idnits 2.17.1 draft-hixie-thewebsocketprotocol-23.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** The document seems to lack a License Notice according IETF Trust Provisions of 28 Dec 2009, Section 6.b.ii or Provisions of 12 Sep 2009 Section 6.b -- however, there's a paragraph with a matching beginning. Boilerplate error? (You're using the IETF Trust Provisions' Section 6.b License Notice from 12 Feb 2009 rather than one of the newer Notices. See https://trustee.ietf.org/license-info/.) 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 abstract seems to contain references ([HTML5]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document doesn't use any RFC 2119 keywords, yet has text resembling RFC 2119 boilerplate text. -- The document date (July 27, 2009) is 5358 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) -- Possible downref: Non-RFC (?) normative reference: ref. 'HTML5' ** Obsolete normative reference: RFC 2109 (Obsoleted by RFC 2965) ** Obsolete normative reference: RFC 2246 (Obsoleted by RFC 4346) ** Obsolete normative reference: RFC 2616 (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) ** Obsolete normative reference: RFC 2965 (Obsoleted by RFC 6265) Summary: 6 errors (**), 0 flaws (~~), 2 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group I. Hickson 3 Internet-Draft Google, Inc. 4 Intended status: Standards Track July 27, 2009 5 Expires: January 28, 2010 7 The Web Socket protocol 8 draft-hixie-thewebsocketprotocol-23 10 Status of this Memo 12 This Internet-Draft is submitted to IETF in full conformance with the 13 provisions of BCP 78 and BCP 79. 15 Internet-Drafts are working documents of the Internet Engineering 16 Task Force (IETF), its areas, and its working groups. Note that 17 other groups may also distribute working documents as Internet- 18 Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet-Drafts as reference 23 material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt. 28 The list of Internet-Draft Shadow Directories can be accessed at 29 http://www.ietf.org/shadow.html. 31 This Internet-Draft will expire on January 28, 2010. 33 Copyright Notice 35 Copyright (c) 2009 IETF Trust and the persons identified as the 36 document authors. All rights reserved. 38 This document is subject to BCP 78 and the IETF Trust's Legal 39 Provisions Relating to IETF Documents in effect on the date of 40 publication of this document (http://trustee.ietf.org/license-info). 41 Please review these documents carefully, as they describe your rights 42 and restrictions with respect to this document. 44 Abstract 46 This protocol enables two-way communication between a user agent 47 running untrusted code running in a controlled environment to a 48 remote host that understands the protocol. It is intended to fail to 49 communicate with servers of pre-existing protocols like SMTP or HTTP, 50 while allowing HTTP servers to opt-in to supporting this protocol if 51 desired. It is designed to be easy to implement on the server side. 53 Author's note 55 This document is automatically generated from the same source 56 document as the HTML5 specification. [HTML5] 58 Table of Contents 60 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 61 1.1. Security model . . . . . . . . . . . . . . . . . . . . . . 4 62 1.2. Relationship to TCP/IP and HTTP . . . . . . . . . . . . . 4 63 1.3. Establishing a connection . . . . . . . . . . . . . . . . 4 64 2. Conformance requirements . . . . . . . . . . . . . . . . . . . 6 65 3. Client-side requirements . . . . . . . . . . . . . . . . . . . 7 66 3.1. Handshake . . . . . . . . . . . . . . . . . . . . . . . . 7 67 3.2. Data framing . . . . . . . . . . . . . . . . . . . . . . . 14 68 3.3. Handling errors in UTF-8 . . . . . . . . . . . . . . . . . 15 69 4. Server-side requirements . . . . . . . . . . . . . . . . . . . 16 70 4.1. Minimal handshake . . . . . . . . . . . . . . . . . . . . 16 71 4.2. Handshake details . . . . . . . . . . . . . . . . . . . . 17 72 4.3. Data framing . . . . . . . . . . . . . . . . . . . . . . . 18 73 5. Closing the connection . . . . . . . . . . . . . . . . . . . . 19 74 6. Security considerations . . . . . . . . . . . . . . . . . . . 20 75 7. IANA considerations . . . . . . . . . . . . . . . . . . . . . 21 76 8. Normative References . . . . . . . . . . . . . . . . . . . . . 22 77 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 23 79 1. Introduction 81 The Web Socket protocol is designed on the principle that there 82 should be minimal framing (the only framing that exists is to make 83 the protocol frame-based instead of stream-based, and to support a 84 distinction between Unicode text and binary frames). It is expected 85 that metadata would be layered on top of Web Socket by the 86 application layer, in the same way that metadata is layered on top of 87 TCP/IP by the application layer (HTTP). 89 Conceptually, Web Socket is really just a layer on top of TCP/IP that 90 adds a Web "origin"-based security model for browsers; adds an 91 addressing and protocol naming mechanism to support multiple services 92 on one port and multiple host names on one IP address; and layers a 93 framing mechanism on top of TCP to get back to the IP packet 94 mechanism that TCP is built on, but without length limits. Other 95 than that, it adds nothing. Basically it is intended to be as close 96 as possible to just exposing raw TCP/IP to script as possible given 97 the constraints of the Web. It's also designed in such a way that its 98 servers can share a port with HTTP servers, by having its handshake 99 be a valid HTTP Upgrade handshake also. 101 1.1. Security model 103 The Web Socket protocol uses the origin model used by Web browsers to 104 restrict which Web pages can contact a Web Socket server when the Web 105 Socket protocol is used from a Web page. Naturally, when the Web 106 Socket protocol is used directly (not from a Web page), the origin 107 model is not useful, as the client can provide any arbitrary origin 108 string. 110 1.2. Relationship to TCP/IP and HTTP 112 The Web Socket protocol is an independent TCP-based protocol. It's 113 only relationship to HTTP is that its handshake is interpreted by 114 HTTP servers as an Upgrade request. 116 The Web Socket protocol by default uses port 81 for regular Web 117 Socket connections and port 815 for Web Socket connections tunneled 118 over TLS. 120 1.3. Establishing a connection 122 There are several options for establishing a Web Socket connection. 124 The simplest method is to use port 81 to get a direct connection to a 125 Web Socket server. However, this port may be blocked by firewalls. 127 The second simplest method is to use TLS encryption and port 815 to 128 connect directly to a Web Socket server. This is the preferred 129 solution, as it is secure and correct. However, TLS encryption can 130 be computationally expensive, and port 815 might also be blocked by 131 firewalls. 133 To avoid firewalls, ports 80 and 443 might be used instead. These 134 are the HTTP and HTTPS ports. Port 80 traffic, however, will often 135 be intercepted by HTTP proxies, which can lead to the connection 136 failing to be established. 138 Port 443, using encryption, is therefore the most reliable solution. 139 It is unlikely to be blocked by a firewall or intercepted by a proxy. 140 However, again, TLS encryption can be computationally expensive. 142 When a connection is to be made to a port that is shared by an HTTP 143 server (a situation that is quite likely to occur with traffic to 144 ports 80 and 443), the connection will appear to the HTTP server to 145 be a regular GET request with an Upgrade offer. In relatively simple 146 setups with just one IP address and a single server for all traffic 147 to a single hostname, this might allow a practical way for systems 148 based on the Web Socket protocol to be deployed. In more elaborate 149 setups (e.g. with load balancers and multiple servers), a dedicated 150 set of hosts for Web Socket connections separate from the HTTP 151 servers is probably easier to manage. 153 2. Conformance requirements 155 All diagrams, examples, and notes in this specification are non- 156 normative, as are all sections explicitly marked non-normative. 157 Everything else in this specification is normative. 159 The key words "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT", 160 "RECOMMENDED", "MAY", and "OPTIONAL" in the normative parts of this 161 document are to be interpreted as described in RFC2119. For 162 readability, these words do not appear in all uppercase letters in 163 this specification. [RFC2119] 165 Requirements phrased in the imperative as part of algorithms (such as 166 "strip any leading space characters" or "return false and abort these 167 steps") are to be interpreted with the meaning of the key word 168 ("must", "should", "may", etc) used in introducing the algorithm. 170 Conformance requirements phrased as algorithms or specific steps may 171 be implemented in any manner, so long as the end result is 172 equivalent. (In particular, the algorithms defined in this 173 specification are intended to be easy to follow, and not intended to 174 be performant.) 176 Implementations may impose implementation-specific limits on 177 otherwise unconstrained inputs, e.g. to prevent denial of service 178 attacks, to guard against running out of memory, or to work around 179 platform-specific limitations. 181 The conformance classes defined by this specification are user agents 182 and servers. 184 3. Client-side requirements 186 _This section only applies to user agents, not to servers._ 188 NOTE: This specification doesn't currently define a limit to the 189 number of simultaneous connections that a client can establish to a 190 server. 192 3.1. Handshake 194 When the user agent is to *establish a Web Socket connection* to a 195 host /host/, optionally on port /port/, from an origin /origin/, with 196 a flag /secure/, with a particular /resource name/, and optionally 197 with a particular /protocol/, it must run the following steps. 199 NOTE: The /host/ and /origin/ strings will be all-lowercase when this 200 algorithm is invoked. 202 1. If there is no explicit /port/, then: if /secure/ is false, let 203 /port/ be 81, otherwise let /port/ be 815. 205 2. If the user agent already has a Web Socket connection to the 206 remote host identified by /host/ (even if known by another 207 name), wait until that connection has been established or for 208 that connection to have failed. 210 NOTE: This makes it harder for a script to perform a denial of 211 service attack by just opening a large number of Web Socket 212 connections to a remote host. 214 NOTE: There is no limit to the number of established Web Socket 215 connections a user agent can have with a single remote host. 216 Servers can refuse to connect users with an excessive number of 217 connections, or disconnect resource-hogging users when suffering 218 high load. 220 3. If the user agent is configured to use a proxy when using the 221 Web Socket protocol to connect to host /host/ and/or port 222 /port/, then connect to that proxy and ask it to open a TCP/IP 223 connection to the host given by /host/ and the port given by 224 /port/. 226 EXAMPLE: For example, if the user agent uses an HTTP proxy 227 for all traffic, then if it was to try to connect to port 80 228 on server example.com, it might send the following lines to 229 the proxy server: 231 CONNECT example.com:80 HTTP/1.1 232 Host: example.com 234 If there was a password, the connection might look like: 236 CONNECT example.com:80 HTTP/1.1 237 Host: example.com 238 Proxy-authorization: Basic ZWRuYW1vZGU6bm9jYXBlcyE= 240 Otherwise, if the user agent is not configured to use a proxy, 241 then open a TCP/IP connection to the host given by /host/ and 242 the port given by /port/. 244 NOTE: Implementations that do not expose explicit UI for 245 selecting a proxy for Web Socket connections separate from other 246 proxies are encouraged to use a SOCKS proxy for Web Socket 247 connections, if available, or failing that, to prefer an HTTPS 248 proxy over an HTTP proxy. 250 4. If the connection could not be opened, then fail the Web Socket 251 connection and abort these steps. 253 5. If /secure/ is true, perform a TLS handshake over the 254 connection. If this fails (e.g. the server's certificate could 255 not be verified), then fail the Web Socket connection and abort 256 these steps. Otherwise, all further communication on this 257 channel must run through the encrypted tunnel. [RFC2246] 259 6. Send the following bytes to the remote side (the server): 261 47 45 54 20 263 Send the /resource name/ value, encoded as US-ASCII. 265 Send the following bytes: 267 20 48 54 54 50 2f 31 2e 31 0d 0a 55 70 67 72 61 268 64 65 3a 20 57 65 62 53 6f 63 6b 65 74 0d 0a 43 269 6f 6e 6e 65 63 74 69 6f 6e 3a 20 55 70 67 72 61 270 64 65 0d 0a 272 NOTE: The string "GET ", the path, " HTTP/1.1", CRLF, the string 273 "Upgrade: WebSocket", CRLF, and the string "Connection: 274 Upgrade", CRLF. 276 7. Send the following bytes: 278 48 6f 73 74 3a 20 280 Send the /host/ value, encoded as US-ASCII. 282 Send the following bytes: 284 0d 0a 286 NOTE: The string "Host: ", the host, and CRLF. 288 8. Send the following bytes: 290 4f 72 69 67 69 6e 3a 20 292 Send the /origin/ value, encoded as US-ASCII. 294 NOTE: The /origin/ value is a string that was passed to this 295 algorithm. 297 Send the following bytes: 299 0d 0a 301 NOTE: The string "Origin: ", the origin, and CRLF. 303 9. If there is no /protocol/, then skip this step. 305 Otherwise, send the following bytes: 307 57 65 62 53 6f 63 6b 65 74 2d 50 72 6f 74 6f 63 308 6f 6c 3a 20 310 Send the /protocol/ value, encoded as US-ASCII. 312 Send the following bytes: 314 0d 0a 316 NOTE: The string "WebSocket-Protocol: ", the protocol, and CRLF. 318 10. If the client has any authentication information or cookies that 319 would be relevant to a resource accessed over HTTP, if /secure/ 320 is false, or HTTPS, if it is true, on host /host/, port /port/, 321 with /resource name/ as the path (and possibly query 322 parameters), then HTTP headers that would be appropriate for 323 that information should be sent at this point. [RFC2616] 324 [RFC2109] [RFC2965] 326 Each header must be on a line of its own (each ending with a CR 327 LF sequence). For the purposes of this step, each header must 328 not be split into multiple lines (despite HTTP otherwise 329 allowing this with continuation lines). 331 EXAMPLE: For example, if the server had a username and 332 password that applied to |http://example.com/socket|, and the 333 Web Socket was being opened to |ws://example.com:80/socket|, 334 it could send them: 336 Authorization: Basic d2FsbGU6ZXZl 338 However, it would not send them if the Web Socket was being 339 opened to |ws://example.com/socket|, as that uses a different 340 port (81, not 80). 342 11. Send the following bytes: 344 0d 0a 346 NOTE: Just a CRLF (a blank line). 348 12. Read the first 85 bytes from the server. If the connection 349 closes before 85 bytes are received, or if the first 85 bytes 350 aren't exactly equal to the following bytes, then fail the Web 351 Socket connection and abort these steps. 353 48 54 54 50 2f 31 2e 31 20 31 30 31 20 57 65 62 354 20 53 6f 63 6b 65 74 20 50 72 6f 74 6f 63 6f 6c 355 20 48 61 6e 64 73 68 61 6b 65 0d 0a 55 70 67 72 356 61 64 65 3a 20 57 65 62 53 6f 63 6b 65 74 0d 0a 357 43 6f 6e 6e 65 63 74 69 6f 6e 3a 20 55 70 67 72 358 61 64 65 0d 0a 360 NOTE: The string "HTTP/1.1 101 Web Socket Protocol Handshake", 361 CRLF, the string "Upgrade: WebSocket", CRLF, the string 362 "Connection: Upgrade", CRLF. 364 User agents may apply a timeout to this step, failing the Web 365 Socket connection if the server does not respond with the above 366 bytes within a suitable time period. 368 13. Let /headers/ be a list of name-value pairs, initially empty. 370 14. _Header_: Let /name/ and /value/ be empty byte arrays. 372 15. Read a byte from the server. 374 If the connection closes before this byte is received, then fail 375 the Web Socket connection and abort these steps. 377 Otherwise, handle the byte as described in the appropriate entry 378 below: 380 -> If the byte is 0x0d (ASCII CR) 381 If the /name/ byte array is empty, then jump to the headers 382 processing step. Otherwise, fail the Web Socket connection 383 and abort these steps. 385 -> If the byte is 0x0a (ASCII LF) 386 Fail the Web Socket connection and abort these steps. 388 -> If the byte is 0x3a (ASCII ":") 389 Move on to the next step. 391 -> If the byte is in the range 0x41 .. 0x5a (ASCII "A" .. "Z") 392 Append a byte whose value is the byte's value plus 0x20 to 393 the /name/ byte array and redo this step for the next byte. 395 -> Otherwise 396 Append the byte to the /name/ byte array and redo this step 397 for the next byte. 399 NOTE: This reads a header name, terminated by a colon, 400 converting upper-case ASCII letters to lowercase, and aborting 401 if a stray CR or LF is found. 403 16. Read a byte from the server. 405 If the connection closes before this byte is received, then fail 406 the Web Socket connection and abort these steps. 408 Otherwise, handle the byte as described in the appropriate entry 409 below: 411 -> If the byte is 0x20 (ASCII space) 412 Ignore the byte and move on to the next step. 414 -> Otherwise 415 Treat the byte as described by the list in the next step, 416 then move on to that next step for real. 418 NOTE: This skips past a space character after the colon, if 419 necessary. 421 17. Read a byte from the server. 423 If the connection closes before this byte is received, then fail 424 the Web Socket connection and abort these steps. 426 Otherwise, handle the byte as described in the appropriate entry 427 below: 429 -> If the byte is 0x0d (ASCII CR) 430 Move on to the next step. 432 -> If the byte is 0x0a (ASCII LF) 433 Fail the Web Socket connection and abort these steps. 435 -> Otherwise 436 Append the byte to the /value/ byte array and redo this step 437 for the next byte. 439 NOTE: This reads a header value, terminated by a CRLF. 441 18. Read a byte from the server. 443 If the connection closes before this byte is received, or if the 444 byte is not a 0x0a byte (ASCII LF), then fail the Web Socket 445 connection and abort these steps. 447 NOTE: This skips past the LF byte of the CRLF after the header. 449 19. Append an entry to the /headers/ list that has the name given by 450 the string obtained by interpreting the /name/ byte array as a 451 UTF-8 byte stream and the value given by the string obtained by 452 interpreting the /value/ byte array as a UTF-8 byte stream. 454 20. Return to the "Header" step above. 456 21. _Headers processing_: If there is not exactly one entry in the 457 /headers/ list whose name is "websocket-origin", or if there is 458 not exactly one entry in the /headers/ list whose name is 459 "websocket-location", or if the /protocol/ was specified but 460 there is not exactly one entry in the /headers/ list whose name 461 is "websocket-protocol", or if there are any entries in the 462 /headers/ list whose names are the empty string, then fail the 463 Web Socket connection and abort these steps. 465 22. Read a byte from the server. 467 If the connection closes before this byte is received, or if the 468 byte is not a 0x0a byte (ASCII LF), then fail the Web Socket 469 connection and abort these steps. 471 NOTE: This skips past the LF byte of the CRLF after the blank 472 line after the headers. 474 23. Handle each entry in the /headers/ list as follows: 476 -> If the entry's name is "websocket-origin" 477 If the value is not exactly equal to /origin/, converted to 478 ASCII lowercase, then fail the Web Socket connection and 479 abort these steps. 481 -> If the entry's name is "websocket-location" 482 If the value is not exactly equal to a string consisting of 483 the following components in the same order, then fail the Web 484 Socket connection and abort these steps: 486 1. The string "ws" if /secure/ is false and "wss" if 487 /secure/ is true 489 2. The three characters "://". 491 3. The value of /host/. 493 4. If /secure/ is false and /port/ is not 81, or if /secure/ 494 is true and /port/ is not 815: a ":" character followed 495 by the value of /port/. 497 5. The value of /resource name/. 499 -> If the entry's name is "websocket-protocol" 500 If there was a /protocol/ specified, and the value is not 501 exactly equal to /protocol/, then fail the Web Socket 502 connection and abort these steps. (If no /protocol/ was 503 specified, the header is ignored.) 505 -> If the entry's name is "set-cookie" or "set-cookie2" or 506 another cookie-related header name 507 Handle the cookie as defined by the appropriate spec, with 508 the resource being the one with the host /host/, the port 509 /port/, the path (and possibly query parameters) /resource 510 name/, and the scheme |http| if /secure/ is false and |https| 511 if /secure/ is true. [RFC2109] [RFC2965] 513 -> Any other name 514 Ignore it. 516 24. The *Web Socket connection is established*. Now the user agent 517 must send and receive to and from the connection as described in 518 the next section. 520 To *fail the Web Socket connection*, the user agent must close the 521 Web Socket connection, and may report the problem to the user (which 522 would be especially useful for developers). However, user agents 523 must not convey the failure information to the script that attempted 524 the connection in a way distinguishable from the Web Socket being 525 closed normally. 527 3.2. Data framing 529 Once a Web Socket connection is established, the user agent must run 530 through the following state machine for the bytes sent by the server. 532 1. Try to read a byte from the server. Let /frame type/ be that 533 byte. 535 If no byte could be read because the Web Socket connection is 536 closed, then abort. 538 2. Handle the /frame type/ byte as follows: 540 If the high-order bit of the /frame type/ byte is set (i.e. if 541 /frame type/ _and_ed with 0x80 returns 0x80) 542 Run these steps. If at any point during these steps a read is 543 attempted but fails because the Web Socket connection is 544 closed, then abort. 546 1. Let /length/ be zero. 548 2. _Length_: Read a byte, let /b/ be that byte. 550 3. Let /b_v/ be integer corresponding to the low 7 bits of 551 /b/ (the value you would get by _and_ing /b/ with 0x7f). 553 4. Multiply /length/ by 128, add /b_v/ to that result, and 554 store the final result in /length/. 556 5. If the high-order bit of /b/ is set (i.e. if /b/ _and_ed 557 with 0x80 returns 0x80), then return to the step above 558 labeled _length_. 560 6. Read /length/ bytes. 562 7. Discard the read bytes. 564 If the high-order bit of the /frame type/ byte is _not_ set (i.e. 565 if /frame type/ _and_ed with 0x80 returns 0x00) 566 Run these steps. If at any point during these steps a read is 567 attempted but fails because the Web Socket connection is 568 closed, then abort. 570 1. Let /raw data/ be an empty byte array. 572 2. _Data_: Read a byte, let /b/ be that byte. 574 3. If /b/ is not 0xff, then append /b/ to /raw data/ and 575 return to the previous step (labeled _data_). 577 4. Interpret /raw data/ as a UTF-8 string, and store that 578 string in /data/. 580 5. If /frame type/ is 0x00, then *a message has been 581 received* with text /data/. Otherwise, discard the data. 583 3. Return to the first step to read the next byte. 585 If the user agent is faced with content that is too large to be 586 handled appropriately, then it must fail the Web Socket connection. 588 Once a Web Socket connection is established, the user agent must use 589 the following steps to *send /data/ using the Web Socket*: 591 1. Send a 0x00 byte to the server. 593 2. Encode /data/ using UTF-8 and send the resulting byte stream to 594 the server. 596 3. Send a 0xff byte to the server. 598 3.3. Handling errors in UTF-8 600 When a client is to interpret a byte stream as UTF-8 but finds that 601 the byte stream is not in fact a valid UTF-8 stream, then any bytes 602 or sequences of bytes that are not valid UTF-8 sequences must be 603 interpreted as a U+FFFD REPLACEMENT CHARACTER. 605 4. Server-side requirements 607 _This section only applies to servers._ 609 4.1. Minimal handshake 611 NOTE: This section describes the minimal requirements for a server- 612 side implementation of Web Sockets. 614 Listen on a port for TCP/IP. Upon receiving a connection request, 615 open a connection and send the following bytes back to the client: 617 48 54 54 50 2f 31 2e 31 20 31 30 31 20 57 65 62 618 20 53 6f 63 6b 65 74 20 50 72 6f 74 6f 63 6f 6c 619 20 48 61 6e 64 73 68 61 6b 65 0d 0a 55 70 67 72 620 61 64 65 3a 20 57 65 62 53 6f 63 6b 65 74 0d 0a 621 43 6f 6e 6e 65 63 74 69 6f 6e 3a 20 55 70 67 72 622 61 64 65 0d 0a 624 Send the string "WebSocket-Origin" followed by a U+003A COLON (":") 625 followed by the ASCII serialization of the origin from which the 626 server is willing to accept connections, followed by a CRLF pair 627 (0x0d 0x0a). 629 For instance: 631 WebSocket-Origin: http://example.com 633 Send the string "WebSocket-Location" followed by a U+003A COLON (":") 634 followed by the URL of the Web Socket script, followed by a CRLF pair 635 (0x0d 0x0a). 637 For instance: 639 WebSocket-Location: ws://example.com:80/demo 641 Send another CRLF pair (0x0d 0x0a). 643 Read data from the client until four bytes 0x0d 0x0a 0x0d 0x0a are 644 read. This data must either be discarded or handled as described in 645 the following section describing the handshake details. 647 If the connection isn't dropped at this point, go to the data framing 648 section. 650 4.2. Handshake details 652 The previous section ignores the data that is transmitted by the 653 client during the handshake. 655 The data sent by the client consists of a number of fields separated 656 by CR LF pairs (bytes 0x0d 0x0a). 658 The first field consists of three tokens separated by space 659 characters (byte 0x20). The middle token is the path being opened. 660 If the server supports multiple paths, then the server should echo 661 the value of this field in the initial handshake, as part of the URL 662 given on the |WebSocket-Location| line (after the appropriate scheme 663 and host). 665 If the first field does not have three tokens, the server should 666 abort the connection as it probably represents an errorneous client. 668 The remaining fields consist of name-value pairs, with the name part 669 separated from the value part by a colon and a space (bytes 0x3a 670 0x20). Of these, several are interesting: 672 Host (bytes 48 6f 73 74) 673 The value gives the hostname that the client intended to use when 674 opening the Web Socket. It would be of interest in particular to 675 virtual hosting environments, where one server might serve 676 multiple hosts, and might therefore want to return different data. 678 The right host has to be output as part of the URL given on the 679 |WebSocket-Location| line of the handshake described above, to 680 verify that the server knows that it is really representing that 681 host. 683 Origin (bytes 4f 72 69 67 69 6e) 684 The value gives the scheme, hostname, and port (if it's not the 685 default port for the given scheme) of the page that asked the 686 client to open the Web Socket. It would be interesting if the 687 server's operator had deals with operators of other sites, since 688 the server could then decide how to respond (or indeed, _whether_ 689 to respond) based on which site was requesting a connection. 691 If the server supports connections from more than one origin, then 692 the server should echo the value of this field in the initial 693 handshake, on the |WebSocket-Origin| line. 695 Other fields 696 Other fields can be used, such as "Cookie" or "Authorization", for 697 authentication purposes. 699 Any fields that lack the colon-space separator should be discarded 700 and may cause the server to disconnect. 702 4.3. Data framing 704 NOTE: This section only describes how to handle content that this 705 specification allows user agents to send (text). It doesn't handle 706 any arbitrary content in the same way that the requirements on user 707 agents defined earlier handle any content including possible future 708 extensions to the protocols. 710 The server must run through the following steps to process the bytes 711 sent by the client: 713 1. Read a byte from the client. Assuming everything is going 714 according to plan, it will be a 0x00 byte. If the byte is not a 715 0x00 byte, then the server may disconnect. 717 2. Let /raw data/ be an empty byte array. 719 3. _Data_: Read a byte, let /b/ be that byte. 721 4. If /b/ is not 0xff, then append /b/ to /raw data/ and return to 722 the previous step (labeled _data_). 724 5. Interpret /raw data/ as a UTF-8 string, and apply whatever 725 server-specific processing is to occur for the resulting string. 727 6. Return to the first step to read the next byte. 729 The server must run through the following steps to send strings to 730 the client: 732 1. Send a 0x00 byte to the client to indicate the start of a string. 734 2. Encode /data/ using UTF-8 and send the resulting byte stream to 735 the client. 737 3. Send a 0xff byte to the client to indicate the end of the 738 message. 740 5. Closing the connection 742 To *close the Web Socket connection*, either the user agent or the 743 server closes the TCP/IP connection. There is no closing handshake. 744 Whether the user agent or the server closes the connection, it is 745 said that the *Web Socket connection is closed*. 747 Servers may close the Web Socket connection whenever desired. 749 User agents should not close the Web Socket connection arbitrarily. 751 6. Security considerations 753 ** ISSUE ** ... 755 7. IANA considerations 757 ** ISSUE ** ...(two URI schemes, two ports, HTTP Upgrade keyword) 759 8. Normative References 761 [HTML5] Hickson, I., "HTML5", July 2009. 763 [RFC2109] Kristol, D. and L. Montulli, "HTTP State Management 764 Mechanism", RFC 2109, February 1997. 766 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 767 Requirement Levels", BCP 14, RFC 2119, March 1997. 769 [RFC2246] Dierks, T. and C. Allen, "The TLS Protocol Version 1.0", 770 RFC 2246, January 1999. 772 [RFC2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., 773 Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext 774 Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999. 776 [RFC2965] Kristol, D. and L. Montulli, "HTTP State Management 777 Mechanism", RFC 2965, October 2000. 779 Author's Address 781 Ian Hickson 782 Google, Inc. 784 Email: ian@hixie.ch 785 URI: http://ln.hixie.ch/