idnits 2.17.1 

draft-ietf-hybi-thewebsocketprotocol-09.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 :
  ----------------------------------------------------------------------------

  ** There are 10 instances of too long lines in the document, the longest
     one being 9 characters in excess of 72.


  Miscellaneous warnings:
  ----------------------------------------------------------------------------

  == The copyright year in the IETF Trust and authors Copyright Line does not
     match the current year

  == The document seems to lack the recommended RFC 2119 boilerplate, even if
     it appears to use RFC 2119 keywords -- however, there's a paragraph with
     a matching beginning. Boilerplate error?

     (The document does seem to have the reference to RFC 2119 which the
     ID-Checklist requires).
  -- The document date (June 13, 2011) is 4694 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)

  ** Downref: Normative reference to an Informational RFC: RFC 1951

  ** Obsolete normative reference: RFC 2616 (Obsoleted by RFC 7230, RFC 7231,
     RFC 7232, RFC 7233, RFC 7234, RFC 7235)

  ** Obsolete normative reference: RFC 2818 (Obsoleted by RFC 9110)

  ** Obsolete normative reference: RFC 3490 (Obsoleted by RFC 5890, RFC 5891)

  ** Obsolete normative reference: RFC 3548 (Obsoleted by RFC 4648)

  ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446)

  ** Obsolete normative reference: RFC 5226 (Obsoleted by RFC 8126)

  == Outdated reference: A later version (-23) exists of
     draft-ietf-httpstate-cookie-20

  == Outdated reference: A later version (-06) exists of
     draft-ietf-websec-origin-00


     Summary: 8 errors (**), 0 flaws (~~), 4 warnings (==), 1 comment (--).

     Run idnits with the --verbose option for more detailed information about
     the items above.

--------------------------------------------------------------------------------


2	HyBi Working Group                                              I. Fette
3	Internet-Draft                                              Google, Inc.
4	Intended status: Standards Track                           June 13, 2011
5	Expires: December 15, 2011

7	                         The WebSocket protocol
8	                draft-ietf-hybi-thewebsocketprotocol-09

10	Abstract

12	   The WebSocket protocol enables two-way communication between a client
13	   running untrusted code running in a controlled environment to a
14	   remote host that has opted-in to communications from that code.  The
15	   security model used for this is the Origin-based security model
16	   commonly used by Web browsers.  The protocol consists of an opening
17	   handshake followed by basic message framing, layered over TCP.  (In
18	   theory, any transport protocol could be used so long as it provides
19	   for reliable transport, is byte clean, and supports relatively large
20	   message sizes.  However, for this document, we consider only TCP.)
21	   The goal of this technology is to provide a mechanism for browser-
22	   based applications that need two-way communication with servers that
23	   does not rely on opening multiple HTTP connections (e.g. using
24	   XMLHttpRequest or <iframe>s and long polling).

26	   Please send feedback to the hybi@ietf.org mailing list.

28	Status of this Memo

30	   This Internet-Draft is submitted in full conformance with the
31	   provisions of BCP 78 and BCP 79.

33	   Internet-Drafts are working documents of the Internet Engineering
34	   Task Force (IETF).  Note that other groups may also distribute
35	   working documents as Internet-Drafts.  The list of current Internet-
36	   Drafts is at http://datatracker.ietf.org/drafts/current/.

38	   Internet-Drafts are draft documents valid for a maximum of six months
39	   and may be updated, replaced, or obsoleted by other documents at any
40	   time.  It is inappropriate to use Internet-Drafts as reference
41	   material or to cite them other than as "work in progress."

43	   This Internet-Draft will expire on December 15, 2011.

45	Copyright Notice

47	   Copyright (c) 2011 IETF Trust and the persons identified as the
48	   document authors.  All rights reserved.

50	   This document is subject to BCP 78 and the IETF Trust's Legal
51	   Provisions Relating to IETF Documents
52	   (http://trustee.ietf.org/license-info) in effect on the date of
53	   publication of this document.  Please review these documents
54	   carefully, as they describe your rights and restrictions with respect
55	   to this document.  Code Components extracted from this document must
56	   include Simplified BSD License text as described in Section 4.e of
57	   the Trust Legal Provisions and are provided without warranty as
58	   described in the Simplified BSD License.

60	Table of Contents

62	   1.  Introduction . . . . . . . . . . . . . . . . . . . . . . . . .  4
63	     1.1.   Background  . . . . . . . . . . . . . . . . . . . . . . .  4
64	     1.2.   Protocol Overview . . . . . . . . . . . . . . . . . . . .  4
65	     1.3.   Opening Handshake . . . . . . . . . . . . . . . . . . . .  6
66	     1.4.   Closing Handshake . . . . . . . . . . . . . . . . . . . .  9
67	     1.5.   Design Philosophy . . . . . . . . . . . . . . . . . . . .  9
68	     1.6.   Security Model  . . . . . . . . . . . . . . . . . . . . . 10
69	     1.7.   Relationship to TCP and HTTP  . . . . . . . . . . . . . . 11
70	     1.8.   Establishing a Connection . . . . . . . . . . . . . . . . 11
71	     1.9.   Subprotocols Using the WebSocket protocol . . . . . . . . 11
72	   2.  Conformance Requirements . . . . . . . . . . . . . . . . . . . 13
73	     2.1.   Terminology . . . . . . . . . . . . . . . . . . . . . . . 13
74	   3.  WebSocket URIs . . . . . . . . . . . . . . . . . . . . . . . . 15
75	   4.  Data Framing . . . . . . . . . . . . . . . . . . . . . . . . . 16
76	     4.1.   Overview  . . . . . . . . . . . . . . . . . . . . . . . . 16
77	     4.2.   Base Framing Protocol . . . . . . . . . . . . . . . . . . 16
78	     4.3.   Client-to-Server Masking  . . . . . . . . . . . . . . . . 20
79	     4.4.   Fragmentation . . . . . . . . . . . . . . . . . . . . . . 21
80	     4.5.   Control Frames  . . . . . . . . . . . . . . . . . . . . . 22
81	       4.5.1.  Close  . . . . . . . . . . . . . . . . . . . . . . . . 23
82	       4.5.2.  Ping . . . . . . . . . . . . . . . . . . . . . . . . . 24
83	       4.5.3.  Pong . . . . . . . . . . . . . . . . . . . . . . . . . 24
84	     4.6.   Data Frames . . . . . . . . . . . . . . . . . . . . . . . 24
85	     4.7.   Examples  . . . . . . . . . . . . . . . . . . . . . . . . 25
86	     4.8.   Extensibility . . . . . . . . . . . . . . . . . . . . . . 25
87	   5.  Opening Handshake  . . . . . . . . . . . . . . . . . . . . . . 27
88	     5.1.   Client Requirements . . . . . . . . . . . . . . . . . . . 27
89	     5.2.   Server-side Requirements  . . . . . . . . . . . . . . . . 32
90	       5.2.1.  Reading the Client's Opening Handshake . . . . . . . . 32
91	       5.2.2.  Sending the Server's Opening Handshake . . . . . . . . 33
92	   6.  Sending and Receiving Data . . . . . . . . . . . . . . . . . . 37
93	     6.1.   Sending Data  . . . . . . . . . . . . . . . . . . . . . . 37
94	     6.2.   Receiving Data  . . . . . . . . . . . . . . . . . . . . . 37
95	   7.  Closing the connection . . . . . . . . . . . . . . . . . . . . 39
96	     7.1.   Definitions . . . . . . . . . . . . . . . . . . . . . . . 39
97	       7.1.1.  Close the WebSocket Connection . . . . . . . . . . . . 39
98	       7.1.2.  Start the WebSocket Closing Handshake  . . . . . . . . 39
99	       7.1.3.  The WebSocket Closing Handshake is Started . . . . . . 39
100	       7.1.4.  The WebSocket Connection is Closed . . . . . . . . . . 40
101	       7.1.5.  The WebSocket Connection Close Code  . . . . . . . . . 40
102	       7.1.6.  The WebSocket Connection Close Reason  . . . . . . . . 40
103	       7.1.7.  Fail the WebSocket Connection  . . . . . . . . . . . . 41
104	     7.2.   Abnormal Closures . . . . . . . . . . . . . . . . . . . . 41
105	       7.2.1.  Client-Initiated Closure . . . . . . . . . . . . . . . 41
106	       7.2.2.  Server-initiated closure . . . . . . . . . . . . . . . 42
107	     7.3.   Normal Closure of Connections . . . . . . . . . . . . . . 42
108	     7.4.   Status Codes  . . . . . . . . . . . . . . . . . . . . . . 42
109	       7.4.1.  Defined Status Codes . . . . . . . . . . . . . . . . . 42
110	       7.4.2.  Reserved Status Code Ranges  . . . . . . . . . . . . . 43
111	   8.  Error Handling . . . . . . . . . . . . . . . . . . . . . . . . 45
112	     8.1.   Handling Errors in UTF-8 from the Server  . . . . . . . . 45
113	     8.2.   Handling Errors in UTF-8 from the Client  . . . . . . . . 45
114	   9.  Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . 46
115	     9.1.   Negotiating Extensions  . . . . . . . . . . . . . . . . . 46
116	     9.2.   Known Extensions  . . . . . . . . . . . . . . . . . . . . 47
117	       9.2.1.  Compression  . . . . . . . . . . . . . . . . . . . . . 47
118	   10. Security Considerations  . . . . . . . . . . . . . . . . . . . 49
119	   11. IANA Considerations  . . . . . . . . . . . . . . . . . . . . . 51
120	     11.1.  Registration of "ws:" Scheme  . . . . . . . . . . . . . . 51
121	     11.2.  Registration of "wss:" Scheme . . . . . . . . . . . . . . 52
122	     11.3.  Registration of the "WebSocket" HTTP Upgrade Keyword  . . 53
123	     11.4.  Sec-WebSocket-Key . . . . . . . . . . . . . . . . . . . . 53
124	     11.5.  Sec-WebSocket-Extensions  . . . . . . . . . . . . . . . . 54
125	     11.6.  WebSocket Extension Name Registry . . . . . . . . . . . . 55
126	     11.7.  Sec-WebSocket-Accept  . . . . . . . . . . . . . . . . . . 56
127	     11.8.  Sec-WebSocket-Origin  . . . . . . . . . . . . . . . . . . 56
128	     11.9.  Sec-WebSocket-Protocol  . . . . . . . . . . . . . . . . . 57
129	     11.10. WebSocket Subprotocol Name Registry . . . . . . . . . . . 57
130	     11.11. Sec-WebSocket-Version . . . . . . . . . . . . . . . . . . 58
131	     11.12. WebSocket Version Number Registry . . . . . . . . . . . . 59
132	     11.13. WebSocket Close Code Number Registry  . . . . . . . . . . 60
133	     11.14. WebSocket Opcode Registry . . . . . . . . . . . . . . . . 61
134	     11.15. WebSocket Framing Header Bits Registry  . . . . . . . . . 62
135	   12. Using the WebSocket protocol from Other Specifications . . . . 63
136	   13. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 64
137	   14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 65
138	     14.1.  Normative References  . . . . . . . . . . . . . . . . . . 65
139	     14.2.  Informative References  . . . . . . . . . . . . . . . . . 66
140	   Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 68

142	1.  Introduction

144	1.1.  Background

146	   _This section is non-normative._

148	   Historically, creating an instant messenger chat client as a Web
149	   application has required an abuse of HTTP to poll the server for
150	   updates while sending upstream notifications as distinct HTTP
151	   calls.[RFC6202]

153	   This results in a variety of problems:

155	   o  The server is forced to use a number of different underlying TCP
156	      connections for each client: one for sending information to the
157	      client, and a new one for each incoming message.

159	   o  The wire protocol has a high overhead, with each client-to-server
160	      message having an HTTP header.

162	   o  The client-side script is forced to maintain a mapping from the
163	      outgoing connections to the incoming connection to track replies.

165	   A simpler solution would be to use a single TCP connection for
166	   traffic in both directions.  This is what the WebSocket protocol
167	   provides.  Combined with the WebSocket API, it provides an
168	   alternative to HTTP polling for two-way communication from a Web page
169	   to a remote server.  [WSAPI]

171	   The same technique can be used for a variety of Web applications:
172	   games, stock tickers, multiuser applications with simultaneous
173	   editing, user interfaces exposing server-side services in real time,
174	   etc.

176	1.2.  Protocol Overview

178	   _This section is non-normative._

180	   The protocol has two parts: a handshake, and then the data transfer.

182	   The handshake from the client looks as follows:

184	        GET /chat HTTP/1.1
185	        Host: server.example.com
186	        Upgrade: websocket
187	        Connection: Upgrade
188	        Sec-WebSocket-Key: dGhlIHNhbXBsZSBub25jZQ==
189	        Sec-WebSocket-Origin: http://example.com
190	        Sec-WebSocket-Protocol: chat, superchat
191	        Sec-WebSocket-Version: 8

193	   The handshake from the server looks as follows:

195	        HTTP/1.1 101 Switching Protocols
196	        Upgrade: websocket
197	        Connection: Upgrade
198	        Sec-WebSocket-Accept: s3pPLMBiTxaQ9kYGzzhZRbK+xOo=
199	        Sec-WebSocket-Protocol: chat

201	   The leading line from the client follows the Request-Line format.
202	   The leading line from the server follows the Status-Line format.  The
203	   Request-Line and Status-Line productions are defined in [RFC2616].

205	   After the leading line in both cases come an unordered set of header
206	   fields.  The meaning of these header fields is specified in Section 5
207	   of this document.  Additional header fields may also be present, such
208	   as cookies [I-D.ietf-httpstate-cookie] required to identify the user.
209	   The format and parsing of headers is as defined in [RFC2616].

211	   Once the client and server have both sent their handshakes, and if
212	   the handshake was successful, then the data transfer part starts.
213	   This is a two-way communication channel where each side can,
214	   independently from the other, send data at will.

216	   Clients and servers, after a successful handshake, transfer data back
217	   and forth in conceptual units referred to in this specification as
218	   "messages".  A message is a complete unit of data at an application
219	   level, with the expectation that many or most applications
220	   implementing this protocol (such as web user agents) provide APIs in
221	   terms of sending and receiving messages.  The WebSocket message does
222	   not necessarily correspond to a particular network layer framing, as
223	   a fragmented message may be coalesced, or vice versa, e.g. by an
224	   intermediary.

226	   Data is sent on the wire in the form of frames that have an
227	   associated type.  A message is composed of one or more frames, all of
228	   which contain the same type of data.  Broadly speaking, there are
229	   types for textual data, which is interpreted as UTF-8 [RFC3629] text,
230	   binary data (whose interpretation is left up to the application), and
231	   control frames, which are not intended to carry data for the
232	   application, but instead for protocol-level signaling, such as to
233	   signal that the connection should be closed.  This version of the
234	   protocol defines six frame types and leaves ten reserved for future
235	   use.

237	   The WebSocket protocol uses this framing so that specifications that
238	   use the WebSocket protocol can expose such connections using an
239	   event-based mechanism instead of requiring users of those
240	   specifications to implement buffering and piecing together of
241	   messages manually.

243	1.3.  Opening Handshake

245	   _This section is non-normative._

247	   The opening handshake is intended to be compatible with HTTP-based
248	   server-side software and intermediaries, so that a single port can be
249	   used by both HTTP clients talking to that server and WebSocket
250	   clients talking to that server.  To this end, the WebSocket client's
251	   handshake is an HTTP Upgrade request:

253	        GET /chat HTTP/1.1
254	        Host: server.example.com
255	        Upgrade: websocket
256	        Connection: Upgrade
257	        Sec-WebSocket-Key: dGhlIHNhbXBsZSBub25jZQ==
258	        Sec-WebSocket-Origin: http://example.com
259	        Sec-WebSocket-Protocol: chat, superchat
260	        Sec-WebSocket-Version: 8

262	   Headers in the handshake are sent by the client in a random order;
263	   the order is not meaningful.

265	   The "Request-URI" of the GET method [RFC2616] is used to identify the
266	   endpoint of the WebSocket connection, both to allow multiple domains
267	   to be served from one IP address and to allow multiple WebSocket
268	   endpoints to be served by a single server.

270	   The client includes the hostname in the Host header of its handshake
271	   as per [RFC2616], so that both the client and the server can verify
272	   that they agree on which host is in use.

274	   Additional headers are used to select options in the WebSocket
275	   protocol.  Options available in this version are the subprotocol
276	   selector, |Sec-WebSocket-Protocol|, and |Cookie|, which can used for
277	   sending cookies to the server (e.g. as an authentication mechanism).

279	   The |Sec-WebSocket-Protocol| request-header field can be used to
280	   indicate what subprotocols (application-level protocols layered over
281	   the WebSocket protocol) are acceptable to the client.  The server
282	   selects one of the acceptable protocols and echoes that value in its
283	   handshake to indicate that it has selected that protocol.
284	        Sec-WebSocket-Protocol: chat

286	   The |Sec-WebSocket-Origin| header is used to protect against
287	   unauthorized cross-origin use of a WebSocket server by scripts using
288	   the |WebSocket| API in a Web browser.  The server is informed of the
289	   script origin generating the WebSocket connection request.  If the
290	   server does not wish to accept connections from this origin, it can
291	   choose to reject the connection by sending an appropriate HTTP error
292	   code.  This header is sent by browser clients, for non-browser
293	   clients this header may be sent if it makes sense in the context of
294	   those clients.

296	   NOTE: It is worth noting that for the attack cases this header
297	   protects against, the untrusted party is typically the author of a
298	   JavaScript application that is executing in the context of the
299	   client.  The client itself can contact the server and via the
300	   mechanism of the |Sec-WebSocket-Origin| header, determine whether to
301	   extend those communication privileges to the JavaScript application.
302	   A JavaScript application cannot set a header starting with "Sec-" via
303	   XHR.  The intent is not to prevent non-browsers from establishing
304	   connections, but rather to ensure that browsers under the control of
305	   potentially malicious JavaScript cannot fake a WebSocket handshake.

307	   Finally, the server has to prove to the client that it received the
308	   client's WebSocket handshake, so that the server doesn't accept
309	   connections that are not WebSocket connections.  This prevents an
310	   attacker from tricking a WebSocket server by sending it carefully-
311	   crafted packets using |XMLHttpRequest| or a |form| submission.

313	   To prove that the handshake was received, the server has to take two
314	   pieces of information and combine them to form a response.  The first
315	   piece of information comes from the |Sec-WebSocket-Key| header in the
316	   client handshake:

318	        Sec-WebSocket-Key: dGhlIHNhbXBsZSBub25jZQ==

320	   For this header, the server has to take the value (as present in the
321	   header, e.g. the base64-encoded [RFC4648] version minus leading and
322	   trailing whitespace), and concatenate this with the GUID "258EAFA5-
323	   E914-47DA-95CA-C5AB0DC85B11" in string form, which is unlikely to be
324	   used by network endpoints that do not understand the WebSocket
325	   protocol.  A SHA-1 hash (160 bits), base64-encoded, of this
326	   concatenation is then returned in the server's handshake

328	   [FIPS.180-2.2002].

330	   Concretely, if as in the example above, header |Sec-WebSocket-Key|
331	   had the value "dGhlIHNhbXBsZSBub25jZQ==", the server would
332	   concatenate the string "258EAFA5-E914-47DA-95CA-C5AB0DC85B11" to form
333	   the string "dGhlIHNhbXBsZSBub25jZQ==258EAFA5-E914-47DA-95CA-
334	   C5AB0DC85B11".  The server would then take the SHA-1 hash of this,
335	   giving the value 0xb3 0x7a 0x4f 0x2c 0xc0 0x62 0x4f 0x16 0x90 0xf6
336	   0x46 0x06 0xcf 0x38 0x59 0x45 0xb2 0xbe 0xc4 0xea.  This value is
337	   then base64-encoded, to give the value "s3pPLMBiTxaQ9kYGzzhZRbK+
338	   xOo=".  This value would then be echoed in the header |Sec-WebSocket-
339	   Accept|.

341	   The handshake from the server is much simpler than the client
342	   handshake.  The first line is an HTTP Status-Line, with the status
343	   code 101:

345	        HTTP/1.1 101 Switching Protocols

347	   Any status code other than 101 indicates that the WebSocket handshake
348	   has not completed, and that the semantics of HTTP still apply.  The
349	   headers follow the status code.

351	   The |Connection| and |Upgrade| headers complete the HTTP Upgrade.
352	   The |Sec-WebSocket-Accept| header indicates whether the server is
353	   willing to accept the connection.  If present, this header must
354	   include a hash of the client's nonce sent in |Sec-WebSocket-Key|
355	   along with a predefined GUID.  Any other value must not be
356	   interpreted as an acceptance of the connection by the server.

358	        HTTP/1.1 101 Switching Protocols
359	        Upgrade: websocket
360	        Connection: Upgrade
361	        Sec-WebSocket-Accept: s3pPLMBiTxaQ9kYGzzhZRbK+xOo=

363	   These fields are checked by the Web browser when it is acting as a
364	   |WebSocket| client for scripted pages.  If the |Sec-WebSocket-Accept|
365	   value does not match the expected value, or if the header is missing,
366	   or if the HTTP status code is not 101, the connection will not be
367	   established and WebSocket frames will not be sent.

369	   Option fields can also be included.  In this version of the protocol,
370	   the main option field is |Sec-WebSocket-Protocol|, which indicates
371	   the subprotocol that the server has selected.  Web browsers verify
372	   that the server included one of the values as was specified in the
373	   WebSocket client's handshake.  A server that speaks multiple
374	   subprotocols has to make sure it selects one based on the client's
375	   handshake and specifies it in its handshake.

377	        Sec-WebSocket-Protocol: chat

379	   The server can also set cookie-related option fields to _set_
380	   cookies, as in HTTP.

382	1.4.  Closing Handshake

384	   _This section is non-normative._

386	   The closing handshake is far simpler than the opening handshake.

388	   Either peer can send a control frame with data containing a specified
389	   control sequence to begin the closing handshake (detailed in
390	   Section 4.5.1).  Upon receiving such a frame, the other peer sends a
391	   close frame in response, if it hasn't already sent one.  Upon
392	   receiving _that_ control frame, the first peer then closes the
393	   connection, safe in the knowledge that no further data is
394	   forthcoming.

396	   After sending a control frame indicating the connection should be
397	   closed, a peer does not send any further data; after receiving a
398	   control frame indicating the connection should be closed, a peer
399	   discards any further data received.

401	   It is safe for both peers to initiate this handshake simultaneously.

403	   The closing handshake is intended to complement the TCP closing
404	   handshake (FIN/ACK), on the basis that the TCP closing handshake is
405	   not always reliable end-to-end, especially in the presence of man-in-
406	   the-middle proxies and other intermediaries.

408	   By sending a close frame and waiting for a close frame in response,
409	   certain cases are avoided where data may be unnecessarily lost.  For
410	   instance, on some platforms, if a socket is closed with data in the
411	   receive queue, a RST packet is sent, which will then cause recv() to
412	   fail for the party that received the RST, even if there was data
413	   waiting to be read.

415	1.5.  Design Philosophy

417	   _This section is non-normative._

419	   The WebSocket protocol is designed on the principle that there should
420	   be minimal framing (the only framing that exists is to make the
421	   protocol frame-based instead of stream-based, and to support a
422	   distinction between Unicode text and binary frames).  It is expected
423	   that metadata would be layered on top of WebSocket by the application
424	   layer, in the same way that metadata is layered on top of TCP by the
425	   application layer (HTTP).

427	   Conceptually, WebSocket is really just a layer on top of TCP that
428	   adds a Web "origin"-based security model for browsers; adds an
429	   addressing and protocol naming mechanism to support multiple services
430	   on one port and multiple host names on one IP address; layers a
431	   framing mechanism on top of TCP to get back to the IP packet
432	   mechanism that TCP is built on, but without length limits; and
433	   includes an additional closing handshake in-band that is designed to
434	   work in the presence of proxies and other intermediaries.  Other than
435	   that, it adds nothing.  Basically it is intended to be as close to
436	   just exposing raw TCP to script as possible given the constraints of
437	   the Web. It's also designed in such a way that its servers can share
438	   a port with HTTP servers, by having its handshake be a valid HTTP
439	   Upgrade request mechanism also.

441	   The protocol is intended to be extensible; future versions will
442	   likely introduce additional concepts such as multiplexing.

444	1.6.  Security Model

446	   _This section is non-normative._

448	   The WebSocket protocol uses the origin model used by Web browsers to
449	   restrict which Web pages can contact a WebSocket server when the
450	   WebSocket protocol is used from a Web page.  Naturally, when the
451	   WebSocket protocol is used by a dedicated client directly (i.e. not
452	   from a Web page through a Web browser), the origin model is not
453	   useful, as the client can provide any arbitrary origin string.

455	   This protocol is intended to fail to establish a connection with
456	   servers of pre-existing protocols like SMTP [RFC5321] and HTTP, while
457	   allowing HTTP servers to opt-in to supporting this protocol if
458	   desired.  This is achieved by having a strict and elaborate
459	   handshake, and by limiting the data that can be inserted into the
460	   connection before the handshake is finished (thus limiting how much
461	   the server can be influenced).

463	   It is similarly intended to fail to establish a connection when data
464	   from other protocols, especially HTTP, is sent to a WebSocket server,
465	   for example as might happen if an HTML |form| were submitted to a
466	   WebSocket server.  This is primarily achieved by requiring that the
467	   server prove that it read the handshake, which it can only do if the
468	   handshake contains the appropriate parts which themselves can only be
469	   sent by a WebSocket handshake.  In particular, at the time of writing
470	   of this specification, fields starting with |Sec-| cannot be set by
471	   an attacker from a Web browser using only HTML and JavaScript APIs
472	   such as |XMLHttpRequest|.

474	1.7.  Relationship to TCP and HTTP

476	   _This section is non-normative._

478	   The WebSocket protocol is an independent TCP-based protocol.  Its
479	   only relationship to HTTP is that its handshake is interpreted by
480	   HTTP servers as an Upgrade request.

482	   By default the WebSocket protocol uses port 80 for regular WebSocket
483	   connections and port 443 for WebSocket connections tunneled over TLS
484	   [RFC2818].

486	1.8.  Establishing a Connection

488	   _This section is non-normative._

490	   When a connection is to be made to a port that is shared by an HTTP
491	   server (a situation that is quite likely to occur with traffic to
492	   ports 80 and 443), the connection will appear to the HTTP server to
493	   be a regular GET request with an Upgrade offer.  In relatively simple
494	   setups with just one IP address and a single server for all traffic
495	   to a single hostname, this might allow a practical way for systems
496	   based on the WebSocket protocol to be deployed.  In more elaborate
497	   setups (e.g. with load balancers and multiple servers), a dedicated
498	   set of hosts for WebSocket connections separate from the HTTP servers
499	   is probably easier to manage.  At the time of writing of this
500	   specification, it should be noted that connections on port 80 and 443
501	   have significantly different success rates, with connections on port
502	   443 being significantly more likely to succeed, though this may
503	   change with time.

505	1.9.  Subprotocols Using the WebSocket protocol

507	   _This section is non-normative._

509	   The client can request that the server use a specific subprotocol by
510	   including the |Sec-WebSocket-Protocol| field in its handshake.  If it
511	   is specified, the server needs to include the same field and one of
512	   the selected subprotocol values in its response for the connection to
513	   be established.

515	   These subprotocol names should be registered as per Section 11.10.
516	   To avoid potential collisions, it is recommended to use names that
517	   contain the domain name of the subprotocol's originator.  For
518	   example, if Example Corporation were to create a Chat subprotocol to
519	   be implemented by many servers around the Web, they could name it
520	   "chat.example.com".  If the Example Organization called their
521	   competing subprotocol "chat.example.org", then the two subprotocols
522	   could be implemented by servers simultaneously, with the server
523	   dynamically selecting which subprotocol to use based on the value
524	   sent by the client.

526	   Subprotocols can be versioned in backwards-incompatible ways by
527	   changing the subprotocol name, e.g. going from "bookings.example.net"
528	   to "v2.bookings.example.net".  These subprotocols would be considered
529	   completely separate by WebSocket clients.  Backwards-compatible
530	   versioning can be implemented by reusing the same subprotocol string
531	   but carefully designing the actual subprotocol to support this kind
532	   of extensibility.

534	2.  Conformance Requirements

536	   All diagrams, examples, and notes in this specification are non-
537	   normative, as are all sections explicitly marked non-normative.
538	   Everything else in this specification is normative.

540	   The key words "MUST", "MUST NOT", "REQUIRED", "SHOULD", "SHOULD NOT",
541	   "RECOMMENDED", "MAY", and "OPTIONAL" in the normative parts of this
542	   document are to be interpreted as described in RFC2119.  [RFC2119]

544	   Requirements phrased in the imperative as part of algorithms (such as
545	   "strip any leading space characters" or "return false and abort these
546	   steps") are to be interpreted with the meaning of the key word
547	   ("must", "should", "may", etc) used in introducing the algorithm.

549	   Conformance requirements phrased as algorithms or specific steps MAY
550	   be implemented in any manner, so long as the end result is
551	   equivalent.  (In particular, the algorithms defined in this
552	   specification are intended to be easy to follow, and not intended to
553	   be performant.)

555	   Implementations MAY impose implementation-specific limits on
556	   otherwise unconstrained inputs, e.g. to prevent denial of service
557	   attacks, to guard against running out of memory, or to work around
558	   platform-specific limitations.

560	   The conformance classes defined by this specification are clients and
561	   servers.

563	2.1.  Terminology

565	   _ASCII_ shall mean the character-encoding scheme defined in
566	   [ANSI.X3-4.1986].

568	   This document makes reference to UTF-8 values and uses UTF-8
569	   notational formats as defined in STD 63 [RFC3629].

571	   Key Terms such as named algorithms or definitions are indicated like
572	   _this_.

574	   Names of headers or variables are indicated like |this|.

576	   Variable values are indicated like /this/.

578	   _Converting a string to ASCII lowercase_ means replacing all
579	   characters in the range U+0041 to U+005A (i.e.  LATIN CAPITAL LETTER
580	   A to LATIN CAPITAL LETTER Z) with the corresponding characters in the
581	   range U+0061 to U+007A (i.e.  LATIN SMALL LETTER A to LATIN SMALL
582	   LETTER Z).

584	   Comparing two strings in an _ASCII case-insensitive_ manner means
585	   comparing them exactly, code point for code point, except that the
586	   characters in the range U+0041 to U+005A (i.e.  LATIN CAPITAL LETTER
587	   A to LATIN CAPITAL LETTER Z) and the corresponding characters in the
588	   range U+0061 to U+007A (i.e.  LATIN SMALL LETTER A to LATIN SMALL
589	   LETTER Z) are considered to also match.

591	   The term "URI" is used in this document as defined in [RFC3986].

593	   When an implementation is required to _send_ data as part of the
594	   WebSocket protocol, the implementation MAY delay the actual
595	   transmission arbitrarily, e.g. buffering data so as to send fewer IP
596	   packets.

598	3.  WebSocket URIs

600	   This specification defines two URI schemes, using the ABNF syntax
601	   defined in RFC 5234 [RFC5234], and terminology and ABNF productions
602	   defined by the URI specification RFC 3986 [RFC3986].

604	          ws-URI = "ws:" "//" host [ ":" port ] path [ "?" query ]
605	          wss-URI = "wss:" "//" host [ ":" port ] path [ "?" query ]

607	          host = <host, defined in [RFC3986], Section 3.2.2>
608	          port = <port, defined in [RFC3986], Section 3.2.3>
609	          path = <path-abempty, defined in [RFC3986], Section 3.3>
610	          query = <query, defined in [RFC3986], Section 3.4>

612	   The port component is OPTIONAL; the default for "ws" is port 80,
613	   while the default for "wss" is port 443.

615	   The URI is called "secure" if the scheme component matches "wss"
616	   case-insensitively.

618	   The "resource-name" can be constructed by concatenating

620	   o  "/" if the path component is empty

622	   o  the path component

624	   o  "?" if the query component is non-empty

626	   o  the query component

628	   Fragment identifiers are meaningless in the context of WebSocket
629	   URIs, and MUST NOT be used on these URIs.  The character "#" in URIs
630	   MUST be escaped as %23 if used as part of the query component.

632	4.  Data Framing

634	4.1.  Overview

636	   In the WebSocket protocol, data is transmitted using a sequence of
637	   frames.  Frames sent from the client to the server are masked to
638	   avoid confusing network intermediaries, such as intercepting proxies.
639	   Frames sent from the server to the client are not masked.

641	   The base framing protocol defines a frame type with an opcode, a
642	   payload length, and designated locations for extension and
643	   application data, which together define the _payload_ data.  Certain
644	   bits and opcodes are reserved for future expansion of the protocol.

646	   A data frame MAY be transmitted by either the client or the server at
647	   any time after opening handshake completion and before that endpoint
648	   has sent a close frame (Section 4.5.1).

650	4.2.  Base Framing Protocol

652	   This wire format for the data transfer part is described by the ABNF
653	   [RFC5234] given in detail in this section.  A high level overview of
654	   the framing is given in the following figure.

656	      0                   1                   2                   3
657	      0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
658	     +-+-+-+-+-------+-+-------------+-------------------------------+
659	     |F|R|R|R| opcode|M| Payload len |    Extended payload length    |
660	     |I|S|S|S|  (4)  |A|     (7)     |             (16/63)           |
661	     |N|V|V|V|       |S|             |   (if payload len==126/127)   |
662	     | |1|2|3|       |K|             |                               |
663	     +-+-+-+-+-------+-+-------------+ - - - - - - - - - - - - - - - +
664	     |     Extended payload length continued, if payload len == 127  |
665	     + - - - - - - - - - - - - - - - +-------------------------------+
666	     |                               |Masking-key, if MASK set to 1  |
667	     +-------------------------------+-------------------------------+
668	     | Masking-key (continued)       |          Payload Data         |
669	     +-------------------------------- - - - - - - - - - - - - - - - +
670	     :                     Payload Data continued ...                :
671	     + - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - +
672	     |                     Payload Data continued ...                |
673	     +---------------------------------------------------------------+

675	   FIN:  1 bit

677	      Indicates that this is the final fragment in a message.  The first
678	      fragment MAY also be the final fragment.

680	   RSV1, RSV2, RSV3:  1 bit each

682	      MUST be 0 unless an extension is negotiated which defines meanings
683	      for non-zero values.  If a nonzero value is received and none of
684	      the negotiated extensions defines the meaning of such a nonzero
685	      value, the receiving endpoint MUST _Fail the WebSocket
686	      Connection_.

688	   Opcode:  4 bits

690	      Defines the interpretation of the payload data.  If an unknown
691	      opcode is received, the receiving endpoint MUST ignore that frame.
692	      The following values are defined.

694	      *  %x0 denotes a continuation frame

696	      *  %x1 denotes a text frame

698	      *  %x2 denotes a binary frame

700	      *  %x3-7 are reserved for further non-control frames

702	      *  %x8 denotes a connection close

704	      *  %x9 denotes a ping

706	      *  %xA denotes a pong

708	      *  %xB-F are reserved for further control frames

710	   Mask:  1 bit

712	      Defines whether the payload data is masked.  If set to 1, a
713	      masking key is present in masking-key, and this is used to unmask
714	      the payload data as per Section 4.3.  All frames sent from client
715	      to server have this bit set to 1.

717	   Payload length:  7 bits, 7+16 bits, or 7+64 bits

719	      The length of the payload data, in bytes: if 0-125, that is the
720	      payload length.  If 126, the following 2 bytes interpreted as a 16
721	      bit unsigned integer are the payload length.  If 127, the
722	      following 8 bytes interpreted as a 64-bit unsigned integer (the
723	      most significant bit MUST be 0) are the payload length.  Multibyte
724	      length quantities are expressed in network byte order.  The
725	      payload length is the length of the extension data + the length of
726	      the application data.  The length of the extension data may be
727	      zero, in which case the payload length is the length of the
728	      application data.

730	   Masking-key:  0 or 4 bytes

732	      All frames sent from the client to the server are masked by a 32-
733	      bit value that is contained within the frame.  This field is
734	      present if the mask bit is set to 1, and is absent if the mask bit
735	      is set to 0.  See Section 4.3 for further information on client-
736	      to-server masking.

738	   Payload data:  (x+y) bytes

740	      The payload data is defined as extension data concatenated with
741	      application data.

743	   Extension data:  x bytes

745	      The extension data is 0 bytes unless an extension has been
746	      negotiated.  Any extension MUST specify the length of the
747	      extension data, or how that length may be calculated, and how the
748	      extension use MUST be negotiated during the opening handshake.  If
749	      present, the extension data is included in the total payload
750	      length.

752	   Application data:  y bytes

754	      Arbitrary application data, taking up the remainder of the frame
755	      after any extension data.  The length of the application data is
756	      equal to the payload length minus the length of the extension
757	      data.

759	   The base framing protocol is formally defined by the following ABNF
760	   [RFC5234]:

762	   ws-frame                = frame-fin
763	                             frame-rsv1
764	                             frame-rsv2
765	                             frame-rsv3
766	                             frame-opcode
767	                             frame-masked
768	                             frame-payload-length
769	                             [ frame-masking-key ]
770	                             frame-payload-data

772	   frame-fin               = %x0 ; more frames of this message follow
773	                           / %x1 ; final frame of this message

775	   frame-rsv1              = %x0 ; 1 bit, MUST be 0

777	   frame-rsv2              = %x0 ; 1 bit, MUST be 0

779	   frame-rsv3              = %x0 ; 1 bit, MUST be 0

781	   frame-opcode            = %x0 ; continuation frame
782	                           / %x1 ; text frame
783	                           / %x2 ; binary frame
784	                           / %x3-7 ; reserved for further non-control frames
785	                           / %x8 ; connection close
786	                           / %x9 ; ping
787	                           / %xA ; pong
788	                           / %xB-F ; reserved for further control frames

790	   frame-masked            = %x0 ; frame is not masked, no frame-masking-key
791	                           / %x1 ; frame is masked, frame-masking-key present

793	   frame-payload-length    = %x00-7D
794	                           / %x7E frame-payload-length-16
795	                           / %x7F frame-payload-length-63

797	   frame-payload-length-16 = %x0000-FFFF

799	   frame-payload-length-63 = %x0000000000000000-7FFFFFFFFFFFFFFF

801	   frame-masking-key       = 4( %0x00-FF ) ; present only if frame-masked is 1

803	   frame-payload-data      = (frame-masked-extension-data
804	                              frame-masked-application-data)   ; frame-masked 1
805	                           / (frame-unmasked-extension-data
806	                              frame-unmasked-application-data) ; frame-masked 0

808	   frame-masked-extension-data     = *( %x00-FF ) ; to be defined later

810	   frame-masked-application-data   = *( %x00-FF )

812	   frame-unmasked-extension-data   = *( %x00-FF ) ; to be defined later

814	   frame-unmasked-application-data = *( %x00-FF )

816	4.3.  Client-to-Server Masking

818	   The client MUST mask all frames sent to the server.  A server MUST
819	   close the connection upon receiving a frame with the MASK bit set to
820	   0.  In this case, a server MAY send a close frame with a status code
821	   of 1002 (protocol error) as defined in Section 7.4.1.

823	   A masked frame MUST have the field frame-masked set to 1, as defined
824	   in Section 4.2.

826	   The masking key is contained completely within the frame, as defined
827	   in Section 4.2 as frame-masking-key.  It is used to mask the payload
828	   data defined in the same section as frame-payload-data, which
829	   includes extension and application data.

831	   The masking key is a 32-bit value chosen at random by the client.
832	   The masking key MUST be derived from a strong source of entropy, and
833	   the masking key for a given frame MUST NOT make it simple for a
834	   server to predict the masking key for a subsequent frame.  RFC 4086
835	   [RFC4086] discusses what entails a suitable source of entropy for
836	   security-sensitive applications.

838	   The masking does not affect the length of the payload data.  To
839	   convert masked data into unmasked data, or vice versa, the following
840	   algorithm is applied.  The same algorithm applies regardless of the
841	   direction of the translation - e.g. the same steps are applied to
842	   mask the data as to unmask the data.

844	   Octet i of the transformed data ("transformed-octet-i") is the XOR of
845	   octet i of the original data ("original-octet-i") with octet i modulo
846	   4 of the masking key ("masking-key-octet-j"):

848	     j                   = i MOD 4
849	     transformed-octet-i = original-octet-i XOR masking-key-octet-j

851	   When preparing a masked frame, the client MUST pick a fresh masking
852	   key uniformly at random from the set of allowed 32-bit values.  The
853	   unpredictability of the masking key is essential to prevent the
854	   author of malicious applications from selecting the bytes that appear
855	   on the wire.

857	   The payload length, indicated in the framing as frame-payload-length,
858	   does NOT include the length of the masking key.  It is the length of
859	   the payload data, e.g. the number of bytes following the masking key.

861	4.4.  Fragmentation

863	   The primary purpose of fragmentation is to allow sending a message
864	   that is of unknown size when the message is started without having to
865	   buffer that message.  If messages couldn't be fragmented, then an
866	   endpoint would have to buffer the entire message so its length could
867	   be counted before first byte is sent.  With fragmentation, a server
868	   or intermediary may choose a reasonable size buffer, and when the
869	   buffer is full write a fragment to the network.

871	   A secondary use-case for fragmentation is for multiplexing, where it
872	   is not desirable for a large message on one logical channel to
873	   monopolize the output channel, so the MUX needs to be free to split
874	   the message into smaller fragments to better share the output
875	   channel.

877	   The following rules apply to fragmentation:

879	   o  An unfragmented message consists of a single frame with the FIN
880	      bit set and an opcode other than 0.

882	   o  A fragmented message consists of a single frame with the FIN bit
883	      clear and an opcode other than 0, followed by zero or more frames
884	      with the FIN bit clear and the opcode set to 0, and terminated by
885	      a single frame with the FIN bit set and an opcode of 0.  A
886	      fragmented message is conceptually equivalent to a single larger
887	      message whose payload is equal to the concatenation of the
888	      payloads of the fragments in order, however in the presence of
889	      extensions this may not hold true as the extension defines the
890	      interpretation of the extension data present.  For instance,
891	      extension data may only be present at the beginning of the first
892	      fragment and apply to subsequent fragments, or there may be
893	      extension data present in each of the fragments that applies only
894	      to that particular fragment.  Setting aside the issue of
895	      extensions, the following example demonstrates how fragmentation
896	      works.

898	      EXAMPLE: For a text message sent as three fragments, the first
899	      fragment would have an opcode of 0x1 and a FIN bit clear, the
900	      second fragment would have an opcode of 0x0 and a FIN bit clear,
901	      and the third fragment would have an opcode of 0x0 and a FIN bit
902	      that is set.

904	   o  Control frames MAY be injected in the middle of a fragmented
905	      message.  Control frames themselves MUST NOT be fragmented.

907	   o  Message fragments MUST be delivered to the recipient in the order
908	      sent by the sender.

910	   o  The fragments of one message MUST NOT be interleaved between the
911	      fragments of another message unless an extension has been
912	      negotiated that can interpret the interleaving.

914	   o  An endpoint MUST be capable of handling control frames in the
915	      middle of a fragmented message.

917	   o  A sender MAY create fragments of any size for non-control
918	      messages.

920	   o  Clients and servers MUST support receiving both fragmented and
921	      unfragmented messages.

923	   o  As control frames cannot be fragmented, an intermediary MUST NOT
924	      attempt to change the fragmentation of a control frame.

926	   o  An intermediary MUST NOT change the fragmentation of a message if
927	      any reserved bit values are used and the meaning of these values
928	      is not known to the intermediary.

930	   o  An intermediary MUST NOT change the fragmentation of any message
931	      in the context of a connection where extensions have been
932	      negotiated and the intermediary is not aware of the semantics of
933	      the negotiated extensions.

935	   o  As a consequence of these rules, all fragments of a message are of
936	      the same type, as set by the first fragment's opcode.  Since
937	      Control frames cannot be fragmented, the type for all fragments in
938	      a message MUST be either text or binary, or one of the reserved
939	      opcodes.

941	   _Note: if control frames could not be interjected, the latency of a
942	   ping, for example, would be very long if behind a large message.
943	   Hence, the requirement of handling control frames in the middle of a
944	   fragmented message._

946	4.5.  Control Frames

948	   Control frames are identified by opcodes where the most significant
949	   bit of the opcode is 1.  Currently defined opcodes for control frames
950	   include 0x8 (Close), 0x9 (Ping), and 0xA (Pong).  Opcodes 0xB-0xF are
951	   reserved for further control frames yet to be defined.

953	   Control frames are used to communicate state about the WebSocket.
954	   Control frames can be interjected in the middle of a fragmented
955	   message.

957	   All control frames MUST have a payload length of 125 bytes or less
958	   and MUST NOT be fragmented.

960	4.5.1.  Close

962	   The Close frame contains an opcode of 0x8.

964	   The Close frame MAY contain a body (the "application data" portion of
965	   the frame) that indicates a reason for closing, such as an endpoint
966	   shutting down, an endpoint having received a frame too large, or an
967	   endpoint having received a frame that does not conform to the format
968	   expected by the other endpoint.  If there is a body, the first two
969	   bytes of the body MUST be a 2-byte unsigned integer (in network byte
970	   order) representing a status code with value /code/ defined in
971	   Section 7.4.  Following the 2-byte integer the body MAY contain UTF-8
972	   encoded data with value /reason/, the interpretation of which is not
973	   defined by this specification.  This data is not necessarily human
974	   readable, but may be useful for debugging or passing information
975	   relevant to the script that opened the connection.

977	   Close frames sent from client to server must be masked as per
978	   Section 4.3.

980	   The application MUST NOT send any more data frames after sending a
981	   close frame.

983	   If an endpoint receives a Close frame and that endpoint did not
984	   previously send a Close frame, the endpoint MUST send a Close frame
985	   in response.  It SHOULD do so as soon as is practical.  An endpoint
986	   MAY delay sending a close frame until its current message is sent
987	   (for instance, if the majority of a fragmented message is already
988	   sent, an endpoint MAY send the remaining fragments before sending a
989	   Close frame).  However, there is no guarantee that the endpoint which
990	   has already sent a Close frame will continue to process data.

992	   After both sending and receiving a close message, an endpoint
993	   considers the WebSocket connection closed, and MUST close the
994	   underlying TCP connection.  The server MUST close the underlying TCP
995	   connection immediately; the client SHOULD wait for the server to
996	   close the connection but MAY close the connection at any time after
997	   sending and receiving a close message, e.g. if it has not received a
998	   TCP close from the server in a reasonable time period.

1000	   If a client and server both send a Close message at the same time,
1001	   both endpoints will have sent and received a Close message and should
1002	   consider the WebSocket connection closed and close the underlying TCP
1003	   connection.

1005	4.5.2.  Ping

1007	   The Ping frame contains an opcode of 0x9.

1009	   Upon receipt of a Ping frame, an endpoint MUST send a Pong frame in
1010	   response.  It SHOULD do so as soon as is practical.  Pong frames are
1011	   discussed in Section 4.5.3.

1013	   An endpoint MAY send a Ping frame any time after the connection is
1014	   established and before the connection is closed.  NOTE: A ping frame
1015	   may serve either as a keepalive, or to verify that the remote
1016	   endpoint is still responsive.

1018	4.5.3.  Pong

1020	   The Pong frame contains an opcode of 0xA.

1022	   Section 4.5.2 details requirements that apply to both Ping and Pong
1023	   frames.

1025	   A Pong frame sent in response to a Ping frame must have identical
1026	   Application Data as found in the message body of the Ping frame being
1027	   replied to.

1029	   If an endpoint receives a Ping frame and has not yet sent Pong
1030	   frame(s) in response to previous Ping frame(s), the endpoint MAY
1031	   elect to send a Pong frame for only the most recently processed Ping
1032	   frame.

1034	   A Pong frame MAY be sent unsolicited.  This serves as a
1035	   unidirectional heartbeat.  A response to an unsolicited pong is not
1036	   expected.

1038	4.6.  Data Frames

1040	   Data frames (e.g. non-control frames) are identified by opcodes where
1041	   the most significant bit of the opcode is 0.  Currently defined
1042	   opcodes for data frames include 0x1 (Text), 0x2 (Binary).  Opcodes
1043	   0x3-0x7 are reserved for further non-control frames yet to be
1044	   defined.

1046	   Data frames carry application-layer or extension-layer data.  The
1047	   opcode determines the interpretation of the data:

1049	   Text

1051	      The payload data is text data encoded as UTF-8.

1053	   Binary

1055	      The payload data is arbitrary binary data whose interpretation is
1056	      solely up to the application layer.

1058	4.7.  Examples

1060	   _This section is non-normative._

1062	   o  A single-frame unmasked text message

1064	      *  0x81 0x05 0x48 0x65 0x6c 0x6c 0x6f (contains "Hello")

1066	   o  A single-frame masked text message

1068	      *  0x81 0x85 0x37 0xfa 0x21 0x3d 0x7f 0x9f 0x4d 0x51 0x58
1069	         (contains "Hello")

1071	   o  A fragmented unmasked text message

1073	      *  0x01 0x03 0x48 0x65 0x6c (contains "Hel")

1075	      *  0x80 0x02 0x6c 0x6f (contains "lo")

1077	   o  Ping request and response

1079	      *  0x89 0x05 0x48 0x65 0x6c 0x6c 0x6f (contains a body of "Hello",
1080	         but the contents of the body are arbitrary)

1082	      *  0x8a 0x05 0x48 0x65 0x6c 0x6c 0x6f (contains a body of "Hello",
1083	         matching the body of the ping)

1085	   o  256 bytes binary message in a single unmasked frame

1087	      *  0x82 0x7E 0x0100 [256 bytes of binary data]

1089	   o  64KiB binary message in a single unmasked frame

1091	      *  0x82 0x7F 0x0000000000010000 [65536 bytes of binary data]

1093	4.8.  Extensibility

1095	   The protocol is designed to allow for extensions, which will add
1096	   capabilities to the base protocols.  The endpoints of a connection
1097	   MUST negotiate the use of any extensions during the opening
1098	   handshake.  This specification provides opcodes 0x3 through 0x7 and
1099	   0xB through 0xF, the extension data field, and the frame-rsv1, frame-
1100	   rsv2, and frame-rsv3 bits of the frame header for use by extensions.

1102	   The negotiation of extensions is discussed in further detail in
1103	   Section 9.1.  Below are some anticipated uses of extensions.  This
1104	   list is neither complete nor proscriptive.

1106	   o  Extension data may be placed in the payload data before the
1107	      application data.

1109	   o  Reserved bits can be allocated for per-frame needs.

1111	   o  Reserved opcode values can be defined.

1113	   o  Reserved bits can be allocated to the opcode field if more opcode
1114	      values are needed.

1116	   o  A reserved bit or an "extension" opcode can be defined which
1117	      allocates additional bits out of the payload data to define larger
1118	      opcodes or more per-frame bits.

1120	5.  Opening Handshake

1122	5.1.  Client Requirements

1124	   To _Establish a WebSocket Connection_, a client opens a connection
1125	   and sends a handshake as defined in this section.  A connection is
1126	   defined to initially be in a CONNECTING state.  A client will need to
1127	   supply a /host/, /port/, /resource name/, and a /secure/ flag, which
1128	   are the components of a WebSocket URI as discussed in Section 3,
1129	   along with a list of /protocols/ and /extensions/ to be used.
1130	   Additionally, if the client is a web browser, an /origin/ MUST be
1131	   supplied.

1133	   Clients running in controlled environments, e.g. browsers on mobile
1134	   handsets tied to specific carriers, may offload the management of the
1135	   connection to another agent on the network.  In such a situation, the
1136	   client for the purposes of conformance is considered to include both
1137	   the handset software and any such agents.

1139	   When the client is to _Establish a WebSocket Connection_ given a set
1140	   of (/host/, /port/, /resource name/, and /secure/ flag), along with a
1141	   list of /protocols/ and /extensions/ to be used, and an /origin/ in
1142	   the case of web browsers, it MUST open a connection, send an opening
1143	   handshake, and read the server's handshake in response.  The exact
1144	   requirements of how the connection should be opened, what should be
1145	   sent in the opening handshake, and how the server's response should
1146	   be interpreted, are as follows in this section.  In the following
1147	   text, we will use terms from Section 3 such as "/host/" and "/secure/
1148	   flag" as defined in that section.

1150	   1.  The components of the WebSocket URI passed into this algorithm
1151	       (/host/, /port/, /resource name/ and /secure/ flag) MUST be valid
1152	       according to the specification of WebSocket URIs specified in
1153	       Section 3.  If any of the components are invalid, the client MUST
1154	       _Fail the WebSocket Connection_ and abort these steps.

1156	   2.  If the client already has a WebSocket connection to the remote
1157	       host (IP address) identified by /host/ and port /port/ pair, even
1158	       if the remote host is known by another name, the client MUST wait
1159	       until that connection has been established or for that connection
1160	       to have failed.  There MUST be no more than one connection in a
1161	       CONNECTING state.  If multiple connections to the same IP address
1162	       are attempted simultaneously, the client MUST serialize them so
1163	       that there is no more than one connection at a time running
1164	       through the following steps.

1166	       If the client cannot determine the IP address of the remote host
1167	       (for example because all communication is being done through a
1168	       proxy server that performs DNS queries itself), then the client
1169	       MUST assume for the purposes of this step that each host name
1170	       refers to a distinct remote host, and should instead limit the
1171	       total number of simultaneous connections that are not established
1172	       to a reasonably low number (e.g., in a Web browser, simultaneous
1173	       pending connections to a.example.com and b.example.com would be
1174	       allowed, but if thirty connections are requested, that may not be
1175	       allowed.  The limit should consider the number of tabs the user
1176	       has open.

1178	       NOTE: This makes it harder for a script to perform a denial of
1179	       service attack by just opening a large number of WebSocket
1180	       connections to a remote host.  A server can further reduce the
1181	       load on itself when attacked by making use of this by pausing
1182	       before closing the connection, as that will reduce the rate at
1183	       which the client reconnects.

1185	       NOTE: There is no limit to the number of established WebSocket
1186	       connections a client can have with a single remote host.  Servers
1187	       can refuse to accept connections from hosts with an excessive
1188	       number of existing connections, or disconnect resource-hogging
1189	       connections when suffering high load.

1191	   3.  _Proxy Usage_: If the client is configured to use a proxy when
1192	       using the WebSocket protocol to connect to host /host/ and/or
1193	       port /port/, then the client SHOULD connect to that proxy and ask
1194	       it to open a TCP connection to the host given by /host/ and the
1195	       port given by /port/.

1197	          EXAMPLE: For example, if the client uses an HTTP proxy for all
1198	          traffic, then if it was to try to connect to port 80 on server
1199	          example.com, it might send the following lines to the proxy
1200	          server:

1202	              CONNECT example.com:80 HTTP/1.1
1203	              Host: example.com

1205	          If there was a password, the connection might look like:

1207	              CONNECT example.com:80 HTTP/1.1
1208	              Host: example.com
1209	              Proxy-authorization: Basic ZWRuYW1vZGU6bm9jYXBlcyE=

1211	       If the client is not configured to use a proxy, then a direct TCP
1212	       connection SHOULD be opened to the host given by /host/ and the
1213	       port given by /port/.

1215	       NOTE: Implementations that do not expose explicit UI for
1216	       selecting a proxy for WebSocket connections separate from other
1217	       proxies are encouraged to use a SOCKS proxy for WebSocket
1218	       connections, if available, or failing that, to prefer the proxy
1219	       configured for HTTPS connections over the proxy configured for
1220	       HTTP connections.

1222	       For the purpose of proxy autoconfiguration scripts, the URI to
1223	       pass the function MUST be constructed from /host/, /port/,
1224	       /resource name/, and the /secure/ flag using the definition of a
1225	       WebSocket URI as given in Section 3.

1227	       NOTE: The WebSocket protocol can be identified in proxy
1228	       autoconfiguration scripts from the scheme ("ws:" for unencrypted
1229	       connections and "wss:" for encrypted connections).

1231	   4.  If the connection could not be opened, either because a direct
1232	       connection failed or because any proxy used returned an error,
1233	       then the client MUST _Fail the WebSocket Connection_ and abort
1234	       the connection attempt.

1236	   5.  If /secure/ is true, the client MUST perform a TLS handshake over
1237	       the connection after opening the connection and before sending
1238	       the handshake data [RFC2818].  If this fails (e.g. the server's
1239	       certificate could not be verified), then the client MUST _Fail
1240	       the WebSocket Connection_ and abort the connection.  Otherwise,
1241	       all further communication on this channel MUST run through the
1242	       encrypted tunnel.  [RFC5246]

1244	       Clients MUST use the Server Name Indication extension in the TLS
1245	       handshake.  [RFC6066]

1247	   Once a connection to the server has been established (including a
1248	   connection via a proxy or over a TLS-encrypted tunnel), the client
1249	   MUST send an opening handshake to the server.  The handshake consists
1250	   of an HTTP upgrade request, along with a list of required and
1251	   optional headers.  The requirements for this handshake are as
1252	   follows.

1254	   1.   The handshake MUST be a valid HTTP request as specified by
1255	        [RFC2616].

1257	   2.   The Method of the request MUST be GET and the HTTP version MUST
1258	        be at least 1.1.

1260	        For example, if the WebSocket URI is "ws://example.com/chat",
1261	        The first line sent should be "GET /chat HTTP/1.1"

1263	   3.   The request MUST contain a "Request-URI" as part of the GET
1264	        method.  This MUST match the /resource name/ Section 3 (a
1265	        relative URI), or be an absolute URI that, when parsed, has a
1266	        matching /resource name/ as well as matching /host/, /port/, and
1267	        appropriate scheme (ws or wss).

1269	   4.   The request MUST contain a "Host" header whose value is equal to
1270	        /host/.

1272	   5.   The request MUST contain an "Upgrade" header whose value is
1273	        equal to "websocket".

1275	   6.   The request MUST contain a "Connection" header whose value MUST
1276	        include the "Upgrade" token.

1278	   7.   The request MUST include a header with the name "Sec-WebSocket-
1279	        Key".  The value of this header MUST be a nonce consisting of a
1280	        randomly selected 16-byte value that has been base64-encoded
1281	        [RFC3548].  The nonce MUST be selected randomly for each
1282	        connection.

1284	        NOTE: As an example, if the randomly selected value was the
1285	        sequence of bytes 0x01 0x02 0x03 0x04 0x05 0x06 0x07 0x08 0x09
1286	        0x0a 0x0b 0x0c 0x0d 0x0e 0x0f 0x10, the value of the header
1287	        would be "AQIDBAUGBwgJCgsMDQ4PEC=="

1289	   8.   The request MUST include a header with the name "Sec-WebSocket-
1290	        Origin" if the request is coming from a browser client.  If the
1291	        connection is from a non-browser client, the request MAY include
1292	        this header if the semantics of that client match the use-case
1293	        described here for browser clients.  The value of this header
1294	        MUST be the ASCII serialization of origin of the context in
1295	        which the code establishing the connection is running, and MUST
1296	        be lower-case.  The value MUST NOT contain letters in the range
1297	        U+0041 to U+005A (i.e.  LATIN CAPITAL LETTER A to LATIN CAPITAL
1298	        LETTER Z) [I-D.ietf-websec-origin].  The ABNF is as defined in
1299	        Section 6.1 of [I-D.ietf-websec-origin].

1301	        As an example, if code is running on www.example.com attempting
1302	        to establish a connection to ww2.example.com, the value of the
1303	        header would be "http://www.example.com".

1305	   9.   The request MUST include a header with the name "Sec-WebSocket-
1306	        Version".  The value of this header MUST be 8. _Note: Although a
1307	        draft -09 was published, as -09 was comprised of editorial
1308	        changes and not changes to the wire protocol, 9 was not used as
1309	        a valid value for Sec-WebSocket-Version.  This value was
1310	        reserved in the IANA registry but was not and will not be used.

1312	        If subsequent changes to the wire protocol are necessary, 9 will
1313	        be skipped to prevent confusion with the draft 9 protocol._

1315	   10.  The request MAY include a header with the name "Sec-WebSocket-
1316	        Protocol".  If present, this value indicates the subprotocol(s)
1317	        the client wishes to speak, ordered by preference.  The elements
1318	        that comprise this value MUST be non-empty strings with
1319	        characters in the range U+0021 to U+007E not including separator
1320	        characters as defined in [RFC2616], and MUST all be unique
1321	        strings.  The ABNF for the value of this header is 1#token,
1322	        where the definitions of constructs and rules are as given in
1323	        [RFC2616].

1325	   11.  The request MAY include a header with the name "Sec-WebSocket-
1326	        Extensions".  If present, this value indicates the protocol-
1327	        level extension(s) the client wishes to speak.  The
1328	        interpretation and format of this header is described in
1329	        Section 9.1.

1331	   12.  The request MAY include headers associated with sending cookies,
1332	        as defined by the appropriate specifications
1333	        [I-D.ietf-httpstate-cookie].  These headers are referred to as
1334	        _Headers to Send Appropriate Cookies_.

1336	   Once the client's opening handshake has been sent, the client MUST
1337	   wait for a response from the server before sending any further data.
1338	   The client MUST validate the server's response as follows:

1340	   1.  If the status code received from the server is not 101, the
1341	       client handles the response per HTTP procedures.  Otherwise,
1342	       proceed as follows.

1344	   2.  If the response lacks an "Upgrade" header or the "Upgrade" header
1345	       contains a value that is not an ASCII case-insensitive match for
1346	       the value "websocket", the client MUST _Fail the WebSocket
1347	       Connection _.

1349	   3.  If the response lacks a "Connection" header or the "Connection"
1350	       header contains a value that is not an ASCII case-insensitive
1351	       match for the value "Upgrade", the client MUST _Fail the
1352	       WebSocket Connection_.

1354	   4.  If the response lacks a "Sec-WebSocket-Accept" header or the
1355	       "Sec-WebSocket-Accept" contains a value other than the base64-
1356	       encoded SHA-1 of the concatenation of the "Sec-WebSocket-Key" (as
1357	       a string, not base64-decoded) with the string "258EAFA5-E914-
1358	       47DA-95CA-C5AB0DC85B11", the client MUST _Fail the WebSocket
1359	       Connection_

1361	   5.  If the response includes a "Sec-WebSocket-Extensions" header, and
1362	       this header indicates the use of an extension that was not
1363	       present in the client' handshake (the server has indicated an
1364	       extension not requested by the client), the client MUST _Fail the
1365	       WebSocket Connection_.  (The parsing of this header to determine
1366	       which extensions are requested is discussed in Section 9.1.)

1368	   If the server's response is validated as provided for above, it is
1369	   said that _The WebSocket Connection is Established_ and that the
1370	   WebSocket Connection is in the OPEN state.  The _Extensions In Use_
1371	   is defined to be a (possibly empty) string, the value of which is
1372	   equal to the value of the |Sec-WebSocket-Extensions| header supplied
1373	   by the server's handshake, or the null value if that header was not
1374	   present in the server's handshake.  The _Subprotocol In Use_ is
1375	   defined to be the value of the |Sec-WebSocket-Protocol| header in the
1376	   server's handshake, or the null value if that header was not present
1377	   in the server's handshake.  Additionally, if any headers in the
1378	   server's handshake indicate that cookies should be set (as defined by
1379	   [I-D.ietf-httpstate-cookie]), these cookies are referred to as
1380	   _Cookies Set During the Server's Opening Handshake_.

1382	5.2.  Server-side Requirements

1384	   _This section only applies to servers._

1386	   Servers MAY offload the management of the connection to other agents
1387	   on the network, for example load balancers and reverse proxies.  In
1388	   such a situation, the server for the purposes of conformance is
1389	   considered to include all parts of the server-side infrastructure
1390	   from the first device to terminate the TCP connection all the way to
1391	   the server that processes requests and sends responses.

1393	   EXAMPLE: For example, a data center might have a server that responds
1394	   to WebSocket requests with an appropriate handshake, and then passes
1395	   the connection to another server to actually process the data frames.
1396	   For the purposes of this specification, the "server" is the
1397	   combination of both computers.

1399	5.2.1.  Reading the Client's Opening Handshake

1401	   When a client starts a WebSocket connection, it sends its part of the
1402	   opening handshake.  The server must parse at least part of this
1403	   handshake in order to obtain the necessary information to generate
1404	   the server part of the handshake.

1406	   The client's opening handshake consists of the following parts.  If
1407	   the server, while reading the handshake, finds that the client did
1408	   not send a handshake that matches the description below, the server
1409	   MUST stop processing the client's handshake, and return an HTTP
1410	   response with an appropriate error code (such as 400 Bad Request).

1412	   1.  An HTTP/1.1 or higher GET request, including a "Request-URI"
1413	       [RFC2616] that should be interpreted as a /resource name/
1414	       Section 3.

1416	   2.  A "Host" header containing the server's authority.

1418	   3.  A "Sec-WebSocket-Key" header with a base64-encoded value that,
1419	       when decoded, is 16 bytes in length.

1421	   4.  A "Sec-WebSocket-Version" header, with a value of 8.

1423	   5.  Optionally, a "Sec-WebSocket-Origin" header.  This header is sent
1424	       by all browser clients.  A connection attempt lacking this header
1425	       SHOULD NOT be interpreted as coming from a browser client.

1427	   6.  Optionally, a "Sec-WebSocket-Protocol" header, with a list of
1428	       values indicating which protocols the client would like to speak,
1429	       ordered by preference.

1431	   7.  Optionally, a "Sec-WebSocket-Extensions" header, with a list of
1432	       values indicating which extensions the client would like to
1433	       speak.  The interpretation of this header is discussed in
1434	       Section 9.1.

1436	   8.  Optionally, other headers, such as those used to send cookies to
1437	       a server.  Unknown headers MUST be ignored.

1439	5.2.2.  Sending the Server's Opening Handshake

1441	   When a client establishes a WebSocket connection to a server, the
1442	   server MUST complete the following steps to accept the connection and
1443	   send the server's opening handshake.

1445	   1.  If the server supports encryption, perform a TLS handshake over
1446	       the connection.  If this fails (e.g. the client indicated a host
1447	       name in the extended client hello "server_name" extension that
1448	       the server does not host), then close the connection; otherwise,
1449	       all further communication for the connection (including the
1450	       server's handshake) MUST run through the encrypted tunnel.
1451	       [RFC5246]

1453	   2.  Establish the following information:

1455	       /origin/
1456	          The |Sec-WebSocket-Origin| header in the client's handshake
1457	          indicates the origin of the script establishing the
1458	          connection.  The origin is serialized to ASCII and converted
1459	          to lowercase.  The server MAY use this information as part of
1460	          a determination of whether to accept the incoming connection.
1461	          If the server does not validate the origin, it will accept
1462	          connections from anywhere.  If the server does not wish to
1463	          accept this connection, it MUST return an appropriate HTTP
1464	          error code (e.g. 403 Forbidden) and abort the WebSocket
1465	          handshake described in this section.  For more detail, refer
1466	          to Section 10.

1468	       /key/
1469	          The |Sec-WebSocket-Key| header in the client's handshake
1470	          includes a base64-encoded value that, if decoded, is 16 bytes
1471	          in length.  This (encoded) value is used in the creation of
1472	          the server's handshake to indicate an acceptance of the
1473	          connection.  It is not necessary for the server to base64-
1474	          decode the "Sec-WebSocket-Key" value.

1476	       /version/
1477	          The |Sec-WebSocket-Version| header in the client's handshake
1478	          includes the version of the WebSocket protocol the client is
1479	          attempting to communicate with.  If this version does not
1480	          match a version understood by the server, the server MUST
1481	          abort the websocket handshake described in this section and
1482	          instead send an appropriate HTTP error code (such as 426
1483	          Upgrade Required), and a |Sec-WebSocket-Version| header
1484	          indicating the version(s) the server is capable of
1485	          understanding.

1487	       /resource name/
1488	          An identifier for the service provided by the server.  If the
1489	          server provides multiple services, then the value should be
1490	          derived from the resource name given in the client's handshake
1491	          from the Request-URI [RFC2616] of the GET method.  If the
1492	          requested service is not available, the server MUST send an
1493	          appropriate HTTP error code (such as 404 Not Found) and abort
1494	          the WebSocket handshake.

1496	       /subprotocol/
1497	          Either a single value or null, representing the subprotocol
1498	          the server is ready to use.  If the server supports multiple
1499	          subprotocols, then the value MUST be derived from the client's
1500	          handshake, specifically by selecting one of the values from
1501	          the "Sec-WebSocket-Protocol" field.  The absence of such a
1502	          field is equivalent to the null value.  The empty string is
1503	          not the same as the null value for these purposes, and is not
1504	          a legal value for this field.  The ABNF for the value of this
1505	          header is (token), where the definitions of constructs and
1506	          rules are as given in [RFC2616].

1508	       /extensions/
1509	          A (possibly empty) list representing the protocol-level
1510	          extensions the server is ready to use.  If the server supports
1511	          multiple extensions, then the value MUST be derived from the
1512	          client's handshake, specifically by selecting one or more of
1513	          the values from the "Sec-WebSocket-Extensions" field.  The
1514	          absence of such a field is equivalent to the null value.  The
1515	          empty string is not the same as the null value for these
1516	          purposes.  Extensions not listed by the client MUST NOT be
1517	          listed.  The method by which these values should be selected
1518	          and interpreted is discussed in Section 9.1.

1520	   3.  If the server chooses to accept the incoming connection, it MUST
1521	       reply with a valid HTTP response indicating the following.

1523	       1.  A Status-Line with a 101 response code as per RFC 2616
1524	           [RFC2616].  Such a response could look like "HTTP/1.1 101
1525	           Switching Protocols"

1527	       2.  An "Upgrade" header with value "websocket" as per RFC 2616
1528	           [RFC2616].

1530	       3.  A "Connection" header with value "Upgrade"

1532	       4.  A "Sec-WebSocket-Accept" header.  The value of this header is
1533	           constructed by concatenating /key/, defined above in
1534	           Paragraph 2 of Section 5.2.2, with the string "258EAFA5-E914-
1535	           47DA-95CA-C5AB0DC85B11", taking the SHA-1 hash of this
1536	           concatenated value to obtain a 20-byte value, and base64-
1537	           encoding this 20-byte hash.

1539	           The ABNF of this header is defined as follows:

1541	      accept-value     = base64-value
1542	      base64-value     = *base64-data [ base64-padding ]
1543	      base64-data      = 4base64-character
1544	      base64-padding   = (2base64-character "==") / (3base64-character "=")
1545	      base64-character = ALPHA / DIGIT / "+" / "/"

1547	           NOTE: As an example, if the value of the "Sec-WebSocket-Key"
1548	           header in the client's handshake were
1549	           "dGhlIHNhbXBsZSBub25jZQ==", the server would append the
1550	           string "258EAFA5-E914-47DA-95CA-C5AB0DC85B11" to form the
1551	           string "dGhlIHNhbXBsZSBub25jZQ==258EAFA5-E914-47DA-95CA-
1552	           C5AB0DC85B11".  The server would then take the SHA-1 hash of
1553	           this string, giving the value 0xb3 0x7a 0x4f 0x2c 0xc0 0x62
1554	           0x4f 0x16 0x90 0xf6 0x46 0x06 0xcf 0x38 0x59 0x45 0xb2 0xbe
1555	           0xc4 0xea.  This value is then base64-encoded, to give the
1556	           value "s3pPLMBiTxaQ9kYGzzhZRbK+xOo=", which would be returned
1557	           in the "Sec-WebSocket-Accept" header.

1559	       5.  Optionally, a "Sec-WebSocket-Protocol" header, with a value
1560	           /subprotocol/ as defined in Paragraph 2 of Section 5.2.2.

1562	       6.  Optionally, a "Sec-WebSocket-Extensions" header, with a value
1563	           /extensions/ as defined in Paragraph 2 of Section 5.2.2.  If
1564	           multiple extensions are to be used, they must all be listed
1565	           in a single Sec-WebSocket-Extensions header.  This header
1566	           MUST NOT be repeated.

1568	   This completes the server's handshake.  If the server finishes these
1569	   steps without aborting the WebSocket handshake, the server considers
1570	   the WebSocket connection to be established and that the WebSocket
1571	   connection is in the OPEN state.  At this point, the server may begin
1572	   sending (and receiving) data.

1574	6.  Sending and Receiving Data

1576	6.1.  Sending Data

1578	   To _Send a WebSocket Message_ comprising of /data/ over a WebSocket
1579	   connection, an endpoint MUST perform the following steps.

1581	   1.  The endpoint MUST ensure the WebSocket connection is in the OPEN
1582	       state (cf. Section 5.1 and Section 5.2.2.)  If at any point the
1583	       state of the WebSocket connection changes, the endpoint MUST
1584	       abort the following steps.

1586	   2.  An endpoint MUST encapsulate the /data/ in a WebSocket frame as
1587	       defined in Section 4.2.  If the data to be sent is large, or if
1588	       the data is not available in its entirety at the point the
1589	       endpoint wishes to begin sending the data, the endpoint MAY
1590	       alternately encapsulate the data in a series of frames as defined
1591	       in Section 4.4.

1593	   3.  The opcode (frame-opcode) of the first frame containing the data
1594	       MUST be set to the appropriate value from Section 4.2 for data
1595	       that is to be interpreted by the recipient as text or binary
1596	       data.

1598	   4.  The FIN bit (frame-fin) of the last frame containing the data
1599	       MUST be set to 1 as defined in Section 4.2.

1601	   5.  If the data is being sent by the client, the frame(s) MUST be
1602	       masked as defined in Section 4.3.

1604	   6.  If any extensions (Section 9) have been negotiated for the
1605	       WebSocket connection, additional considerations may apply as per
1606	       the definition of those extensions.

1608	   7.  The frame(s) that have been formed MUST be transmitted over the
1609	       underlying network connection.

1611	6.2.  Receiving Data

1613	   To receive WebSocket data, an endpoint listens on the underlying
1614	   network connection.  Incoming data MUST be parsed as WebSocket frames
1615	   as defined in Section 4.2.  If a control frame (Section 4.5) is
1616	   received, the frame MUST be handled as defined by Section 4.5.  Upon
1617	   receiving a data frame (Section 4.6), the endpoint MUST note the
1618	   /type/ of the data as defined by the Opcode (frame-opcode) from
1619	   Section 4.2.  The _Application Data_ from this frame is defined as
1620	   the /data/ of the message.  If the frame comprises an unfragmented
1621	   message (Section 4.4), it is said that _A WebSocket Message Has Been
1622	   Received_ with type /type/ and data /data/.  If the frame is part of
1623	   a fragmented message, the _Application Data_ of the subsequent data
1624	   frames is concatenated to form the /data/.  When the last fragment is
1625	   received as indicated by the FIN bit (frame-fin), it is said that _A
1626	   WebSocket Message Has Been Received_ with data /data/ (comprised of
1627	   the concatenation of the _Application Data_ of the fragments) and
1628	   type /type/ (noted from the first frame of the fragmented message).
1629	   Subsequent data frames MUST be interpreted as belonging to a new
1630	   WebSocket Message.

1632	   Extensions (Section 9) MAY change the semantics of how data is read,
1633	   specifically including what comprises a message boundary.
1634	   Extensions, in addition to adding "Extension data" before the
1635	   "Application data" in a payload, MAY also modify the "Application
1636	   data" (such as by compressing it).

1638	   Data frames received by a server from a client MUST be unmasked as
1639	   described in Section 4.3.

1641	7.  Closing the connection

1643	7.1.  Definitions

1645	7.1.1.  Close the WebSocket Connection

1647	   To _Close the WebSocket Connection_, an endpoint closes the
1648	   underlying TCP connection.  An endpoint SHOULD use a method that
1649	   cleanly closes the TCP connection, as well as the TLS session, if
1650	   applicable, discarding any trailing bytes that may be received.  An
1651	   endpoint MAY close the connection via any means available when
1652	   necessary, such as when under attack.

1654	   The underlying TCP connection, in most normal cases, SHOULD be closed
1655	   first by the server, so that it holds the TIME_WAIT state and not the
1656	   client (as this would prevent it from re-opening the connection for 2
1657	   MSL, while there is no corresponding server impact as a TIME_WAIT
1658	   connection is immediately reopened upon a new SYN with a higher seq
1659	   number).  In abnormal cases (such as not having received a TCP Close
1660	   from the server after a reasonable amount of time) a client MAY
1661	   initiate the TCP Close.  As such, when a server is instructed to
1662	   _Close the WebSocket Connection_ it SHOULD initiate a TCP Close
1663	   immediately, and when a client is instructed to do the same, it
1664	   SHOULD wait for a TCP Close from the server.

1666	   As an example of how to obtain a clean closure in C using Berkeley
1667	   sockets, one would call shutdown() with SHUT_WR on the socket, call
1668	   recv() until obtaining a return value of 0 indicating that the peer
1669	   has also performed an orderly shutdown, and finally calling close()
1670	   on the socket.

1672	7.1.2.  Start the WebSocket Closing Handshake

1674	   To _Start the WebSocket Closing Handshake_ with a status code
1675	   (Section 7.4) /code/ and an optional close reason (Section 7.1.6)
1676	   /reason/, an endpoint MUST send a Close control frame, as described
1677	   in Section 4.5.1 whose status code is set to /code/ and whose close
1678	   reason is set to /reason/.  Once an endpoint has both sent and
1679	   received a Close control frame, that endpoint SHOULD _Close the
1680	   WebSocket Connection_ as defined in Section 7.1.1.

1682	7.1.3.  The WebSocket Closing Handshake is Started

1684	   Upon either sending or receiving a Close control frame, it is said
1685	   that _The WebSocket Closing Handshake is Started_ and that the
1686	   WebSocket connection is in the CLOSING state.

1688	7.1.4.  The WebSocket Connection is Closed

1690	   When the underlying TCP connection is closed, it is said that _The
1691	   WebSocket Connection is Closed_ and that the WebSocket connection is
1692	   in the CLOSED state.  If the tcp connection was closed after the
1693	   WebSocket closing handshake was completed, the WebSocket connection
1694	   is said to have been closed _cleanly_.

1696	   If the WebSocket connection could not be established, it is also said
1697	   that _The WebSocket Connection is Closed_, but not cleanly.

1699	7.1.5.  The WebSocket Connection Close Code

1701	   As defined in Section 4.5.1 and Section 7.4, a Close control frame
1702	   may contain a status code indicating a reason for closure.  A closing
1703	   of the WebSocket connection may be initiated by either endpoint,
1704	   potentially simultaneously. _The WebSocket Connection Close Code_ is
1705	   defined as the status code (Section 7.4) contained in the first Close
1706	   control frame received by the application implementing this protocol.
1707	   If this Close control frame contains no status code, _The WebSocket
1708	   Connection Close Code_ is considered to be 1005.  If _The WebSocket
1709	   Connection is Closed_ and no Close control frame was received by the
1710	   endpoint (such as could occur if the underlying transport connection
1711	   is lost), _The WebSocket Connection Close Code_ is considered to be
1712	   1006.

1714	   NOTE: Two endpoints may not agree on the value of _The WebSocket
1715	   Connection Close Code_.  As an example, if the remote endpoint sent a
1716	   Close frame but the local application has not yet read the data
1717	   containing the Close frame from its socket's receive buffer, and the
1718	   local application independently decided to close the connection and
1719	   send a Close frame, both endpoints will have sent and received a
1720	   Close frame, and will not send further Close frames.  Each endpoint
1721	   will see the Connection Close Code sent by the other end as the
1722	   _WebSocket Connection Close Code_.  As such, it is possible that the
1723	   two endpoints may not agree on the value of _The WebSocket Connection
1724	   Close Code_ in the case that both endpoints _Start the WebSocket
1725	   Closing Handshake_ independently and at roughly the same time.

1727	7.1.6.  The WebSocket Connection Close Reason

1729	   As defined in Section 4.5.1 and Section 7.4, a Close control frame
1730	   may contain a status code indicating a reason for closure, followed
1731	   by UTF-8 encoded data, the interpretation of said data being left to
1732	   the endpoints and not defined by this protocol.  A closing of the
1733	   WebSocket connection may be initiated by either endpoint, potentially
1734	   simultaneously. _The WebSocket Connection Close Reason_ is defined as
1735	   the UTF-8 encoded data following the status code (Section 7.4)
1736	   contained in the first Close control frame received by the
1737	   application implementing this protocol.  If there is no such data in
1738	   the Close control frame, _The WebSocket Connection Close Reason_ is
1739	   the empty string.

1741	   NOTE: Following the same logic as noted in Section 7.1.5, two
1742	   endpoints may not agree on _The WebSocket Connection Close Reason_.

1744	7.1.7.  Fail the WebSocket Connection

1746	   Certain algorithms and specifications require an endpoint to _Fail
1747	   the WebSocket Connection_.  To do so, the client MUST _Close the
1748	   WebSocket Connection_, and MAY report the problem to the user (which
1749	   would be especially useful for developers) in an appropriate manner.

1751	   If _The WebSocket Connection is Established_ prior to the point where
1752	   the endpoint is required to _Fail the WebSocket Connection_, the
1753	   endpoint SHOULD send a Close frame with an appropriate status code
1754	   Section 7.4 before proceeding to _Close the WebSocket Connection_.
1755	   An endpoint MAY omit sending a Close frame if it believes the other
1756	   side is unlikely to be able to receive and process the close frame,
1757	   due to the nature of the error that led to the WebSocket connection
1758	   being failed in the first place.  An endpoint MUST NOT continue to
1759	   attempt to process data (including a responding Close frame) from the
1760	   remote endpoint after being instruted to _Fail the WebSocket
1761	   Connection_.

1763	   Except as indicated above or as specified by the application layer
1764	   (e.g. a script using the WebSocket API), clients SHOULD NOT close the
1765	   connection.

1767	7.2.  Abnormal Closures

1769	7.2.1.  Client-Initiated Closure

1771	   Certain algorithms, namely during the opening handshake, require the
1772	   client to _Fail the WebSocket Connection_.  To do so, the client MUST
1773	   _Fail the WebSocket Connection_ as defined in Section 7.1.7.

1775	   If at any point the underlying transport layer connection is
1776	   unexpectedly lost, the client MUST _Fail the WebSocket Connection_.

1778	   Except as indicated above or as specified by the application layer
1779	   (e.g. a script using the WebSocket API), clients SHOULD NOT close the
1780	   connection.

1782	7.2.2.  Server-initiated closure

1784	   Certain algorithms require or recommend that the server _Abort the
1785	   WebSocket Connection_ during the opening handshake.  To do so, the
1786	   server MUST simply _Close the WebSocket Connection_ (Section 7.1.1).

1788	7.3.  Normal Closure of Connections

1790	   Servers MAY close the WebSocket connection whenever desired.  Clients
1791	   SHOULD NOT close the WebSocket connection arbitrarily.  In either
1792	   case, an endpoint initiates a closure by following the procedures to
1793	   _Start the WebSocket Closing Handshake_ (Section 7.1.2).

1795	7.4.  Status Codes

1797	   When closing an established connection (e.g. when sending a Close
1798	   frame, after the opening handshake has completed), an endpoint MAY
1799	   indicate a reason for closure.  The interpretation of this reason by
1800	   an endpoint, and the action an endpoint should take given this
1801	   reason, are left undefined by this specification.  This specification
1802	   defines a set of pre-defined status codes, and specifies which ranges
1803	   may be used by extensions, frameworks, and end applications.  The
1804	   status code and any associated textual message are optional
1805	   components of a Close frame.

1807	7.4.1.  Defined Status Codes

1809	   Endpoints MAY use the following pre-defined status codes when sending
1810	   a Close frame.

1812	   1000

1814	      1000 indicates a normal closure, meaning whatever purpose the
1815	      connection was established for has been fulfilled.

1817	   1001

1819	      1001 indicates that an endpoint is "going away", such as a server
1820	      going down, or a browser having navigated away from a page.

1822	   1002

1824	      1002 indicates that an endpoint is terminating the connection due
1825	      to a protocol error.

1827	   1003

1829	      1003 indicates that an endpoint is terminating the connection
1830	      because it has received a type of data it cannot accept (e.g. an
1831	      endpoint that understands only text data MAY send this if it
1832	      receives a binary message).

1834	   1004

1836	      1004 indicates that an endpoint is terminating the connection
1837	      because it has received a frame that is too large.

1839	   1005

1841	      1005 is a reserved value and MUST NOT be set as a status code in a
1842	      Close control frame by an endpoint.  It is designated for use in
1843	      applications expecting a status code to indicate that no status
1844	      code was actually present.

1846	   1006

1848	      1006 is a reserved value and MUST NOT be set as a status code in a
1849	      Close control frame by an endpoint.  It is designated for use in
1850	      applications expecting a status code to indicate that the
1851	      connection was closed abnormally, e.g. without sending or
1852	      receiving a Close control frame.

1854	7.4.2.  Reserved Status Code Ranges

1856	   0-999

1858	      Status codes in the range 0-999 are not used.

1860	   1000-1999

1862	      Status codes in the range 1000-1999 are reserved for definition by
1863	      this protocol.

1865	   2000-2999

1867	      Status codes in the range 2000-2999 are reserved for use by
1868	      extensions.

1870	   3000-3999

1872	      Status codes in the range 3000-3999 MAY be used by libraries and
1873	      frameworks.  The interpretation of these codes is undefined by
1874	      this protocol.  End applications MUST NOT use status codes in this
1875	      range.

1877	   4000-4999

1879	      Status codes in the range 4000-4999 MAY be used by application
1880	      code.  The interpretation of these codes is undefined by this
1881	      protocol.

1883	8.  Error Handling

1885	8.1.  Handling Errors in UTF-8 from the Server

1887	   When a client is to interpret a byte stream as UTF-8 but finds that
1888	   the byte stream is not in fact a valid UTF-8 stream, then any bytes
1889	   or sequences of bytes that are not valid UTF-8 sequences MUST be
1890	   interpreted as a U+FFFD REPLACEMENT CHARACTER.

1892	8.2.  Handling Errors in UTF-8 from the Client

1894	   When a server is to interpret a byte stream as UTF-8 but finds that
1895	   the byte stream is not in fact a valid UTF-8 stream, behavior is
1896	   undefined.  A server could close the connection, convert invalid byte
1897	   sequences to U+FFFD REPLACEMENT CHARACTERs, store the data verbatim,
1898	   or perform application-specific processing.  Subprotocols layered on
1899	   the WebSocket protocol might define specific behavior for servers.

1901	9.  Extensions

1903	   WebSocket clients MAY request extensions to this specification, and
1904	   WebSocket servers MAY accept some or all extensions requested by the
1905	   client.  A server MUST NOT respond with any extension not requested
1906	   by the client.  If extension parameters are included in negotiations
1907	   between the client and the server, those parameters MUST be chosen in
1908	   accordance with the specification of the extension to which the
1909	   parameters apply.

1911	9.1.  Negotiating Extensions

1913	   A client requests extensions by including a "Sec-WebSocket-
1914	   Extensions" header, which follows the normal rules for HTTP headers
1915	   (see [RFC2616] section 4.2) and the value of the header is defined by
1916	   the following ABNF.  Note that unlike other section of the document
1917	   this section is using ABNF syntax/rules from [RFC2616].  If a value
1918	   is received by either the client or the server during negotiation
1919	   that does not conform to the ABNF below, the recipient of such
1920	   malformed data MUST immediately _Fail the WebSocket Connection_.

1922	         extension-list = 1#extension
1923	         extension = extension-token *( ";" extension-param )
1924	         extension-token = registered-token / private-use-token
1925	         registered-token = token
1926	         private-use-token = "x-" token
1927	         extension-param = token [ "=" token ]

1929	   Note that like other HTTP headers, this header MAY be split or
1930	   combined across multiple lines.  Ergo, the following are equivalent:

1932	         Sec-WebSocket-Extensions: foo
1933	         Sec-WebSocket-Extensions: bar; baz=2

1935	   is exactly equivalent to

1937	         Sec-WebSocket-Extensions: foo, bar; baz=2

1939	   Any extension-token used MUST either be a registered token
1940	   (registration TBD), or have a prefix of "x-" to indicate a private-
1941	   use token.  The parameters supplied with any given extension MUST be
1942	   defined for that extension.  Note that the client is only offering to
1943	   use any advertised extensions, and MUST NOT use them unless the
1944	   server indicates that it wishes to use the extension.

1946	   Note that the order of extensions is significant.  Any interactions
1947	   between multiple extensions MAY be defined in the documents defining
1948	   the extensions.  In the absence of such definition, the
1949	   interpretation is that the headers listed by the client in its
1950	   request represent a preference of the headers it wishes to use, with
1951	   the first options listed being most preferable.  The extensions
1952	   listed by the server in response represent the extensions actually in
1953	   use for the connection.  Should the extensions modify the data and/or
1954	   framing, the order of operations on the data should be assumed to be
1955	   the same as the order in which the extensions are listed in the
1956	   server's response in the opening handshake.

1958	   For example, if there are two extensions "foo" and "bar", if the
1959	   header |Sec-WebSocket-Extensions| sent by the server has the value
1960	   "foo, bar" then operations on the data will be made as
1961	   bar(foo(data)), be those changes to the data itself (such as
1962	   compression) or changes to the framing thay may "stack".

1964	   Non-normative examples of acceptable extension headers:

1966	      Sec-WebSocket-Extensions: deflate-stream
1967	      Sec-WebSocket-Extensions: mux; max-channels=4; flow-control, deflate-stream
1968	      Sec-WebSocket-Extensions: x-private-extension

1970	   A server accepts one or more extensions by including a |Sec-
1971	   WebSocket-Extensions| header containing one or more extensions which
1972	   were requested by the client.  The interpretation of any extension
1973	   parameters, and what constitutes a valid response by a server to a
1974	   requested set of parameters by a client, will be defined by each such
1975	   extension.

1977	9.2.  Known Extensions

1979	   Extensions provide a mechanism for implementations to opt-in to
1980	   additional protocol features.  This section defines the meaning of
1981	   well-known extensions but implementations MAY use extensions defined
1982	   separately as well.

1984	9.2.1.  Compression

1986	   The registered extension token for this compression extension is
1987	   "deflate-stream".

1989	   The extension does not have any per frame extension data and it does
1990	   not define the use of any WebSocket reserved bits or opcodes.

1992	   Senders using this extension MUST apply [RFC1951] encodings to all
1993	   bytes of the data stream following the opening handshake including
1994	   both data and control frames.  The data stream MAY include multiple
1995	   blocks of both compressed and uncompressed types as defined by
1996	   [RFC1951].

1998	   Senders MUST NOT delay the transmission of any portion of a WebSocket
1999	   frame because the deflate encoding of the frame does not end on a
2000	   byte boundary.  The encodings for adjacent frames MAY appear in the
2001	   same byte if no delay in transmission is occurred by doing so.

2003	   Historically there have been some confusion and interoperability
2004	   problems around the specification of compression algorithms.  In this
2005	   specification "deflate-stream" requires a [RFC1951] deflate encoding.
2006	   It MUST NOT be wrapped in any of the header formats often associated
2007	   with RFC 1951 such as "zlib" [RFC1950].  This requirement is given
2008	   special attention with this note because of confusion in this area,
2009	   the presence of some popular open source libraries that create both
2010	   formats under a single API call with confusing naming conventions,
2011	   and the fact that the popular HTTP [RFC2616] specification defines
2012	   "deflate" compression differently than this specification.

2014	10.  Security Considerations

2016	   While this protocol is intended to be used by scripts in Web pages,
2017	   it can also be used directly by hosts.  Such hosts are acting on
2018	   their own behalf, and can therefore send fake "Origin" fields,
2019	   misleading the server.  Servers should therefore be careful about
2020	   assuming that they are talking directly to scripts from known
2021	   origins, and must consider that they might be accessed in unexpected
2022	   ways.  In particular, a server should not trust that any input is
2023	   valid.

2025	   EXAMPLE: For example, if the server uses input as part of SQL
2026	   queries, all input text should be escaped before being passed to the
2027	   SQL server, lest the server be susceptible to SQL injection.

2029	   Servers that are not intended to process input from any Web page but
2030	   only for certain sites SHOULD verify the "Origin" field is an origin
2031	   they expect, and should only respond with the corresponding "Sec-
2032	   WebSocket-Origin" if it is an accepted origin.  Servers that only
2033	   accept input from one origin can just send back that value in the
2034	   "Sec-WebSocket-Origin" field, without bothering to check the client's
2035	   value.

2037	   If at any time a server is faced with data that it does not
2038	   understand, or that violates some criteria by which the server
2039	   determines safety of input, or when the server sees an opening
2040	   handshake that does not correspond to the values the server is
2041	   expecting (e.g. incorrect path or origin), the server SHOULD just
2042	   disconnect.  It is always safe to disconnect.

2044	   A common class of security problems arise when sending text data
2045	   using using the wrong encoding.  This protocol specifies that
2046	   messages with a Text data type (as opposed to Binary or other types)
2047	   contain UTF-8 encoded data.  Although the length is still indicated
2048	   and applications implementing this protocol should use the length to
2049	   determine where the frame actually ends, sending data in an improper
2050	   encoding may still break assumptions applications built on top of
2051	   this protocol may make, leading from anything to misinterpretation of
2052	   data to loss of data to potential security bugs.

2054	   In addition to endpoints being the target of attacks via WebSockets,
2055	   other parts of web infrastructure, such as proxies, may be the
2056	   subject of an attack.  In particular, an intermediary may interpret a
2057	   WebSocket frame from a client as a request, and a frame from the
2058	   server as a response to that request.  For instance, an attacker
2059	   could get a browser to establish a connection to its server, get the
2060	   browser to send a frame that looks to an intermediary like a GET
2061	   request for a common piece of JavaScript on another domain, and send
2062	   back a frame that is interpreted as a cacheable response to that
2063	   request, thus poisioning the cache for other users.  To prevent this
2064	   attack, frames sent from clients are masked on the wire with a 32-bit
2065	   value, to prevent an attacker from controlling the bits on the wire
2066	   and thus lessen the probability of an attacker being able to
2067	   construct a frame that can be misinterpreted by a proxy as a non-
2068	   WebSocket request.

2070	   As mentioned in Section 8.2, servers must be extremely cautious
2071	   interpreting invalid UTF-8 data from the client.  A naive UTF-8
2072	   parsing implementation can result in buffer overflows in the case of
2073	   invalid input data.

2075	   For connections using TLS (wss: URIs), the amount of benefit provided
2076	   by TLS depends greatly on the strength of the algorithms negotiated
2077	   during the TLS handshake.  To achieve reasonable levels of
2078	   protections, clients should use only Strong TLS algorithms.  "Web
2079	   Security Context: User Interface Guidelines"
2080	   [W3C.REC-wsc-ui-20100812] discusses what constitutes Strong TLS
2081	   algorithms.

2083	11.  IANA Considerations

2085	11.1.  Registration of "ws:" Scheme

2087	   A |ws:| URI identifies a WebSocket server and resource name.

2089	   URI scheme name.
2090	      ws

2092	   Status.
2093	      Permanent.

2095	   URI scheme syntax.
2096	      In ABNF terms using the terminals from the URI specifications:
2097	      [RFC5234] [RFC3986]

2099	           "ws" ":" hier-part [ "?" query ]

2101	      The <path> [RFC3986] and <query> components form the resource name
2102	      sent to the server to identify the kind of service desired.  Other
2103	      components have the meanings described in RFC3986.

2105	   URI scheme semantics.
2106	      The only operation for this scheme is to open a connection using
2107	      the WebSocket protocol.

2109	   Encoding considerations.
2110	      Characters in the host component that are excluded by the syntax
2111	      defined above MUST be converted from Unicode to ASCII by applying
2112	      the IDNA ToASCII algorithm to the Unicode host name, with both the
2113	      AllowUnassigned and UseSTD3ASCIIRules flags set, and using the
2114	      result of this algorithm as the host in the URI.  [RFC3490]

2116	      Characters in other components that are excluded by the syntax
2117	      defined above MUST be converted from Unicode to ASCII by first
2118	      encoding the characters as UTF-8 and then replacing the
2119	      corresponding bytes using their percent-encoded form as defined in
2120	      the URI and IRI specifications.  [RFC3986] [RFC3987]

2122	   Applications/protocols that use this URI scheme name.
2123	      WebSocket protocol.

2125	   Interoperability considerations.
2126	      None.

2128	   Security considerations.
2129	      See "Security considerations" section above.

2131	   Contact.
2132	      HYBI WG <hybi@ietf.org>

2134	   Author/Change controller.
2135	      IETF <iesg@ietf.org>

2137	   References.
2138	      RFC XXXX

2140	11.2.  Registration of "wss:" Scheme

2142	   A |wss:| URI identifies a WebSocket server and resource name, and
2143	   indicates that traffic over that connection is to be protected via
2144	   TLS (including standard benefits of TLS such as confidentiality,
2145	   integrity, and authentication).

2147	   URI scheme name.
2148	      wss

2150	   Status.
2151	      Permanent.

2153	   URI scheme syntax.
2154	      In ABNF terms using the terminals from the URI specifications:
2155	      [RFC5234] [RFC3986]

2157	           "wss" ":" hier-part [ "?" query ]

2159	      The <path> and <query> components form the resource name sent to
2160	      the server to identify the kind of service desired.  Other
2161	      components have the meanings described in RFC3986.

2163	   URI scheme semantics.
2164	      The only operation for this scheme is to open a connection using
2165	      the WebSocket protocol, encrypted using TLS.

2167	   Encoding considerations.
2168	      Characters in the host component that are excluded by the syntax
2169	      defined above MUST be converted from Unicode to ASCII by applying
2170	      the IDNA ToASCII algorithm to the Unicode host name, with both the
2171	      AllowUnassigned and UseSTD3ASCIIRules flags set, and using the
2172	      result of this algorithm as the host in the URI.  [RFC3490]

2174	      Characters in other components that are excluded by the syntax
2175	      defined above MUST be converted from Unicode to ASCII by first
2176	      encoding the characters as UTF-8 and then replacing the
2177	      corresponding bytes using their percent-encoded form as defined in
2178	      the URI and IRI specification.  [RFC3986] [RFC3987]

2180	   Applications/protocols that use this URI scheme name.
2181	      WebSocket protocol over TLS.

2183	   Interoperability considerations.
2184	      None.

2186	   Security considerations.
2187	      See "Security considerations" section above.

2189	   Contact.
2190	      HYBI WG <hybi@ietf.org>

2192	   Author/Change controller.
2193	      IETF <iesg@ietf.org>

2195	   References.
2196	      RFC XXXX

2198	11.3.  Registration of the "WebSocket" HTTP Upgrade Keyword

2200	   This section defines a keyword for registration in the "HTTP Upgrade
2201	   Tokens" registry as per RFC 2817 [RFC2817].

2203	   Name of token.
2204	      WebSocket

2206	   Author/Change controller.
2207	      IETF <iesg@ietf.org>

2209	   Contact.
2210	      HYBI <hybi@ietf.org>

2212	   References.
2213	      RFC XXXX

2215	11.4.  Sec-WebSocket-Key

2217	   This section describes a header field for registration in the
2218	   Permanent Message Header Field Registry.  [RFC3864]
2219	   Header field name
2220	      Sec-WebSocket-Key

2222	   Applicable protocol
2223	      http

2225	   Status
2226	      standard

2228	   Author/Change controller
2229	      IETF

2231	   Specification document(s)
2232	      RFC XXXX

2234	   Related information
2235	      This header field is only used for WebSocket opening handshake.

2237	   The |Sec-WebSocket-Key| header is used in the WebSocket opening
2238	   handshake.  It is sent from the client to the server to provide part
2239	   of the information used by the server to prove that it received a
2240	   valid WebSocket opening handshake.  This helps ensure that the server
2241	   does not accept connections from non-WebSocket clients (e.g.  HTTP
2242	   clients) that are being abused to send data to unsuspecting WebSocket
2243	   servers.

2245	11.5.  Sec-WebSocket-Extensions

2247	   This section describes a header field for registration in the
2248	   Permanent Message Header Field Registry.  [RFC3864]

2250	   Header field name
2251	      Sec-WebSocket-Extensions

2253	   Applicable protocol
2254	      http

2256	   Status
2257	      standard

2259	   Author/Change controller
2260	      IETF

2262	   Specification document(s)
2263	      RFC XXXX

2265	   Related information
2266	      This header field is only used for WebSocket opening handshake.

2268	   The |Sec-WebSocket-Extensions| header is used in the WebSocket
2269	   opening handshake.  It is initially sent from the client to the
2270	   server, and then subsequently sent from the server to the client, to
2271	   agree on a set of protocol-level extensions to use for the duration
2272	   of the connection.

2274	11.6.  WebSocket Extension Name Registry

2276	   This specification requests the creation of a new IANA registry for
2277	   WebSocket Extension names to be used with the WebSocket protocol in
2278	   accordance with the principles set out in RFC 5226 [RFC5226].

2280	   As part of this registry IANA will maintain the following
2281	   information:

2283	   Extension Identifier
2284	      The identifier of the extension, as will be used in the Sec-
2285	      WebSocket-Extension header registered in Section 11.5 of this
2286	      specification.  The value must conform to the requirements for an
2287	      extension-token as defined in Section 9.1 of this specification.

2289	   Extension Common Name
2290	      The name of the extension, as the extension is generally referred
2291	      to.

2293	   Extension Definition
2294	      A reference to the document in which the extension being used with
2295	      the WebSocket protocol is defined.

2297	   Known Incompatible Extensions
2298	      A list of extension identifiers with which this extension is known
2299	      to be incompatible.

2301	   WebSocket Extension names are to be subject to First Come First Serve
2302	   as per RFC5226 [RFC5226], with the exception of WebSocket Extension
2303	   names whose Extension Identifier matches a private-use-token as
2304	   defined in Section 9.1 (values beginning with "x-").  These Extension
2305	   Identifiers matching private-use-token are reserved for Experimental
2306	   Use as per RFC5226 [RFC5226].

2308	   IANA is asked to add an initial value to the registry (there are no
2309	   known incompatible extensions for this initial entry):

2311	   Extension Identifier |  Extension Common Name |  Extension Definition
2312	 -+---------------------+------------------------+-----------------------
2313	  | deflate-stream      | Deflate Stream         | Section 9.2.1 of this
2314	  |                     | Compression            | document.
2315	 -+---------------------+------------------------+-----------------------

2317	11.7.  Sec-WebSocket-Accept

2319	   This section describes a header field for registration in the
2320	   Permanent Message Header Field Registry.  [RFC3864]

2322	   Header field name
2323	      Sec-WebSocket-Accept

2325	   Applicable protocol
2326	      http

2328	   Status
2329	      standard

2331	   Author/Change controller
2332	      IETF

2334	   Specification document(s)
2335	      RFC XXXX

2337	   Related information
2338	      This header field is only used for WebSocket opening handshake.

2340	   The |Sec-WebSocket-Accept| header is used in the WebSocket opening
2341	   handshake.  It is sent from the server to the client to confirm that
2342	   the server is willing to initiate the connection.

2344	11.8.  Sec-WebSocket-Origin

2346	   This section describes a header field for registration in the
2347	   Permanent Message Header Field Registry.  [RFC3864]

2349	   Header field name
2350	      Sec-WebSocket-Origin

2352	   Applicable protocol
2353	      http

2355	   Status
2356	      standard

2358	   Author/Change controller
2359	      IETF

2361	   Specification document(s)
2362	      RFC XXXX

2364	   Related information
2365	      This header field is only used for WebSocket opening handshake.

2367	   The |Sec-WebSocket-Origin| header is used in the WebSocket opening
2368	   handshake.  It is sent from the server to the client to confirm the
2369	   origin of the script that opened the connection.  This enables
2370	   clients to verify that the server is willing to serve the script that
2371	   opened the connection.

2373	11.9.  Sec-WebSocket-Protocol

2375	   This section describes a header field for registration in the
2376	   Permanent Message Header Field Registry.  [RFC3864]

2378	   Header field name
2379	      Sec-WebSocket-Protocol

2381	   Applicable protocol
2382	      http

2384	   Status
2385	      standard

2387	   Author/Change controller
2388	      IETF

2390	   Specification document(s)
2391	      RFC XXXX

2393	   Related information
2394	      This header field is only used for WebSocket opening handshake.

2396	   The |Sec-WebSocket-Protocol| header is used in the WebSocket opening
2397	   handshake.  It is sent from the client to the server and back from
2398	   the server to the client to confirm the subprotocol of the
2399	   connection.  This enables scripts to both select a subprotocol and be
2400	   sure that the server agreed to serve that subprotocol.

2402	11.10.  WebSocket Subprotocol Name Registry

2404	   This specification requests the creation of a new IANA registry for
2405	   WebSocket Subprotocol names to be used with the WebSocket protocol in
2406	   accordance with the principles set out in RFC 5226 [RFC5226].

2408	   As part of this registry IANA will maintain the following
2409	   information:

2411	   Subprotocol Identifier
2412	      The identifier of the subprotocol, as will be used in the Sec-
2413	      WebSocket-Protocol header registered in Section 11.9 of this
2414	      specification.  The value must conform to the requirements given
2415	      in Paragraph 10 of Paragraph 4 of Section 5.1 of this
2416	      specification, namely the value must be a token as defined by RFC
2417	      2616 [RFC2616].

2419	   Subprotocol Common Name
2420	      The name of the subprotocol, as the subprotocol is generally
2421	      referred to.

2423	   Subprotocol Definition
2424	      A reference to the document in which the subprotocol being used
2425	      with the WebSocket protocol is defined.

2427	   WebSocket Subprotocol names are to be subject to First Come First
2428	   Serve as per RFC5226 [RFC5226].

2430	11.11.  Sec-WebSocket-Version

2432	   This section describes a header field for registration in the
2433	   Permanent Message Header Field Registry.  [RFC3864]

2435	   Header field name
2436	      Sec-WebSocket-Version

2438	   Applicable protocol
2439	      http

2441	   Status
2442	      standard

2444	   Author/Change controller
2445	      IETF

2447	   Specification document(s)
2448	      RFC XXXX

2450	   Related information
2451	      This header field is only used for WebSocket opening handshake.

2453	   The |Sec-WebSocket-Version| header is used in the WebSocket opening
2454	   handshake.  It is sent from the client to the server to indicate the
2455	   protocol version of the connection.  This enables servers to
2456	   correctly interpret the opening handshake and subsequent data being
2457	   sent from the data, and close the connection if the server cannot
2458	   interpret that data in a safe manner.

2460	11.12.  WebSocket Version Number Registry

2462	   This specification requests the creation of a new IANA registry for
2463	   WebSocket Version Numbers to be used with the WebSocket protocol in
2464	   accordance with the principles set out in RFC 5226 [RFC5226].

2466	   As part of this registry IANA will maintain the following
2467	   information:

2469	   Version Number
2470	      The version number to be used in the Sec-WebSocket-Version as
2471	      specified in Section 5.1 of this specification.  The value must be
2472	      a whole number (zero or higher).

2474	   Reference
2475	      The RFC requesting a new version number.

2477	   WebSocket Version Numbers are to be subject to RFC Required as per
2478	   RFC5226 [RFC5226].

2480	   IANA is asked to add initial values to the registry, with suggested
2481	   numerical values as these have been used in past versions of this
2482	   protocol.

2484	      Version Number  | Reference
2485	    -+----------------+-----------------------------------------+-
2486	     | 0              + draft-ietf-hybi-thewebsocketprotocol-00 |
2487	    -+----------------+-----------------------------------------+-
2488	     | 1              + draft-ietf-hybi-thewebsocketprotocol-01 |
2489	    -+----------------+-----------------------------------------+-
2490	     | 2              + draft-ietf-hybi-thewebsocketprotocol-02 |
2491	    -+----------------+-----------------------------------------+-
2492	     | 3              + draft-ietf-hybi-thewebsocketprotocol-03 |
2493	    -+----------------+-----------------------------------------+-
2494	     | 4              + draft-ietf-hybi-thewebsocketprotocol-04 |
2495	    -+----------------+-----------------------------------------+-
2496	     | 5              + draft-ietf-hybi-thewebsocketprotocol-05 |
2497	    -+----------------+-----------------------------------------+-
2498	     | 6              + draft-ietf-hybi-thewebsocketprotocol-06 |
2499	    -+----------------+-----------------------------------------+-
2500	     | 7              + draft-ietf-hybi-thewebsocketprotocol-07 |
2501	    -+----------------+-----------------------------------------+-
2502	     | 8              + draft-ietf-hybi-thewebsocketprotocol-08 |
2503	    -+----------------+-----------------------------------------+-
2504	     | 9              + draft-ietf-hybi-thewebsocketprotocol-09 |
2505	    -+----------------+-----------------------------------------+-

2507	11.13.  WebSocket Close Code Number Registry

2509	   This specification requests the creation of a new IANA registry for
2510	   WebSocket Connection Close Code Numbers in accordance with the
2511	   principles set out in RFC 5226 [RFC5226].

2513	   As part of this registry IANA will maintain the following
2514	   information:

2516	   Status Code
2517	      The Status Code which denotes a reson for a WebSocket connection
2518	      closure as per Section 7.4 of this document.  The status code is
2519	      an integer number.

2521	   Meaning
2522	      The meaning of the status code.

2524	   Contact
2525	      A contact for the entity reserving the status code.

2527	   Reference
2528	      The stable document requesting the status codes and defining their
2529	      meaning, required for status codes in the range 1000-2999.

2531	   WebSocket Close Code Numbers are to be subject to different
2532	   registration requirements depending on their range.  Unless otherwise
2533	   specified, requests are subject to Standards Action as per RFC5226
2534	   [RFC5226].  Requests for status codes for use by this protocol or
2535	   subsequent versions are subject to Standards Action and should be
2536	   granted status codes in the range 1000-1999.  Requests for status
2537	   codes for use by extensions are subject to Specification Required and
2538	   should be granted Status Codes in the range 2000-2999.  Requests for
2539	   status codes for use by libraries and frameworks are subject to First
2540	   Come First Served and should be granted in the range 3000-3999.  The
2541	   range of status codes from 4000-4999 is designated for Private Use by
2542	   application code.  Requests should indicate whether they are
2543	   requesting status codes for use by the WebSocket protocol (or a
2544	   future version of the protocol), by extensions, or by libraries and
2545	   frameworks.

2547	   IANA is asked to add initial values to the registry, with suggested
2548	   numerical values as these have been used in past versions of this
2549	   protocol.

2551	     |Status Code | Meaning         | Contact       | Reference |
2552	    -+------------+-----------------+---------------+-----------|
2553	     | 1000       | Normal Closure  | hybi@ietf.org | RFC XXXX  |
2554	    -+------------+-----------------+---------------+-----------|
2555	     | 1001       | Going Away      | hybi@ietf.org | RFC XXXX  |
2556	    -+------------+-----------------+---------------+-----------|
2557	     | 1002       | Protocol error  | hybi@ietf.org | RFC XXXX  |
2558	    -+------------+-----------------+---------------+-----------|
2559	     | 1003       | Unsupported Data| hybi@ietf.org | RFC XXXX  |
2560	    -+------------+-----------------+---------------+-----------|
2561	     | 1004       | Frame Too Large | hybi@ietf.org | RFC XXXX  |
2562	    -+------------+-----------------+---------------+-----------|
2563	     | 1005       | No Status Rcvd  | hybi@ietf.org | RFC XXXX  |
2564	    -+------------+-----------------+---------------+-----------|
2565	     | 1006       | Abnormal Closure| hybi@ietf.org | RFC XXXX  |
2566	    -+------------+-----------------+---------------+-----------|

2568	11.14.  WebSocket Opcode Registry

2570	   This specification requests the creation of a new IANA registry for
2571	   WebSocket Opcodes in accordance with the principles set out in RFC
2572	   5226 [RFC5226].

2574	   As part of this registry IANA will maintain the following
2575	   information:

2577	   Opcode
2578	      The opcode denotes the frame type of the WebSocket frame, as
2579	      defined in Section 4.2.  The status code is an integer number
2580	      between 0 and 15, inclusive.

2582	   Meaning
2583	      The meaning of the opcode code.

2585	   Reference
2586	      The specification requesting the opcode.

2588	   WebSocket Opcode numbers are subject to Standards Action as per
2589	   RFC5226 [RFC5226].

2591	   IANA is asked to add initial values to the registry, with suggested
2592	   numerical values as these have been used in past versions of this
2593	   protocol.

2595	     |Opcode  | Meaning                             | Reference |
2596	    -+--------+-------------------------------------+-----------|
2597	     | 0      | Continuation Frame                  | RFC XXXX  |
2598	    -+--------+-------------------------------------+-----------|
2599	     | 1      | Text Frame                          | RFC XXXX  |
2600	    -+--------+-------------------------------------+-----------|
2601	     | 2      | Binary Frame                        | RFC XXXX  |
2602	    -+--------+-------------------------------------+-----------|
2603	     | 8      | Connection Close Frame              | RFC XXXX  |
2604	    -+--------+-------------------------------------+-----------|
2605	     | 9      | Ping Frame                          | RFC XXXX  |
2606	    -+--------+-------------------------------------+-----------|
2607	     | 10     | Pong Frame                          | RFC XXXX  |
2608	    -+--------+-------------------------------------+-----------|

2610	11.15.  WebSocket Framing Header Bits Registry

2612	   This specification requests the creation of a new IANA registry for
2613	   WebSocket Framing Header Bits in accordance with the principles set
2614	   out in RFC 5226 [RFC5226].  This registry controls assignment of the
2615	   bits marked RSV1, RSV2, and RSV3 in Section 4.2.

2617	   These bits are reserved for future versions or extensions of this
2618	   specification.

2620	   WebSocket Framing Header Bits assignments are subject to Standards
2621	   Action per RFC 5226 [RFC5226].

2623	12.  Using the WebSocket protocol from Other Specifications

2625	   The WebSocket protocol is intended to be used by another
2626	   specification to provide a generic mechanism for dynamic author-
2627	   defined content, e.g. in a specification defining a scripted API.

2629	   Such a specification first needs to _Establish a WebSocket
2630	   Connection_, providing that algorithm with:

2632	   o  The destination, consisting of a /host/ and a /port/.

2634	   o  A /resource name/, which allows for multiple services to be
2635	      identified at one host and port.

2637	   o  A /secure/ flag, which is true if the connection is to be
2638	      encrypted, and false otherwise.

2640	   o  An ASCII serialization of an origin that is being made responsible
2641	      for the connection.  [I-D.ietf-websec-origin]

2643	   o  Optionally a string identifying a protocol that is to be layered
2644	      over the WebSocket connection.

2646	   The /host/, /port/, /resource name/, and /secure/ flag are usually
2647	   obtained from a URI using the steps to parse a WebSocket URI's
2648	   components.  These steps fail if the URI does not specify a
2649	   WebSocket.

2651	   If at any time the connection is to be closed, then the specification
2652	   needs to use the _Close the WebSocket Connection_ algorithm
2653	   (Section 7.1.1).

2655	   Section 7.1.4 defines when _The WebSocket Connection is Closed_.

2657	   While a connection is open, the specification will need to handle the
2658	   cases when _A WebSocket Message Has Been Received_ (Section 6.2).

2660	   To send some data /data/ to an open connection, the specification
2661	   needs to _Send a WebSocket Message_ (Section 6.1).

2663	13.  Acknowledgements

2665	   Special thanks are due to Ian Hickson, who was the original author
2666	   and editor of this protocol.  The initial design of this
2667	   specification benefitted from the participation of many people in the
2668	   WHATWG and WHATWG mailing list.  Contributions to that specification
2669	   are not tracked by section, but a list of all who contributed to that
2670	   specification is given in the WHATWG HTML specification at
2671	   http://whatwg.org/html5.

2673	   Special thanks also to John Tamplin for providing a significant
2674	   amount of text for the Data Framing section of this specification.

2676	   Special thanks also to Adam Barth for providing a significant amount
2677	   of text and background research for the Data Masking section of this
2678	   specification.

2680	14.  References

2682	14.1.  Normative References

2684	   [ANSI.X3-4.1986]
2685	              American National Standards Institute, "Coded Character
2686	              Set - 7-bit American Standard Code for Information
2687	              Interchange", ANSI X3.4, 1986.

2689	   [FIPS.180-2.2002]
2690	              National Institute of Standards and Technology, "Secure
2691	              Hash Standard", FIPS PUB 180-2, August 2002, <http://
2692	              csrc.nist.gov/publications/fips/fips180-2/fips180-2.pdf>.

2694	   [RFC1951]  Deutsch, P., "DEFLATE Compressed Data Format Specification
2695	              version 1.3", RFC 1951, May 1996.

2697	   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
2698	              Requirement Levels", BCP 14, RFC 2119, March 1997.

2700	   [RFC2616]  Fielding, R., Gettys, J., Mogul, J., Frystyk, H.,
2701	              Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext
2702	              Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999.

2704	   [RFC2817]  Khare, R. and S. Lawrence, "Upgrading to TLS Within
2705	              HTTP/1.1", RFC 2817, May 2000.

2707	   [RFC2818]  Rescorla, E., "HTTP Over TLS", RFC 2818, May 2000.

2709	   [RFC3490]  Faltstrom, P., Hoffman, P., and A. Costello,
2710	              "Internationalizing Domain Names in Applications (IDNA)",
2711	              RFC 3490, March 2003.

2713	   [RFC3548]  Josefsson, S., "The Base16, Base32, and Base64 Data
2714	              Encodings", RFC 3548, July 2003.

2716	   [RFC3629]  Yergeau, F., "UTF-8, a transformation format of ISO
2717	              10646", STD 63, RFC 3629, November 2003.

2719	   [RFC3864]  Klyne, G., Nottingham, M., and J. Mogul, "Registration
2720	              Procedures for Message Header Fields", BCP 90, RFC 3864,
2721	              September 2004.

2723	   [RFC3986]  Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform
2724	              Resource Identifier (URI): Generic Syntax", STD 66,
2725	              RFC 3986, January 2005.

2727	   [RFC3987]  Duerst, M. and M. Suignard, "Internationalized Resource
2728	              Identifiers (IRIs)", RFC 3987, January 2005.

2730	   [RFC4086]  Eastlake, D., Schiller, J., and S. Crocker, "Randomness
2731	              Requirements for Security", BCP 106, RFC 4086, June 2005.

2733	   [RFC5246]  Dierks, T. and E. Rescorla, "The Transport Layer Security
2734	              (TLS) Protocol Version 1.2", RFC 5246, August 2008.

2736	   [RFC6066]  Eastlake, D., "Transport Layer Security (TLS) Extensions:
2737	              Extension Definitions", RFC 6066, January 2011.

2739	   [RFC4648]  Josefsson, S., "The Base16, Base32, and Base64 Data
2740	              Encodings", RFC 4648, October 2006.

2742	   [RFC5226]  Narten, T. and H. Alvestrand, "Guidelines for Writing an
2743	              IANA Considerations Section in RFCs", BCP 26, RFC 5226,
2744	              May 2008.

2746	   [RFC5234]  Crocker, D. and P. Overell, "Augmented BNF for Syntax
2747	              Specifications: ABNF", STD 68, RFC 5234, January 2008.

2749	14.2.  Informative References

2751	   [WSAPI]    Hickson, I., "The Web Sockets API", August 2010,
2752	              <http://dev.w3.org/html5/websockets/>.

2754	   [I-D.ietf-httpstate-cookie]
2755	              Barth, A., "HTTP State Management Mechanism",
2756	              draft-ietf-httpstate-cookie-20 (work in progress),
2757	              December 2010.

2759	   [I-D.ietf-websec-origin]
2760	              Barth, A., "The Web Origin Concept",
2761	              draft-ietf-websec-origin-00 (work in progress),
2762	              December 2010.

2764	   [RFC1950]  Deutsch, L. and J-L. Gailly, "ZLIB Compressed Data Format
2765	              Specification version 3.3", RFC 1950, May 1996.

2767	   [RFC5321]  Klensin, J., "Simple Mail Transfer Protocol", RFC 5321,
2768	              October 2008.

2770	   [RFC6202]  Loreto, S., Saint-Andre, P., Salsano, S., and G. Wilkins,
2771	              "Known Issues and Best Practices for the Use of Long
2772	              Polling and Streaming in Bidirectional HTTP", RFC 6202,
2773	              April 2011.

2775	   [W3C.REC-wsc-ui-20100812]
2776	              Saldhana, A. and T. Roessler, "Web Security Context: User
2777	              Interface Guidelines", World Wide Web Consortium
2778	              Recommendation REC-wsc-ui-20100812, August 2010,
2779	              <http://www.w3.org/TR/2010/REC-wsc-ui-20100812>.

2781	Author's Address

2783	   Ian Fette
2784	   Google, Inc.

2786	   Email: ifette+ietf@google.com
2787	   URI:   http://www.ianfette.com/