source: draft-ietf-httpbis/latest/p1-messaging.xml @ 29

Last change on this file since 29 was 29, checked in by fielding@…, 12 years ago

Partition RFC 2616 into seven (mostly) independent documents.
No semantic changes. Some meaningless crossreferences to prior
editorial decisions have been removed from appendices.

Structural changes minimized to simplify diff versus rfc2616.
This was a lot harder than it looks.

Part 8 (Cookies) is for future specification based on RFC 2965.

  • Property svn:eol-style set to native
File size: 167.7 KB
Line 
1<?xml version="1.0" encoding="utf-8"?>
2<!DOCTYPE rfc [
3  <!ENTITY MAY "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>MAY</bcp14>">
4  <!ENTITY MUST "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>MUST</bcp14>">
5  <!ENTITY MUST-NOT "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>MUST NOT</bcp14>">
6  <!ENTITY OPTIONAL "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>OPTIONAL</bcp14>">
7  <!ENTITY RECOMMENDED "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>RECOMMENDED</bcp14>">
8  <!ENTITY REQUIRED "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>REQUIRED</bcp14>">
9  <!ENTITY SHALL "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHALL</bcp14>">
10  <!ENTITY SHALL-NOT "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHALL NOT</bcp14>">
11  <!ENTITY SHOULD "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHOULD</bcp14>">
12  <!ENTITY SHOULD-NOT "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHOULD NOT</bcp14>">
13  <!ENTITY ID-VERSION "latest">
14  <!ENTITY caching                "[Part 6]">
15  <!ENTITY payload                "[Part 3]">
16  <!ENTITY CONNECT                "[Part 2]">
17  <!ENTITY content.negotiation    "[Part 3]">
18  <!ENTITY diff2045entity         "[Part 3]">
19  <!ENTITY entity                 "[Part 3]">
20  <!ENTITY entity-header-fields   "[Part 3]">
21  <!ENTITY header-cache-control   "[Part 6]">
22  <!ENTITY header-expect          "[Part 2]">
23  <!ENTITY header-pragma          "[Part 6]">
24  <!ENTITY header-warning         "[Part 6]">
25  <!ENTITY idempotent-methods     "[Part 2]">
26  <!ENTITY qvalue                 "[Part 3]">
27  <!ENTITY request-header-fields  "[Part 2]">
28  <!ENTITY response-header-fields "[Part 2]">
29  <!ENTITY method                 "[Part 2]">
30  <!ENTITY status-codes           "[Part 2]">
31  <!ENTITY status-100             "[Part 2]">
32  <!ENTITY status-1xx             "[Part 2]">
33  <!ENTITY status-414             "[Part 2]">
34]>
35<?rfc toc="yes" ?>
36<?rfc symrefs="yes" ?>
37<?rfc sortrefs="yes" ?>
38<?rfc compact="yes"?>
39<?rfc subcompact="no" ?>
40<?rfc linkmailto="no" ?>
41<?rfc editing="no" ?>
42<?rfc-ext allow-markup-in-artwork="yes" ?>
43<?rfc-ext include-references-in-index="yes" ?>
44<rfc obsoletes="2068, 2616, 2617" category="std"
45     ipr="full3978" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
46     xmlns:x='http://purl.org/net/xml2rfc/ext' xmlns:ed="http://greenbytes.de/2002/rfcedit">
47<front>
48
49  <title abbrev="HTTP/1.1">HTTP/1.1, part 1: URIs, Connections, and Message Parsing</title>
50
51  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
52    <organization abbrev="Day Software">Day Software</organization>
53    <address>
54      <postal>
55        <street>23 Corporate Plaza DR, Suite 280</street>
56        <city>Newport Beach</city>
57        <region>CA</region>
58        <code>92660</code>
59        <country>USA</country>
60      </postal>
61      <phone>+1-949-706-5300</phone>
62      <facsimile>+1-949-706-5305</facsimile>
63      <email>fielding@gbiv.com</email>
64      <uri>http://roy.gbiv.com/</uri>
65    </address>
66  </author>
67
68  <author initials="J." surname="Gettys" fullname="Jim Gettys">
69    <organization>One Laptop per Child</organization>
70    <address>
71      <postal>
72        <street>21 Oak Knoll Road</street>
73        <city>Carlisle</city>
74        <region>MA</region>
75        <code>01741</code>
76        <country>USA</country>
77      </postal>
78      <email>jg@laptop.org</email>
79      <uri>http://www.laptop.org/</uri>
80    </address>
81  </author>
82 
83  <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
84    <organization abbrev="HP">Hewlett-Packard Company</organization>
85    <address>
86      <postal>
87        <street>HP Labs, Large Scale Systems Group</street>
88        <street>1501 Page Mill Road, MS 1177</street>
89        <city>Palo Alto</city>
90        <region>CA</region>
91        <code>94304</code>
92        <country>USA</country>
93      </postal>
94      <email>JeffMogul@acm.org</email>
95    </address>
96  </author>
97
98  <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
99    <organization abbrev="Microsoft">Microsoft Corporation</organization>
100    <address>
101      <postal>
102        <street>1 Microsoft Way</street>
103        <city>Redmond</city>
104        <region>WA</region>
105        <code>98052</code>
106        <country>USA</country>
107      </postal>
108      <email>henrikn@microsoft.com</email>
109    </address>
110  </author>
111
112  <author initials="L." surname="Masinter" fullname="Larry Masinter">
113    <organization abbrev="Adobe Systems">Adobe Systems, Incorporated</organization>
114    <address>
115      <postal>
116        <street>345 Park Ave</street>
117        <city>San Jose</city>
118        <region>CA</region>
119        <code>95110</code>
120        <country>USA</country>
121      </postal>
122      <email>LMM@acm.org</email>
123      <uri>http://larry.masinter.net/</uri>
124    </address>
125  </author>
126 
127  <author initials="P." surname="Leach" fullname="Paul J. Leach">
128    <organization abbrev="Microsoft">Microsoft Corporation</organization>
129    <address>
130      <postal>
131        <street>1 Microsoft Way</street>
132        <city>Redmond</city>
133        <region>WA</region>
134        <code>98052</code>
135      </postal>
136      <email>paulle@microsoft.com</email>
137    </address>
138  </author>
139   
140  <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
141    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
142    <address>
143      <postal>
144        <street>MIT Laboratory for Computer Science</street>
145        <street>545 Technology Square</street>
146        <city>Cambridge</city>
147        <region>MA</region>
148        <code>02139</code>
149        <country>USA</country>
150      </postal>
151      <facsimile>+1 (617) 258 8682</facsimile>
152      <email>timbl@w3.org</email>
153    </address>
154  </author>
155
156  <date month="December" year="2007"/>
157
158<abstract>
159<t>
160   The Hypertext Transfer Protocol (HTTP) is an application-level
161   protocol for distributed, collaborative, hypermedia information
162   systems. HTTP has been in use by the World Wide Web global information
163   initiative since 1990. This document is Part 1 of the eight-part specification
164   that defines the protocol referred to as "HTTP/1.1" and, taken together,
165   updates RFC 2616 and RFC 2617.  Part 1 provides an overview of HTTP and
166   its associated terminology, defines the "http" and "https" Uniform
167   Resource Identifier (URI) schemes, defines the generic message syntax
168   and parsing requirements for HTTP message frames, and describes
169   general security concerns for implementations.
170</t>
171</abstract>
172</front>
173<middle>
174<section title="Introduction" anchor="introduction">
175<t>
176   This document will define aspects of HTTP related to overall network
177   operation, message framing, interaction with transport protocols, and
178   URI schemes. Right now it only includes the extracted relevant sections
179   of <xref target="RFC2616"/> and <xref target="RFC2617"/>.
180</t>
181<section title="Purpose" anchor="intro.purpose">
182<t>
183   The Hypertext Transfer Protocol (HTTP) is an application-level
184   protocol for distributed, collaborative, hypermedia information
185   systems. HTTP has been in use by the World-Wide Web global
186   information initiative since 1990. The first version of HTTP,
187   referred to as HTTP/0.9, was a simple protocol for raw data transfer
188   across the Internet. HTTP/1.0, as defined by RFC 1945 <xref target="RFC1945"/>, improved
189   the protocol by allowing messages to be in the format of MIME-like
190   messages, containing metainformation about the data transferred and
191   modifiers on the request/response semantics. However, HTTP/1.0 does
192   not sufficiently take into consideration the effects of hierarchical
193   proxies, caching, the need for persistent connections, or virtual
194   hosts. In addition, the proliferation of incompletely-implemented
195   applications calling themselves "HTTP/1.0" has necessitated a
196   protocol version change in order for two communicating applications
197   to determine each other's true capabilities.
198</t>
199<t>
200   This specification defines the protocol referred to as "HTTP/1.1".
201   This protocol includes more stringent requirements than HTTP/1.0 in
202   order to ensure reliable implementation of its features.
203</t>
204<t>
205   Practical information systems require more functionality than simple
206   retrieval, including search, front-end update, and annotation. HTTP
207   allows an open-ended set of methods and headers that indicate the
208   purpose of a request <xref target="RFC2324"/>. It builds on the discipline of reference
209   provided by the Uniform Resource Identifier (URI) <xref target="RFC1630"/>, as a location
210   (URL) <xref target="RFC1738"/> or name (URN) <xref target="RFC1737"/>, for indicating the resource to which a
211   method is to be applied. Messages are passed in a format similar to
212   that used by Internet mail <xref target="RFC822"/> as defined by the Multipurpose
213   Internet Mail Extensions (MIME) <xref target="RFC2045"/>.
214</t>
215<t>
216   HTTP is also used as a generic protocol for communication between
217   user agents and proxies/gateways to other Internet systems, including
218   those supported by the SMTP <xref target="RFC821"/>, NNTP <xref target="RFC977"/>, FTP <xref target="RFC959"/>, Gopher <xref target="RFC1436"/>,
219   and WAIS <xref target="WAIS"/> protocols. In this way, HTTP allows basic hypermedia
220   access to resources available from diverse applications.
221</t>
222</section>
223
224<section title="Requirements" anchor="intro.requirements">
225<t>
226   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
227   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
228   document are to be interpreted as described in RFC 2119 <xref target="RFC2119"/>.
229</t>
230<t>
231   An implementation is not compliant if it fails to satisfy one or more
232   of the &MUST; or &REQUIRED; level requirements for the protocols it
233   implements. An implementation that satisfies all the &MUST; or &REQUIRED;
234   level and all the &SHOULD; level requirements for its protocols is said
235   to be "unconditionally compliant"; one that satisfies all the &MUST;
236   level requirements but not all the &SHOULD; level requirements for its
237   protocols is said to be "conditionally compliant."
238</t>
239</section>
240
241<section title="Terminology" anchor="intro.terminology">
242<t>
243   This specification uses a number of terms to refer to the roles
244   played by participants in, and objects of, the HTTP communication.
245</t>
246<t>
247  <iref item="connection"/>
248  <x:dfn>connection</x:dfn>
249  <list>
250    <t>
251      A transport layer virtual circuit established between two programs
252      for the purpose of communication.
253    </t>
254  </list>
255</t>
256<t>
257  <iref item="message"/>
258  <x:dfn>message</x:dfn>
259  <list>
260    <t>
261      The basic unit of HTTP communication, consisting of a structured
262      sequence of octets matching the syntax defined in <xref target="http.message"/> and
263      transmitted via the connection.
264    </t>
265  </list>
266</t>
267<t>
268  <iref item="request"/>
269  <x:dfn>request</x:dfn>
270  <list>
271    <t>
272      An HTTP request message, as defined in <xref target="request"/>.
273    </t>
274  </list>
275</t>
276<t>
277  <iref item="response"/>
278  <x:dfn>response</x:dfn>
279  <list>
280    <t>
281      An HTTP response message, as defined in <xref target="response"/>.
282    </t>
283  </list>
284</t>
285<t>
286  <iref item="resource"/>
287  <x:dfn>resource</x:dfn>
288  <list>
289    <t>
290      A network data object or service that can be identified by a URI,
291      as defined in <xref target="uri"/>. Resources may be available in multiple
292      representations (e.g. multiple languages, data formats, size, and
293      resolutions) or vary in other ways.
294    </t>
295  </list>
296</t>
297<t>
298  <iref item="entity"/>
299  <x:dfn>entity</x:dfn>
300  <list>
301    <t>
302      The information transferred as the payload of a request or
303      response. An entity consists of metainformation in the form of
304      entity-header fields and content in the form of an entity-body, as
305      described in &entity;.
306    </t>
307  </list>
308</t>
309<t>
310  <iref item="representation"/>
311  <x:dfn>representation</x:dfn>
312  <list>
313    <t>
314      An entity included with a response that is subject to content
315      negotiation, as described in &content.negotiation;. There may exist multiple
316      representations associated with a particular response status.
317    </t>
318  </list>
319</t>
320<t>
321  <iref item="content negotiation"/>
322  <x:dfn>content negotiation</x:dfn>
323  <list>
324    <t>
325      The mechanism for selecting the appropriate representation when
326      servicing a request, as described in &content.negotiation;. The
327      representation of entities in any response can be negotiated
328      (including error responses).
329    </t>
330  </list>
331</t>
332<t>
333  <iref item="variant"/>
334  <x:dfn>variant</x:dfn>
335  <list>
336    <t>
337      A resource may have one, or more than one, representation(s)
338      associated with it at any given instant. Each of these
339      representations is termed a `varriant'.  Use of the term `variant'
340      does not necessarily imply that the resource is subject to content
341      negotiation.
342    </t>
343  </list>
344</t>
345<t>
346  <iref item="client"/>
347  <x:dfn>client</x:dfn>
348  <list>
349    <t>
350      A program that establishes connections for the purpose of sending
351      requests.
352    </t>
353  </list>
354</t>
355<t>
356  <iref item="user agent"/>
357  <x:dfn>user agent</x:dfn>
358  <list>
359    <t>
360      The client which initiates a request. These are often browsers,
361      editors, spiders (web-traversing robots), or other end user tools.
362    </t>
363  </list>
364</t>
365<t>
366  <iref item="server"/>
367  <x:dfn>server</x:dfn>
368  <list>
369    <t>
370      An application program that accepts connections in order to
371      service requests by sending back responses. Any given program may
372      be capable of being both a client and a server; our use of these
373      terms refers only to the role being performed by the program for a
374      particular connection, rather than to the program's capabilities
375      in general. Likewise, any server may act as an origin server,
376      proxy, gateway, or tunnel, switching behavior based on the nature
377      of each request.
378    </t>
379  </list>
380</t>
381<t>
382  <iref item="origin server"/>
383  <x:dfn>origin server</x:dfn>
384  <list>
385    <t>
386      The server on which a given resource resides or is to be created.
387    </t>
388  </list>
389</t>
390<t>
391  <iref item="proxy"/>
392  <x:dfn>proxy</x:dfn>
393  <list>
394    <t>
395      An intermediary program which acts as both a server and a client
396      for the purpose of making requests on behalf of other clients.
397      Requests are serviced internally or by passing them on, with
398      possible translation, to other servers. A proxy &MUST; implement
399      both the client and server requirements of this specification. A
400      "transparent proxy" is a proxy that does not modify the request or
401      response beyond what is required for proxy authentication and
402      identification. A "non-transparent proxy" is a proxy that modifies
403      the request or response in order to provide some added service to
404      the user agent, such as group annotation services, media type
405      transformation, protocol reduction, or anonymity filtering. Except
406      where either transparent or non-transparent behavior is explicitly
407      stated, the HTTP proxy requirements apply to both types of
408      proxies.
409    </t>
410  </list>
411</t>
412<t>
413  <iref item="gateway"/>
414  <x:dfn>gateway</x:dfn>
415  <list>
416    <t>
417      A server which acts as an intermediary for some other server.
418      Unlike a proxy, a gateway receives requests as if it were the
419      origin server for the requested resource; the requesting client
420      may not be aware that it is communicating with a gateway.
421    </t>
422  </list>
423</t>
424<t>
425  <iref item="tunnel"/>
426  <x:dfn>tunnel</x:dfn>
427  <list>
428    <t>
429      An intermediary program which is acting as a blind relay between
430      two connections. Once active, a tunnel is not considered a party
431      to the HTTP communication, though the tunnel may have been
432      initiated by an HTTP request. The tunnel ceases to exist when both
433      ends of the relayed connections are closed.
434    </t>
435  </list>
436</t>
437<t>
438  <iref item="cache"/>
439  <x:dfn>cache</x:dfn>
440  <list>
441    <t>
442      A program's local store of response messages and the subsystem
443      that controls its message storage, retrieval, and deletion. A
444      cache stores cacheable responses in order to reduce the response
445      time and network bandwidth consumption on future, equivalent
446      requests. Any client or server may include a cache, though a cache
447      cannot be used by a server that is acting as a tunnel.
448    </t>
449  </list>
450</t>
451<t>
452  <iref item="cacheable"/>
453  <x:dfn>cacheable</x:dfn>
454  <list>
455    <t>
456      A response is cacheable if a cache is allowed to store a copy of
457      the response message for use in answering subsequent requests. The
458      rules for determining the cacheability of HTTP responses are
459      defined in &caching;. Even if a resource is cacheable, there may
460      be additional constraints on whether a cache can use the cached
461      copy for a particular request.
462    </t>
463  </list>
464</t>
465<t>
466  <iref item="upstream"/>
467  <iref item="downstream"/>
468  <x:dfn>upstream</x:dfn>/<x:dfn>downstream</x:dfn>
469  <list>
470    <t>
471      Upstream and downstream describe the flow of a message: all
472      messages flow from upstream to downstream.
473    </t>
474  </list>
475</t>
476<t>
477  <iref item="inbound"/>
478  <iref item="outbound"/>
479  <x:dfn>inbound</x:dfn>/<x:dfn>outbound</x:dfn>
480  <list>
481    <t>
482      Inbound and outbound refer to the request and response paths for
483      messages: "inbound" means "traveling toward the origin server",
484      and "outbound" means "traveling toward the user agent"
485    </t>
486  </list>
487</t>
488</section>
489
490<section title="Overall Operation" anchor="intro.overall.operation">
491<t>
492   The HTTP protocol is a request/response protocol. A client sends a
493   request to the server in the form of a request method, URI, and
494   protocol version, followed by a MIME-like message containing request
495   modifiers, client information, and possible body content over a
496   connection with a server. The server responds with a status line,
497   including the message's protocol version and a success or error code,
498   followed by a MIME-like message containing server information, entity
499   metainformation, and possible entity-body content. The relationship
500   between HTTP and MIME is described in &diff2045entity;.
501</t>
502<t>
503   Most HTTP communication is initiated by a user agent and consists of
504   a request to be applied to a resource on some origin server. In the
505   simplest case, this may be accomplished via a single connection (v)
506   between the user agent (UA) and the origin server (O).
507</t>
508<figure><artwork type="drawing">
509       request chain ------------------------&gt;
510    UA -------------------v------------------- O
511       &lt;----------------------- response chain
512</artwork></figure>
513<t>
514   A more complicated situation occurs when one or more intermediaries
515   are present in the request/response chain. There are three common
516   forms of intermediary: proxy, gateway, and tunnel. A proxy is a
517   forwarding agent, receiving requests for a URI in its absolute form,
518   rewriting all or part of the message, and forwarding the reformatted
519   request toward the server identified by the URI. A gateway is a
520   receiving agent, acting as a layer above some other server(s) and, if
521   necessary, translating the requests to the underlying server's
522   protocol. A tunnel acts as a relay point between two connections
523   without changing the messages; tunnels are used when the
524   communication needs to pass through an intermediary (such as a
525   firewall) even when the intermediary cannot understand the contents
526   of the messages.
527</t>
528<figure><artwork type="drawing">
529       request chain --------------------------------------&gt;
530    UA -----v----- A -----v----- B -----v----- C -----v----- O
531       &lt;------------------------------------- response chain
532</artwork></figure>
533<t>
534   The figure above shows three intermediaries (A, B, and C) between the
535   user agent and origin server. A request or response message that
536   travels the whole chain will pass through four separate connections.
537   This distinction is important because some HTTP communication options
538   may apply only to the connection with the nearest, non-tunnel
539   neighbor, only to the end-points of the chain, or to all connections
540   along the chain. Although the diagram is linear, each participant may
541   be engaged in multiple, simultaneous communications. For example, B
542   may be receiving requests from many clients other than A, and/or
543   forwarding requests to servers other than C, at the same time that it
544   is handling A's request.
545</t>
546<t>
547   Any party to the communication which is not acting as a tunnel may
548   employ an internal cache for handling requests. The effect of a cache
549   is that the request/response chain is shortened if one of the
550   participants along the chain has a cached response applicable to that
551   request. The following illustrates the resulting chain if B has a
552   cached copy of an earlier response from O (via C) for a request which
553   has not been cached by UA or A.
554</t>
555<figure><artwork type="drawing">
556          request chain ----------&gt;
557       UA -----v----- A -----v----- B - - - - - - C - - - - - - O
558          &lt;--------- response chain
559</artwork></figure>
560<t>
561   Not all responses are usefully cacheable, and some requests may
562   contain modifiers which place special requirements on cache behavior.
563   HTTP requirements for cache behavior and cacheable responses are
564   defined in &caching;.
565</t>
566<t>
567   In fact, there are a wide variety of architectures and configurations
568   of caches and proxies currently being experimented with or deployed
569   across the World Wide Web. These systems include national hierarchies
570   of proxy caches to save transoceanic bandwidth, systems that
571   broadcast or multicast cache entries, organizations that distribute
572   subsets of cached data via CD-ROM, and so on. HTTP systems are used
573   in corporate intranets over high-bandwidth links, and for access via
574   PDAs with low-power radio links and intermittent connectivity. The
575   goal of HTTP/1.1 is to support the wide diversity of configurations
576   already deployed while introducing protocol constructs that meet the
577   needs of those who build web applications that require high
578   reliability and, failing that, at least reliable indications of
579   failure.
580</t>
581<t>
582   HTTP communication usually takes place over TCP/IP connections. The
583   default port is TCP 80 <xref target="RFC1700"/>, but other ports can be used. This does
584   not preclude HTTP from being implemented on top of any other protocol
585   on the Internet, or on other networks. HTTP only presumes a reliable
586   transport; any protocol that provides such guarantees can be used;
587   the mapping of the HTTP/1.1 request and response structures onto the
588   transport data units of the protocol in question is outside the scope
589   of this specification.
590</t>
591<t>
592   In HTTP/1.0, most implementations used a new connection for each
593   request/response exchange. In HTTP/1.1, a connection may be used for
594   one or more request/response exchanges, although connections may be
595   closed for a variety of reasons (see <xref target="persistent.connections"/>).
596</t>
597</section>
598</section>
599
600<section title="Notational Conventions and Generic Grammar" anchor="notation">
601
602<section title="Augmented BNF" anchor="notation.abnf">
603<t>
604   All of the mechanisms specified in this document are described in
605   both prose and an augmented Backus-Naur Form (BNF) similar to that
606   used by RFC 822 <xref target="RFC822"/>. Implementors will need to be familiar with the
607   notation in order to understand this specification. The augmented BNF
608   includes the following constructs:
609</t>
610<t>
611   name = definition
612  <list>
613    <t>
614      The name of a rule is simply the name itself (without any
615      enclosing "&lt;" and "&gt;") and is separated from its definition by the
616      equal "=" character. White space is only significant in that
617      indentation of continuation lines is used to indicate a rule
618      definition that spans more than one line. Certain basic rules are
619      in uppercase, such as SP, LWS, HT, CRLF, DIGIT, ALPHA, etc. Angle
620      brackets are used within definitions whenever their presence will
621      facilitate discerning the use of rule names.
622    </t>
623  </list>
624</t>
625<t>
626   "literal"
627  <list>
628    <t>
629      Quotation marks surround literal text. Unless stated otherwise,
630      the text is case-insensitive.
631    </t>
632  </list>
633</t>
634<t>
635   rule1 | rule2
636  <list>
637    <t>
638      Elements separated by a bar ("|") are alternatives, e.g., "yes |
639      no" will accept yes or no.
640    </t>
641  </list>
642</t>
643<t>
644   (rule1 rule2)
645  <list>
646    <t>
647      Elements enclosed in parentheses are treated as a single element.
648      Thus, "(elem (foo | bar) elem)" allows the token sequences "elem
649      foo elem" and "elem bar elem".
650    </t>
651  </list>
652</t>
653<t>
654   *rule
655  <list>
656    <t>
657      The character "*" preceding an element indicates repetition. The
658      full form is "&lt;n&gt;*&lt;m&gt;element" indicating at least &lt;n&gt; and at most
659      &lt;m&gt; occurrences of element. Default values are 0 and infinity so
660      that "*(element)" allows any number, including zero; "1*element"
661      requires at least one; and "1*2element" allows one or two.
662    </t>
663  </list>
664</t>
665<t>
666   [rule]
667  <list>
668    <t>
669      Square brackets enclose optional elements; "[foo bar]" is
670      equivalent to "*1(foo bar)".
671    </t>
672  </list>
673</t>
674<t>
675   N rule
676  <list>
677    <t>
678      Specific repetition: "&lt;n&gt;(element)" is equivalent to
679      "&lt;n&gt;*&lt;n&gt;(element)"; that is, exactly &lt;n&gt; occurrences of (element).
680      Thus 2DIGIT is a 2-digit number, and 3ALPHA is a string of three
681      alphabetic characters.
682    </t>
683  </list>
684</t>
685<t>
686   #rule
687  <list>
688    <t>
689      A construct "#" is defined, similar to "*", for defining lists of
690      elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating at least
691      &lt;n&gt; and at most &lt;m&gt; elements, each separated by one or more commas
692      (",") and &OPTIONAL; linear white space (LWS). This makes the usual
693      form of lists very easy; a rule such as
694    </t>
695    <t>
696         ( *LWS element *( *LWS "," *LWS element ))
697    </t>
698    <t>
699      can be shown as
700    </t>
701    <t>
702         1#element
703    </t>
704    <t>
705      Wherever this construct is used, null elements are allowed, but do
706      not contribute to the count of elements present. That is,
707      "(element), , (element) " is permitted, but counts as only two
708      elements. Therefore, where at least one element is required, at
709      least one non-null element &MUST; be present. Default values are 0
710      and infinity so that "#element" allows any number, including zero;
711      "1#element" requires at least one; and "1#2element" allows one or
712      two.
713    </t>
714  </list>
715</t>
716<t>
717   ; comment
718  <list>
719    <t>
720      A semi-colon, set off some distance to the right of rule text,
721      starts a comment that continues to the end of line. This is a
722      simple way of including useful notes in parallel with the
723      specifications.
724    </t>
725  </list>
726</t>
727<t>
728   implied *LWS
729  <list>
730    <t>
731      The grammar described by this specification is word-based. Except
732      where noted otherwise, linear white space (LWS) can be included
733      between any two adjacent words (token or quoted-string), and
734      between adjacent words and separators, without changing the
735      interpretation of a field. At least one delimiter (LWS and/or
736      separators) &MUST; exist between any two tokens (for the definition
737      of "token" below), since they would otherwise be interpreted as a
738      single token.
739    </t>
740  </list>
741</t>
742</section>
743
744<section title="Basic Rules" anchor="basic.rules">
745<x:anchor-alias value="OCTET"/>
746<x:anchor-alias value="CHAR"/>
747<x:anchor-alias value="UPALPHA"/>
748<x:anchor-alias value="LOALPHA"/>
749<x:anchor-alias value="ALPHA"/>
750<x:anchor-alias value="DIGIT"/>
751<x:anchor-alias value="CTL"/>
752<x:anchor-alias value="CR"/>
753<x:anchor-alias value="LF"/>
754<x:anchor-alias value="SP"/>
755<x:anchor-alias value="HT"/>
756<x:anchor-alias value="CRLF"/>
757<x:anchor-alias value="LWS"/>
758<x:anchor-alias value="TEXT"/>
759<x:anchor-alias value="HEX"/>
760<x:anchor-alias value="token"/>
761<x:anchor-alias value="separators"/>
762<x:anchor-alias value="comment"/>
763<x:anchor-alias value="ctext"/>
764<x:anchor-alias value="quoted-string"/>
765<x:anchor-alias value="qdtext"/>
766<x:anchor-alias value="quoted-pair"/>
767<t>
768   The following rules are used throughout this specification to
769   describe basic parsing constructs. The US-ASCII coded character set
770   is defined by ANSI X3.4-1986 <xref target="USASCII"/>.
771</t>
772<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="OCTET"/><iref primary="true" item="Grammar" subitem="CHAR"/><iref primary="true" item="Grammar" subitem="UPALPHA"/><iref primary="true" item="Grammar" subitem="LOALPHA"/><iref primary="true" item="Grammar" subitem="ALPHA"/><iref primary="true" item="Grammar" subitem="DIGIT"/><iref primary="true" item="Grammar" subitem="CTL"/><iref primary="true" item="Grammar" subitem="CR"/><iref primary="true" item="Grammar" subitem="LF"/><iref primary="true" item="Grammar" subitem="SP"/><iref primary="true" item="Grammar" subitem="HT"/>
773    OCTET          = &lt;any 8-bit sequence of data&gt;
774    CHAR           = &lt;any US-ASCII character (octets 0 - 127)&gt;
775    UPALPHA        = &lt;any US-ASCII uppercase letter "A".."Z"&gt;
776    LOALPHA        = &lt;any US-ASCII lowercase letter "a".."z"&gt;
777    ALPHA          = UPALPHA | LOALPHA
778    DIGIT          = &lt;any US-ASCII digit "0".."9"&gt;
779    CTL            = &lt;any US-ASCII control character
780                     (octets 0 - 31) and DEL (127)&gt;
781    CR             = &lt;US-ASCII CR, carriage return (13)&gt;
782    LF             = &lt;US-ASCII LF, linefeed (10)&gt;
783    SP             = &lt;US-ASCII SP, space (32)&gt;
784    HT             = &lt;US-ASCII HT, horizontal-tab (9)&gt;
785    &lt;"&gt;            = &lt;US-ASCII double-quote mark (34)&gt;
786</artwork></figure>
787<t>
788   HTTP/1.1 defines the sequence CR LF as the end-of-line marker for all
789   protocol elements except the entity-body (see <xref target="tolerant.applications"/> for
790   tolerant applications). The end-of-line marker within an entity-body
791   is defined by its associated media type, as described in &payload;.
792</t>
793<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="CRLF"/>
794    CRLF           = CR LF
795</artwork></figure>
796<t>
797   HTTP/1.1 header field values can be folded onto multiple lines if the
798   continuation line begins with a space or horizontal tab. All linear
799   white space, including folding, has the same semantics as SP. A
800   recipient &MAY; replace any linear white space with a single SP before
801   interpreting the field value or forwarding the message downstream.
802</t>
803<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="LWS"/>
804    LWS            = [CRLF] 1*( SP | HT )
805</artwork></figure>
806<t>
807   The TEXT rule is only used for descriptive field contents and values
808   that are not intended to be interpreted by the message parser. Words
809   of *TEXT &MAY; contain characters from character sets other than ISO-8859-1
810   <xref target="ISO-8859"/> only when encoded according to the rules of RFC 2047
811   <xref target="RFC2047"/>.
812</t>
813<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TEXT"/>
814    TEXT           = &lt;any OCTET except CTLs,
815                     but including LWS&gt;
816</artwork></figure>
817<t>
818   A CRLF is allowed in the definition of TEXT only as part of a header
819   field continuation. It is expected that the folding LWS will be
820   replaced with a single SP before interpretation of the TEXT value.
821</t>
822<t>
823   Hexadecimal numeric characters are used in several protocol elements.
824</t>
825<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HEX"/>
826    HEX            = "A" | "B" | "C" | "D" | "E" | "F"
827                   | "a" | "b" | "c" | "d" | "e" | "f" | DIGIT
828</artwork></figure>
829<t>
830   Many HTTP/1.1 header field values consist of words separated by LWS
831   or special characters. These special characters &MUST; be in a quoted
832   string to be used within a parameter value (as defined in
833   <xref target="transfer.codings"/>).
834</t>
835<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="separators"/>
836    token          = 1*&lt;any CHAR except CTLs or separators&gt;
837    separators     = "(" | ")" | "&lt;" | "&gt;" | "@"
838                   | "," | ";" | ":" | "\" | &lt;"&gt;
839                   | "/" | "[" | "]" | "?" | "="
840                   | "{" | "}" | SP | HT
841</artwork></figure>
842<t>
843   Comments can be included in some HTTP header fields by surrounding
844   the comment text with parentheses. Comments are only allowed in
845   fields containing "comment" as part of their field value definition.
846   In all other fields, parentheses are considered part of the field
847   value.
848</t>
849<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
850    comment        = "(" *( ctext | quoted-pair | comment ) ")"
851    ctext          = &lt;any TEXT excluding "(" and ")"&gt;
852</artwork></figure>
853<t>
854   A string of text is parsed as a single word if it is quoted using
855   double-quote marks.
856</t>
857<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-string"/><iref primary="true" item="Grammar" subitem="qdtext"/>
858    quoted-string  = ( &lt;"&gt; *(qdtext | quoted-pair ) &lt;"&gt; )
859    qdtext         = &lt;any TEXT except &lt;"&gt;&gt;
860</artwork></figure>
861<t>
862   The backslash character ("\") &MAY; be used as a single-character
863   quoting mechanism only within quoted-string and comment constructs.
864</t>
865<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
866    quoted-pair    = "\" CHAR
867</artwork></figure>
868</section>
869</section>
870
871<section title="Protocol Parameters" anchor="protocol.parameters">
872
873<section title="HTTP Version" anchor="http.version">
874<t>
875   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate versions
876   of the protocol. The protocol versioning policy is intended to allow
877   the sender to indicate the format of a message and its capacity for
878   understanding further HTTP communication, rather than the features
879   obtained via that communication. No change is made to the version
880   number for the addition of message components which do not affect
881   communication behavior or which only add to extensible field values.
882   The &lt;minor&gt; number is incremented when the changes made to the
883   protocol add features which do not change the general message parsing
884   algorithm, but which may add to the message semantics and imply
885   additional capabilities of the sender. The &lt;major&gt; number is
886   incremented when the format of a message within the protocol is
887   changed. See RFC 2145 <xref target="RFC2145"/> for a fuller explanation.
888</t>
889<t>
890   The version of an HTTP message is indicated by an HTTP-Version field
891   in the first line of the message.
892</t>
893<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-Version"/>
894       HTTP-Version   = "HTTP" "/" 1*DIGIT "." 1*DIGIT
895</artwork></figure>
896<t>
897   Note that the major and minor numbers &MUST; be treated as separate
898   integers and that each &MAY; be incremented higher than a single digit.
899   Thus, HTTP/2.4 is a lower version than HTTP/2.13, which in turn is
900   lower than HTTP/12.3. Leading zeros &MUST; be ignored by recipients and
901   &MUST-NOT; be sent.
902</t>
903<t>
904   An application that sends a request or response message that includes
905   HTTP-Version of "HTTP/1.1" &MUST; be at least conditionally compliant
906   with this specification. Applications that are at least conditionally
907   compliant with this specification &SHOULD; use an HTTP-Version of
908   "HTTP/1.1" in their messages, and &MUST; do so for any message that is
909   not compatible with HTTP/1.0. For more details on when to send
910   specific HTTP-Version values, see RFC 2145 <xref target="RFC2145"/>.
911</t>
912<t>
913   The HTTP version of an application is the highest HTTP version for
914   which the application is at least conditionally compliant.
915</t>
916<t>
917   Proxy and gateway applications need to be careful when forwarding
918   messages in protocol versions different from that of the application.
919   Since the protocol version indicates the protocol capability of the
920   sender, a proxy/gateway &MUST-NOT; send a message with a version
921   indicator which is greater than its actual version. If a higher
922   version request is received, the proxy/gateway &MUST; either downgrade
923   the request version, or respond with an error, or switch to tunnel
924   behavior.
925</t>
926<t>
927   Due to interoperability problems with HTTP/1.0 proxies discovered
928   since the publication of RFC 2068 <xref target="RFC2068"/>, caching proxies MUST, gateways
929   &MAY;, and tunnels &MUST-NOT; upgrade the request to the highest version
930   they support. The proxy/gateway's response to that request &MUST; be in
931   the same major version as the request.
932</t>
933<t>
934  <list>
935    <t>
936      <x:h>Note:</x:h> Converting between versions of HTTP may involve modification
937      of header fields required or forbidden by the versions involved.
938    </t>
939  </list>
940</t>
941</section>
942
943<section title="Uniform Resource Identifiers" anchor="uri">
944<t>
945   URIs have been known by many names: WWW addresses, Universal Document
946   Identifiers, Universal Resource Identifiers <xref target="RFC1630"/>, and finally the
947   combination of Uniform Resource Locators (URL) <xref target="RFC1738"/> and Names (URN)
948   <xref target="RFC1737"/>. As far as HTTP is concerned, Uniform Resource Identifiers are
949   simply formatted strings which identify--via name, location, or any
950   other characteristic--a resource.
951</t>
952
953<section title="General Syntax" anchor="general.syntax">
954<t>
955   URIs in HTTP can be represented in absolute form or relative to some
956   known base URI <xref target="RFC1808"/>, depending upon the context of their use. The two
957   forms are differentiated by the fact that absolute URIs always begin
958   with a scheme name followed by a colon. For definitive information on
959   URL syntax and semantics, see "Uniform Resource Identifiers (URI):
960   Generic Syntax and Semantics," RFC 2396 <xref target="RFC2396"/> (which replaces RFCs
961   1738 <xref target="RFC1738"/> and RFC 1808 <xref target="RFC1808"/>). This specification adopts the
962   definitions of "URI-reference", "absoluteURI", "relativeURI", "port",
963   "host","abs_path", "rel_path", and "authority" from that
964   specification.
965</t>
966<t>
967   The HTTP protocol does not place any a priori limit on the length of
968   a URI. Servers &MUST; be able to handle the URI of any resource they
969   serve, and &SHOULD; be able to handle URIs of unbounded length if they
970   provide GET-based forms that could generate such URIs. A server
971   &SHOULD; return 414 (Request-URI Too Long) status if a URI is longer
972   than the server can handle (see &status-414;).
973</t>
974<t>
975  <list>
976    <t>
977      <x:h>Note:</x:h> Servers ought to be cautious about depending on URI lengths
978      above 255 bytes, because some older client or proxy
979      implementations might not properly support these lengths.
980    </t>
981  </list>
982</t>
983</section>
984
985<section title="http URL" anchor="http.url">
986<t>
987   The "http" scheme is used to locate network resources via the HTTP
988   protocol. This section defines the scheme-specific syntax and
989   semantics for http URLs.
990</t>
991<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http_URL"/>
992http_URL = "http:" "//" host [ ":" port ] [ abs_path [ "?" query ]]
993</artwork></figure>
994<t>
995   If the port is empty or not given, port 80 is assumed. The semantics
996   are that the identified resource is located at the server listening
997   for TCP connections on that port of that host, and the Request-URI
998   for the resource is abs_path (<xref target="request-uri"/>). The use of IP addresses
999   in URLs &SHOULD; be avoided whenever possible (see RFC 1900 <xref target="RFC1900"/>). If
1000   the abs_path is not present in the URL, it &MUST; be given as "/" when
1001   used as a Request-URI for a resource (<xref target="request-uri"/>). If a proxy
1002
1003   receives a host name which is not a fully qualified domain name, it
1004   &MAY; add its domain to the host name it received. If a proxy receives
1005   a fully qualified domain name, the proxy &MUST-NOT; change the host
1006   name.
1007</t>
1008</section>
1009
1010<section title="URI Comparison" anchor="uri.comparison">
1011<t>
1012   When comparing two URIs to decide if they match or not, a client
1013   &SHOULD; use a case-sensitive octet-by-octet comparison of the entire
1014   URIs, with these exceptions:
1015  <list style="symbols">
1016    <t>A port that is empty or not given is equivalent to the default
1017        port for that URI-reference;</t>
1018    <t>Comparisons of host names &MUST; be case-insensitive;</t>
1019    <t>Comparisons of scheme names &MUST; be case-insensitive;</t>
1020    <t>An empty abs_path is equivalent to an abs_path of "/".</t>
1021  </list>
1022</t>
1023<t>
1024   Characters other than those in the "reserved" and "unsafe" sets (see
1025   RFC 2396 <xref target="RFC2396"/>) are equivalent to their ""%" HEX HEX" encoding.
1026</t>
1027<t>
1028   For example, the following three URIs are equivalent:
1029</t>
1030<figure><artwork type="example">
1031   http://abc.com:80/~smith/home.html
1032   http://ABC.com/%7Esmith/home.html
1033   http://ABC.com:/%7esmith/home.html
1034</artwork></figure>
1035</section>
1036</section>
1037
1038<section title="Date/Time Formats" anchor="date.time.formats">
1039<section title="Full Date" anchor="full.date">
1040<t>
1041   HTTP applications have historically allowed three different formats
1042   for the representation of date/time stamps:
1043</t>
1044<figure><artwork type="example">
1045   Sun, 06 Nov 1994 08:49:37 GMT  ; RFC 822, updated by RFC 1123
1046   Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
1047   Sun Nov  6 08:49:37 1994       ; ANSI C's asctime() format
1048</artwork></figure>
1049<t>
1050   The first format is preferred as an Internet standard and represents
1051   a fixed-length subset of that defined by RFC 1123 <xref target="RFC1123"/> (an update to
1052   RFC 822 <xref target="RFC822"/>). The second format is in common use, but is based on the
1053   obsolete RFC 850 <xref target="RFC1036"/> date format and lacks a four-digit year.
1054   HTTP/1.1 clients and servers that parse the date value &MUST; accept
1055   all three formats (for compatibility with HTTP/1.0), though they &MUST;
1056   only generate the RFC 1123 format for representing HTTP-date values
1057   in header fields. See <xref target="tolerant.applications"/> for further information.
1058</t>
1059<t><list><t>
1060      <x:h>Note:</x:h> Recipients of date values are encouraged to be robust in
1061      accepting date values that may have been sent by non-HTTP
1062      applications, as is sometimes the case when retrieving or posting
1063      messages via proxies/gateways to SMTP or NNTP.
1064</t></list></t>
1065<t>
1066   All HTTP date/time stamps &MUST; be represented in Greenwich Mean Time
1067   (GMT), without exception. For the purposes of HTTP, GMT is exactly
1068   equal to UTC (Coordinated Universal Time). This is indicated in the
1069   first two formats by the inclusion of "GMT" as the three-letter
1070   abbreviation for time zone, and &MUST; be assumed when reading the
1071   asctime format. HTTP-date is case sensitive and &MUST-NOT; include
1072   additional LWS beyond that specifically included as SP in the
1073   grammar.
1074</t>
1075<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-date"/><iref primary="true" item="Grammar" subitem="rfc1123-date"/><iref primary="true" item="Grammar" subitem="rfc850-date"/><iref primary="true" item="Grammar" subitem="asctime-date"/><iref primary="true" item="Grammar" subitem="date1"/><iref primary="true" item="Grammar" subitem="date2"/><iref primary="true" item="Grammar" subitem="date3"/><iref primary="true" item="Grammar" subitem="time"/><iref primary="true" item="Grammar" subitem="wkday"/><iref primary="true" item="Grammar" subitem="weekday"/><iref primary="true" item="Grammar" subitem="month"/>
1076    HTTP-date    = rfc1123-date | rfc850-date | asctime-date
1077    rfc1123-date = wkday "," SP date1 SP time SP "GMT"
1078    rfc850-date  = weekday "," SP date2 SP time SP "GMT"
1079    asctime-date = wkday SP date3 SP time SP 4DIGIT
1080    date1        = 2DIGIT SP month SP 4DIGIT
1081                   ; day month year (e.g., 02 Jun 1982)
1082    date2        = 2DIGIT "-" month "-" 2DIGIT
1083                   ; day-month-year (e.g., 02-Jun-82)
1084    date3        = month SP ( 2DIGIT | ( SP 1DIGIT ))
1085                   ; month day (e.g., Jun  2)
1086    time         = 2DIGIT ":" 2DIGIT ":" 2DIGIT
1087                   ; 00:00:00 - 23:59:59
1088    wkday        = "Mon" | "Tue" | "Wed"
1089                 | "Thu" | "Fri" | "Sat" | "Sun"
1090    weekday      = "Monday" | "Tuesday" | "Wednesday"
1091                 | "Thursday" | "Friday" | "Saturday" | "Sunday"
1092    month        = "Jan" | "Feb" | "Mar" | "Apr"
1093                 | "May" | "Jun" | "Jul" | "Aug"
1094                 | "Sep" | "Oct" | "Nov" | "Dec"
1095</artwork></figure>
1096<t>
1097      <x:h>Note:</x:h> HTTP requirements for the date/time stamp format apply only
1098      to their usage within the protocol stream. Clients and servers are
1099      not required to use these formats for user presentation, request
1100      logging, etc.
1101</t>
1102</section>
1103</section>
1104
1105<section title="Transfer Codings" anchor="transfer.codings">
1106<t>
1107   Transfer-coding values are used to indicate an encoding
1108   transformation that has been, can be, or may need to be applied to an
1109   entity-body in order to ensure "safe transport" through the network.
1110   This differs from a content coding in that the transfer-coding is a
1111   property of the message, not of the original entity.
1112</t>
1113<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1114    transfer-coding         = "chunked" | transfer-extension
1115    transfer-extension      = token *( ";" parameter )
1116</artwork></figure>
1117<t>
1118   Parameters are in  the form of attribute/value pairs.
1119</t>
1120<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="parameter"/><iref primary="true" item="Grammar" subitem="attribute"/><iref primary="true" item="Grammar" subitem="value"/>
1121    parameter               = attribute "=" value
1122    attribute               = token
1123    value                   = token | quoted-string
1124</artwork></figure>
1125<t>
1126   All transfer-coding values are case-insensitive. HTTP/1.1 uses
1127   transfer-coding values in the TE header field (<xref target="header.te"/>) and in
1128   the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1129</t>
1130<t>
1131   Whenever a transfer-coding is applied to a message-body, the set of
1132   transfer-codings &MUST; include "chunked", unless the message is
1133   terminated by closing the connection. When the "chunked" transfer-coding
1134   is used, it &MUST; be the last transfer-coding applied to the
1135   message-body. The "chunked" transfer-coding &MUST-NOT; be applied more
1136   than once to a message-body. These rules allow the recipient to
1137   determine the transfer-length of the message (<xref target="message.length"/>).
1138</t>
1139<t>
1140   Transfer-codings are analogous to the Content-Transfer-Encoding
1141   values of MIME <xref target="RFC2045"/>, which were designed to enable safe transport of
1142   binary data over a 7-bit transport service. However, safe transport
1143   has a different focus for an 8bit-clean transfer protocol. In HTTP,
1144   the only unsafe characteristic of message-bodies is the difficulty in
1145   determining the exact body length (<xref target="message.length"/>), or the desire to
1146   encrypt data over a shared transport.
1147</t>
1148<t>
1149   The Internet Assigned Numbers Authority (IANA) acts as a registry for
1150   transfer-coding value tokens. Initially, the registry contains the
1151   following tokens: "chunked" (<xref target="chunked.transfer.encoding"/>), "identity" (section
1152   3.6.2), "gzip" (&payload;), "compress" (&payload;), and "deflate"
1153   (&payload;).
1154</t>
1155<t>
1156   New transfer-coding value tokens &SHOULD; be registered in the same way
1157   as new content-coding value tokens (&payload;).
1158</t>
1159<t>
1160   A server which receives an entity-body with a transfer-coding it does
1161   not understand &SHOULD; return 501 (Unimplemented), and close the
1162   connection. A server &MUST-NOT; send transfer-codings to an HTTP/1.0
1163   client.
1164</t>
1165
1166<section title="Chunked Transfer Coding" anchor="chunked.transfer.encoding">
1167<t>
1168   The chunked encoding modifies the body of a message in order to
1169   transfer it as a series of chunks, each with its own size indicator,
1170   followed by an &OPTIONAL; trailer containing entity-header fields. This
1171   allows dynamically produced content to be transferred along with the
1172   information necessary for the recipient to verify that it has
1173   received the full message.
1174</t>
1175<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Chunked-Body"/><iref primary="true" item="Grammar" subitem="chunk"/><iref primary="true" item="Grammar" subitem="chunk-size"/><iref primary="true" item="Grammar" subitem="last-chunk"/><iref primary="true" item="Grammar" subitem="chunk-extension"/><iref primary="true" item="Grammar" subitem="chunk-ext-name"/><iref primary="true" item="Grammar" subitem="chunk-ext-val"/><iref primary="true" item="Grammar" subitem="chunk-data"/><iref primary="true" item="Grammar" subitem="trailer"/>
1176    Chunked-Body   = *chunk
1177                     last-chunk
1178                     trailer
1179                     CRLF
1180
1181    chunk          = chunk-size [ chunk-extension ] CRLF
1182                     chunk-data CRLF
1183    chunk-size     = 1*HEX
1184    last-chunk     = 1*("0") [ chunk-extension ] CRLF
1185
1186    chunk-extension= *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
1187    chunk-ext-name = token
1188    chunk-ext-val  = token | quoted-string
1189    chunk-data     = chunk-size(OCTET)
1190    trailer        = *(entity-header CRLF)
1191</artwork></figure>
1192<t>
1193   The chunk-size field is a string of hex digits indicating the size of
1194   the chunk. The chunked encoding is ended by any chunk whose size is
1195   zero, followed by the trailer, which is terminated by an empty line.
1196</t>
1197<t>
1198   The trailer allows the sender to include additional HTTP header
1199   fields at the end of the message. The Trailer header field can be
1200   used to indicate which header fields are included in a trailer (see
1201   <xref target="header.trailer"/>).
1202</t>
1203<t>
1204   A server using chunked transfer-coding in a response &MUST-NOT; use the
1205   trailer for any header fields unless at least one of the following is
1206   true:
1207  <list style="numbers">
1208    <t>the request included a TE header field that indicates "trailers" is
1209     acceptable in the transfer-coding of the  response, as described in
1210     <xref target="header.te"/>; or,</t>
1211
1212    <t>the server is the origin server for the response, the trailer
1213     fields consist entirely of optional metadata, and the recipient
1214     could use the message (in a manner acceptable to the origin server)
1215     without receiving this metadata.  In other words, the origin server
1216     is willing to accept the possibility that the trailer fields might
1217     be silently discarded along the path to the client.</t>
1218  </list>
1219</t>
1220<t>
1221   This requirement prevents an interoperability failure when the
1222   message is being received by an HTTP/1.1 (or later) proxy and
1223   forwarded to an HTTP/1.0 recipient. It avoids a situation where
1224   compliance with the protocol would have necessitated a possibly
1225   infinite buffer on the proxy.
1226</t>
1227<t>
1228   A process for decoding the "chunked" transfer-coding
1229   can be represented in pseudo-code as:
1230</t>
1231<figure><artwork type="code">
1232    length := 0
1233    read chunk-size, chunk-extension (if any) and CRLF
1234    while (chunk-size &gt; 0) {
1235       read chunk-data and CRLF
1236       append chunk-data to entity-body
1237       length := length + chunk-size
1238       read chunk-size and CRLF
1239    }
1240    read entity-header
1241    while (entity-header not empty) {
1242       append entity-header to existing header fields
1243       read entity-header
1244    }
1245    Content-Length := length
1246    Remove "chunked" from Transfer-Encoding
1247</artwork></figure>
1248<t>
1249   All HTTP/1.1 applications &MUST; be able to receive and decode the
1250   "chunked" transfer-coding, and &MUST; ignore chunk-extension extensions
1251   they do not understand.
1252</t>
1253</section>
1254</section>
1255
1256</section>
1257
1258<section title="HTTP Message" anchor="http.message">
1259
1260<section title="Message Types" anchor="message.types">
1261<t>
1262   HTTP messages consist of requests from client to server and responses
1263   from server to client.
1264</t>
1265<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"/>
1266    HTTP-message   = Request | Response     ; HTTP/1.1 messages
1267</artwork></figure>
1268<t>
1269   Request (<xref target="request"/>) and Response (<xref target="response"/>) messages use the generic
1270   message format of RFC 822 <xref target="RFC822"/> for transferring entities (the payload
1271   of the message). Both types of message consist of a start-line, zero
1272   or more header fields (also known as "headers"), an empty line (i.e.,
1273   a line with nothing preceding the CRLF) indicating the end of the
1274   header fields, and possibly a message-body.
1275</t>
1276<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="generic-message"/><iref primary="true" item="Grammar" subitem="start-line"/>
1277     generic-message = start-line
1278                       *(message-header CRLF)
1279                       CRLF
1280                       [ message-body ]
1281     start-line      = Request-Line | Status-Line
1282</artwork></figure>
1283<t>
1284   In the interest of robustness, servers &SHOULD; ignore any empty
1285   line(s) received where a Request-Line is expected. In other words, if
1286   the server is reading the protocol stream at the beginning of a
1287   message and receives a CRLF first, it should ignore the CRLF.
1288</t>
1289<t>
1290   Certain buggy HTTP/1.0 client implementations generate extra CRLF's
1291   after a POST request. To restate what is explicitly forbidden by the
1292   BNF, an HTTP/1.1 client &MUST-NOT; preface or follow a request with an
1293   extra CRLF.
1294</t>
1295</section>
1296
1297<section title="Message Headers" anchor="message.headers">
1298<t>
1299   HTTP header fields, which include general-header (<xref target="general.header.fields"/>),
1300   request-header (&request-header-fields;), response-header (&response-header-fields;), and
1301   entity-header (&entity-header-fields;) fields, follow the same generic format as
1302   that given in <xref target="RFC822" x:fmt="sec" x:sec="3.1"/> of RFC 822 <xref target="RFC822"/>. Each header field consists
1303   of a name followed by a colon (":") and the field value. Field names
1304   are case-insensitive. The field value &MAY; be preceded by any amount
1305   of LWS, though a single SP is preferred. Header fields can be
1306   extended over multiple lines by preceding each extra line with at
1307   least one SP or HT. Applications ought to follow "common form", where
1308   one is known or indicated, when generating HTTP constructs, since
1309   there might exist some implementations that fail to accept anything
1310   beyond the common forms.
1311</t>
1312<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-header"/><iref primary="true" item="Grammar" subitem="field-name"/><iref primary="true" item="Grammar" subitem="field-value"/><iref primary="true" item="Grammar" subitem="field-content"/>
1313    message-header = field-name ":" [ field-value ]
1314    field-name     = token
1315    field-value    = *( field-content | LWS )
1316    field-content  = &lt;the OCTETs making up the field-value
1317                     and consisting of either *TEXT or combinations
1318                     of token, separators, and quoted-string&gt;
1319</artwork></figure>
1320<t>
1321   The field-content does not include any leading or trailing LWS:
1322   linear white space occurring before the first non-whitespace
1323   character of the field-value or after the last non-whitespace
1324   character of the field-value. Such leading or trailing LWS &MAY; be
1325   removed without changing the semantics of the field value. Any LWS
1326   that occurs between field-content &MAY; be replaced with a single SP
1327   before interpreting the field value or forwarding the message
1328   downstream.
1329</t>
1330<t>
1331   The order in which header fields with differing field names are
1332   received is not significant. However, it is "good practice" to send
1333   general-header fields first, followed by request-header or response-header
1334   fields, and ending with the entity-header fields.
1335</t>
1336<t>
1337   Multiple message-header fields with the same field-name &MAY; be
1338   present in a message if and only if the entire field-value for that
1339   header field is defined as a comma-separated list [i.e., #(values)].
1340   It &MUST; be possible to combine the multiple header fields into one
1341   "field-name: field-value" pair, without changing the semantics of the
1342   message, by appending each subsequent field-value to the first, each
1343   separated by a comma. The order in which header fields with the same
1344   field-name are received is therefore significant to the
1345   interpretation of the combined field value, and thus a proxy &MUST-NOT;
1346   change the order of these field values when a message is forwarded.
1347</t>
1348</section>
1349
1350<section title="Message Body" anchor="message.body">
1351<t>
1352   The message-body (if any) of an HTTP message is used to carry the
1353   entity-body associated with the request or response. The message-body
1354   differs from the entity-body only when a transfer-coding has been
1355   applied, as indicated by the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
1356</t>
1357<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1358    message-body = entity-body
1359                 | &lt;entity-body encoded as per Transfer-Encoding&gt;
1360</artwork></figure>
1361<t>
1362   Transfer-Encoding &MUST; be used to indicate any transfer-codings
1363   applied by an application to ensure safe and proper transfer of the
1364   message. Transfer-Encoding is a property of the message, not of the
1365   entity, and thus &MAY; be added or removed by any application along the
1366   request/response chain. (However, <xref target="transfer.codings"/> places restrictions on
1367   when certain transfer-codings may be used.)
1368</t>
1369<t>
1370   The rules for when a message-body is allowed in a message differ for
1371   requests and responses.
1372</t>
1373<t>
1374   The presence of a message-body in a request is signaled by the
1375   inclusion of a Content-Length or Transfer-Encoding header field in
1376   the request's message-headers. A message-body &MUST-NOT; be included in
1377   a request if the specification of the request method (&method;)
1378   does not allow sending an entity-body in requests. A server &SHOULD;
1379   read and forward a message-body on any request; if the request method
1380   does not include defined semantics for an entity-body, then the
1381   message-body &SHOULD; be ignored when handling the request.
1382</t>
1383<t>
1384   For response messages, whether or not a message-body is included with
1385   a message is dependent on both the request method and the response
1386   status code (<xref target="status.code.and.reason.phrase"/>). All responses to the HEAD request method
1387   &MUST-NOT; include a message-body, even though the presence of entity-header
1388   fields might lead one to believe they do. All 1xx
1389   (informational), 204 (no content), and 304 (not modified) responses
1390   &MUST-NOT; include a message-body. All other responses do include a
1391   message-body, although it &MAY; be of zero length.
1392</t>
1393</section>
1394
1395<section title="Message Length" anchor="message.length">
1396<t>
1397   The transfer-length of a message is the length of the message-body as
1398   it appears in the message; that is, after any transfer-codings have
1399   been applied. When a message-body is included with a message, the
1400   transfer-length of that body is determined by one of the following
1401   (in order of precedence):
1402</t>
1403<t>
1404  <list style="numbers">
1405    <x:lt><t>
1406     Any response message which "&MUST-NOT;" include a message-body (such
1407     as the 1xx, 204, and 304 responses and any response to a HEAD
1408     request) is always terminated by the first empty line after the
1409     header fields, regardless of the entity-header fields present in
1410     the message.
1411    </t></x:lt>
1412    <x:lt><t>
1413     If a Transfer-Encoding header field (<xref target="header.transfer-encoding"/>) is present and
1414     has any value other than "identity", then the transfer-length is
1415     defined by use of the "chunked" transfer-coding (<xref target="transfer.codings"/>),
1416     unless the message is terminated by closing the connection.
1417    </t></x:lt>
1418    <x:lt><t>
1419     If a Content-Length header field (<xref target="header.content-length"/>) is present, its
1420     decimal value in OCTETs represents both the entity-length and the
1421     transfer-length. The Content-Length header field &MUST-NOT; be sent
1422     if these two lengths are different (i.e., if a Transfer-Encoding
1423     header field is present). If a message is received with both a
1424     Transfer-Encoding header field and a Content-Length header field,
1425     the latter &MUST; be ignored.
1426    </t></x:lt>
1427    <x:lt><t>
1428     If the message uses the media type "multipart/byteranges", and the
1429     ransfer-length is not otherwise specified, then this self-elimiting
1430     media type defines the transfer-length. This media type
1431     UST NOT be used unless the sender knows that the recipient can arse
1432     it; the presence in a request of a Range header with ultiple byte-range
1433     specifiers from a 1.1 client implies that the lient can parse
1434     multipart/byteranges responses.
1435    <list style="empty"><t>
1436       A range header might be forwarded by a 1.0 proxy that does not
1437       understand multipart/byteranges; in this case the server &MUST;
1438       delimit the message using methods defined in items 1, 3 or 5 of
1439       this section.
1440    </t></list>
1441    </t></x:lt>
1442    <x:lt><t>
1443     By the server closing the connection. (Closing the connection
1444     cannot be used to indicate the end of a request body, since that
1445     would leave no possibility for the server to send back a response.)
1446    </t></x:lt>
1447  </list>
1448</t>
1449<t>
1450   For compatibility with HTTP/1.0 applications, HTTP/1.1 requests
1451   containing a message-body &MUST; include a valid Content-Length header
1452   field unless the server is known to be HTTP/1.1 compliant. If a
1453   request contains a message-body and a Content-Length is not given,
1454   the server &SHOULD; respond with 400 (bad request) if it cannot
1455   determine the length of the message, or with 411 (length required) if
1456   it wishes to insist on receiving a valid Content-Length.
1457</t>
1458<t>
1459   All HTTP/1.1 applications that receive entities &MUST; accept the
1460   "chunked" transfer-coding (<xref target="transfer.codings"/>), thus allowing this mechanism
1461   to be used for messages when the message length cannot be determined
1462   in advance.
1463</t>
1464<t>
1465   Messages &MUST-NOT; include both a Content-Length header field and a
1466   non-identity transfer-coding. If the message does include a non-identity
1467   transfer-coding, the Content-Length &MUST; be ignored.
1468</t>
1469<t>
1470   When a Content-Length is given in a message where a message-body is
1471   allowed, its field value &MUST; exactly match the number of OCTETs in
1472   the message-body. HTTP/1.1 user agents &MUST; notify the user when an
1473   invalid length is received and detected.
1474</t>
1475</section>
1476
1477<section title="General Header Fields" anchor="general.header.fields">
1478<t>
1479   There are a few header fields which have general applicability for
1480   both request and response messages, but which do not apply to the
1481   entity being transferred. These header fields apply only to the
1482   message being transmitted.
1483</t>
1484<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="general-header"/>
1485    general-header = Cache-Control            ; &header-cache-control;
1486                   | Connection               ; <xref target="header.connection"/>
1487                   | Date                     ; <xref target="header.date"/>
1488                   | Pragma                   ; &header-pragma;
1489                   | Trailer                  ; <xref target="header.trailer"/>
1490                   | Transfer-Encoding        ; <xref target="header.transfer-encoding"/>
1491                   | Upgrade                  ; <xref target="header.upgrade"/>
1492                   | Via                      ; <xref target="header.via"/>
1493                   | Warning                  ; &header-warning;
1494</artwork></figure>
1495<t>
1496   General-header field names can be extended reliably only in
1497   combination with a change in the protocol version. However, new or
1498   experimental header fields may be given the semantics of general
1499   header fields if all parties in the communication recognize them to
1500   be general-header fields. Unrecognized header fields are treated as
1501   entity-header fields.
1502</t>
1503</section>
1504</section>
1505
1506<section title="Request" anchor="request">
1507<t>
1508   A request message from a client to a server includes, within the
1509   first line of that message, the method to be applied to the resource,
1510   the identifier of the resource, and the protocol version in use.
1511</t>
1512<!--                 Host                      ; should be moved here eventually -->
1513<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request"/>
1514     Request       = Request-Line              ; <xref target="request-line"/>
1515                     *(( general-header        ; <xref target="general.header.fields"/>
1516                      | request-header         ; &request-header-fields;
1517                      | entity-header ) CRLF)  ; &entity-header-fields;
1518                     CRLF
1519                     [ message-body ]          ; <xref target="message.body"/>
1520</artwork></figure>
1521
1522<section title="Request-Line" anchor="request-line">
1523<t>
1524   The Request-Line begins with a method token, followed by the
1525   Request-URI and the protocol version, and ending with CRLF. The
1526   elements are separated by SP characters. No CR or LF is allowed
1527   except in the final CRLF sequence.
1528</t>
1529<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-Line"/>
1530     Request-Line   = Method SP Request-URI SP HTTP-Version CRLF
1531</artwork></figure>
1532
1533<section title="Method" anchor="method">
1534<t>
1535   The Method  token indicates the method to be performed on the
1536   resource identified by the Request-URI. The method is case-sensitive.
1537</t>
1538<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Method"/><iref primary="true" item="Grammar" subitem="extension-method"/>
1539    Method         = token
1540</artwork></figure>
1541</section>
1542
1543<section title="Request-URI" anchor="request-uri">
1544<t>
1545   The Request-URI is a Uniform Resource Identifier (<xref target="uri"/>) and
1546   identifies the resource upon which to apply the request.
1547</t>
1548<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-URI"/>
1549    Request-URI    = "*" | absoluteURI | abs_path | authority
1550</artwork></figure>
1551<t>
1552   The four options for Request-URI are dependent on the nature of the
1553   request. The asterisk "*" means that the request does not apply to a
1554   particular resource, but to the server itself, and is only allowed
1555   when the method used does not necessarily apply to a resource. One
1556   example would be
1557</t>
1558<figure><artwork type="example">
1559    OPTIONS * HTTP/1.1
1560</artwork></figure>
1561<t>
1562   The absoluteURI form is &REQUIRED; when the request is being made to a
1563   proxy. The proxy is requested to forward the request or service it
1564   from a valid cache, and return the response. Note that the proxy &MAY;
1565   forward the request on to another proxy or directly to the server
1566   specified by the absoluteURI. In order to avoid request loops, a
1567   proxy &MUST; be able to recognize all of its server names, including
1568   any aliases, local variations, and the numeric IP address. An example
1569   Request-Line would be:
1570</t>
1571<figure><artwork type="example">
1572    GET http://www.w3.org/pub/WWW/TheProject.html HTTP/1.1
1573</artwork></figure>
1574<t>
1575   To allow for transition to absoluteURIs in all requests in future
1576   versions of HTTP, all HTTP/1.1 servers &MUST; accept the absoluteURI
1577   form in requests, even though HTTP/1.1 clients will only generate
1578   them in requests to proxies.
1579</t>
1580<t>
1581   The authority form is only used by the CONNECT method (&CONNECT;).
1582</t>
1583<t>
1584   The most common form of Request-URI is that used to identify a
1585   resource on an origin server or gateway. In this case the absolute
1586   path of the URI &MUST; be transmitted (see <xref target="general.syntax"/>, abs_path) as
1587   the Request-URI, and the network location of the URI (authority) &MUST;
1588   be transmitted in a Host header field. For example, a client wishing
1589   to retrieve the resource above directly from the origin server would
1590   create a TCP connection to port 80 of the host "www.w3.org" and send
1591   the lines:
1592</t>
1593<figure><artwork type="example">
1594    GET /pub/WWW/TheProject.html HTTP/1.1
1595    Host: www.w3.org
1596</artwork></figure>
1597<t>
1598   followed by the remainder of the Request. Note that the absolute path
1599   cannot be empty; if none is present in the original URI, it &MUST; be
1600   given as "/" (the server root).
1601</t>
1602<t>
1603   The Request-URI is transmitted in the format specified in
1604   <xref target="general.syntax"/>. If the Request-URI is encoded using the "% HEX HEX" encoding
1605   <xref target="RFC2396"/>, the origin server &MUST; decode the Request-URI in order to
1606   properly interpret the request. Servers &SHOULD; respond to invalid
1607   Request-URIs with an appropriate status code.
1608</t>
1609<t>
1610   A transparent proxy &MUST-NOT; rewrite the "abs_path" part of the
1611   received Request-URI when forwarding it to the next inbound server,
1612   except as noted above to replace a null abs_path with "/".
1613</t>
1614<t>
1615  <list><t>
1616      <x:h>Note:</x:h> The "no rewrite" rule prevents the proxy from changing the
1617      meaning of the request when the origin server is improperly using
1618      a non-reserved URI character for a reserved purpose.  Implementors
1619      should be aware that some pre-HTTP/1.1 proxies have been known to
1620      rewrite the Request-URI.
1621  </t></list>
1622</t>
1623</section>
1624</section>
1625
1626<section title="The Resource Identified by a Request" anchor="the.resource.identified.by.a.request">
1627<t>
1628   The exact resource identified by an Internet request is determined by
1629   examining both the Request-URI and the Host header field.
1630</t>
1631<t>
1632   An origin server that does not allow resources to differ by the
1633   requested host &MAY; ignore the Host header field value when
1634   determining the resource identified by an HTTP/1.1 request. (But see
1635   <xref target="changes.to.simplify.multi-homed.web.servers.and.conserve.ip.addresses"/>
1636   for other requirements on Host support in HTTP/1.1.)
1637</t>
1638<t>
1639   An origin server that does differentiate resources based on the host
1640   requested (sometimes referred to as virtual hosts or vanity host
1641   names) &MUST; use the following rules for determining the requested
1642   resource on an HTTP/1.1 request:
1643  <list style="numbers">
1644    <t>If Request-URI is an absoluteURI, the host is part of the
1645     Request-URI. Any Host header field value in the request &MUST; be
1646     ignored.</t>
1647    <t>If the Request-URI is not an absoluteURI, and the request includes
1648     a Host header field, the host is determined by the Host header
1649     field value.</t>
1650    <t>If the host as determined by rule 1 or 2 is not a valid host on
1651     the server, the response &MUST; be a 400 (Bad Request) error message.</t>
1652  </list>
1653</t>
1654<t>
1655   Recipients of an HTTP/1.0 request that lacks a Host header field &MAY;
1656   attempt to use heuristics (e.g., examination of the URI path for
1657   something unique to a particular host) in order to determine what
1658   exact resource is being requested.
1659</t>
1660</section>
1661
1662</section>
1663
1664
1665<section title="Response" anchor="response">
1666<t>
1667   After receiving and interpreting a request message, a server responds
1668   with an HTTP response message.
1669</t>
1670<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Response"/>
1671    Response      = Status-Line               ; <xref target="status-line"/>
1672                    *(( general-header        ; <xref target="general.header.fields"/>
1673                     | response-header        ; &response-header-fields;
1674                     | entity-header ) CRLF)  ; &entity-header-fields;
1675                    CRLF
1676                    [ message-body ]          ; <xref target="message.body"/>
1677</artwork></figure>
1678
1679<section title="Status-Line" anchor="status-line">
1680<t>
1681   The first line of a Response message is the Status-Line, consisting
1682   of the protocol version followed by a numeric status code and its
1683   associated textual phrase, with each element separated by SP
1684   characters. No CR or LF is allowed except in the final CRLF sequence.
1685</t>
1686<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Line"/>
1687    Status-Line = HTTP-Version SP Status-Code SP Reason-Phrase CRLF
1688</artwork></figure>
1689
1690<section title="Status Code and Reason Phrase" anchor="status.code.and.reason.phrase">
1691<t>
1692   The Status-Code element is a 3-digit integer result code of the
1693   attempt to understand and satisfy the request. These codes are fully
1694   defined in &status-codes;. The Reason-Phrase is intended to give a short
1695   textual description of the Status-Code. The Status-Code is intended
1696   for use by automata and the Reason-Phrase is intended for the human
1697   user. The client is not required to examine or display the Reason-Phrase.
1698</t>
1699<t>
1700   The first digit of the Status-Code defines the class of response. The
1701   last two digits do not have any categorization role. There are 5
1702   values for the first digit:
1703  <list style="symbols">
1704    <t>
1705      1xx: Informational - Request received, continuing process
1706    </t>
1707    <t>
1708      2xx: Success - The action was successfully received,
1709        understood, and accepted
1710    </t>
1711    <t>
1712      3xx: Redirection - Further action must be taken in order to
1713        complete the request
1714    </t>
1715    <t>
1716      4xx: Client Error - The request contains bad syntax or cannot
1717        be fulfilled
1718    </t>
1719    <t>
1720      5xx: Server Error - The server failed to fulfill an apparently
1721        valid request
1722    </t>
1723  </list>
1724</t>
1725<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Code"/><iref primary="true" item="Grammar" subitem="extension-code"/><iref primary="true" item="Grammar" subitem="Reason-Phrase"/>
1726   Status-Code    = 3DIGIT
1727   Reason-Phrase  = *&lt;TEXT, excluding CR, LF&gt;
1728</artwork></figure>
1729</section>
1730</section>
1731
1732</section>
1733
1734
1735<section title="Connections" anchor="connections">
1736
1737<section title="Persistent Connections" anchor="persistent.connections">
1738
1739<section title="Purpose" anchor="persistent.purpose">
1740<t>
1741   Prior to persistent connections, a separate TCP connection was
1742   established to fetch each URL, increasing the load on HTTP servers
1743   and causing congestion on the Internet. The use of inline images and
1744   other associated data often require a client to make multiple
1745   requests of the same server in a short amount of time. Analysis of
1746   these performance problems and results from a prototype
1747   implementation are available <xref target="Pad1995"/> <xref target="Spe"/>. Implementation experience and
1748   measurements of actual HTTP/1.1 (RFC 2068) implementations show good
1749   results <xref target="Nie1997"/>. Alternatives have also been explored, for example,
1750   T/TCP <xref target="Tou1998"/>.
1751</t>
1752<t>
1753   Persistent HTTP connections have a number of advantages:
1754  <list style="symbols">
1755      <t>
1756        By opening and closing fewer TCP connections, CPU time is saved
1757        in routers and hosts (clients, servers, proxies, gateways,
1758        tunnels, or caches), and memory used for TCP protocol control
1759        blocks can be saved in hosts.
1760      </t>
1761      <t>
1762        HTTP requests and responses can be pipelined on a connection.
1763        Pipelining allows a client to make multiple requests without
1764        waiting for each response, allowing a single TCP connection to
1765        be used much more efficiently, with much lower elapsed time.
1766      </t>
1767      <t>
1768        Network congestion is reduced by reducing the number of packets
1769        caused by TCP opens, and by allowing TCP sufficient time to
1770        determine the congestion state of the network.
1771      </t>
1772      <t>
1773        Latency on subsequent requests is reduced since there is no time
1774        spent in TCP's connection opening handshake.
1775      </t>
1776      <t>
1777        HTTP can evolve more gracefully, since errors can be reported
1778        without the penalty of closing the TCP connection. Clients using
1779        future versions of HTTP might optimistically try a new feature,
1780        but if communicating with an older server, retry with old
1781        semantics after an error is reported.
1782      </t>
1783    </list>
1784</t>
1785<t>
1786   HTTP implementations &SHOULD; implement persistent connections.
1787</t>
1788</section>
1789
1790<section title="Overall Operation" anchor="persistent.overall">
1791<t>
1792   A significant difference between HTTP/1.1 and earlier versions of
1793   HTTP is that persistent connections are the default behavior of any
1794   HTTP connection. That is, unless otherwise indicated, the client
1795   &SHOULD; assume that the server will maintain a persistent connection,
1796   even after error responses from the server.
1797</t>
1798<t>
1799   Persistent connections provide a mechanism by which a client and a
1800   server can signal the close of a TCP connection. This signaling takes
1801   place using the Connection header field (<xref target="header.connection"/>). Once a close
1802   has been signaled, the client &MUST-NOT; send any more requests on that
1803   connection.
1804</t>
1805
1806<section title="Negotiation" anchor="persistent.negotiation">
1807<t>
1808   An HTTP/1.1 server &MAY; assume that a HTTP/1.1 client intends to
1809   maintain a persistent connection unless a Connection header including
1810   the connection-token "close" was sent in the request. If the server
1811   chooses to close the connection immediately after sending the
1812   response, it &SHOULD; send a Connection header including the
1813   connection-token close.
1814</t>
1815<t>
1816   An HTTP/1.1 client &MAY; expect a connection to remain open, but would
1817   decide to keep it open based on whether the response from a server
1818   contains a Connection header with the connection-token close. In case
1819   the client does not want to maintain a connection for more than that
1820   request, it &SHOULD; send a Connection header including the
1821   connection-token close.
1822</t>
1823<t>
1824   If either the client or the server sends the close token in the
1825   Connection header, that request becomes the last one for the
1826   connection.
1827</t>
1828<t>
1829   Clients and servers &SHOULD-NOT;  assume that a persistent connection is
1830   maintained for HTTP versions less than 1.1 unless it is explicitly
1831   signaled. See <xref target="compatibility.with.http.1.0.persistent.connections"/> for more information on backward
1832   compatibility with HTTP/1.0 clients.
1833</t>
1834<t>
1835   In order to remain persistent, all messages on the connection &MUST;
1836   have a self-defined message length (i.e., one not defined by closure
1837   of the connection), as described in <xref target="message.length"/>.
1838</t>
1839</section>
1840
1841<section title="Pipelining" anchor="pipelining">
1842<t>
1843   A client that supports persistent connections &MAY; "pipeline" its
1844   requests (i.e., send multiple requests without waiting for each
1845   response). A server &MUST; send its responses to those requests in the
1846   same order that the requests were received.
1847</t>
1848<t>
1849   Clients which assume persistent connections and pipeline immediately
1850   after connection establishment &SHOULD; be prepared to retry their
1851   connection if the first pipelined attempt fails. If a client does
1852   such a retry, it &MUST-NOT; pipeline before it knows the connection is
1853   persistent. Clients &MUST; also be prepared to resend their requests if
1854   the server closes the connection before sending all of the
1855   corresponding responses.
1856</t>
1857<t>
1858   Clients &SHOULD-NOT;  pipeline requests using non-idempotent methods or
1859   non-idempotent sequences of methods (see &idempotent-methods;). Otherwise, a
1860   premature termination of the transport connection could lead to
1861   indeterminate results. A client wishing to send a non-idempotent
1862   request &SHOULD; wait to send that request until it has received the
1863   response status for the previous request.
1864</t>
1865</section>
1866</section>
1867
1868<section title="Proxy Servers" anchor="persistent.proxy">
1869<t>
1870   It is especially important that proxies correctly implement the
1871   properties of the Connection header field as specified in <xref target="header.connection"/>.
1872</t>
1873<t>
1874   The proxy server &MUST; signal persistent connections separately with
1875   its clients and the origin servers (or other proxy servers) that it
1876   connects to. Each persistent connection applies to only one transport
1877   link.
1878</t>
1879<t>
1880   A proxy server &MUST-NOT; establish a HTTP/1.1 persistent connection
1881   with an HTTP/1.0 client (but see RFC 2068 <xref target="RFC2068"/> for information and
1882   discussion of the problems with the Keep-Alive header implemented by
1883   many HTTP/1.0 clients).
1884</t>
1885</section>
1886
1887<section title="Practical Considerations" anchor="persistent.practical">
1888<t>
1889   Servers will usually have some time-out value beyond which they will
1890   no longer maintain an inactive connection. Proxy servers might make
1891   this a higher value since it is likely that the client will be making
1892   more connections through the same server. The use of persistent
1893   connections places no requirements on the length (or existence) of
1894   this time-out for either the client or the server.
1895</t>
1896<t>
1897   When a client or server wishes to time-out it &SHOULD; issue a graceful
1898   close on the transport connection. Clients and servers &SHOULD; both
1899   constantly watch for the other side of the transport close, and
1900   respond to it as appropriate. If a client or server does not detect
1901   the other side's close promptly it could cause unnecessary resource
1902   drain on the network.
1903</t>
1904<t>
1905   A client, server, or proxy &MAY; close the transport connection at any
1906   time. For example, a client might have started to send a new request
1907   at the same time that the server has decided to close the "idle"
1908   connection. From the server's point of view, the connection is being
1909   closed while it was idle, but from the client's point of view, a
1910   request is in progress.
1911</t>
1912<t>
1913   This means that clients, servers, and proxies &MUST; be able to recover
1914   from asynchronous close events. Client software &SHOULD; reopen the
1915   transport connection and retransmit the aborted sequence of requests
1916   without user interaction so long as the request sequence is
1917   idempotent (see &idempotent-methods;). Non-idempotent methods or sequences
1918   &MUST-NOT; be automatically retried, although user agents &MAY; offer a
1919   human operator the choice of retrying the request(s). Confirmation by
1920   user-agent software with semantic understanding of the application
1921   &MAY; substitute for user confirmation. The automatic retry &SHOULD-NOT; 
1922   be repeated if the second sequence of requests fails.
1923</t>
1924<t>
1925   Servers &SHOULD; always respond to at least one request per connection,
1926   if at all possible. Servers &SHOULD-NOT;  close a connection in the
1927   middle of transmitting a response, unless a network or client failure
1928   is suspected.
1929</t>
1930<t>
1931   Clients that use persistent connections &SHOULD; limit the number of
1932   simultaneous connections that they maintain to a given server. A
1933   single-user client &SHOULD-NOT; maintain more than 2 connections with
1934   any server or proxy. A proxy &SHOULD; use up to 2*N connections to
1935   another server or proxy, where N is the number of simultaneously
1936   active users. These guidelines are intended to improve HTTP response
1937   times and avoid congestion.
1938</t>
1939</section>
1940</section>
1941
1942<section title="Message Transmission Requirements" anchor="message.transmission.requirements">
1943
1944<section title="Persistent Connections and Flow Control" anchor="persistent.flow">
1945<t>
1946   HTTP/1.1 servers &SHOULD; maintain persistent connections and use TCP's
1947   flow control mechanisms to resolve temporary overloads, rather than
1948   terminating connections with the expectation that clients will retry.
1949   The latter technique can exacerbate network congestion.
1950</t>
1951</section>
1952
1953<section title="Monitoring Connections for Error Status Messages" anchor="persistent.monitor">
1954<t>
1955   An HTTP/1.1 (or later) client sending a message-body &SHOULD; monitor
1956   the network connection for an error status while it is transmitting
1957   the request. If the client sees an error status, it &SHOULD;
1958   immediately cease transmitting the body. If the body is being sent
1959   using a "chunked" encoding (<xref target="transfer.codings"/>), a zero length chunk and
1960   empty trailer &MAY; be used to prematurely mark the end of the message.
1961   If the body was preceded by a Content-Length header, the client &MUST;
1962   close the connection.
1963</t>
1964</section>
1965
1966<section title="Use of the 100 (Continue) Status" anchor="use.of.the.100.status">
1967<t>
1968   The purpose of the 100 (Continue) status (see &status-100;) is to
1969   allow a client that is sending a request message with a request body
1970   to determine if the origin server is willing to accept the request
1971   (based on the request headers) before the client sends the request
1972   body. In some cases, it might either be inappropriate or highly
1973   inefficient for the client to send the body if the server will reject
1974   the message without looking at the body.
1975</t>
1976<t>
1977   Requirements for HTTP/1.1 clients:
1978  <list style="symbols">
1979    <t>
1980        If a client will wait for a 100 (Continue) response before
1981        sending the request body, it &MUST; send an Expect request-header
1982        field (&header-expect;) with the "100-continue" expectation.
1983    </t>
1984    <t>
1985        A client &MUST-NOT; send an Expect request-header field (&header-expect;)
1986        with the "100-continue" expectation if it does not intend
1987        to send a request body.
1988    </t>
1989  </list>
1990</t>
1991<t>
1992   Because of the presence of older implementations, the protocol allows
1993   ambiguous situations in which a client may send "Expect: 100-continue"
1994   without receiving either a 417 (Expectation Failed) status
1995   or a 100 (Continue) status. Therefore, when a client sends this
1996   header field to an origin server (possibly via a proxy) from which it
1997   has never seen a 100 (Continue) status, the client &SHOULD-NOT;  wait
1998   for an indefinite period before sending the request body.
1999</t>
2000<t>
2001   Requirements for HTTP/1.1 origin servers:
2002  <list style="symbols">
2003    <t> Upon receiving a request which includes an Expect request-header
2004        field with the "100-continue" expectation, an origin server &MUST;
2005        either respond with 100 (Continue) status and continue to read
2006        from the input stream, or respond with a final status code. The
2007        origin server &MUST-NOT; wait for the request body before sending
2008        the 100 (Continue) response. If it responds with a final status
2009        code, it &MAY; close the transport connection or it &MAY; continue
2010        to read and discard the rest of the request.  It &MUST-NOT;
2011        perform the requested method if it returns a final status code.
2012    </t>
2013    <t> An origin server &SHOULD-NOT;  send a 100 (Continue) response if
2014        the request message does not include an Expect request-header
2015        field with the "100-continue" expectation, and &MUST-NOT; send a
2016        100 (Continue) response if such a request comes from an HTTP/1.0
2017        (or earlier) client. There is an exception to this rule: for
2018        compatibility with RFC 2068, a server &MAY; send a 100 (Continue)
2019        status in response to an HTTP/1.1 PUT or POST request that does
2020        not include an Expect request-header field with the "100-continue"
2021        expectation. This exception, the purpose of which is
2022        to minimize any client processing delays associated with an
2023        undeclared wait for 100 (Continue) status, applies only to
2024        HTTP/1.1 requests, and not to requests with any other HTTP-version
2025        value.
2026    </t>
2027    <t> An origin server &MAY; omit a 100 (Continue) response if it has
2028        already received some or all of the request body for the
2029        corresponding request.
2030    </t>
2031    <t> An origin server that sends a 100 (Continue) response &MUST;
2032    ultimately send a final status code, once the request body is
2033        received and processed, unless it terminates the transport
2034        connection prematurely.
2035    </t>
2036    <t> If an origin server receives a request that does not include an
2037        Expect request-header field with the "100-continue" expectation,
2038        the request includes a request body, and the server responds
2039        with a final status code before reading the entire request body
2040        from the transport connection, then the server &SHOULD-NOT;  close
2041        the transport connection until it has read the entire request,
2042        or until the client closes the connection. Otherwise, the client
2043        might not reliably receive the response message. However, this
2044        requirement is not be construed as preventing a server from
2045        defending itself against denial-of-service attacks, or from
2046        badly broken client implementations.
2047      </t>
2048    </list>
2049</t>
2050<t>
2051   Requirements for HTTP/1.1 proxies:
2052  <list style="symbols">
2053    <t> If a proxy receives a request that includes an Expect request-header
2054        field with the "100-continue" expectation, and the proxy
2055        either knows that the next-hop server complies with HTTP/1.1 or
2056        higher, or does not know the HTTP version of the next-hop
2057        server, it &MUST; forward the request, including the Expect header
2058        field.
2059    </t>
2060    <t> If the proxy knows that the version of the next-hop server is
2061        HTTP/1.0 or lower, it &MUST-NOT; forward the request, and it &MUST;
2062        respond with a 417 (Expectation Failed) status.
2063    </t>
2064    <t> Proxies &SHOULD; maintain a cache recording the HTTP version
2065        numbers received from recently-referenced next-hop servers.
2066    </t>
2067    <t> A proxy &MUST-NOT; forward a 100 (Continue) response if the
2068        request message was received from an HTTP/1.0 (or earlier)
2069        client and did not include an Expect request-header field with
2070        the "100-continue" expectation. This requirement overrides the
2071        general rule for forwarding of 1xx responses (see &status-1xx;).
2072    </t>
2073  </list>
2074</t>
2075</section>
2076
2077<section title="Client Behavior if Server Prematurely Closes Connection" anchor="connection.premature">
2078<t>
2079   If an HTTP/1.1 client sends a request which includes a request body,
2080   but which does not include an Expect request-header field with the
2081   "100-continue" expectation, and if the client is not directly
2082   connected to an HTTP/1.1 origin server, and if the client sees the
2083   connection close before receiving any status from the server, the
2084   client &SHOULD; retry the request.  If the client does retry this
2085   request, it &MAY; use the following "binary exponential backoff"
2086   algorithm to be assured of obtaining a reliable response:
2087  <list style="numbers">
2088    <t>
2089      Initiate a new connection to the server
2090    </t>
2091    <t>
2092      Transmit the request-headers
2093    </t>
2094    <t>
2095      Initialize a variable R to the estimated round-trip time to the
2096         server (e.g., based on the time it took to establish the
2097         connection), or to a constant value of 5 seconds if the round-trip
2098         time is not available.
2099    </t>
2100    <t>
2101       Compute T = R * (2**N), where N is the number of previous
2102         retries of this request.
2103    </t>
2104    <t>
2105       Wait either for an error response from the server, or for T
2106         seconds (whichever comes first)
2107    </t>
2108    <t>
2109       If no error response is received, after T seconds transmit the
2110         body of the request.
2111    </t>
2112    <t>
2113       If client sees that the connection is closed prematurely,
2114         repeat from step 1 until the request is accepted, an error
2115         response is received, or the user becomes impatient and
2116         terminates the retry process.
2117    </t>
2118  </list>
2119</t>
2120<t>
2121   If at any point an error status is received, the client
2122  <list style="symbols">
2123      <t>&SHOULD-NOT;  continue and</t>
2124
2125      <t>&SHOULD; close the connection if it has not completed sending the
2126        request message.</t>
2127    </list>
2128</t>
2129</section>
2130</section>
2131</section>
2132
2133
2134<section title="Header Field Definitions" anchor="header.fields">
2135<t>
2136   This section defines the syntax and semantics of all standard
2137   HTTP/1.1 header fields. For entity-header fields, both sender and
2138   recipient refer to either the client or the server, depending on who
2139   sends and who receives the entity.
2140</t>
2141
2142<section title="Connection" anchor="header.connection">
2143  <iref primary="true" item="Connection header" x:for-anchor=""/>
2144  <iref primary="true" item="Headers" subitem="Connection" x:for-anchor=""/>
2145<t>
2146   The Connection general-header field allows the sender to specify
2147   options that are desired for that particular connection and &MUST-NOT;
2148   be communicated by proxies over further connections.
2149</t>
2150<t>
2151   The Connection header has the following grammar:
2152</t>
2153<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-token"/>
2154    Connection = "Connection" ":" 1#(connection-token)
2155    connection-token  = token
2156</artwork></figure>
2157<t>
2158   HTTP/1.1 proxies &MUST; parse the Connection header field before a
2159   message is forwarded and, for each connection-token in this field,
2160   remove any header field(s) from the message with the same name as the
2161   connection-token. Connection options are signaled by the presence of
2162   a connection-token in the Connection header field, not by any
2163   corresponding additional header field(s), since the additional header
2164   field may not be sent if there are no parameters associated with that
2165   connection option.
2166</t>
2167<t>
2168   Message headers listed in the Connection header &MUST-NOT; include
2169   end-to-end headers, such as Cache-Control.
2170</t>
2171<t>
2172   HTTP/1.1 defines the "close" connection option for the sender to
2173   signal that the connection will be closed after completion of the
2174   response. For example,
2175</t>
2176<figure><artwork type="example">
2177    Connection: close
2178</artwork></figure>
2179<t>
2180   in either the request or the response header fields indicates that
2181   the connection &SHOULD-NOT;  be considered `persistent' (<xref target="persistent.connections"/>)
2182   after the current request/response is complete.
2183</t>
2184<t>
2185   HTTP/1.1 applications that do not support persistent connections &MUST;
2186   include the "close" connection option in every message.
2187</t>
2188<t>
2189   A system receiving an HTTP/1.0 (or lower-version) message that
2190   includes a Connection header MUST, for each connection-token in this
2191   field, remove and ignore any header field(s) from the message with
2192   the same name as the connection-token. This protects against mistaken
2193   forwarding of such header fields by pre-HTTP/1.1 proxies. See <xref target="compatibility.with.http.1.0.persistent.connections"/>.
2194</t>
2195</section>
2196
2197<section title="Content-Length" anchor="header.content-length">
2198  <iref primary="true" item="Content-Length header" x:for-anchor=""/>
2199  <iref primary="true" item="Headers" subitem="Content-Length" x:for-anchor=""/>
2200<t>
2201   The Content-Length entity-header field indicates the size of the
2202   entity-body, in decimal number of OCTETs, sent to the recipient or,
2203   in the case of the HEAD method, the size of the entity-body that
2204   would have been sent had the request been a GET.
2205</t>
2206<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
2207    Content-Length    = "Content-Length" ":" 1*DIGIT
2208</artwork></figure>
2209<t>
2210   An example is
2211</t>
2212<figure><artwork type="example">
2213    Content-Length: 3495
2214</artwork></figure>
2215<t>
2216   Applications &SHOULD; use this field to indicate the transfer-length of
2217   the message-body, unless this is prohibited by the rules in <xref target="message.length"/>.
2218</t>
2219<t>
2220   Any Content-Length greater than or equal to zero is a valid value.
2221   <xref target="message.length"/> describes how to determine the length of a message-body
2222   if a Content-Length is not given.
2223</t>
2224<t>
2225   Note that the meaning of this field is significantly different from
2226   the corresponding definition in MIME, where it is an optional field
2227   used within the "message/external-body" content-type. In HTTP, it
2228   &SHOULD; be sent whenever the message's length can be determined prior
2229   to being transferred, unless this is prohibited by the rules in
2230   <xref target="message.length"/>.
2231</t>
2232</section>
2233
2234<section title="Date" anchor="header.date">
2235  <iref primary="true" item="Date header" x:for-anchor=""/>
2236  <iref primary="true" item="Headers" subitem="Date" x:for-anchor=""/>
2237<t>
2238   The Date general-header field represents the date and time at which
2239   the message was originated, having the same semantics as orig-date in
2240   RFC 822. The field value is an HTTP-date, as described in <xref target="full.date"/>;
2241   it &MUST; be sent in RFC 1123 <xref target="RFC1123"/>-date format.
2242</t>
2243<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Date"/>
2244    Date  = "Date" ":" HTTP-date
2245</artwork></figure>
2246<t>
2247   An example is
2248</t>
2249<figure><artwork type="example">
2250    Date: Tue, 15 Nov 1994 08:12:31 GMT
2251</artwork></figure>
2252<t>
2253   Origin servers &MUST; include a Date header field in all responses,
2254   except in these cases:
2255  <list style="numbers">
2256      <t>If the response status code is 100 (Continue) or 101 (Switching
2257         Protocols), the response &MAY; include a Date header field, at
2258         the server's option.</t>
2259
2260      <t>If the response status code conveys a server error, e.g. 500
2261         (Internal Server Error) or 503 (Service Unavailable), and it is
2262         inconvenient or impossible to generate a valid Date.</t>
2263
2264      <t>If the server does not have a clock that can provide a
2265         reasonable approximation of the current time, its responses
2266         &MUST-NOT; include a Date header field. In this case, the rules
2267         in <xref target="clockless.origin.server.operation"/> &MUST; be followed.</t>
2268  </list>
2269</t>
2270<t>
2271   A received message that does not have a Date header field &MUST; be
2272   assigned one by the recipient if the message will be cached by that
2273   recipient or gatewayed via a protocol which requires a Date. An HTTP
2274   implementation without a clock &MUST-NOT; cache responses without
2275   revalidating them on every use. An HTTP cache, especially a shared
2276   cache, &SHOULD; use a mechanism, such as NTP <xref target="RFC1305"/>, to synchronize its
2277   clock with a reliable external standard.
2278</t>
2279<t>
2280   Clients &SHOULD; only send a Date header field in messages that include
2281   an entity-body, as in the case of the PUT and POST requests, and even
2282   then it is optional. A client without a clock &MUST-NOT; send a Date
2283   header field in a request.
2284</t>
2285<t>
2286   The HTTP-date sent in a Date header &SHOULD-NOT;  represent a date and
2287   time subsequent to the generation of the message. It &SHOULD; represent
2288   the best available approximation of the date and time of message
2289   generation, unless the implementation has no means of generating a
2290   reasonably accurate date and time. In theory, the date ought to
2291   represent the moment just before the entity is generated. In
2292   practice, the date can be generated at any time during the message
2293   origination without affecting its semantic value.
2294</t>
2295
2296<section title="Clockless Origin Server Operation" anchor="clockless.origin.server.operation">
2297<t>
2298   Some origin server implementations might not have a clock available.
2299   An origin server without a clock &MUST-NOT; assign Expires or Last-Modified
2300   values to a response, unless these values were associated
2301   with the resource by a system or user with a reliable clock. It &MAY;
2302   assign an Expires value that is known, at or before server
2303   configuration time, to be in the past (this allows "pre-expiration"
2304   of responses without storing separate Expires values for each
2305   resource).
2306</t>
2307</section>
2308</section>
2309
2310<section title="Host" anchor="header.host">
2311  <iref primary="true" item="Host header" x:for-anchor=""/>
2312  <iref primary="true" item="Headers" subitem="Host" x:for-anchor=""/>
2313<t>
2314   The Host request-header field specifies the Internet host and port
2315   number of the resource being requested, as obtained from the original
2316   URI given by the user or referring resource (generally an HTTP URL,
2317   as described in <xref target="http.url"/>). The Host field value &MUST; represent
2318   the naming authority of the origin server or gateway given by the
2319   original URL. This allows the origin server or gateway to
2320   differentiate between internally-ambiguous URLs, such as the root "/"
2321   URL of a server for multiple host names on a single IP address.
2322</t>
2323<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2324    Host = "Host" ":" host [ ":" port ] ; <xref target="http.url"/>
2325</artwork></figure>
2326<t>
2327   A "host" without any trailing port information implies the default
2328   port for the service requested (e.g., "80" for an HTTP URL). For
2329   example, a request on the origin server for
2330   &lt;http://www.w3.org/pub/WWW/&gt; would properly include:
2331</t>
2332<figure><artwork type="example">
2333    GET /pub/WWW/ HTTP/1.1
2334    Host: www.w3.org
2335</artwork></figure>
2336<t>
2337   A client &MUST; include a Host header field in all HTTP/1.1 request
2338   messages . If the requested URI does not include an Internet host
2339   name for the service being requested, then the Host header field &MUST;
2340   be given with an empty value. An HTTP/1.1 proxy &MUST; ensure that any
2341   request message it forwards does contain an appropriate Host header
2342   field that identifies the service being requested by the proxy. All
2343   Internet-based HTTP/1.1 servers &MUST; respond with a 400 (Bad Request)
2344   status code to any HTTP/1.1 request message which lacks a Host header
2345   field.
2346</t>
2347<t>
2348   See sections <xref target="the.resource.identified.by.a.request" format="counter"/>
2349   and <xref target="changes.to.simplify.multi-homed.web.servers.and.conserve.ip.addresses" format="counter"/>
2350   for other requirements relating to Host.
2351</t>
2352</section>
2353
2354<section title="TE" anchor="header.te">
2355  <iref primary="true" item="TE header" x:for-anchor=""/>
2356  <iref primary="true" item="Headers" subitem="TE" x:for-anchor=""/>
2357<t>
2358   The TE request-header field indicates what extension transfer-codings
2359   it is willing to accept in the response and whether or not it is
2360   willing to accept trailer fields in a chunked transfer-coding. Its
2361   value may consist of the keyword "trailers" and/or a comma-separated
2362   list of extension transfer-coding names with optional accept
2363   parameters (as described in <xref target="transfer.codings"/>).
2364</t>
2365<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TE"/><iref primary="true" item="Grammar" subitem="t-codings"/>
2366    TE        = "TE" ":" #( t-codings )
2367    t-codings = "trailers" | ( transfer-extension [ accept-params ] )
2368</artwork></figure>
2369<t>
2370   The presence of the keyword "trailers" indicates that the client is
2371   willing to accept trailer fields in a chunked transfer-coding, as
2372   defined in <xref target="chunked.transfer.encoding"/>. This keyword is reserved for use with
2373   transfer-coding values even though it does not itself represent a
2374   transfer-coding.
2375</t>
2376<t>
2377   Examples of its use are:
2378</t>
2379<figure><artwork type="example">
2380    TE: deflate
2381    TE:
2382    TE: trailers, deflate;q=0.5
2383</artwork></figure>
2384<t>
2385   The TE header field only applies to the immediate connection.
2386   Therefore, the keyword &MUST; be supplied within a Connection header
2387   field (<xref target="header.connection"/>) whenever TE is present in an HTTP/1.1 message.
2388</t>
2389<t>
2390   A server tests whether a transfer-coding is acceptable, according to
2391   a TE field, using these rules:
2392  <list style="numbers">
2393    <x:lt>
2394      <t>The "chunked" transfer-coding is always acceptable. If the
2395         keyword "trailers" is listed, the client indicates that it is
2396         willing to accept trailer fields in the chunked response on
2397         behalf of itself and any downstream clients. The implication is
2398         that, if given, the client is stating that either all
2399         downstream clients are willing to accept trailer fields in the
2400         forwarded response, or that it will attempt to buffer the
2401         response on behalf of downstream recipients.
2402      </t><t>
2403         <x:h>Note:</x:h> HTTP/1.1 does not define any means to limit the size of a
2404         chunked response such that a client can be assured of buffering
2405         the entire response.</t>
2406    </x:lt>
2407    <x:lt>
2408      <t>If the transfer-coding being tested is one of the transfer-codings
2409         listed in the TE field, then it is acceptable unless it
2410         is accompanied by a qvalue of 0. (As defined in &qvalue;, a
2411         qvalue of 0 means "not acceptable.")</t>
2412    </x:lt>
2413    <x:lt>
2414      <t>If multiple transfer-codings are acceptable, then the
2415         acceptable transfer-coding with the highest non-zero qvalue is
2416         preferred.  The "chunked" transfer-coding always has a qvalue
2417         of 1.</t>
2418    </x:lt>
2419  </list>
2420</t>
2421<t>
2422   If the TE field-value is empty or if no TE field is present, the only
2423   transfer-coding  is "chunked". A message with no transfer-coding is
2424   always acceptable.
2425</t>
2426</section>
2427
2428<section title="Trailer" anchor="header.trailer">
2429  <iref primary="true" item="Trailer header" x:for-anchor=""/>
2430  <iref primary="true" item="Headers" subitem="Trailer" x:for-anchor=""/>
2431<t>
2432   The Trailer general field value indicates that the given set of
2433   header fields is present in the trailer of a message encoded with
2434   chunked transfer-coding.
2435</t>
2436<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
2437    Trailer  = "Trailer" ":" 1#field-name
2438</artwork></figure>
2439<t>
2440   An HTTP/1.1 message &SHOULD; include a Trailer header field in a
2441   message using chunked transfer-coding with a non-empty trailer. Doing
2442   so allows the recipient to know which header fields to expect in the
2443   trailer.
2444</t>
2445<t>
2446   If no Trailer header field is present, the trailer &SHOULD-NOT;  include
2447   any header fields. See <xref target="chunked.transfer.encoding"/> for restrictions on the use of
2448   trailer fields in a "chunked" transfer-coding.
2449</t>
2450<t>
2451   Message header fields listed in the Trailer header field &MUST-NOT;
2452   include the following header fields:
2453  <list style="symbols">
2454    <t>Transfer-Encoding</t>
2455    <t>Content-Length</t>
2456    <t>Trailer</t>
2457  </list>
2458</t>
2459</section>
2460
2461<section title="Transfer-Encoding" anchor="header.transfer-encoding">
2462  <iref primary="true" item="Transfer-Encoding header" x:for-anchor=""/>
2463  <iref primary="true" item="Headers" subitem="Transfer-Encoding" x:for-anchor=""/>
2464<t>
2465   The Transfer-Encoding general-header field indicates what (if any)
2466   type of transformation has been applied to the message body in order
2467   to safely transfer it between the sender and the recipient. This
2468   differs from the content-coding in that the transfer-coding is a
2469   property of the message, not of the entity.
2470</t>
2471<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
2472  Transfer-Encoding       = "Transfer-Encoding" ":" 1#transfer-coding
2473</artwork></figure>
2474<t>
2475   Transfer-codings are defined in <xref target="transfer.codings"/>. An example is:
2476</t>
2477<figure><artwork type="example">
2478  Transfer-Encoding: chunked
2479</artwork></figure>
2480<t>
2481   If multiple encodings have been applied to an entity, the transfer-codings
2482   &MUST; be listed in the order in which they were applied.
2483   Additional information about the encoding parameters &MAY; be provided
2484   by other entity-header fields not defined by this specification.
2485</t>
2486<t>
2487   Many older HTTP/1.0 applications do not understand the Transfer-Encoding
2488   header.
2489</t>
2490</section>
2491
2492<section title="Upgrade" anchor="header.upgrade">
2493  <iref primary="true" item="Upgrade header" x:for-anchor=""/>
2494  <iref primary="true" item="Headers" subitem="Upgrade" x:for-anchor=""/>
2495<t>
2496   The Upgrade general-header allows the client to specify what
2497   additional communication protocols it supports and would like to use
2498   if the server finds it appropriate to switch protocols. The server
2499   &MUST; use the Upgrade header field within a 101 (Switching Protocols)
2500   response to indicate which protocol(s) are being switched.
2501</t>
2502<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
2503    Upgrade        = "Upgrade" ":" 1#product
2504</artwork></figure>
2505<t>
2506   For example,
2507</t>
2508<figure><artwork type="example">
2509    Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
2510</artwork></figure>
2511<t>
2512   The Upgrade header field is intended to provide a simple mechanism
2513   for transition from HTTP/1.1 to some other, incompatible protocol. It
2514   does so by allowing the client to advertise its desire to use another
2515   protocol, such as a later version of HTTP with a higher major version
2516   number, even though the current request has been made using HTTP/1.1.
2517   This eases the difficult transition between incompatible protocols by
2518   allowing the client to initiate a request in the more commonly
2519   supported protocol while indicating to the server that it would like
2520   to use a "better" protocol if available (where "better" is determined
2521   by the server, possibly according to the nature of the method and/or
2522   resource being requested).
2523</t>
2524<t>
2525   The Upgrade header field only applies to switching application-layer
2526   protocols upon the existing transport-layer connection. Upgrade
2527   cannot be used to insist on a protocol change; its acceptance and use
2528   by the server is optional. The capabilities and nature of the
2529   application-layer communication after the protocol change is entirely
2530   dependent upon the new protocol chosen, although the first action
2531   after changing the protocol &MUST; be a response to the initial HTTP
2532   request containing the Upgrade header field.
2533</t>
2534<t>
2535   The Upgrade header field only applies to the immediate connection.
2536   Therefore, the upgrade keyword &MUST; be supplied within a Connection
2537   header field (<xref target="header.connection"/>) whenever Upgrade is present in an
2538   HTTP/1.1 message.
2539</t>
2540<t>
2541   The Upgrade header field cannot be used to indicate a switch to a
2542   protocol on a different connection. For that purpose, it is more
2543   appropriate to use a 301, 302, 303, or 305 redirection response.
2544</t>
2545<t>
2546   This specification only defines the protocol name "HTTP" for use by
2547   the family of Hypertext Transfer Protocols, as defined by the HTTP
2548   version rules of <xref target="http.version"/> and future updates to this
2549   specification. Any token can be used as a protocol name; however, it
2550   will only be useful if both the client and server associate the name
2551   with the same protocol.
2552</t>
2553</section>
2554
2555<section title="Via" anchor="header.via">
2556  <iref primary="true" item="Via header" x:for-anchor=""/>
2557  <iref primary="true" item="Headers" subitem="Via" x:for-anchor=""/>
2558<t>
2559   The Via general-header field &MUST; be used by gateways and proxies to
2560   indicate the intermediate protocols and recipients between the user
2561   agent and the server on requests, and between the origin server and
2562   the client on responses. It is analogous to the "Received" field of
2563   RFC 822 <xref target="RFC822"/> and is intended to be used for tracking message forwards,
2564   avoiding request loops, and identifying the protocol capabilities of
2565   all senders along the request/response chain.
2566</t>
2567<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Via"/><iref primary="true" item="Grammar" subitem="received-protocol"/><iref primary="true" item="Grammar" subitem="protocol-name"/><iref primary="true" item="Grammar" subitem="protocol-version"/><iref primary="true" item="Grammar" subitem="received-by"/><iref primary="true" item="Grammar" subitem="pseudonym"/>
2568   Via =  "Via" ":" 1#( received-protocol received-by [ comment ] )
2569   received-protocol = [ protocol-name "/" ] protocol-version
2570   protocol-name     = token
2571   protocol-version  = token
2572   received-by       = ( host [ ":" port ] ) | pseudonym
2573   pseudonym         = token
2574</artwork></figure>
2575<t>
2576   The received-protocol indicates the protocol version of the message
2577   received by the server or client along each segment of the
2578   request/response chain. The received-protocol version is appended to
2579   the Via field value when the message is forwarded so that information
2580   about the protocol capabilities of upstream applications remains
2581   visible to all recipients.
2582</t>
2583<t>
2584   The protocol-name is optional if and only if it would be "HTTP". The
2585   received-by field is normally the host and optional port number of a
2586   recipient server or client that subsequently forwarded the message.
2587   However, if the real host is considered to be sensitive information,
2588   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
2589   be assumed to be the default port of the received-protocol.
2590</t>
2591<t>
2592   Multiple Via field values represents each proxy or gateway that has
2593   forwarded the message. Each recipient &MUST; append its information
2594   such that the end result is ordered according to the sequence of
2595   forwarding applications.
2596</t>
2597<t>
2598   Comments &MAY; be used in the Via header field to identify the software
2599   of the recipient proxy or gateway, analogous to the User-Agent and
2600   Server header fields. However, all comments in the Via field are
2601   optional and &MAY; be removed by any recipient prior to forwarding the
2602   message.
2603</t>
2604<t>
2605   For example, a request message could be sent from an HTTP/1.0 user
2606   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2607   forward the request to a public proxy at nowhere.com, which completes
2608   the request by forwarding it to the origin server at www.ics.uci.edu.
2609   The request received by www.ics.uci.edu would then have the following
2610   Via header field:
2611</t>
2612<figure><artwork type="example">
2613    Via: 1.0 fred, 1.1 nowhere.com (Apache/1.1)
2614</artwork></figure>
2615<t>
2616   Proxies and gateways used as a portal through a network firewall
2617   &SHOULD-NOT;, by default, forward the names and ports of hosts within
2618   the firewall region. This information &SHOULD; only be propagated if
2619   explicitly enabled. If not enabled, the received-by host of any host
2620   behind the firewall &SHOULD; be replaced by an appropriate pseudonym
2621   for that host.
2622</t>
2623<t>
2624   For organizations that have strong privacy requirements for hiding
2625   internal structures, a proxy &MAY; combine an ordered subsequence of
2626   Via header field entries with identical received-protocol values into
2627   a single such entry. For example,
2628</t>
2629<figure><artwork type="example">
2630    Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2631</artwork></figure>
2632<t>
2633        could be collapsed to
2634</t>
2635<figure><artwork type="example">
2636    Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2637</artwork></figure>
2638<t>
2639   Applications &SHOULD-NOT;  combine multiple entries unless they are all
2640   under the same organizational control and the hosts have already been
2641   replaced by pseudonyms. Applications &MUST-NOT; combine entries which
2642   have different received-protocol values.
2643</t>
2644</section>
2645
2646</section>
2647
2648<section title="IANA Considerations" anchor="IANA.considerations">
2649<t>
2650   TBD.
2651</t>
2652</section>
2653
2654<section title="Security Considerations" anchor="security.considerations">
2655<t>
2656   This section is meant to inform application developers, information
2657   providers, and users of the security limitations in HTTP/1.1 as
2658   described by this document. The discussion does not include
2659   definitive solutions to the problems revealed, though it does make
2660   some suggestions for reducing security risks.
2661</t>
2662
2663<section title="Personal Information" anchor="personal.information">
2664<t>
2665   HTTP clients are often privy to large amounts of personal information
2666   (e.g. the user's name, location, mail address, passwords, encryption
2667   keys, etc.), and &SHOULD; be very careful to prevent unintentional
2668   leakage of this information via the HTTP protocol to other sources.
2669   We very strongly recommend that a convenient interface be provided
2670   for the user to control dissemination of such information, and that
2671   designers and implementors be particularly careful in this area.
2672   History shows that errors in this area often create serious security
2673   and/or privacy problems and generate highly adverse publicity for the
2674   implementor's company.
2675</t>
2676</section>
2677
2678<section title="Abuse of Server Log Information" anchor="abuse.of.server.log.information">
2679<t>
2680   A server is in the position to save personal data about a user's
2681   requests which might identify their reading patterns or subjects of
2682   interest. This information is clearly confidential in nature and its
2683   handling can be constrained by law in certain countries. People using
2684   the HTTP protocol to provide data are responsible for ensuring that
2685   such material is not distributed without the permission of any
2686   individuals that are identifiable by the published results.
2687</t>
2688</section>
2689
2690<section title="Attacks Based On File and Path Names" anchor="attack.pathname">
2691<t>
2692   Implementations of HTTP origin servers &SHOULD; be careful to restrict
2693   the documents returned by HTTP requests to be only those that were
2694   intended by the server administrators. If an HTTP server translates
2695   HTTP URIs directly into file system calls, the server &MUST; take
2696   special care not to serve files that were not intended to be
2697   delivered to HTTP clients. For example, UNIX, Microsoft Windows, and
2698   other operating systems use ".." as a path component to indicate a
2699   directory level above the current one. On such a system, an HTTP
2700   server &MUST; disallow any such construct in the Request-URI if it
2701   would otherwise allow access to a resource outside those intended to
2702   be accessible via the HTTP server. Similarly, files intended for
2703   reference only internally to the server (such as access control
2704   files, configuration files, and script code) &MUST; be protected from
2705   inappropriate retrieval, since they might contain sensitive
2706   information. Experience has shown that minor bugs in such HTTP server
2707   implementations have turned into security risks.
2708</t>
2709</section>
2710
2711<section title="DNS Spoofing" anchor="dns.spoofing">
2712<t>
2713   Clients using HTTP rely heavily on the Domain Name Service, and are
2714   thus generally prone to security attacks based on the deliberate
2715   mis-association of IP addresses and DNS names. Clients need to be
2716   cautious in assuming the continuing validity of an IP number/DNS name
2717   association.
2718</t>
2719<t>
2720   In particular, HTTP clients &SHOULD; rely on their name resolver for
2721   confirmation of an IP number/DNS name association, rather than
2722   caching the result of previous host name lookups. Many platforms
2723   already can cache host name lookups locally when appropriate, and
2724   they &SHOULD; be configured to do so. It is proper for these lookups to
2725   be cached, however, only when the TTL (Time To Live) information
2726   reported by the name server makes it likely that the cached
2727   information will remain useful.
2728</t>
2729<t>
2730   If HTTP clients cache the results of host name lookups in order to
2731   achieve a performance improvement, they &MUST; observe the TTL
2732   information reported by DNS.
2733</t>
2734<t>
2735   If HTTP clients do not observe this rule, they could be spoofed when
2736   a previously-accessed server's IP address changes. As network
2737   renumbering is expected to become increasingly common <xref target="RFC1900"/>, the
2738   possibility of this form of attack will grow. Observing this
2739   requirement thus reduces this potential security vulnerability.
2740</t>
2741<t>
2742   This requirement also improves the load-balancing behavior of clients
2743   for replicated servers using the same DNS name and reduces the
2744   likelihood of a user's experiencing failure in accessing sites which
2745   use that strategy.
2746</t>
2747</section>
2748
2749<section title="Proxies and Caching" anchor="attack.proxies">
2750<t>
2751   By their very nature, HTTP proxies are men-in-the-middle, and
2752   represent an opportunity for man-in-the-middle attacks. Compromise of
2753   the systems on which the proxies run can result in serious security
2754   and privacy problems. Proxies have access to security-related
2755   information, personal information about individual users and
2756   organizations, and proprietary information belonging to users and
2757   content providers. A compromised proxy, or a proxy implemented or
2758   configured without regard to security and privacy considerations,
2759   might be used in the commission of a wide range of potential attacks.
2760</t>
2761<t>
2762   Proxy operators should protect the systems on which proxies run as
2763   they would protect any system that contains or transports sensitive
2764   information. In particular, log information gathered at proxies often
2765   contains highly sensitive personal information, and/or information
2766   about organizations. Log information should be carefully guarded, and
2767   appropriate guidelines for use developed and followed. (<xref target="abuse.of.server.log.information"/>).
2768</t>
2769<t>
2770   Proxy implementors should consider the privacy and security
2771   implications of their design and coding decisions, and of the
2772   configuration options they provide to proxy operators (especially the
2773   default configuration).
2774</t>
2775<t>
2776   Users of a proxy need to be aware that they are no trustworthier than
2777   the people who run the proxy; HTTP itself cannot solve this problem.
2778</t>
2779<t>
2780   The judicious use of cryptography, when appropriate, may suffice to
2781   protect against a broad range of security and privacy attacks. Such
2782   cryptography is beyond the scope of the HTTP/1.1 specification.
2783</t>
2784</section>
2785
2786<section title="Denial of Service Attacks on Proxies" anchor="attack.DoS">
2787<t>
2788   They exist. They are hard to defend against. Research continues.
2789   Beware.
2790</t>
2791</section>
2792</section>
2793
2794<section title="Acknowledgments" anchor="ack">
2795<t>
2796   This specification makes heavy use of the augmented BNF and generic
2797   constructs defined by David H. Crocker for RFC 822 <xref target="RFC822"/>. Similarly, it
2798   reuses many of the definitions provided by Nathaniel Borenstein and
2799   Ned Freed for MIME <xref target="RFC2045"/>. We hope that their inclusion in this
2800   specification will help reduce past confusion over the relationship
2801   between HTTP and Internet mail message formats.
2802</t>
2803<t>
2804   The HTTP protocol has evolved considerably over the years. It has
2805   benefited from a large and active developer community--the many
2806   people who have participated on the www-talk mailing list--and it is
2807   that community which has been most responsible for the success of
2808   HTTP and of the World-Wide Web in general. Marc Andreessen, Robert
2809   Cailliau, Daniel W. Connolly, Bob Denny, John Franks, Jean-Francois
2810   Groff, Phillip M. Hallam-Baker, Hakon W. Lie, Ari Luotonen, Rob
2811   McCool, Lou Montulli, Dave Raggett, Tony Sanders, and Marc
2812   VanHeyningen deserve special recognition for their efforts in
2813   defining early aspects of the protocol.
2814</t>
2815<t>
2816   This document has benefited greatly from the comments of all those
2817   participating in the HTTP-WG. In addition to those already mentioned,
2818   the following individuals have contributed to this specification:
2819</t>
2820<figure><artwork type="text/plain">
2821    Gary Adams                  Ross Patterson
2822    Harald Tveit Alvestrand     Albert Lunde
2823    Keith Ball                  John C. Mallery
2824    Brian Behlendorf            Jean-Philippe Martin-Flatin
2825    Paul Burchard               Mitra
2826    Maurizio Codogno            David Morris
2827    Mike Cowlishaw              Gavin Nicol
2828    Roman Czyborra              Bill Perry
2829    Michael A. Dolan            Jeffrey Perry
2830    David J. Fiander            Scott Powers
2831    Alan Freier                 Owen Rees
2832    Marc Hedlund                Luigi Rizzo
2833    Greg Herlihy                David Robinson
2834    Koen Holtman                Marc Salomon
2835    Alex Hopmann                Rich Salz
2836    Bob Jernigan                Allan M. Schiffman
2837    Shel Kaphan                 Jim Seidman
2838    Rohit Khare                 Chuck Shotton
2839    John Klensin                Eric W. Sink
2840    Martijn Koster              Simon E. Spero
2841    Alexei Kosut                Richard N. Taylor
2842    David M. Kristol            Robert S. Thau
2843    Daniel LaLiberte            Bill (BearHeart) Weinman
2844    Ben Laurie                  Francois Yergeau
2845    Paul J. Leach               Mary Ellen Zurko
2846    Daniel DuBois               Josh Cohen
2847</artwork></figure>
2848<t>
2849   Based on an XML translation of RFC 2616 by Julian Reschke.
2850</t>
2851</section>
2852
2853</middle>
2854<back>
2855<references>
2856
2857<reference anchor="RFC1436">
2858<front>
2859<title abbrev="Gopher">The Internet Gopher Protocol (a distributed document search and retrieval protocol)</title>
2860<author initials="F." surname="Anklesaria" fullname="Farhad Anklesaria">
2861<organization>University of Minnesota, Computer and Information Services</organization>
2862<address>
2863<postal>
2864<street>100 Union Street SE</street>
2865<street>Room 152 Shepherd Labs</street>
2866<city>Minneapolis</city>
2867<region>MN</region>
2868<code>55455</code>
2869<country>US</country></postal>
2870<phone>+1 612 625 1300</phone>
2871<email>fxa@boombox.micro.umn.edu</email></address></author>
2872<author initials="M." surname="McCahill" fullname="Mark McCahill">
2873<organization>University of Minnesota, Computer and Information Services</organization>
2874<address>
2875<postal>
2876<street>100 Union Street SE</street>
2877<street>Room 152 Shepherd Labs</street>
2878<city>Minneapolis</city>
2879<region>MN</region>
2880<code>55455</code>
2881<country>US</country></postal>
2882<phone>+1 612 625 1300</phone>
2883<email>mpm@boombox.micro.umn.edu</email></address></author>
2884<author initials="P." surname="Lindner" fullname="Paul Lindner">
2885<organization>University of Minnesota, Computer and Information Services</organization>
2886<address>
2887<postal>
2888<street>100 Union Street SE</street>
2889<street>Room 152 Shepherd Labs</street>
2890<city>Minneapolis</city>
2891<region>MN</region>
2892<code>55455</code>
2893<country>US</country></postal>
2894<phone>+1 612 625 1300</phone>
2895<email>lindner@boombox.micro.umn.edu</email></address></author>
2896<author initials="D." surname="Johnson" fullname="David Johnson">
2897<organization>University of Minnesota, Computer and Information Services</organization>
2898<address>
2899<postal>
2900<street>100 Union Street SE</street>
2901<street>Room 152 Shepherd Labs</street>
2902<city>Minneapolis</city>
2903<region>MN</region>
2904<code>55455</code>
2905<country>US</country></postal>
2906<phone>+1 612 625 1300</phone>
2907<email>dmj@boombox.micro.umn.edu</email></address></author>
2908<author initials="D." surname="Torrey" fullname="Daniel Torrey">
2909<organization>University of Minnesota, Computer and Information Services</organization>
2910<address>
2911<postal>
2912<street>100 Union Street SE</street>
2913<street>Room 152 Shepherd Labs</street>
2914<city>Minneapolis</city>
2915<region>MN</region>
2916<code>55455</code>
2917<country>US</country></postal>
2918<phone>+1 612 625 1300</phone>
2919<email>daniel@boombox.micro.umn.edu</email></address></author>
2920<author initials="B." surname="Alberti" fullname="Bob Alberti">
2921<organization>University of Minnesota, Computer and Information Services</organization>
2922<address>
2923<postal>
2924<street>100 Union Street SE</street>
2925<street>Room 152 Shepherd Labs</street>
2926<city>Minneapolis</city>
2927<region>MN</region>
2928<code>55455</code>
2929<country>US</country></postal>
2930<phone>+1 612 625 1300</phone>
2931<email>alberti@boombox.micro.umn.edu</email></address></author>
2932<date month="March" year="1993"/>
2933</front>
2934<seriesInfo name="RFC" value="1436"/>
2935</reference>
2936
2937<reference anchor="RFC1630">
2938<front>
2939<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>
2940<author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
2941<organization>CERN, World-Wide Web project</organization>
2942<address>
2943<postal>
2944<street>1211 Geneva 23</street>
2945<city/>
2946<region/>
2947<code/>
2948<country>CH</country></postal>
2949<phone>+41 22 7673755</phone>
2950<facsimile>+41 22 7677155</facsimile>
2951<email>timbl@info.cern.ch</email></address></author>
2952<date month="June" year="1994"/></front>
2953<seriesInfo name="RFC" value="1630"/>
2954</reference>
2955
2956<reference anchor="RFC1738">
2957<front>
2958<title>Uniform Resource Locators (URL)</title>
2959<author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
2960<organization>CERN, World-Wide Web project</organization>
2961<address>
2962<postal>
2963<street>1211 Geneva 23</street>
2964<city/>
2965<region/>
2966<code/>
2967<country>CH</country></postal>
2968<phone>+41 22 7673755</phone>
2969<facsimile>+41 22 7677155</facsimile>
2970<email>timbl@info.cern.ch</email></address></author>
2971<author initials="L." surname="Masinter" fullname="Larry Masinter">
2972<organization>Xerox PARC</organization>
2973<address>
2974<postal>
2975<street>3333 Coyote Hill Road</street>
2976<city>Palo Alto</city>
2977<region>CA</region>
2978<code>94034</code>
2979<country>US</country></postal>
2980<phone>+1 415 812 4365</phone>
2981<facsimile>+1 415 812 4333</facsimile>
2982<email>masinter@parc.xerox.com</email></address></author>
2983<author initials="M." surname="McCahill" fullname="Mark McCahill">
2984<organization>University of Minnesota, Computer and Information Services</organization>
2985<address>
2986<postal>
2987<street>100 Union Street SE, Shepherd Labs</street>
2988<street>Room 152</street>
2989<city>Minneapolis</city>
2990<region>MN</region>
2991<code>55455</code>
2992<country>US</country></postal>
2993<phone>+1 612 625 1300</phone>
2994<email>mpm@boombox.micro.umn.edu</email></address></author>
2995<date month="December" year="1994"/>
2996</front>
2997<seriesInfo name="RFC" value="1738"/>
2998</reference>
2999
3000<reference anchor="RFC1945">
3001<front>
3002<title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
3003<author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3004<organization>MIT, Laboratory for Computer Science</organization>
3005<address>
3006<postal>
3007<street>545 Technology Square</street>
3008<city>Cambridge</city>
3009<region>MA</region>
3010<code>02139</code>
3011<country>US</country></postal>
3012<phone/>
3013<facsimile>+1 617 258 8682</facsimile>
3014<email>timbl@w3.org</email></address></author>
3015<author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3016<organization>University of California, Irvine, Department of Information and Computer Science</organization>
3017<address>
3018<postal>
3019<street/>
3020<city>Irvine</city>
3021<region>CA</region>
3022<code>92717-3425</code>
3023<country>US</country></postal>
3024<phone/>
3025<facsimile>+1 714 824 4056</facsimile>
3026<email>fielding@ics.uci.edu</email></address></author>
3027<author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3028<organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
3029<address>
3030<postal>
3031<street>545 Technology Square</street>
3032<city>Cambridge</city>
3033<region>MA</region>
3034<code>02139</code>
3035<country>US</country></postal>
3036<phone/>
3037<facsimile>+1 617 258 8682</facsimile>
3038<email>frystyk@w3.org</email></address></author>
3039<date month="May" year="1996"/>
3040</front>
3041<seriesInfo name="RFC" value="1945"/>
3042</reference>
3043
3044<reference anchor="RFC2045">
3045<front>
3046<title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
3047<author initials="N." surname="Freed" fullname="Ned Freed">
3048<organization>Innosoft International, Inc.</organization>
3049<address>
3050<postal>
3051<street>1050 East Garvey Avenue South</street>
3052<city>West Covina</city>
3053<region>CA</region>
3054<code>91790</code>
3055<country>US</country></postal>
3056<phone>+1 818 919 3600</phone>
3057<facsimile>+1 818 919 3614</facsimile>
3058<email>ned@innosoft.com</email></address></author>
3059<author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
3060<organization>First Virtual Holdings</organization>
3061<address>
3062<postal>
3063<street>25 Washington Avenue</street>
3064<city>Morristown</city>
3065<region>NJ</region>
3066<code>07960</code>
3067<country>US</country></postal>
3068<phone>+1 201 540 8967</phone>
3069<facsimile>+1 201 993 3032</facsimile>
3070<email>nsb@nsb.fv.com</email></address></author>
3071<date month="November" year="1996"/>
3072</front>
3073<seriesInfo name="RFC" value="2045"/>
3074</reference>
3075
3076<reference anchor="RFC1123">
3077<front>
3078<title>Requirements for Internet Hosts - Application and Support</title>
3079<author initials="R." surname="Braden" fullname="Robert Braden">
3080<organization>University of Southern California (USC), Information Sciences Institute</organization>
3081<address>
3082<postal>
3083<street>4676 Admiralty Way</street>
3084<city>Marina del Rey</city>
3085<region>CA</region>
3086<code>90292-6695</code>
3087<country>US</country></postal>
3088<phone>+1 213 822 1511</phone>
3089<email>Braden@ISI.EDU</email></address></author>
3090<date month="October" year="1989"/></front>
3091<seriesInfo name="STD" value="3"/>
3092<seriesInfo name="RFC" value="1123"/>
3093</reference>
3094
3095<reference anchor="RFC822">
3096<front>
3097<title abbrev="Standard for ARPA Internet Text Messages">Standard for the format of ARPA Internet text messages</title>
3098<author initials="D.H." surname="Crocker" fullname="David H. Crocker">
3099<organization>University of Delaware, Dept. of Electrical Engineering</organization>
3100<address>
3101<postal>
3102<street/>
3103<city>Newark</city>
3104<region>DE</region>
3105<code>19711</code>
3106<country>US</country></postal>
3107<email>DCrocker@UDel-Relay</email></address></author>
3108<date month="August" day="13" year="1982"/></front>
3109<seriesInfo name="STD" value="11"/>
3110<seriesInfo name="RFC" value="822"/>
3111</reference>
3112
3113<reference anchor="WAIS">
3114<front>
3115<title>WAIS Interface Protocol Prototype Functional Specification (v1.5)</title>
3116<author initials="F." surname="Davis" fullname="F. Davis">
3117<organization>Thinking Machines Corporation</organization>
3118</author>
3119<author initials="B." surname="Kahle" fullname="B. Kahle">
3120<organization>Thinking Machines Corporation</organization>
3121</author>
3122<author initials="H." surname="Morris" fullname="H. Morris">
3123<organization>Thinking Machines Corporation</organization>
3124</author>
3125<author initials="J." surname="Salem" fullname="J. Salem">
3126<organization>Thinking Machines Corporation</organization>
3127</author>
3128<author initials="T." surname="Shen" fullname="T. Shen">
3129<organization>Thinking Machines Corporation</organization>
3130</author>
3131<author initials="R." surname="Wang" fullname="R. Wang">
3132<organization>Thinking Machines Corporation</organization>
3133</author>
3134<author initials="J." surname="Sui" fullname="J. Sui">
3135<organization>Thinking Machines Corporation</organization>
3136</author>
3137<author initials="M." surname="Grinbaum" fullname="M. Grinbaum">
3138<organization>Thinking Machines Corporation</organization>
3139</author>
3140<date month="April" year="1990"/></front>
3141<seriesInfo name="Thinking Machines Corporation" value=""/>
3142</reference>
3143
3144<reference anchor="RFC1808">
3145<front>
3146<title>Relative Uniform Resource Locators</title>
3147<author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3148<organization>University of California Irvine, Department of Information and Computer Science</organization>
3149<address>
3150<postal>
3151<street/>
3152<city>Irvine</city>
3153<region>CA</region>
3154<code>92717-3425</code>
3155<country>US</country></postal>
3156<phone>+1 714 824 4049</phone>
3157<facsimile>+1 714 824 4056</facsimile>
3158<email>fielding@ics.uci.edu</email></address></author>
3159<date month="June" year="1995"/>
3160</front>
3161<seriesInfo name="RFC" value="1808"/>
3162</reference>
3163
3164<reference anchor="RFC1036">
3165<front>
3166<title abbrev="Standard for USENET Messages">Standard for interchange of USENET messages</title>
3167<author initials="M." surname="Horton" fullname="M. Horton">
3168<organization>AT&amp;amp;T Bell Laboratories</organization></author>
3169<author initials="R." surname="Adams" fullname="R. Adams">
3170<organization>Center for Seismic Studies</organization></author>
3171<date month="December" year="1987"/></front>
3172<seriesInfo name="RFC" value="1036"/>
3173</reference>
3174
3175<reference anchor="RFC977">
3176<front>
3177<title abbrev="Network News Transfer Protocol">Network News Transfer Protocol</title>
3178<author initials="B." surname="Kantor" fullname="Brian Kantor">
3179<organization>University of California, San Diego (UCSD)</organization></author>
3180<author initials="P." surname="Lapsley" fullname="Phil Lapsley">
3181<organization>University of California, Berkeley (UCB)</organization></author>
3182<date month="February" year="1986"/></front>
3183<seriesInfo name="RFC" value="977"/>
3184</reference>
3185
3186<reference anchor="RFC2047">
3187<front>
3188<title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
3189<author initials="K." surname="Moore" fullname="Keith Moore">
3190<organization>University of Tennessee</organization>
3191<address>
3192<postal>
3193<street>107 Ayres Hall</street>
3194<street>Knoxville TN 37996-1301</street></postal>
3195<email>moore@cs.utk.edu</email></address></author>
3196<date month="November" year="1996"/>
3197<area>Applications</area>
3198<keyword>Amercian Standard Code for Information Interchange</keyword>
3199<keyword>mail</keyword>
3200<keyword>multipurpose internet mail extensions</keyword>
3201</front>
3202<seriesInfo name="RFC" value="2047"/>
3203</reference>
3204
3205<reference anchor="RFC821">
3206<front>
3207<title>Simple Mail Transfer Protocol</title>
3208<author initials="J.B." surname="Postel" fullname="Jonathan B. Postel">
3209<organization>University of Southern California (USC)/Information Sciences Institute</organization>
3210<address>
3211<postal>
3212<street>4676 Admiralty Way</street>
3213<city>Marina del Rey</city>
3214<region>CA</region>
3215<code>90291</code>
3216<country>US</country></postal>
3217<phone>+1 213 822 1511</phone></address></author>
3218<date month="August" year="1982"/></front>
3219<seriesInfo name="STD" value="10"/>
3220<seriesInfo name="RFC" value="821"/>
3221</reference>
3222
3223<reference anchor="RFC1590">
3224<front>
3225<title>Media Type Registration Procedure</title>
3226<author initials="J." surname="Postel" fullname="Jon Postel">
3227<organization>USC/Information Sciences Institute</organization>
3228<address>
3229<postal>
3230<street>4676 Admiralty Way</street>
3231<city>Marina del Rey</city>
3232<region>CA</region>
3233<code>90292</code>
3234<country>US</country></postal>
3235<phone>+1 310 822 1511</phone>
3236<facsimile>+1 310 823 6714</facsimile>
3237<email>Postel@ISI.EDU</email></address></author>
3238<date month="November" year="1996"/>
3239</front>
3240<seriesInfo name="RFC" value="1590"/>
3241</reference>
3242
3243<reference anchor="RFC959">
3244<front>
3245<title abbrev="File Transfer Protocol">File Transfer Protocol</title>
3246<author initials="J." surname="Postel" fullname="J. Postel">
3247<organization>Information Sciences Institute (ISI)</organization></author>
3248<author initials="J." surname="Reynolds" fullname="J. Reynolds">
3249<organization/></author>
3250<date month="October" year="1985"/></front>
3251<seriesInfo name="STD" value="9"/>
3252<seriesInfo name="RFC" value="959"/>
3253</reference>
3254
3255<reference anchor="RFC1700">
3256<front>
3257<title abbrev="Assigned Numbers">Assigned Numbers</title>
3258<author initials="J." surname="Reynolds" fullname="Joyce K. Reynolds">
3259<organization>USC/Information Sciences Institute</organization>
3260<address>
3261<postal>
3262<street>4676 Admiralty Way</street>
3263<city>Marina del Rey</city>
3264<region>CA</region>
3265<code>90292-6695</code>
3266<country>US</country></postal>
3267<phone>+1 310 822 1511</phone>
3268<email>jkrey@isi.edu</email></address></author>
3269<author initials="J." surname="Postel" fullname="Jon Postel">
3270<organization>USC/Information Sciences Institute</organization>
3271<address>
3272<postal>
3273<street>4676 Admiralty Way</street>
3274<city>Marina del Rey</city>
3275<region>CA</region>
3276<code>90292-6695</code>
3277<country>US</country></postal>
3278<phone>+1 310 822 1511</phone>
3279<email>postel@isi.edu</email></address></author>
3280<date month="October" year="1994"/>
3281<abstract>
3282<t/></abstract></front>
3283<seriesInfo name="STD" value="2"/>
3284<seriesInfo name="RFC" value="1700"/>
3285</reference>
3286
3287<reference anchor="RFC1737">
3288<front>
3289<title abbrev="Requirements for Uniform Resource Names">Functional Requirements for Uniform Resource Names</title>
3290<author initials="L." surname="Masinter" fullname="Larry Masinter">
3291<organization>Xerox Palo Alto Research Center</organization>
3292<address>
3293<postal>
3294<street>3333 Coyote Hill Road</street>
3295<city>Palo Alto</city>
3296<region>CA</region>
3297<code>94304</code>
3298<country>US</country></postal>
3299<phone>+1 415 812 4365</phone>
3300<facsimile>+1 415 812 4333</facsimile>
3301<email>masinter@parc.xerox.com</email></address></author>
3302<author initials="K." surname="Sollins" fullname="Karen Sollins">
3303<organization>MIT Laboratory for Computer Science</organization>
3304<address>
3305<postal>
3306<street>545 Technology Square</street>
3307<city>Cambridge</city>
3308<region>MA</region>
3309<code>02139</code>
3310<country>US</country></postal>
3311<phone>+1 617 253 2673</phone>
3312<email>sollins@lcs.mit.edu</email></address></author>
3313<date month="December" year="1994"/></front>
3314<seriesInfo name="RFC" value="1737"/>
3315</reference>
3316
3317<reference anchor="USASCII">
3318  <front>
3319    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
3320    <author>
3321      <organization>American National Standards Institute</organization>
3322    </author>
3323    <date year="1986"/>
3324  </front>
3325  <seriesInfo name="ANSI" value="X3.4"/>
3326</reference>
3327
3328<reference anchor="ISO-8859">
3329<front>
3330<title>Information technology - 8-bit single byte coded graphic - character sets</title>
3331<author>
3332<organization>International Organization for Standardization</organization>
3333</author>
3334<date year="1987-1990"/>
3335</front>
3336<annotation>
3337  Part 1: Latin alphabet No. 1, ISO-8859-1:1987.
3338  Part 2: Latin alphabet No. 2, ISO-8859-2, 1987.
3339  Part 3: Latin alphabet No. 3, ISO-8859-3, 1988.
3340  Part 4: Latin alphabet No. 4, ISO-8859-4, 1988.
3341  Part 5: Latin/Cyrillic alphabet, ISO-8859-5, 1988.
3342  Part 6: Latin/Arabic alphabet, ISO-8859-6, 1987.
3343  Part 7: Latin/Greek alphabet, ISO-8859-7, 1987.
3344  Part 8: Latin/Hebrew alphabet, ISO-8859-8, 1988.
3345  Part 9: Latin alphabet No. 5, ISO-8859-9, 1990.
3346</annotation>
3347</reference>
3348
3349<reference anchor="RFC1900">
3350<front>
3351<title>Renumbering Needs Work</title>
3352<author initials="B." surname="Carpenter" fullname="Brian E. Carpenter">
3353<organization>CERN, Computing and Networks Division</organization>
3354<address>
3355<postal>
3356<street>1211 Geneva 23</street>
3357<country>CH</country></postal>
3358<phone>+41 22 7674967</phone>
3359<facsimile>+41 22 7677155</facsimile>
3360<email>brian@dxcoms.cern.ch</email></address></author>
3361<author initials="Y." surname="Rekhter" fullname="Yakov Rekhter">
3362<organization>cisco Systems</organization>
3363<address>
3364<postal>
3365<street>170 West Tasman Drive</street>
3366<city>San Jose</city>
3367<region>CA</region>
3368<code>95134</code>
3369<country>US</country></postal>
3370<phone>+1 914 528 0090</phone>
3371<email>yakov@cisco.com</email></address></author>
3372<date month="February" year="1996"/>
3373</front>
3374<seriesInfo name="RFC" value="1900"/>
3375</reference>
3376
3377<reference anchor="Pad1995">
3378<front>
3379<title>Improving HTTP Latency</title>
3380<author initials="V.N." surname="Padmanabhan" fullname="Venkata N. Padmanabhan"><organization/></author>
3381<author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul"><organization/></author>
3382<date year="1995" month="Dec"/>
3383</front>
3384<seriesInfo name="Computer Networks and ISDN Systems" value="v. 28, pp. 25-35"/>
3385<annotation>
3386  Slightly revised version of paper in Proc. 2nd International WWW Conference '94: Mosaic and the Web, Oct. 1994,
3387  which is available at <eref target="http://www.ncsa.uiuc.edu/SDG/IT94/Proceedings/DDay/mogul/HTTPLatency.html"/>.
3388</annotation>
3389</reference>
3390
3391<reference anchor="Tou1998" target="http://www.isi.edu/touch/pubs/http-perf96/">
3392<front>
3393<title>Analysis of HTTP Performance</title>
3394<author initials="J." surname="Touch" fullname="Joe Touch">
3395  <organization>USC/Information Sciences Institute</organization>
3396  <address><email>touch@isi.edu</email></address>
3397</author>
3398<author initials="J." surname="Heidemann" fullname="John Heidemann">
3399  <organization>USC/Information Sciences Institute</organization>
3400  <address><email>johnh@isi.edu</email></address>
3401</author>
3402<author initials="K." surname="Obraczka" fullname="Katia Obraczka">
3403  <organization>USC/Information Sciences Institute</organization>
3404  <address><email>katia@isi.edu</email></address>
3405</author>
3406<date year="1998" month="Aug"/>
3407</front>
3408<seriesInfo name="ISI Research Report" value="ISI/RR-98-463 (original report dated Aug.1996)"/>
3409</reference>
3410
3411<reference anchor="RFC1305">
3412<front>
3413<title>Network Time Protocol (Version 3) Specification, Implementation</title>
3414<author initials="D." surname="Mills" fullname="David L. Mills">
3415<organization>University of Delaware, Electrical Engineering Department</organization>
3416<address>
3417<postal>
3418<street/>
3419<city>Newark</city>
3420<region>DE</region>
3421<code>19716</code>
3422<country>US</country></postal>
3423<phone>+1 302 451 8247</phone>
3424<email>mills@udel.edu</email></address></author>
3425<date month="March" year="1992"/>
3426<abstract>
3427<t>This document describes the Network Time Protocol (NTP), specifies its normal structure and summarizes information useful for its implementation. NTP provides the mechanisms to synchronize time and coordinate time distribution in a large, diverse internet operating at rates from mundane to lightwave. It uses a returnable-time design in which a distributed subnet of time servers operating in a self-organizing, hierarchical-master-slave configuration synchronizes local clocks within the subnet and to national time standards via wire or radio. The servers can also redistribute reference time via local routing algorithms and time daemons.</t></abstract></front>
3428<seriesInfo name="RFC" value="1305"/>
3429</reference>
3430
3431<reference anchor="Spe" target="http://sunsite.unc.edu/mdma-release/http-prob.html">
3432<front>
3433<title>Analysis of HTTP Performance Problems</title>
3434<author initials="S." surname="Spero" fullname="Simon E. Spero">
3435  <organization/>
3436</author>
3437<date/>
3438</front>
3439</reference>
3440
3441<reference anchor="RFC2068">
3442<front>
3443<title abbrev="HTTP/1.1">Hypertext Transfer Protocol -- HTTP/1.1</title>
3444<author initials="R." surname="Fielding" fullname="Roy T. Fielding">
3445<organization>University of California, Irvine, Department of Information and Computer Science</organization>
3446<address>
3447<postal>
3448<street/>
3449<city>Irvine</city>
3450<region>CA</region>
3451<code>92717-3425</code>
3452<country>US</country></postal>
3453<facsimile>+1 714 824 4056</facsimile>
3454<email>fielding@ics.uci.edu</email></address></author>
3455<author initials="J." surname="Gettys" fullname="Jim Gettys">
3456<organization>MIT Laboratory for Computer Science</organization>
3457<address>
3458<postal>
3459<street>545 Technology Square</street>
3460<city>Cambridge</city>
3461<region>MA</region>
3462<code>02139</code>
3463<country>US</country></postal>
3464<facsimile>+1 617 258 8682</facsimile>
3465<email>jg@w3.org</email></address></author>
3466<author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
3467<organization>Digital Equipment Corporation, Western Research Laboratory</organization>
3468<address>
3469<postal>
3470<street>250 University Avenue</street>
3471<city>Palo Alto</city>
3472<region>CA</region>
3473<code>94301</code>
3474<country>US</country></postal>
3475<email>mogul@wrl.dec.com</email></address></author>
3476<author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3477<organization>MIT Laboratory for Computer Science</organization>
3478<address>
3479<postal>
3480<street>545 Technology Square</street>
3481<city>Cambridge</city>
3482<region>MA</region>
3483<code>02139</code>
3484<country>US</country></postal>
3485<facsimile>+1 617 258 8682</facsimile>
3486<email>frystyk@w3.org</email></address></author>
3487<author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3488<organization>MIT Laboratory for Computer Science</organization>
3489<address>
3490<postal>
3491<street>545 Technology Square</street>
3492<city>Cambridge</city>
3493<region>MA</region>
3494<code>02139</code>
3495<country>US</country></postal>
3496<facsimile>+1 617 258 8682</facsimile>
3497<email>timbl@w3.org</email></address></author>
3498<date month="January" year="1997"/>
3499<abstract>
3500<t>The Hypertext Transfer Protocol (HTTP) is an application-level protocol for distributed, collaborative, hypermedia information systems. It is a generic, stateless, object-oriented protocol which can be used for many tasks, such as name servers and distributed object management systems, through extension of its request methods. A feature of HTTP is the typing and negotiation of data representation, allowing systems to be built independently of the data being transferred.</t>
3501<t>HTTP has been in use by the World-Wide Web global information initiative since 1990. This specification defines the protocol referred to as "HTTP/1.1".</t></abstract></front>
3502<seriesInfo name="RFC" value="2068"/>
3503</reference>
3504
3505<reference anchor="RFC2119">
3506<front>
3507<title abbrev="RFC Key Words">Key words for use in RFCs to Indicate Requirement Levels</title>
3508<author initials="S." surname="Bradner" fullname="Scott Bradner">
3509<organization>Harvard University</organization>
3510<address>
3511<postal>
3512<street>1350 Mass. Ave.</street>
3513<street>Cambridge</street>
3514<street>MA 02138</street></postal>
3515<phone>- +1 617 495 3864</phone>
3516<email>-</email></address></author>
3517<date month="March" year="1997"/>
3518<area>General</area>
3519<keyword>keyword</keyword>
3520</front>
3521
3522<seriesInfo name="BCP" value="14"/>
3523<seriesInfo name="RFC" value="2119"/>
3524</reference>
3525
3526<reference anchor="RFC2145">
3527<front>
3528<title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
3529<author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
3530<organization>Western Research Laboratory</organization>
3531<address>
3532<postal>
3533<street>Digital Equipment Corporation</street>
3534<street>250 University Avenue</street>
3535<street>Palo Alto</street>
3536<street>California</street>
3537<street>94305</street>
3538<country>USA</country></postal>
3539<email>mogul@wrl.dec.com</email></address></author>
3540<author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3541<organization>Department of Information and Computer Science</organization>
3542<address>
3543<postal>
3544<street>University of California</street>
3545<street>Irvine</street>
3546<street>CA 92717-3425</street>
3547<country>USA</country></postal>
3548<facsimile>+1 (714) 824-4056</facsimile>
3549<email>fielding@ics.uci.edu</email></address></author>
3550<author initials="J." surname="Gettys" fullname="Jim Gettys">
3551<organization>MIT Laboratory for Computer Science</organization>
3552<address>
3553<postal>
3554<street>545 Technology Square</street>
3555<street>Cambridge</street>
3556<street>MA 02139</street>
3557<country>USA</country></postal>
3558<facsimile>+1 (617) 258 8682</facsimile>
3559<email>jg@w3.org</email></address></author>
3560<author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
3561<organization>W3 Consortium</organization>
3562<address>
3563<postal>
3564<street>MIT Laboratory for Computer Science</street>
3565<street>545 Technology Square</street>
3566<street>Cambridge</street>
3567<street>MA 02139</street>
3568<country>USA</country></postal>
3569<facsimile>+1 (617) 258 8682</facsimile>
3570<email>frystyk@w3.org</email></address></author>
3571<date month="May" year="1997"/>
3572<area>Applications</area>
3573<keyword>HTTP</keyword>
3574<keyword>hypertext transfer protocol</keyword>
3575<abstract>
3576<t>
3577   HTTP request and response messages include an HTTP protocol version
3578   number.  Some confusion exists concerning the proper use and
3579   interpretation of HTTP version numbers, and concerning
3580   interoperability of HTTP implementations of different protocol
3581   versions.  This document is an attempt to clarify the situation.  It
3582   is not a modification of the intended meaning of the existing
3583   HTTP/1.0 and HTTP/1.1 documents, but it does describe the intention
3584   of the authors of those documents, and can be considered definitive
3585   when there is any ambiguity in those documents concerning HTTP
3586   version numbers, for all versions of HTTP.
3587</t></abstract></front>
3588<seriesInfo name="RFC" value="2145"/>
3589</reference>
3590
3591<reference anchor="Nie1997">
3592<front>
3593<title>Network Performance Effects of HTTP/1.1, CSS1, and PNG</title>
3594<author initials="H.F.." surname="Nielsen" fullname="H.F. Nielsen">
3595  <organization/>
3596</author>
3597<author initials="J." surname="Gettys" fullname="J. Gettys">
3598  <organization/>
3599</author>
3600<author initials="E." surname="Prud'hommeaux" fullname="E. Prud'hommeaux">
3601  <organization/>
3602</author>
3603<author initials="H." surname="Lie" fullname="H. Lie">
3604  <organization/>
3605</author>
3606<author initials="C." surname="Lilley" fullname="C. Lilley">
3607  <organization/>
3608</author>
3609<date year="1997" month="Sep"/>
3610</front>
3611<seriesInfo name="Proceedings of ACM SIGCOMM '97, Cannes France" value=""/>
3612</reference>
3613
3614<reference anchor="RFC2396">
3615<front>
3616<title abbrev="URI Generic Syntax">Uniform Resource Identifiers (URI): Generic Syntax</title>
3617<author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
3618<organization abbrev="MIT/LCS">World Wide Web Consortium</organization>
3619<address>
3620<postal>
3621<street>MIT Laboratory for Computer Science, NE43-356</street>
3622<street>545 Technology Square</street>
3623<city>Cambridge</city>
3624<region>MA</region>
3625<code>02139</code></postal>
3626<facsimile>+1(617)258-8682</facsimile>
3627<email>timbl@w3.org</email></address></author>
3628<author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
3629<organization abbrev="U.C. Irvine">Department of Information and Computer Science</organization>
3630<address>
3631<postal>
3632<street>University of California, Irvine</street>
3633<city>Irvine</city>
3634<region>CA</region>
3635<code>92697-3425</code></postal>
3636<facsimile>+1(949)824-1715</facsimile>
3637<email>fielding@ics.uci.edu</email></address></author>
3638<author initials="L." surname="Masinter" fullname="Larry Masinter">
3639<organization abbrev="Xerox Corporation">Xerox PARC</organization>
3640<address>
3641<postal>
3642<street>3333 Coyote Hill Road</street>
3643<city>Palo Alto</city>
3644<region>CA</region>
3645<code>94034</code></postal>
3646<facsimile>+1(415)812-4333</facsimile>
3647<email>masinter@parc.xerox.com</email></address></author>
3648<date month="August" year="1998"/>
3649<area>Applications</area>
3650<keyword>uniform resource</keyword>
3651<keyword>URI</keyword>
3652</front>
3653<seriesInfo name="RFC" value="2396"/>
3654</reference>
3655
3656<reference anchor="RFC2616">
3657  <front>
3658    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
3659    <author initials="R." surname="Fielding" fullname="R. Fielding">
3660      <organization>University of California, Irvine</organization>
3661      <address><email>fielding@ics.uci.edu</email></address>
3662    </author>
3663    <author initials="J." surname="Gettys" fullname="J. Gettys">
3664      <organization>W3C</organization>
3665      <address><email>jg@w3.org</email></address>
3666    </author>
3667    <author initials="J." surname="Mogul" fullname="J. Mogul">
3668      <organization>Compaq Computer Corporation</organization>
3669      <address><email>mogul@wrl.dec.com</email></address>
3670    </author>
3671    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
3672      <organization>MIT Laboratory for Computer Science</organization>
3673      <address><email>frystyk@w3.org</email></address>
3674    </author>
3675    <author initials="L." surname="Masinter" fullname="L. Masinter">
3676      <organization>Xerox Corporation</organization>
3677      <address><email>masinter@parc.xerox.com</email></address>
3678    </author>
3679    <author initials="P." surname="Leach" fullname="P. Leach">
3680      <organization>Microsoft Corporation</organization>
3681      <address><email>paulle@microsoft.com</email></address>
3682    </author>
3683    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
3684      <organization>W3C</organization>
3685      <address><email>timbl@w3.org</email></address>
3686    </author>
3687    <date month="June" year="1999"/>
3688  </front>
3689  <seriesInfo name="RFC" value="2616"/>
3690</reference>
3691
3692<reference anchor="RFC2617">
3693<front>
3694<title abbrev="HTTP Authentication">HTTP Authentication: Basic and Digest Access Authentication</title>
3695<author initials="J." surname="Franks" fullname="John Franks">
3696<organization>Northwestern University, Department of Mathematics</organization>
3697<address>
3698<postal>
3699<street/>
3700<city>Evanston</city>
3701<region>IL</region>
3702<code>60208-2730</code>
3703<country>US</country></postal>
3704<email>john@math.nwu.edu</email></address></author>
3705<author initials="P.M." surname="Hallam-Baker" fullname="Phillip M. Hallam-Baker">
3706<organization>Verisign Inc.</organization>
3707<address>
3708<postal>
3709<street>301 Edgewater Place</street>
3710<street>Suite 210</street>
3711<city>Wakefield</city>
3712<region>MA</region>
3713<code>01880</code>
3714<country>US</country></postal>
3715<email>pbaker@verisign.com</email></address></author>
3716<author initials="J.L." surname="Hostetler" fullname="Jeffery L. Hostetler">
3717<organization>AbiSource, Inc.</organization>
3718<address>
3719<postal>
3720<street>6 Dunlap Court</street>
3721<city>Savoy</city>
3722<region>IL</region>
3723<code>61874</code>
3724<country>US</country></postal>
3725<email>jeff@AbiSource.com</email></address></author>
3726<author initials="S.D." surname="Lawrence" fullname="Scott D. Lawrence">
3727<organization>Agranat Systems, Inc.</organization>
3728<address>
3729<postal>
3730<street>5 Clocktower Place</street>
3731<street>Suite 400</street>
3732<city>Maynard</city>
3733<region>MA</region>
3734<code>01754</code>
3735<country>US</country></postal>
3736<email>lawrence@agranat.com</email></address></author>
3737<author initials="P.J." surname="Leach" fullname="Paul J. Leach">
3738<organization>Microsoft Corporation</organization>
3739<address>
3740<postal>
3741<street>1 Microsoft Way</street>
3742<city>Redmond</city>
3743<region>WA</region>
3744<code>98052</code>
3745<country>US</country></postal>
3746<email>paulle@microsoft.com</email></address></author>
3747<author initials="A." surname="Luotonen" fullname="Ari Luotonen">
3748<organization>Netscape Communications Corporation</organization>
3749<address>
3750<postal>
3751<street>501 East Middlefield Road</street>
3752<city>Mountain View</city>
3753<region>CA</region>
3754<code>94043</code>
3755<country>US</country></postal></address></author>
3756<author initials="L." surname="Stewart" fullname="Lawrence C. Stewart">
3757<organization>Open Market, Inc.</organization>
3758<address>
3759<postal>
3760<street>215 First Street</street>
3761<city>Cambridge</city>
3762<region>MA</region>
3763<code>02142</code>
3764<country>US</country></postal>
3765<email>stewart@OpenMarket.com</email></address></author>
3766<date month="June" year="1999"/>
3767</front>
3768<seriesInfo name="RFC" value="2617"/>
3769</reference>
3770
3771<reference anchor="RFC2324">
3772<front>
3773<title abbrev="HTCPCP/1.0">Hyper Text Coffee Pot Control Protocol (HTCPCP/1.0)</title>
3774<author initials="L." surname="Masinter" fullname="Larry Masinter">
3775<organization>Xerox Palo Alto Research Center</organization>
3776<address>
3777<postal>
3778<street>3333 Coyote Hill Road</street>
3779<city>Palo Alto</city>
3780<region>CA</region>
3781<code>94304</code></postal>
3782<email>masinter@parc.xerox.com</email></address></author>
3783<date month="April" day="1" year="1998"/>
3784<area>General</area>
3785<keyword>control protocol</keyword>
3786<keyword>coffee</keyword>
3787</front>
3788<seriesInfo name="RFC" value="2324"/>
3789</reference>
3790
3791</references>
3792
3793<section title="Internet Media Type message/http and application/http" anchor="internet.media.type.http">
3794<iref item="Media Type" subitem="message/http" primary="true"/>
3795<iref item="message/http Media Type" primary="true"/>
3796<iref item="Media Type" subitem="application/http" primary="true"/>
3797<iref item="application/http Media Type" primary="true"/>
3798<t>
3799   In addition to defining the HTTP/1.1 protocol, this document serves
3800   as the specification for the Internet media type "message/http" and
3801   "application/http". The message/http type can be used to enclose a
3802   single HTTP request or response message, provided that it obeys the
3803   MIME restrictions for all "message" types regarding line length and
3804   encodings. The application/http type can be used to enclose a
3805   pipeline of one or more HTTP request or response messages (not
3806   intermixed). The following is to be registered with IANA <xref target="RFC1590"/>.
3807</t>
3808<t>
3809  <list style="hanging" x:indent="12em">
3810    <t hangText="Media Type name:">
3811      message
3812    </t>
3813    <t hangText="Media subtype name:">
3814      http
3815    </t>
3816    <t hangText="Required parameters:">
3817      none
3818    </t>
3819    <t hangText="Optional parameters:">
3820      version, msgtype
3821      <list style="hanging">
3822        <t hangText="version:">
3823          The HTTP-Version number of the enclosed message
3824          (e.g., "1.1"). If not present, the version can be
3825          determined from the first line of the body.
3826        </t>
3827        <t hangText="msgtype:">
3828          The message type -- "request" or "response". If not
3829          present, the type can be determined from the first
3830          line of the body.
3831        </t>
3832      </list>
3833    </t>
3834    <t hangText="Encoding considerations:">
3835      only "7bit", "8bit", or "binary" are permitted
3836    </t>
3837    <t hangText="Security considerations:">
3838      none
3839    </t>
3840  </list>
3841</t>
3842<t>
3843  <list style="hanging" x:indent="12em">
3844    <t hangText="Media Type name:">
3845      application
3846    </t>
3847    <t hangText="Media subtype name:">
3848      http
3849    </t>
3850    <t hangText="Required parameters:">
3851      none
3852    </t>
3853    <t hangText="Optional parameters:">
3854      version, msgtype
3855      <list style="hanging">
3856        <t hangText="version:">
3857          The HTTP-Version number of the enclosed messages
3858          (e.g., "1.1"). If not present, the version can be
3859          determined from the first line of the body.
3860        </t>
3861        <t hangText="msgtype:">
3862          The message type -- "request" or "response". If not
3863          present, the type can be determined from the first
3864          line of the body.
3865        </t>
3866      </list>
3867    </t>
3868    <t hangText="Encoding considerations:">
3869      HTTP messages enclosed by this type
3870      are in "binary" format; use of an appropriate
3871      Content-Transfer-Encoding is required when
3872      transmitted via E-mail.
3873    </t>
3874    <t hangText="Security considerations:">
3875      none
3876    </t>
3877  </list>
3878</t>
3879</section>
3880
3881<section title="Tolerant Applications" anchor="tolerant.applications">
3882<t>
3883   Although this document specifies the requirements for the generation
3884   of HTTP/1.1 messages, not all applications will be correct in their
3885   implementation. We therefore recommend that operational applications
3886   be tolerant of deviations whenever those deviations can be
3887   interpreted unambiguously.
3888</t>
3889<t>
3890   Clients &SHOULD; be tolerant in parsing the Status-Line and servers
3891   tolerant when parsing the Request-Line. In particular, they &SHOULD;
3892   accept any amount of SP or HT characters between fields, even though
3893   only a single SP is required.
3894</t>
3895<t>
3896   The line terminator for message-header fields is the sequence CRLF.
3897   However, we recommend that applications, when parsing such headers,
3898   recognize a single LF as a line terminator and ignore the leading CR.
3899</t>
3900<t>
3901   The character set of an entity-body &SHOULD; be labeled as the lowest
3902   common denominator of the character codes used within that body, with
3903   the exception that not labeling the entity is preferred over labeling
3904   the entity with the labels US-ASCII or ISO-8859-1. See &payload;.
3905</t>
3906<t>
3907   Additional rules for requirements on parsing and encoding of dates
3908   and other potential problems with date encodings include:
3909</t>
3910<t>
3911  <list style="symbols">
3912     <t>HTTP/1.1 clients and caches &SHOULD; assume that an RFC-850 date
3913        which appears to be more than 50 years in the future is in fact
3914        in the past (this helps solve the "year 2000" problem).</t>
3915
3916     <t>An HTTP/1.1 implementation &MAY; internally represent a parsed
3917        Expires date as earlier than the proper value, but &MUST-NOT;
3918        internally represent a parsed Expires date as later than the
3919        proper value.</t>
3920
3921     <t>All expiration-related calculations &MUST; be done in GMT. The
3922        local time zone &MUST-NOT; influence the calculation or comparison
3923        of an age or expiration time.</t>
3924
3925     <t>If an HTTP header incorrectly carries a date value with a time
3926        zone other than GMT, it &MUST; be converted into GMT using the
3927        most conservative possible conversion.</t>
3928  </list>
3929</t>
3930</section>
3931
3932<section title="Conversion of Date Formats" anchor="conversion.of.date.formats">
3933<t>
3934   HTTP/1.1 uses a restricted set of date formats (<xref target="full.date"/>) to
3935   simplify the process of date comparison. Proxies and gateways from
3936   other protocols &SHOULD; ensure that any Date header field present in a
3937   message conforms to one of the HTTP/1.1 formats and rewrite the date
3938   if necessary.
3939</t>
3940</section>
3941
3942<section title="Compatibility with Previous Versions" anchor="compatibility">
3943<t>
3944   It is beyond the scope of a protocol specification to mandate
3945   compliance with previous versions. HTTP/1.1 was deliberately
3946   designed, however, to make supporting previous versions easy. It is
3947   worth noting that, at the time of composing this specification
3948   (1996), we would expect commercial HTTP/1.1 servers to:
3949  <list style="symbols">
3950     <t>recognize the format of the Request-Line for HTTP/0.9, 1.0, and
3951        1.1 requests;</t>
3952
3953     <t>understand any valid request in the format of HTTP/0.9, 1.0, or
3954        1.1;</t>
3955
3956     <t>respond appropriately with a message in the same major version
3957        used by the client.</t>
3958  </list>
3959</t>
3960<t>
3961   And we would expect HTTP/1.1 clients to:
3962  <list style="symbols">
3963     <t>recognize the format of the Status-Line for HTTP/1.0 and 1.1
3964        responses;</t>
3965
3966     <t>understand any valid response in the format of HTTP/0.9, 1.0, or
3967        1.1.</t>
3968  </list>
3969</t>
3970<t>
3971   For most implementations of HTTP/1.0, each connection is established
3972   by the client prior to the request and closed by the server after
3973   sending the response. Some implementations implement the Keep-Alive
3974   version of persistent connections described in <xref x:sec="19.7.1" x:fmt="sec" target="RFC2068"/> of RFC
3975   2068 <xref target="RFC2068"/>.
3976</t>
3977
3978<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
3979<t>
3980   This section summarizes major differences between versions HTTP/1.0
3981   and HTTP/1.1.
3982</t>
3983
3984<section title="Changes to Simplify Multi-homed Web Servers and Conserve IP Addresses" anchor="changes.to.simplify.multi-homed.web.servers.and.conserve.ip.addresses">
3985<t>
3986   The requirements that clients and servers support the Host request-header,
3987   report an error if the Host request-header (<xref target="header.host"/>) is
3988   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-uri"/>)
3989   are among the most important changes defined by this
3990   specification.
3991</t>
3992<t>
3993   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
3994   addresses and servers; there was no other established mechanism for
3995   distinguishing the intended server of a request than the IP address
3996   to which that request was directed. The changes outlined above will
3997   allow the Internet, once older HTTP clients are no longer common, to
3998   support multiple Web sites from a single IP address, greatly
3999   simplifying large operational Web servers, where allocation of many
4000   IP addresses to a single host has created serious problems. The
4001   Internet will also be able to recover the IP addresses that have been
4002   allocated for the sole purpose of allowing special-purpose domain
4003   names to be used in root-level HTTP URLs. Given the rate of growth of
4004   the Web, and the number of servers already deployed, it is extremely
4005   important that all implementations of HTTP (including updates to
4006   existing HTTP/1.0 applications) correctly implement these
4007   requirements:
4008  <list style="symbols">
4009     <t>Both clients and servers &MUST; support the Host request-header.</t>
4010
4011     <t>A client that sends an HTTP/1.1 request &MUST; send a Host header.</t>
4012
4013     <t>Servers &MUST; report a 400 (Bad Request) error if an HTTP/1.1
4014        request does not include a Host request-header.</t>
4015
4016     <t>Servers &MUST; accept absolute URIs.</t>
4017  </list>
4018</t>
4019</section>
4020</section>
4021
4022<section title="Compatibility with HTTP/1.0 Persistent Connections" anchor="compatibility.with.http.1.0.persistent.connections">
4023<t>
4024   Some clients and servers might wish to be compatible with some
4025   previous implementations of persistent connections in HTTP/1.0
4026   clients and servers. Persistent connections in HTTP/1.0 are
4027   explicitly negotiated as they are not the default behavior. HTTP/1.0
4028   experimental implementations of persistent connections are faulty,
4029   and the new facilities in HTTP/1.1 are designed to rectify these
4030   problems. The problem was that some existing 1.0 clients may be
4031   sending Keep-Alive to a proxy server that doesn't understand
4032   Connection, which would then erroneously forward it to the next
4033   inbound server, which would establish the Keep-Alive connection and
4034   result in a hung HTTP/1.0 proxy waiting for the close on the
4035   response. The result is that HTTP/1.0 clients must be prevented from
4036   using Keep-Alive when talking to proxies.
4037</t>
4038<t>
4039   However, talking to proxies is the most important use of persistent
4040   connections, so that prohibition is clearly unacceptable. Therefore,
4041   we need some other mechanism for indicating a persistent connection
4042   is desired, which is safe to use even when talking to an old proxy
4043   that ignores Connection. Persistent connections are the default for
4044   HTTP/1.1 messages; we introduce a new keyword (Connection: close) for
4045   declaring non-persistence. See <xref target="header.connection"/>.
4046</t>
4047<t>
4048   The original HTTP/1.0 form of persistent connections (the Connection:
4049   Keep-Alive and Keep-Alive header) is documented in RFC 2068. <xref target="RFC2068"/>
4050</t>
4051</section>
4052
4053<section title="Changes from RFC 2068" anchor="changes.from.rfc.2068">
4054<t>
4055   This specification has been carefully audited to correct and
4056   disambiguate key word usage; RFC 2068 had many problems in respect to
4057   the conventions laid out in RFC 2119 <xref target="RFC2119"/>.
4058</t>
4059<t>
4060   Transfer-coding and message lengths all interact in ways that
4061   required fixing exactly when chunked encoding is used (to allow for
4062   transfer encoding that may not be self delimiting); it was important
4063   to straighten out exactly how message lengths are computed.
4064</t>
4065<t>
4066   The use and interpretation of HTTP version numbers has been clarified
4067   by RFC 2145. Require proxies to upgrade requests to highest protocol
4068   version they support to deal with problems discovered in HTTP/1.0
4069   implementations (<xref target="http.version"/>)
4070</t>
4071<t>
4072   Proxies should be able to add Content-Length when appropriate.
4073</t>
4074<t>
4075   Transfer-coding had significant problems, particularly with
4076   interactions with chunked encoding. The solution is that transfer-codings
4077   become as full fledged as content-codings. This involves
4078   adding an IANA registry for transfer-codings (separate from content
4079   codings), a new header field (TE) and enabling trailer headers in the
4080   future. Transfer encoding is a major performance benefit, so it was
4081   worth fixing <xref target="Nie1997"/>. TE also solves another, obscure, downward
4082   interoperability problem that could have occurred due to interactions
4083   between authentication trailers, chunked encoding and HTTP/1.0
4084   clients.(Section <xref target="transfer.codings" format="counter"/>, <xref target="chunked.transfer.encoding" format="counter"/>,
4085   and <xref target="header.te" format="counter"/>)
4086</t>
4087</section>
4088</section>
4089</back>
4090</rfc>
Note: See TracBrowser for help on using the repository browser.