source: draft-ietf-httpbis/01/draft-ietf-httpbis-p1-messaging-01.xml @ 2762

Last change on this file since 2762 was 1500, checked in by julian.reschke@…, 11 years ago

fix mime types

  • Property svn:eol-style set to native
  • Property svn:mime-type set to text/xml
File size: 171.5 KB
1<?xml version="1.0" encoding="UTF-8"?>
3    This XML document is the output of clean-for-DTD.xslt; a tool that strips
4    extensions to RFC2629(bis) from documents for processing with xml2rfc.
6<?xml-stylesheet type='text/xsl' href='../myxml2rfc.xslt'?>
7<?rfc toc="yes" ?>
8<?rfc symrefs="yes" ?>
9<?rfc sortrefs="yes" ?>
10<?rfc compact="yes"?>
11<?rfc subcompact="no" ?>
12<?rfc linkmailto="no" ?>
13<?rfc editing="no" ?>
14<!DOCTYPE rfc
15  PUBLIC "" "rfc2629.dtd">
16<rfc obsoletes="2616" category="std" ipr="full3978" docName="draft-ietf-httpbis-p1-messaging-01">
19  <title abbrev="HTTP/1.1, Part 1">HTTP/1.1, part 1: URIs, Connections, and Message Parsing</title>
21  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
22    <organization abbrev="Day Software">Day Software</organization>
23    <address>
24      <postal>
25        <street>23 Corporate Plaza DR, Suite 280</street>
26        <city>Newport Beach</city>
27        <region>CA</region>
28        <code>92660</code>
29        <country>USA</country>
30      </postal>
31      <phone>+1-949-706-5300</phone>
32      <facsimile>+1-949-706-5305</facsimile>
33      <email></email>
34      <uri></uri>
35    </address>
36  </author>
38  <author initials="J." surname="Gettys" fullname="Jim Gettys">
39    <organization>One Laptop per Child</organization>
40    <address>
41      <postal>
42        <street>21 Oak Knoll Road</street>
43        <city>Carlisle</city>
44        <region>MA</region>
45        <code>01741</code>
46        <country>USA</country>
47      </postal>
48      <email></email>
49      <uri></uri>
50    </address>
51  </author>
53  <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
54    <organization abbrev="HP">Hewlett-Packard Company</organization>
55    <address>
56      <postal>
57        <street>HP Labs, Large Scale Systems Group</street>
58        <street>1501 Page Mill Road, MS 1177</street>
59        <city>Palo Alto</city>
60        <region>CA</region>
61        <code>94304</code>
62        <country>USA</country>
63      </postal>
64      <email></email>
65    </address>
66  </author>
68  <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
69    <organization abbrev="Microsoft">Microsoft Corporation</organization>
70    <address>
71      <postal>
72        <street>1 Microsoft Way</street>
73        <city>Redmond</city>
74        <region>WA</region>
75        <code>98052</code>
76        <country>USA</country>
77      </postal>
78      <email></email>
79    </address>
80  </author>
82  <author initials="L." surname="Masinter" fullname="Larry Masinter">
83    <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
84    <address>
85      <postal>
86        <street>345 Park Ave</street>
87        <city>San Jose</city>
88        <region>CA</region>
89        <code>95110</code>
90        <country>USA</country>
91      </postal>
92      <email></email>
93      <uri></uri>
94    </address>
95  </author>
97  <author initials="P." surname="Leach" fullname="Paul J. Leach">
98    <organization abbrev="Microsoft">Microsoft Corporation</organization>
99    <address>
100      <postal>
101        <street>1 Microsoft Way</street>
102        <city>Redmond</city>
103        <region>WA</region>
104        <code>98052</code>
105      </postal>
106      <email></email>
107    </address>
108  </author>
110  <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
111    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
112    <address>
113      <postal>
114        <street>MIT Computer Science and Artificial Intelligence Laboratory</street>
115        <street>The Stata Center, Building 32</street>
116        <street>32 Vassar Street</street>
117        <city>Cambridge</city>
118        <region>MA</region>
119        <code>02139</code>
120        <country>USA</country>
121      </postal>
122      <email></email>
123      <uri></uri>
124    </address>
125  </author>
127  <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
128    <organization abbrev="W3C">World Wide Web Consortium</organization>
129    <address>
130      <postal>
131        <street>W3C / ERCIM</street>
132        <street>2004, rte des Lucioles</street>
133        <city>Sophia-Antipolis</city>
134        <region>AM</region>
135        <code>06902</code>
136        <country>France</country>
137      </postal>
138      <email></email>
139      <uri></uri>
140    </address>
141  </author>
143  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
144    <organization abbrev="greenbytes">greenbytes GmbH</organization>
145    <address>
146      <postal>
147        <street>Hafenweg 16</street>
148        <city>Muenster</city><region>NW</region><code>48155</code>
149        <country>Germany</country>
150      </postal>
151      <phone>+49 251 2807760</phone>   
152      <facsimile>+49 251 2807761</facsimile>   
153      <email></email>       
154      <uri></uri>     
155    </address>
156  </author>
158  <date month="January" year="2008" day="12"/>
162   The Hypertext Transfer Protocol (HTTP) is an application-level
163   protocol for distributed, collaborative, hypermedia information
164   systems. HTTP has been in use by the World Wide Web global information
165   initiative since 1990. This document is Part 1 of the seven-part specification
166   that defines the protocol referred to as "HTTP/1.1" and, taken together,
167   obsoletes RFC 2616.  Part 1 provides an overview of HTTP and
168   its associated terminology, defines the "http" and "https" Uniform
169   Resource Identifier (URI) schemes, defines the generic message syntax
170   and parsing requirements for HTTP message frames, and describes
171   general security concerns for implementations.
175<note title="Editorial Note (To be removed by RFC Editor)">
176  <t>
177    Discussion of this draft should take place on the HTTPBIS working group
178    mailing list ( The current issues list is
179    at <eref target=""/>
180    and related documents (including fancy diffs) can be found at
181    <eref target=""/>.
182  </t>
183  <t>
184    This draft incorporates those issue resolutions that were either
185    collected in the original RFC2616 errata list (<eref target=""/>),
186    or which were agreed upon on the mailing list between October 2006 and
187    November 2007 (as published in "draft-lafon-rfc2616bis-03").
188  </t>
192<section title="Introduction" anchor="introduction">
194   The Hypertext Transfer Protocol (HTTP) is an application-level
195   protocol for distributed, collaborative, hypermedia information
196   systems. HTTP has been in use by the World-Wide Web global
197   information initiative since 1990. The first version of HTTP, commonly
198   referred to as HTTP/0.9, was a simple protocol for raw data transfer
199   across the Internet with only a single method and no metadata.
200   HTTP/1.0, as defined by <xref target="RFC1945"/>, improved
201   the protocol by allowing messages to be in the format of MIME-like
202   messages, containing metadata about the data transferred and
203   modifiers on the request/response semantics. However, HTTP/1.0 did
204   not sufficiently take into consideration the effects of hierarchical
205   proxies, caching, the need for persistent connections, or name-based
206   virtual hosts. In addition, the proliferation of incompletely-implemented
207   applications calling themselves "HTTP/1.0" necessitated a
208   protocol version change in order for two communicating applications
209   to determine each other's true capabilities.
212   This document is Part 1 of the seven-part specification that defines
213   the protocol referred to as "HTTP/1.1", obsoleting <xref target="RFC2616"/>.
214   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
215   requirements that enable reliable implementations and adding only
216   those new features that will either be safely ignored by an HTTP/1.0
217   recipient or only sent when communicating with a party advertising
218   compliance with HTTP/1.1.
219   Part 1 defines those aspects of HTTP/1.1 related to overall network
220   operation, message framing, interaction with transport protocols, and
221   URI schemes.
224   This document is currently disorganized in order to minimize the changes
225   between drafts and enable reviewers to see the smaller errata changes.
226   The next draft will reorganize the sections to better reflect the content.
227   In particular, the sections will be organized according to the typical
228   process of deciding when to use HTTP (URI schemes), overall network operation,
229   connection management, message framing, and generic message parsing.
230   The current mess reflects how widely dispersed these topics and associated
231   requirements had become in <xref target="RFC2616"/>.
234<section title="Purpose" anchor="intro.purpose">
236   Practical information systems require more functionality than simple
237   retrieval, including search, front-end update, and annotation. HTTP
238   allows an open-ended set of methods and headers that indicate the
239   purpose of a request <xref target="RFC2324"/>. It builds on the discipline of reference
240   provided by the Uniform Resource Identifier (URI) <xref target="RFC1630"/>, as a location
241   (URL) <xref target="RFC1738"/> or name (URN) <xref target="RFC1737"/>, for indicating the resource to which a
242   method is to be applied. Messages are passed in a format similar to
243   that used by Internet mail <xref target="RFC2822"/> as defined by the Multipurpose
244   Internet Mail Extensions (MIME) <xref target="RFC2045"/>.
247   HTTP is also used as a generic protocol for communication between
248   user agents and proxies/gateways to other Internet systems, including
249   those supported by the SMTP <xref target="RFC2821"/>, NNTP <xref target="RFC3977"/>, FTP <xref target="RFC959"/>, Gopher <xref target="RFC1436"/>,
250   and WAIS <xref target="WAIS"/> protocols. In this way, HTTP allows basic hypermedia
251   access to resources available from diverse applications.
255<section title="Requirements" anchor="intro.requirements">
257   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
258   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
259   document are to be interpreted as described in <xref target="RFC2119"/>.
262   An implementation is not compliant if it fails to satisfy one or more
263   of the MUST or REQUIRED level requirements for the protocols it
264   implements. An implementation that satisfies all the MUST or REQUIRED
265   level and all the SHOULD level requirements for its protocols is said
266   to be "unconditionally compliant"; one that satisfies all the MUST
267   level requirements but not all the SHOULD level requirements for its
268   protocols is said to be "conditionally compliant."
272<section title="Terminology" anchor="intro.terminology">
274   This specification uses a number of terms to refer to the roles
275   played by participants in, and objects of, the HTTP communication.
278  <iref item="connection"/>
279  connection
280  <list>
281    <t>
282      A transport layer virtual circuit established between two programs
283      for the purpose of communication.
284    </t>
285  </list>
288  <iref item="message"/>
289  message
290  <list>
291    <t>
292      The basic unit of HTTP communication, consisting of a structured
293      sequence of octets matching the syntax defined in <xref target="http.message"/> and
294      transmitted via the connection.
295    </t>
296  </list>
299  <iref item="request"/>
300  request
301  <list>
302    <t>
303      An HTTP request message, as defined in <xref target="request"/>.
304    </t>
305  </list>
308  <iref item="response"/>
309  response
310  <list>
311    <t>
312      An HTTP response message, as defined in <xref target="response"/>.
313    </t>
314  </list>
317  <iref item="resource"/>
318  resource
319  <list>
320    <t>
321      A network data object or service that can be identified by a URI,
322      as defined in <xref target="uri"/>. Resources may be available in multiple
323      representations (e.g. multiple languages, data formats, size, and
324      resolutions) or vary in other ways.
325    </t>
326  </list>
329  <iref item="entity"/>
330  entity
331  <list>
332    <t>
333      The information transferred as the payload of a request or
334      response. An entity consists of metainformation in the form of
335      entity-header fields and content in the form of an entity-body, as
336      described in Section 3 of <xref target="Part3"/>.
337    </t>
338  </list>
341  <iref item="representation"/>
342  representation
343  <list>
344    <t>
345      An entity included with a response that is subject to content
346      negotiation, as described in Section 4 of <xref target="Part3"/>. There may exist multiple
347      representations associated with a particular response status.
348    </t>
349  </list>
352  <iref item="content negotiation"/>
353  content negotiation
354  <list>
355    <t>
356      The mechanism for selecting the appropriate representation when
357      servicing a request, as described in Section 4 of <xref target="Part3"/>. The
358      representation of entities in any response can be negotiated
359      (including error responses).
360    </t>
361  </list>
364  <iref item="variant"/>
365  variant
366  <list>
367    <t>
368      A resource may have one, or more than one, representation(s)
369      associated with it at any given instant. Each of these
370      representations is termed a `variant'.  Use of the term `variant'
371      does not necessarily imply that the resource is subject to content
372      negotiation.
373    </t>
374  </list>
377  <iref item="client"/>
378  client
379  <list>
380    <t>
381      A program that establishes connections for the purpose of sending
382      requests.
383    </t>
384  </list>
387  <iref item="user agent"/>
388  user agent
389  <list>
390    <t>
391      The client which initiates a request. These are often browsers,
392      editors, spiders (web-traversing robots), or other end user tools.
393    </t>
394  </list>
397  <iref item="server"/>
398  server
399  <list>
400    <t>
401      An application program that accepts connections in order to
402      service requests by sending back responses. Any given program may
403      be capable of being both a client and a server; our use of these
404      terms refers only to the role being performed by the program for a
405      particular connection, rather than to the program's capabilities
406      in general. Likewise, any server may act as an origin server,
407      proxy, gateway, or tunnel, switching behavior based on the nature
408      of each request.
409    </t>
410  </list>
413  <iref item="origin server"/>
414  origin server
415  <list>
416    <t>
417      The server on which a given resource resides or is to be created.
418    </t>
419  </list>
422  <iref item="proxy"/>
423  proxy
424  <list>
425    <t>
426      An intermediary program which acts as both a server and a client
427      for the purpose of making requests on behalf of other clients.
428      Requests are serviced internally or by passing them on, with
429      possible translation, to other servers. A proxy MUST implement
430      both the client and server requirements of this specification. A
431      "transparent proxy" is a proxy that does not modify the request or
432      response beyond what is required for proxy authentication and
433      identification. A "non-transparent proxy" is a proxy that modifies
434      the request or response in order to provide some added service to
435      the user agent, such as group annotation services, media type
436      transformation, protocol reduction, or anonymity filtering. Except
437      where either transparent or non-transparent behavior is explicitly
438      stated, the HTTP proxy requirements apply to both types of
439      proxies.
440    </t>
441  </list>
444  <iref item="gateway"/>
445  gateway
446  <list>
447    <t>
448      A server which acts as an intermediary for some other server.
449      Unlike a proxy, a gateway receives requests as if it were the
450      origin server for the requested resource; the requesting client
451      may not be aware that it is communicating with a gateway.
452    </t>
453  </list>
456  <iref item="tunnel"/>
457  tunnel
458  <list>
459    <t>
460      An intermediary program which is acting as a blind relay between
461      two connections. Once active, a tunnel is not considered a party
462      to the HTTP communication, though the tunnel may have been
463      initiated by an HTTP request. The tunnel ceases to exist when both
464      ends of the relayed connections are closed.
465    </t>
466  </list>
469  <iref item="cache"/>
470  cache
471  <list>
472    <t>
473      A program's local store of response messages and the subsystem
474      that controls its message storage, retrieval, and deletion. A
475      cache stores cacheable responses in order to reduce the response
476      time and network bandwidth consumption on future, equivalent
477      requests. Any client or server may include a cache, though a cache
478      cannot be used by a server that is acting as a tunnel.
479    </t>
480  </list>
483  <iref item="cacheable"/>
484  cacheable
485  <list>
486    <t>
487      A response is cacheable if a cache is allowed to store a copy of
488      the response message for use in answering subsequent requests. The
489      rules for determining the cacheability of HTTP responses are
490      defined in Section 1 of <xref target="Part6"/>. Even if a resource is cacheable, there may
491      be additional constraints on whether a cache can use the cached
492      copy for a particular request.
493    </t>
494  </list>
497  <iref item="upstream"/>
498  <iref item="downstream"/>
499  upstream/downstream
500  <list>
501    <t>
502      Upstream and downstream describe the flow of a message: all
503      messages flow from upstream to downstream.
504    </t>
505  </list>
508  <iref item="inbound"/>
509  <iref item="outbound"/>
510  inbound/outbound
511  <list>
512    <t>
513      Inbound and outbound refer to the request and response paths for
514      messages: "inbound" means "traveling toward the origin server",
515      and "outbound" means "traveling toward the user agent"
516    </t>
517  </list>
521<section title="Overall Operation" anchor="intro.overall.operation">
523   The HTTP protocol is a request/response protocol. A client sends a
524   request to the server in the form of a request method, URI, and
525   protocol version, followed by a MIME-like message containing request
526   modifiers, client information, and possible body content over a
527   connection with a server. The server responds with a status line,
528   including the message's protocol version and a success or error code,
529   followed by a MIME-like message containing server information, entity
530   metainformation, and possible entity-body content. The relationship
531   between HTTP and MIME is described in Appendix A of <xref target="Part3"/>.
534   Most HTTP communication is initiated by a user agent and consists of
535   a request to be applied to a resource on some origin server. In the
536   simplest case, this may be accomplished via a single connection (v)
537   between the user agent (UA) and the origin server (O).
539<figure><artwork type="drawing"><![CDATA[
540       request chain ------------------------>
541    UA -------------------v------------------- O
542       <----------------------- response chain
545   A more complicated situation occurs when one or more intermediaries
546   are present in the request/response chain. There are three common
547   forms of intermediary: proxy, gateway, and tunnel. A proxy is a
548   forwarding agent, receiving requests for a URI in its absolute form,
549   rewriting all or part of the message, and forwarding the reformatted
550   request toward the server identified by the URI. A gateway is a
551   receiving agent, acting as a layer above some other server(s) and, if
552   necessary, translating the requests to the underlying server's
553   protocol. A tunnel acts as a relay point between two connections
554   without changing the messages; tunnels are used when the
555   communication needs to pass through an intermediary (such as a
556   firewall) even when the intermediary cannot understand the contents
557   of the messages.
559<figure><artwork type="drawing"><![CDATA[
560       request chain -------------------------------------->
561    UA -----v----- A -----v----- B -----v----- C -----v----- O
562       <------------------------------------- response chain
565   The figure above shows three intermediaries (A, B, and C) between the
566   user agent and origin server. A request or response message that
567   travels the whole chain will pass through four separate connections.
568   This distinction is important because some HTTP communication options
569   may apply only to the connection with the nearest, non-tunnel
570   neighbor, only to the end-points of the chain, or to all connections
571   along the chain. Although the diagram is linear, each participant may
572   be engaged in multiple, simultaneous communications. For example, B
573   may be receiving requests from many clients other than A, and/or
574   forwarding requests to servers other than C, at the same time that it
575   is handling A's request.
578   Any party to the communication which is not acting as a tunnel may
579   employ an internal cache for handling requests. The effect of a cache
580   is that the request/response chain is shortened if one of the
581   participants along the chain has a cached response applicable to that
582   request. The following illustrates the resulting chain if B has a
583   cached copy of an earlier response from O (via C) for a request which
584   has not been cached by UA or A.
586<figure><artwork type="drawing"><![CDATA[
587          request chain ---------->
588       UA -----v----- A -----v----- B - - - - - - C - - - - - - O
589          <--------- response chain
592   Not all responses are usefully cacheable, and some requests may
593   contain modifiers which place special requirements on cache behavior.
594   HTTP requirements for cache behavior and cacheable responses are
595   defined in Section 1 of <xref target="Part6"/>.
598   In fact, there are a wide variety of architectures and configurations
599   of caches and proxies currently being experimented with or deployed
600   across the World Wide Web. These systems include national hierarchies
601   of proxy caches to save transoceanic bandwidth, systems that
602   broadcast or multicast cache entries, organizations that distribute
603   subsets of cached data via CD-ROM, and so on. HTTP systems are used
604   in corporate intranets over high-bandwidth links, and for access via
605   PDAs with low-power radio links and intermittent connectivity. The
606   goal of HTTP/1.1 is to support the wide diversity of configurations
607   already deployed while introducing protocol constructs that meet the
608   needs of those who build web applications that require high
609   reliability and, failing that, at least reliable indications of
610   failure.
613   HTTP communication usually takes place over TCP/IP connections. The
614   default port is TCP 80 (<eref target=""/>), but other ports can be used. This does
615   not preclude HTTP from being implemented on top of any other protocol
616   on the Internet, or on other networks. HTTP only presumes a reliable
617   transport; any protocol that provides such guarantees can be used;
618   the mapping of the HTTP/1.1 request and response structures onto the
619   transport data units of the protocol in question is outside the scope
620   of this specification.
623   In HTTP/1.0, most implementations used a new connection for each
624   request/response exchange. In HTTP/1.1, a connection may be used for
625   one or more request/response exchanges, although connections may be
626   closed for a variety of reasons (see <xref target="persistent.connections"/>).
631<section title="Notational Conventions and Generic Grammar" anchor="notation">
633<section title="Augmented BNF" anchor="notation.abnf">
635   All of the mechanisms specified in this document are described in
636   both prose and an augmented Backus-Naur Form (BNF) similar to that
637   used by <xref target="RFC822ABNF"/>. Implementors will need to be familiar with the
638   notation in order to understand this specification. The augmented BNF
639   includes the following constructs:
642   name = definition
643  <list>
644    <t>
645      The name of a rule is simply the name itself (without any
646      enclosing "&lt;" and "&gt;") and is separated from its definition by the
647      equal "=" character. White space is only significant in that
648      indentation of continuation lines is used to indicate a rule
649      definition that spans more than one line. Certain basic rules are
650      in uppercase, such as SP, LWS, HTAB, CRLF, DIGIT, ALPHA, etc. Angle
651      brackets are used within definitions whenever their presence will
652      facilitate discerning the use of rule names.
653    </t>
654  </list>
657   "literal"
658  <list>
659    <t>
660      Quotation marks surround literal text. Unless stated otherwise,
661      the text is case-insensitive.
662    </t>
663  </list>
666   rule1 | rule2
667  <list>
668    <t>
669      Elements separated by a bar ("|") are alternatives, e.g., "yes |
670      no" will accept yes or no.
671    </t>
672  </list>
675   (rule1 rule2)
676  <list>
677    <t>
678      Elements enclosed in parentheses are treated as a single element.
679      Thus, "(elem (foo | bar) elem)" allows the token sequences "elem
680      foo elem" and "elem bar elem".
681    </t>
682  </list>
685   *rule
686  <list>
687    <t>
688      The character "*" preceding an element indicates repetition. The
689      full form is "&lt;n&gt;*&lt;m&gt;element" indicating at least &lt;n&gt; and at most
690      &lt;m&gt; occurrences of element. Default values are 0 and infinity so
691      that "*(element)" allows any number, including zero; "1*element"
692      requires at least one; and "1*2element" allows one or two.
693    </t>
694  </list>
697   [rule]
698  <list>
699    <t>
700      Square brackets enclose optional elements; "[foo bar]" is
701      equivalent to "*1(foo bar)".
702    </t>
703  </list>
706   N rule
707  <list>
708    <t>
709      Specific repetition: "&lt;n&gt;(element)" is equivalent to
710      "&lt;n&gt;*&lt;n&gt;(element)"; that is, exactly &lt;n&gt; occurrences of (element).
711      Thus 2DIGIT is a 2-digit number, and 3ALPHA is a string of three
712      alphabetic characters.
713    </t>
714  </list>
717   #rule
718  <list>
719    <t>
720      A construct "#" is defined, similar to "*", for defining lists of
721      elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating at least
722      &lt;n&gt; and at most &lt;m&gt; elements, each separated by one or more commas
723      (",") and OPTIONAL linear white space (LWS). This makes the usual
724      form of lists very easy; a rule such as
725    </t>
726    <t>
727         ( *LWS element *( *LWS "," *LWS element ))
728    </t>
729    <t>
730      can be shown as
731    </t>
732    <t>
733         1#element
734    </t>
735    <t>
736      Wherever this construct is used, null elements are allowed, but do
737      not contribute to the count of elements present. That is,
738      "(element), , (element) " is permitted, but counts as only two
739      elements. Therefore, where at least one element is required, at
740      least one non-null element MUST be present. Default values are 0
741      and infinity so that "#element" allows any number, including zero;
742      "1#element" requires at least one; and "1#2element" allows one or
743      two.
744    </t>
745  </list>
748   ; comment
749  <list>
750    <t>
751      A semi-colon, set off some distance to the right of rule text,
752      starts a comment that continues to the end of line. This is a
753      simple way of including useful notes in parallel with the
754      specifications.
755    </t>
756  </list>
759   implied *LWS
760  <list>
761    <t>
762      The grammar described by this specification is word-based. Except
763      where noted otherwise, linear white space (LWS) can be included
764      between any two adjacent words (token or quoted-string), and
765      between adjacent words and separators, without changing the
766      interpretation of a field. At least one delimiter (LWS and/or
767      separators) MUST exist between any two tokens (for the definition
768      of "token" below), since they would otherwise be interpreted as a
769      single token.
770    </t>
771  </list>
775<section title="Basic Rules" anchor="basic.rules">
799   The following rules are used throughout this specification to
800   describe basic parsing constructs. The US-ASCII coded character set
801   is defined by ANSI X3.4-1986 <xref target="USASCII"/>.
803<figure><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"/><artwork type="abnf2616"><![CDATA[
804  OCTET          = <any 8-bit sequence of data>
805  CHAR           = <any US-ASCII character (octets 0 - 127)>
806  UPALPHA        = <any US-ASCII uppercase letter "A".."Z">
807  LOALPHA        = <any US-ASCII lowercase letter "a".."z">
808  ALPHA          = UPALPHA | LOALPHA
809  DIGIT          = <any US-ASCII digit "0".."9">
810  CTL            = <any US-ASCII control character
811                   (octets 0 - 31) and DEL (127)>
812  CR             = <US-ASCII CR, carriage return (13)>
813  LF             = <US-ASCII LF, linefeed (10)>
814  SP             = <US-ASCII SP, space (32)>
815  HTAB           = <US-ASCII HT, horizontal-tab (9)>
816  DQUOTE         = <US-ASCII double-quote mark (34)>
819   HTTP/1.1 defines the sequence CR LF as the end-of-line marker for all
820   protocol elements except the entity-body (see <xref target="tolerant.applications"/> for
821   tolerant applications). The end-of-line marker within an entity-body
822   is defined by its associated media type, as described in Section 2.3 of <xref target="Part3"/>.
824<figure><iref primary="true" item="Grammar" subitem="CRLF"/><artwork type="abnf2616"><![CDATA[
825  CRLF           = CR LF
828   HTTP/1.1 header field values can be folded onto multiple lines if the
829   continuation line begins with a space or horizontal tab. All linear
830   white space, including folding, has the same semantics as SP. A
831   recipient MAY replace any linear white space with a single SP before
832   interpreting the field value or forwarding the message downstream.
834<figure><iref primary="true" item="Grammar" subitem="LWS"/><artwork type="abnf2616"><![CDATA[
835  LWS            = [CRLF] 1*( SP | HTAB )
838   The TEXT rule is only used for descriptive field contents and values
839   that are not intended to be interpreted by the message parser. Words
840   of *TEXT MAY contain characters from character sets other than ISO-8859-1
841   <xref target="ISO-8859-1"/> only when encoded according to the rules of
842   <xref target="RFC2047"/>.
844<figure><iref primary="true" item="Grammar" subitem="TEXT"/><artwork type="abnf2616"><![CDATA[
845  TEXT           = <any OCTET except CTLs,
846                   but including LWS>
849   A CRLF is allowed in the definition of TEXT only as part of a header
850   field continuation. It is expected that the folding LWS will be
851   replaced with a single SP before interpretation of the TEXT value.
854   Hexadecimal numeric characters are used in several protocol elements.
856<figure><iref primary="true" item="Grammar" subitem="HEX"/><artwork type="abnf2616"><![CDATA[
857  HEX            = "A" | "B" | "C" | "D" | "E" | "F"
858                 | "a" | "b" | "c" | "d" | "e" | "f" | DIGIT
861   Many HTTP/1.1 header field values consist of words separated by LWS
862   or special characters. These special characters MUST be in a quoted
863   string to be used within a parameter value (as defined in
864   <xref target="transfer.codings"/>).
866<figure><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="separators"/><artwork type="abnf2616"><![CDATA[
867  token          = 1*<any CHAR except CTLs or separators>
868  separators     = "(" | ")" | "<" | ">" | "@"
869                 | "," | ";" | ":" | "\" | DQUOTE
870                 | "/" | "[" | "]" | "?" | "="
871                 | "{" | "}" | SP | HTAB
874   Comments can be included in some HTTP header fields by surrounding
875   the comment text with parentheses. Comments are only allowed in
876   fields containing "comment" as part of their field value definition.
877   In all other fields, parentheses are considered part of the field
878   value.
880<figure><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/><artwork type="abnf2616"><![CDATA[
881  comment        = "(" *( ctext | quoted-pair | comment ) ")"
882  ctext          = <any TEXT excluding "(" and ")">
885   A string of text is parsed as a single word if it is quoted using
886   double-quote marks.
888<figure><iref primary="true" item="Grammar" subitem="quoted-string"/><iref primary="true" item="Grammar" subitem="qdtext"/><artwork type="abnf2616"><![CDATA[
889  quoted-string  = ( DQUOTE *(qdtext | quoted-pair ) DQUOTE )
890  qdtext         = <any TEXT excluding DQUOTE and "\">
893   The backslash character ("\") MAY be used as a single-character
894   quoting mechanism only within quoted-string and comment constructs.
896<figure><iref primary="true" item="Grammar" subitem="quoted-pair"/><artwork type="abnf2616"><![CDATA[
897  quoted-pair    = "\" CHAR
902<section title="Protocol Parameters" anchor="protocol.parameters">
904<section title="HTTP Version" anchor="http.version">
906   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate versions
907   of the protocol. The protocol versioning policy is intended to allow
908   the sender to indicate the format of a message and its capacity for
909   understanding further HTTP communication, rather than the features
910   obtained via that communication. No change is made to the version
911   number for the addition of message components which do not affect
912   communication behavior or which only add to extensible field values.
913   The &lt;minor&gt; number is incremented when the changes made to the
914   protocol add features which do not change the general message parsing
915   algorithm, but which may add to the message semantics and imply
916   additional capabilities of the sender. The &lt;major&gt; number is
917   incremented when the format of a message within the protocol is
918   changed. See <xref target="RFC2145"/> for a fuller explanation.
921   The version of an HTTP message is indicated by an HTTP-Version field
922   in the first line of the message. HTTP-Version is case-sensitive.
924<figure><iref primary="true" item="Grammar" subitem="HTTP-Version"/><artwork type="abnf2616"><![CDATA[
925  HTTP-Version   = "HTTP" "/" 1*DIGIT "." 1*DIGIT
928   Note that the major and minor numbers MUST be treated as separate
929   integers and that each MAY be incremented higher than a single digit.
930   Thus, HTTP/2.4 is a lower version than HTTP/2.13, which in turn is
931   lower than HTTP/12.3. Leading zeros MUST be ignored by recipients and
932   MUST NOT be sent.
935   An application that sends a request or response message that includes
936   HTTP-Version of "HTTP/1.1" MUST be at least conditionally compliant
937   with this specification. Applications that are at least conditionally
938   compliant with this specification SHOULD use an HTTP-Version of
939   "HTTP/1.1" in their messages, and MUST do so for any message that is
940   not compatible with HTTP/1.0. For more details on when to send
941   specific HTTP-Version values, see <xref target="RFC2145"/>.
944   The HTTP version of an application is the highest HTTP version for
945   which the application is at least conditionally compliant.
948   Proxy and gateway applications need to be careful when forwarding
949   messages in protocol versions different from that of the application.
950   Since the protocol version indicates the protocol capability of the
951   sender, a proxy/gateway MUST NOT send a message with a version
952   indicator which is greater than its actual version. If a higher
953   version request is received, the proxy/gateway MUST either downgrade
954   the request version, or respond with an error, or switch to tunnel
955   behavior.
958   Due to interoperability problems with HTTP/1.0 proxies discovered
959   since the publication of <xref target="RFC2068"/>, caching proxies MUST, gateways
960   MAY, and tunnels MUST NOT upgrade the request to the highest version
961   they support. The proxy/gateway's response to that request MUST be in
962   the same major version as the request.
965  <list>
966    <t>
967      Note: Converting between versions of HTTP may involve modification
968      of header fields required or forbidden by the versions involved.
969    </t>
970  </list>
974<section title="Uniform Resource Identifiers" anchor="uri">
976   URIs have been known by many names: WWW addresses, Universal Document
977   Identifiers, Universal Resource Identifiers <xref target="RFC1630"/>, and finally the
978   combination of Uniform Resource Locators (URL) <xref target="RFC1738"/> and Names (URN)
979   <xref target="RFC1737"/>. As far as HTTP is concerned, Uniform Resource Identifiers are
980   simply formatted strings which identify--via name, location, or any
981   other characteristic--a resource.
984<section title="General Syntax" anchor="general.syntax">
986   URIs in HTTP can be represented in absolute form or relative to some
987   known base URI <xref target="RFC1808"/>, depending upon the context of their use. The two
988   forms are differentiated by the fact that absolute URIs always begin
989   with a scheme name followed by a colon. For definitive information on
990   URL syntax and semantics, see "Uniform Resource Identifiers (URI):
991   Generic Syntax and Semantics," <xref target="RFC2396"/> (which replaces <xref target="RFC1738"/>
992   and <xref target="RFC1808"/>). This specification adopts the
993   definitions of "URI-reference", "absoluteURI", "relativeURI", "port",
994   "host", "abs_path", "rel_path", "query", and "authority" from that
995   specification.
998   The HTTP protocol does not place any a priori limit on the length of
999   a URI. Servers MUST be able to handle the URI of any resource they
1000   serve, and SHOULD be able to handle URIs of unbounded length if they
1001   provide GET-based forms that could generate such URIs. A server
1002   SHOULD return 414 (Request-URI Too Long) status if a URI is longer
1003   than the server can handle (see Section 9.4.15 of <xref target="Part2"/>).
1006  <list>
1007    <t>
1008      Note: Servers ought to be cautious about depending on URI lengths
1009      above 255 bytes, because some older client or proxy
1010      implementations might not properly support these lengths.
1011    </t>
1012  </list>
1016<section title="http URL" anchor="http.url">
1018   The "http" scheme is used to locate network resources via the HTTP
1019   protocol. This section defines the scheme-specific syntax and
1020   semantics for http URLs.
1022<figure><iref primary="true" item="Grammar" subitem="http_URL"/><artwork type="abnf2616"><![CDATA[
1023  http_URL = "http:" "//" host [ ":" port ] [ abs_path [ "?" query ]]
1026   If the port is empty or not given, port 80 is assumed. The semantics
1027   are that the identified resource is located at the server listening
1028   for TCP connections on that port of that host, and the Request-URI
1029   for the resource is abs_path (<xref target="request-uri"/>). The use of IP addresses
1030   in URLs SHOULD be avoided whenever possible (see <xref target="RFC1900"/>). If
1031   the abs_path is not present in the URL, it MUST be given as "/" when
1032   used as a Request-URI for a resource (<xref target="request-uri"/>). If a proxy
1034   receives a host name which is not a fully qualified domain name, it
1035   MAY add its domain to the host name it received. If a proxy receives
1036   a fully qualified domain name, the proxy MUST NOT change the host
1037   name.
1041<section title="URI Comparison" anchor="uri.comparison">
1043   When comparing two URIs to decide if they match or not, a client
1044   SHOULD use a case-sensitive octet-by-octet comparison of the entire
1045   URIs, with these exceptions:
1046  <list style="symbols">
1047    <t>A port that is empty or not given is equivalent to the default
1048        port for that URI-reference;</t>
1049    <t>Comparisons of host names MUST be case-insensitive;</t>
1050    <t>Comparisons of scheme names MUST be case-insensitive;</t>
1051    <t>An empty abs_path is equivalent to an abs_path of "/".</t>
1052  </list>
1055   Characters other than those in the "reserved" set (see
1056   <xref target="RFC2396"/>) are equivalent to their ""%" HEX HEX" encoding.
1059   For example, the following three URIs are equivalent:
1061<figure><artwork type="example"><![CDATA[
1069<section title="Date/Time Formats" anchor="date.time.formats">
1070<section title="Full Date" anchor="">
1072   HTTP applications have historically allowed three different formats
1073   for the representation of date/time stamps:
1075<figure><artwork type="example"><![CDATA[
1076   Sun, 06 Nov 1994 08:49:37 GMT  ; RFC 822, updated by RFC 1123
1077   Sunday, 06-Nov-94 08:49:37 GMT ; obsolete RFC 850 format
1078   Sun Nov  6 08:49:37 1994       ; ANSI C's asctime() format
1081   The first format is preferred as an Internet standard and represents
1082   a fixed-length subset of that defined by <xref target="RFC1123"/> (an update to
1083   <xref target="RFC822"/>). The other formats are described here only for
1084   compatibility with obsolete implementations.
1085   HTTP/1.1 clients and servers that parse the date value MUST accept
1086   all three formats (for compatibility with HTTP/1.0), though they MUST
1087   only generate the RFC 1123 format for representing HTTP-date values
1088   in header fields. See <xref target="tolerant.applications"/> for further information.
1091      Note: Recipients of date values are encouraged to be robust in
1092      accepting date values that may have been sent by non-HTTP
1093      applications, as is sometimes the case when retrieving or posting
1094      messages via proxies/gateways to SMTP or NNTP.
1097   All HTTP date/time stamps MUST be represented in Greenwich Mean Time
1098   (GMT), without exception. For the purposes of HTTP, GMT is exactly
1099   equal to UTC (Coordinated Universal Time). This is indicated in the
1100   first two formats by the inclusion of "GMT" as the three-letter
1101   abbreviation for time zone, and MUST be assumed when reading the
1102   asctime format. HTTP-date is case sensitive and MUST NOT include
1103   additional LWS beyond that specifically included as SP in the
1104   grammar.
1106<figure><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"/><artwork type="abnf2616"><![CDATA[
1107  HTTP-date    = rfc1123-date | rfc850-date | asctime-date
1108  rfc1123-date = wkday "," SP date1 SP time SP "GMT"
1109  rfc850-date  = weekday "," SP date2 SP time SP "GMT"
1110  asctime-date = wkday SP date3 SP time SP 4DIGIT
1111  date1        = 2DIGIT SP month SP 4DIGIT
1112                 ; day month year (e.g., 02 Jun 1982)
1113  date2        = 2DIGIT "-" month "-" 2DIGIT
1114                 ; day-month-year (e.g., 02-Jun-82)
1115  date3        = month SP ( 2DIGIT | ( SP 1DIGIT ))
1116                 ; month day (e.g., Jun  2)
1117  time         = 2DIGIT ":" 2DIGIT ":" 2DIGIT
1118                 ; 00:00:00 - 23:59:59
1119  wkday        = "Mon" | "Tue" | "Wed"
1120               | "Thu" | "Fri" | "Sat" | "Sun"
1121  weekday      = "Monday" | "Tuesday" | "Wednesday"
1122               | "Thursday" | "Friday" | "Saturday" | "Sunday"
1123  month        = "Jan" | "Feb" | "Mar" | "Apr"
1124               | "May" | "Jun" | "Jul" | "Aug"
1125               | "Sep" | "Oct" | "Nov" | "Dec"
1128      Note: HTTP requirements for the date/time stamp format apply only
1129      to their usage within the protocol stream. Clients and servers are
1130      not required to use these formats for user presentation, request
1131      logging, etc.
1136<section title="Transfer Codings" anchor="transfer.codings">
1138   Transfer-coding values are used to indicate an encoding
1139   transformation that has been, can be, or may need to be applied to an
1140   entity-body in order to ensure "safe transport" through the network.
1141   This differs from a content coding in that the transfer-coding is a
1142   property of the message, not of the original entity.
1144<figure><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/><artwork type="abnf2616"><![CDATA[
1145  transfer-coding         = "chunked" | transfer-extension
1146  transfer-extension      = token *( ";" parameter )
1149   Parameters are in  the form of attribute/value pairs.
1151<figure><iref primary="true" item="Grammar" subitem="parameter"/><iref primary="true" item="Grammar" subitem="attribute"/><iref primary="true" item="Grammar" subitem="value"/><artwork type="abnf2616"><![CDATA[
1152  parameter               = attribute "=" value
1153  attribute               = token
1154  value                   = token | quoted-string
1157   All transfer-coding values are case-insensitive. HTTP/1.1 uses
1158   transfer-coding values in the TE header field (<xref target="header.te"/>) and in
1159   the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1162   Whenever a transfer-coding is applied to a message-body, the set of
1163   transfer-codings MUST include "chunked", unless the message is
1164   terminated by closing the connection. When the "chunked" transfer-coding
1165   is used, it MUST be the last transfer-coding applied to the
1166   message-body. The "chunked" transfer-coding MUST NOT be applied more
1167   than once to a message-body. These rules allow the recipient to
1168   determine the transfer-length of the message (<xref target="message.length"/>).
1171   Transfer-codings are analogous to the Content-Transfer-Encoding
1172   values of MIME <xref target="RFC2045"/>, which were designed to enable safe transport of
1173   binary data over a 7-bit transport service. However, safe transport
1174   has a different focus for an 8bit-clean transfer protocol. In HTTP,
1175   the only unsafe characteristic of message-bodies is the difficulty in
1176   determining the exact body length (<xref target="message.length"/>), or the desire to
1177   encrypt data over a shared transport.
1180   The Internet Assigned Numbers Authority (IANA) acts as a registry for
1181   transfer-coding value tokens. Initially, the registry contains the
1182   following tokens: "chunked" (<xref target="chunked.transfer.encoding"/>),
1183   "gzip", "compress", and "deflate" (Section 2.2 of <xref target="Part3"/>).
1186   New transfer-coding value tokens SHOULD be registered in the same way
1187   as new content-coding value tokens (Section 2.2 of <xref target="Part3"/>).
1190   A server which receives an entity-body with a transfer-coding it does
1191   not understand SHOULD return 501 (Not Implemented), and close the
1192   connection. A server MUST NOT send transfer-codings to an HTTP/1.0
1193   client.
1196<section title="Chunked Transfer Coding" anchor="chunked.transfer.encoding">
1198   The chunked encoding modifies the body of a message in order to
1199   transfer it as a series of chunks, each with its own size indicator,
1200   followed by an OPTIONAL trailer containing entity-header fields. This
1201   allows dynamically produced content to be transferred along with the
1202   information necessary for the recipient to verify that it has
1203   received the full message.
1205<figure><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"/><artwork type="abnf2616"><![CDATA[
1206  Chunked-Body   = *chunk
1207                   last-chunk
1208                   trailer
1209                   CRLF
1211  chunk          = chunk-size [ chunk-extension ] CRLF
1212                   chunk-data CRLF
1213  chunk-size     = 1*HEX
1214  last-chunk     = 1*("0") [ chunk-extension ] CRLF
1216  chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
1217  chunk-ext-name = token
1218  chunk-ext-val  = token | quoted-string
1219  chunk-data     = 1*OCTET ; a sequence of chunk-size octets
1220  trailer        = *(entity-header CRLF)
1223   The chunk-size field is a string of hex digits indicating the size of
1224   the chunk-data in octets. The chunked encoding is ended by any chunk whose size is
1225   zero, followed by the trailer, which is terminated by an empty line.
1228   The trailer allows the sender to include additional HTTP header
1229   fields at the end of the message. The Trailer header field can be
1230   used to indicate which header fields are included in a trailer (see
1231   <xref target="header.trailer"/>).
1234   A server using chunked transfer-coding in a response MUST NOT use the
1235   trailer for any header fields unless at least one of the following is
1236   true:
1237  <list style="numbers">
1238    <t>the request included a TE header field that indicates "trailers" is
1239     acceptable in the transfer-coding of the  response, as described in
1240     <xref target="header.te"/>; or,</t>
1242    <t>the server is the origin server for the response, the trailer
1243     fields consist entirely of optional metadata, and the recipient
1244     could use the message (in a manner acceptable to the origin server)
1245     without receiving this metadata.  In other words, the origin server
1246     is willing to accept the possibility that the trailer fields might
1247     be silently discarded along the path to the client.</t>
1248  </list>
1251   This requirement prevents an interoperability failure when the
1252   message is being received by an HTTP/1.1 (or later) proxy and
1253   forwarded to an HTTP/1.0 recipient. It avoids a situation where
1254   compliance with the protocol would have necessitated a possibly
1255   infinite buffer on the proxy.
1258   A process for decoding the "chunked" transfer-coding
1259   can be represented in pseudo-code as:
1261<figure><artwork type="code"><![CDATA[
1262    length := 0
1263    read chunk-size, chunk-extension (if any) and CRLF
1264    while (chunk-size > 0) {
1265       read chunk-data and CRLF
1266       append chunk-data to entity-body
1267       length := length + chunk-size
1268       read chunk-size and CRLF
1269    }
1270    read entity-header
1271    while (entity-header not empty) {
1272       append entity-header to existing header fields
1273       read entity-header
1274    }
1275    Content-Length := length
1276    Remove "chunked" from Transfer-Encoding
1279   All HTTP/1.1 applications MUST be able to receive and decode the
1280   "chunked" transfer-coding, and MUST ignore chunk-extension extensions
1281   they do not understand.
1288<section title="HTTP Message" anchor="http.message">
1290<section title="Message Types" anchor="message.types">
1292   HTTP messages consist of requests from client to server and responses
1293   from server to client.
1295<figure><iref primary="true" item="Grammar" subitem="HTTP-message"/><artwork type="abnf2616"><![CDATA[
1296  HTTP-message   = Request | Response     ; HTTP/1.1 messages
1299   Request (<xref target="request"/>) and Response (<xref target="response"/>) messages use the generic
1300   message format of <xref target="RFC2822"/> for transferring entities (the payload
1301   of the message). Both types of message consist of a start-line, zero
1302   or more header fields (also known as "headers"), an empty line (i.e.,
1303   a line with nothing preceding the CRLF) indicating the end of the
1304   header fields, and possibly a message-body.
1306<figure><iref primary="true" item="Grammar" subitem="generic-message"/><iref primary="true" item="Grammar" subitem="start-line"/><artwork type="abnf2616"><![CDATA[
1307  generic-message = start-line
1308                    *(message-header CRLF)
1309                    CRLF
1310                    [ message-body ]
1311  start-line      = Request-Line | Status-Line
1314   In the interest of robustness, servers SHOULD ignore any empty
1315   line(s) received where a Request-Line is expected. In other words, if
1316   the server is reading the protocol stream at the beginning of a
1317   message and receives a CRLF first, it should ignore the CRLF.
1320   Certain buggy HTTP/1.0 client implementations generate extra CRLF's
1321   after a POST request. To restate what is explicitly forbidden by the
1322   BNF, an HTTP/1.1 client MUST NOT preface or follow a request with an
1323   extra CRLF.
1327<section title="Message Headers" anchor="message.headers">
1329   HTTP header fields, which include general-header (<xref target="general.header.fields"/>),
1330   request-header (Section 4 of <xref target="Part2"/>), response-header (Section 6 of <xref target="Part2"/>), and
1331   entity-header (Section 3.1 of <xref target="Part3"/>) fields, follow the same generic format as
1332   that given in Section 2.1 of <xref target="RFC2822"/>. Each header field consists
1333   of a name followed by a colon (":") and the field value. Field names
1334   are case-insensitive. The field value MAY be preceded by any amount
1335   of LWS, though a single SP is preferred. Header fields can be
1336   extended over multiple lines by preceding each extra line with at
1337   least one SP or HTAB. Applications ought to follow "common form", where
1338   one is known or indicated, when generating HTTP constructs, since
1339   there might exist some implementations that fail to accept anything
1340   beyond the common forms.
1342<figure><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"/><artwork type="abnf2616"><![CDATA[
1343  message-header = field-name ":" [ field-value ]
1344  field-name     = token
1345  field-value    = *( field-content | LWS )
1346  field-content  = <the OCTETs making up the field-value
1347                   and consisting of either *TEXT or combinations
1348                   of token, separators, and quoted-string>
1351   The field-content does not include any leading or trailing LWS:
1352   linear white space occurring before the first non-whitespace
1353   character of the field-value or after the last non-whitespace
1354   character of the field-value. Such leading or trailing LWS MAY be
1355   removed without changing the semantics of the field value. Any LWS
1356   that occurs between field-content MAY be replaced with a single SP
1357   before interpreting the field value or forwarding the message
1358   downstream.
1361   The order in which header fields with differing field names are
1362   received is not significant. However, it is "good practice" to send
1363   general-header fields first, followed by request-header or response-header
1364   fields, and ending with the entity-header fields.
1367   Multiple message-header fields with the same field-name MAY be
1368   present in a message if and only if the entire field-value for that
1369   header field is defined as a comma-separated list [i.e., #(values)].
1370   It MUST be possible to combine the multiple header fields into one
1371   "field-name: field-value" pair, without changing the semantics of the
1372   message, by appending each subsequent field-value to the first, each
1373   separated by a comma. The order in which header fields with the same
1374   field-name are received is therefore significant to the
1375   interpretation of the combined field value, and thus a proxy MUST NOT
1376   change the order of these field values when a message is forwarded.
1380<section title="Message Body" anchor="message.body">
1382   The message-body (if any) of an HTTP message is used to carry the
1383   entity-body associated with the request or response. The message-body
1384   differs from the entity-body only when a transfer-coding has been
1385   applied, as indicated by the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1387<figure><iref primary="true" item="Grammar" subitem="message-body"/><artwork type="abnf2616"><![CDATA[
1388  message-body = entity-body
1389               | <entity-body encoded as per Transfer-Encoding>
1392   Transfer-Encoding MUST be used to indicate any transfer-codings
1393   applied by an application to ensure safe and proper transfer of the
1394   message. Transfer-Encoding is a property of the message, not of the
1395   entity, and thus MAY be added or removed by any application along the
1396   request/response chain. (However, <xref target="transfer.codings"/> places restrictions on
1397   when certain transfer-codings may be used.)
1400   The rules for when a message-body is allowed in a message differ for
1401   requests and responses.
1404   The presence of a message-body in a request is signaled by the
1405   inclusion of a Content-Length or Transfer-Encoding header field in
1406   the request's message-headers. A message-body MUST NOT be included in
1407   a request if the specification of the request method (Section 3 of <xref target="Part2"/>)
1408   does not allow sending an entity-body in requests. A server SHOULD
1409   read and forward a message-body on any request; if the request method
1410   does not include defined semantics for an entity-body, then the
1411   message-body SHOULD be ignored when handling the request.
1414   For response messages, whether or not a message-body is included with
1415   a message is dependent on both the request method and the response
1416   status code (<xref target="status.code.and.reason.phrase"/>). All responses to the HEAD request method
1417   MUST NOT include a message-body, even though the presence of entity-header
1418   fields might lead one to believe they do. All 1xx
1419   (informational), 204 (No Content), and 304 (Not Modified) responses
1420   MUST NOT include a message-body. All other responses do include a
1421   message-body, although it MAY be of zero length.
1425<section title="Message Length" anchor="message.length">
1427   The transfer-length of a message is the length of the message-body as
1428   it appears in the message; that is, after any transfer-codings have
1429   been applied. When a message-body is included with a message, the
1430   transfer-length of that body is determined by one of the following
1431   (in order of precedence):
1434  <list style="numbers">
1435    <t>
1436     Any response message which "MUST NOT" include a message-body (such
1437     as the 1xx, 204, and 304 responses and any response to a HEAD
1438     request) is always terminated by the first empty line after the
1439     header fields, regardless of the entity-header fields present in
1440     the message.
1441    </t>
1442    <t>
1443     If a Transfer-Encoding header field (<xref target="header.transfer-encoding"/>)
1444     is present, then the transfer-length is
1445     defined by use of the "chunked" transfer-coding (<xref target="transfer.codings"/>),
1446     unless the message is terminated by closing the connection.
1447    </t>
1448    <t>
1449     If a Content-Length header field (<xref target="header.content-length"/>) is present, its
1450     decimal value in OCTETs represents both the entity-length and the
1451     transfer-length. The Content-Length header field MUST NOT be sent
1452     if these two lengths are different (i.e., if a Transfer-Encoding
1453     header field is present). If a message is received with both a
1454     Transfer-Encoding header field and a Content-Length header field,
1455     the latter MUST be ignored.
1456    </t>
1457    <t>
1458     If the message uses the media type "multipart/byteranges", and the
1459     transfer-length is not otherwise specified, then this self-delimiting
1460     media type defines the transfer-length. This media type
1461     MUST NOT be used unless the sender knows that the recipient can parse
1462     it; the presence in a request of a Range header with multiple byte-range
1463     specifiers from a 1.1 client implies that the client can parse
1464     multipart/byteranges responses.
1465    <list style="empty"><t>
1466       A range header might be forwarded by a 1.0 proxy that does not
1467       understand multipart/byteranges; in this case the server MUST
1468       delimit the message using methods defined in items 1, 3 or 5 of
1469       this section.
1470    </t></list>
1471    </t>
1472    <t>
1473     By the server closing the connection. (Closing the connection
1474     cannot be used to indicate the end of a request body, since that
1475     would leave no possibility for the server to send back a response.)
1476    </t>
1477  </list>
1480   For compatibility with HTTP/1.0 applications, HTTP/1.1 requests
1481   containing a message-body MUST include a valid Content-Length header
1482   field unless the server is known to be HTTP/1.1 compliant. If a
1483   request contains a message-body and a Content-Length is not given,
1484   the server SHOULD respond with 400 (Bad Request) if it cannot
1485   determine the length of the message, or with 411 (Length Required) if
1486   it wishes to insist on receiving a valid Content-Length.
1489   All HTTP/1.1 applications that receive entities MUST accept the
1490   "chunked" transfer-coding (<xref target="transfer.codings"/>), thus allowing this mechanism
1491   to be used for messages when the message length cannot be determined
1492   in advance.
1495   Messages MUST NOT include both a Content-Length header field and a
1496   transfer-coding. If the message does include a
1497   transfer-coding, the Content-Length MUST be ignored.
1500   When a Content-Length is given in a message where a message-body is
1501   allowed, its field value MUST exactly match the number of OCTETs in
1502   the message-body. HTTP/1.1 user agents MUST notify the user when an
1503   invalid length is received and detected.
1507<section title="General Header Fields" anchor="general.header.fields">
1509   There are a few header fields which have general applicability for
1510   both request and response messages, but which do not apply to the
1511   entity being transferred. These header fields apply only to the
1512   message being transmitted.
1514<figure><iref primary="true" item="Grammar" subitem="general-header"/><artwork type="abnf2616"><![CDATA[
1515  general-header = Cache-Control            ; [Part6], Section 15.2
1516                 | Connection               ; Section 8.1
1517                 | Date                     ; Section 8.3
1518                 | Pragma                   ; [Part6], Section 15.4
1519                 | Trailer                  ; Section 8.6
1520                 | Transfer-Encoding        ; Section 8.7
1521                 | Upgrade                  ; Section 8.8
1522                 | Via                      ; Section 8.9
1523                 | Warning                  ; [Part6], Section 15.6
1526   General-header field names can be extended reliably only in
1527   combination with a change in the protocol version. However, new or
1528   experimental header fields may be given the semantics of general
1529   header fields if all parties in the communication recognize them to
1530   be general-header fields. Unrecognized header fields are treated as
1531   entity-header fields.
1536<section title="Request" anchor="request">
1538   A request message from a client to a server includes, within the
1539   first line of that message, the method to be applied to the resource,
1540   the identifier of the resource, and the protocol version in use.
1542<!--                 Host                      ; should be moved here eventually -->
1543<figure><iref primary="true" item="Grammar" subitem="Request"/><artwork type="abnf2616"><![CDATA[
1544  Request       = Request-Line              ; Section 5.1
1545                  *(( general-header        ; Section 4.5
1546                   | request-header         ; [Part2], Section 4
1547                   | entity-header ) CRLF)  ; [Part3], Section 3.1
1548                  CRLF
1549                  [ message-body ]          ; Section 4.3
1552<section title="Request-Line" anchor="request-line">
1554   The Request-Line begins with a method token, followed by the
1555   Request-URI and the protocol version, and ending with CRLF. The
1556   elements are separated by SP characters. No CR or LF is allowed
1557   except in the final CRLF sequence.
1559<figure><iref primary="true" item="Grammar" subitem="Request-Line"/><artwork type="abnf2616"><![CDATA[
1560  Request-Line   = Method SP Request-URI SP HTTP-Version CRLF
1563<section title="Method" anchor="method">
1565   The Method  token indicates the method to be performed on the
1566   resource identified by the Request-URI. The method is case-sensitive.
1568<figure><iref primary="true" item="Grammar" subitem="Method"/><iref primary="true" item="Grammar" subitem="extension-method"/><artwork type="abnf2616"><![CDATA[
1569  Method         = token
1573<section title="Request-URI" anchor="request-uri">
1575   The Request-URI is a Uniform Resource Identifier (<xref target="uri"/>) and
1576   identifies the resource upon which to apply the request.
1578<figure><iref primary="true" item="Grammar" subitem="Request-URI"/><artwork type="abnf2616"><![CDATA[
1579  Request-URI    = "*"
1580                 | absoluteURI
1581                 | ( abs_path [ "?" query ] )
1582                 | authority
1585   The four options for Request-URI are dependent on the nature of the
1586   request. The asterisk "*" means that the request does not apply to a
1587   particular resource, but to the server itself, and is only allowed
1588   when the method used does not necessarily apply to a resource. One
1589   example would be
1591<figure><artwork type="example"><![CDATA[
1592    OPTIONS * HTTP/1.1
1595   The absoluteURI form is REQUIRED when the request is being made to a
1596   proxy. The proxy is requested to forward the request or service it
1597   from a valid cache, and return the response. Note that the proxy MAY
1598   forward the request on to another proxy or directly to the server
1599   specified by the absoluteURI. In order to avoid request loops, a
1600   proxy MUST be able to recognize all of its server names, including
1601   any aliases, local variations, and the numeric IP address. An example
1602   Request-Line would be:
1604<figure><artwork type="example"><![CDATA[
1605    GET HTTP/1.1
1608   To allow for transition to absoluteURIs in all requests in future
1609   versions of HTTP, all HTTP/1.1 servers MUST accept the absoluteURI
1610   form in requests, even though HTTP/1.1 clients will only generate
1611   them in requests to proxies.
1614   The authority form is only used by the CONNECT method (Section 8.9 of <xref target="Part2"/>).
1617   The most common form of Request-URI is that used to identify a
1618   resource on an origin server or gateway. In this case the absolute
1619   path of the URI MUST be transmitted (see <xref target="general.syntax"/>, abs_path) as
1620   the Request-URI, and the network location of the URI (authority) MUST
1621   be transmitted in a Host header field. For example, a client wishing
1622   to retrieve the resource above directly from the origin server would
1623   create a TCP connection to port 80 of the host "" and send
1624   the lines:
1626<figure><artwork type="example"><![CDATA[
1627    GET /pub/WWW/TheProject.html HTTP/1.1
1628    Host:
1631   followed by the remainder of the Request. Note that the absolute path
1632   cannot be empty; if none is present in the original URI, it MUST be
1633   given as "/" (the server root).
1636   The Request-URI is transmitted in the format specified in
1637   <xref target="general.syntax"/>. If the Request-URI is encoded using the "% HEX HEX" encoding
1638   <xref target="RFC2396"/>, the origin server MUST decode the Request-URI in order to
1639   properly interpret the request. Servers SHOULD respond to invalid
1640   Request-URIs with an appropriate status code.
1643   A transparent proxy MUST NOT rewrite the "abs_path" part of the
1644   received Request-URI when forwarding it to the next inbound server,
1645   except as noted above to replace a null abs_path with "/".
1648  <list><t>
1649      Note: The "no rewrite" rule prevents the proxy from changing the
1650      meaning of the request when the origin server is improperly using
1651      a non-reserved URI character for a reserved purpose.  Implementors
1652      should be aware that some pre-HTTP/1.1 proxies have been known to
1653      rewrite the Request-URI.
1654  </t></list>
1659<section title="The Resource Identified by a Request" anchor="">
1661   The exact resource identified by an Internet request is determined by
1662   examining both the Request-URI and the Host header field.
1665   An origin server that does not allow resources to differ by the
1666   requested host MAY ignore the Host header field value when
1667   determining the resource identified by an HTTP/1.1 request. (But see
1668   <xref target=""/>
1669   for other requirements on Host support in HTTP/1.1.)
1672   An origin server that does differentiate resources based on the host
1673   requested (sometimes referred to as virtual hosts or vanity host
1674   names) MUST use the following rules for determining the requested
1675   resource on an HTTP/1.1 request:
1676  <list style="numbers">
1677    <t>If Request-URI is an absoluteURI, the host is part of the
1678     Request-URI. Any Host header field value in the request MUST be
1679     ignored.</t>
1680    <t>If the Request-URI is not an absoluteURI, and the request includes
1681     a Host header field, the host is determined by the Host header
1682     field value.</t>
1683    <t>If the host as determined by rule 1 or 2 is not a valid host on
1684     the server, the response MUST be a 400 (Bad Request) error message.</t>
1685  </list>
1688   Recipients of an HTTP/1.0 request that lacks a Host header field MAY
1689   attempt to use heuristics (e.g., examination of the URI path for
1690   something unique to a particular host) in order to determine what
1691   exact resource is being requested.
1698<section title="Response" anchor="response">
1700   After receiving and interpreting a request message, a server responds
1701   with an HTTP response message.
1703<figure><iref primary="true" item="Grammar" subitem="Response"/><artwork type="abnf2616"><![CDATA[
1704  Response      = Status-Line               ; Section 6.1
1705                  *(( general-header        ; Section 4.5
1706                   | response-header        ; [Part2], Section 6
1707                   | entity-header ) CRLF)  ; [Part3], Section 3.1
1708                  CRLF
1709                  [ message-body ]          ; Section 4.3
1712<section title="Status-Line" anchor="status-line">
1714   The first line of a Response message is the Status-Line, consisting
1715   of the protocol version followed by a numeric status code and its
1716   associated textual phrase, with each element separated by SP
1717   characters. No CR or LF is allowed except in the final CRLF sequence.
1719<figure><iref primary="true" item="Grammar" subitem="Status-Line"/><artwork type="abnf2616"><![CDATA[
1720  Status-Line = HTTP-Version SP Status-Code SP Reason-Phrase CRLF
1723<section title="Status Code and Reason Phrase" anchor="status.code.and.reason.phrase">
1725   The Status-Code element is a 3-digit integer result code of the
1726   attempt to understand and satisfy the request. These codes are fully
1727   defined in Section 9 of <xref target="Part2"/>. The Reason-Phrase is intended to give a short
1728   textual description of the Status-Code. The Status-Code is intended
1729   for use by automata and the Reason-Phrase is intended for the human
1730   user. The client is not required to examine or display the Reason-Phrase.
1733   The first digit of the Status-Code defines the class of response. The
1734   last two digits do not have any categorization role. There are 5
1735   values for the first digit:
1736  <list style="symbols">
1737    <t>
1738      1xx: Informational - Request received, continuing process
1739    </t>
1740    <t>
1741      2xx: Success - The action was successfully received,
1742        understood, and accepted
1743    </t>
1744    <t>
1745      3xx: Redirection - Further action must be taken in order to
1746        complete the request
1747    </t>
1748    <t>
1749      4xx: Client Error - The request contains bad syntax or cannot
1750        be fulfilled
1751    </t>
1752    <t>
1753      5xx: Server Error - The server failed to fulfill an apparently
1754        valid request
1755    </t>
1756  </list>
1758<figure><iref primary="true" item="Grammar" subitem="Status-Code"/><iref primary="true" item="Grammar" subitem="extension-code"/><iref primary="true" item="Grammar" subitem="Reason-Phrase"/><artwork type="abnf2616"><![CDATA[
1759  Status-Code    = 3DIGIT
1760  Reason-Phrase  = *<TEXT, excluding CR, LF>
1768<section title="Connections" anchor="connections">
1770<section title="Persistent Connections" anchor="persistent.connections">
1772<section title="Purpose" anchor="persistent.purpose">
1774   Prior to persistent connections, a separate TCP connection was
1775   established to fetch each URL, increasing the load on HTTP servers
1776   and causing congestion on the Internet. The use of inline images and
1777   other associated data often require a client to make multiple
1778   requests of the same server in a short amount of time. Analysis of
1779   these performance problems and results from a prototype
1780   implementation are available <xref target="Pad1995"/> <xref target="Spe"/>. Implementation experience and
1781   measurements of actual HTTP/1.1 (RFC 2068) implementations show good
1782   results <xref target="Nie1997"/>. Alternatives have also been explored, for example,
1783   T/TCP <xref target="Tou1998"/>.
1786   Persistent HTTP connections have a number of advantages:
1787  <list style="symbols">
1788      <t>
1789        By opening and closing fewer TCP connections, CPU time is saved
1790        in routers and hosts (clients, servers, proxies, gateways,
1791        tunnels, or caches), and memory used for TCP protocol control
1792        blocks can be saved in hosts.
1793      </t>
1794      <t>
1795        HTTP requests and responses can be pipelined on a connection.
1796        Pipelining allows a client to make multiple requests without
1797        waiting for each response, allowing a single TCP connection to
1798        be used much more efficiently, with much lower elapsed time.
1799      </t>
1800      <t>
1801        Network congestion is reduced by reducing the number of packets
1802        caused by TCP opens, and by allowing TCP sufficient time to
1803        determine the congestion state of the network.
1804      </t>
1805      <t>
1806        Latency on subsequent requests is reduced since there is no time
1807        spent in TCP's connection opening handshake.
1808      </t>
1809      <t>
1810        HTTP can evolve more gracefully, since errors can be reported
1811        without the penalty of closing the TCP connection. Clients using
1812        future versions of HTTP might optimistically try a new feature,
1813        but if communicating with an older server, retry with old
1814        semantics after an error is reported.
1815      </t>
1816    </list>
1819   HTTP implementations SHOULD implement persistent connections.
1823<section title="Overall Operation" anchor="persistent.overall">
1825   A significant difference between HTTP/1.1 and earlier versions of
1826   HTTP is that persistent connections are the default behavior of any
1827   HTTP connection. That is, unless otherwise indicated, the client
1828   SHOULD assume that the server will maintain a persistent connection,
1829   even after error responses from the server.
1832   Persistent connections provide a mechanism by which a client and a
1833   server can signal the close of a TCP connection. This signaling takes
1834   place using the Connection header field (<xref target="header.connection"/>). Once a close
1835   has been signaled, the client MUST NOT send any more requests on that
1836   connection.
1839<section title="Negotiation" anchor="persistent.negotiation">
1841   An HTTP/1.1 server MAY assume that a HTTP/1.1 client intends to
1842   maintain a persistent connection unless a Connection header including
1843   the connection-token "close" was sent in the request. If the server
1844   chooses to close the connection immediately after sending the
1845   response, it SHOULD send a Connection header including the
1846   connection-token close.
1849   An HTTP/1.1 client MAY expect a connection to remain open, but would
1850   decide to keep it open based on whether the response from a server
1851   contains a Connection header with the connection-token close. In case
1852   the client does not want to maintain a connection for more than that
1853   request, it SHOULD send a Connection header including the
1854   connection-token close.
1857   If either the client or the server sends the close token in the
1858   Connection header, that request becomes the last one for the
1859   connection.
1862   Clients and servers SHOULD NOT  assume that a persistent connection is
1863   maintained for HTTP versions less than 1.1 unless it is explicitly
1864   signaled. See <xref target="compatibility.with.http.1.0.persistent.connections"/> for more information on backward
1865   compatibility with HTTP/1.0 clients.
1868   In order to remain persistent, all messages on the connection MUST
1869   have a self-defined message length (i.e., one not defined by closure
1870   of the connection), as described in <xref target="message.length"/>.
1874<section title="Pipelining" anchor="pipelining">
1876   A client that supports persistent connections MAY "pipeline" its
1877   requests (i.e., send multiple requests without waiting for each
1878   response). A server MUST send its responses to those requests in the
1879   same order that the requests were received.
1882   Clients which assume persistent connections and pipeline immediately
1883   after connection establishment SHOULD be prepared to retry their
1884   connection if the first pipelined attempt fails. If a client does
1885   such a retry, it MUST NOT pipeline before it knows the connection is
1886   persistent. Clients MUST also be prepared to resend their requests if
1887   the server closes the connection before sending all of the
1888   corresponding responses.
1891   Clients SHOULD NOT  pipeline requests using non-idempotent methods or
1892   non-idempotent sequences of methods (see Section 8.1.2 of <xref target="Part2"/>). Otherwise, a
1893   premature termination of the transport connection could lead to
1894   indeterminate results. A client wishing to send a non-idempotent
1895   request SHOULD wait to send that request until it has received the
1896   response status for the previous request.
1901<section title="Proxy Servers" anchor="persistent.proxy">
1903   It is especially important that proxies correctly implement the
1904   properties of the Connection header field as specified in <xref target="header.connection"/>.
1907   The proxy server MUST signal persistent connections separately with
1908   its clients and the origin servers (or other proxy servers) that it
1909   connects to. Each persistent connection applies to only one transport
1910   link.
1913   A proxy server MUST NOT establish a HTTP/1.1 persistent connection
1914   with an HTTP/1.0 client (but see <xref target="RFC2068"/> for information and
1915   discussion of the problems with the Keep-Alive header implemented by
1916   many HTTP/1.0 clients).
1920<section title="Practical Considerations" anchor="persistent.practical">
1922   Servers will usually have some time-out value beyond which they will
1923   no longer maintain an inactive connection. Proxy servers might make
1924   this a higher value since it is likely that the client will be making
1925   more connections through the same server. The use of persistent
1926   connections places no requirements on the length (or existence) of
1927   this time-out for either the client or the server.
1930   When a client or server wishes to time-out it SHOULD issue a graceful
1931   close on the transport connection. Clients and servers SHOULD both
1932   constantly watch for the other side of the transport close, and
1933   respond to it as appropriate. If a client or server does not detect
1934   the other side's close promptly it could cause unnecessary resource
1935   drain on the network.
1938   A client, server, or proxy MAY close the transport connection at any
1939   time. For example, a client might have started to send a new request
1940   at the same time that the server has decided to close the "idle"
1941   connection. From the server's point of view, the connection is being
1942   closed while it was idle, but from the client's point of view, a
1943   request is in progress.
1946   This means that clients, servers, and proxies MUST be able to recover
1947   from asynchronous close events. Client software SHOULD reopen the
1948   transport connection and retransmit the aborted sequence of requests
1949   without user interaction so long as the request sequence is
1950   idempotent (see Section 8.1.2 of <xref target="Part2"/>). Non-idempotent methods or sequences
1951   MUST NOT be automatically retried, although user agents MAY offer a
1952   human operator the choice of retrying the request(s). Confirmation by
1953   user-agent software with semantic understanding of the application
1954   MAY substitute for user confirmation. The automatic retry SHOULD NOT
1955   be repeated if the second sequence of requests fails.
1958   Servers SHOULD always respond to at least one request per connection,
1959   if at all possible. Servers SHOULD NOT  close a connection in the
1960   middle of transmitting a response, unless a network or client failure
1961   is suspected.
1964   Clients that use persistent connections SHOULD limit the number of
1965   simultaneous connections that they maintain to a given server. A
1966   single-user client SHOULD NOT maintain more than 2 connections with
1967   any server or proxy. A proxy SHOULD use up to 2*N connections to
1968   another server or proxy, where N is the number of simultaneously
1969   active users. These guidelines are intended to improve HTTP response
1970   times and avoid congestion.
1975<section title="Message Transmission Requirements" anchor="message.transmission.requirements">
1977<section title="Persistent Connections and Flow Control" anchor="persistent.flow">
1979   HTTP/1.1 servers SHOULD maintain persistent connections and use TCP's
1980   flow control mechanisms to resolve temporary overloads, rather than
1981   terminating connections with the expectation that clients will retry.
1982   The latter technique can exacerbate network congestion.
1986<section title="Monitoring Connections for Error Status Messages" anchor="persistent.monitor">
1988   An HTTP/1.1 (or later) client sending a message-body SHOULD monitor
1989   the network connection for an error status while it is transmitting
1990   the request. If the client sees an error status, it SHOULD
1991   immediately cease transmitting the body. If the body is being sent
1992   using a "chunked" encoding (<xref target="transfer.codings"/>), a zero length chunk and
1993   empty trailer MAY be used to prematurely mark the end of the message.
1994   If the body was preceded by a Content-Length header, the client MUST
1995   close the connection.
1999<section title="Use of the 100 (Continue) Status" anchor="use.of.the.100.status">
2001   The purpose of the 100 (Continue) status (see Section 9.1.1 of <xref target="Part2"/>) is to
2002   allow a client that is sending a request message with a request body
2003   to determine if the origin server is willing to accept the request
2004   (based on the request headers) before the client sends the request
2005   body. In some cases, it might either be inappropriate or highly
2006   inefficient for the client to send the body if the server will reject
2007   the message without looking at the body.
2010   Requirements for HTTP/1.1 clients:
2011  <list style="symbols">
2012    <t>
2013        If a client will wait for a 100 (Continue) response before
2014        sending the request body, it MUST send an Expect request-header
2015        field (Section 10.2 of <xref target="Part2"/>) with the "100-continue" expectation.
2016    </t>
2017    <t>
2018        A client MUST NOT send an Expect request-header field (Section 10.2 of <xref target="Part2"/>)
2019        with the "100-continue" expectation if it does not intend
2020        to send a request body.
2021    </t>
2022  </list>
2025   Because of the presence of older implementations, the protocol allows
2026   ambiguous situations in which a client may send "Expect: 100-continue"
2027   without receiving either a 417 (Expectation Failed) status
2028   or a 100 (Continue) status. Therefore, when a client sends this
2029   header field to an origin server (possibly via a proxy) from which it
2030   has never seen a 100 (Continue) status, the client SHOULD NOT  wait
2031   for an indefinite period before sending the request body.
2034   Requirements for HTTP/1.1 origin servers:
2035  <list style="symbols">
2036    <t> Upon receiving a request which includes an Expect request-header
2037        field with the "100-continue" expectation, an origin server MUST
2038        either respond with 100 (Continue) status and continue to read
2039        from the input stream, or respond with a final status code. The
2040        origin server MUST NOT wait for the request body before sending
2041        the 100 (Continue) response. If it responds with a final status
2042        code, it MAY close the transport connection or it MAY continue
2043        to read and discard the rest of the request.  It MUST NOT
2044        perform the requested method if it returns a final status code.
2045    </t>
2046    <t> An origin server SHOULD NOT  send a 100 (Continue) response if
2047        the request message does not include an Expect request-header
2048        field with the "100-continue" expectation, and MUST NOT send a
2049        100 (Continue) response if such a request comes from an HTTP/1.0
2050        (or earlier) client. There is an exception to this rule: for
2051        compatibility with <xref target="RFC2068"/>, a server MAY send a 100 (Continue)
2052        status in response to an HTTP/1.1 PUT or POST request that does
2053        not include an Expect request-header field with the "100-continue"
2054        expectation. This exception, the purpose of which is
2055        to minimize any client processing delays associated with an
2056        undeclared wait for 100 (Continue) status, applies only to
2057        HTTP/1.1 requests, and not to requests with any other HTTP-version
2058        value.
2059    </t>
2060    <t> An origin server MAY omit a 100 (Continue) response if it has
2061        already received some or all of the request body for the
2062        corresponding request.
2063    </t>
2064    <t> An origin server that sends a 100 (Continue) response MUST
2065    ultimately send a final status code, once the request body is
2066        received and processed, unless it terminates the transport
2067        connection prematurely.
2068    </t>
2069    <t> If an origin server receives a request that does not include an
2070        Expect request-header field with the "100-continue" expectation,
2071        the request includes a request body, and the server responds
2072        with a final status code before reading the entire request body
2073        from the transport connection, then the server SHOULD NOT  close
2074        the transport connection until it has read the entire request,
2075        or until the client closes the connection. Otherwise, the client
2076        might not reliably receive the response message. However, this
2077        requirement is not be construed as preventing a server from
2078        defending itself against denial-of-service attacks, or from
2079        badly broken client implementations.
2080      </t>
2081    </list>
2084   Requirements for HTTP/1.1 proxies:
2085  <list style="symbols">
2086    <t> If a proxy receives a request that includes an Expect request-header
2087        field with the "100-continue" expectation, and the proxy
2088        either knows that the next-hop server complies with HTTP/1.1 or
2089        higher, or does not know the HTTP version of the next-hop
2090        server, it MUST forward the request, including the Expect header
2091        field.
2092    </t>
2093    <t> If the proxy knows that the version of the next-hop server is
2094        HTTP/1.0 or lower, it MUST NOT forward the request, and it MUST
2095        respond with a 417 (Expectation Failed) status.
2096    </t>
2097    <t> Proxies SHOULD maintain a cache recording the HTTP version
2098        numbers received from recently-referenced next-hop servers.
2099    </t>
2100    <t> A proxy MUST NOT forward a 100 (Continue) response if the
2101        request message was received from an HTTP/1.0 (or earlier)
2102        client and did not include an Expect request-header field with
2103        the "100-continue" expectation. This requirement overrides the
2104        general rule for forwarding of 1xx responses (see Section 9.1 of <xref target="Part2"/>).
2105    </t>
2106  </list>
2110<section title="Client Behavior if Server Prematurely Closes Connection" anchor="connection.premature">
2112   If an HTTP/1.1 client sends a request which includes a request body,
2113   but which does not include an Expect request-header field with the
2114   "100-continue" expectation, and if the client is not directly
2115   connected to an HTTP/1.1 origin server, and if the client sees the
2116   connection close before receiving any status from the server, the
2117   client SHOULD retry the request.  If the client does retry this
2118   request, it MAY use the following "binary exponential backoff"
2119   algorithm to be assured of obtaining a reliable response:
2120  <list style="numbers">
2121    <t>
2122      Initiate a new connection to the server
2123    </t>
2124    <t>
2125      Transmit the request-headers
2126    </t>
2127    <t>
2128      Initialize a variable R to the estimated round-trip time to the
2129         server (e.g., based on the time it took to establish the
2130         connection), or to a constant value of 5 seconds if the round-trip
2131         time is not available.
2132    </t>
2133    <t>
2134       Compute T = R * (2**N), where N is the number of previous
2135         retries of this request.
2136    </t>
2137    <t>
2138       Wait either for an error response from the server, or for T
2139         seconds (whichever comes first)
2140    </t>
2141    <t>
2142       If no error response is received, after T seconds transmit the
2143         body of the request.
2144    </t>
2145    <t>
2146       If client sees that the connection is closed prematurely,
2147         repeat from step 1 until the request is accepted, an error
2148         response is received, or the user becomes impatient and
2149         terminates the retry process.
2150    </t>
2151  </list>
2154   If at any point an error status is received, the client
2155  <list style="symbols">
2156      <t>SHOULD NOT  continue and</t>
2158      <t>SHOULD close the connection if it has not completed sending the
2159        request message.</t>
2160    </list>
2167<section title="Header Field Definitions" anchor="header.fields">
2169   This section defines the syntax and semantics of HTTP/1.1 header fields
2170   related to message framing and transport protocols.
2173   For entity-header fields, both sender and recipient refer to either the
2174   client or the server, depending on who sends and who receives the entity.
2177<section title="Connection" anchor="header.connection">
2178  <iref primary="true" item="Connection header"/>
2179  <iref primary="true" item="Headers" subitem="Connection"/>
2181   The Connection general-header field allows the sender to specify
2182   options that are desired for that particular connection and MUST NOT
2183   be communicated by proxies over further connections.
2186   The Connection header has the following grammar:
2188<figure><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-token"/><artwork type="abnf2616"><![CDATA[
2189  Connection = "Connection" ":" 1#(connection-token)
2190  connection-token  = token
2193   HTTP/1.1 proxies MUST parse the Connection header field before a
2194   message is forwarded and, for each connection-token in this field,
2195   remove any header field(s) from the message with the same name as the
2196   connection-token. Connection options are signaled by the presence of
2197   a connection-token in the Connection header field, not by any
2198   corresponding additional header field(s), since the additional header
2199   field may not be sent if there are no parameters associated with that
2200   connection option.
2203   Message headers listed in the Connection header MUST NOT include
2204   end-to-end headers, such as Cache-Control.
2207   HTTP/1.1 defines the "close" connection option for the sender to
2208   signal that the connection will be closed after completion of the
2209   response. For example,
2211<figure><artwork type="example"><![CDATA[
2212    Connection: close
2215   in either the request or the response header fields indicates that
2216   the connection SHOULD NOT  be considered `persistent' (<xref target="persistent.connections"/>)
2217   after the current request/response is complete.
2220   An HTTP/1.1 client that does not support persistent connections MUST
2221   include the "close" connection option in every request message.
2224   An HTTP/1.1 server that does not support persistent connections MUST
2225   include the "close" connection option in every response message that
2226   does not have a 1xx (informational) status code.
2229   A system receiving an HTTP/1.0 (or lower-version) message that
2230   includes a Connection header MUST, for each connection-token in this
2231   field, remove and ignore any header field(s) from the message with
2232   the same name as the connection-token. This protects against mistaken
2233   forwarding of such header fields by pre-HTTP/1.1 proxies. See <xref target="compatibility.with.http.1.0.persistent.connections"/>.
2237<section title="Content-Length" anchor="header.content-length">
2238  <iref primary="true" item="Content-Length header"/>
2239  <iref primary="true" item="Headers" subitem="Content-Length"/>
2241   The Content-Length entity-header field indicates the size of the
2242   entity-body, in decimal number of OCTETs, sent to the recipient or,
2243   in the case of the HEAD method, the size of the entity-body that
2244   would have been sent had the request been a GET.
2246<figure><iref primary="true" item="Grammar" subitem="Content-Length"/><artwork type="abnf2616"><![CDATA[
2247  Content-Length    = "Content-Length" ":" 1*DIGIT
2250   An example is
2252<figure><artwork type="example"><![CDATA[
2253    Content-Length: 3495
2256   Applications SHOULD use this field to indicate the transfer-length of
2257   the message-body, unless this is prohibited by the rules in <xref target="message.length"/>.
2260   Any Content-Length greater than or equal to zero is a valid value.
2261   <xref target="message.length"/> describes how to determine the length of a message-body
2262   if a Content-Length is not given.
2265   Note that the meaning of this field is significantly different from
2266   the corresponding definition in MIME, where it is an optional field
2267   used within the "message/external-body" content-type. In HTTP, it
2268   SHOULD be sent whenever the message's length can be determined prior
2269   to being transferred, unless this is prohibited by the rules in
2270   <xref target="message.length"/>.
2274<section title="Date" anchor="">
2275  <iref primary="true" item="Date header"/>
2276  <iref primary="true" item="Headers" subitem="Date"/>
2278   The Date general-header field represents the date and time at which
2279   the message was originated, having the same semantics as orig-date in
2280   Section 3.6.1 of <xref target="RFC2822"/>. The field value is an HTTP-date, as described in <xref target=""/>;
2281   it MUST be sent in rfc1123-date format.
2283<figure><iref primary="true" item="Grammar" subitem="Date"/><artwork type="abnf2616"><![CDATA[
2284  Date  = "Date" ":" HTTP-date
2287   An example is
2289<figure><artwork type="example"><![CDATA[
2290    Date: Tue, 15 Nov 1994 08:12:31 GMT
2293   Origin servers MUST include a Date header field in all responses,
2294   except in these cases:
2295  <list style="numbers">
2296      <t>If the response status code is 100 (Continue) or 101 (Switching
2297         Protocols), the response MAY include a Date header field, at
2298         the server's option.</t>
2300      <t>If the response status code conveys a server error, e.g. 500
2301         (Internal Server Error) or 503 (Service Unavailable), and it is
2302         inconvenient or impossible to generate a valid Date.</t>
2304      <t>If the server does not have a clock that can provide a
2305         reasonable approximation of the current time, its responses
2306         MUST NOT include a Date header field. In this case, the rules
2307         in <xref target="clockless.origin.server.operation"/> MUST be followed.</t>
2308  </list>
2311   A received message that does not have a Date header field MUST be
2312   assigned one by the recipient if the message will be cached by that
2313   recipient or gatewayed via a protocol which requires a Date. An HTTP
2314   implementation without a clock MUST NOT cache responses without
2315   revalidating them on every use. An HTTP cache, especially a shared
2316   cache, SHOULD use a mechanism, such as NTP <xref target="RFC1305"/>, to synchronize its
2317   clock with a reliable external standard.
2320   Clients SHOULD only send a Date header field in messages that include
2321   an entity-body, as in the case of the PUT and POST requests, and even
2322   then it is optional. A client without a clock MUST NOT send a Date
2323   header field in a request.
2326   The HTTP-date sent in a Date header SHOULD NOT  represent a date and
2327   time subsequent to the generation of the message. It SHOULD represent
2328   the best available approximation of the date and time of message
2329   generation, unless the implementation has no means of generating a
2330   reasonably accurate date and time. In theory, the date ought to
2331   represent the moment just before the entity is generated. In
2332   practice, the date can be generated at any time during the message
2333   origination without affecting its semantic value.
2336<section title="Clockless Origin Server Operation" anchor="clockless.origin.server.operation">
2338   Some origin server implementations might not have a clock available.
2339   An origin server without a clock MUST NOT assign Expires or Last-Modified
2340   values to a response, unless these values were associated
2341   with the resource by a system or user with a reliable clock. It MAY
2342   assign an Expires value that is known, at or before server
2343   configuration time, to be in the past (this allows "pre-expiration"
2344   of responses without storing separate Expires values for each
2345   resource).
2350<section title="Host" anchor="">
2351  <iref primary="true" item="Host header"/>
2352  <iref primary="true" item="Headers" subitem="Host"/>
2354   The Host request-header field specifies the Internet host and port
2355   number of the resource being requested, as obtained from the original
2356   URI given by the user or referring resource (generally an HTTP URL,
2357   as described in <xref target="http.url"/>). The Host field value MUST represent
2358   the naming authority of the origin server or gateway given by the
2359   original URL. This allows the origin server or gateway to
2360   differentiate between internally-ambiguous URLs, such as the root "/"
2361   URL of a server for multiple host names on a single IP address.
2363<figure><iref primary="true" item="Grammar" subitem="Host"/><artwork type="abnf2616"><![CDATA[
2364  Host = "Host" ":" host [ ":" port ] ; Section 3.2.2
2367   A "host" without any trailing port information implies the default
2368   port for the service requested (e.g., "80" for an HTTP URL). For
2369   example, a request on the origin server for
2370   &lt;; would properly include:
2372<figure><artwork type="example"><![CDATA[
2373    GET /pub/WWW/ HTTP/1.1
2374    Host:
2377   A client MUST include a Host header field in all HTTP/1.1 request
2378   messages. If the requested URI does not include an Internet host
2379   name for the service being requested, then the Host header field MUST
2380   be given with an empty value. An HTTP/1.1 proxy MUST ensure that any
2381   request message it forwards does contain an appropriate Host header
2382   field that identifies the service being requested by the proxy. All
2383   Internet-based HTTP/1.1 servers MUST respond with a 400 (Bad Request)
2384   status code to any HTTP/1.1 request message which lacks a Host header
2385   field.
2388   See Sections <xref target="" format="counter"/>
2389   and <xref target="" format="counter"/>
2390   for other requirements relating to Host.
2394<section title="TE" anchor="header.te">
2395  <iref primary="true" item="TE header"/>
2396  <iref primary="true" item="Headers" subitem="TE"/>
2398   The TE request-header field indicates what extension transfer-codings
2399   it is willing to accept in the response and whether or not it is
2400   willing to accept trailer fields in a chunked transfer-coding. Its
2401   value may consist of the keyword "trailers" and/or a comma-separated
2402   list of extension transfer-coding names with optional accept
2403   parameters (as described in <xref target="transfer.codings"/>).
2405<figure><iref primary="true" item="Grammar" subitem="TE"/><iref primary="true" item="Grammar" subitem="t-codings"/><artwork type="abnf2616"><![CDATA[
2406  TE        = "TE" ":" #( t-codings )
2407  t-codings = "trailers" | ( transfer-extension [ accept-params ] )
2410   The presence of the keyword "trailers" indicates that the client is
2411   willing to accept trailer fields in a chunked transfer-coding, as
2412   defined in <xref target="chunked.transfer.encoding"/>. This keyword is reserved for use with
2413   transfer-coding values even though it does not itself represent a
2414   transfer-coding.
2417   Examples of its use are:
2419<figure><artwork type="example"><![CDATA[
2420    TE: deflate
2421    TE:
2422    TE: trailers, deflate;q=0.5
2425   The TE header field only applies to the immediate connection.
2426   Therefore, the keyword MUST be supplied within a Connection header
2427   field (<xref target="header.connection"/>) whenever TE is present in an HTTP/1.1 message.
2430   A server tests whether a transfer-coding is acceptable, according to
2431   a TE field, using these rules:
2432  <list style="numbers">
2433    <t>The "chunked" transfer-coding is always acceptable. If the
2434         keyword "trailers" is listed, the client indicates that it is
2435         willing to accept trailer fields in the chunked response on
2436         behalf of itself and any downstream clients. The implication is
2437         that, if given, the client is stating that either all
2438         downstream clients are willing to accept trailer fields in the
2439         forwarded response, or that it will attempt to buffer the
2440         response on behalf of downstream recipients.
2441      <vspace blankLines="1"/>
2442         Note: HTTP/1.1 does not define any means to limit the size of a
2443         chunked response such that a client can be assured of buffering
2444         the entire response.</t>
2445    <t>If the transfer-coding being tested is one of the transfer-codings
2446         listed in the TE field, then it is acceptable unless it
2447         is accompanied by a qvalue of 0. (As defined in Section 2.4 of <xref target="Part3"/>, a
2448         qvalue of 0 means "not acceptable.")</t>
2449    <t>If multiple transfer-codings are acceptable, then the
2450         acceptable transfer-coding with the highest non-zero qvalue is
2451         preferred.  The "chunked" transfer-coding always has a qvalue
2452         of 1.</t>
2453  </list>
2456   If the TE field-value is empty or if no TE field is present, the only
2457   transfer-coding  is "chunked". A message with no transfer-coding is
2458   always acceptable.
2462<section title="Trailer" anchor="header.trailer">
2463  <iref primary="true" item="Trailer header"/>
2464  <iref primary="true" item="Headers" subitem="Trailer"/>
2466   The Trailer general field value indicates that the given set of
2467   header fields is present in the trailer of a message encoded with
2468   chunked transfer-coding.
2470<figure><iref primary="true" item="Grammar" subitem="Trailer"/><artwork type="abnf2616"><![CDATA[
2471  Trailer  = "Trailer" ":" 1#field-name
2474   An HTTP/1.1 message SHOULD include a Trailer header field in a
2475   message using chunked transfer-coding with a non-empty trailer. Doing
2476   so allows the recipient to know which header fields to expect in the
2477   trailer.
2480   If no Trailer header field is present, the trailer SHOULD NOT  include
2481   any header fields. See <xref target="chunked.transfer.encoding"/> for restrictions on the use of
2482   trailer fields in a "chunked" transfer-coding.
2485   Message header fields listed in the Trailer header field MUST NOT
2486   include the following header fields:
2487  <list style="symbols">
2488    <t>Transfer-Encoding</t>
2489    <t>Content-Length</t>
2490    <t>Trailer</t>
2491  </list>
2495<section title="Transfer-Encoding" anchor="header.transfer-encoding">
2496  <iref primary="true" item="Transfer-Encoding header"/>
2497  <iref primary="true" item="Headers" subitem="Transfer-Encoding"/>
2499   The Transfer-Encoding general-header field indicates what (if any)
2500   type of transformation has been applied to the message body in order
2501   to safely transfer it between the sender and the recipient. This
2502   differs from the content-coding in that the transfer-coding is a
2503   property of the message, not of the entity.
2505<figure><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/><artwork type="abnf2616"><![CDATA[
2506  Transfer-Encoding       = "Transfer-Encoding" ":" 1#transfer-coding
2509   Transfer-codings are defined in <xref target="transfer.codings"/>. An example is:
2511<figure><artwork type="example"><![CDATA[
2512  Transfer-Encoding: chunked
2515   If multiple encodings have been applied to an entity, the transfer-codings
2516   MUST be listed in the order in which they were applied.
2517   Additional information about the encoding parameters MAY be provided
2518   by other entity-header fields not defined by this specification.
2521   Many older HTTP/1.0 applications do not understand the Transfer-Encoding
2522   header.
2526<section title="Upgrade" anchor="header.upgrade">
2527  <iref primary="true" item="Upgrade header"/>
2528  <iref primary="true" item="Headers" subitem="Upgrade"/>
2530   The Upgrade general-header allows the client to specify what
2531   additional communication protocols it supports and would like to use
2532   if the server finds it appropriate to switch protocols. The server
2533   MUST use the Upgrade header field within a 101 (Switching Protocols)
2534   response to indicate which protocol(s) are being switched.
2536<figure><iref primary="true" item="Grammar" subitem="Upgrade"/><artwork type="abnf2616"><![CDATA[
2537  Upgrade        = "Upgrade" ":" 1#product
2540   For example,
2542<figure><artwork type="example"><![CDATA[
2543    Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
2546   The Upgrade header field is intended to provide a simple mechanism
2547   for transition from HTTP/1.1 to some other, incompatible protocol. It
2548   does so by allowing the client to advertise its desire to use another
2549   protocol, such as a later version of HTTP with a higher major version
2550   number, even though the current request has been made using HTTP/1.1.
2551   This eases the difficult transition between incompatible protocols by
2552   allowing the client to initiate a request in the more commonly
2553   supported protocol while indicating to the server that it would like
2554   to use a "better" protocol if available (where "better" is determined
2555   by the server, possibly according to the nature of the method and/or
2556   resource being requested).
2559   The Upgrade header field only applies to switching application-layer
2560   protocols upon the existing transport-layer connection. Upgrade
2561   cannot be used to insist on a protocol change; its acceptance and use
2562   by the server is optional. The capabilities and nature of the
2563   application-layer communication after the protocol change is entirely
2564   dependent upon the new protocol chosen, although the first action
2565   after changing the protocol MUST be a response to the initial HTTP
2566   request containing the Upgrade header field.
2569   The Upgrade header field only applies to the immediate connection.
2570   Therefore, the upgrade keyword MUST be supplied within a Connection
2571   header field (<xref target="header.connection"/>) whenever Upgrade is present in an
2572   HTTP/1.1 message.
2575   The Upgrade header field cannot be used to indicate a switch to a
2576   protocol on a different connection. For that purpose, it is more
2577   appropriate to use a 301, 302, 303, or 305 redirection response.
2580   This specification only defines the protocol name "HTTP" for use by
2581   the family of Hypertext Transfer Protocols, as defined by the HTTP
2582   version rules of <xref target="http.version"/> and future updates to this
2583   specification. Any token can be used as a protocol name; however, it
2584   will only be useful if both the client and server associate the name
2585   with the same protocol.
2589<section title="Via" anchor="header.via">
2590  <iref primary="true" item="Via header"/>
2591  <iref primary="true" item="Headers" subitem="Via"/>
2593   The Via general-header field MUST be used by gateways and proxies to
2594   indicate the intermediate protocols and recipients between the user
2595   agent and the server on requests, and between the origin server and
2596   the client on responses. It is analogous to the "Received" field of
2597   <xref target="RFC2822"/> and is intended to be used for tracking message forwards,
2598   avoiding request loops, and identifying the protocol capabilities of
2599   all senders along the request/response chain.
2601<figure><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"/><artwork type="abnf2616"><![CDATA[
2602  Via =  "Via" ":" 1#( received-protocol received-by [ comment ] )
2603  received-protocol = [ protocol-name "/" ] protocol-version
2604  protocol-name     = token
2605  protocol-version  = token
2606  received-by       = ( host [ ":" port ] ) | pseudonym
2607  pseudonym         = token
2610   The received-protocol indicates the protocol version of the message
2611   received by the server or client along each segment of the
2612   request/response chain. The received-protocol version is appended to
2613   the Via field value when the message is forwarded so that information
2614   about the protocol capabilities of upstream applications remains
2615   visible to all recipients.
2618   The protocol-name is optional if and only if it would be "HTTP". The
2619   received-by field is normally the host and optional port number of a
2620   recipient server or client that subsequently forwarded the message.
2621   However, if the real host is considered to be sensitive information,
2622   it MAY be replaced by a pseudonym. If the port is not given, it MAY
2623   be assumed to be the default port of the received-protocol.
2626   Multiple Via field values represents each proxy or gateway that has
2627   forwarded the message. Each recipient MUST append its information
2628   such that the end result is ordered according to the sequence of
2629   forwarding applications.
2632   Comments MAY be used in the Via header field to identify the software
2633   of the recipient proxy or gateway, analogous to the User-Agent and
2634   Server header fields. However, all comments in the Via field are
2635   optional and MAY be removed by any recipient prior to forwarding the
2636   message.
2639   For example, a request message could be sent from an HTTP/1.0 user
2640   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2641   forward the request to a public proxy at, which completes
2642   the request by forwarding it to the origin server at
2643   The request received by would then have the following
2644   Via header field:
2646<figure><artwork type="example"><![CDATA[
2647    Via: 1.0 fred, 1.1 (Apache/1.1)
2650   Proxies and gateways used as a portal through a network firewall
2651   SHOULD NOT, by default, forward the names and ports of hosts within
2652   the firewall region. This information SHOULD only be propagated if
2653   explicitly enabled. If not enabled, the received-by host of any host
2654   behind the firewall SHOULD be replaced by an appropriate pseudonym
2655   for that host.
2658   For organizations that have strong privacy requirements for hiding
2659   internal structures, a proxy MAY combine an ordered subsequence of
2660   Via header field entries with identical received-protocol values into
2661   a single such entry. For example,
2663<figure><artwork type="example"><![CDATA[
2664    Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2667        could be collapsed to
2669<figure><artwork type="example"><![CDATA[
2670    Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2673   Applications SHOULD NOT  combine multiple entries unless they are all
2674   under the same organizational control and the hosts have already been
2675   replaced by pseudonyms. Applications MUST NOT combine entries which
2676   have different received-protocol values.
2682<section title="IANA Considerations" anchor="IANA.considerations">
2684   TBD.
2688<section title="Security Considerations" anchor="security.considerations">
2690   This section is meant to inform application developers, information
2691   providers, and users of the security limitations in HTTP/1.1 as
2692   described by this document. The discussion does not include
2693   definitive solutions to the problems revealed, though it does make
2694   some suggestions for reducing security risks.
2697<section title="Personal Information" anchor="personal.information">
2699   HTTP clients are often privy to large amounts of personal information
2700   (e.g. the user's name, location, mail address, passwords, encryption
2701   keys, etc.), and SHOULD be very careful to prevent unintentional
2702   leakage of this information via the HTTP protocol to other sources.
2703   We very strongly recommend that a convenient interface be provided
2704   for the user to control dissemination of such information, and that
2705   designers and implementors be particularly careful in this area.
2706   History shows that errors in this area often create serious security
2707   and/or privacy problems and generate highly adverse publicity for the
2708   implementor's company.
2712<section title="Abuse of Server Log Information" anchor="abuse.of.server.log.information">
2714   A server is in the position to save personal data about a user's
2715   requests which might identify their reading patterns or subjects of
2716   interest. This information is clearly confidential in nature and its
2717   handling can be constrained by law in certain countries. People using
2718   the HTTP protocol to provide data are responsible for ensuring that
2719   such material is not distributed without the permission of any
2720   individuals that are identifiable by the published results.
2724<section title="Attacks Based On File and Path Names" anchor="attack.pathname">
2726   Implementations of HTTP origin servers SHOULD be careful to restrict
2727   the documents returned by HTTP requests to be only those that were
2728   intended by the server administrators. If an HTTP server translates
2729   HTTP URIs directly into file system calls, the server MUST take
2730   special care not to serve files that were not intended to be
2731   delivered to HTTP clients. For example, UNIX, Microsoft Windows, and
2732   other operating systems use ".." as a path component to indicate a
2733   directory level above the current one. On such a system, an HTTP
2734   server MUST disallow any such construct in the Request-URI if it
2735   would otherwise allow access to a resource outside those intended to
2736   be accessible via the HTTP server. Similarly, files intended for
2737   reference only internally to the server (such as access control
2738   files, configuration files, and script code) MUST be protected from
2739   inappropriate retrieval, since they might contain sensitive
2740   information. Experience has shown that minor bugs in such HTTP server
2741   implementations have turned into security risks.
2745<section title="DNS Spoofing" anchor="dns.spoofing">
2747   Clients using HTTP rely heavily on the Domain Name Service, and are
2748   thus generally prone to security attacks based on the deliberate
2749   mis-association of IP addresses and DNS names. Clients need to be
2750   cautious in assuming the continuing validity of an IP number/DNS name
2751   association.
2754   In particular, HTTP clients SHOULD rely on their name resolver for
2755   confirmation of an IP number/DNS name association, rather than
2756   caching the result of previous host name lookups. Many platforms
2757   already can cache host name lookups locally when appropriate, and
2758   they SHOULD be configured to do so. It is proper for these lookups to
2759   be cached, however, only when the TTL (Time To Live) information
2760   reported by the name server makes it likely that the cached
2761   information will remain useful.
2764   If HTTP clients cache the results of host name lookups in order to
2765   achieve a performance improvement, they MUST observe the TTL
2766   information reported by DNS.
2769   If HTTP clients do not observe this rule, they could be spoofed when
2770   a previously-accessed server's IP address changes. As network
2771   renumbering is expected to become increasingly common <xref target="RFC1900"/>, the
2772   possibility of this form of attack will grow. Observing this
2773   requirement thus reduces this potential security vulnerability.
2776   This requirement also improves the load-balancing behavior of clients
2777   for replicated servers using the same DNS name and reduces the
2778   likelihood of a user's experiencing failure in accessing sites which
2779   use that strategy.
2783<section title="Proxies and Caching" anchor="attack.proxies">
2785   By their very nature, HTTP proxies are men-in-the-middle, and
2786   represent an opportunity for man-in-the-middle attacks. Compromise of
2787   the systems on which the proxies run can result in serious security
2788   and privacy problems. Proxies have access to security-related
2789   information, personal information about individual users and
2790   organizations, and proprietary information belonging to users and
2791   content providers. A compromised proxy, or a proxy implemented or
2792   configured without regard to security and privacy considerations,
2793   might be used in the commission of a wide range of potential attacks.
2796   Proxy operators should protect the systems on which proxies run as
2797   they would protect any system that contains or transports sensitive
2798   information. In particular, log information gathered at proxies often
2799   contains highly sensitive personal information, and/or information
2800   about organizations. Log information should be carefully guarded, and
2801   appropriate guidelines for use developed and followed. (<xref target="abuse.of.server.log.information"/>).
2804   Proxy implementors should consider the privacy and security
2805   implications of their design and coding decisions, and of the
2806   configuration options they provide to proxy operators (especially the
2807   default configuration).
2810   Users of a proxy need to be aware that they are no trustworthier than
2811   the people who run the proxy; HTTP itself cannot solve this problem.
2814   The judicious use of cryptography, when appropriate, may suffice to
2815   protect against a broad range of security and privacy attacks. Such
2816   cryptography is beyond the scope of the HTTP/1.1 specification.
2820<section title="Denial of Service Attacks on Proxies" anchor="attack.DoS">
2822   They exist. They are hard to defend against. Research continues.
2823   Beware.
2828<section title="Acknowledgments" anchor="ack">
2830   This specification makes heavy use of the augmented BNF and generic
2831   constructs defined by David H. Crocker for <xref target="RFC822ABNF"/>. Similarly, it
2832   reuses many of the definitions provided by Nathaniel Borenstein and
2833   Ned Freed for MIME <xref target="RFC2045"/>. We hope that their inclusion in this
2834   specification will help reduce past confusion over the relationship
2835   between HTTP and Internet mail message formats.
2838   The HTTP protocol has evolved considerably over the years. It has
2839   benefited from a large and active developer community--the many
2840   people who have participated on the www-talk mailing list--and it is
2841   that community which has been most responsible for the success of
2842   HTTP and of the World-Wide Web in general. Marc Andreessen, Robert
2843   Cailliau, Daniel W. Connolly, Bob Denny, John Franks, Jean-Francois
2844   Groff, Phillip M. Hallam-Baker, Hakon W. Lie, Ari Luotonen, Rob
2845   McCool, Lou Montulli, Dave Raggett, Tony Sanders, and Marc
2846   VanHeyningen deserve special recognition for their efforts in
2847   defining early aspects of the protocol.
2850   This document has benefited greatly from the comments of all those
2851   participating in the HTTP-WG. In addition to those already mentioned,
2852   the following individuals have contributed to this specification:
2855   Gary Adams, Harald Tveit Alvestrand, Keith Ball, Brian Behlendorf,
2856   Paul Burchard, Maurizio Codogno, Mike Cowlishaw, Roman Czyborra,
2857   Michael A. Dolan, Daniel DuBois, David J. Fiander, Alan Freier, Marc Hedlund, Greg Herlihy,
2858   Koen Holtman, Alex Hopmann, Bob Jernigan, Shel Kaphan, Rohit Khare,
2859   John Klensin, Martijn Koster, Alexei Kosut, David M. Kristol,
2860   Daniel LaLiberte, Ben Laurie, Paul J. Leach, Albert Lunde,
2861   John C. Mallery, Jean-Philippe Martin-Flatin, Mitra, David Morris,
2862   Gavin Nicol, Ross Patterson, Bill Perry, Jeffrey Perry, Scott Powers, Owen Rees,
2863   Luigi Rizzo, David Robinson, Marc Salomon, Rich Salz,
2864   Allan M. Schiffman, Jim Seidman, Chuck Shotton, Eric W. Sink,
2865   Simon E. Spero, Richard N. Taylor, Robert S. Thau,
2866   Bill (BearHeart) Weinman, Francois Yergeau, Mary Ellen Zurko,
2867   Josh Cohen.
2870   Thanks to the "cave men" of Palo Alto. You know who you are.
2873   Jim Gettys (the editor of <xref target="RFC2616"/>) wishes particularly
2874   to thank Roy Fielding, the editor of <xref target="RFC2068"/>, along
2875   with John Klensin, Jeff Mogul, Paul Leach, Dave Kristol, Koen
2876   Holtman, John Franks, Josh Cohen, Alex Hopmann, Scott Lawrence, and
2877   Larry Masinter for their help. And thanks go particularly to Jeff
2878   Mogul and Scott Lawrence for performing the "MUST/MAY/SHOULD" audit.
2881   The Apache Group, Anselm Baird-Smith, author of Jigsaw, and Henrik
2882   Frystyk implemented RFC 2068 early, and we wish to thank them for the
2883   discovery of many of the problems that this document attempts to
2884   rectify.
2891<references title="Normative References">
2893<reference anchor="ISO-8859-1">
2894  <front>
2895    <title>
2896     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
2897    </title>
2898    <author>
2899      <organization>International Organization for Standardization</organization>
2900    </author>
2901    <date year="1998"/>
2902  </front>
2903  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
2906<reference anchor="Part2">
2907  <front>
2908    <title abbrev="HTTP/1.1">HTTP/1.1, part 2: Message Semantics</title>
2909    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
2910      <organization abbrev="Day Software">Day Software</organization>
2911      <address><email></email></address>
2912    </author>
2913    <author initials="J." surname="Gettys" fullname="Jim Gettys">
2914      <organization>One Laptop per Child</organization>
2915      <address><email></email></address>
2916    </author>
2917    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
2918      <organization abbrev="HP">Hewlett-Packard Company</organization>
2919      <address><email></email></address>
2920    </author>
2921    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
2922      <organization abbrev="Microsoft">Microsoft Corporation</organization>
2923      <address><email></email></address>
2924    </author>
2925    <author initials="L." surname="Masinter" fullname="Larry Masinter">
2926      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
2927      <address><email></email></address>
2928    </author>
2929    <author initials="P." surname="Leach" fullname="Paul J. Leach">
2930      <organization abbrev="Microsoft">Microsoft Corporation</organization>
2931      <address><email></email></address>
2932    </author>
2933    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
2934      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
2935      <address><email></email></address>
2936    </author>
2937    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
2938      <organization abbrev="W3C">World Wide Web Consortium</organization>
2939      <address><email></email></address>
2940    </author>
2941    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
2942      <organization abbrev="greenbytes">greenbytes GmbH</organization>
2943      <address><email></email></address>
2944    </author>
2945    <date month="January" year="2008"/>
2946  </front>
2947  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-01"/>
2951<reference anchor="Part3">
2952  <front>
2953    <title abbrev="HTTP/1.1">HTTP/1.1, part 3: Message Payload and Content Negotiation</title>
2954    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
2955      <organization abbrev="Day Software">Day Software</organization>
2956      <address><email></email></address>
2957    </author>
2958    <author initials="J." surname="Gettys" fullname="Jim Gettys">
2959      <organization>One Laptop per Child</organization>
2960      <address><email></email></address>
2961    </author>
2962    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
2963      <organization abbrev="HP">Hewlett-Packard Company</organization>
2964      <address><email></email></address>
2965    </author>
2966    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
2967      <organization abbrev="Microsoft">Microsoft Corporation</organization>
2968      <address><email></email></address>
2969    </author>
2970    <author initials="L." surname="Masinter" fullname="Larry Masinter">
2971      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
2972      <address><email></email></address>
2973    </author>
2974    <author initials="P." surname="Leach" fullname="Paul J. Leach">
2975      <organization abbrev="Microsoft">Microsoft Corporation</organization>
2976      <address><email></email></address>
2977    </author>
2978    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
2979      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
2980      <address><email></email></address>
2981    </author>
2982    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
2983      <organization abbrev="W3C">World Wide Web Consortium</organization>
2984      <address><email></email></address>
2985    </author>
2986    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
2987      <organization abbrev="greenbytes">greenbytes GmbH</organization>
2988      <address><email></email></address>
2989    </author>
2990    <date month="January" year="2008"/>
2991  </front>
2992  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p3-payload-01"/>
2996<reference anchor="Part5">
2997  <front>
2998    <title abbrev="HTTP/1.1">HTTP/1.1, part 5: Range Requests and Partial Responses</title>
2999    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3000      <organization abbrev="Day Software">Day Software</organization>
3001      <address><email></email></address>
3002    </author>
3003    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3004      <organization>One Laptop per Child</organization>
3005      <address><email></email></address>
3006    </author>
3007    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3008      <organization abbrev="HP">Hewlett-Packard Company</organization>
3009      <address><email></email></address>
3010    </author>
3011    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3012      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3013      <address><email></email></address>
3014    </author>
3015    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3016      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3017      <address><email></email></address>
3018    </author>
3019    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3020      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3021      <address><email></email></address>
3022    </author>
3023    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3024      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3025      <address><email></email></address>
3026    </author>
3027    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3028      <organization abbrev="W3C">World Wide Web Consortium</organization>
3029      <address><email></email></address>
3030    </author>
3031    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3032      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3033      <address><email></email></address>
3034    </author>
3035    <date month="January" year="2008"/>
3036  </front>
3037  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-01"/>
3041<reference anchor="Part6">
3042  <front>
3043    <title abbrev="HTTP/1.1">HTTP/1.1, part 6: Caching</title>
3044    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3045      <organization abbrev="Day Software">Day Software</organization>
3046      <address><email></email></address>
3047    </author>
3048    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3049      <organization>One Laptop per Child</organization>
3050      <address><email></email></address>
3051    </author>
3052    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3053      <organization abbrev="HP">Hewlett-Packard Company</organization>
3054      <address><email></email></address>
3055    </author>
3056    <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
3057      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3058      <address><email></email></address>
3059    </author>
3060    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3061      <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
3062      <address><email></email></address>
3063    </author>
3064    <author initials="P." surname="Leach" fullname="Paul J. Leach">
3065      <organization abbrev="Microsoft">Microsoft Corporation</organization>
3066      <address><email></email></address>
3067    </author>
3068    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3069      <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
3070      <address><email></email></address>
3071    </author>
3072    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
3073      <organization abbrev="W3C">World Wide Web Consortium</organization>
3074      <address><email></email></address>
3075    </author>
3076    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3077      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3078      <address><email></email></address>
3079    </author>
3080    <date month="January" year="2008"/>
3081  </front>
3082  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-01"/>
3086<reference anchor="RFC822ABNF">
3087  <front>
3088    <title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3089    <author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3090      <organization>University of Delaware, Dept. of Electrical Engineering</organization>
3091      <address><email>DCrocker@UDel-Relay</email></address>
3092    </author>
3093    <date month="August" day="13" year="1982"/>
3094  </front>
3095  <seriesInfo name="STD" value="11"/>
3096  <seriesInfo name="RFC" value="822"/>
3099<reference anchor="RFC2045">
3100  <front>
3101    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
3102    <author initials="N." surname="Freed" fullname="Ned Freed">
3103      <organization>Innosoft International, Inc.</organization>
3104      <address><email></email></address>
3105    </author>
3106    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
3107      <organization>First Virtual Holdings</organization>
3108      <address><email></email></address>
3109    </author>
3110    <date month="November" year="1996"/>
3111  </front>
3112  <seriesInfo name="RFC" value="2045"/>
3115<reference anchor="RFC2047">
3116  <front>
3117    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
3118    <author initials="K." surname="Moore" fullname="Keith Moore">
3119      <organization>University of Tennessee</organization>
3120      <address><email></email></address>
3121    </author>
3122    <date month="November" year="1996"/>
3123  </front>
3124  <seriesInfo name="RFC" value="2047"/>
3127<reference anchor="RFC2119">
3128  <front>
3129    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
3130    <author initials="S." surname="Bradner" fullname="Scott Bradner">
3131      <organization>Harvard University</organization>
3132      <address><email></email></address>
3133    </author>
3134    <date month="March" year="1997"/>
3135  </front>
3136  <seriesInfo name="BCP" value="14"/>
3137  <seriesInfo name="RFC" value="2119"/>
3140<reference anchor="RFC2396">
3141  <front>
3142    <title abbrev="URI Generic Syntax">Uniform Resource Identifiers (URI): Generic Syntax</title>
3143    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3144      <organization abbrev="MIT/LCS">World Wide Web Consortium</organization>
3145      <address><email></email></address>
3146    </author>
3147    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3148      <organization abbrev="U.C. Irvine">Department of Information and Computer Science</organization>
3149      <address><email></email></address>
3150    </author>
3151    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3152      <organization abbrev="Xerox Corporation">Xerox PARC</organization>
3153      <address><email></email></address>
3154    </author>
3155    <date month="August" year="1998"/>
3156  </front>
3157  <seriesInfo name="RFC" value="2396"/>
3160<reference anchor="RFC4288">
3161  <front>
3162    <title>Media Type Specifications and Registration Procedures</title>
3163    <author initials="N." surname="Freed" fullname="N. Freed">
3164      <organization>Sun Microsystems</organization>
3165      <address>
3166        <email></email>
3167      </address>
3168    </author>
3169    <author initials="J." surname="Klensin" fullname="J. Klensin">
3170      <organization/>
3171      <address>
3172        <email></email>
3173      </address>
3174    </author>
3175    <date year="2005" month="December"/>
3176  </front>
3177  <seriesInfo name="BCP" value="13"/>
3178  <seriesInfo name="RFC" value="4288"/>
3181<reference anchor="USASCII">
3182  <front>
3183    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
3184    <author>
3185      <organization>American National Standards Institute</organization>
3186    </author>
3187    <date year="1986"/>
3188  </front>
3189  <seriesInfo name="ANSI" value="X3.4"/>
3194<references title="Informative References">
3196<reference anchor="Nie1997" target="">
3197  <front>
3198    <title>Network Performance Effects of HTTP/1.1, CSS1, and PNG</title>
3199    <author initials="H.F.." surname="Nielsen" fullname="H.F. Nielsen">
3200      <organization/>
3201    </author>
3202    <author initials="J." surname="Gettys" fullname="J. Gettys">
3203      <organization/>
3204    </author>
3205    <author initials="E." surname="Prud'hommeaux" fullname="E. Prud'hommeaux">
3206      <organization/>
3207    </author>
3208    <author initials="H." surname="Lie" fullname="H. Lie">
3209      <organization/>
3210    </author>
3211    <author initials="C." surname="Lilley" fullname="C. Lilley">
3212      <organization/>
3213    </author>
3214    <date year="1997" month="September"/>
3215  </front>
3216  <seriesInfo name="ACM" value="Proceedings of the ACM SIGCOMM '97 conference on Applications, technologies, architectures, and protocols for computer communication SIGCOMM '97"/>
3219<reference anchor="Pad1995">
3220  <front>
3221    <title>Improving HTTP Latency</title>
3222    <author initials="V.N." surname="Padmanabhan" fullname="Venkata N. Padmanabhan">
3223      <organization/>
3224    </author>
3225    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3226      <organization/>
3227    </author>
3228    <date year="1995" month="December"/>
3229  </front>
3230  <seriesInfo name="Computer Networks and ISDN Systems" value="v. 28, pp. 25-35"/>
3231  <annotation>
3232    Slightly revised version of paper in Proc. 2nd International WWW Conference '94: Mosaic and the Web, Oct. 1994,
3233    which is available at <eref target=""/>.
3234  </annotation>
3237<reference anchor="RFC822">
3238  <front>
3239    <title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3240    <author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3241      <organization>University of Delaware, Dept. of Electrical Engineering</organization>
3242      <address><email>DCrocker@UDel-Relay</email></address>
3243    </author>
3244    <date month="August" day="13" year="1982"/>
3245  </front>
3246  <seriesInfo name="STD" value="11"/>
3247  <seriesInfo name="RFC" value="822"/>
3250<reference anchor="RFC959">
3251  <front>
3252    <title abbrev="File Transfer Protocol">File Transfer Protocol</title>
3253    <author initials="J." surname="Postel" fullname="J. Postel">
3254      <organization>Information Sciences Institute (ISI)</organization>
3255    </author>
3256    <author initials="J." surname="Reynolds" fullname="J. Reynolds">
3257      <organization/>
3258    </author>
3259    <date month="October" year="1985"/>
3260  </front>
3261  <seriesInfo name="STD" value="9"/>
3262  <seriesInfo name="RFC" value="959"/>
3265<reference anchor="RFC1123">
3266  <front>
3267    <title>Requirements for Internet Hosts - Application and Support</title>
3268    <author initials="R." surname="Braden" fullname="Robert Braden">
3269      <organization>University of Southern California (USC), Information Sciences Institute</organization>
3270      <address><email>Braden@ISI.EDU</email></address>
3271    </author>
3272    <date month="October" year="1989"/>
3273  </front>
3274  <seriesInfo name="STD" value="3"/>
3275  <seriesInfo name="RFC" value="1123"/>
3278<reference anchor="RFC1305">
3279  <front>
3280    <title>Network Time Protocol (Version 3) Specification, Implementation</title>
3281    <author initials="D." surname="Mills" fullname="David L. Mills">
3282      <organization>University of Delaware, Electrical Engineering Department</organization>
3283      <address><email></email></address>
3284    </author>
3285    <date month="March" year="1992"/>
3286  </front>
3287  <seriesInfo name="RFC" value="1305"/>
3290<reference anchor="RFC1436">
3291  <front>
3292    <title abbrev="Gopher">The Internet Gopher Protocol (a distributed document search and retrieval protocol)</title>
3293    <author initials="F." surname="Anklesaria" fullname="Farhad Anklesaria">
3294      <organization>University of Minnesota, Computer and Information Services</organization>
3295      <address><email></email></address>
3296    </author>
3297    <author initials="M." surname="McCahill" fullname="Mark McCahill">
3298      <organization>University of Minnesota, Computer and Information Services</organization>
3299      <address><email></email></address>
3300    </author>
3301    <author initials="P." surname="Lindner" fullname="Paul Lindner">
3302      <organization>University of Minnesota, Computer and Information Services</organization>
3303      <address><email></email></address>
3304    </author>
3305    <author initials="D." surname="Johnson" fullname="David Johnson">
3306      <organization>University of Minnesota, Computer and Information Services</organization>
3307      <address><email></email></address>
3308    </author>
3309    <author initials="D." surname="Torrey" fullname="Daniel Torrey">
3310      <organization>University of Minnesota, Computer and Information Services</organization>
3311      <address><email></email></address>
3312    </author>
3313    <author initials="B." surname="Alberti" fullname="Bob Alberti">
3314      <organization>University of Minnesota, Computer and Information Services</organization>
3315      <address><email></email></address>
3316    </author>
3317    <date month="March" year="1993"/>
3318  </front>
3319  <seriesInfo name="RFC" value="1436"/>
3322<reference anchor="RFC1630">
3323  <front>
3324    <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>
3325    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3326      <organization>CERN, World-Wide Web project</organization>
3327      <address><email></email></address>
3328    </author>
3329    <date month="June" year="1994"/>
3330  </front>
3331  <seriesInfo name="RFC" value="1630"/>
3334<reference anchor="RFC1737">
3335  <front>
3336    <title abbrev="Requirements for Uniform Resource Names">Functional Requirements for Uniform Resource Names</title>
3337    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3338      <organization>Xerox Palo Alto Research Center</organization>
3339      <address><email></email></address>
3340    </author>
3341    <author initials="K." surname="Sollins" fullname="Karen Sollins">
3342      <organization>MIT Laboratory for Computer Science</organization>
3343      <address><email></email></address>
3344    </author>
3345    <date month="December" year="1994"/>
3346  </front>
3347  <seriesInfo name="RFC" value="1737"/>
3350<reference anchor="RFC1738">
3351  <front>
3352    <title>Uniform Resource Locators (URL)</title>
3353    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3354      <organization>CERN, World-Wide Web project</organization>
3355      <address><email></email></address>
3356    </author>
3357    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3358      <organization>Xerox PARC</organization>
3359      <address><email></email></address>
3360    </author>
3361    <author initials="M." surname="McCahill" fullname="Mark McCahill">
3362      <organization>University of Minnesota, Computer and Information Services</organization>
3363      <address><email></email></address>
3364    </author>
3365    <date month="December" year="1994"/>
3366  </front>
3367  <seriesInfo name="RFC" value="1738"/>
3370<reference anchor="RFC1808">
3371  <front>
3372    <title>Relative Uniform Resource Locators</title>
3373    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3374      <organization>University of California Irvine, Department of Information and Computer Science</organization>
3375      <address><email></email></address>
3376    </author>
3377    <date month="June" year="1995"/>
3378  </front>
3379  <seriesInfo name="RFC" value="1808"/>
3382<reference anchor="RFC1900">
3383  <front>
3384    <title>Renumbering Needs Work</title>
3385    <author initials="B." surname="Carpenter" fullname="Brian E. Carpenter">
3386      <organization>CERN, Computing and Networks Division</organization>
3387      <address><email></email></address>
3388    </author>
3389    <author initials="Y." surname="Rekhter" fullname="Yakov Rekhter">
3390      <organization>cisco Systems</organization>
3391      <address><email></email></address>
3392    </author>
3393    <date month="February" year="1996"/>
3394  </front>
3395  <seriesInfo name="RFC" value="1900"/>
3398<reference anchor="RFC1945">
3399  <front>
3400    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
3401    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3402      <organization>MIT, Laboratory for Computer Science</organization>
3403      <address><email></email></address>
3404    </author>
3405    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3406      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
3407      <address><email></email></address>
3408    </author>
3409    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3410      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
3411      <address><email></email></address>
3412    </author>
3413    <date month="May" year="1996"/>
3414  </front>
3415  <seriesInfo name="RFC" value="1945"/>
3418<reference anchor="RFC2068">
3419  <front>
3420    <title abbrev="HTTP/1.1">Hypertext Transfer Protocol -- HTTP/1.1</title>
3421    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3422      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
3423      <address><email></email></address>
3424    </author>
3425    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3426      <organization>MIT Laboratory for Computer Science</organization>
3427      <address><email></email></address>
3428    </author>
3429    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3430      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
3431      <address><email></email></address>
3432    </author>
3433    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3434      <organization>MIT Laboratory for Computer Science</organization>
3435      <address><email></email></address>
3436    </author>
3437    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3438      <organization>MIT Laboratory for Computer Science</organization>
3439      <address><email></email></address>
3440    </author>
3441    <date month="January" year="1997"/>
3442  </front>
3443  <seriesInfo name="RFC" value="2068"/>
3446<reference anchor="RFC2145">
3447  <front>
3448    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
3449    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3450      <organization>Western Research Laboratory</organization>
3451      <address><email></email></address>
3452    </author>
3453    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3454      <organization>Department of Information and Computer Science</organization>
3455      <address><email></email></address>
3456    </author>
3457    <author initials="J." surname="Gettys" fullname="Jim Gettys">
3458      <organization>MIT Laboratory for Computer Science</organization>
3459      <address><email></email></address>
3460    </author>
3461    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3462      <organization>W3 Consortium</organization>
3463      <address><email></email></address>
3464    </author>
3465    <date month="May" year="1997"/>
3466  </front>
3467  <seriesInfo name="RFC" value="2145"/>
3470<reference anchor="RFC2324">
3471  <front>
3472    <title abbrev="HTCPCP/1.0">Hyper Text Coffee Pot Control Protocol (HTCPCP/1.0)</title>
3473    <author initials="L." surname="Masinter" fullname="Larry Masinter">
3474      <organization>Xerox Palo Alto Research Center</organization>
3475      <address><email></email></address>
3476    </author>
3477    <date month="April" day="1" year="1998"/>
3478  </front>
3479  <seriesInfo name="RFC" value="2324"/>
3482<reference anchor="RFC2616">
3483  <front>
3484    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
3485    <author initials="R." surname="Fielding" fullname="R. Fielding">
3486      <organization>University of California, Irvine</organization>
3487      <address><email></email></address>
3488    </author>
3489    <author initials="J." surname="Gettys" fullname="J. Gettys">
3490      <organization>W3C</organization>
3491      <address><email></email></address>
3492    </author>
3493    <author initials="J." surname="Mogul" fullname="J. Mogul">
3494      <organization>Compaq Computer Corporation</organization>
3495      <address><email></email></address>
3496    </author>
3497    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
3498      <organization>MIT Laboratory for Computer Science</organization>
3499      <address><email></email></address>
3500    </author>
3501    <author initials="L." surname="Masinter" fullname="L. Masinter">
3502      <organization>Xerox Corporation</organization>
3503      <address><email></email></address>
3504    </author>
3505    <author initials="P." surname="Leach" fullname="P. Leach">
3506      <organization>Microsoft Corporation</organization>
3507      <address><email></email></address>
3508    </author>
3509    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
3510      <organization>W3C</organization>
3511      <address><email></email></address>
3512    </author>
3513    <date month="June" year="1999"/>
3514  </front>
3515  <seriesInfo name="RFC" value="2616"/>
3518<reference anchor="RFC2821">
3519  <front>
3520    <title>Simple Mail Transfer Protocol</title>
3521    <author initials="J." surname="Klensin" fullname="J. Klensin">
3522      <organization>AT&amp;T Laboratories</organization>
3523      <address><email></email></address>
3524    </author>
3525    <date year="2001" month="April"/>
3526  </front>
3527  <seriesInfo name="RFC" value="2821"/>
3530<reference anchor="RFC2822">
3531  <front>
3532    <title>Internet Message Format</title>
3533    <author initials="P." surname="Resnick" fullname="P. Resnick">
3534      <organization>QUALCOMM Incorporated</organization>
3535    </author>
3536    <date year="2001" month="April"/>
3537  </front>
3538  <seriesInfo name="RFC" value="2822"/>
3541<reference anchor="RFC3977">
3542  <front>
3543    <title>Network News Transfer Protocol (NNTP)</title>
3544    <author initials="C." surname="Feather" fullname="C. Feather">
3545      <organization>THUS plc</organization>
3546      <address><email></email></address>
3547    </author>
3548    <date year="2006" month="October"/>
3549  </front>
3550  <seriesInfo name="RFC" value="3977"/>
3553<reference anchor="Spe" target="">
3554  <front>
3555  <title>Analysis of HTTP Performance Problems</title>
3556  <author initials="S." surname="Spero" fullname="Simon E. Spero">
3557    <organization/>
3558  </author>
3559  <date/>
3560  </front>
3563<reference anchor="Tou1998" target="">
3564  <front>
3565  <title>Analysis of HTTP Performance</title>
3566  <author initials="J." surname="Touch" fullname="Joe Touch">
3567    <organization>USC/Information Sciences Institute</organization>
3568    <address><email></email></address>
3569  </author>
3570  <author initials="J." surname="Heidemann" fullname="John Heidemann">
3571    <organization>USC/Information Sciences Institute</organization>
3572    <address><email></email></address>
3573  </author>
3574  <author initials="K." surname="Obraczka" fullname="Katia Obraczka">
3575    <organization>USC/Information Sciences Institute</organization>
3576    <address><email></email></address>
3577  </author>
3578  <date year="1998" month="Aug"/>
3579  </front>
3580  <seriesInfo name="ISI Research Report" value="ISI/RR-98-463"/>
3581  <annotation>(original report dated Aug. 1996)</annotation>
3584<reference anchor="WAIS">
3585  <front>
3586    <title>WAIS Interface Protocol Prototype Functional Specification (v1.5)</title>
3587    <author initials="F." surname="Davis" fullname="F. Davis">
3588      <organization>Thinking Machines Corporation</organization>
3589    </author>
3590    <author initials="B." surname="Kahle" fullname="B. Kahle">
3591      <organization>Thinking Machines Corporation</organization>
3592    </author>
3593    <author initials="H." surname="Morris" fullname="H. Morris">
3594      <organization>Thinking Machines Corporation</organization>
3595    </author>
3596    <author initials="J." surname="Salem" fullname="J. Salem">
3597      <organization>Thinking Machines Corporation</organization>
3598    </author>
3599    <author initials="T." surname="Shen" fullname="T. Shen">
3600      <organization>Thinking Machines Corporation</organization>
3601    </author>
3602    <author initials="R." surname="Wang" fullname="R. Wang">
3603      <organization>Thinking Machines Corporation</organization>
3604    </author>
3605    <author initials="J." surname="Sui" fullname="J. Sui">
3606      <organization>Thinking Machines Corporation</organization>
3607    </author>
3608    <author initials="M." surname="Grinbaum" fullname="M. Grinbaum">
3609      <organization>Thinking Machines Corporation</organization>
3610    </author>
3611    <date month="April" year="1990"/>
3612  </front>
3613  <seriesInfo name="Thinking Machines Corporation" value=""/>
3619<section title="Internet Media Types" anchor="">
3621   In addition to defining the HTTP/1.1 protocol, this document serves
3622   as the specification for the Internet media type "message/http" and
3623   "application/http". The following is to be registered with IANA <xref target="RFC4288"/>.
3625<section title="Internet Media Type message/http" anchor="">
3626<iref item="Media Type" subitem="message/http" primary="true"/>
3627<iref item="message/http Media Type" primary="true"/>
3629   The message/http type can be used to enclose a single HTTP request or
3630   response message, provided that it obeys the MIME restrictions for all
3631   "message" types regarding line length and encodings.
3634  <list style="hanging">
3635    <t hangText="Type name:">
3636      message
3637    </t>
3638    <t hangText="Subtype name:">
3639      http
3640    </t>
3641    <t hangText="Required parameters:">
3642      none
3643    </t>
3644    <t hangText="Optional parameters:">
3645      version, msgtype
3646      <list style="hanging">
3647        <t hangText="version:">
3648          The HTTP-Version number of the enclosed message
3649          (e.g., "1.1"). If not present, the version can be
3650          determined from the first line of the body.
3651        </t>
3652        <t hangText="msgtype:">
3653          The message type -- "request" or "response". If not
3654          present, the type can be determined from the first
3655          line of the body.
3656        </t>
3657      </list>
3658    </t>
3659    <t hangText="Encoding considerations:">
3660      only "7bit", "8bit", or "binary" are permitted
3661    </t>
3662    <t hangText="Security considerations:">
3663      none
3664    </t>
3665    <t hangText="Interoperability considerations:">
3666      none
3667    </t>
3668    <t hangText="Published specification:">
3669      This specification (see <xref target=""/>).
3670    </t>
3671    <t hangText="Applications that use this media type:">
3672    </t>
3673    <t hangText="Additional information:">
3674      <list style="hanging">
3675        <t hangText="Magic number(s):">none</t>
3676        <t hangText="File extension(s):">none</t>
3677        <t hangText="Macintosh file type code(s):">none</t>
3678      </list>
3679    </t>
3680    <t hangText="Person and email address to contact for further information:">
3681      See Authors Section.
3682    </t>
3683                <t hangText="Intended usage:">
3684                  COMMON
3685    </t>
3686                <t hangText="Restrictions on usage:">
3687                  none
3688    </t>
3689    <t hangText="Author/Change controller:">
3690      IESG
3691    </t>
3692  </list>
3695<section title="Internet Media Type application/http" anchor="">
3696<iref item="Media Type" subitem="application/http" primary="true"/>
3697<iref item="application/http Media Type" primary="true"/>
3699   The application/http type can be used to enclose a pipeline of one or more
3700   HTTP request or response messages (not intermixed).
3703  <list style="hanging">
3704    <t hangText="Type name:">
3705      application
3706    </t>
3707    <t hangText="Subtype name:">
3708      http
3709    </t>
3710    <t hangText="Required parameters:">
3711      none
3712    </t>
3713    <t hangText="Optional parameters:">
3714      version, msgtype
3715      <list style="hanging">
3716        <t hangText="version:">
3717          The HTTP-Version number of the enclosed messages
3718          (e.g., "1.1"). If not present, the version can be
3719          determined from the first line of the body.
3720        </t>
3721        <t hangText="msgtype:">
3722          The message type -- "request" or "response". If not
3723          present, the type can be determined from the first
3724          line of the body.
3725        </t>
3726      </list>
3727    </t>
3728    <t hangText="Encoding considerations:">
3729      HTTP messages enclosed by this type
3730      are in "binary" format; use of an appropriate
3731      Content-Transfer-Encoding is required when
3732      transmitted via E-mail.
3733    </t>
3734    <t hangText="Security considerations:">
3735      none
3736    </t>
3737    <t hangText="Interoperability considerations:">
3738      none
3739    </t>
3740    <t hangText="Published specification:">
3741      This specification (see <xref target=""/>).
3742    </t>
3743    <t hangText="Applications that use this media type:">
3744    </t>
3745    <t hangText="Additional information:">
3746      <list style="hanging">
3747        <t hangText="Magic number(s):">none</t>
3748        <t hangText="File extension(s):">none</t>
3749        <t hangText="Macintosh file type code(s):">none</t>
3750      </list>
3751    </t>
3752    <t hangText="Person and email address to contact for further information:">
3753      See Authors Section.
3754    </t>
3755                <t hangText="Intended usage:">
3756                  COMMON
3757    </t>
3758                <t hangText="Restrictions on usage:">
3759                  none
3760    </t>
3761    <t hangText="Author/Change controller:">
3762      IESG
3763    </t>
3764  </list>
3769<section title="Tolerant Applications" anchor="tolerant.applications">
3771   Although this document specifies the requirements for the generation
3772   of HTTP/1.1 messages, not all applications will be correct in their
3773   implementation. We therefore recommend that operational applications
3774   be tolerant of deviations whenever those deviations can be
3775   interpreted unambiguously.
3778   Clients SHOULD be tolerant in parsing the Status-Line and servers
3779   tolerant when parsing the Request-Line. In particular, they SHOULD
3780   accept any amount of SP or HTAB characters between fields, even though
3781   only a single SP is required.
3784   The line terminator for message-header fields is the sequence CRLF.
3785   However, we recommend that applications, when parsing such headers,
3786   recognize a single LF as a line terminator and ignore the leading CR.
3789   The character set of an entity-body SHOULD be labeled as the lowest
3790   common denominator of the character codes used within that body, with
3791   the exception that not labeling the entity is preferred over labeling
3792   the entity with the labels US-ASCII or ISO-8859-1. See <xref target="Part3"/>.
3795   Additional rules for requirements on parsing and encoding of dates
3796   and other potential problems with date encodings include:
3799  <list style="symbols">
3800     <t>HTTP/1.1 clients and caches SHOULD assume that an RFC-850 date
3801        which appears to be more than 50 years in the future is in fact
3802        in the past (this helps solve the "year 2000" problem).</t>
3804     <t>An HTTP/1.1 implementation MAY internally represent a parsed
3805        Expires date as earlier than the proper value, but MUST NOT
3806        internally represent a parsed Expires date as later than the
3807        proper value.</t>
3809     <t>All expiration-related calculations MUST be done in GMT. The
3810        local time zone MUST NOT influence the calculation or comparison
3811        of an age or expiration time.</t>
3813     <t>If an HTTP header incorrectly carries a date value with a time
3814        zone other than GMT, it MUST be converted into GMT using the
3815        most conservative possible conversion.</t>
3816  </list>
3820<section title="Conversion of Date Formats" anchor="">
3822   HTTP/1.1 uses a restricted set of date formats (<xref target=""/>) to
3823   simplify the process of date comparison. Proxies and gateways from
3824   other protocols SHOULD ensure that any Date header field present in a
3825   message conforms to one of the HTTP/1.1 formats and rewrite the date
3826   if necessary.
3830<section title="Compatibility with Previous Versions" anchor="compatibility">
3832   It is beyond the scope of a protocol specification to mandate
3833   compliance with previous versions. HTTP/1.1 was deliberately
3834   designed, however, to make supporting previous versions easy. It is
3835   worth noting that, at the time of composing this specification
3836   (1996), we would expect commercial HTTP/1.1 servers to:
3837  <list style="symbols">
3838     <t>recognize the format of the Request-Line for HTTP/0.9, 1.0, and
3839        1.1 requests;</t>
3841     <t>understand any valid request in the format of HTTP/0.9, 1.0, or
3842        1.1;</t>
3844     <t>respond appropriately with a message in the same major version
3845        used by the client.</t>
3846  </list>
3849   And we would expect HTTP/1.1 clients to:
3850  <list style="symbols">
3851     <t>recognize the format of the Status-Line for HTTP/1.0 and 1.1
3852        responses;</t>
3854     <t>understand any valid response in the format of HTTP/0.9, 1.0, or
3855        1.1.</t>
3856  </list>
3859   For most implementations of HTTP/1.0, each connection is established
3860   by the client prior to the request and closed by the server after
3861   sending the response. Some implementations implement the Keep-Alive
3862   version of persistent connections described in Section 19.7.1 of <xref target="RFC2068"/>.
3865<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
3867   This section summarizes major differences between versions HTTP/1.0
3868   and HTTP/1.1.
3871<section title="Changes to Simplify Multi-homed Web Servers and Conserve IP Addresses" anchor="">
3873   The requirements that clients and servers support the Host request-header,
3874   report an error if the Host request-header (<xref target=""/>) is
3875   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-uri"/>)
3876   are among the most important changes defined by this
3877   specification.
3880   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
3881   addresses and servers; there was no other established mechanism for
3882   distinguishing the intended server of a request than the IP address
3883   to which that request was directed. The changes outlined above will
3884   allow the Internet, once older HTTP clients are no longer common, to
3885   support multiple Web sites from a single IP address, greatly
3886   simplifying large operational Web servers, where allocation of many
3887   IP addresses to a single host has created serious problems. The
3888   Internet will also be able to recover the IP addresses that have been
3889   allocated for the sole purpose of allowing special-purpose domain
3890   names to be used in root-level HTTP URLs. Given the rate of growth of
3891   the Web, and the number of servers already deployed, it is extremely
3892   important that all implementations of HTTP (including updates to
3893   existing HTTP/1.0 applications) correctly implement these
3894   requirements:
3895  <list style="symbols">
3896     <t>Both clients and servers MUST support the Host request-header.</t>
3898     <t>A client that sends an HTTP/1.1 request MUST send a Host header.</t>
3900     <t>Servers MUST report a 400 (Bad Request) error if an HTTP/1.1
3901        request does not include a Host request-header.</t>
3903     <t>Servers MUST accept absolute URIs.</t>
3904  </list>
3909<section title="Compatibility with HTTP/1.0 Persistent Connections" anchor="compatibility.with.http.1.0.persistent.connections">
3911   Some clients and servers might wish to be compatible with some
3912   previous implementations of persistent connections in HTTP/1.0
3913   clients and servers. Persistent connections in HTTP/1.0 are
3914   explicitly negotiated as they are not the default behavior. HTTP/1.0
3915   experimental implementations of persistent connections are faulty,
3916   and the new facilities in HTTP/1.1 are designed to rectify these
3917   problems. The problem was that some existing 1.0 clients may be
3918   sending Keep-Alive to a proxy server that doesn't understand
3919   Connection, which would then erroneously forward it to the next
3920   inbound server, which would establish the Keep-Alive connection and
3921   result in a hung HTTP/1.0 proxy waiting for the close on the
3922   response. The result is that HTTP/1.0 clients must be prevented from
3923   using Keep-Alive when talking to proxies.
3926   However, talking to proxies is the most important use of persistent
3927   connections, so that prohibition is clearly unacceptable. Therefore,
3928   we need some other mechanism for indicating a persistent connection
3929   is desired, which is safe to use even when talking to an old proxy
3930   that ignores Connection. Persistent connections are the default for
3931   HTTP/1.1 messages; we introduce a new keyword (Connection: close) for
3932   declaring non-persistence. See <xref target="header.connection"/>.
3935   The original HTTP/1.0 form of persistent connections (the Connection:
3936   Keep-Alive and Keep-Alive header) is documented in <xref target="RFC2068"/>.
3940<section title="Changes from RFC 2068" anchor="changes.from.rfc.2068">
3942   This specification has been carefully audited to correct and
3943   disambiguate key word usage; RFC 2068 had many problems in respect to
3944   the conventions laid out in <xref target="RFC2119"/>.
3947   Transfer-coding and message lengths all interact in ways that
3948   required fixing exactly when chunked encoding is used (to allow for
3949   transfer encoding that may not be self delimiting); it was important
3950   to straighten out exactly how message lengths are computed. (Sections
3951   <xref target="transfer.codings" format="counter"/>, <xref target="message.length" format="counter"/>,
3952   <xref target="header.content-length" format="counter"/>,
3953   see also <xref target="Part3"/>, <xref target="Part5"/> and <xref target="Part6"/>)
3956   The use and interpretation of HTTP version numbers has been clarified
3957   by <xref target="RFC2145"/>. Require proxies to upgrade requests to highest protocol
3958   version they support to deal with problems discovered in HTTP/1.0
3959   implementations (<xref target="http.version"/>)
3962   Transfer-coding had significant problems, particularly with
3963   interactions with chunked encoding. The solution is that transfer-codings
3964   become as full fledged as content-codings. This involves
3965   adding an IANA registry for transfer-codings (separate from content
3966   codings), a new header field (TE) and enabling trailer headers in the
3967   future. Transfer encoding is a major performance benefit, so it was
3968   worth fixing <xref target="Nie1997"/>. TE also solves another, obscure, downward
3969   interoperability problem that could have occurred due to interactions
3970   between authentication trailers, chunked encoding and HTTP/1.0
3971   clients.(Section <xref target="transfer.codings" format="counter"/>, <xref target="chunked.transfer.encoding" format="counter"/>,
3972   and <xref target="header.te" format="counter"/>)
3976<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
3978  Clarify that HTTP-Version is case sensitive.
3979  (<xref target="http.version"/>)
3982  Remove reference to non-existant identity transfer-coding value tokens.
3983  (Sections <xref format="counter" target="transfer.codings"/> and
3984  <xref format="counter" target="message.length"/>)
3987  Clarification that the chunk length does not include
3988  the count of the octets in the chunk header and trailer.
3989  (<xref target="chunked.transfer.encoding"/>)
3992  Fix BNF to add query, as the abs_path production in
3993  Section 3 of <xref target="RFC2396"/> doesn't define it.
3994  (<xref target="request-uri"/>)
3997  Clarify exactly when close connection options must be sent.
3998  (<xref target="header.connection"/>)
4003<section title="Change Log (to be removed by RFC Editor before publication)">
4005<section title="Since RFC2616">
4007  Extracted relevant partitions from <xref target="RFC2616"/>.
4011<section title="Since draft-ietf-httpbis-p1-messaging-00">
4013  Closed issues:
4014  <list style="symbols">
4015    <t>
4016      <eref target=""/>:
4017      "HTTP Version should be case sensitive"
4018      (<eref target=""/>)
4019    </t>
4020    <t>
4021      <eref target=""/>:
4022      "'unsafe' characters"
4023      (<eref target=""/>)
4024    </t>
4025    <t>
4026      <eref target=""/>:
4027      "Chunk Size Definition"
4028      (<eref target=""/>)
4029    </t>
4030    <t>
4031      <eref target=""/>:
4032      "Message Length"
4033      (<eref target=""/>)
4034    </t>
4035    <t>
4036      <eref target=""/>:
4037      "Media Type Registrations"
4038      (<eref target=""/>)
4039    </t>
4040    <t>
4041      <eref target=""/>:
4042      "URI includes query"
4043      (<eref target=""/>)
4044    </t>
4045    <t>
4046      <eref target=""/>:
4047      "No close on 1xx responses"
4048      (<eref target=""/>)
4049    </t>
4050    <t>
4051      <eref target=""/>:
4052      "Remove 'identity' token references"
4053      (<eref target=""/>)
4054    </t>
4055    <t>
4056      <eref target=""/>:
4057      "Import query BNF"
4058    </t>
4059    <t>
4060      <eref target=""/>:
4061      "qdtext BNF"
4062    </t>
4063    <t>
4064      <eref target=""/>:
4065      "Normative and Informative references"
4066    </t>
4067    <t>
4068      <eref target=""/>:
4069      "RFC2606 Compliance"
4070    </t>
4071    <t>
4072      <eref target=""/>:
4073      "RFC977 reference"
4074    </t>
4075    <t>
4076      <eref target=""/>:
4077      "RFC1700 references"
4078    </t>
4079    <t>
4080      <eref target=""/>:
4081      "inconsistency in date format explanation"
4082    </t>
4083    <t>
4084      <eref target=""/>:
4085      "Date reference typo"
4086    </t>
4087    <t>
4088      <eref target=""/>:
4089      "Informative references"
4090    </t>
4091    <t>
4092      <eref target=""/>:
4093      "ISO-8859-1 Reference"
4094    </t>
4095    <t>
4096      <eref target=""/>:
4097      "Normative up-to-date references"
4098    </t>
4099  </list>
4102  Other changes:
4103  <list style="symbols">
4104    <t>
4105      Update media type registrations to use RFC4288 template.
4106    </t>
4107    <t>
4108      Use names of RFC4234 core rules DQUOTE and HTAB,
4109      fix broken ABNF for chunk-data
4110      (work in progress on <eref target=""/>)
4111    </t>
4112  </list>
Note: See TracBrowser for help on using the repository browser.