idnits 2.17.1 draft-ietf-httpbis-messaging-07.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The draft header indicates that this document obsoletes RFC7230, but the abstract doesn't seem to directly say this. It does mention RFC7230 though, so this could be OK. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document seems to contain a disclaimer for pre-RFC5378 work, but was first submitted on or after 10 November 2008. The disclaimer is usually necessary only for documents that revise or obsolete older RFCs, and that take significant amounts of text from those RFCs. If you can contact all authors of the source material and they are willing to grant the BCP78 rights to the IETF Trust, you can and should remove the disclaimer. Otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (March 7, 2020) is 1510 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'RFC3986' is defined on line 2070, but no explicit reference was found in the text == Unused Reference: 'RFC7231' is defined on line 2157, but no explicit reference was found in the text == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-cache-07 -- Possible downref: Normative reference to a draft: ref. 'Caching' ** Downref: Normative reference to an Informational RFC: RFC 1950 ** Downref: Normative reference to an Informational RFC: RFC 1951 ** Downref: Normative reference to an Informational RFC: RFC 1952 == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-semantics-07 -- Possible downref: Normative reference to a draft: ref. 'Semantics' -- Possible downref: Non-RFC (?) normative reference: ref. 'USASCII' -- Possible downref: Non-RFC (?) normative reference: ref. 'Welch' -- Obsolete informational reference (is this intentional?): RFC 7230 (ref. 'Err4667') (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 2068 (Obsoleted by RFC 2616) -- Duplicate reference: RFC7230, mentioned in 'RFC7230', was also mentioned in 'Err4667'. -- Obsolete informational reference (is this intentional?): RFC 7230 (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 7231 (Obsoleted by RFC 9110) Summary: 3 errors (**), 0 flaws (~~), 6 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 HTTP Working Group R. Fielding, Ed. 3 Internet-Draft Adobe 4 Obsoletes: 7230 (if approved) M. Nottingham, Ed. 5 Intended status: Standards Track Fastly 6 Expires: September 8, 2020 J. Reschke, Ed. 7 greenbytes 8 March 7, 2020 10 HTTP/1.1 Messaging 11 draft-ietf-httpbis-messaging-07 13 Abstract 15 The Hypertext Transfer Protocol (HTTP) is a stateless application- 16 level protocol for distributed, collaborative, hypertext information 17 systems. This document specifies the HTTP/1.1 message syntax, 18 message parsing, connection management, and related security 19 concerns. 21 This document obsoletes portions of RFC 7230. 23 Editorial Note 25 This note is to be removed before publishing as an RFC. 27 Discussion of this draft takes place on the HTTP working group 28 mailing list (ietf-http-wg@w3.org), which is archived at 29 . 31 Working Group information can be found at ; 32 source code and issues list for this draft can be found at 33 . 35 The changes in this draft are summarized in Appendix D.8. 37 Status of This Memo 39 This Internet-Draft is submitted in full conformance with the 40 provisions of BCP 78 and BCP 79. 42 Internet-Drafts are working documents of the Internet Engineering 43 Task Force (IETF). Note that other groups may also distribute 44 working documents as Internet-Drafts. The list of current Internet- 45 Drafts is at https://datatracker.ietf.org/drafts/current/. 47 Internet-Drafts are draft documents valid for a maximum of six months 48 and may be updated, replaced, or obsoleted by other documents at any 49 time. It is inappropriate to use Internet-Drafts as reference 50 material or to cite them other than as "work in progress." 52 This Internet-Draft will expire on September 8, 2020. 54 Copyright Notice 56 Copyright (c) 2020 IETF Trust and the persons identified as the 57 document authors. All rights reserved. 59 This document is subject to BCP 78 and the IETF Trust's Legal 60 Provisions Relating to IETF Documents 61 (https://trustee.ietf.org/license-info) in effect on the date of 62 publication of this document. Please review these documents 63 carefully, as they describe your rights and restrictions with respect 64 to this document. Code Components extracted from this document must 65 include Simplified BSD License text as described in Section 4.e of 66 the Trust Legal Provisions and are provided without warranty as 67 described in the Simplified BSD License. 69 This document may contain material from IETF Documents or IETF 70 Contributions published or made publicly available before November 71 10, 2008. The person(s) controlling the copyright in some of this 72 material may not have granted the IETF Trust the right to allow 73 modifications of such material outside the IETF Standards Process. 74 Without obtaining an adequate license from the person(s) controlling 75 the copyright in such materials, this document may not be modified 76 outside the IETF Standards Process, and derivative works of it may 77 not be created outside the IETF Standards Process, except to format 78 it for publication as an RFC or to translate it into languages other 79 than English. 81 Table of Contents 83 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 84 1.1. Requirements Notation . . . . . . . . . . . . . . . . . . 5 85 1.2. Syntax Notation . . . . . . . . . . . . . . . . . . . . . 5 86 2. Message . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 87 2.1. Message Format . . . . . . . . . . . . . . . . . . . . . 6 88 2.2. Message Parsing . . . . . . . . . . . . . . . . . . . . . 7 89 2.3. HTTP Version . . . . . . . . . . . . . . . . . . . . . . 8 90 3. Request Line . . . . . . . . . . . . . . . . . . . . . . . . 9 91 3.1. Method . . . . . . . . . . . . . . . . . . . . . . . . . 9 92 3.2. Request Target . . . . . . . . . . . . . . . . . . . . . 10 93 3.2.1. origin-form . . . . . . . . . . . . . . . . . . . . . 10 94 3.2.2. absolute-form . . . . . . . . . . . . . . . . . . . . 11 95 3.2.3. authority-form . . . . . . . . . . . . . . . . . . . 11 96 3.2.4. asterisk-form . . . . . . . . . . . . . . . . . . . . 11 98 3.3. Effective Request URI . . . . . . . . . . . . . . . . . . 12 99 4. Status Line . . . . . . . . . . . . . . . . . . . . . . . . . 13 100 5. Field Syntax . . . . . . . . . . . . . . . . . . . . . . . . 14 101 5.1. Field Line Parsing . . . . . . . . . . . . . . . . . . . 15 102 5.2. Obsolete Line Folding . . . . . . . . . . . . . . . . . . 16 103 6. Message Body . . . . . . . . . . . . . . . . . . . . . . . . 16 104 6.1. Transfer-Encoding . . . . . . . . . . . . . . . . . . . . 17 105 6.2. Content-Length . . . . . . . . . . . . . . . . . . . . . 18 106 6.3. Message Body Length . . . . . . . . . . . . . . . . . . . 19 107 7. Transfer Codings . . . . . . . . . . . . . . . . . . . . . . 21 108 7.1. Chunked Transfer Coding . . . . . . . . . . . . . . . . . 22 109 7.1.1. Chunk Extensions . . . . . . . . . . . . . . . . . . 23 110 7.1.2. Chunked Trailer Section . . . . . . . . . . . . . . . 23 111 7.1.3. Decoding Chunked . . . . . . . . . . . . . . . . . . 24 112 7.2. Transfer Codings for Compression . . . . . . . . . . . . 24 113 7.3. Transfer Coding Registry . . . . . . . . . . . . . . . . 25 114 7.4. TE . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 115 8. Handling Incomplete Messages . . . . . . . . . . . . . . . . 27 116 9. Connection Management . . . . . . . . . . . . . . . . . . . . 27 117 9.1. Connection . . . . . . . . . . . . . . . . . . . . . . . 28 118 9.2. Establishment . . . . . . . . . . . . . . . . . . . . . . 29 119 9.3. Associating a Response to a Request . . . . . . . . . . . 29 120 9.4. Persistence . . . . . . . . . . . . . . . . . . . . . . . 30 121 9.4.1. Retrying Requests . . . . . . . . . . . . . . . . . . 31 122 9.4.2. Pipelining . . . . . . . . . . . . . . . . . . . . . 31 123 9.5. Concurrency . . . . . . . . . . . . . . . . . . . . . . . 32 124 9.6. Failures and Timeouts . . . . . . . . . . . . . . . . . . 32 125 9.7. Tear-down . . . . . . . . . . . . . . . . . . . . . . . . 33 126 9.8. TLS Connection Closure . . . . . . . . . . . . . . . . . 34 127 9.9. Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . 35 128 9.9.1. Upgrade Protocol Names . . . . . . . . . . . . . . . 37 129 9.9.2. Upgrade Token Registry . . . . . . . . . . . . . . . 37 130 10. Enclosing Messages as Data . . . . . . . . . . . . . . . . . 38 131 10.1. Media Type message/http . . . . . . . . . . . . . . . . 38 132 10.2. Media Type application/http . . . . . . . . . . . . . . 39 133 11. Security Considerations . . . . . . . . . . . . . . . . . . . 41 134 11.1. Response Splitting . . . . . . . . . . . . . . . . . . . 41 135 11.2. Request Smuggling . . . . . . . . . . . . . . . . . . . 42 136 11.3. Message Integrity . . . . . . . . . . . . . . . . . . . 42 137 11.4. Message Confidentiality . . . . . . . . . . . . . . . . 42 138 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 43 139 12.1. Field Name Registration . . . . . . . . . . . . . . . . 43 140 12.2. Media Type Registration . . . . . . . . . . . . . . . . 43 141 12.3. Transfer Coding Registration . . . . . . . . . . . . . . 43 142 12.4. Upgrade Token Registration . . . . . . . . . . . . . . . 43 143 12.5. ALPN Protocol ID Registration . . . . . . . . . . . . . 43 144 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 44 145 13.1. Normative References . . . . . . . . . . . . . . . . . . 44 146 13.2. Informative References . . . . . . . . . . . . . . . . . 45 147 Appendix A. Collected ABNF . . . . . . . . . . . . . . . . . . . 47 148 Appendix B. Differences between HTTP and MIME . . . . . . . . . 48 149 B.1. MIME-Version . . . . . . . . . . . . . . . . . . . . . . 49 150 B.2. Conversion to Canonical Form . . . . . . . . . . . . . . 49 151 B.3. Conversion of Date Formats . . . . . . . . . . . . . . . 49 152 B.4. Conversion of Content-Encoding . . . . . . . . . . . . . 50 153 B.5. Conversion of Content-Transfer-Encoding . . . . . . . . . 50 154 B.6. MHTML and Line Length Limitations . . . . . . . . . . . . 50 155 Appendix C. HTTP Version History . . . . . . . . . . . . . . . . 50 156 C.1. Changes from HTTP/1.0 . . . . . . . . . . . . . . . . . . 51 157 C.1.1. Multihomed Web Servers . . . . . . . . . . . . . . . 51 158 C.1.2. Keep-Alive Connections . . . . . . . . . . . . . . . 52 159 C.1.3. Introduction of Transfer-Encoding . . . . . . . . . . 52 160 C.2. Changes from RFC 7230 . . . . . . . . . . . . . . . . . . 52 161 Appendix D. Change Log . . . . . . . . . . . . . . . . . . . . . 53 162 D.1. Between RFC7230 and draft 00 . . . . . . . . . . . . . . 53 163 D.2. Since draft-ietf-httpbis-messaging-00 . . . . . . . . . . 53 164 D.3. Since draft-ietf-httpbis-messaging-01 . . . . . . . . . . 54 165 D.4. Since draft-ietf-httpbis-messaging-02 . . . . . . . . . . 55 166 D.5. Since draft-ietf-httpbis-messaging-03 . . . . . . . . . . 55 167 D.6. Since draft-ietf-httpbis-messaging-04 . . . . . . . . . . 55 168 D.7. Since draft-ietf-httpbis-messaging-05 . . . . . . . . . . 55 169 D.8. Since draft-ietf-httpbis-messaging-06 . . . . . . . . . . 56 170 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 171 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 59 172 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 59 174 1. Introduction 176 The Hypertext Transfer Protocol (HTTP) is a stateless application- 177 level request/response protocol that uses extensible semantics and 178 self-descriptive messages for flexible interaction with network-based 179 hypertext information systems. HTTP is defined by a series of 180 documents that collectively form the HTTP/1.1 specification: 182 o "HTTP Semantics" [Semantics] 184 o "HTTP Caching" [Caching] 186 o "HTTP/1.1 Messaging" (this document) 188 This document defines HTTP/1.1 message syntax and framing 189 requirements and their associated connection management. Our goal is 190 to define all of the mechanisms necessary for HTTP/1.1 message 191 handling that are independent of message semantics, thereby defining 192 the complete set of requirements for message parsers and message- 193 forwarding intermediaries. 195 This document obsoletes the portions of RFC 7230 related to HTTP/1.1 196 messaging and connection management, with the changes being 197 summarized in Appendix C.2. The other parts of RFC 7230 are 198 obsoleted by "HTTP Semantics" [Semantics]. 200 1.1. Requirements Notation 202 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 203 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 204 "OPTIONAL" in this document are to be interpreted as described in BCP 205 14 [RFC2119] [RFC8174] when, and only when, they appear in all 206 capitals, as shown here. 208 Conformance criteria and considerations regarding error handling are 209 defined in Section 3 of [Semantics]. 211 1.2. Syntax Notation 213 This specification uses the Augmented Backus-Naur Form (ABNF) 214 notation of [RFC5234], extended with the notation for case- 215 sensitivity in strings defined in [RFC7405]. 217 It also uses a list extension, defined in Section 4.5 of [Semantics], 218 that allows for compact definition of comma-separated lists using a 219 '#' operator (similar to how the '*' operator indicates repetition). 220 Appendix A shows the collected grammar with all list operators 221 expanded to standard ABNF notation. 223 As a convention, ABNF rule names prefixed with "obs-" denote 224 "obsolete" grammar rules that appear for historical reasons. 226 The following core rules are included by reference, as defined in 227 [RFC5234], Appendix B.1: ALPHA (letters), CR (carriage return), CRLF 228 (CR LF), CTL (controls), DIGIT (decimal 0-9), DQUOTE (double quote), 229 HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line 230 feed), OCTET (any 8-bit sequence of data), SP (space), and VCHAR (any 231 visible [USASCII] character). 233 The rules below are defined in [Semantics]: 235 BWS = 236 OWS = 237 RWS = 238 absolute-URI = 239 absolute-path = 240 authority = 241 comment = 242 field-name = 243 field-value = 244 obs-text = 245 port = 246 query = 247 quoted-string = 248 token = 249 uri-host = 251 2. Message 253 2.1. Message Format 255 An HTTP/1.1 message consists of a start-line followed by a CRLF and a 256 sequence of octets in a format similar to the Internet Message Format 257 [RFC5322]: zero or more header field lines (collectively referred to 258 as the "headers" or the "header section"), an empty line indicating 259 the end of the header section, and an optional message body. 261 HTTP-message = start-line CRLF 262 *( field-line CRLF ) 263 CRLF 264 [ message-body ] 266 A message can be either a request from client to server or a response 267 from server to client. Syntactically, the two types of message 268 differ only in the start-line, which is either a request-line (for 269 requests) or a status-line (for responses), and in the algorithm for 270 determining the length of the message body (Section 6). 272 start-line = request-line / status-line 274 In theory, a client could receive requests and a server could receive 275 responses, distinguishing them by their different start-line formats. 276 In practice, servers are implemented to only expect a request (a 277 response is interpreted as an unknown or invalid request method) and 278 clients are implemented to only expect a response. 280 Although HTTP makes use of some protocol elements similar to the 281 Multipurpose Internet Mail Extensions (MIME) [RFC2045], see 282 Appendix B for the differences between HTTP and MIME messages. 284 2.2. Message Parsing 286 The normal procedure for parsing an HTTP message is to read the 287 start-line into a structure, read each header field line into a hash 288 table by field name until the empty line, and then use the parsed 289 data to determine if a message body is expected. If a message body 290 has been indicated, then it is read as a stream until an amount of 291 octets equal to the message body length is read or the connection is 292 closed. 294 A recipient MUST parse an HTTP message as a sequence of octets in an 295 encoding that is a superset of US-ASCII [USASCII]. Parsing an HTTP 296 message as a stream of Unicode characters, without regard for the 297 specific encoding, creates security vulnerabilities due to the 298 varying ways that string processing libraries handle invalid 299 multibyte character sequences that contain the octet LF (%x0A). 300 String-based parsers can only be safely used within protocol elements 301 after the element has been extracted from the message, such as within 302 a header field line value after message parsing has delineated the 303 individual field lines. 305 Although the line terminator for the start-line and header fields is 306 the sequence CRLF, a recipient MAY recognize a single LF as a line 307 terminator and ignore any preceding CR. 309 Older HTTP/1.0 user agent implementations might send an extra CRLF 310 after a POST request as a workaround for some early server 311 applications that failed to read message body content that was not 312 terminated by a line-ending. An HTTP/1.1 user agent MUST NOT preface 313 or follow a request with an extra CRLF. If terminating the request 314 message body with a line-ending is desired, then the user agent MUST 315 count the terminating CRLF octets as part of the message body length. 317 In the interest of robustness, a server that is expecting to receive 318 and parse a request-line SHOULD ignore at least one empty line (CRLF) 319 received prior to the request-line. 321 A sender MUST NOT send whitespace between the start-line and the 322 first header field. A recipient that receives whitespace between the 323 start-line and the first header field MUST either reject the message 324 as invalid or consume each whitespace-preceded line without further 325 processing of it (i.e., ignore the entire line, along with any 326 subsequent lines preceded by whitespace, until a properly formed 327 header field is received or the header section is terminated). 329 The presence of such whitespace in a request might be an attempt to 330 trick a server into ignoring that field line or processing the line 331 after it as a new request, either of which might result in a security 332 vulnerability if other implementations within the request chain 333 interpret the same message differently. Likewise, the presence of 334 such whitespace in a response might be ignored by some clients or 335 cause others to cease parsing. 337 When a server listening only for HTTP request messages, or processing 338 what appears from the start-line to be an HTTP request message, 339 receives a sequence of octets that does not match the HTTP-message 340 grammar aside from the robustness exceptions listed above, the server 341 SHOULD respond with a 400 (Bad Request) response. 343 2.3. HTTP Version 345 HTTP uses a "." numbering scheme to indicate versions 346 of the protocol. This specification defines version "1.1". 347 Section 3.5 of [Semantics] specifies the semantics of HTTP version 348 numbers. 350 The version of an HTTP/1.x message is indicated by an HTTP-version 351 field in the start-line. HTTP-version is case-sensitive. 353 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 354 HTTP-name = %s"HTTP" 356 When an HTTP/1.1 message is sent to an HTTP/1.0 recipient [RFC1945] 357 or a recipient whose version is unknown, the HTTP/1.1 message is 358 constructed such that it can be interpreted as a valid HTTP/1.0 359 message if all of the newer features are ignored. This specification 360 places recipient-version requirements on some new features so that a 361 conformant sender will only use compatible features until it has 362 determined, through configuration or the receipt of a message, that 363 the recipient supports HTTP/1.1. 365 Intermediaries that process HTTP messages (i.e., all intermediaries 366 other than those acting as tunnels) MUST send their own HTTP-version 367 in forwarded messages. In other words, they are not allowed to 368 blindly forward the start-line without ensuring that the protocol 369 version in that message matches a version to which that intermediary 370 is conformant for both the receiving and sending of messages. 371 Forwarding an HTTP message without rewriting the HTTP-version might 372 result in communication errors when downstream recipients use the 373 message sender's version to determine what features are safe to use 374 for later communication with that sender. 376 A server MAY send an HTTP/1.0 response to an HTTP/1.1 request if it 377 is known or suspected that the client incorrectly implements the HTTP 378 specification and is incapable of correctly processing later version 379 responses, such as when a client fails to parse the version number 380 correctly or when an intermediary is known to blindly forward the 381 HTTP-version even when it doesn't conform to the given minor version 382 of the protocol. Such protocol downgrades SHOULD NOT be performed 383 unless triggered by specific client attributes, such as when one or 384 more of the request header fields (e.g., User-Agent) uniquely match 385 the values sent by a client known to be in error. 387 3. Request Line 389 A request-line begins with a method token, followed by a single space 390 (SP), the request-target, another single space (SP), and ends with 391 the protocol version. 393 request-line = method SP request-target SP HTTP-version 395 Although the request-line grammar rule requires that each of the 396 component elements be separated by a single SP octet, recipients MAY 397 instead parse on whitespace-delimited word boundaries and, aside from 398 the CRLF terminator, treat any form of whitespace as the SP separator 399 while ignoring preceding or trailing whitespace; such whitespace 400 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 401 (%x0C), or bare CR. However, lenient parsing can result in request 402 smuggling security vulnerabilities if there are multiple recipients 403 of the message and each has its own unique interpretation of 404 robustness (see Section 11.2). 406 HTTP does not place a predefined limit on the length of a request- 407 line, as described in Section 3 of [Semantics]. A server that 408 receives a method longer than any that it implements SHOULD respond 409 with a 501 (Not Implemented) status code. A server that receives a 410 request-target longer than any URI it wishes to parse MUST respond 411 with a 414 (URI Too Long) status code (see Section 9.5.15 of 412 [Semantics]). 414 Various ad hoc limitations on request-line length are found in 415 practice. It is RECOMMENDED that all HTTP senders and recipients 416 support, at a minimum, request-line lengths of 8000 octets. 418 3.1. Method 420 The method token indicates the request method to be performed on the 421 target resource. The request method is case-sensitive. 423 method = token 425 The request methods defined by this specification can be found in 426 Section 7 of [Semantics], along with information regarding the HTTP 427 method registry and considerations for defining new methods. 429 3.2. Request Target 431 The request-target identifies the target resource upon which to apply 432 the request. The client derives a request-target from its desired 433 target URI. There are four distinct formats for the request-target, 434 depending on both the method being requested and whether the request 435 is to a proxy. 437 request-target = origin-form 438 / absolute-form 439 / authority-form 440 / asterisk-form 442 No whitespace is allowed in the request-target. Unfortunately, some 443 user agents fail to properly encode or exclude whitespace found in 444 hypertext references, resulting in those disallowed characters being 445 sent as the request-target in a malformed request-line. 447 Recipients of an invalid request-line SHOULD respond with either a 448 400 (Bad Request) error or a 301 (Moved Permanently) redirect with 449 the request-target properly encoded. A recipient SHOULD NOT attempt 450 to autocorrect and then process the request without a redirect, since 451 the invalid request-line might be deliberately crafted to bypass 452 security filters along the request chain. 454 3.2.1. origin-form 456 The most common form of request-target is the origin-form. 458 origin-form = absolute-path [ "?" query ] 460 When making a request directly to an origin server, other than a 461 CONNECT or server-wide OPTIONS request (as detailed below), a client 462 MUST send only the absolute path and query components of the target 463 URI as the request-target. If the target URI's path component is 464 empty, the client MUST send "/" as the path within the origin-form of 465 request-target. A Host header field is also sent, as defined in 466 Section 5.6 of [Semantics]. 468 For example, a client wishing to retrieve a representation of the 469 resource identified as 471 http://www.example.org/where?q=now 473 directly from the origin server would open (or reuse) a TCP 474 connection to port 80 of the host "www.example.org" and send the 475 lines: 477 GET /where?q=now HTTP/1.1 478 Host: www.example.org 480 followed by the remainder of the request message. 482 3.2.2. absolute-form 484 When making a request to a proxy, other than a CONNECT or server-wide 485 OPTIONS request (as detailed below), a client MUST send the target 486 URI in absolute-form as the request-target. 488 absolute-form = absolute-URI 490 The proxy is requested to either service that request from a valid 491 cache, if possible, or make the same request on the client's behalf 492 to either the next inbound proxy server or directly to the origin 493 server indicated by the request-target. Requirements on such 494 "forwarding" of messages are defined in Section 5.7 of [Semantics]. 496 An example absolute-form of request-line would be: 498 GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1 500 To allow for transition to the absolute-form for all requests in some 501 future version of HTTP, a server MUST accept the absolute-form in 502 requests, even though HTTP/1.1 clients will only send them in 503 requests to proxies. 505 3.2.3. authority-form 507 The authority-form of request-target is only used for CONNECT 508 requests (Section 7.3.6 of [Semantics]). 510 authority-form = authority 512 When making a CONNECT request to establish a tunnel through one or 513 more proxies, a client MUST send only the target URI's authority 514 component (excluding any userinfo and its "@" delimiter) as the 515 request-target. For example, 517 CONNECT www.example.com:80 HTTP/1.1 519 3.2.4. asterisk-form 521 The asterisk-form of request-target is only used for a server-wide 522 OPTIONS request (Section 7.3.7 of [Semantics]). 524 asterisk-form = "*" 526 When a client wishes to request OPTIONS for the server as a whole, as 527 opposed to a specific named resource of that server, the client MUST 528 send only "*" (%x2A) as the request-target. For example, 530 OPTIONS * HTTP/1.1 532 If a proxy receives an OPTIONS request with an absolute-form of 533 request-target in which the URI has an empty path and no query 534 component, then the last proxy on the request chain MUST send a 535 request-target of "*" when it forwards the request to the indicated 536 origin server. 538 For example, the request 540 OPTIONS http://www.example.org:8001 HTTP/1.1 542 would be forwarded by the final proxy as 544 OPTIONS * HTTP/1.1 545 Host: www.example.org:8001 547 after connecting to port 8001 of host "www.example.org". 549 3.3. Effective Request URI 551 Since the request-target often contains only part of the user agent's 552 target URI, a server reconstructs the intended target as an effective 553 request URI to properly service the request (Section 5.5 of 554 [Semantics]). 556 If the request-target is in absolute-form, the effective request URI 557 is the same as the request-target. Otherwise, the effective request 558 URI is constructed as follows: 560 If the server's configuration (or outbound gateway) provides a 561 fixed URI scheme, that scheme is used for the effective request 562 URI. Otherwise, if the request is received over a TLS-secured TCP 563 connection, the effective request URI's scheme is "https"; if not, 564 the scheme is "http". 566 If the server's configuration (or outbound gateway) provides a 567 fixed URI authority component, that authority is used for the 568 effective request URI. If not, then if the request-target is in 569 authority-form, the effective request URI's authority component is 570 the same as the request-target. If not, then if a Host header 571 field is supplied with a non-empty field-value, the authority 572 component is the same as the Host field-value. Otherwise, the 573 authority component is assigned the default name configured for 574 the server and, if the connection's incoming TCP port number 575 differs from the default port for the effective request URI's 576 scheme, then a colon (":") and the incoming port number (in 577 decimal form) are appended to the authority component. 579 If the request-target is in authority-form or asterisk-form, the 580 effective request URI's combined path and query component is 581 empty. Otherwise, the combined path and query component is the 582 same as the request-target. 584 The components of the effective request URI, once determined as 585 above, can be combined into absolute-URI form by concatenating the 586 scheme, "://", authority, and combined path and query component. 588 Example 1: the following message received over an insecure TCP 589 connection 591 GET /pub/WWW/TheProject.html HTTP/1.1 592 Host: www.example.org:8080 594 has an effective request URI of 596 http://www.example.org:8080/pub/WWW/TheProject.html 598 Example 2: the following message received over a TLS-secured TCP 599 connection 601 OPTIONS * HTTP/1.1 602 Host: www.example.org 604 has an effective request URI of 606 https://www.example.org 608 Recipients of an HTTP/1.0 request that lacks a Host header field 609 might need to use heuristics (e.g., examination of the URI path for 610 something unique to a particular host) in order to guess the 611 effective request URI's authority component. 613 4. Status Line 615 The first line of a response message is the status-line, consisting 616 of the protocol version, a space (SP), the status code, another 617 space, and ending with an OPTIONAL textual phrase describing the 618 status code. 620 status-line = HTTP-version SP status-code SP [reason-phrase] 622 Although the status-line grammar rule requires that each of the 623 component elements be separated by a single SP octet, recipients MAY 624 instead parse on whitespace-delimited word boundaries and, aside from 625 the line terminator, treat any form of whitespace as the SP separator 626 while ignoring preceding or trailing whitespace; such whitespace 627 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 628 (%x0C), or bare CR. However, lenient parsing can result in response 629 splitting security vulnerabilities if there are multiple recipients 630 of the message and each has its own unique interpretation of 631 robustness (see Section 11.1). 633 The status-code element is a 3-digit integer code describing the 634 result of the server's attempt to understand and satisfy the client's 635 corresponding request. The rest of the response message is to be 636 interpreted in light of the semantics defined for that status code. 637 See Section 9 of [Semantics] for information about the semantics of 638 status codes, including the classes of status code (indicated by the 639 first digit), the status codes defined by this specification, 640 considerations for the definition of new status codes, and the IANA 641 registry. 643 status-code = 3DIGIT 645 The reason-phrase element exists for the sole purpose of providing a 646 textual description associated with the numeric status code, mostly 647 out of deference to earlier Internet application protocols that were 648 more frequently used with interactive text clients. 650 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 652 A client SHOULD ignore the reason-phrase content because it is not a 653 reliable channel for information (it might be translated for a given 654 locale, overwritten by intermediaries, or discarded when the message 655 is forwarded via other versions of HTTP). A server MUST send the 656 space that separates status-code from the reason-phrase even when the 657 reason-phrase is absent (i.e., the status-line would end with the 658 three octets SP CR LF). 660 5. Field Syntax 662 Each field line consists of a case-insensitive field name followed by 663 a colon (":"), optional leading whitespace, the field line value, and 664 optional trailing whitespace. 666 field-line = field-name ":" OWS field-value OWS 668 Most HTTP field names and the rules for parsing within field values 669 are defined in Section 4 of [Semantics]. This section covers the 670 generic syntax for header field inclusion within, and extraction 671 from, HTTP/1.1 messages. In addition, the following header fields 672 are defined by this document because they are specific to HTTP/1.1 673 message processing: 675 +-------------------+----------+---------------+ 676 | Field Name | Status | Reference | 677 +-------------------+----------+---------------+ 678 | Connection | standard | Section 9.1 | 679 | MIME-Version | standard | Appendix B.1 | 680 | TE | standard | Section 7.4 | 681 | Transfer-Encoding | standard | Section 6.1 | 682 | Upgrade | standard | Section 9.9 | 683 +-------------------+----------+---------------+ 685 Table 1 687 Furthermore, the field name "Close" is reserved, since using that 688 name as an HTTP header field might conflict with the "close" 689 connection option of the Connection header field (Section 9.1). 691 +-------------------+----------+----------+------------+ 692 | Header Field Name | Protocol | Status | Reference | 693 +-------------------+----------+----------+------------+ 694 | Close | http | reserved | Section 5 | 695 +-------------------+----------+----------+------------+ 697 5.1. Field Line Parsing 699 Messages are parsed using a generic algorithm, independent of the 700 individual field names. The contents within a given field line value 701 are not parsed until a later stage of message interpretation (usually 702 after the message's entire header section has been processed). 704 No whitespace is allowed between the field name and colon. In the 705 past, differences in the handling of such whitespace have led to 706 security vulnerabilities in request routing and response handling. A 707 server MUST reject any received request message that contains 708 whitespace between a header field name and colon with a response 709 status code of 400 (Bad Request). A proxy MUST remove any such 710 whitespace from a response message before forwarding the message 711 downstream. 713 A field line value might be preceded and/or followed by optional 714 whitespace (OWS); a single SP preceding the field line value is 715 preferred for consistent readability by humans. The field line value 716 does not include any leading or trailing whitespace: OWS occurring 717 before the first non-whitespace octet of the field line value or 718 after the last non-whitespace octet of the field line value ought to 719 be excluded by parsers when extracting the field line value from a 720 header field line. 722 5.2. Obsolete Line Folding 724 Historically, HTTP header field line values could be extended over 725 multiple lines by preceding each extra line with at least one space 726 or horizontal tab (obs-fold). This specification deprecates such 727 line folding except within the message/http media type 728 (Section 10.1). 730 obs-fold = OWS CRLF RWS 731 ; obsolete line folding 733 A sender MUST NOT generate a message that includes line folding 734 (i.e., that has any field line value that contains a match to the 735 obs-fold rule) unless the message is intended for packaging within 736 the message/http media type. 738 A server that receives an obs-fold in a request message that is not 739 within a message/http container MUST either reject the message by 740 sending a 400 (Bad Request), preferably with a representation 741 explaining that obsolete line folding is unacceptable, or replace 742 each received obs-fold with one or more SP octets prior to 743 interpreting the field value or forwarding the message downstream. 745 A proxy or gateway that receives an obs-fold in a response message 746 that is not within a message/http container MUST either discard the 747 message and replace it with a 502 (Bad Gateway) response, preferably 748 with a representation explaining that unacceptable line folding was 749 received, or replace each received obs-fold with one or more SP 750 octets prior to interpreting the field value or forwarding the 751 message downstream. 753 A user agent that receives an obs-fold in a response message that is 754 not within a message/http container MUST replace each received obs- 755 fold with one or more SP octets prior to interpreting the field 756 value. 758 6. Message Body 760 The message body (if any) of an HTTP message is used to carry the 761 payload body (Section 6.3.3 of [Semantics]) of that request or 762 response. The message body is identical to the payload body unless a 763 transfer coding has been applied, as described in Section 6.1. 765 message-body = *OCTET 767 The rules for determining when a message body is present in an 768 HTTP/1.1 message differ for requests and responses. 770 The presence of a message body in a request is signaled by a Content- 771 Length or Transfer-Encoding header field. Request message framing is 772 independent of method semantics, even if the method does not define 773 any use for a message body. 775 The presence of a message body in a response depends on both the 776 request method to which it is responding and the response status code 777 (Section 4), and corresponds to when a payload body is allowed; see 778 Section 6.3.3 of [Semantics]. 780 6.1. Transfer-Encoding 782 The Transfer-Encoding header field lists the transfer coding names 783 corresponding to the sequence of transfer codings that have been (or 784 will be) applied to the payload body in order to form the message 785 body. Transfer codings are defined in Section 7. 787 Transfer-Encoding = 1#transfer-coding 789 Transfer-Encoding is analogous to the Content-Transfer-Encoding field 790 of MIME, which was designed to enable safe transport of binary data 791 over a 7-bit transport service ([RFC2045], Section 6). However, safe 792 transport has a different focus for an 8bit-clean transfer protocol. 793 In HTTP's case, Transfer-Encoding is primarily intended to accurately 794 delimit a dynamically generated payload and to distinguish payload 795 encodings that are only applied for transport efficiency or security 796 from those that are characteristics of the selected resource. 798 A recipient MUST be able to parse the chunked transfer coding 799 (Section 7.1) because it plays a crucial role in framing messages 800 when the payload body size is not known in advance. A sender MUST 801 NOT apply chunked more than once to a message body (i.e., chunking an 802 already chunked message is not allowed). If any transfer coding 803 other than chunked is applied to a request payload body, the sender 804 MUST apply chunked as the final transfer coding to ensure that the 805 message is properly framed. If any transfer coding other than 806 chunked is applied to a response payload body, the sender MUST either 807 apply chunked as the final transfer coding or terminate the message 808 by closing the connection. 810 For example, 812 Transfer-Encoding: gzip, chunked 814 indicates that the payload body has been compressed using the gzip 815 coding and then chunked using the chunked coding while forming the 816 message body. 818 Unlike Content-Encoding (Section 6.1.2 of [Semantics]), Transfer- 819 Encoding is a property of the message, not of the representation, and 820 any recipient along the request/response chain MAY decode the 821 received transfer coding(s) or apply additional transfer coding(s) to 822 the message body, assuming that corresponding changes are made to the 823 Transfer-Encoding field value. Additional information about the 824 encoding parameters can be provided by other header fields not 825 defined by this specification. 827 Transfer-Encoding MAY be sent in a response to a HEAD request or in a 828 304 (Not Modified) response (Section 9.4.5 of [Semantics]) to a GET 829 request, neither of which includes a message body, to indicate that 830 the origin server would have applied a transfer coding to the message 831 body if the request had been an unconditional GET. This indication 832 is not required, however, because any recipient on the response chain 833 (including the origin server) can remove transfer codings when they 834 are not needed. 836 A server MUST NOT send a Transfer-Encoding header field in any 837 response with a status code of 1xx (Informational) or 204 (No 838 Content). A server MUST NOT send a Transfer-Encoding header field in 839 any 2xx (Successful) response to a CONNECT request (Section 7.3.6 of 840 [Semantics]). 842 Transfer-Encoding was added in HTTP/1.1. It is generally assumed 843 that implementations advertising only HTTP/1.0 support will not 844 understand how to process a transfer-encoded payload. A client MUST 845 NOT send a request containing Transfer-Encoding unless it knows the 846 server will handle HTTP/1.1 (or later) requests; such knowledge might 847 be in the form of specific user configuration or by remembering the 848 version of a prior received response. A server MUST NOT send a 849 response containing Transfer-Encoding unless the corresponding 850 request indicates HTTP/1.1 (or later). 852 A server that receives a request message with a transfer coding it 853 does not understand SHOULD respond with 501 (Not Implemented). 855 6.2. Content-Length 857 When a message does not have a Transfer-Encoding header field, a 858 Content-Length header field can provide the anticipated size, as a 859 decimal number of octets, for a potential payload body. For messages 860 that do include a payload body, the Content-Length field value 861 provides the framing information necessary for determining where the 862 body (and message) ends. For messages that do not include a payload 863 body, the Content-Length indicates the size of the selected 864 representation (Section 6.2.4 of [Semantics]). 866 Note: HTTP's use of Content-Length for message framing differs 867 significantly from the same field's use in MIME, where it is an 868 optional field used only within the "message/external-body" media- 869 type. 871 6.3. Message Body Length 873 The length of a message body is determined by one of the following 874 (in order of precedence): 876 1. Any response to a HEAD request and any response with a 1xx 877 (Informational), 204 (No Content), or 304 (Not Modified) status 878 code is always terminated by the first empty line after the 879 header fields, regardless of the header fields present in the 880 message, and thus cannot contain a message body. 882 2. Any 2xx (Successful) response to a CONNECT request implies that 883 the connection will become a tunnel immediately after the empty 884 line that concludes the header fields. A client MUST ignore any 885 Content-Length or Transfer-Encoding header fields received in 886 such a message. 888 3. If a Transfer-Encoding header field is present and the chunked 889 transfer coding (Section 7.1) is the final encoding, the message 890 body length is determined by reading and decoding the chunked 891 data until the transfer coding indicates the data is complete. 893 If a Transfer-Encoding header field is present in a response and 894 the chunked transfer coding is not the final encoding, the 895 message body length is determined by reading the connection until 896 it is closed by the server. If a Transfer-Encoding header field 897 is present in a request and the chunked transfer coding is not 898 the final encoding, the message body length cannot be determined 899 reliably; the server MUST respond with the 400 (Bad Request) 900 status code and then close the connection. 902 If a message is received with both a Transfer-Encoding and a 903 Content-Length header field, the Transfer-Encoding overrides the 904 Content-Length. Such a message might indicate an attempt to 905 perform request smuggling (Section 11.2) or response splitting 906 (Section 11.1) and ought to be handled as an error. A sender 907 MUST remove the received Content-Length field prior to forwarding 908 such a message downstream. 910 4. If a message is received without Transfer-Encoding and with 911 either multiple Content-Length header fields having differing 912 field values or a single Content-Length header field having an 913 invalid value, then the message framing is invalid and the 914 recipient MUST treat it as an unrecoverable error. If this is a 915 request message, the server MUST respond with a 400 (Bad Request) 916 status code and then close the connection. If this is a response 917 message received by a proxy, the proxy MUST close the connection 918 to the server, discard the received response, and send a 502 (Bad 919 Gateway) response to the client. If this is a response message 920 received by a user agent, the user agent MUST close the 921 connection to the server and discard the received response. 923 5. If a valid Content-Length header field is present without 924 Transfer-Encoding, its decimal value defines the expected message 925 body length in octets. If the sender closes the connection or 926 the recipient times out before the indicated number of octets are 927 received, the recipient MUST consider the message to be 928 incomplete and close the connection. 930 6. If this is a request message and none of the above are true, then 931 the message body length is zero (no message body is present). 933 7. Otherwise, this is a response message without a declared message 934 body length, so the message body length is determined by the 935 number of octets received prior to the server closing the 936 connection. 938 Since there is no way to distinguish a successfully completed, close- 939 delimited message from a partially received message interrupted by 940 network failure, a server SHOULD generate encoding or length- 941 delimited messages whenever possible. The close-delimiting feature 942 exists primarily for backwards compatibility with HTTP/1.0. 944 A server MAY reject a request that contains a message body but not a 945 Content-Length by responding with 411 (Length Required). 947 Unless a transfer coding other than chunked has been applied, a 948 client that sends a request containing a message body SHOULD use a 949 valid Content-Length header field if the message body length is known 950 in advance, rather than the chunked transfer coding, since some 951 existing services respond to chunked with a 411 (Length Required) 952 status code even though they understand the chunked transfer coding. 953 This is typically because such services are implemented via a gateway 954 that requires a content-length in advance of being called and the 955 server is unable or unwilling to buffer the entire request before 956 processing. 958 A user agent that sends a request containing a message body MUST send 959 a valid Content-Length header field if it does not know the server 960 will handle HTTP/1.1 (or later) requests; such knowledge can be in 961 the form of specific user configuration or by remembering the version 962 of a prior received response. 964 If the final response to the last request on a connection has been 965 completely received and there remains additional data to read, a user 966 agent MAY discard the remaining data or attempt to determine if that 967 data belongs as part of the prior response body, which might be the 968 case if the prior message's Content-Length value is incorrect. A 969 client MUST NOT process, cache, or forward such extra data as a 970 separate response, since such behavior would be vulnerable to cache 971 poisoning. 973 7. Transfer Codings 975 Transfer coding names are used to indicate an encoding transformation 976 that has been, can be, or might need to be applied to a payload body 977 in order to ensure "safe transport" through the network. This 978 differs from a content coding in that the transfer coding is a 979 property of the message rather than a property of the representation 980 that is being transferred. 982 transfer-coding = token *( OWS ";" OWS transfer-parameter ) 984 Parameters are in the form of a name=value pair. 986 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 988 All transfer-coding names are case-insensitive and ought to be 989 registered within the HTTP Transfer Coding registry, as defined in 990 Section 7.3. They are used in the TE (Section 7.4) and Transfer- 991 Encoding (Section 6.1) header fields. 993 +------------+------------------------------------------+-----------+ 994 | Name | Description | Reference | 995 +------------+------------------------------------------+-----------+ 996 | chunked | Transfer in a series of chunks | Section 7 | 997 | | | .1 | 998 | compress | UNIX "compress" data format [Welch] | Section 7 | 999 | | | .2 | 1000 | deflate | "deflate" compressed data ([RFC1951]) | Section 7 | 1001 | | inside the "zlib" data format | .2 | 1002 | | ([RFC1950]) | | 1003 | gzip | GZIP file format [RFC1952] | Section 7 | 1004 | | | .2 | 1005 | trailers | (reserved) | Section 7 | 1006 | x-compress | Deprecated (alias for compress) | Section 7 | 1007 | | | .2 | 1008 | x-gzip | Deprecated (alias for gzip) | Section 7 | 1009 | | | .2 | 1010 +------------+------------------------------------------+-----------+ 1012 Table 2 1014 Note: the coding name "trailers" is reserved because its use would 1015 conflict with the keyword "trailers" in the TE header field 1016 (Section 7.4). 1018 7.1. Chunked Transfer Coding 1020 The chunked transfer coding wraps the payload body in order to 1021 transfer it as a series of chunks, each with its own size indicator, 1022 followed by an OPTIONAL trailer section containing trailer fields. 1023 Chunked enables content streams of unknown size to be transferred as 1024 a sequence of length-delimited buffers, which enables the sender to 1025 retain connection persistence and the recipient to know when it has 1026 received the entire message. 1028 chunked-body = *chunk 1029 last-chunk 1030 trailer-section 1031 CRLF 1033 chunk = chunk-size [ chunk-ext ] CRLF 1034 chunk-data CRLF 1035 chunk-size = 1*HEXDIG 1036 last-chunk = 1*("0") [ chunk-ext ] CRLF 1038 chunk-data = 1*OCTET ; a sequence of chunk-size octets 1040 The chunk-size field is a string of hex digits indicating the size of 1041 the chunk-data in octets. The chunked transfer coding is complete 1042 when a chunk with a chunk-size of zero is received, possibly followed 1043 by a trailer section, and finally terminated by an empty line. 1045 A recipient MUST be able to parse and decode the chunked transfer 1046 coding. 1048 The chunked encoding does not define any parameters. Their presence 1049 SHOULD be treated as an error. 1051 7.1.1. Chunk Extensions 1053 The chunked encoding allows each chunk to include zero or more chunk 1054 extensions, immediately following the chunk-size, for the sake of 1055 supplying per-chunk metadata (such as a signature or hash), mid- 1056 message control information, or randomization of message body size. 1058 chunk-ext = *( BWS ";" BWS chunk-ext-name 1059 [ BWS "=" BWS chunk-ext-val ] ) 1061 chunk-ext-name = token 1062 chunk-ext-val = token / quoted-string 1064 The chunked encoding is specific to each connection and is likely to 1065 be removed or recoded by each recipient (including intermediaries) 1066 before any higher-level application would have a chance to inspect 1067 the extensions. Hence, use of chunk extensions is generally limited 1068 to specialized HTTP services such as "long polling" (where client and 1069 server can have shared expectations regarding the use of chunk 1070 extensions) or for padding within an end-to-end secured connection. 1072 A recipient MUST ignore unrecognized chunk extensions. A server 1073 ought to limit the total length of chunk extensions received in a 1074 request to an amount reasonable for the services provided, in the 1075 same way that it applies length limitations and timeouts for other 1076 parts of a message, and generate an appropriate 4xx (Client Error) 1077 response if that amount is exceeded. 1079 7.1.2. Chunked Trailer Section 1081 A trailer section allows the sender to include additional fields at 1082 the end of a chunked message in order to supply metadata that might 1083 be dynamically generated while the message body is sent, such as a 1084 message integrity check, digital signature, or post-processing 1085 status. The proper use and limitations of trailer fields are defined 1086 in Section 4.6 of [Semantics]. 1088 trailer-section = *( field-line CRLF ) 1090 A recipient that decodes and removes the chunked encoding from a 1091 message (e.g., for storage or forwarding to a non-HTTP/1.1 peer) MUST 1092 discard any received trailer fields, store/forward them separately 1093 from the header fields, or selectively merge into the header section 1094 only those trailer fields corresponding to header field definitions 1095 that are understood by the recipient to explicitly permit and define 1096 how their corresponding trailer field value can be safely merged. 1098 7.1.3. Decoding Chunked 1100 A process for decoding the chunked transfer coding can be represented 1101 in pseudo-code as: 1103 length := 0 1104 read chunk-size, chunk-ext (if any), and CRLF 1105 while (chunk-size > 0) { 1106 read chunk-data and CRLF 1107 append chunk-data to decoded-body 1108 length := length + chunk-size 1109 read chunk-size, chunk-ext (if any), and CRLF 1110 } 1111 read trailer field 1112 while (trailer field is not empty) { 1113 if (trailer fields are stored/forwarded separately) { 1114 append trailer field to existing trailer fields 1115 } 1116 else if (trailer field is understood and defined as mergeable) { 1117 merge trailer field with existing header fields 1118 } 1119 else { 1120 discard trailer field 1121 } 1122 read trailer field 1123 } 1124 Content-Length := length 1125 Remove "chunked" from Transfer-Encoding 1126 Remove Trailer from existing header fields 1128 7.2. Transfer Codings for Compression 1130 The following transfer coding names for compression are defined by 1131 the same algorithm as their corresponding content coding: 1133 compress (and x-compress) 1134 See Section 6.1.2.1 of [Semantics]. 1136 deflate 1137 See Section 6.1.2.2 of [Semantics]. 1139 gzip (and x-gzip) 1140 See Section 6.1.2.3 of [Semantics]. 1142 The compression codings do not define any parameters. Their presence 1143 SHOULD be treated as an error. 1145 7.3. Transfer Coding Registry 1147 The "HTTP Transfer Coding Registry" defines the namespace for 1148 transfer coding names. It is maintained at 1149 . 1151 Registrations MUST include the following fields: 1153 o Name 1155 o Description 1157 o Pointer to specification text 1159 Names of transfer codings MUST NOT overlap with names of content 1160 codings (Section 6.1.2 of [Semantics]) unless the encoding 1161 transformation is identical, as is the case for the compression 1162 codings defined in Section 7.2. 1164 The TE header field (Section 7.4) uses a pseudo parameter named "q" 1165 as rank value when multiple transfer codings are acceptable. Future 1166 registrations of transfer codings SHOULD NOT define parameters called 1167 "q" (case-insensitively) in order to avoid ambiguities. 1169 Values to be added to this namespace require IETF Review (see 1170 Section 4.8 of [RFC8126]), and MUST conform to the purpose of 1171 transfer coding defined in this specification. 1173 Use of program names for the identification of encoding formats is 1174 not desirable and is discouraged for future encodings. 1176 7.4. TE 1178 The "TE" header field in a request indicates what transfer codings, 1179 besides chunked, the client is willing to accept in response, and 1180 whether or not the client is willing to accept trailer fields in a 1181 chunked transfer coding. 1183 The TE field-value consists of a list of transfer coding names, each 1184 allowing for optional parameters (as described in Section 7), and/or 1185 the keyword "trailers". A client MUST NOT send the chunked transfer 1186 coding name in TE; chunked is always acceptable for HTTP/1.1 1187 recipients. 1189 TE = #t-codings 1190 t-codings = "trailers" / ( transfer-coding [ t-ranking ] ) 1191 t-ranking = OWS ";" OWS "q=" rank 1192 rank = ( "0" [ "." 0*3DIGIT ] ) 1193 / ( "1" [ "." 0*3("0") ] ) 1195 Three examples of TE use are below. 1197 TE: deflate 1198 TE: 1199 TE: trailers, deflate;q=0.5 1201 The presence of the keyword "trailers" indicates that the client is 1202 willing to accept trailer fields in a chunked transfer coding, as 1203 defined in Section 7.1.2, on behalf of itself and any downstream 1204 clients. For requests from an intermediary, this implies that 1205 either: (a) all downstream clients are willing to accept trailer 1206 fields in the forwarded response; or, (b) the intermediary will 1207 attempt to buffer the response on behalf of downstream recipients. 1208 Note that HTTP/1.1 does not define any means to limit the size of a 1209 chunked response such that an intermediary can be assured of 1210 buffering the entire response. 1212 When multiple transfer codings are acceptable, the client MAY rank 1213 the codings by preference using a case-insensitive "q" parameter 1214 (similar to the qvalues used in content negotiation fields, 1215 Section 8.4.1 of [Semantics]). The rank value is a real number in 1216 the range 0 through 1, where 0.001 is the least preferred and 1 is 1217 the most preferred; a value of 0 means "not acceptable". 1219 If the TE field value is empty or if no TE field is present, the only 1220 acceptable transfer coding is chunked. A message with no transfer 1221 coding is always acceptable. 1223 Since the TE header field only applies to the immediate connection, a 1224 sender of TE MUST also send a "TE" connection option within the 1225 Connection header field (Section 9.1) in order to prevent the TE 1226 field from being forwarded by intermediaries that do not support its 1227 semantics. 1229 8. Handling Incomplete Messages 1231 A server that receives an incomplete request message, usually due to 1232 a canceled request or a triggered timeout exception, MAY send an 1233 error response prior to closing the connection. 1235 A client that receives an incomplete response message, which can 1236 occur when a connection is closed prematurely or when decoding a 1237 supposedly chunked transfer coding fails, MUST record the message as 1238 incomplete. Cache requirements for incomplete responses are defined 1239 in Section 3 of [Caching]. 1241 If a response terminates in the middle of the header section (before 1242 the empty line is received) and the status code might rely on header 1243 fields to convey the full meaning of the response, then the client 1244 cannot assume that meaning has been conveyed; the client might need 1245 to repeat the request in order to determine what action to take next. 1247 A message body that uses the chunked transfer coding is incomplete if 1248 the zero-sized chunk that terminates the encoding has not been 1249 received. A message that uses a valid Content-Length is incomplete 1250 if the size of the message body received (in octets) is less than the 1251 value given by Content-Length. A response that has neither chunked 1252 transfer coding nor Content-Length is terminated by closure of the 1253 connection and, thus, is considered complete regardless of the number 1254 of message body octets received, provided that the header section was 1255 received intact. 1257 9. Connection Management 1259 HTTP messaging is independent of the underlying transport- or 1260 session-layer connection protocol(s). HTTP only presumes a reliable 1261 transport with in-order delivery of requests and the corresponding 1262 in-order delivery of responses. The mapping of HTTP request and 1263 response structures onto the data units of an underlying transport 1264 protocol is outside the scope of this specification. 1266 As described in Section 5.3 of [Semantics], the specific connection 1267 protocols to be used for an HTTP interaction are determined by client 1268 configuration and the target URI. For example, the "http" URI scheme 1269 (Section 2.5.1 of [Semantics]) indicates a default connection of TCP 1270 over IP, with a default TCP port of 80, but the client might be 1271 configured to use a proxy via some other connection, port, or 1272 protocol. 1274 HTTP implementations are expected to engage in connection management, 1275 which includes maintaining the state of current connections, 1276 establishing a new connection or reusing an existing connection, 1277 processing messages received on a connection, detecting connection 1278 failures, and closing each connection. Most clients maintain 1279 multiple connections in parallel, including more than one connection 1280 per server endpoint. Most servers are designed to maintain thousands 1281 of concurrent connections, while controlling request queues to enable 1282 fair use and detect denial-of-service attacks. 1284 9.1. Connection 1286 The "Connection" header field allows the sender to list desired 1287 control options for the current connection. In order to avoid 1288 confusing downstream recipients, a proxy or gateway MUST remove or 1289 replace any received connection options before forwarding the 1290 message. 1292 When a field aside from Connection is used to supply control 1293 information for or about the current connection, the sender MUST list 1294 the corresponding field name within the Connection header field. A 1295 proxy or gateway MUST parse a received Connection header field before 1296 a message is forwarded and, for each connection-option in this field, 1297 remove any header or trailer field(s) from the message with the same 1298 name as the connection-option, and then remove the Connection header 1299 field itself (or replace it with the intermediary's own connection 1300 options for the forwarded message). 1302 Hence, the Connection header field provides a declarative way of 1303 distinguishing fields that are only intended for the immediate 1304 recipient ("hop-by-hop") from those fields that are intended for all 1305 recipients on the chain ("end-to-end"), enabling the message to be 1306 self-descriptive and allowing future connection-specific extensions 1307 to be deployed without fear that they will be blindly forwarded by 1308 older intermediaries. 1310 The Connection header field's value has the following grammar: 1312 Connection = 1#connection-option 1313 connection-option = token 1315 Connection options are case-insensitive. 1317 A sender MUST NOT send a connection option corresponding to a field 1318 that is intended for all recipients of the payload. For example, 1319 Cache-Control is never appropriate as a connection option 1320 (Section 5.2 of [Caching]). 1322 The connection options do not always correspond to a field present in 1323 the message, since a connection-specific field might not be needed if 1324 there are no parameters associated with a connection option. In 1325 contrast, a connection-specific field that is received without a 1326 corresponding connection option usually indicates that the field has 1327 been improperly forwarded by an intermediary and ought to be ignored 1328 by the recipient. 1330 When defining new connection options, specification authors ought to 1331 document it as reserved field name and register that definition in 1332 the Hypertext Transfer Protocol (HTTP) Field Name Registry 1333 (Section 4.3.2 of [Semantics]), to avoid collisions. 1335 The "close" connection option is defined for a sender to signal that 1336 this connection will be closed after completion of the response. For 1337 example, 1339 Connection: close 1341 in either the request or the response header fields indicates that 1342 the sender is going to close the connection after the current 1343 request/response is complete (Section 9.7). 1345 A client that does not support persistent connections MUST send the 1346 "close" connection option in every request message. 1348 A server that does not support persistent connections MUST send the 1349 "close" connection option in every response message that does not 1350 have a 1xx (Informational) status code. 1352 9.2. Establishment 1354 It is beyond the scope of this specification to describe how 1355 connections are established via various transport- or session-layer 1356 protocols. Each connection applies to only one transport link. 1358 9.3. Associating a Response to a Request 1360 HTTP/1.1 does not include a request identifier for associating a 1361 given request message with its corresponding one or more response 1362 messages. Hence, it relies on the order of response arrival to 1363 correspond exactly to the order in which requests are made on the 1364 same connection. More than one response message per request only 1365 occurs when one or more informational responses (1xx, see Section 9.2 1366 of [Semantics]) precede a final response to the same request. 1368 A client that has more than one outstanding request on a connection 1369 MUST maintain a list of outstanding requests in the order sent and 1370 MUST associate each received response message on that connection to 1371 the highest ordered request that has not yet received a final (non- 1372 1xx) response. 1374 If an HTTP/1.1 client receives data on a connection that doesn't have 1375 any outstanding requests, it MUST NOT consider them to be a response 1376 to a not-yet-issued request; it SHOULD close the connection, since 1377 message delimitation is now ambiguous, unless the data consists only 1378 of one or more CRLF (which can be discarded, as per Section 2.2). 1380 9.4. Persistence 1382 HTTP/1.1 defaults to the use of "persistent connections", allowing 1383 multiple requests and responses to be carried over a single 1384 connection. The "close" connection option is used to signal that a 1385 connection will not persist after the current request/response. HTTP 1386 implementations SHOULD support persistent connections. 1388 A recipient determines whether a connection is persistent or not 1389 based on the most recently received message's protocol version and 1390 Connection header field (if any): 1392 o If the "close" connection option is present, the connection will 1393 not persist after the current response; else, 1395 o If the received protocol is HTTP/1.1 (or later), the connection 1396 will persist after the current response; else, 1398 o If the received protocol is HTTP/1.0, the "keep-alive" connection 1399 option is present, either the recipient is not a proxy or the 1400 message is a response, and the recipient wishes to honor the 1401 HTTP/1.0 "keep-alive" mechanism, the connection will persist after 1402 the current response; otherwise, 1404 o The connection will close after the current response. 1406 A client MAY send additional requests on a persistent connection 1407 until it sends or receives a "close" connection option or receives an 1408 HTTP/1.0 response without a "keep-alive" connection option. 1410 In order to remain persistent, all messages on a connection need to 1411 have a self-defined message length (i.e., one not defined by closure 1412 of the connection), as described in Section 6. A server MUST read 1413 the entire request message body or close the connection after sending 1414 its response, since otherwise the remaining data on a persistent 1415 connection would be misinterpreted as the next request. Likewise, a 1416 client MUST read the entire response message body if it intends to 1417 reuse the same connection for a subsequent request. 1419 A proxy server MUST NOT maintain a persistent connection with an 1420 HTTP/1.0 client (see Section 19.7.1 of [RFC2068] for information and 1421 discussion of the problems with the Keep-Alive header field 1422 implemented by many HTTP/1.0 clients). 1424 See Appendix C.1.2 for more information on backwards compatibility 1425 with HTTP/1.0 clients. 1427 9.4.1. Retrying Requests 1429 Connections can be closed at any time, with or without intention. 1430 Implementations ought to anticipate the need to recover from 1431 asynchronous close events. The conditions under which a client can 1432 automatically retry a sequence of outstanding requests are defined in 1433 Section 7.2.2 of [Semantics]. 1435 9.4.2. Pipelining 1437 A client that supports persistent connections MAY "pipeline" its 1438 requests (i.e., send multiple requests without waiting for each 1439 response). A server MAY process a sequence of pipelined requests in 1440 parallel if they all have safe methods (Section 7.2.1 of 1441 [Semantics]), but it MUST send the corresponding responses in the 1442 same order that the requests were received. 1444 A client that pipelines requests SHOULD retry unanswered requests if 1445 the connection closes before it receives all of the corresponding 1446 responses. When retrying pipelined requests after a failed 1447 connection (a connection not explicitly closed by the server in its 1448 last complete response), a client MUST NOT pipeline immediately after 1449 connection establishment, since the first remaining request in the 1450 prior pipeline might have caused an error response that can be lost 1451 again if multiple requests are sent on a prematurely closed 1452 connection (see the TCP reset problem described in Section 9.7). 1454 Idempotent methods (Section 7.2.2 of [Semantics]) are significant to 1455 pipelining because they can be automatically retried after a 1456 connection failure. A user agent SHOULD NOT pipeline requests after 1457 a non-idempotent method, until the final response status code for 1458 that method has been received, unless the user agent has a means to 1459 detect and recover from partial failure conditions involving the 1460 pipelined sequence. 1462 An intermediary that receives pipelined requests MAY pipeline those 1463 requests when forwarding them inbound, since it can rely on the 1464 outbound user agent(s) to determine what requests can be safely 1465 pipelined. If the inbound connection fails before receiving a 1466 response, the pipelining intermediary MAY attempt to retry a sequence 1467 of requests that have yet to receive a response if the requests all 1468 have idempotent methods; otherwise, the pipelining intermediary 1469 SHOULD forward any received responses and then close the 1470 corresponding outbound connection(s) so that the outbound user 1471 agent(s) can recover accordingly. 1473 9.5. Concurrency 1475 A client ought to limit the number of simultaneous open connections 1476 that it maintains to a given server. 1478 Previous revisions of HTTP gave a specific number of connections as a 1479 ceiling, but this was found to be impractical for many applications. 1480 As a result, this specification does not mandate a particular maximum 1481 number of connections but, instead, encourages clients to be 1482 conservative when opening multiple connections. 1484 Multiple connections are typically used to avoid the "head-of-line 1485 blocking" problem, wherein a request that takes significant server- 1486 side processing and/or has a large payload blocks subsequent requests 1487 on the same connection. However, each connection consumes server 1488 resources. Furthermore, using multiple connections can cause 1489 undesirable side effects in congested networks. 1491 Note that a server might reject traffic that it deems abusive or 1492 characteristic of a denial-of-service attack, such as an excessive 1493 number of open connections from a single client. 1495 9.6. Failures and Timeouts 1497 Servers will usually have some timeout value beyond which they will 1498 no longer maintain an inactive connection. Proxy servers might make 1499 this a higher value since it is likely that the client will be making 1500 more connections through the same proxy server. The use of 1501 persistent connections places no requirements on the length (or 1502 existence) of this timeout for either the client or the server. 1504 A client or server that wishes to time out SHOULD issue a graceful 1505 close on the connection. Implementations SHOULD constantly monitor 1506 open connections for a received closure signal and respond to it as 1507 appropriate, since prompt closure of both sides of a connection 1508 enables allocated system resources to be reclaimed. 1510 A client, server, or proxy MAY close the transport connection at any 1511 time. For example, a client might have started to send a new request 1512 at the same time that the server has decided to close the "idle" 1513 connection. From the server's point of view, the connection is being 1514 closed while it was idle, but from the client's point of view, a 1515 request is in progress. 1517 A server SHOULD sustain persistent connections, when possible, and 1518 allow the underlying transport's flow-control mechanisms to resolve 1519 temporary overloads, rather than terminate connections with the 1520 expectation that clients will retry. The latter technique can 1521 exacerbate network congestion. 1523 A client sending a message body SHOULD monitor the network connection 1524 for an error response while it is transmitting the request. If the 1525 client sees a response that indicates the server does not wish to 1526 receive the message body and is closing the connection, the client 1527 SHOULD immediately cease transmitting the body and close its side of 1528 the connection. 1530 9.7. Tear-down 1532 The Connection header field (Section 9.1) provides a "close" 1533 connection option that a sender SHOULD send when it wishes to close 1534 the connection after the current request/response pair. 1536 A client that sends a "close" connection option MUST NOT send further 1537 requests on that connection (after the one containing "close") and 1538 MUST close the connection after reading the final response message 1539 corresponding to this request. 1541 A server that receives a "close" connection option MUST initiate a 1542 close of the connection (see below) after it sends the final response 1543 to the request that contained "close". The server SHOULD send a 1544 "close" connection option in its final response on that connection. 1545 The server MUST NOT process any further requests received on that 1546 connection. 1548 A server that sends a "close" connection option MUST initiate a close 1549 of the connection (see below) after it sends the response containing 1550 "close". The server MUST NOT process any further requests received 1551 on that connection. 1553 A client that receives a "close" connection option MUST cease sending 1554 requests on that connection and close the connection after reading 1555 the response message containing the "close"; if additional pipelined 1556 requests had been sent on the connection, the client SHOULD NOT 1557 assume that they will be processed by the server. 1559 If a server performs an immediate close of a TCP connection, there is 1560 a significant risk that the client will not be able to read the last 1561 HTTP response. If the server receives additional data from the 1562 client on a fully closed connection, such as another request that was 1563 sent by the client before receiving the server's response, the 1564 server's TCP stack will send a reset packet to the client; 1565 unfortunately, the reset packet might erase the client's 1566 unacknowledged input buffers before they can be read and interpreted 1567 by the client's HTTP parser. 1569 To avoid the TCP reset problem, servers typically close a connection 1570 in stages. First, the server performs a half-close by closing only 1571 the write side of the read/write connection. The server then 1572 continues to read from the connection until it receives a 1573 corresponding close by the client, or until the server is reasonably 1574 certain that its own TCP stack has received the client's 1575 acknowledgement of the packet(s) containing the server's last 1576 response. Finally, the server fully closes the connection. 1578 It is unknown whether the reset problem is exclusive to TCP or might 1579 also be found in other transport connection protocols. 1581 9.8. TLS Connection Closure 1583 TLS provides a facility for secure connection closure. When a valid 1584 closure alert is received, an implementation can be assured that no 1585 further data will be received on that connection. TLS 1586 implementations MUST initiate an exchange of closure alerts before 1587 closing a connection. A TLS implementation MAY, after sending a 1588 closure alert, close the connection without waiting for the peer to 1589 send its closure alert, generating an "incomplete close". Note that 1590 an implementation which does this MAY choose to reuse the session. 1591 This SHOULD only be done when the application knows (typically 1592 through detecting HTTP message boundaries) that it has received all 1593 the message data that it cares about. 1595 As specified in [RFC8446], any implementation which receives a 1596 connection close without first receiving a valid closure alert (a 1597 "premature close") MUST NOT reuse that session. Note that a 1598 premature close does not call into question the security of the data 1599 already received, but simply indicates that subsequent data might 1600 have been truncated. Because TLS is oblivious to HTTP request/ 1601 response boundaries, it is necessary to examine the HTTP data itself 1602 (specifically the Content-Length header) to determine whether the 1603 truncation occurred inside a message or between messages. 1605 When encountering a premature close, a client SHOULD treat as 1606 completed all requests for which it has received as much data as 1607 specified in the Content-Length header. 1609 A client detecting an incomplete close SHOULD recover gracefully. It 1610 MAY resume a TLS session closed in this fashion. 1612 Clients MUST send a closure alert before closing the connection. 1613 Clients which are unprepared to receive any more data MAY choose not 1614 to wait for the server's closure alert and simply close the 1615 connection, thus generating an incomplete close on the server side. 1617 Servers SHOULD be prepared to receive an incomplete close from the 1618 client, since the client can often determine when the end of server 1619 data is. Servers SHOULD be willing to resume TLS sessions closed in 1620 this fashion. 1622 Servers MUST attempt to initiate an exchange of closure alerts with 1623 the client before closing the connection. Servers MAY close the 1624 connection after sending the closure alert, thus generating an 1625 incomplete close on the client side. 1627 9.9. Upgrade 1629 The "Upgrade" header field is intended to provide a simple mechanism 1630 for transitioning from HTTP/1.1 to some other protocol on the same 1631 connection. 1633 A client MAY send a list of protocol names in the Upgrade header 1634 field of a request to invite the server to switch to one or more of 1635 the named protocols, in order of descending preference, before 1636 sending the final response. A server MAY ignore a received Upgrade 1637 header field if it wishes to continue using the current protocol on 1638 that connection. Upgrade cannot be used to insist on a protocol 1639 change. 1641 Upgrade = 1#protocol 1643 protocol = protocol-name ["/" protocol-version] 1644 protocol-name = token 1645 protocol-version = token 1647 Although protocol names are registered with a preferred case, 1648 recipients SHOULD use case-insensitive comparison when matching each 1649 protocol-name to supported protocols. 1651 A server that sends a 101 (Switching Protocols) response MUST send an 1652 Upgrade header field to indicate the new protocol(s) to which the 1653 connection is being switched; if multiple protocol layers are being 1654 switched, the sender MUST list the protocols in layer-ascending 1655 order. A server MUST NOT switch to a protocol that was not indicated 1656 by the client in the corresponding request's Upgrade header field. A 1657 server MAY choose to ignore the order of preference indicated by the 1658 client and select the new protocol(s) based on other factors, such as 1659 the nature of the request or the current load on the server. 1661 A server that sends a 426 (Upgrade Required) response MUST send an 1662 Upgrade header field to indicate the acceptable protocols, in order 1663 of descending preference. 1665 A server MAY send an Upgrade header field in any other response to 1666 advertise that it implements support for upgrading to the listed 1667 protocols, in order of descending preference, when appropriate for a 1668 future request. 1670 The following is a hypothetical example sent by a client: 1672 GET /hello HTTP/1.1 1673 Host: www.example.com 1674 Connection: upgrade 1675 Upgrade: websocket, IRC/6.9, RTA/x11 1677 The capabilities and nature of the application-level communication 1678 after the protocol change is entirely dependent upon the new 1679 protocol(s) chosen. However, immediately after sending the 101 1680 (Switching Protocols) response, the server is expected to continue 1681 responding to the original request as if it had received its 1682 equivalent within the new protocol (i.e., the server still has an 1683 outstanding request to satisfy after the protocol has been changed, 1684 and is expected to do so without requiring the request to be 1685 repeated). 1687 For example, if the Upgrade header field is received in a GET request 1688 and the server decides to switch protocols, it first responds with a 1689 101 (Switching Protocols) message in HTTP/1.1 and then immediately 1690 follows that with the new protocol's equivalent of a response to a 1691 GET on the target resource. This allows a connection to be upgraded 1692 to protocols with the same semantics as HTTP without the latency cost 1693 of an additional round trip. A server MUST NOT switch protocols 1694 unless the received message semantics can be honored by the new 1695 protocol; an OPTIONS request can be honored by any protocol. 1697 The following is an example response to the above hypothetical 1698 request: 1700 HTTP/1.1 101 Switching Protocols 1701 Connection: upgrade 1702 Upgrade: websocket 1704 [... data stream switches to websocket with an appropriate response 1705 (as defined by new protocol) to the "GET /hello" request ...] 1707 When Upgrade is sent, the sender MUST also send a Connection header 1708 field (Section 9.1) that contains an "upgrade" connection option, in 1709 order to prevent Upgrade from being accidentally forwarded by 1710 intermediaries that might not implement the listed protocols. A 1711 server MUST ignore an Upgrade header field that is received in an 1712 HTTP/1.0 request. 1714 A client cannot begin using an upgraded protocol on the connection 1715 until it has completely sent the request message (i.e., the client 1716 can't change the protocol it is sending in the middle of a message). 1717 If a server receives both an Upgrade and an Expect header field with 1718 the "100-continue" expectation (Section 8.1.1 of [Semantics]), the 1719 server MUST send a 100 (Continue) response before sending a 101 1720 (Switching Protocols) response. 1722 The Upgrade header field only applies to switching protocols on top 1723 of the existing connection; it cannot be used to switch the 1724 underlying connection (transport) protocol, nor to switch the 1725 existing communication to a different connection. For those 1726 purposes, it is more appropriate to use a 3xx (Redirection) response 1727 (Section 9.4 of [Semantics]). 1729 9.9.1. Upgrade Protocol Names 1731 This specification only defines the protocol name "HTTP" for use by 1732 the family of Hypertext Transfer Protocols, as defined by the HTTP 1733 version rules of Section 3.5 of [Semantics] and future updates to 1734 this specification. Additional protocol names ought to be registered 1735 using the registration procedure defined in Section 9.9.2. 1737 +------+-------------------+--------------------+-------------------+ 1738 | Name | Description | Expected Version | Reference | 1739 | | | Tokens | | 1740 +------+-------------------+--------------------+-------------------+ 1741 | HTTP | Hypertext | any DIGIT.DIGIT | Section 3.5 of | 1742 | | Transfer Protocol | (e.g, "2.0") | [Semantics] | 1743 +------+-------------------+--------------------+-------------------+ 1745 9.9.2. Upgrade Token Registry 1747 The "Hypertext Transfer Protocol (HTTP) Upgrade Token Registry" 1748 defines the namespace for protocol-name tokens used to identify 1749 protocols in the Upgrade header field. The registry is maintained at 1750 . 1752 Each registered protocol name is associated with contact information 1753 and an optional set of specifications that details how the connection 1754 will be processed after it has been upgraded. 1756 Registrations happen on a "First Come First Served" basis (see 1757 Section 4.4 of [RFC8126]) and are subject to the following rules: 1759 1. A protocol-name token, once registered, stays registered forever. 1761 2. A protocol-name token is case-insensitive and registered with the 1762 preferred case to be generated by senders. 1764 3. The registration MUST name a responsible party for the 1765 registration. 1767 4. The registration MUST name a point of contact. 1769 5. The registration MAY name a set of specifications associated with 1770 that token. Such specifications need not be publicly available. 1772 6. The registration SHOULD name a set of expected "protocol-version" 1773 tokens associated with that token at the time of registration. 1775 7. The responsible party MAY change the registration at any time. 1776 The IANA will keep a record of all such changes, and make them 1777 available upon request. 1779 8. The IESG MAY reassign responsibility for a protocol token. This 1780 will normally only be used in the case when a responsible party 1781 cannot be contacted. 1783 10. Enclosing Messages as Data 1785 10.1. Media Type message/http 1787 The message/http media type can be used to enclose a single HTTP 1788 request or response message, provided that it obeys the MIME 1789 restrictions for all "message" types regarding line length and 1790 encodings. 1792 Type name: message 1794 Subtype name: http 1796 Required parameters: N/A 1798 Optional parameters: version, msgtype 1799 version: The HTTP-version number of the enclosed message (e.g., 1800 "1.1"). If not present, the version can be determined from the 1801 first line of the body. 1803 msgtype: The message type -- "request" or "response". If not 1804 present, the type can be determined from the first line of the 1805 body. 1807 Encoding considerations: only "7bit", "8bit", or "binary" are 1808 permitted 1810 Security considerations: see Section 11 1812 Interoperability considerations: N/A 1814 Published specification: This specification (see Section 10.1). 1816 Applications that use this media type: N/A 1818 Fragment identifier considerations: N/A 1820 Additional information: 1822 Magic number(s): N/A 1824 Deprecated alias names for this type: N/A 1826 File extension(s): N/A 1828 Macintosh file type code(s): N/A 1830 Person and email address to contact for further information: 1831 See Authors' Addresses section. 1833 Intended usage: COMMON 1835 Restrictions on usage: N/A 1837 Author: See Authors' Addresses section. 1839 Change controller: IESG 1841 10.2. Media Type application/http 1843 The application/http media type can be used to enclose a pipeline of 1844 one or more HTTP request or response messages (not intermixed). 1846 Type name: application 1847 Subtype name: http 1849 Required parameters: N/A 1851 Optional parameters: version, msgtype 1853 version: The HTTP-version number of the enclosed messages (e.g., 1854 "1.1"). If not present, the version can be determined from the 1855 first line of the body. 1857 msgtype: The message type -- "request" or "response". If not 1858 present, the type can be determined from the first line of the 1859 body. 1861 Encoding considerations: HTTP messages enclosed by this type are in 1862 "binary" format; use of an appropriate Content-Transfer-Encoding 1863 is required when transmitted via email. 1865 Security considerations: see Section 11 1867 Interoperability considerations: N/A 1869 Published specification: This specification (see Section 10.2). 1871 Applications that use this media type: N/A 1873 Fragment identifier considerations: N/A 1875 Additional information: 1877 Deprecated alias names for this type: N/A 1879 Magic number(s): N/A 1881 File extension(s): N/A 1883 Macintosh file type code(s): N/A 1885 Person and email address to contact for further information: 1886 See Authors' Addresses section. 1888 Intended usage: COMMON 1890 Restrictions on usage: N/A 1892 Author: See Authors' Addresses section. 1894 Change controller: IESG 1896 11. Security Considerations 1898 This section is meant to inform developers, information providers, 1899 and users of known security considerations relevant to HTTP message 1900 syntax, parsing, and routing. Security considerations about HTTP 1901 semantics and payloads are addressed in [Semantics]. 1903 11.1. Response Splitting 1905 Response splitting (a.k.a, CRLF injection) is a common technique, 1906 used in various attacks on Web usage, that exploits the line-based 1907 nature of HTTP message framing and the ordered association of 1908 requests to responses on persistent connections [Klein]. This 1909 technique can be particularly damaging when the requests pass through 1910 a shared cache. 1912 Response splitting exploits a vulnerability in servers (usually 1913 within an application server) where an attacker can send encoded data 1914 within some parameter of the request that is later decoded and echoed 1915 within any of the response header fields of the response. If the 1916 decoded data is crafted to look like the response has ended and a 1917 subsequent response has begun, the response has been split and the 1918 content within the apparent second response is controlled by the 1919 attacker. The attacker can then make any other request on the same 1920 persistent connection and trick the recipients (including 1921 intermediaries) into believing that the second half of the split is 1922 an authoritative answer to the second request. 1924 For example, a parameter within the request-target might be read by 1925 an application server and reused within a redirect, resulting in the 1926 same parameter being echoed in the Location header field of the 1927 response. If the parameter is decoded by the application and not 1928 properly encoded when placed in the response field, the attacker can 1929 send encoded CRLF octets and other content that will make the 1930 application's single response look like two or more responses. 1932 A common defense against response splitting is to filter requests for 1933 data that looks like encoded CR and LF (e.g., "%0D" and "%0A"). 1934 However, that assumes the application server is only performing URI 1935 decoding, rather than more obscure data transformations like charset 1936 transcoding, XML entity translation, base64 decoding, sprintf 1937 reformatting, etc. A more effective mitigation is to prevent 1938 anything other than the server's core protocol libraries from sending 1939 a CR or LF within the header section, which means restricting the 1940 output of header fields to APIs that filter for bad octets and not 1941 allowing application servers to write directly to the protocol 1942 stream. 1944 11.2. Request Smuggling 1946 Request smuggling ([Linhart]) is a technique that exploits 1947 differences in protocol parsing among various recipients to hide 1948 additional requests (which might otherwise be blocked or disabled by 1949 policy) within an apparently harmless request. Like response 1950 splitting, request smuggling can lead to a variety of attacks on HTTP 1951 usage. 1953 This specification has introduced new requirements on request 1954 parsing, particularly with regard to message framing in Section 6.3, 1955 to reduce the effectiveness of request smuggling. 1957 11.3. Message Integrity 1959 HTTP does not define a specific mechanism for ensuring message 1960 integrity, instead relying on the error-detection ability of 1961 underlying transport protocols and the use of length or chunk- 1962 delimited framing to detect completeness. Additional integrity 1963 mechanisms, such as hash functions or digital signatures applied to 1964 the content, can be selectively added to messages via extensible 1965 metadata fields. Historically, the lack of a single integrity 1966 mechanism has been justified by the informal nature of most HTTP 1967 communication. However, the prevalence of HTTP as an information 1968 access mechanism has resulted in its increasing use within 1969 environments where verification of message integrity is crucial. 1971 User agents are encouraged to implement configurable means for 1972 detecting and reporting failures of message integrity such that those 1973 means can be enabled within environments for which integrity is 1974 necessary. For example, a browser being used to view medical history 1975 or drug interaction information needs to indicate to the user when 1976 such information is detected by the protocol to be incomplete, 1977 expired, or corrupted during transfer. Such mechanisms might be 1978 selectively enabled via user agent extensions or the presence of 1979 message integrity metadata in a response. At a minimum, user agents 1980 ought to provide some indication that allows a user to distinguish 1981 between a complete and incomplete response message (Section 8) when 1982 such verification is desired. 1984 11.4. Message Confidentiality 1986 HTTP relies on underlying transport protocols to provide message 1987 confidentiality when that is desired. HTTP has been specifically 1988 designed to be independent of the transport protocol, such that it 1989 can be used over many different forms of encrypted connection, with 1990 the selection of such transports being identified by the choice of 1991 URI scheme or within user agent configuration. 1993 The "https" scheme can be used to identify resources that require a 1994 confidential connection, as described in Section 2.5.2 of 1995 [Semantics]. 1997 12. IANA Considerations 1999 The change controller for the following registrations is: "IETF 2000 (iesg@ietf.org) - Internet Engineering Task Force". 2002 12.1. Field Name Registration 2004 Please update the "Hypertext Transfer Protocol (HTTP) Field Name 2005 Registry" at with the 2006 field names listed in the two tables of Section 5. 2008 12.2. Media Type Registration 2010 Please update the "Media Types" registry at 2011 with the registration 2012 information in Section 10.1 and Section 10.2 for the media types 2013 "message/http" and "application/http", respectively. 2015 12.3. Transfer Coding Registration 2017 Please update the "HTTP Transfer Coding Registry" at 2018 with the 2019 registration procedure of Section 7.3 and the content coding names 2020 summarized in the table of Section 7. 2022 12.4. Upgrade Token Registration 2024 Please update the "Hypertext Transfer Protocol (HTTP) Upgrade Token 2025 Registry" at 2026 with the registration procedure of Section 9.9.2 and the upgrade 2027 token names summarized in the table of Section 9.9.1. 2029 12.5. ALPN Protocol ID Registration 2031 Please update the "TLS Application-Layer Protocol Negotiation (ALPN) 2032 Protocol IDs" registry at with the 2034 registration below: 2036 +----------+--------------------------------------+-----------------+ 2037 | Protocol | Identification Sequence | Reference | 2038 +----------+--------------------------------------+-----------------+ 2039 | HTTP/1.1 | 0x68 0x74 0x74 0x70 0x2f 0x31 0x2e | (this | 2040 | | 0x31 ("http/1.1") | specification) | 2041 +----------+--------------------------------------+-----------------+ 2043 13. References 2045 13.1. Normative References 2047 [Caching] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 2048 Ed., "HTTP Caching", draft-ietf-httpbis-cache-07 (work in 2049 progress), March 2020. 2051 [RFC1950] Deutsch, L. and J-L. Gailly, "ZLIB Compressed Data Format 2052 Specification version 3.3", RFC 1950, 2053 DOI 10.17487/RFC1950, May 1996, 2054 . 2056 [RFC1951] Deutsch, P., "DEFLATE Compressed Data Format Specification 2057 version 1.3", RFC 1951, DOI 10.17487/RFC1951, May 1996, 2058 . 2060 [RFC1952] Deutsch, P., Gailly, J-L., Adler, M., Deutsch, L., and G. 2061 Randers-Pehrson, "GZIP file format specification version 2062 4.3", RFC 1952, DOI 10.17487/RFC1952, May 1996, 2063 . 2065 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2066 Requirement Levels", BCP 14, RFC 2119, 2067 DOI 10.17487/RFC2119, March 1997, 2068 . 2070 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 2071 Resource Identifier (URI): Generic Syntax", STD 66, 2072 RFC 3986, DOI 10.17487/RFC3986, January 2005, 2073 . 2075 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 2076 Specifications: ABNF", STD 68, RFC 5234, 2077 DOI 10.17487/RFC5234, January 2008, 2078 . 2080 [RFC7405] Kyzivat, P., "Case-Sensitive String Support in ABNF", 2081 RFC 7405, DOI 10.17487/RFC7405, December 2014, 2082 . 2084 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2085 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 2086 May 2017, . 2088 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 2089 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 2090 . 2092 [Semantics] 2093 Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 2094 Ed., "HTTP Semantics", draft-ietf-httpbis-semantics-07 2095 (work in progress), March 2020. 2097 [USASCII] American National Standards Institute, "Coded Character 2098 Set -- 7-bit American Standard Code for Information 2099 Interchange", ANSI X3.4, 1986. 2101 [Welch] Welch, T., "A Technique for High-Performance Data 2102 Compression", IEEE Computer 17(6), June 1984. 2104 13.2. Informative References 2106 [Err4667] RFC Errata, Erratum ID 4667, RFC 7230, 2107 . 2109 [Klein] Klein, A., "Divide and Conquer - HTTP Response Splitting, 2110 Web Cache Poisoning Attacks, and Related Topics", March 2111 2004, . 2114 [Linhart] Linhart, C., Klein, A., Heled, R., and S. Orrin, "HTTP 2115 Request Smuggling", June 2005, 2116 . 2118 [RFC1945] Berners-Lee, T., Fielding, R., and H. Nielsen, "Hypertext 2119 Transfer Protocol -- HTTP/1.0", RFC 1945, 2120 DOI 10.17487/RFC1945, May 1996, 2121 . 2123 [RFC2045] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2124 Extensions (MIME) Part One: Format of Internet Message 2125 Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996, 2126 . 2128 [RFC2046] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2129 Extensions (MIME) Part Two: Media Types", RFC 2046, 2130 DOI 10.17487/RFC2046, November 1996, 2131 . 2133 [RFC2049] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2134 Extensions (MIME) Part Five: Conformance Criteria and 2135 Examples", RFC 2049, DOI 10.17487/RFC2049, November 1996, 2136 . 2138 [RFC2068] Fielding, R., Gettys, J., Mogul, J., Nielsen, H., and T. 2139 Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", 2140 RFC 2068, DOI 10.17487/RFC2068, January 1997, 2141 . 2143 [RFC2557] Palme, F., Hopmann, A., Shelness, N., and E. Stefferud, 2144 "MIME Encapsulation of Aggregate Documents, such as HTML 2145 (MHTML)", RFC 2557, DOI 10.17487/RFC2557, March 1999, 2146 . 2148 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 2149 DOI 10.17487/RFC5322, October 2008, 2150 . 2152 [RFC7230] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2153 Protocol (HTTP/1.1): Message Syntax and Routing", 2154 RFC 7230, DOI 10.17487/RFC7230, June 2014, 2155 . 2157 [RFC7231] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2158 Protocol (HTTP/1.1): Semantics and Content", RFC 7231, 2159 DOI 10.17487/RFC7231, June 2014, 2160 . 2162 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2163 Writing an IANA Considerations Section in RFCs", BCP 26, 2164 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2165 . 2167 Appendix A. Collected ABNF 2169 In the collected ABNF below, list rules are expanded as per 2170 Section 4.5 of [Semantics]. 2172 BWS = 2174 Connection = [ connection-option ] *( OWS "," OWS [ connection-option 2175 ] ) 2177 HTTP-message = start-line CRLF *( field-line CRLF ) CRLF [ 2178 message-body ] 2179 HTTP-name = %x48.54.54.50 ; HTTP 2180 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 2182 OWS = 2184 RWS = 2186 TE = [ t-codings ] *( OWS "," OWS [ t-codings ] ) 2187 Transfer-Encoding = [ transfer-coding ] *( OWS "," OWS [ 2188 transfer-coding ] ) 2190 Upgrade = [ protocol ] *( OWS "," OWS [ protocol ] ) 2192 absolute-URI = 2193 absolute-form = absolute-URI 2194 absolute-path = 2195 asterisk-form = "*" 2196 authority = 2197 authority-form = authority 2199 chunk = chunk-size [ chunk-ext ] CRLF chunk-data CRLF 2200 chunk-data = 1*OCTET 2201 chunk-ext = *( BWS ";" BWS chunk-ext-name [ BWS "=" BWS chunk-ext-val 2202 ] ) 2203 chunk-ext-name = token 2204 chunk-ext-val = token / quoted-string 2205 chunk-size = 1*HEXDIG 2206 chunked-body = *chunk last-chunk trailer-section CRLF 2207 comment = 2208 connection-option = token 2210 field-line = field-name ":" OWS field-value OWS 2211 field-name = 2212 field-value = 2214 last-chunk = 1*"0" [ chunk-ext ] CRLF 2215 message-body = *OCTET 2216 method = token 2218 obs-fold = OWS CRLF RWS 2219 obs-text = 2220 origin-form = absolute-path [ "?" query ] 2222 port = 2223 protocol = protocol-name [ "/" protocol-version ] 2224 protocol-name = token 2225 protocol-version = token 2227 query = 2228 quoted-string = 2230 rank = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] ) 2231 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 2232 request-line = method SP request-target SP HTTP-version 2233 request-target = origin-form / absolute-form / authority-form / 2234 asterisk-form 2236 start-line = request-line / status-line 2237 status-code = 3DIGIT 2238 status-line = HTTP-version SP status-code SP [ reason-phrase ] 2240 t-codings = "trailers" / ( transfer-coding [ t-ranking ] ) 2241 t-ranking = OWS ";" OWS "q=" rank 2242 token = 2243 trailer-section = *( field-line CRLF ) 2244 transfer-coding = token *( OWS ";" OWS transfer-parameter ) 2245 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 2247 uri-host = 2249 Appendix B. Differences between HTTP and MIME 2251 HTTP/1.1 uses many of the constructs defined for the Internet Message 2252 Format [RFC5322] and the Multipurpose Internet Mail Extensions (MIME) 2253 [RFC2045] to allow a message body to be transmitted in an open 2254 variety of representations and with extensible fields. However, RFC 2255 2045 is focused only on email; applications of HTTP have many 2256 characteristics that differ from email; hence, HTTP has features that 2257 differ from MIME. These differences were carefully chosen to 2258 optimize performance over binary connections, to allow greater 2259 freedom in the use of new media types, to make date comparisons 2260 easier, and to acknowledge the practice of some early HTTP servers 2261 and clients. 2263 This appendix describes specific areas where HTTP differs from MIME. 2264 Proxies and gateways to and from strict MIME environments need to be 2265 aware of these differences and provide the appropriate conversions 2266 where necessary. 2268 B.1. MIME-Version 2270 HTTP is not a MIME-compliant protocol. However, messages can include 2271 a single MIME-Version header field to indicate what version of the 2272 MIME protocol was used to construct the message. Use of the MIME- 2273 Version header field indicates that the message is in full 2274 conformance with the MIME protocol (as defined in [RFC2045]). 2275 Senders are responsible for ensuring full conformance (where 2276 possible) when exporting HTTP messages to strict MIME environments. 2278 B.2. Conversion to Canonical Form 2280 MIME requires that an Internet mail body part be converted to 2281 canonical form prior to being transferred, as described in Section 4 2282 of [RFC2049]. Section 6.1.1.2 of [Semantics] describes the forms 2283 allowed for subtypes of the "text" media type when transmitted over 2284 HTTP. [RFC2046] requires that content with a type of "text" 2285 represent line breaks as CRLF and forbids the use of CR or LF outside 2286 of line break sequences. HTTP allows CRLF, bare CR, and bare LF to 2287 indicate a line break within text content. 2289 A proxy or gateway from HTTP to a strict MIME environment ought to 2290 translate all line breaks within text media types to the RFC 2049 2291 canonical form of CRLF. Note, however, this might be complicated by 2292 the presence of a Content-Encoding and by the fact that HTTP allows 2293 the use of some charsets that do not use octets 13 and 10 to 2294 represent CR and LF, respectively. 2296 Conversion will break any cryptographic checksums applied to the 2297 original content unless the original content is already in canonical 2298 form. Therefore, the canonical form is recommended for any content 2299 that uses such checksums in HTTP. 2301 B.3. Conversion of Date Formats 2303 HTTP/1.1 uses a restricted set of date formats (Section 10.1.1.1 of 2304 [Semantics]) to simplify the process of date comparison. Proxies and 2305 gateways from other protocols ought to ensure that any Date header 2306 field present in a message conforms to one of the HTTP/1.1 formats 2307 and rewrite the date if necessary. 2309 B.4. Conversion of Content-Encoding 2311 MIME does not include any concept equivalent to HTTP/1.1's Content- 2312 Encoding header field. Since this acts as a modifier on the media 2313 type, proxies and gateways from HTTP to MIME-compliant protocols 2314 ought to either change the value of the Content-Type header field or 2315 decode the representation before forwarding the message. (Some 2316 experimental applications of Content-Type for Internet mail have used 2317 a media-type parameter of ";conversions=" to perform 2318 a function equivalent to Content-Encoding. However, this parameter 2319 is not part of the MIME standards). 2321 B.5. Conversion of Content-Transfer-Encoding 2323 HTTP does not use the Content-Transfer-Encoding field of MIME. 2324 Proxies and gateways from MIME-compliant protocols to HTTP need to 2325 remove any Content-Transfer-Encoding prior to delivering the response 2326 message to an HTTP client. 2328 Proxies and gateways from HTTP to MIME-compliant protocols are 2329 responsible for ensuring that the message is in the correct format 2330 and encoding for safe transport on that protocol, where "safe 2331 transport" is defined by the limitations of the protocol being used. 2332 Such a proxy or gateway ought to transform and label the data with an 2333 appropriate Content-Transfer-Encoding if doing so will improve the 2334 likelihood of safe transport over the destination protocol. 2336 B.6. MHTML and Line Length Limitations 2338 HTTP implementations that share code with MHTML [RFC2557] 2339 implementations need to be aware of MIME line length limitations. 2340 Since HTTP does not have this limitation, HTTP does not fold long 2341 lines. MHTML messages being transported by HTTP follow all 2342 conventions of MHTML, including line length limitations and folding, 2343 canonicalization, etc., since HTTP transfers message-bodies as 2344 payload and, aside from the "multipart/byteranges" type 2345 (Section 6.3.5 of [Semantics]), does not interpret the content or any 2346 MIME header lines that might be contained therein. 2348 Appendix C. HTTP Version History 2350 HTTP has been in use since 1990. The first version, later referred 2351 to as HTTP/0.9, was a simple protocol for hypertext data transfer 2352 across the Internet, using only a single request method (GET) and no 2353 metadata. HTTP/1.0, as defined by [RFC1945], added a range of 2354 request methods and MIME-like messaging, allowing for metadata to be 2355 transferred and modifiers placed on the request/response semantics. 2356 However, HTTP/1.0 did not sufficiently take into consideration the 2357 effects of hierarchical proxies, caching, the need for persistent 2358 connections, or name-based virtual hosts. The proliferation of 2359 incompletely implemented applications calling themselves "HTTP/1.0" 2360 further necessitated a protocol version change in order for two 2361 communicating applications to determine each other's true 2362 capabilities. 2364 HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent 2365 requirements that enable reliable implementations, adding only those 2366 features that can either be safely ignored by an HTTP/1.0 recipient 2367 or only be sent when communicating with a party advertising 2368 conformance with HTTP/1.1. 2370 HTTP/1.1 has been designed to make supporting previous versions easy. 2371 A general-purpose HTTP/1.1 server ought to be able to understand any 2372 valid request in the format of HTTP/1.0, responding appropriately 2373 with an HTTP/1.1 message that only uses features understood (or 2374 safely ignored) by HTTP/1.0 clients. Likewise, an HTTP/1.1 client 2375 can be expected to understand any valid HTTP/1.0 response. 2377 Since HTTP/0.9 did not support header fields in a request, there is 2378 no mechanism for it to support name-based virtual hosts (selection of 2379 resource by inspection of the Host header field). Any server that 2380 implements name-based virtual hosts ought to disable support for 2381 HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in fact, 2382 badly constructed HTTP/1.x requests caused by a client failing to 2383 properly encode the request-target. 2385 C.1. Changes from HTTP/1.0 2387 This section summarizes major differences between versions HTTP/1.0 2388 and HTTP/1.1. 2390 C.1.1. Multihomed Web Servers 2392 The requirements that clients and servers support the Host header 2393 field (Section 5.6 of [Semantics]), report an error if it is missing 2394 from an HTTP/1.1 request, and accept absolute URIs (Section 3.2) are 2395 among the most important changes defined by HTTP/1.1. 2397 Older HTTP/1.0 clients assumed a one-to-one relationship of IP 2398 addresses and servers; there was no other established mechanism for 2399 distinguishing the intended server of a request than the IP address 2400 to which that request was directed. The Host header field was 2401 introduced during the development of HTTP/1.1 and, though it was 2402 quickly implemented by most HTTP/1.0 browsers, additional 2403 requirements were placed on all HTTP/1.1 requests in order to ensure 2404 complete adoption. At the time of this writing, most HTTP-based 2405 services are dependent upon the Host header field for targeting 2406 requests. 2408 C.1.2. Keep-Alive Connections 2410 In HTTP/1.0, each connection is established by the client prior to 2411 the request and closed by the server after sending the response. 2412 However, some implementations implement the explicitly negotiated 2413 ("Keep-Alive") version of persistent connections described in 2414 Section 19.7.1 of [RFC2068]. 2416 Some clients and servers might wish to be compatible with these 2417 previous approaches to persistent connections, by explicitly 2418 negotiating for them with a "Connection: keep-alive" request header 2419 field. However, some experimental implementations of HTTP/1.0 2420 persistent connections are faulty; for example, if an HTTP/1.0 proxy 2421 server doesn't understand Connection, it will erroneously forward 2422 that header field to the next inbound server, which would result in a 2423 hung connection. 2425 One attempted solution was the introduction of a Proxy-Connection 2426 header field, targeted specifically at proxies. In practice, this 2427 was also unworkable, because proxies are often deployed in multiple 2428 layers, bringing about the same problem discussed above. 2430 As a result, clients are encouraged not to send the Proxy-Connection 2431 header field in any requests. 2433 Clients are also encouraged to consider the use of Connection: keep- 2434 alive in requests carefully; while they can enable persistent 2435 connections with HTTP/1.0 servers, clients using them will need to 2436 monitor the connection for "hung" requests (which indicate that the 2437 client ought stop sending the header field), and this mechanism ought 2438 not be used by clients at all when a proxy is being used. 2440 C.1.3. Introduction of Transfer-Encoding 2442 HTTP/1.1 introduces the Transfer-Encoding header field (Section 6.1). 2443 Transfer codings need to be decoded prior to forwarding an HTTP 2444 message over a MIME-compliant protocol. 2446 C.2. Changes from RFC 7230 2448 Most of the sections introducing HTTP's design goals, history, 2449 architecture, conformance criteria, protocol versioning, URIs, 2450 message routing, and header fields have been moved to [Semantics]. 2451 This document has been reduced to just the messaging syntax and 2452 connection management requirements specific to HTTP/1.1. 2454 In the ABNF for chunked extensions, re-introduced (bad) whitespace 2455 around ";" and "=". Whitespace was removed in [RFC7230], but that 2456 change was found to break existing implementations (see [Err4667]). 2457 (Section 7.1.1) 2459 Trailer field semantics now transcend the specifics of chunked 2460 encoding. The decoding algorithm for chunked (Section 7.1.3) has 2461 been updated to encourage storage/forwarding of trailer fields 2462 separately from the header section, to only allow merging into the 2463 header section if the recipient knows the corresponding field 2464 definition permits and defines how to merge, and otherwise to discard 2465 the trailer fields instead of merging. The trailer part is now 2466 called the trailer section to be more consistent with the header 2467 section and more distinct from a body part. (Section 7.1.2) 2469 Disallowed transfer coding parameters called "q" in order to avoid 2470 conflicts with the use of ranks in the TE header field. 2471 (Section 7.3) 2473 Appendix D. Change Log 2475 This section is to be removed before publishing as an RFC. 2477 D.1. Between RFC7230 and draft 00 2479 The changes were purely editorial: 2481 o Change boilerplate and abstract to indicate the "draft" status, 2482 and update references to ancestor specifications. 2484 o Adjust historical notes. 2486 o Update links to sibling specifications. 2488 o Replace sections listing changes from RFC 2616 by new empty 2489 sections referring to RFC 723x. 2491 o Remove acknowledgements specific to RFC 723x. 2493 o Move "Acknowledgements" to the very end and make them unnumbered. 2495 D.2. Since draft-ietf-httpbis-messaging-00 2497 The changes in this draft are editorial, with respect to HTTP as a 2498 whole, to move all core HTTP semantics into [Semantics]: 2500 o Moved introduction, architecture, conformance, and ABNF extensions 2501 from RFC 7230 (Messaging) to semantics [Semantics]. 2503 o Moved discussion of MIME differences from RFC 7231 (Semantics) to 2504 Appendix B since they mostly cover transforming 1.1 messages. 2506 o Moved all extensibility tips, registration procedures, and 2507 registry tables from the IANA considerations to normative 2508 sections, reducing the IANA considerations to just instructions 2509 that will be removed prior to publication as an RFC. 2511 D.3. Since draft-ietf-httpbis-messaging-01 2513 o Cite RFC 8126 instead of RFC 5226 () 2516 o Resolved erratum 4779, no change needed here 2517 (, 2518 ) 2520 o In Section 7, fixed prose claiming transfer parameters allow bare 2521 names (, 2522 ) 2524 o Resolved erratum 4225, no change needed here 2525 (, 2526 ) 2528 o Replace "response code" with "response status code" 2529 (, 2530 ) 2532 o In Section 9.4, clarify statement about HTTP/1.0 keep-alive 2533 (, 2534 ) 2536 o In Section 7.1.1, re-introduce (bad) whitespace around ";" and "=" 2537 (, 2538 , ) 2541 o In Section 7.3, state that transfer codings should not use 2542 parameters named "q" (, ) 2545 o In Section 7, mark coding name "trailers" as reserved in the IANA 2546 registry () 2548 D.4. Since draft-ietf-httpbis-messaging-02 2550 o In Section 4, explain why the reason phrase should be ignored by 2551 clients (). 2553 o Add Section 9.3 to explain how request/response correlation is 2554 performed () 2556 D.5. Since draft-ietf-httpbis-messaging-03 2558 o In Section 9.3, caution against treating data on a connection as 2559 part of a not-yet-issued request () 2562 o In Section 7, remove the predefined codings from the ABNF and make 2563 it generic instead () 2566 o Use RFC 7405 ABNF notation for case-sensitive string constants 2567 () 2569 D.6. Since draft-ietf-httpbis-messaging-04 2571 o In Section 9.9, clarify that protocol-name is to be matched case- 2572 insensitively () 2574 o In Section 5.2, add leading optional whitespace to obs-fold ABNF 2575 (, 2576 ) 2578 o In Section 4, add clarifications about empty reason phrases 2579 () 2581 o Move discussion of retries from Section 9.4.1 into [Semantics] 2582 () 2584 D.7. Since draft-ietf-httpbis-messaging-05 2586 o In Section 7.1.2, the trailer part has been renamed the trailer 2587 section (for consistency with the header section) and trailers are 2588 no longer merged as header fields by default, but rather can be 2589 discarded, kept separate from header fields, or merged with header 2590 fields only if understood and defined as being mergeable 2591 () 2593 o In Section 2.1 and related Sections, move the trailing CRLF from 2594 the line grammars into the message format 2595 () 2597 o Moved Section 2.3 down () 2600 o In Section 9.9, use 'websocket' instead of 'HTTP/2.0' in examples 2601 () 2603 o Move version non-specific text from Section 6 into semantics as 2604 "payload body" () 2606 o In Section 9.8, add text from RFC 2818 2607 () 2609 D.8. Since draft-ietf-httpbis-messaging-06 2611 o In Section 12.5, update the APLN protocol id for HTTP/1.1 2612 () 2614 o In Section 5, align with updates to field terminology in semantics 2615 () 2617 o In Section 9.1, clarify that new connection options indeed need to 2618 be registered () 2620 o In Section 1.1, reference RFC 8174 as well 2621 () 2623 Index 2625 A 2626 absolute-form (of request-target) 11 2627 application/http Media Type 39 2628 asterisk-form (of request-target) 11 2629 authority-form (of request-target) 11 2631 C 2632 Connection header field 28, 33 2633 Content-Length header field 18 2634 Content-Transfer-Encoding header field 50 2635 chunked (Coding Format) 17, 19 2636 chunked (transfer coding) 22 2637 close 28, 33 2638 compress (transfer coding) 24 2640 D 2641 deflate (transfer coding) 24 2643 E 2644 effective request URI 12 2646 F 2647 Fields 2648 Connection 28 2649 MIME-Version 49 2650 TE 25 2651 Transfer-Encoding 17 2652 Upgrade 35 2654 G 2655 Grammar 2656 absolute-form 10-11 2657 ALPHA 5 2658 asterisk-form 10-11 2659 authority-form 10-11 2660 chunk 22 2661 chunk-data 22 2662 chunk-ext 22-23 2663 chunk-ext-name 23 2664 chunk-ext-val 23 2665 chunk-size 22 2666 chunked-body 22 2667 Connection 28 2668 connection-option 28 2669 CR 5 2670 CRLF 5 2671 CTL 5 2672 DIGIT 5 2673 DQUOTE 5 2674 field-line 14, 24 2675 field-name 14 2676 field-value 14 2677 HEXDIG 5 2678 HTAB 5 2679 HTTP-message 6 2680 HTTP-name 8 2681 HTTP-version 8 2682 last-chunk 22 2683 LF 5 2684 message-body 16 2685 method 9 2686 obs-fold 16 2687 OCTET 5 2688 origin-form 10 2689 rank 26 2690 reason-phrase 14 2691 request-line 9 2692 request-target 10 2693 SP 5 2694 start-line 6 2695 status-code 14 2696 status-line 13 2697 t-codings 26 2698 t-ranking 26 2699 TE 26 2700 trailer-section 22, 24 2701 transfer-coding 21 2702 Transfer-Encoding 17 2703 transfer-parameter 21 2704 Upgrade 35 2705 VCHAR 5 2706 gzip (transfer coding) 24 2708 H 2709 Header Fields 2710 Connection 28 2711 MIME-Version 49 2712 TE 25 2713 Transfer-Encoding 17 2714 Upgrade 35 2715 header line 6 2716 header section 6 2717 headers 6 2719 M 2720 MIME-Version header field 49 2721 Media Type 2722 application/http 39 2723 message/http 38 2724 message/http Media Type 38 2725 method 9 2727 O 2728 origin-form (of request-target) 10 2730 R 2731 request-target 10 2733 T 2734 TE header field 25 2735 Transfer-Encoding header field 17 2737 U 2738 Upgrade header field 35 2740 X 2741 x-compress (transfer coding) 24 2742 x-gzip (transfer coding) 24 2744 Acknowledgments 2746 See Appendix "Acknowledgments" of [Semantics]. 2748 Authors' Addresses 2750 Roy T. Fielding (editor) 2751 Adobe 2752 345 Park Ave 2753 San Jose, CA 95110 2754 United States of America 2756 EMail: fielding@gbiv.com 2757 URI: https://roy.gbiv.com/ 2759 Mark Nottingham (editor) 2760 Fastly 2762 EMail: mnot@mnot.net 2763 URI: https://www.mnot.net/ 2765 Julian F. Reschke (editor) 2766 greenbytes GmbH 2767 Hafenweg 16 2768 Muenster 48155 2769 Germany 2771 EMail: julian.reschke@greenbytes.de 2772 URI: https://greenbytes.de/tech/webdav/