIT | ( SP 1DIGIT ))
1150                 ; month day (e.g., Jun  2)
1151  time         = 2DIGIT ":" 2DIGIT ":" 2DIGIT
1152                 ; 00:00:00 - 23:59:59
1153  wkday        = "Mon" | "Tue" | "Wed"
1154               | "Thu" | "Fri" | "Sat" | "Sun"
1155  weekday      = "Monday" | "Tuesday" | "Wednesday"
1156               | "Thursday" | "Friday" | "Saturday" | "Sunday"
1157  month        = "Jan" | "Feb" | "Mar" | "Apr"
1158               | "May" | "Jun" | "Jul" | "Aug"
1159               | "Sep" | "Oct" | "Nov" | "Dec"
1162      <x:h>Note:</x:h> HTTP requirements for the date/time stamp format apply only
1163      to their usage within the protocol stream. Clients and servers are
1164      not required to use these formats for user presentation, request
1165      logging, etc.
1170<section title="Transfer Codings" anchor="transfer.codings">
1172   Transfer-coding values are used to indicate an encoding
1173   transformation that has been, can be, or may need to be applied to an
1174   entity-body in order to ensure "safe transport" through the network.
1175   This differs from a content coding in that the transfer-coding is a
1176   property of the message, not of the original entity.
1178<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1179  transfer-coding         = "chunked" | transfer-extension
1180  transfer-extension      = token *( ";" parameter )
1183   Parameters are in  the form of attribute/value pairs.
1185<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"/>
1186  parameter               = attribute "=" value
1187  attribute               = token
1188  value                   = token | quoted-string
1191   All transfer-coding values are case-insensitive. HTTP/1.1 uses
1192   transfer-coding values in the TE header field (<xref target="header.te"/>) and in
1193   the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1196   Whenever a transfer-coding is applied to a message-body, the set of
1197   transfer-codings &MUST; include "chunked", unless the message is
1198   terminated by closing the connection. When the "chunked" transfer-coding
1199   is used, it &MUST; be the last transfer-coding applied to the
1200   message-body. The "chunked" transfer-coding &MUST-NOT; be applied more
1201   than once to a message-body. These rules allow the recipient to
1202   determine the transfer-length of the message (<xref target="message.length"/>).
1205   Transfer-codings are analogous to the Content-Transfer-Encoding
1206   values of MIME <xref target="RFC2045"/>, which were designed to enable safe transport of
1207   binary data over a 7-bit transport service. However, safe transport
1208   has a different focus for an 8bit-clean transfer protocol. In HTTP,
1209   the only unsafe characteristic of message-bodies is the difficulty in
1210   determining the exact body length (<xref target="message.length"/>), or the desire to
1211   encrypt data over a shared transport.
1214   The Internet Assigned Numbers Authority (IANA) acts as a registry for
1215   transfer-coding value tokens. Initially, the registry contains the
1216   following tokens: "chunked" (<xref target="chunked.transfer.encoding"/>),
1217   "gzip", "compress", and "deflate" (&content-codings;).
1220   New transfer-coding value tokens &SHOULD; be registered in the same way
1221   as new content-coding value tokens (&content-codings;).
1224   A server which receives an entity-body with a transfer-coding it does
1225   not understand &SHOULD; return 501 (Not Implemented), and close the
1226   connection. A server &MUST-NOT; send transfer-codings to an HTTP/1.0
1227   client.
1230<section title="Chunked Transfer Coding" anchor="chunked.transfer.encoding">
1232   The chunked encoding modifies the body of a message in order to
1233   transfer it as a series of chunks, each with its own size indicator,
1234   followed by an &OPTIONAL; trailer containing entity-header fields. This
1235   allows dynamically produced content to be transferred along with the
1236   information necessary for the recipient to verify that it has
1237   received the full message.
1239<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"/>
1240  Chunked-Body   = *chunk
1241                   last-chunk
1242                   trailer
1243                   CRLF
1245  chunk          = chunk-size [ chunk-extension ] CRLF
1246                   chunk-data CRLF
1247  chunk-size     = 1*HEX
1248  last-chunk     = 1*("0") [ chunk-extension ] CRLF
1250  chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
1251  chunk-ext-name = token
1252  chunk-ext-val  = token | quoted-string
1253  chunk-data     = 1*OCTET ; a sequence of chunk-size octets
1254  trailer        = *(entity-header CRLF)
1257   The chunk-size field is a string of hex digits indicating the size of
1258   the chunk-data in octets. The chunked encoding is ended by any chunk whose size is
1259   zero, followed by the trailer, which is terminated by an empty line.
1262   The trailer allows the sender to include additional HTTP header
1263   fields at the end of the message. The Trailer header field can be
1264   used to indicate which header fields are included in a trailer (see
1265   <xref target="header.trailer"/>).
1268   A server using chunked transfer-coding in a response &MUST-NOT; use the
1269   trailer for any header fields unless at least one of the following is
1270   true:
1271  <list style="numbers">
1272    <t>the request included a TE header field that indicates "trailers" is
1273     acceptable in the transfer-coding of the  response, as described in
1274     <xref target="header.te"/>; or,</t>
1276    <t>the server is the origin server for the response, the trailer
1277     fields consist entirely of optional metadata, and the recipient
1278     could use the message (in a manner acceptable to the origin server)
1279     without receiving this metadata.  In other words, the origin server
1280     is willing to accept the possibility that the trailer fields might
1281     be silently discarded along the path to the client.</t>
1282  </list>
1285   This requirement prevents an interoperability failure when the
1286   message is being received by an HTTP/1.1 (or later) proxy and
1287   forwarded to an HTTP/1.0 recipient. It avoids a situation where
1288   compliance with the protocol would have necessitated a possibly
1289   infinite buffer on the proxy.
1292   A process for decoding the "chunked" transfer-coding
1293   can be represented in pseudo-code as:
1295<figure><artwork type="code">
1296    length := 0
1297    read chunk-size, chunk-extension (if any) and CRLF
1298    while (chunk-size &gt; 0) {
1299       read chunk-data and CRLF
1300       append chunk-data to entity-body
1301       length := length + chunk-size
1302       read chunk-size and CRLF
1303    }
1304    read entity-header
1305    while (entity-header not empty) {
1306       append entity-header to existing header fields
1307       read entity-header
1308    }
1309    Content-Length := length
1310    Remove "chunked" from Transfer-Encoding
1313   All HTTP/1.1 applications &MUST; be able to receive and decode the
1314   "chunked" transfer-coding, and &MUST; ignore chunk-extension extensions
1315   they do not understand.
1322<section title="HTTP Message" anchor="http.message">
1324<section title="Message Types" anchor="message.types">
1326   HTTP messages consist of requests from client to server and responses
1327   from server to client.
1329<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"/>
1330  HTTP-message   = Request | Response     ; HTTP/1.1 messages
1333   Request (<xref target="request"/>) and Response (<xref target="response"/>) messages use the generic
1334   message format of <xref target="RFC2822"/> for transferring entities (the payload
1335   of the message). Both types of message consist of a start-line, zero
1336   or more header fields (also known as "headers"), an empty line (i.e.,
1337   a line with nothing preceding the CRLF) indicating the end of the
1338   header fields, and possibly a message-body.
1340<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="generic-message"/><iref primary="true" item="Grammar" subitem="start-line"/>
1341  generic-message = start-line
1342                    *(message-header CRLF)
1343                    CRLF
1344                    [ message-body ]
1345  start-line      = Request-Line | Status-Line
1348   In the interest of robustness, servers &SHOULD; ignore any empty
1349   line(s) received where a Request-Line is expected. In other words, if
1350   the server is reading the protocol stream at the beginning of a
1351   message and receives a CRLF first, it should ignore the CRLF.
1354   Certain buggy HTTP/1.0 client implementations generate extra CRLF's
1355   after a POST request. To restate what is explicitly forbidden by the
1356   BNF, an HTTP/1.1 client &MUST-NOT; preface or follow a request with an
1357   extra CRLF.
1361<section title="Message Headers" anchor="message.headers">
1363   HTTP header fields, which include general-header (<xref target="general.header.fields"/>),
1364   request-header (&request-header-fields;), response-header (&response-header-fields;), and
1365   entity-header (&entity-header-fields;) fields, follow the same generic format as
1366   that given in <xref target="RFC2822" x:fmt="of" x:sec="2.1"/>. Each header field consists
1367   of a name followed by a colon (":") and the field value. Field names
1368   are case-insensitive. The field value &MAY; be preceded by any amount
1369   of LWS, though a single SP is preferred. Header fields can be
1370   extended over multiple lines by preceding each extra line with at
1371   least one SP or HTAB. Applications ought to follow "common form", where
1372   one is known or indicated, when generating HTTP constructs, since
1373   there might exist some implementations that fail to accept anything
1374   beyond the common forms.
1376<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"/>
1377  message-header = field-name ":" [ field-value ]
1378  field-name     = token
1379  field-value    = *( field-content | LWS )
1380  field-content  = &lt;the OCTETs making up the field-value
1381                   and consisting of either *TEXT or combinations
1382                   of token, separators, and quoted-string&gt;
1385   The field-content does not include any leading or trailing LWS:
1386   linear white space occurring before the first non-whitespace
1387   character of the field-value or after the last non-whitespace
1388   character of the field-value. Such leading or trailing LWS &MAY; be
1389   removed without changing the semantics of the field value. Any LWS
1390   that occurs between field-content &MAY; be replaced with a single SP
1391   before interpreting the field value or forwarding the message
1392   downstream.
1395   The order in which header fields with differing field names are
1396   received is not significant. However, it is "good practice" to send
1397   general-header fields first, followed by request-header or response-header
1398   fields, and ending with the entity-header fields.
1401   Multiple message-header fields with the same field-name &MAY; be
1402   present in a message if and only if the entire field-value for that
1403   header field is defined as a comma-separated list [i.e., #(values)].
1404   It &MUST; be possible to combine the multiple header fields into one
1405   "field-name: field-value" pair, without changing the semantics of the
1406   message, by appending each subsequent field-value to the first, each
1407   separated by a comma. The order in which header fields with the same
1408   field-name are received is therefore significant to the
1409   interpretation of the combined field value, and thus a proxy &MUST-NOT;
1410   change the order of these field values when a message is forwarded.
1414<section title="Message Body" anchor="message.body">
1416   The message-body (if any) of an HTTP message is used to carry the
1417   entity-body associated with the request or response. The message-body
1418   differs from the entity-body only when a transfer-coding has been
1419   applied, as indicated by the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1421<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1422  message-body = entity-body
1423               | &lt;entity-body encoded as per Transfer-Encoding&gt;
1426   Transfer-Encoding &MUST; be used to indicate any transfer-codings
1427   applied by an application to ensure safe and proper transfer of the
1428   message. Transfer-Encoding is a property of the message, not of the
1429   entity, and thus &MAY; be added or removed by any application along the
1430   request/response chain. (However, <xref target="transfer.codings"/> places restrictions on
1431   when certain transfer-codings may be used.)
1434   The rules for when a message-body is allowed in a message differ for
1435   requests and responses.
1438   The presence of a message-body in a request is signaled by the
1439   inclusion of a Content-Length or Transfer-Encoding header field in
1440   the request's message-headers. A message-body &MUST-NOT; be included in
1441   a request if the specification of the request method (&method;)
1442   explicitly disallows an entity-body in requests.
1443   When a request message contains both a message-body of non-zero
1444   length and a method that does not define any semantics for that
1445   request message-body, then an origin server &SHOULD; either ignore
1446   the message-body or respond with an appropriate error message
1447   (e.g., 413).  A proxy or gateway, when presented the same request,
1448   &SHOULD; either forward the request inbound with the message-body or
1449   ignore the message-body when determining a response.
1452   For response messages, whether or not a message-body is included with
1453   a message is dependent on both the request method and the response
1454   status code (<xref target="status.code.and.reason.phrase"/>). All responses to the HEAD request method
1455   &MUST-NOT; include a message-body, even though the presence of entity-header
1456   fields might lead one to believe they do. All 1xx
1457   (informational), 204 (No Content), and 304 (Not Modified) responses
1458   &MUST-NOT; include a message-body. All other responses do include a
1459   message-body, although it &MAY; be of zero length.
1463<section title="Message Length" anchor="message.length">
1465   The transfer-length of a message is the length of the message-body as
1466   it appears in the message; that is, after any transfer-codings have
1467   been applied. When a message-body is included with a message, the
1468   transfer-length of that body is determined by one of the following
1469   (in order of precedence):
1472  <list style="numbers">
1473    <x:lt><t>
1474     Any response message which "&MUST-NOT;" include a message-body (such
1475     as the 1xx, 204, and 304 responses and any response to a HEAD
1476     request) is always terminated by the first empty line after the
1477     header fields, regardless of the entity-header fields present in
1478     the message.
1479    </t></x:lt>
1480    <x:lt><t>
1481     If a Transfer-Encoding header field (<xref target="header.transfer-encoding"/>)
1482     is present, then the transfer-length is
1483     defined by use of the "chunked" transfer-coding (<xref target="transfer.codings"/>),
1484     unless the message is terminated by closing the connection.
1485    </t></x:lt>
1486    <x:lt><t>
1487     If a Content-Length header field (<xref target="header.content-length"/>) is present, its
1488     decimal value in OCTETs represents both the entity-length and the
1489     transfer-length. The Content-Length header field &MUST-NOT; be sent
1490     if these two lengths are different (i.e., if a Transfer-Encoding
1491     header field is present). If a message is received with both a
1492     Transfer-Encoding header field and a Content-Length header field,
1493     the latter &MUST; be ignored.
1494    </t></x:lt>
1495    <x:lt><t>
1496     If the message uses the media type "multipart/byteranges", and the
1497     transfer-length is not otherwise specified, then this self-delimiting
1498     media type defines the transfer-length. This media type
1499     &MUST-NOT; be used unless the sender knows that the recipient can parse
1500     it; the presence in a request of a Range header with multiple byte-range
1501     specifiers from a 1.1 client implies that the client can parse
1502     multipart/byteranges responses.
1503    <list style="empty"><t>
1504       A range header might be forwarded by a 1.0 proxy that does not
1505       understand multipart/byteranges; in this case the server &MUST;
1506       delimit the message using methods defined in items 1, 3 or 5 of
1507       this section.
1508    </t></list>
1509    </t></x:lt>
1510    <x:lt><t>
1511     By the server closing the connection. (Closing the connection
1512     cannot be used to indicate the end of a request body, since that
1513     would leave no possibility for the server to send back a response.)
1514    </t></x:lt>
1515  </list>
1518   For compatibility with HTTP/1.0 applications, HTTP/1.1 requests
1519   containing a message-body &MUST; include a valid Content-Length header
1520   field unless the server is known to be HTTP/1.1 compliant. If a
1521   request contains a message-body and a Content-Length is not given,
1522   the server &SHOULD; respond with 400 (Bad Request) if it cannot
1523   determine the length of the message, or with 411 (Length Required) if
1524   it wishes to insist on receiving a valid Content-Length.
1527   All HTTP/1.1 applications that receive entities &MUST; accept the
1528   "chunked" transfer-coding (<xref target="transfer.codings"/>), thus allowing this mechanism
1529   to be used for messages when the message length cannot be determined
1530   in advance.
1533   Messages &MUST-NOT; include both a Content-Length header field and a
1534   transfer-coding. If the message does include a
1535   transfer-coding, the Content-Length &MUST; be ignored.
1538   When a Content-Length is given in a message where a message-body is
1539   allowed, its field value &MUST; exactly match the number of OCTETs in
1540   the message-body. HTTP/1.1 user agents &MUST; notify the user when an
1541   invalid length is received and detected.
1545<section title="General Header Fields" anchor="general.header.fields">
1547   There are a few header fields which have general applicability for
1548   both request and response messages, but which do not apply to the
1549   entity being transferred. These header fields apply only to the
1550   message being transmitted.
1552<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="general-header"/>
1553  general-header = Cache-Control            ; &header-cache-control;
1554                 | Connection               ; <xref target="header.connection"/>
1555                 | Date                     ; <xref target=""/>
1556                 | Pragma                   ; &header-pragma;
1557                 | Trailer                  ; <xref target="header.trailer"/>
1558                 | Transfer-Encoding        ; <xref target="header.transfer-encoding"/>
1559                 | Upgrade                  ; <xref target="header.upgrade"/>
1560                 | Via                      ; <xref target="header.via"/>
1561                 | Warning                  ; &header-warning;
1564   General-header field names can be extended reliably only in
1565   combination with a change in the protocol version. However, new or
1566   experimental header fields may be given the semantics of general
1567   header fields if all parties in the communication recognize them to
1568   be general-header fields. Unrecognized header fields are treated as
1569   entity-header fields.
1574<section title="Request" anchor="request">
1576   A request message from a client to a server includes, within the
1577   first line of that message, the method to be applied to the resource,
1578   the identifier of the resource, and the protocol version in use.
1580<!--                 Host                      ; should be moved here eventually -->
1581<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request"/>
1582  Request       = Request-Line              ; <xref target="request-line"/>
1583                  *(( general-header        ; <xref target="general.header.fields"/>
1584                   | request-header         ; &request-header-fields;
1585                   | entity-header ) CRLF)  ; &entity-header-fields;
1586                  CRLF
1587                  [ message-body ]          ; <xref target="message.body"/>
1590<section title="Request-Line" anchor="request-line">
1592   The Request-Line begins with a method token, followed by the
1593   Request-URI and the protocol version, and ending with CRLF. The
1594   elements are separated by SP characters. No CR or LF is allowed
1595   except in the final CRLF sequence.
1597<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-Line"/>
1598  Request-Line   = Method SP Request-URI SP HTTP-Version CRLF
1601<section title="Method" anchor="method">
1603   The Method  token indicates the method to be performed on the
1604   resource identified by the Request-URI. The method is case-sensitive.
1606<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Method"/><iref primary="true" item="Grammar" subitem="extension-method"/>
1607  Method         = token
1611<section title="Request-URI" anchor="request-uri">
1613   The Request-URI is a Uniform Resource Identifier (<xref target="uri"/>) and
1614   identifies the resource upon which to apply the request.
1616<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-URI"/>
1617  Request-URI    = "*"
1618                 | absoluteURI
1619                 | ( abs_path [ "?" query ] )
1620                 | authority
1623   The four options for Request-URI are dependent on the nature of the
1624   request. The asterisk "*" means that the request does not apply to a
1625   particular resource, but to the server itself, and is only allowed
1626   when the method used does not necessarily apply to a resource. One
1627   example would be
1629<figure><artwork type="example">
1630    OPTIONS * HTTP/1.1
1633   The absoluteURI form is &REQUIRED; when the request is being made to a
1634   proxy. The proxy is requested to forward the request or service it
1635   from a valid cache, and return the response. Note that the proxy &MAY;
1636   forward the request on to another proxy or directly to the server
1637   specified by the absoluteURI. In order to avoid request loops, a
1638   proxy &MUST; be able to recognize all of its server names, including
1639   any aliases, local variations, and the numeric IP address. An example
1640   Request-Line would be:
1642<figure><artwork type="example">
1643    GET HTTP/1.1
1646   To allow for transition to absoluteURIs in all requests in future
1647   versions of HTTP, all HTTP/1.1 servers &MUST; accept the absoluteURI
1648   form in requests, even though HTTP/1.1 clients will only generate
1649   them in requests to proxies.
1652   The authority form is only used by the CONNECT method (&CONNECT;).
1655   The most common form of Request-URI is that used to identify a
1656   resource on an origin server or gateway. In this case the absolute
1657   path of the URI &MUST; be transmitted (see <xref target="general.syntax"/>, abs_path) as
1658   the Request-URI, and the network location of the URI (authority) &MUST;
1659   be transmitted in a Host header field. For example, a client wishing
1660   to retrieve the resource above directly from the origin server would
1661   create a TCP connection to port 80 of the host "" and send
1662   the lines:
1664<figure><artwork type="example">
1665    GET /pub/WWW/TheProject.html HTTP/1.1
1666    Host:
1669   followed by the remainder of the Request. Note that the absolute path
1670   cannot be empty; if none is present in the original URI, it &MUST; be
1671   given as "/" (the server root).
1674   The Request-URI is transmitted in the format specified in
1675   <xref target="general.syntax"/>. If the Request-URI is encoded using the "% HEX HEX" encoding
1676   <xref target="RFC2396"/>, the origin server &MUST; decode the Request-URI in order to
1677   properly interpret the request. Servers &SHOULD; respond to invalid
1678   Request-URIs with an appropriate status code.
1681   A transparent proxy &MUST-NOT; rewrite the "abs_path" part of the
1682   received Request-URI when forwarding it to the next inbound server,
1683   except as noted above to replace a null abs_path with "/".
1686  <list><t>
1687      <x:h>Note:</x:h> The "no rewrite" rule prevents the proxy from changing the
1688      meaning of the request when the origin server is improperly using
1689      a non-reserved URI character for a reserved purpose.  Implementors
1690      should be aware that some pre-HTTP/1.1 proxies have been known to
1691      rewrite the Request-URI.
1692  </t></list>
1697<section title="The Resource Identified by a Request" anchor="">
1699   The exact resource identified by an Internet request is determined by
1700   examining both the Request-URI and the Host header field.
1703   An origin server that does not allow resources to differ by the
1704   requested host &MAY; ignore the Host header field value when
1705   determining the resource identified by an HTTP/1.1 request. (But see
1706   <xref target=""/>
1707   for other requirements on Host support in HTTP/1.1.)
1710   An origin server that does differentiate resources based on the host
1711   requested (sometimes referred to as virtual hosts or vanity host
1712   names) &MUST; use the following rules for determining the requested
1713   resource on an HTTP/1.1 request:
1714  <list style="numbers">
1715    <t>If Request-URI is an absoluteURI, the host is part of the
1716     Request-URI. Any Host header field value in the request &MUST; be
1717     ignored.</t>
1718    <t>If the Request-URI is not an absoluteURI, and the request includes
1719     a Host header field, the host is determined by the Host header
1720     field value.</t>
1721    <t>If the host as determined by rule 1 or 2 is not a valid host on
1722     the server, the response &MUST; be a 400 (Bad Request) error message.</t>
1723  </list>
1726   Recipients of an HTTP/1.0 request that lacks a Host header field &MAY;
1727   attempt to use heuristics (e.g., examination of the URI path for
1728   something unique to a particular host) in order to determine what
1729   exact resource is being requested.
1736<section title="Response" anchor="response">
1738   After receiving and interpreting a request message, a server responds
1739   with an HTTP response message.
1741<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Response"/>
1742  Response      = Status-Line               ; <xref target="status-line"/>
1743                  *(( general-header        ; <xref target="general.header.fields"/>
1744                   | response-header        ; &response-header-fields;
1745                   | entity-header ) CRLF)  ; &entity-header-fields;
1746                  CRLF
1747                  [ message-body ]          ; <xref target="message.body"/>
1750<section title="Status-Line" anchor="status-line">
1752   The first line of a Response message is the Status-Line, consisting
1753   of the protocol version followed by a numeric status code and its
1754   associated textual phrase, with each element separated by SP
1755   characters. No CR or LF is allowed except in the final CRLF sequence.
1757<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Line"/>
1758  Status-Line = HTTP-Version SP Status-Code SP Reason-Phrase CRLF
1761<section title="Status Code and Reason Phrase" anchor="status.code.and.reason.phrase">
1763   The Status-Code element is a 3-digit integer result code of the
1764   attempt to understand and satisfy the request. These codes are fully
1765   defined in &status-codes;. The Reason-Phrase is intended to give a short
1766   textual description of the Status-Code. The Status-Code is intended
1767   for use by automata and the Reason-Phrase is intended for the human
1768   user. The client is not required to examine or display the Reason-Phrase.
1771   The first digit of the Status-Code defines the class of response. The
1772   last two digits do not have any categorization role. There are 5
1773   values for the first digit:
1774  <list style="symbols">
1775    <t>
1776      1xx: Informational - Request received, continuing process
1777    </t>
1778    <t>
1779      2xx: Success - The action was successfully received,
1780        understood, and accepted
1781    </t>
1782    <t>
1783      3xx: Redirection - Further action must be taken in order to
1784        complete the request
1785    </t>
1786    <t>
1787      4xx: Client Error - The request contains bad syntax or cannot
1788        be fulfilled
1789    </t>
1790    <t>
1791      5xx: Server Error - The server failed to fulfill an apparently
1792        valid request
1793    </t>
1794  </list>
1796<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"/>
1797  Status-Code    = 3DIGIT
1798  Reason-Phrase  = *&lt;TEXT, excluding CR, LF&gt;
1806<section title="Connections" anchor="connections">
1808<section title="Persistent Connections" anchor="persistent.connections">
1810<section title="Purpose" anchor="persistent.purpose">
1812   Prior to persistent connections, a separate TCP connection was
1813   established to fetch each URL, increasing the load on HTTP servers
1814   and causing congestion on the Internet. The use of inline images and
1815   other associated data often require a client to make multiple
1816   requests of the same server in a short amount of time. Analysis of
1817   these performance problems and results from a prototype
1818   implementation are available <xref target="Pad1995"/> <xref target="Spe"/>. Implementation experience and
1819   measurements of actual HTTP/1.1 (<xref target="RFC2068" x:fmt="none">RFC 2068</xref>) implementations show good
1820   results <xref target="Nie1997"/>. Alternatives have also been explored, for example,
1821   T/TCP <xref target="Tou1998"/>.
1824   Persistent HTTP connections have a number of advantages:
1825  <list style="symbols">
1826      <t>
1827        By opening and closing fewer TCP connections, CPU time is saved
1828        in routers and hosts (clients, servers, proxies, gateways,
1829        tunnels, or caches), and memory used for TCP protocol control
1830        blocks can be saved in hosts.
1831      </t>
1832      <t>
1833        HTTP requests and responses can be pipelined on a connection.
1834        Pipelining allows a client to make multiple requests without
1835        waiting for each response, allowing a single TCP connection to
1836        be used much more efficiently, with much lower elapsed time.
1837      </t>
1838      <t>
1839        Network congestion is reduced by reducing the number of packets
1840        caused by TCP opens, and by allowing TCP sufficient time to
1841        determine the congestion state of the network.
1842      </t>
1843      <t>
1844        Latency on subsequent requests is reduced since there is no time
1845        spent in TCP's connection opening handshake.
1846      </t>
1847      <t>
1848        HTTP can evolve more gracefully, since errors can be reported
1849        without the penalty of closing the TCP connection. Clients using
1850        future versions of HTTP might optimistically try a new feature,
1851        but if communicating with an older server, retry with old
1852        semantics after an error is reported.
1853      </t>
1854    </list>
1857   HTTP implementations &SHOULD; implement persistent connections.
1861<section title="Overall Operation" anchor="persistent.overall">
1863   A significant difference between HTTP/1.1 and earlier versions of
1864   HTTP is that persistent connections are the default behavior of any
1865   HTTP connection. That is, unless otherwise indicated, the client
1866   &SHOULD; assume that the server will maintain a persistent connection,
1867   even after error responses from the server.
1870   Persistent connections provide a mechanism by which a client and a
1871   server can signal the close of a TCP connection. This signaling takes
1872   place using the Connection header field (<xref target="header.connection"/>). Once a close
1873   has been signaled, the client &MUST-NOT; send any more requests on that
1874   connection.
1877<section title="Negotiation" anchor="persistent.negotiation">
1879   An HTTP/1.1 server &MAY; assume that a HTTP/1.1 client intends to
1880   maintain a persistent connection unless a Connection header including
1881   the connection-token "close" was sent in the request. If the server
1882   chooses to close the connection immediately after sending the
1883   response, it &SHOULD; send a Connection header including the
1884   connection-token close.
1887   An HTTP/1.1 client &MAY; expect a connection to remain open, but would
1888   decide to keep it open based on whether the response from a server
1889   contains a Connection header with the connection-token close. In case
1890   the client does not want to maintain a connection for more than that
1891   request, it &SHOULD; send a Connection header including the
1892   connection-token close.
1895   If either the client or the server sends the close token in the
1896   Connection header, that request becomes the last one for the
1897   connection.
1900   Clients and servers &SHOULD-NOT;  assume that a persistent connection is
1901   maintained for HTTP versions less than 1.1 unless it is explicitly
1902   signaled. See <xref target="compatibility.with.http.1.0.persistent.connections"/> for more information on backward
1903   compatibility with HTTP/1.0 clients.
1906   In order to remain persistent, all messages on the connection &MUST;
1907   have a self-defined message length (i.e., one not defined by closure
1908   of the connection), as described in <xref target="message.length"/>.
1912<section title="Pipelining" anchor="pipelining">
1914   A client that supports persistent connections &MAY; "pipeline" its
1915   requests (i.e., send multiple requests without waiting for each
1916   response). A server &MUST; send its responses to those requests in the
1917   same order that the requests were received.
1920   Clients which assume persistent connections and pipeline immediately
1921   after connection establishment &SHOULD; be prepared to retry their
1922   connection if the first pipelined attempt fails. If a client does
1923   such a retry, it &MUST-NOT; pipeline before it knows the connection is
1924   persistent. Clients &MUST; also be prepared to resend their requests if
1925   the server closes the connection before sending all of the
1926   corresponding responses.
1929   Clients &SHOULD-NOT;  pipeline requests using non-idempotent methods or
1930   non-idempotent sequences of methods (see &idempotent-methods;). Otherwise, a
1931   premature termination of the transport connection could lead to
1932   indeterminate results. A client wishing to send a non-idempotent
1933   request &SHOULD; wait to send that request until it has received the
1934   response status for the previous request.
1939<section title="Proxy Servers" anchor="persistent.proxy">
1941   It is especially important that proxies correctly implement the
1942   properties of the Connection header field as specified in <xref target="header.connection"/>.
1945   The proxy server &MUST; signal persistent connections separately with
1946   its clients and the origin servers (or other proxy servers) that it
1947   connects to. Each persistent connection applies to only one transport
1948   link.
1951   A proxy server &MUST-NOT; establish a HTTP/1.1 persistent connection
1952   with an HTTP/1.0 client (but see <xref target="RFC2068"/> for information and
1953   discussion of the problems with the Keep-Alive header implemented by
1954   many HTTP/1.0 clients).
1958<section title="Practical Considerations" anchor="persistent.practical">
1960   Servers will usually have some time-out value beyond which they will
1961   no longer maintain an inactive connection. Proxy servers might make
1962   this a higher value since it is likely that the client will be making
1963   more connections through the same server. The use of persistent
1964   connections places no requirements on the length (or existence) of
1965   this time-out for either the client or the server.
1968   When a client or server wishes to time-out it &SHOULD; issue a graceful
1969   close on the transport connection. Clients and servers &SHOULD; both
1970   constantly watch for the other side of the transport close, and
1971   respond to it as appropriate. If a client or server does not detect
1972   the other side's close promptly it could cause unnecessary resource
1973   drain on the network.
1976   A client, server, or proxy &MAY; close the transport connection at any
1977   time. For example, a client might have started to send a new request
1978   at the same time that the server has decided to close the "idle"
1979   connection. From the server's point of view, the connection is being
1980   closed while it was idle, but from the client's point of view, a
1981   request is in progress.
1984   This means that clients, servers, and proxies &MUST; be able to recover
1985   from asynchronous close events. Client software &SHOULD; reopen the
1986   transport connection and retransmit the aborted sequence of requests
1987   without user interaction so long as the request sequence is
1988   idempotent (see &idempotent-methods;). Non-idempotent methods or sequences
1989   &MUST-NOT; be automatically retried, although user agents &MAY; offer a
1990   human operator the choice of retrying the request(s). Confirmation by
1991   user-agent software with semantic understanding of the application
1992   &MAY; substitute for user confirmation. The automatic retry &SHOULD-NOT;
1993   be repeated if the second sequence of requests fails.
1996   Servers &SHOULD; always respond to at least one request per connection,
1997   if at all possible. Servers &SHOULD-NOT;  close a connection in the
1998   middle of transmitting a response, unless a network or client failure
1999   is suspected.
2002   Clients that use persistent connections &SHOULD; limit the number of
2003   simultaneous connections that they maintain to a given server. A
2004   single-user client &SHOULD-NOT; maintain more than 2 connections with
2005   any server or proxy. A proxy &SHOULD; use up to 2*N connections to
2006   another server or proxy, where N is the number of simultaneously
2007   active users. These guidelines are intended to improve HTTP response
2008   times and avoid congestion.
2013<section title="Message Transmission Requirements" anchor="message.transmission.requirements">
2015<section title="Persistent Connections and Flow Control" anchor="persistent.flow">
2017   HTTP/1.1 servers &SHOULD; maintain persistent connections and use TCP's
2018   flow control mechanisms to resolve temporary overloads, rather than
2019   terminating connections with the expectation that clients will retry.
2020   The latter technique can exacerbate network congestion.
2024<section title="Monitoring Connections for Error Status Messages" anchor="persistent.monitor">
2026   An HTTP/1.1 (or later) client sending a message-body &SHOULD; monitor
2027   the network connection for an error status while it is transmitting
2028   the request. If the client sees an error status, it &SHOULD;
2029   immediately cease transmitting the body. If the body is being sent
2030   using a "chunked" encoding (<xref target="transfer.codings"/>), a zero length chunk and
2031   empty trailer &MAY; be used to prematurely mark the end of the message.
2032   If the body was preceded by a Content-Length header, the client &MUST;
2033   close the connection.
2037<section title="Use of the 100 (Continue) Status" anchor="use.of.the.100.status">
2039   The purpose of the 100 (Continue) status (see &status-100;) is to
2040   allow a client that is sending a request message with a request body
2041   to determine if the origin server is willing to accept the request
2042   (based on the request headers) before the client sends the request
2043   body. In some cases, it might either be inappropriate or highly
2044   inefficient for the client to send the body if the server will reject
2045   the message without looking at the body.
2048   Requirements for HTTP/1.1 clients:
2049  <list style="symbols">
2050    <t>
2051        If a client will wait for a 100 (Continue) response before
2052        sending the request body, it &MUST; send an Expect request-header
2053        field (&header-expect;) with the "100-continue" expectation.
2054    </t>
2055    <t>
2056        A client &MUST-NOT; send an Expect request-header field (&header-expect;)
2057        with the "100-continue" expectation if it does not intend
2058        to send a request body.
2059    </t>
2060  </list>
2063   Because of the presence of older implementations, the protocol allows
2064   ambiguous situations in which a client may send "Expect: 100-continue"
2065   without receiving either a 417 (Expectation Failed) status
2066   or a 100 (Continue) status. Therefore, when a client sends this
2067   header field to an origin server (possibly via a proxy) from which it
2068   has never seen a 100 (Continue) status, the client &SHOULD-NOT;  wait
2069   for an indefinite period before sending the request body.
2072   Requirements for HTTP/1.1 origin servers:
2073  <list style="symbols">
2074    <t> Upon receiving a request which includes an Expect request-header
2075        field with the "100-continue" expectation, an origin server &MUST;
2076        either respond with 100 (Continue) status and continue to read
2077        from the input stream, or respond with a final status code. The
2078        origin server &MUST-NOT; wait for the request body before sending
2079        the 100 (Continue) response. If it responds with a final status
2080        code, it &MAY; close the transport connection or it &MAY; continue
2081        to read and discard the rest of the request.  It &MUST-NOT;
2082        perform the requested method if it returns a final status code.
2083    </t>
2084    <t> An origin server &SHOULD-NOT;  send a 100 (Continue) response if
2085        the request message does not include an Expect request-header
2086        field with the "100-continue" expectation, and &MUST-NOT; send a
2087        100 (Continue) response if such a request comes from an HTTP/1.0
2088        (or earlier) client. There is an exception to this rule: for
2089        compatibility with <xref target="RFC2068"/>, a server &MAY; send a 100 (Continue)
2090        status in response to an HTTP/1.1 PUT or POST request that does
2091        not include an Expect request-header field with the "100-continue"
2092        expectation. This exception, the purpose of which is
2093        to minimize any client processing delays associated with an
2094        undeclared wait for 100 (Continue) status, applies only to
2095        HTTP/1.1 requests, and not to requests with any other HTTP-version
2096        value.
2097    </t>
2098    <t> An origin server &MAY; omit a 100 (Continue) response if it has
2099        already received some or all of the request body for the
2100        corresponding request.
2101    </t>
2102    <t> An origin server that sends a 100 (Continue) response &MUST;
2103    ultimately send a final status code, once the request body is
2104        received and processed, unless it terminates the transport
2105        connection prematurely.
2106    </t>
2107    <t> If an origin server receives a request that does not include an
2108        Expect request-header field with the "100-continue" expectation,
2109        the request includes a request body, and the server responds
2110        with a final status code before reading the entire request body
2111        from the transport connection, then the server &SHOULD-NOT;  close
2112        the transport connection until it has read the entire request,
2113        or until the client closes the connection. Otherwise, the client
2114        might not reliably receive the response message. However, this
2115        requirement is not be construed as preventing a server from
2116        defending itself against denial-of-service attacks, or from
2117        badly broken client implementations.
2118      </t>
2119    </list>
2122   Requirements for HTTP/1.1 proxies:
2123  <list style="symbols">
2124    <t> If a proxy receives a request that includes an Expect request-header
2125        field with the "100-continue" expectation, and the proxy
2126        either knows that the next-hop server complies with HTTP/1.1 or
2127        higher, or does not know the HTTP version of the next-hop
2128        server, it &MUST; forward the request, including the Expect header
2129        field.
2130    </t>
2131    <t> If the proxy knows that the version of the next-hop server is
2132        HTTP/1.0 or lower, it &MUST-NOT; forward the request, and it &MUST;
2133        respond with a 417 (Expectation Failed) status.
2134    </t>
2135    <t> Proxies &SHOULD; maintain a cache recording the HTTP version
2136        numbers received from recently-referenced next-hop servers.
2137    </t>
2138    <t> A proxy &MUST-NOT; forward a 100 (Continue) response if the
2139        request message was received from an HTTP/1.0 (or earlier)
2140        client and did not include an Expect request-header field with
2141        the "100-continue" expectation. This requirement overrides the
2142        general rule for forwarding of 1xx responses (see &status-1xx;).
2143    </t>
2144  </list>
2148<section title="Client Behavior if Server Prematurely Closes Connection" anchor="connection.premature">
2150   If an HTTP/1.1 client sends a request which includes a request body,
2151   but which does not include an Expect request-header field with the
2152   "100-continue" expectation, and if the client is not directly
2153   connected to an HTTP/1.1 origin server, and if the client sees the
2154   connection close before receiving any status from the server, the
2155   client &SHOULD; retry the request.  If the client does retry this
2156   request, it &MAY; use the following "binary exponential backoff"
2157   algorithm to be assured of obtaining a reliable response:
2158  <list style="numbers">
2159    <t>
2160      Initiate a new connection to the server
2161    </t>
2162    <t>
2163      Transmit the request-headers
2164    </t>
2165    <t>
2166      Initialize a variable R to the estimated round-trip time to the
2167         server (e.g., based on the time it took to establish the
2168         connection), or to a constant value of 5 seconds if the round-trip
2169         time is not available.
2170    </t>
2171    <t>
2172       Compute T = R * (2**N), where N is the number of previous
2173         retries of this request.
2174    </t>
2175    <t>
2176       Wait either for an error response from the server, or for T
2177         seconds (whichever comes first)
2178    </t>
2179    <t>
2180       If no error response is received, after T seconds transmit the
2181         body of the request.
2182    </t>
2183    <t>
2184       If client sees that the connection is closed prematurely,
2185         repeat from step 1 until the request is accepted, an error
2186         response is received, or the user becomes impatient and
2187         terminates the retry process.
2188    </t>
2189  </list>
2192   If at any point an error status is received, the client
2193  <list style="symbols">
2194      <t>&SHOULD-NOT;  continue and</t>
2196      <t>&SHOULD; close the connection if it has not completed sending the
2197        request message.</t>
2198    </list>
2205<section title="Header Field Definitions" anchor="header.fields">
2207   This section defines the syntax and semantics of HTTP/1.1 header fields
2208   related to message framing and transport protocols.
2211   For entity-header fields, both sender and recipient refer to either the
2212   client or the server, depending on who sends and who receives the entity.
2215<section title="Connection" anchor="header.connection">
2216  <iref primary="true" item="Connection header" x:for-anchor=""/>
2217  <iref primary="true" item="Headers" subitem="Connection" x:for-anchor=""/>
2219   The Connection general-header field allows the sender to specify
2220   options that are desired for that particular connection and &MUST-NOT;
2221   be communicated by proxies over further connections.
2224   The Connection header has the following grammar:
2226<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-token"/>
2227  Connection = "Connection" ":" 1#(connection-token)
2228  connection-token  = token
2231   HTTP/1.1 proxies &MUST; parse the Connection header field before a
2232   message is forwarded and, for each connection-token in this field,
2233   remove any header field(s) from the message with the same name as the
2234   connection-token. Connection options are signaled by the presence of
2235   a connection-token in the Connection header field, not by any
2236   corresponding additional header field(s), since the additional header
2237   field may not be sent if there are no parameters associated with that
2238   connection option.
2241   Message headers listed in the Connection header &MUST-NOT; include
2242   end-to-end headers, such as Cache-Control.
2245   HTTP/1.1 defines the "close" connection option for the sender to
2246   signal that the connection will be closed after completion of the
2247   response. For example,
2249<figure><artwork type="example">
2250    Connection: close
2253   in either the request or the response header fields indicates that
2254   the connection &SHOULD-NOT;  be considered `persistent' (<xref target="persistent.connections"/>)
2255   after the current request/response is complete.
2258   An HTTP/1.1 client that does not support persistent connections &MUST;
2259   include the "close" connection option in every request message.
2262   An HTTP/1.1 server that does not support persistent connections &MUST;
2263   include the "close" connection option in every response message that
2264   does not have a 1xx (informational) status code.
2267   A system receiving an HTTP/1.0 (or lower-version) message that
2268   includes a Connection header &MUST;, for each connection-token in this
2269   field, remove and ignore any header field(s) from the message with
2270   the same name as the connection-token. This protects against mistaken
2271   forwarding of such header fields by pre-HTTP/1.1 proxies. See <xref target="compatibility.with.http.1.0.persistent.connections"/>.
2275<section title="Content-Length" anchor="header.content-length">
2276  <iref primary="true" item="Content-Length header" x:for-anchor=""/>
2277  <iref primary="true" item="Headers" subitem="Content-Length" x:for-anchor=""/>
2279   The Content-Length entity-header field indicates the size of the
2280   entity-body, in decimal number of OCTETs, sent to the recipient or,
2281   in the case of the HEAD method, the size of the entity-body that
2282   would have been sent had the request been a GET.
2284<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
2285  Content-Length    = "Content-Length" ":" 1*DIGIT
2288   An example is
2290<figure><artwork type="example">
2291    Content-Length: 3495
2294   Applications &SHOULD; use this field to indicate the transfer-length of
2295   the message-body, unless this is prohibited by the rules in <xref target="message.length"/>.
2298   Any Content-Length greater than or equal to zero is a valid value.
2299   <xref target="message.length"/> describes how to determine the length of a message-body
2300   if a Content-Length is not given.
2303   Note that the meaning of this field is significantly different from
2304   the corresponding definition in MIME, where it is an optional field
2305   used within the "message/external-body" content-type. In HTTP, it
2306   &SHOULD; be sent whenever the message's length can be determined prior
2307   to being transferred, unless this is prohibited by the rules in
2308   <xref target="message.length"/>.
2312<section title="Date" anchor="">
2313  <iref primary="true" item="Date header" x:for-anchor=""/>
2314  <iref primary="true" item="Headers" subitem="Date" x:for-anchor=""/>
2316   The Date general-header field represents the date and time at which
2317   the message was originated, having the same semantics as orig-date in
2318   <xref target="RFC2822" x:fmt="of" x:sec="3.6.1"/>. The field value is an HTTP-date, as described in <xref target=""/>;
2319   it &MUST; be sent in rfc1123-date format.
2321<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Date"/>
2322  Date  = "Date" ":" HTTP-date
2325   An example is
2327<figure><artwork type="example">
2328    Date: Tue, 15 Nov 1994 08:12:31 GMT
2331   Origin servers &MUST; include a Date header field in all responses,
2332   except in these cases:
2333  <list style="numbers">
2334      <t>If the response status code is 100 (Continue) or 101 (Switching
2335         Protocols), the response &MAY; include a Date header field, at
2336         the server's option.</t>
2338      <t>If the response status code conveys a server error, e.g. 500
2339         (Internal Server Error) or 503 (Service Unavailable), and it is
2340         inconvenient or impossible to generate a valid Date.</t>
2342      <t>If the server does not have a clock that can provide a
2343         reasonable approximation of the current time, its responses
2344         &MUST-NOT; include a Date header field. In this case, the rules
2345         in <xref target="clockless.origin.server.operation"/> &MUST; be followed.</t>
2346  </list>
2349   A received message that does not have a Date header field &MUST; be
2350   assigned one by the recipient if the message will be cached by that
2351   recipient or gatewayed via a protocol which requires a Date. An HTTP
2352   implementation without a clock &MUST-NOT; cache responses without
2353   revalidating them on every use. An HTTP cache, especially a shared
2354   cache, &SHOULD; use a mechanism, such as NTP <xref target="RFC1305"/>, to synchronize its
2355   clock with a reliable external standard.
2358   Clients &SHOULD; only send a Date header field in messages that include
2359   an entity-body, as in the case of the PUT and POST requests, and even
2360   then it is optional. A client without a clock &MUST-NOT; send a Date
2361   header field in a request.
2364   The HTTP-date sent in a Date header &SHOULD-NOT;  represent a date and
2365   time subsequent to the generation of the message. It &SHOULD; represent
2366   the best available approximation of the date and time of message
2367   generation, unless the implementation has no means of generating a
2368   reasonably accurate date and time. In theory, the date ought to
2369   represent the moment just before the entity is generated. In
2370   practice, the date can be generated at any time during the message
2371   origination without affecting its semantic value.
2374<section title="Clockless Origin Server Operation" anchor="clockless.origin.server.operation">
2376   Some origin server implementations might not have a clock available.
2377   An origin server without a clock &MUST-NOT; assign Expires or Last-Modified
2378   values to a response, unless these values were associated
2379   with the resource by a system or user with a reliable clock. It &MAY;
2380   assign an Expires value that is known, at or before server
2381   configuration time, to be in the past (this allows "pre-expiration"
2382   of responses without storing separate Expires values for each
2383   resource).
2388<section title="Host" anchor="">
2389  <iref primary="true" item="Host header" x:for-anchor=""/>
2390  <iref primary="true" item="Headers" subitem="Host" x:for-anchor=""/>
2392   The Host request-header field specifies the Internet host and port
2393   number of the resource being requested, as obtained from the original
2394   URI given by the user or referring resource (generally an HTTP URL,
2395   as described in <xref target="http.url"/>). The Host field value &MUST; represent
2396   the naming authority of the origin server or gateway given by the
2397   original URL. This allows the origin server or gateway to
2398   differentiate between internally-ambiguous URLs, such as the root "/"
2399   URL of a server for multiple host names on a single IP address.
2401<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2402  Host = "Host" ":" host [ ":" port ] ; <xref target="http.url"/>
2405   A "host" without any trailing port information implies the default
2406   port for the service requested (e.g., "80" for an HTTP URL). For
2407   example, a request on the origin server for
2408   &lt;; would properly include:
2410<figure><artwork type="example">
2411    GET /pub/WWW/ HTTP/1.1
2412    Host:
2415   A client &MUST; include a Host header field in all HTTP/1.1 request
2416   messages. If the requested URI does not include an Internet host
2417   name for the service being requested, then the Host header field &MUST;
2418   be given with an empty value. An HTTP/1.1 proxy &MUST; ensure that any
2419   request message it forwards does contain an appropriate Host header
2420   field that identifies the service being requested by the proxy. All
2421   Internet-based HTTP/1.1 servers &MUST; respond with a 400 (Bad Request)
2422   status code to any HTTP/1.1 request message which lacks a Host header
2423   field.
2426   See Sections <xref target="" format="counter"/>
2427   and <xref target="" format="counter"/>
2428   for other requirements relating to Host.
2432<section title="TE" anchor="header.te">
2433  <iref primary="true" item="TE header" x:for-anchor=""/>
2434  <iref primary="true" item="Headers" subitem="TE" x:for-anchor=""/>
2436   The TE request-header field indicates what extension transfer-codings
2437   it is willing to accept in the response and whether or not it is
2438   willing to accept trailer fields in a chunked transfer-coding. Its
2439   value may consist of the keyword "trailers" and/or a comma-separated
2440   list of extension transfer-coding names with optional accept
2441   parameters (as described in <xref target="transfer.codings"/>).
2443<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TE"/><iref primary="true" item="Grammar" subitem="t-codings"/>
2444  TE        = "TE" ":" #( t-codings )
2445  t-codings = "trailers" | ( transfer-extension [ accept-params ] )
2448   The presence of the keyword "trailers" indicates that the client is
2449   willing to accept trailer fields in a chunked transfer-coding, as
2450   defined in <xref target="chunked.transfer.encoding"/>. This keyword is reserved for use with
2451   transfer-coding values even though it does not itself represent a
2452   transfer-coding.
2455   Examples of its use are:
2457<figure><artwork type="example">
2458    TE: deflate
2459    TE:
2460    TE: trailers, deflate;q=0.5
2463   The TE header field only applies to the immediate connection.
2464   Therefore, the keyword &MUST; be supplied within a Connection header
2465   field (<xref target="header.connection"/>) whenever TE is present in an HTTP/1.1 message.
2468   A server tests whether a transfer-coding is acceptable, according to
2469   a TE field, using these rules:
2470  <list style="numbers">
2471    <x:lt>
2472      <t>The "chunked" transfer-coding is always acceptable. If the
2473         keyword "trailers" is listed, the client indicates that it is
2474         willing to accept trailer fields in the chunked response on
2475         behalf of itself and any downstream clients. The implication is
2476         that, if given, the client is stating that either all
2477         downstream clients are willing to accept trailer fields in the
2478         forwarded response, or that it will attempt to buffer the
2479         response on behalf of downstream recipients.
2480      </t><t>
2481         <x:h>Note:</x:h> HTTP/1.1 does not define any means to limit the size of a
2482         chunked response such that a client can be assured of buffering
2483         the entire response.</t>
2484    </x:lt>
2485    <x:lt>
2486      <t>If the transfer-coding being tested is one of the transfer-codings
2487         listed in the TE field, then it is acceptable unless it
2488         is accompanied by a qvalue of 0. (As defined in &qvalue;, a
2489         qvalue of 0 means "not acceptable.")</t>
2490    </x:lt>
2491    <x:lt>
2492      <t>If multiple transfer-codings are acceptable, then the
2493         acceptable transfer-coding with the highest non-zero qvalue is
2494         preferred.  The "chunked" transfer-coding always has a qvalue
2495         of 1.</t>
2496    </x:lt>
2497  </list>
2500   If the TE field-value is empty or if no TE field is present, the only
2501   transfer-coding  is "chunked". A message with no transfer-coding is
2502   always acceptable.
2506<section title="Trailer" anchor="header.trailer">
2507  <iref primary="true" item="Trailer header" x:for-anchor=""/>
2508  <iref primary="true" item="Headers" subitem="Trailer" x:for-anchor=""/>
2510   The Trailer general field value indicates that the given set of
2511   header fields is present in the trailer of a message encoded with
2512   chunked transfer-coding.
2514<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
2515  Trailer  = "Trailer" ":" 1#field-name
2518   An HTTP/1.1 message &SHOULD; include a Trailer header field in a
2519   message using chunked transfer-coding with a non-empty trailer. Doing
2520   so allows the recipient to know which header fields to expect in the
2521   trailer.
2524   If no Trailer header field is present, the trailer &SHOULD-NOT;  include
2525   any header fields. See <xref target="chunked.transfer.encoding"/> for restrictions on the use of
2526   trailer fields in a "chunked" transfer-coding.
2529   Message header fields listed in the Trailer header field &MUST-NOT;
2530   include the following header fields:
2531  <list style="symbols">
2532    <t>Transfer-Encoding</t>
2533    <t>Content-Length</t>
2534    <t>Trailer</t>
2535  </list>
2539<section title="Transfer-Encoding" anchor="header.transfer-encoding">
2540  <iref primary="true" item="Transfer-Encoding header" x:for-anchor=""/>
2541  <iref primary="true" item="Headers" subitem="Transfer-Encoding" x:for-anchor=""/>
2543   The Transfer-Encoding general-header field indicates what (if any)
2544   type of transformation has been applied to the message body in order
2545   to safely transfer it between the sender and the recipient. This
2546   differs from the content-coding in that the transfer-coding is a
2547   property of the message, not of the entity.
2549<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
2550  Transfer-Encoding       = "Transfer-Encoding" ":" 1#transfer-coding
2553   Transfer-codings are defined in <xref target="transfer.codings"/>. An example is:
2555<figure><artwork type="example">
2556  Transfer-Encoding: chunked
2559   If multiple encodings have been applied to an entity, the transfer-codings
2560   &MUST; be listed in the order in which they were applied.
2561   Additional information about the encoding parameters &MAY; be provided
2562   by other entity-header fields not defined by this specification.
2565   Many older HTTP/1.0 applications do not understand the Transfer-Encoding
2566   header.
2570<section title="Upgrade" anchor="header.upgrade">
2571  <iref primary="true" item="Upgrade header" x:for-anchor=""/>
2572  <iref primary="true" item="Headers" subitem="Upgrade" x:for-anchor=""/>
2574   The Upgrade general-header allows the client to specify what
2575   additional communication protocols it supports and would like to use
2576   if the server finds it appropriate to switch protocols. The server
2577   &MUST; use the Upgrade header field within a 101 (Switching Protocols)
2578   response to indicate which protocol(s) are being switched.
2580<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
2581  Upgrade        = "Upgrade" ":" 1#product
2584   For example,
2586<figure><artwork type="example">
2587    Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
2590   The Upgrade header field is intended to provide a simple mechanism
2591   for transition from HTTP/1.1 to some other, incompatible protocol. It
2592   does so by allowing the client to advertise its desire to use another
2593   protocol, such as a later version of HTTP with a higher major version
2594   number, even though the current request has been made using HTTP/1.1.
2595   This eases the difficult transition between incompatible protocols by
2596   allowing the client to initiate a request in the more commonly
2597   supported protocol while indicating to the server that it would like
2598   to use a "better" protocol if available (where "better" is determined
2599   by the server, possibly according to the nature of the method and/or
2600   resource being requested).
2603   The Upgrade header field only applies to switching application-layer
2604   protocols upon the existing transport-layer connection. Upgrade
2605   cannot be used to insist on a protocol change; its acceptance and use
2606   by the server is optional. The capabilities and nature of the
2607   application-layer communication after the protocol change is entirely
2608   dependent upon the new protocol chosen, although the first action
2609   after changing the protocol &MUST; be a response to the initial HTTP
2610   request containing the Upgrade header field.
2613   The Upgrade header field only applies to the immediate connection.
2614   Therefore, the upgrade keyword &MUST; be supplied within a Connection
2615   header field (<xref target="header.connection"/>) whenever Upgrade is present in an
2616   HTTP/1.1 message.
2619   The Upgrade header field cannot be used to indicate a switch to a
2620   protocol on a different connection. For that purpose, it is more
2621   appropriate to use a 301, 302, 303, or 305 redirection response.
2624   This specification only defines the protocol name "HTTP" for use by
2625   the family of Hypertext Transfer Protocols, as defined by the HTTP
2626   version rules of <xref target="http.version"/> and future updates to this
2627   specification. Any token can be used as a protocol name; however, it
2628   will only be useful if both the client and server associate the name
2629   with the same protocol.
2633<section title="Via" anchor="header.via">
2634  <iref primary="true" item="Via header" x:for-anchor=""/>
2635  <iref primary="true" item="Headers" subitem="Via" x:for-anchor=""/>
2637   The Via general-header field &MUST; be used by gateways and proxies to
2638   indicate the intermediate protocols and recipients between the user
2639   agent and the server on requests, and between the origin server and
2640   the client on responses. It is analogous to the "Received" field of
2641   <xref target="RFC2822"/> and is intended to be used for tracking message forwards,
2642   avoiding request loops, and identifying the protocol capabilities of
2643   all senders along the request/response chain.
2645<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"/>
2646  Via =  "Via" ":" 1#( received-protocol received-by [ comment ] )
2647  received-protocol = [ protocol-name "/" ] protocol-version
2648  protocol-name     = token
2649  protocol-version  = token
2650  received-by       = ( host [ ":" port ] ) | pseudonym
2651  pseudonym         = token
2654   The received-protocol indicates the protocol version of the message
2655   received by the server or client along each segment of the
2656   request/response chain. The received-protocol version is appended to
2657   the Via field value when the message is forwarded so that information
2658   about the protocol capabilities of upstream applications remains
2659   visible to all recipients.
2662   The protocol-name is optional if and only if it would be "HTTP". The
2663   received-by field is normally the host and optional port number of a
2664   recipient server or client that subsequently forwarded the message.
2665   However, if the real host is considered to be sensitive information,
2666   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
2667   be assumed to be the default port of the received-protocol.
2670   Multiple Via field values represents each proxy or gateway that has
2671   forwarded the message. Each recipient &MUST; append its information
2672   such that the end result is ordered according to the sequence of
2673   forwarding applications.
2676   Comments &MAY; be used in the Via header field to identify the software
2677   of the recipient proxy or gateway, analogous to the User-Agent and
2678   Server header fields. However, all comments in the Via field are
2679   optional and &MAY; be removed by any recipient prior to forwarding the
2680   message.
2683   For example, a request message could be sent from an HTTP/1.0 user
2684   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2685   forward the request to a public proxy at, which completes
2686   the request by forwarding it to the origin server at
2687   The request received by would then have the following
2688   Via header field:
2690<figure><artwork type="example">
2691    Via: 1.0 fred, 1.1 (Apache/1.1)
2694   Proxies and gateways used as a portal through a network firewall
2695   &SHOULD-NOT;, by default, forward the names and ports of hosts within
2696   the firewall region. This information &SHOULD; only be propagated if
2697   explicitly enabled. If not enabled, the received-by host of any host
2698   behind the firewall &SHOULD; be replaced by an appropriate pseudonym
2699   for that host.
2702   For organizations that have strong privacy requirements for hiding
2703   internal structures, a proxy &MAY; combine an ordered subsequence of
2704   Via header field entries with identical received-protocol values into
2705   a single such entry. For example,
2707<figure><artwork type="example">
2708    Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2711        could be collapsed to
2713<figure><artwork type="example">
2714    Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2717   Applications &SHOULD-NOT;  combine multiple entries unless they are all
2718   under the same organizational control and the hosts have already been
2719   replaced by pseudonyms. Applications &MUST-NOT; combine entries which
2720   have different received-protocol values.
2726<section title="IANA Considerations" anchor="IANA.considerations">
2728   TBD.
2732<section title="Security Considerations" anchor="security.considerations">
2734   This section is meant to inform application developers, information
2735   providers, and users of the security limitations in HTTP/1.1 as
2736   described by this document. The discussion does not include
2737   definitive solutions to the problems revealed, though it does make
2738   some suggestions for reducing security risks.
2741<section title="Personal Information" anchor="personal.information">
2743   HTTP clients are often privy to large amounts of personal information
2744   (e.g. the user's name, location, mail address, passwords, encryption
2745   keys, etc.), and &SHOULD; be very careful to prevent unintentional
2746   leakage of this information.
2747   We very strongly recommend that a convenient interface be provided
2748   for the user to control dissemination of such information, and that
2749   designers and implementors be particularly careful in this area.
2750   History shows that errors in this area often create serious security
2751   and/or privacy problems and generate highly adverse publicity for the
2752   implementor's company.
2756<section title="Abuse of Server Log Information" anchor="abuse.of.server.log.information">
2758   A server is in the position to save personal data about a user's
2759   requests which might identify their reading patterns or subjects of
2760   interest. This information is clearly confidential in nature and its
2761   handling can be constrained by law in certain countries. People using
2762   HTTP to provide data are responsible for ensuring that
2763   such material is not distributed without the permission of any
2764   individuals that are identifiable by the published results.
2768<section title="Attacks Based On File and Path Names" anchor="attack.pathname">
2770   Implementations of HTTP origin servers &SHOULD; be careful to restrict
2771   the documents returned by HTTP requests to be only those that were
2772   intended by the server administrators. If an HTTP server translates
2773   HTTP URIs directly into file system calls, the server &MUST; take
2774   special care not to serve files that were not intended to be
2775   delivered to HTTP clients. For example, UNIX, Microsoft Windows, and
2776   other operating systems use ".." as a path component to indicate a
2777   directory level above the current one. On such a system, an HTTP
2778   server &MUST; disallow any such construct in the Request-URI if it
2779   would otherwise allow access to a resource outside those intended to
2780   be accessible via the HTTP server. Similarly, files intended for
2781   reference only internally to the server (such as access control
2782   files, configuration files, and script code) &MUST; be protected from
2783   inappropriate retrieval, since they might contain sensitive
2784   information. Experience has shown that minor bugs in such HTTP server
2785   implementations have turned into security risks.
2789<section title="DNS Spoofing" anchor="dns.spoofing">
2791   Clients using HTTP rely heavily on the Domain Name Service, and are
2792   thus generally prone to security attacks based on the deliberate
2793   mis-association of IP addresses and DNS names. Clients need to be
2794   cautious in assuming the continuing validity of an IP number/DNS name
2795   association.
2798   In particular, HTTP clients &SHOULD; rely on their name resolver for
2799   confirmation of an IP number/DNS name association, rather than
2800   caching the result of previous host name lookups. Many platforms
2801   already can cache host name lookups locally when appropriate, and
2802   they &SHOULD; be configured to do so. It is proper for these lookups to
2803   be cached, however, only when the TTL (Time To Live) information
2804   reported by the name server makes it likely that the cached
2805   information will remain useful.
2808   If HTTP clients cache the results of host name lookups in order to
2809   achieve a performance improvement, they &MUST; observe the TTL
2810   information reported by DNS.
2813   If HTTP clients do not observe this rule, they could be spoofed when
2814   a previously-accessed server's IP address changes. As network
2815   renumbering is expected to become increasingly common <xref target="RFC1900"/>, the
2816   possibility of this form of attack will grow. Observing this
2817   requirement thus reduces this potential security vulnerability.
2820   This requirement also improves the load-balancing behavior of clients
2821   for replicated servers using the same DNS name and reduces the
2822   likelihood of a user's experiencing failure in accessing sites which
2823   use that strategy.
2827<section title="Proxies and Caching" anchor="attack.proxies">
2829   By their very nature, HTTP proxies are men-in-the-middle, and
2830   represent an opportunity for man-in-the-middle attacks. Compromise of
2831   the systems on which the proxies run can result in serious security
2832   and privacy problems. Proxies have access to security-related
2833   information, personal information about individual users and
2834   organizations, and proprietary information belonging to users and
2835   content providers. A compromised proxy, or a proxy implemented or
2836   configured without regard to security and privacy considerations,
2837   might be used in the commission of a wide range of potential attacks.
2840   Proxy operators should protect the systems on which proxies run as
2841   they would protect any system that contains or transports sensitive
2842   information. In particular, log information gathered at proxies often
2843   contains highly sensitive personal information, and/or information
2844   about organizations. Log information should be carefully guarded, and
2845   appropriate guidelines for use developed and followed. (<xref target="abuse.of.server.log.information"/>).
2848   Proxy implementors should consider the privacy and security
2849   implications of their design and coding decisions, and of the
2850   configuration options they provide to proxy operators (especially the
2851   default configuration).
2854   Users of a proxy need to be aware that they are no trustworthier than
2855   the people who run the proxy; HTTP itself cannot solve this problem.
2858   The judicious use of cryptography, when appropriate, may suffice to
2859   protect against a broad range of security and privacy attacks. Such
2860   cryptography is beyond the scope of the HTTP/1.1 specification.
2864<section title="Denial of Service Attacks on Proxies" anchor="attack.DoS">
2866   They exist. They are hard to defend against. Research continues.
2867   Beware.
2872<section title="Acknowledgments" anchor="ack">
2874   This specification makes heavy use of the augmented BNF and generic
2875   constructs defined by David H. Crocker for <xref target="RFC822ABNF"/>. Similarly, it
2876   reuses many of the definitions provided by Nathaniel Borenstein and
2877   Ned Freed for MIME <xref target="RFC2045"/>. We hope that their inclusion in this
2878   specification will help reduce past confusion over the relationship
2879   between HTTP and Internet mail message formats.
2882   HTTP has evolved considerably over the years. It has
2883   benefited from a large and active developer community--the many
2884   people who have participated on the www-talk mailing list--and it is
2885   that community which has been most responsible for the success of
2886   HTTP and of the World-Wide Web in general. Marc Andreessen, Robert
2887   Cailliau, Daniel W. Connolly, Bob Denny, John Franks, Jean-Francois
2888   Groff, Phillip M. Hallam-Baker, Hakon W. Lie, Ari Luotonen, Rob
2889   McCool, Lou Montulli, Dave Raggett, Tony Sanders, and Marc
2890   VanHeyningen deserve special recognition for their efforts in
2891   defining early aspects of the protocol.
2894   This document has benefited greatly from the comments of all those
2895   participating in the HTTP-WG. In addition to those already mentioned,
2896   the following individuals have contributed to this specification:
2899   Gary Adams, Harald Tveit Alvestrand, Keith Ball, Brian Behlendorf,
2900   Paul Burchard, Maurizio Codogno, Mike Cowlishaw, Roman Czyborra,
2901   Michael A. Dolan, Daniel DuBois, David J. Fiander, Alan Freier, Marc Hedlund, Greg Herlihy,
2902   Koen Holtman, Alex Hopmann, Bob Jernigan, Shel Kaphan, Rohit Khare,
2903   John Klensin, Martijn Koster, Alexei Kosut, David M. Kristol,
2904   Daniel LaLiberte, Ben Laurie, Paul J. Leach, Albert Lunde,
2905   John C. Mallery, Jean-Philippe Martin-Flatin, Mitra, David Morris,
2906   Gavin Nicol, Ross Patterson, Bill Perry, Jeffrey Perry, Scott Powers, Owen Rees,
2907   Luigi Rizzo, David Robinson, Marc Salomon, Rich Salz,
2908   Allan M. Schiffman, Jim Seidman, Chuck Shotton, Eric W. Sink,
2909   Simon E. Spero, Richard N. Taylor, Robert S. Thau,
2910   Bill (BearHeart) Weinman, Francois Yergeau, Mary Ellen Zurko,
2911   Josh Cohen.
2914   Thanks to the "cave men" of Palo Alto. You know who you are.
2917   Jim Gettys (the editor of <xref target="RFC2616"/>) wishes particularly
2918   to thank Roy Fielding, the editor of <xref target="RFC2068"/>, along
2919   with John Klensin, Jeff Mogul, Paul Leach, Dave Kristol, Koen
2920   Holtman, John Franks, Josh Cohen, Alex Hopmann, Scott Lawrence, and
2921   Larry Masinter for their help. And thanks go particularly to Jeff
2922   Mogul and Scott Lawrence for performing the "MUST/MAY/SHOULD" audit.
2925   The Apache Group, Anselm Baird-Smith, author of Jigsaw, and Henrik
2926   Frystyk implemented RFC 2068 early, and we wish to thank them for the
2927   discovery of many of the problems that this document attempts to
2928   rectify.
2935<references title="Normative References">
2937<reference anchor="ISO-8859-1">
2938  <front>
2939    <title>
2940     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
2941    </title>
2942    <author>
2943      <organization>International Organization for Standardization</organization>
2944    </author>
2945    <date year="1998"/>
2946  </front>
2947  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
2950<reference anchor="Part2">
2951  <front>
2952    <title abbrev="HTTP/1.1">HTTP/1.1, part 2: Message Semantics</title>
2953    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
2954      <organization abbrev="Day Software">Day Software</organization>
2955      <address><email></email></address>
2956    </author>
2957    <author initials="J." surname="Gettys" fullname="Jim Gettys">
2958      <organization>One Laptop per Child</organization>
2959      <address><email></email></address>
2960    </author>
2961    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
2962      <organization abbrev="HP">Hewlett-Packard Company</organization>
2963      <address><email></email></address>
2964    </author>
2965    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
2966      <organization abbrev="Microsoft">Microsoft Corporation</organization>
2967      <address><email></email></address>
2968    </author>
2969    <author initials="L." surname="Masinter" fullname="Larry Masinter">
2970      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
2971      <address><email></email></address>
2972    </author>
2973    <author initials="P." surname="Leach" fullname="Paul J. Leach">
2974      <organization abbrev="Microsoft">Microsoft Corporation</organization>
2975      <address><email></email></address>
2976    </author>
2977    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
2978      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
2979      <address><email></email></address>
2980    </author>
2981    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
2982      <organization abbrev="W3C">World Wide Web Consortium</organization>
2983      <address><email></email></address>
2984    </author>
2985    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
2986      <organization abbrev="greenbytes">greenbytes GmbH</organization>
2987      <address><email></email></address>
2988    </author>
2989    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
2990  </front>
2991  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-&ID-VERSION;"/>
2992  <x:source href="p2-semantics.xml" basename="p2-semantics"/>
2995<reference anchor="Part3">
2996  <front>
2997    <title abbrev="HTTP/1.1">HTTP/1.1, part 3: Message Payload and Content Negotiation</title>
2998    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
2999      <organization abbrev="Day Software">Day Software</organization>
3000      <address><email></email></address>
3001    </author>
3002    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3003      <organization>One Laptop per Child</organization>
3004      <address><email></email></address>
3005    </author>
3006    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3007      <organization abbrev="HP">Hewlett-Packard Company</organization>
3008      <address><email></email></address>
3009    </author>
3010    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3011      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3012      <address><email></email></address>
3013    </author>
3014    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3015      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3016      <address><email></email></address>
3017    </author>
3018    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3019      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3020      <address><email></email></address>
3021    </author>
3022    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3023      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3024      <address><email></email></address>
3025    </author>
3026    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3027      <organization abbrev="W3C">World Wide Web Consortium</organization>
3028      <address><email></email></address>
3029    </author>
3030    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3031      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3032      <address><email></email></address>
3033    </author>
3034    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3035  </front>
3036  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p3-payload-&ID-VERSION;"/>
3037  <x:source href="p3-payload.xml" basename="p3-payload"/>
3040<reference anchor="Part5">
3041  <front>
3042    <title abbrev="HTTP/1.1">HTTP/1.1, part 5: Range Requests and Partial Responses</title>
3043    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3044      <organization abbrev="Day Software">Day Software</organization>
3045      <address><email></email></address>
3046    </author>
3047    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3048      <organization>One Laptop per Child</organization>
3049      <address><email></email></address>
3050    </author>
3051    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3052      <organization abbrev="HP">Hewlett-Packard Company</organization>
3053      <address><email></email></address>
3054    </author>
3055    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3056      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3057      <address><email></email></address>
3058    </author>
3059    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3060      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3061      <address><email></email></address>
3062    </author>
3063    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3064      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3065      <address><email></email></address>
3066    </author>
3067    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3068      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3069      <address><email></email></address>
3070    </author>
3071    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3072      <organization abbrev="W3C">World Wide Web Consortium</organization>
3073      <address><email></email></address>
3074    </author>
3075    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3076      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3077      <address><email></email></address>
3078    </author>
3079    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3080  </front>
3081  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-&ID-VERSION;"/>
3082  <x:source href="p5-range.xml" basename="p5-range"/>
3085<reference anchor="Part6">
3086  <front>
3087    <title abbrev="HTTP/1.1">HTTP/1.1, part 6: Caching</title>
3088    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3089      <organization abbrev="Day Software">Day Software</organization>
3090      <address><email></email></address>
3091    </author>
3092    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3093      <organization>One Laptop per Child</organization>
3094      <address><email></email></address>
3095    </author>
3096    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3097      <organization abbrev="HP">Hewlett-Packard Company</organization>
3098      <address><email></email></address>
3099    </author>
3100    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3101      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3102      <address><email></email></address>
3103    </author>
3104    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3105      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3106      <address><email></email></address>
3107    </author>
3108    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3109      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3110      <address><email></email></address>
3111    </author>
3112    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3113      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3114      <address><email></email></address>
3115    </author>
3116    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3117      <organization abbrev="W3C">World Wide Web Consortium</organization>
3118      <address><email></email></address>
3119    </author>
3120    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3121      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3122      <address><email></email></address>
3123    </author>
3124    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3125  </front>
3126  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-&ID-VERSION;"/>
3127  <x:source href="p6-cache.xml" basename="p6-cache"/>
3130<reference anchor="RFC822ABNF">
3131  <front>
3132    <title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3133    <author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3134      <organization>University of Delaware, Dept. of Electrical Engineering</organization>
3135      <address><email>DCrocker@UDel-Relay</email></address>
3136    </author>
3137    <date month="August" day="13" year="1982"/>
3138  </front>
3139  <seriesInfo name="STD" value="11"/>
3140  <seriesInfo name="RFC" value="822"/>
3143<reference anchor="RFC2045">
3144  <front>
3145    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
3146    <author initials="N." surname="Freed" fullname="Ned Freed">
3147      <organization>Innosoft International, Inc.</organization>
3148      <address><email></email></address>
3149    </author>
3150    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
3151      <organization>First Virtual Holdings</organization>
3152      <address><email></email></address>
3153    </author>
3154    <date month="November" year="1996"/>
3155  </front>
3156  <seriesInfo name="RFC" value="2045"/>
3159<reference anchor="RFC2047">
3160  <front>
3161    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
3162    <author initials="K." surname="Moore" fullname="Keith Moore">
3163      <organization>University of Tennessee</organization>
3164      <address><email></email></address>
3165    </author>
3166    <date month="November" year="1996"/>
3167  </front>
3168  <seriesInfo name="RFC" value="2047"/>
3171<reference anchor="RFC2119">
3172  <front>
3173    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
3174    <author initials="S." surname="Bradner" fullname="Scott Bradner">
3175      <organization>Harvard University</organization>
3176      <address><email></email></address>
3177    </author>
3178    <date month="March" year="1997"/>
3179  </front>
3180  <seriesInfo name="BCP" value="14"/>
3181  <seriesInfo name="RFC" value="2119"/>
3184<reference anchor="RFC2396">
3185  <front>
3186    <title abbrev="URI Generic Syntax">Uniform Resource Identifiers (URI): Generic Syntax</title>
3187    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3188      <organization abbrev="MIT/LCS">World Wide Web Consortium</organization>
3189      <address><email></email></address>
3190    </author>
3191    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3192      <organization abbrev="U.C. Irvine">Department of Information and Computer Science</organization>
3193      <address><email></email></address>
3194    </author>
3195    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3196      <organization abbrev="Xerox Corporation">Xerox PARC</organization>
3197      <address><email></email></address>
3198    </author>
3199    <date month="August" year="1998"/>
3200  </front>
3201  <seriesInfo name="RFC" value="2396"/>
3204<reference anchor="RFC4288">
3205  <front>
3206    <title>Media Type Specifications and Registration Procedures</title>
3207    <author initials="N." surname="Freed" fullname="N. Freed">
3208      <organization>Sun Microsystems</organization>
3209      <address>
3210        <email></email>
3211      </address>
3212    </author>
3213    <author initials="J." surname="Klensin" fullname="J. Klensin">
3214      <organization/>
3215      <address>
3216        <email></email>
3217      </address>
3218    </author>
3219    <date year="2005" month="December"/>
3220  </front>
3221  <seriesInfo name="BCP" value="13"/>
3222  <seriesInfo name="RFC" value="4288"/>
3225<reference anchor="USASCII">
3226  <front>
3227    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
3228    <author>
3229      <organization>American National Standards Institute</organization>
3230    </author>
3231    <date year="1986"/>
3232  </front>
3233  <seriesInfo name="ANSI" value="X3.4"/>
3238<references title="Informative References">
3240<reference anchor="Nie1997" target="">
3241  <front>
3242    <title>Network Performance Effects of HTTP/1.1, CSS1, and PNG</title>
3243    <author initials="H.F.." surname="Nielsen" fullname="H.F. Nielsen">
3244      <organization/>
3245    </author>
3246    <author initials="J." surname="Gettys" fullname="J. Gettys">
3247      <organization/>
3248    </author>
3249    <author initials="E." surname="Prud'hommeaux" fullname="E. Prud'hommeaux">
3250      <organization/>
3251    </author>
3252    <author initials="H." surname="Lie" fullname="H. Lie">
3253      <organization/>
3254    </author>
3255    <author initials="C." surname="Lilley" fullname="C. Lilley">
3256      <organization/>
3257    </author>
3258    <date year="1997" month="September"/>
3259  </front>
3260  <seriesInfo name="ACM" value="Proceedings of the ACM SIGCOMM '97 conference on Applications, technologies, architectures, and protocols for computer communication SIGCOMM '97"/>
3263<reference anchor="Pad1995">
3264  <front>
3265    <title>Improving HTTP Latency</title>
3266    <author initials="V.N." surname="Padmanabhan" fullname="Venkata N. Padmanabhan">
3267      <organization/>
3268    </author>
3269    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3270      <organization/>
3271    </author>
3272    <date year="1995" month="December"/>
3273  </front>
3274  <seriesInfo name="Computer Networks and ISDN Systems" value="v. 28, pp. 25-35"/>
3275  <annotation>
3276    Slightly revised version of paper in Proc. 2nd International WWW Conference '94: Mosaic and the Web, Oct. 1994,
3277    which is available at <eref target=""/>.
3278  </annotation>
3281<reference anchor="RFC822">
3282  <front>
3283    <title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3284    <author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3285      <organization>University of Delaware, Dept. of Electrical Engineering</organization>
3286      <address><email>DCrocker@UDel-Relay</email></address>
3287    </author>
3288    <date month="August" day="13" year="1982"/>
3289  </front>
3290  <seriesInfo name="STD" value="11"/>
3291  <seriesInfo name="RFC" value="822"/>
3294<reference anchor="RFC959">
3295  <front>
3296    <title abbrev="File Transfer Protocol">File Transfer Protocol</title>
3297    <author initials="J." surname="Postel" fullname="J. Postel">
3298      <organization>Information Sciences Institute (ISI)</organization>
3299    </author>
3300    <author initials="J." surname="Reynolds" fullname="J. Reynolds">
3301      <organization/>
3302    </author>
3303    <date month="October" year="1985"/>
3304  </front>
3305  <seriesInfo name="STD" value="9"/>
3306  <seriesInfo name="RFC" value="959"/>
3309<reference anchor="RFC1123">
3310  <front>
3311    <title>Requirements for Internet Hosts - Application and Support</title>
3312    <author initials="R." surname="Braden" fullname="Robert Braden">
3313      <organization>University of Southern California (USC), Information Sciences Institute</organization>
3314      <address><email>Braden@ISI.EDU</email></address>
3315    </author>
3316    <date month="October" year="1989"/>
3317  </front>
3318  <seriesInfo name="STD" value="3"/>
3319  <seriesInfo name="RFC" value="1123"/>
3322<reference anchor="RFC1305">
3323  <front>
3324    <title>Network Time Protocol (Version 3) Specification, Implementation</title>
3325    <author initials="D." surname="Mills" fullname="David L. Mills">
3326      <organization>University of Delaware, Electrical Engineering Department</organization>
3327      <address><email></email></address>
3328    </author>
3329    <date month="March" year="1992"/>
3330  </front>
3331  <seriesInfo name="RFC" value="1305"/>
3334<reference anchor="RFC1436">
3335  <front>
3336    <title abbrev="Gopher">The Internet Gopher Protocol (a distributed document search and retrieval protocol)</title>
3337    <author initials="F." surname="Anklesaria" fullname="Farhad Anklesaria">
3338      <organization>University of Minnesota, Computer and Information Services</organization>
3339      <address><email></email></address>
3340    </author>
3341    <author initials="M." surname="McCahill" fullname="Mark McCahill">
3342      <organization>University of Minnesota, Computer and Information Services</organization>
3343      <address><email></email></address>
3344    </author>
3345    <author initials="P." surname="Lindner" fullname="Paul Lindner">
3346      <organization>University of Minnesota, Computer and Information Services</organization>
3347      <address><email></email></address>
3348    </author>
3349    <author initials="D." surname="Johnson" fullname="David Johnson">
3350      <organization>University of Minnesota, Computer and Information Services</organization>
3351      <address><email></email></address>
3352    </author>
3353    <author initials="D." surname="Torrey" fullname="Daniel Torrey">
3354      <organization>University of Minnesota, Computer and Information Services</organization>
3355      <address><email></email></address>
3356    </author>
3357    <author initials="B." surname="Alberti" fullname="Bob Alberti">
3358      <organization>University of Minnesota, Computer and Information Services</organization>
3359      <address><email></email></address>
3360    </author>
3361    <date month="March" year="1993"/>
3362  </front>
3363  <seriesInfo name="RFC" value="1436"/>
3366<reference anchor="RFC1630">
3367  <front>
3368    <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>
3369    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3370      <organization>CERN, World-Wide Web project</organization>
3371      <address><email></email></address>
3372    </author>
3373    <date month="June" year="1994"/>
3374  </front>
3375  <seriesInfo name="RFC" value="1630"/>
3378<reference anchor="RFC1737">
3379  <front>
3380    <title abbrev="Requirements for Uniform Resource Names">Functional Requirements for Uniform Resource Names</title>
3381    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3382      <organization>Xerox Palo Alto Research Center</organization>
3383      <address><email></email></address>
3384    </author>
3385    <author initials="K." surname="Sollins" fullname="Karen Sollins">
3386      <organization>MIT Laboratory for Computer Science</organization>
3387      <address><email></email></address>
3388    </author>
3389    <date month="December" year="1994"/>
3390  </front>
3391  <seriesInfo name="RFC" value="1737"/>
3394<reference anchor="RFC1738">
3395  <front>
3396    <title>Uniform Resource Locators (URL)</title>
3397    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3398      <organization>CERN, World-Wide Web project</organization>
3399      <address><email></email></address>
3400    </author>
3401    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3402      <organization>Xerox PARC</organization>
3403      <address><email></email></address>
3404    </author>
3405    <author initials="M." surname="McCahill" fullname="Mark McCahill">
3406      <organization>University of Minnesota, Computer and Information Services</organization>
3407      <address><email></email></address>
3408    </author>
3409    <date month="December" year="1994"/>
3410  </front>
3411  <seriesInfo name="RFC" value="1738"/>
3414<reference anchor="RFC1808">
3415  <front>
3416    <title>Relative Uniform Resource Locators</title>
3417    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3418      <organization>University of California Irvine, Department of Information and Computer Science</organization>
3419      <address><email></email></address>
3420    </author>
3421    <date month="June" year="1995"/>
3422  </front>
3423  <seriesInfo name="RFC" value="1808"/>
3426<reference anchor="RFC1900">
3427  <front>
3428    <title>Renumbering Needs Work</title>
3429    <author initials="B." surname="Carpenter" fullname="Brian E. Carpenter">
3430      <organization>CERN, Computing and Networks Division</organization>
3431      <address><email></email></address>
3432    </author>
3433    <author initials="Y." surname="Rekhter" fullname="Yakov Rekhter">
3434      <organization>cisco Systems</organization>
3435      <address><email></email></address>
3436    </author>
3437    <date month="February" year="1996"/>
3438  </front>
3439  <seriesInfo name="RFC" value="1900"/>
3442<reference anchor="RFC1945">
3443  <front>
3444    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
3445    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3446      <organization>MIT, Laboratory for Computer Science</organization>
3447      <address><email></email></address>
3448    </author>
3449    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3450      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
3451      <address><email></email></address>
3452    </author>
3453    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3454      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
3455      <address><email></email></address>
3456    </author>
3457    <date month="May" year="1996"/>
3458  </front>
3459  <seriesInfo name="RFC" value="1945"/>
3462<reference anchor="RFC2068">
3463  <front>
3464    <title abbrev="HTTP/1.1">Hypertext Transfer Protocol -- HTTP/1.1</title>
3465    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3466      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
3467      <address><email></email></address>
3468    </author>
3469    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3470      <organization>MIT Laboratory for Computer Science</organization>
3471      <address><email></email></address>
3472    </author>
3473    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3474      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
3475      <address><email></email></address>
3476    </author>
3477    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3478      <organization>MIT Laboratory for Computer Science</organization>
3479      <address><email></email></address>
3480    </author>
3481    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3482      <organization>MIT Laboratory for Computer Science</organization>
3483      <address><email></email></address>
3484    </author>
3485    <date month="January" year="1997"/>
3486  </front>
3487  <seriesInfo name="RFC" value="2068"/>
3490<reference anchor="RFC2145">
3491  <front>
3492    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
3493    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3494      <organization>Western Research Laboratory</organization>
3495      <address><email></email></address>
3496    </author>
3497    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3498      <organization>Department of Information and Computer Science</organization>
3499      <address><email></email></address>
3500    </author>
3501    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3502      <organization>MIT Laboratory for Computer Science</organization>
3503      <address><email></email></address>
3504    </author>
3505    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3506      <organization>W3 Consortium</organization>
3507      <address><email></email></address>
3508    </author>
3509    <date month="May" year="1997"/>
3510  </front>
3511  <seriesInfo name="RFC" value="2145"/>
3514<reference anchor="RFC2324">
3515  <front>
3516    <title abbrev="HTCPCP/1.0">Hyper Text Coffee Pot Control Protocol (HTCPCP/1.0)</title>
3517    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3518      <organization>Xerox Palo Alto Research Center</organization>
3519      <address><email></email></address>
3520    </author>
3521    <date month="April" day="1" year="1998"/>
3522  </front>
3523  <seriesInfo name="RFC" value="2324"/>
3526<reference anchor="RFC2616">
3527  <front>
3528    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
3529    <author initials="R." surname="Fielding" fullname="R. Fielding">
3530      <organization>University of California, Irvine</organization>
3531      <address><email></email></address>
3532    </author>
3533    <author initials="J." surname="Gettys" fullname="J. Gettys">
3534      <organization>W3C</organization>
3535      <address><email></email></address>
3536    </author>
3537    <author initials="J." surname="Mogul" fullname="J. Mogul">
3538      <organization>Compaq Computer Corporation</organization>
3539      <address><email></email></address>
3540    </author>
3541    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
3542      <organization>MIT Laboratory for Computer Science</organization>
3543      <address><email></email></address>
3544    </author>
3545    <author initials="L." surname="Masinter" fullname="L. Masinter">
3546      <organization>Xerox Corporation</organization>
3547      <address><email></email></address>
3548    </author>
3549    <author initials="P." surname="Leach" fullname="P. Leach">
3550      <organization>Microsoft Corporation</organization>
3551      <address><email></email></address>
3552    </author>
3553    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
3554      <organization>W3C</organization>
3555      <address><email></email></address>
3556    </author>
3557    <date month="June" year="1999"/>
3558  </front>
3559  <seriesInfo name="RFC" value="2616"/>
3562<reference anchor="RFC2821">
3563  <front>
3564    <title>Simple Mail Transfer Protocol</title>
3565    <author initials="J." surname="Klensin" fullname="J. Klensin">
3566      <organization>AT&amp;T Laboratories</organization>
3567      <address><email></email></address>
3568    </author>
3569    <date year="2001" month="April"/>
3570  </front>
3571  <seriesInfo name="RFC" value="2821"/>
3574<reference anchor="RFC2822">
3575  <front>
3576    <title>Internet Message Format</title>
3577    <author initials="P." surname="Resnick" fullname="P. Resnick">
3578      <organization>QUALCOMM Incorporated</organization>
3579    </author>
3580    <date year="2001" month="April"/>
3581  </front>
3582  <seriesInfo name="RFC" value="2822"/>
3585<reference anchor='RFC3977'>
3586  <front>
3587    <title>Network News Transfer Protocol (NNTP)</title>
3588    <author initials='C.' surname='Feather' fullname='C. Feather'>
3589      <organization>THUS plc</organization>
3590      <address><email></email></address>
3591    </author>
3592    <date year='2006' month='October' />
3593  </front>
3594  <seriesInfo name="RFC" value="3977"/>
3597<reference anchor="Spe" target="">
3598  <front>
3599  <title>Analysis of HTTP Performance Problems</title>
3600  <author initials="S." surname="Spero" fullname="Simon E. Spero">
3601    <organization/>
3602  </author>
3603  <date/>
3604  </front>
3607<reference anchor="Tou1998" target="">
3608  <front>
3609  <title>Analysis of HTTP Performance</title>
3610  <author initials="J." surname="Touch" fullname="Joe Touch">
3611    <organization>USC/Information Sciences Institute</organization>
3612    <address><email></email></address>
3613  </author>
3614  <author initials="J." surname="Heidemann" fullname="John Heidemann">
3615    <organization>USC/Information Sciences Institute</organization>
3616    <address><email></email></address>
3617  </author>
3618  <author initials="K." surname="Obraczka" fullname="Katia Obraczka">
3619    <organization>USC/Information Sciences Institute</organization>
3620    <address><email></email></address>
3621  </author>
3622  <date year="1998" month="Aug"/>
3623  </front>
3624  <seriesInfo name="ISI Research Report" value="ISI/RR-98-463"/>
3625  <annotation>(original report dated Aug. 1996)</annotation>
3628<reference anchor="WAIS">
3629  <front>
3630    <title>WAIS Interface Protocol Prototype Functional Specification (v1.5)</title>
3631    <author initials="F." surname="Davis" fullname="F. Davis">
3632      <organization>Thinking Machines Corporation</organization>
3633    </author>
3634    <author initials="B." surname="Kahle" fullname="B. Kahle">
3635      <organization>Thinking Machines Corporation</organization>
3636    </author>
3637    <author initials="H." surname="Morris" fullname="H. Morris">
3638      <organization>Thinking Machines Corporation</organization>
3639    </author>
3640    <author initials="J." surname="Salem" fullname="J. Salem">
3641      <organization>Thinking Machines Corporation</organization>
3642    </author>
3643    <author initials="T." surname="Shen" fullname="T. Shen">
3644      <organization>Thinking Machines Corporation</organization>
3645    </author>
3646    <author initials="R." surname="Wang" fullname="R. Wang">
3647      <organization>Thinking Machines Corporation</organization>
3648    </author>
3649    <author initials="J." surname="Sui" fullname="J. Sui">
3650      <organization>Thinking Machines Corporation</organization>
3651    </author>
3652    <author initials="M." surname="Grinbaum" fullname="M. Grinbaum">
3653      <organization>Thinking Machines Corporation</organization>
3654    </author>
3655    <date month="April" year="1990"/>
3656  </front>
3657  <seriesInfo name="Thinking Machines Corporation" value=""/>
3663<section title="Internet Media Types" anchor="">
3665   In addition to defining HTTP/1.1, this document serves
3666   as the specification for the Internet media type "message/http" and
3667   "application/http". The following is to be registered with IANA <xref target="RFC4288"/>.
3669<section title="Internet Media Type message/http" anchor="">
3670<iref item="Media Type" subitem="message/http" primary="true"/>
3671<iref item="message/http Media Type" primary="true"/>
3673   The message/http type can be used to enclose a single HTTP request or
3674   response message, provided that it obeys the MIME restrictions for all
3675   "message" types regarding line length and encodings.
3678  <list style="hanging" x:indent="12em">
3679    <t hangText="Type name:">
3680      message
3681    </t>
3682    <t hangText="Subtype name:">
3683      http
3684    </t>
3685    <t hangText="Required parameters:">
3686      none
3687    </t>
3688    <t hangText="Optional parameters:">
3689      version, msgtype
3690      <list style="hanging">
3691        <t hangText="version:">
3692          The HTTP-Version number of the enclosed message
3693          (e.g., "1.1"). If not present, the version can be
3694          determined from the first line of the body.
3695        </t>
3696        <t hangText="msgtype:">
3697          The message type -- "request" or "response". If not
3698          present, the type can be determined from the first
3699          line of the body.
3700        </t>
3701      </list>
3702    </t>
3703    <t hangText="Encoding considerations:">
3704      only "7bit", "8bit", or "binary" are permitted
3705    </t>
3706    <t hangText="Security considerations:">
3707      none
3708    </t>
3709    <t hangText="Interoperability considerations:">
3710      none
3711    </t>
3712    <t hangText="Published specification:">
3713      This specification (see <xref target=""/>).
3714    </t>
3715    <t hangText="Applications that use this media type:">
3716    </t>
3717    <t hangText="Additional information:">
3718      <list style="hanging">
3719        <t hangText="Magic number(s):">none</t>
3720        <t hangText="File extension(s):">none</t>
3721        <t hangText="Macintosh file type code(s):">none</t>
3722      </list>
3723    </t>
3724    <t hangText="Person and email address to contact for further information:">
3725      See Authors Section.
3726    </t>
3727                <t hangText="Intended usage:">
3728                  COMMON
3729    </t>
3730                <t hangText="Restrictions on usage:">
3731                  none
3732    </t>
3733    <t hangText="Author/Change controller:">
3734      IESG
3735    </t>
3736  </list>
3739<section title="Internet Media Type application/http" anchor="">
3740<iref item="Media Type" subitem="application/http" primary="true"/>
3741<iref item="application/http Media Type" primary="true"/>
3743   The application/http type can be used to enclose a pipeline of one or more
3744   HTTP request or response messages (not intermixed).
3747  <list style="hanging" x:indent="12em">
3748    <t hangText="Type name:">
3749      application
3750    </t>
3751    <t hangText="Subtype name:">
3752      http
3753    </t>
3754    <t hangText="Required parameters:">
3755      none
3756    </t>
3757    <t hangText="Optional parameters:">
3758      version, msgtype
3759      <list style="hanging">
3760        <t hangText="version:">
3761          The HTTP-Version number of the enclosed messages
3762          (e.g., "1.1"). If not present, the version can be
3763          determined from the first line of the body.
3764        </t>
3765        <t hangText="msgtype:">
3766          The message type -- "request" or "response". If not
3767          present, the type can be determined from the first
3768          line of the body.
3769        </t>
3770      </list>
3771    </t>
3772    <t hangText="Encoding considerations:">
3773      HTTP messages enclosed by this type
3774      are in "binary" format; use of an appropriate
3775      Content-Transfer-Encoding is required when
3776      transmitted via E-mail.
3777    </t>
3778    <t hangText="Security considerations:">
3779      none
3780    </t>
3781    <t hangText="Interoperability considerations:">
3782      none
3783    </t>
3784    <t hangText="Published specification:">
3785      This specification (see <xref target=""/>).
3786    </t>
3787    <t hangText="Applications that use this media type:">
3788    </t>
3789    <t hangText="Additional information:">
3790      <list style="hanging">
3791        <t hangText="Magic number(s):">none</t>
3792        <t hangText="File extension(s):">none</t>
3793        <t hangText="Macintosh file type code(s):">none</t>
3794      </list>
3795    </t>
3796    <t hangText="Person and email address to contact for further information:">
3797      See Authors Section.
3798    </t>
3799                <t hangText="Intended usage:">
3800                  COMMON
3801    </t>
3802                <t hangText="Restrictions on usage:">
3803                  none
3804    </t>
3805    <t hangText="Author/Change controller:">
3806      IESG
3807    </t>
3808  </list>
3813<section title="Tolerant Applications" anchor="tolerant.applications">
3815   Although this document specifies the requirements for the generation
3816   of HTTP/1.1 messages, not all applications will be correct in their
3817   implementation. We therefore recommend that operational applications
3818   be tolerant of deviations whenever those deviations can be
3819   interpreted unambiguously.
3822   Clients &SHOULD; be tolerant in parsing the Status-Line and servers
3823   tolerant when parsing the Request-Line. In particular, they &SHOULD;
3824   accept any amount of SP or HTAB characters between fields, even though
3825   only a single SP is required.
3828   The line terminator for message-header fields is the sequence CRLF.
3829   However, we recommend that applications, when parsing such headers,
3830   recognize a single LF as a line terminator and ignore the leading CR.
3833   The character set of an entity-body &SHOULD; be labeled as the lowest
3834   common denominator of the character codes used within that body, with
3835   the exception that not labeling the entity is preferred over labeling
3836   the entity with the labels US-ASCII or ISO-8859-1. See &payload;.
3839   Additional rules for requirements on parsing and encoding of dates
3840   and other potential problems with date encodings include:
3843  <list style="symbols">
3844     <t>HTTP/1.1 clients and caches &SHOULD; assume that an RFC-850 date
3845        which appears to be more than 50 years in the future is in fact
3846        in the past (this helps solve the "year 2000" problem).</t>
3848     <t>An HTTP/1.1 implementation &MAY; internally represent a parsed
3849        Expires date as earlier than the proper value, but &MUST-NOT;
3850        internally represent a parsed Expires date as later than the
3851        proper value.</t>
3853     <t>All expiration-related calculations &MUST; be done in GMT. The
3854        local time zone &MUST-NOT; influence the calculation or comparison
3855        of an age or expiration time.</t>
3857     <t>If an HTTP header incorrectly carries a date value with a time
3858        zone other than GMT, it &MUST; be converted into GMT using the
3859        most conservative possible conversion.</t>
3860  </list>
3864<section title="Conversion of Date Formats" anchor="">
3866   HTTP/1.1 uses a restricted set of date formats (<xref target=""/>) to
3867   simplify the process of date comparison. Proxies and gateways from
3868   other protocols &SHOULD; ensure that any Date header field present in a
3869   message conforms to one of the HTTP/1.1 formats and rewrite the date
3870   if necessary.
3874<section title="Compatibility with Previous Versions" anchor="compatibility">
3876   It is beyond the scope of a protocol specification to mandate
3877   compliance with previous versions. HTTP/1.1 was deliberately
3878   designed, however, to make supporting previous versions easy. It is
3879   worth noting that, at the time of composing this specification
3880   (1996), we would expect commercial HTTP/1.1 servers to:
3881  <list style="symbols">
3882     <t>recognize the format of the Request-Line for HTTP/0.9, 1.0, and
3883        1.1 requests;</t>
3885     <t>understand any valid request in the format of HTTP/0.9, 1.0, or
3886        1.1;</t>
3888     <t>respond appropriately with a message in the same major version
3889        used by the client.</t>
3890  </list>
3893   And we would expect HTTP/1.1 clients to:
3894  <list style="symbols">
3895     <t>recognize the format of the Status-Line for HTTP/1.0 and 1.1
3896        responses;</t>
3898     <t>understand any valid response in the format of HTTP/0.9, 1.0, or
3899        1.1.</t>
3900  </list>
3903   For most implementations of HTTP/1.0, each connection is established
3904   by the client prior to the request and closed by the server after
3905   sending the response. Some implementations implement the Keep-Alive
3906   version of persistent connections described in <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/>.
3909<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
3911   This section summarizes major differences between versions HTTP/1.0
3912   and HTTP/1.1.
3915<section title="Changes to Simplify Multi-homed Web Servers and Conserve IP Addresses" anchor="">
3917   The requirements that clients and servers support the Host request-header,
3918   report an error if the Host request-header (<xref target=""/>) is
3919   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-uri"/>)
3920   are among the most important changes defined by this
3921   specification.
3924   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
3925   addresses and servers; there was no other established mechanism for
3926   distinguishing the intended server of a request than the IP address
3927   to which that request was directed. The changes outlined above will
3928   allow the Internet, once older HTTP clients are no longer common, to
3929   support multiple Web sites from a single IP address, greatly
3930   simplifying large operational Web servers, where allocation of many
3931   IP addresses to a single host has created serious problems. The
3932   Internet will also be able to recover the IP addresses that have been
3933   allocated for the sole purpose of allowing special-purpose domain
3934   names to be used in root-level HTTP URLs. Given the rate of growth of
3935   the Web, and the number of servers already deployed, it is extremely
3936   important that all implementations of HTTP (including updates to
3937   existing HTTP/1.0 applications) correctly implement these
3938   requirements:
3939  <list style="symbols">
3940     <t>Both clients and servers &MUST; support the Host request-header.</t>
3942     <t>A client that sends an HTTP/1.1 request &MUST; send a Host header.</t>
3944     <t>Servers &MUST; report a 400 (Bad Request) error if an HTTP/1.1
3945        request does not include a Host request-header.</t>
3947     <t>Servers &MUST; accept absolute URIs.</t>
3948  </list>
3953<section title="Compatibility with HTTP/1.0 Persistent Connections" anchor="compatibility.with.http.1.0.persistent.connections">
3955   Some clients and servers might wish to be compatible with some
3956   previous implementations of persistent connections in HTTP/1.0
3957   clients and servers. Persistent connections in HTTP/1.0 are
3958   explicitly negotiated as they are not the default behavior. HTTP/1.0
3959   experimental implementations of persistent connections are faulty,
3960   and the new facilities in HTTP/1.1 are designed to rectify these
3961   problems. The problem was that some existing 1.0 clients may be
3962   sending Keep-Alive to a proxy server that doesn't understand
3963   Connection, which would then erroneously forward it to the next
3964   inbound server, which would establish the Keep-Alive connection and
3965   result in a hung HTTP/1.0 proxy waiting for the close on the
3966   response. The result is that HTTP/1.0 clients must be prevented from
3967   using Keep-Alive when talking to proxies.
3970   However, talking to proxies is the most important use of persistent
3971   connections, so that prohibition is clearly unacceptable. Therefore,
3972   we need some other mechanism for indicating a persistent connection
3973   is desired, which is safe to use even when talking to an old proxy
3974   that ignores Connection. Persistent connections are the default for
3975   HTTP/1.1 messages; we introduce a new keyword (Connection: close) for
3976   declaring non-persistence. See <xref target="header.connection"/>.
3979   The original HTTP/1.0 form of persistent connections (the Connection:
3980   Keep-Alive and Keep-Alive header) is documented in <xref target="RFC2068"/>.
3984<section title="Changes from RFC 2068" anchor="changes.from.rfc.2068">
3986   This specification has been carefully audited to correct and
3987   disambiguate key word usage; RFC 2068 had many problems in respect to
3988   the conventions laid out in <xref target="RFC2119"/>.
3991   Transfer-coding and message lengths all interact in ways that
3992   required fixing exactly when chunked encoding is used (to allow for
3993   transfer encoding that may not be self delimiting); it was important
3994   to straighten out exactly how message lengths are computed. (Sections
3995   <xref target="transfer.codings" format="counter"/>, <xref target="message.length" format="counter"/>,
3996   <xref target="header.content-length" format="counter"/>,
3997   see also <xref target="Part3"/>, <xref target="Part5"/> and <xref target="Part6"/>)
4000   The use and interpretation of HTTP version numbers has been clarified
4001   by <xref target="RFC2145"/>. Require proxies to upgrade requests to highest protocol
4002   version they support to deal with problems discovered in HTTP/1.0
4003   implementations (<xref target="http.version"/>)
4006   Transfer-coding had significant problems, particularly with
4007   interactions with chunked encoding. The solution is that transfer-codings
4008   become as full fledged as content-codings. This involves
4009   adding an IANA registry for transfer-codings (separate from content
4010   codings), a new header field (TE) and enabling trailer headers in the
4011   future. Transfer encoding is a major performance benefit, so it was
4012   worth fixing <xref target="Nie1997"/>. TE also solves another, obscure, downward
4013   interoperability problem that could have occurred due to interactions
4014   between authentication trailers, chunked encoding and HTTP/1.0
4015   clients.(Section <xref target="transfer.codings" format="counter"/>, <xref target="chunked.transfer.encoding" format="counter"/>,
4016   and <xref target="header.te" format="counter"/>)
4020<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
4022  Clarify that HTTP-Version is case sensitive.
4023  (<xref target="http.version"/>)
4026  Remove reference to non-existant identity transfer-coding value tokens.
4027  (Sections <xref format="counter" target="transfer.codings"/> and
4028  <xref format="counter" target="message.length"/>)
4031  Clarification that the chunk length does not include
4032  the count of the octets in the chunk header and trailer.
4033  (<xref target="chunked.transfer.encoding"/>)
4036  Fix BNF to add query, as the abs_path production in
4037  <xref x:sec="3" x:fmt="of" target="RFC2396"/> doesn't define it.
4038  (<xref target="request-uri"/>)
4041  Clarify exactly when close connection options must be sent.
4042  (<xref target="header.connection"/>)
4047<section title="Change Log (to be removed by RFC Editor before publication)">
4049<section title="Since RFC2616">
4051  Extracted relevant partitions from <xref target="RFC2616"/>.
4055<section title="Since draft-ietf-httpbis-p1-messaging-00">
4057  Closed issues:
4058  <list style="symbols">
4059    <t>
4060      <eref target=""/>:
4061      "HTTP Version should be case sensitive"
4062      (<eref target=""/>)
4063    </t>
4064    <t>
4065      <eref target=""/>:
4066      "'unsafe' characters"
4067      (<eref target=""/>)
4068    </t>
4069    <t>
4070      <eref target=""/>:
4071      "Chunk Size Definition"
4072      (<eref target=""/>)
4073    </t>
4074    <t>
4075      <eref target=""/>:
4076      "Message Length"
4077      (<eref target=""/>)
4078    </t>
4079    <t>
4080      <eref target=""/>:
4081      "Media Type Registrations"
4082      (<eref target=""/>)
4083    </t>
4084    <t>
4085      <eref target=""/>:
4086      "URI includes query"
4087      (<eref target=""/>)
4088    </t>
4089    <t>
4090      <eref target=""/>:
4091      "No close on 1xx responses"
4092      (<eref target=""/>)
4093    </t>
4094    <t>
4095      <eref target=""/>:
4096      "Remove 'identity' token references"
4097      (<eref target=""/>)
4098    </t>
4099    <t>
4100      <eref target=""/>:
4101      "Import query BNF"
4102    </t>
4103    <t>
4104      <eref target=""/>:
4105      "qdtext BNF"
4106    </t>
4107    <t>
4108      <eref target=""/>:
4109      "Normative and Informative references"
4110    </t>
4111    <t>
4112      <eref target=""/>:
4113      "RFC2606 Compliance"
4114    </t>
4115    <t>
4116      <eref target=""/>:
4117      "RFC977 reference"
4118    </t>
4119    <t>
4120      <eref target=""/>:
4121      "RFC1700 references"
4122    </t>
4123    <t>
4124      <eref target=""/>:
4125      "inconsistency in date format explanation"
4126    </t>
4127    <t>
4128      <eref target=""/>:
4129      "Date reference typo"
4130    </t>
4131    <t>
4132      <eref target=""/>:
4133      "Informative references"
4134    </t>
4135    <t>
4136      <eref target=""/>:
4137      "ISO-8859-1 Reference"
4138    </t>
4139    <t>
4140      <eref target=""/>:
4141      "Normative up-to-date references"
4142    </t>
4143  </list>
4146  Other changes:
4147  <list style="symbols">
4148    <t>
4149      Update media type registrations to use RFC4288 template.
4150    </t>
4151    <t>
4152      Use names of RFC4234 core rules DQUOTE and HTAB,
4153      fix broken ABNF for chunk-data
4154      (work in progress on <eref target=""/>)
4155    </t>
4156  </list>
4160<section title="Since draft-ietf-httpbis-p1-messaging-01">
4162  Closed issues:
4163  <list style="symbols">
4164    <t>
4165      <eref target=""/>:
4166      "Bodies on GET (and other) requests"
4167    </t>
4168    <t>
4169      <eref target=""/>:
4170      "rel_path not used"
4171    </t>
4172  </list>
Note: See TracBrowser for help on using the repository browser.