idnits 2.17.1 draft-ietf-behave-rfc3489bis-18.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 20. -- Found old boilerplate from RFC 3978, Section 5.5, updated by RFC 4748 on line 2328. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 2339. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 2346. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 2352. 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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust 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. (The document does seem to have the reference to RFC 2119 which the ID-Checklist requires). -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (July 28, 2008) is 5748 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) ** Obsolete normative reference: RFC 2818 (Obsoleted by RFC 9110) ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) ** Obsolete normative reference: RFC 2617 (Obsoleted by RFC 7235, RFC 7615, RFC 7616, RFC 7617) ** Obsolete normative reference: RFC 2988 (Obsoleted by RFC 6298) ** Downref: Normative reference to an Informational RFC: RFC 2104 -- Possible downref: Non-RFC (?) normative reference: ref. 'ITU.V42.2002' ** Downref: Normative reference to an Informational RFC: RFC 1321 ** Obsolete normative reference: RFC 4013 (Obsoleted by RFC 7613) -- Obsolete informational reference (is this intentional?): RFC 2616 (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) -- Obsolete informational reference (is this intentional?): RFC 3489 (Obsoleted by RFC 5389) == Outdated reference: A later version (-16) exists of draft-ietf-behave-turn-08 == Outdated reference: A later version (-20) exists of draft-ietf-sip-outbound-15 == Outdated reference: A later version (-08) exists of draft-ietf-behave-nat-behavior-discovery-03 == Outdated reference: A later version (-16) exists of draft-ietf-mmusic-ice-tcp-06 -- Obsolete informational reference (is this intentional?): RFC 5226 (Obsoleted by RFC 8126) Summary: 8 errors (**), 0 flaws (~~), 6 warnings (==), 12 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 BEHAVE Working Group J. Rosenberg 3 Internet-Draft Cisco 4 Obsoletes: 3489 (if approved) R. Mahy 5 Intended status: Standards Track Plantronics 6 Expires: January 29, 2009 P. Matthews 7 Avaya 8 D. Wing 9 Cisco 10 July 28, 2008 12 Session Traversal Utilities for (NAT) (STUN) 13 draft-ietf-behave-rfc3489bis-18 15 Status of this Memo 17 By submitting this Internet-Draft, each author represents that any 18 applicable patent or other IPR claims of which he or she is aware 19 have been or will be disclosed, and any of which he or she becomes 20 aware will be disclosed, in accordance with Section 6 of BCP 79. 22 Internet-Drafts are working documents of the Internet Engineering 23 Task Force (IETF), its areas, and its working groups. Note that 24 other groups may also distribute working documents as Internet- 25 Drafts. 27 Internet-Drafts are draft documents valid for a maximum of six months 28 and may be updated, replaced, or obsoleted by other documents at any 29 time. It is inappropriate to use Internet-Drafts as reference 30 material or to cite them other than as "work in progress." 32 The list of current Internet-Drafts can be accessed at 33 http://www.ietf.org/ietf/1id-abstracts.txt. 35 The list of Internet-Draft Shadow Directories can be accessed at 36 http://www.ietf.org/shadow.html. 38 This Internet-Draft will expire on January 29, 2009. 40 Copyright Notice 42 Copyright (C) The IETF Trust (2008). 44 Abstract 46 Session Traversal Utilities for NAT (STUN) is a protocol that serves 47 as a tool for other protocols in dealing with NAT traversal. It can 48 be used by an endpoint to determine the IP address and port allocated 49 to it by a NAT. It can also be used to check connectivity between 50 two endpoints, and as a keep-alive protocol to maintain NAT bindings. 51 STUN works with many existing NATs, and does not require any special 52 behavior from them. 54 STUN is not a NAT traversal solution by itself. Rather, it is a tool 55 to be used in the context of a NAT traversal solution. This is an 56 important change from the previous version of this specification (RFC 57 3489), which presented STUN as a complete solution. 59 This document obsoletes RFC 3489. 61 Table of Contents 63 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 64 2. Evolution from RFC 3489 . . . . . . . . . . . . . . . . . . . 4 65 3. Overview of Operation . . . . . . . . . . . . . . . . . . . . 5 66 4. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 8 67 5. Definitions . . . . . . . . . . . . . . . . . . . . . . . . . 8 68 6. STUN Message Structure . . . . . . . . . . . . . . . . . . . . 10 69 7. Base Protocol Procedures . . . . . . . . . . . . . . . . . . . 12 70 7.1. Forming a Request or an Indication . . . . . . . . . . . 12 71 7.2. Sending the Request or Indication . . . . . . . . . . . . 13 72 7.2.1. Sending over UDP . . . . . . . . . . . . . . . . . . . 13 73 7.2.2. Sending over TCP or TLS-over-TCP . . . . . . . . . . . 14 74 7.3. Receiving a STUN Message . . . . . . . . . . . . . . . . 16 75 7.3.1. Processing a Request . . . . . . . . . . . . . . . . . 17 76 7.3.1.1. Forming a Success or Error Response . . . . . . . 18 77 7.3.1.2. Sending the Success or Error Response . . . . . . 19 78 7.3.2. Processing an Indication . . . . . . . . . . . . . . . 19 79 7.3.3. Processing a Success Response . . . . . . . . . . . . 19 80 7.3.4. Processing an Error Response . . . . . . . . . . . . . 20 81 8. FINGERPRINT Mechanism . . . . . . . . . . . . . . . . . . . . 20 82 9. DNS Discovery of a Server . . . . . . . . . . . . . . . . . . 21 83 10. Authentication and Message-Integrity Mechanisms . . . . . . . 22 84 10.1. Short-Term Credential Mechanism . . . . . . . . . . . . . 22 85 10.1.1. Forming a Request or Indication . . . . . . . . . . . 23 86 10.1.2. Receiving a Request or Indication . . . . . . . . . . 23 87 10.1.3. Receiving a Response . . . . . . . . . . . . . . . . . 24 88 10.2. Long-term Credential Mechanism . . . . . . . . . . . . . 24 89 10.2.1. Forming a Request . . . . . . . . . . . . . . . . . . 25 90 10.2.1.1. First Request . . . . . . . . . . . . . . . . . . 25 91 10.2.1.2. Subsequent Requests . . . . . . . . . . . . . . . 26 92 10.2.2. Receiving a Request . . . . . . . . . . . . . . . . . 26 93 10.2.3. Receiving a Response . . . . . . . . . . . . . . . . . 27 94 11. ALTERNATE-SERVER Mechanism . . . . . . . . . . . . . . . . . . 27 95 12. Backwards Compatibility with RFC 3489 . . . . . . . . . . . . 28 96 12.1. Changes to Client Processing . . . . . . . . . . . . . . 29 97 12.2. Changes to Server Processing . . . . . . . . . . . . . . 29 98 13. Basic Server Behavior . . . . . . . . . . . . . . . . . . . . 30 99 14. STUN Usages . . . . . . . . . . . . . . . . . . . . . . . . . 30 100 15. STUN Attributes . . . . . . . . . . . . . . . . . . . . . . . 31 101 15.1. MAPPED-ADDRESS . . . . . . . . . . . . . . . . . . . . . 32 102 15.2. XOR-MAPPED-ADDRESS . . . . . . . . . . . . . . . . . . . 33 103 15.3. USERNAME . . . . . . . . . . . . . . . . . . . . . . . . 34 104 15.4. MESSAGE-INTEGRITY . . . . . . . . . . . . . . . . . . . . 34 105 15.5. FINGERPRINT . . . . . . . . . . . . . . . . . . . . . . . 35 106 15.6. ERROR-CODE . . . . . . . . . . . . . . . . . . . . . . . 36 107 15.7. REALM . . . . . . . . . . . . . . . . . . . . . . . . . . 38 108 15.8. NONCE . . . . . . . . . . . . . . . . . . . . . . . . . . 38 109 15.9. UNKNOWN-ATTRIBUTES . . . . . . . . . . . . . . . . . . . 38 110 15.10. SOFTWARE . . . . . . . . . . . . . . . . . . . . . . . . 39 111 15.11. ALTERNATE-SERVER . . . . . . . . . . . . . . . . . . . . 39 112 16. Security Considerations . . . . . . . . . . . . . . . . . . . 39 113 16.1. Attacks against the Protocol . . . . . . . . . . . . . . 39 114 16.1.1. Outside Attacks . . . . . . . . . . . . . . . . . . . 39 115 16.1.2. Inside Attacks . . . . . . . . . . . . . . . . . . . . 40 116 16.2. Attacks Affecting the Usage . . . . . . . . . . . . . . . 40 117 16.2.1. Attack I: DDoS Against a Target . . . . . . . . . . . 41 118 16.2.2. Attack II: Silencing a Client . . . . . . . . . . . . 41 119 16.2.3. Attack III: Assuming the Identity of a Client . . . . 42 120 16.2.4. Attack IV: Eavesdropping . . . . . . . . . . . . . . . 42 121 16.3. Hash Agility Plan . . . . . . . . . . . . . . . . . . . . 42 122 17. IAB Considerations . . . . . . . . . . . . . . . . . . . . . . 42 123 18. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 43 124 18.1. STUN Methods Registry . . . . . . . . . . . . . . . . . . 43 125 18.2. STUN Attribute Registry . . . . . . . . . . . . . . . . . 43 126 18.3. STUN Error Code Registry . . . . . . . . . . . . . . . . 44 127 18.4. STUN UDP and TCP Port Numbers . . . . . . . . . . . . . . 45 128 19. Changes Since RFC 3489 . . . . . . . . . . . . . . . . . . . . 45 129 20. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 47 130 21. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 47 131 22. References . . . . . . . . . . . . . . . . . . . . . . . . . . 47 132 22.1. Normative References . . . . . . . . . . . . . . . . . . 47 133 22.2. Informational References . . . . . . . . . . . . . . . . 48 134 Appendix A. C Snippet to Determine STUN Message Types . . . . . . 49 135 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 49 136 Intellectual Property and Copyright Statements . . . . . . . . . . 51 138 1. Introduction 140 The protocol defined in this specification, Session Traversal 141 Utilities for NAT, provides a tool for dealing with NATs. It 142 provides a means for an endpoint to determine the IP address and port 143 allocated by a NAT that corresponds to its private IP address and 144 port. It also provides a way for an endpoint to keep a NAT binding 145 alive. With some extensions, the protocol can be used to do 146 connectivity checks between two endpoints [I-D.ietf-mmusic-ice], or 147 to relay packets between two endpoints [I-D.ietf-behave-turn]. 149 In keeping with its tool nature, this specification defines an 150 extensible packet format, defines operation over several transport 151 protocols, and provides for two forms of authentication. 153 STUN is intended to be used in context of one or more NAT traversal 154 solutions. These solutions are known as STUN usages. Each usage 155 describes how STUN is utilized to achieve the NAT traversal solution. 156 Typically, a usage indicates when STUN messages get sent, which 157 optional attributes to include, what server is used, and what 158 authentication mechanism is to be used. Interactive Connectivity 159 Establishment (ICE) [I-D.ietf-mmusic-ice] is one usage of STUN. SIP 160 Outbound [I-D.ietf-sip-outbound] is another usage of STUN. In some 161 cases, a usage will require extensions to STUN. A STUN extension can 162 be in the form of new methods, attributes, or error response codes. 163 More information on STUN usages can be found in Section 14. 165 2. Evolution from RFC 3489 167 STUN was originally defined in RFC 3489 [RFC3489]. That 168 specification, sometimes referred to as "classic STUN", represented 169 itself as a complete solution to the NAT traversal problem. In that 170 solution, a client would discover whether it was behind a NAT, 171 determine its NAT type, discover its IP address and port on the 172 public side of the outermost NAT, and then utilize that IP address 173 and port within the body of protocols, such as the Session Initiation 174 Protocol (SIP) [RFC3261]. However, experience since the publication 175 of RFC 3489 has found that classic STUN simply does not work 176 sufficiently well to be a deployable solution. The address and port 177 learned through classic STUN are sometimes usable for communications 178 with a peer, and sometimes not. Classic STUN provided no way to 179 discover whether it would, in fact, work or not, and it provided no 180 remedy in cases where it did not. Furthermore, classic STUN's 181 algorithm for classification of NAT types was found to be faulty, as 182 many NATs did not fit cleanly into the types defined there. 184 Classic STUN also had a security vulnerability - attackers could 185 provide the client with incorrect mapped addresses under certain 186 topologies and constraints, and this was fundamentally not solvable 187 through any cryptographic means. Though this problem remains with 188 this specification, those attacks are now mitigated through the use 189 of more complete solutions that make use of STUN. 191 For these reasons, this specification obsoletes RFC 3489, and instead 192 describes STUN as a tool that is utilized as part of a complete NAT 193 traversal solution. ICE [I-D.ietf-mmusic-ice] is a complete NAT 194 traversal solution for protocols based on the offer/answer [RFC3264] 195 methodology, such as SIP. SIP Outbound [I-D.ietf-sip-outbound] is a 196 complete solution for traversal of SIP signaling, and it uses STUN in 197 a very different way. Though it is possible that a protocol may be 198 able to use STUN by itself (classic STUN) as a traversal solution, 199 such usage is not described here and is strongly discouraged for the 200 reasons described above. 202 The on-the-wire protocol described here is changed only slightly from 203 classic STUN. The protocol now runs over TCP in addition to UDP. 204 Extensibility was added to the protocol in a more structured way. A 205 magic-cookie mechanism for demultiplexing STUN with application 206 protocols was added by stealing 32 bits from the 128 bit transaction 207 ID defined in RFC 3489, allowing the change to be backwards 208 compatible. Mapped addresses are encoded using a new exclusive-or 209 format. There are other, more minor changes. See Section 19 for a 210 more complete listing. 212 Due to the change in scope, STUN has also been renamed from "Simple 213 Traversal of UDP Through NAT" to "Session Traversal Utilities for 214 NAT". The acronym remains STUN, which is all anyone ever remembers 215 anyway. 217 3. Overview of Operation 219 This section is descriptive only. 221 /-----\ 222 // STUN \\ 223 | Server | 224 \\ // 225 \-----/ 227 +--------------+ Public Internet 228 ................| NAT 2 |....................... 229 +--------------+ 231 +--------------+ Private NET 2 232 ................| NAT 1 |....................... 233 +--------------+ 235 /-----\ 236 // STUN \\ 237 | Client | 238 \\ // Private NET 1 239 \-----/ 241 Figure 1: One possible STUN Configuration 243 One possible STUN configuration is shown in Figure 1. In this 244 configuration, there are two entities (called STUN agents) that 245 implement the STUN protocol. The lower agent in the figure is the 246 client, and is connected to private network 1. This network connects 247 to private network 2 through NAT 1. Private network 2 connects to 248 the public Internet through NAT 2. The upper agent in the figure is 249 the server, and resides on the public Internet. 251 STUN is a client-server protocol. It supports two types of 252 transactions. One is a request/response transaction in which a 253 client sends a request to a server, and the server returns a 254 response. The second is an indication transaction in which either 255 agent - client or server - sends an indication which generates no 256 response. Both types of transactions include a transaction ID, which 257 is a randomly selected 96-bit number. For request/response 258 transactions, this transaction ID allows the client to associate the 259 response with the request that generated it; for indications, this 260 simply serves as a debugging aid. 262 All STUN messages start with a fixed header that includes a method, a 263 class, and the transaction ID. The method indicates which of the 264 various requests or indications this is; this specification defines 265 just one method, Binding, but other methods are expected to be 266 defined in other documents. The class indicates whether this is a 267 request, a success response, an error response, or an indication. 268 Following the fixed header comes zero or more attributes, which are 269 type-length-value extensions that convey additional information for 270 the specific message. 272 This document defines a single method called Binding. The Binding 273 method can be used either in request/response transactions or in 274 indication transactions. When used in request/response transactions, 275 the Binding method can be used to determine the particular "binding" 276 a NAT has allocated to a STUN client. When used in either request/ 277 response or in indication transactions, the Binding method can also 278 be used to keep these "bindings" alive. 280 In the Binding request/response transaction, a Binding Request is 281 sent from a STUN client to a STUN server. When the Binding Request 282 arrives at the STUN server, it may have passed through one or more 283 NATs between the STUN client and the STUN server (in Figure 1, there 284 were two such NATs). As the Binding Request message passes through a 285 NAT, the NAT will modify the source transport address (that is, the 286 source IP address and the source port) of the packet. As a result, 287 the source transport address of the request received by the server 288 will be the public IP address and port created by the NAT closest to 289 the server. This is called a reflexive transport address. The STUN 290 server copies that source transport address into an XOR-MAPPED- 291 ADDRESS attribute in the STUN Binding Response and sends the Binding 292 Response back to the STUN client. As this packet passes back through 293 a NAT, the NAT will modify the destination transport address in the 294 IP header, but the transport address in the XOR-MAPPED-ADDRESS 295 attribute within the body of the STUN response will remain untouched. 296 In this way, the client can learn its reflexive transport address 297 allocated by the outermost NAT with respect to the STUN server. 299 In some usages, STUN must be multiplexed with other protocols (e.g., 300 [I-D.ietf-mmusic-ice], [I-D.ietf-sip-outbound]). In these usages, 301 there must be a way to inspect a packet and determine if it is a STUN 302 packet or not. STUN provides three fields in the STUN header with 303 fixed values that can be used for this purpose. If this is not 304 sufficient, then STUN packets can also contain a FINGERPRINT value 305 which can further be used to distinguish the packets. 307 STUN defines a set of optional procedures that a usage can decide to 308 use, called mechanisms. These mechanisms include DNS discovery, a 309 redirection technique to an alternate server, a fingerprint attribute 310 for demultiplexing, and two authentication and message integrity 311 exchanges. The authentication mechanisms revolve around the use of a 312 username, password, and message-integrity value. Two authentication 313 mechanisms, the long-term credential mechanism and the short-term 314 credential mechanism, are defined in this specification. Each usage 315 specifies the mechanisms allowed with that usage. 317 In the long-term credential mechanism, the client and server share a 318 pre-provisioned username and password and perform a digest challenge/ 319 response exchange inspired by (but differing in details) to the one 320 defined for HTTP [RFC2617]. In the short-term credential mechanism, 321 the client and the server exchange a username and password through 322 some out-of-band method prior to the STUN exchange. For example, in 323 the ICE usage [I-D.ietf-mmusic-ice] the two endpoints use out-of-band 324 signaling to exchange a username and password. These are used to 325 integrity protect and authenticate the request and response. There 326 is no challenge or nonce used. 328 4. Terminology 330 In this document, the key words "MUST", "MUST NOT", "REQUIRED", 331 "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", 332 and "OPTIONAL" are to be interpreted as described in BCP 14, RFC 2119 333 [RFC2119] and indicate requirement levels for compliant STUN 334 implementations. 336 5. Definitions 338 STUN Agent: An entity that implements the STUN protocol. The entity 339 can either be a STUN client or a STUN server. 341 STUN Client: A STUN client is an entity that sends STUN requests, 342 and receives STUN responses. STUN clients can also send 343 indications. In this specification, the terms STUN client and 344 client are synonymous. 346 STUN Server: A STUN server is an entity that receives STUN requests 347 and sends STUN responses. A STUN server can also send 348 indications. In this specification, the terms STUN server and 349 server are synonymous. 351 Transport Address: The combination of an IP address and port number 352 (such as a UDP or TCP port number). 354 Reflexive Transport Address: A transport address learned by a client 355 that identifies that client as seen by another host on an IP 356 network, typically a STUN server. When there is an intervening 357 NAT between the client and the other host, the reflexive transport 358 address represents the mapped address allocated to the client on 359 the public side of the NAT. Reflexive transport addresses are 360 learned from the mapped address attribute (MAPPED-ADDRESS or XOR- 361 MAPPED-ADDRESS) in STUN responses. 363 Mapped Address: Same meaning as Reflexive Address. This term is 364 retained only for for historic reasons and due to the naming of 365 the MAPPED-ADDRESS and XOR-MAPPED-ADDRESS attributes. 367 Long Term Credential: A username and associated password that 368 represent a shared secret between client and server. Long term 369 credentials are generally granted to the client when a subscriber 370 enrolls in a service and persist until the subscriber leaves the 371 service or explicitly changes the credential. 373 Long Term Password: The password from a long term credential. 375 Short Term Credential: A temporary username and associated password 376 which represent a shared secret between client and server. Short 377 term credentials are obtained through some kind of protocol 378 mechanism between the client and server, preceding the STUN 379 exchange. A short term credential has an explicit temporal scope, 380 which may be based on a specific amount of time (such as 5 381 minutes) or on an event (such as termination of a SIP dialog). 382 The specific scope of a short term credential is defined by the 383 application usage. 385 Short Term Password: The password component of a short term 386 credential. 388 STUN Indication: A STUN message that does not receive a response 390 Attribute: The STUN term for a Type-Length-Value (TLV) object that 391 can be added to a STUN message. Attributes are divided into two 392 types: comprehension-required and comprehension-optional. STUN 393 agents can safely ignore comprehension-optional attributes they 394 don't understand, but cannot successfully process a message if it 395 contains comprehension-required attributes that are not 396 understood. 398 RTO: Retransmission TimeOut, which defines the initial period of 399 time between transmission of a request and the first retransmit of 400 that request. 402 6. STUN Message Structure 404 STUN messages are encoded in binary using network-oriented format 405 (most significant byte or octet first, also commonly known as big- 406 endian). The transmission order is described in detail in Appendix B 407 of RFC791 [RFC0791]. Unless otherwise noted, numeric constants are 408 in decimal (base 10). 410 All STUN messages MUST start with a 20-byte header followed by zero 411 or more Attributes. The STUN header contains a STUN message type, 412 magic cookie, transaction ID, and message length. 414 0 1 2 3 415 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 1 416 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 417 |0 0| STUN Message Type | Message Length | 418 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 419 | Magic Cookie | 420 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 421 | | 422 | Transaction ID (96 bits) | 423 | | 424 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 426 Figure 2: Format of STUN Message Header 428 The most significant two bits of every STUN message MUST be zeroes. 429 This can be used to differentiate STUN packets from other protocols 430 when STUN is multiplexed with other protocols on the same port. 432 The message type defines the message class (request, success 433 response, failure response, or indication) and the message method 434 (the primary function) of the STUN message. Although there are four 435 message classes, there are only two types of transactions in STUN: 436 request/response transactions (which consist of a request message and 437 a response message), and indication transactions (which consists of a 438 single indication message). Response classes are split into error 439 and success responses to aid in quickly processing the STUN message. 441 The message type field is decomposed further into the following 442 structure: 444 0 1 445 2 3 4 5 6 7 8 9 0 1 2 3 4 5 447 +--+--+-+-+-+-+-+-+-+-+-+-+-+-+ 448 |M |M |M|M|M|C|M|M|M|C|M|M|M|M| 449 |11|10|9|8|7|1|6|5|4|0|3|2|1|0| 450 +--+--+-+-+-+-+-+-+-+-+-+-+-+-+ 452 Figure 3: Format of STUN Message Type Field 454 Here the bits in the message type field are shown as most-significant 455 (M11) through least-significant (M0). M11 through M0 represent a 12- 456 bit encoding of the method. C1 and C0 represent a 2 bit encoding of 457 the class. A class of 0b00 is a Request, a class of 0b01 is an 458 indication, a class of 0b10 is a success response, and a class of 459 0b11 is an error response. This specification defines a single 460 method, Binding. The method and class are orthogonal, so that for 461 each method, a request, success response, error response and 462 indication are defined for that method. 464 For example, a Binding Request has class=0b00 (request) and 465 method=0b000000000001 (Binding), and is encoded into the first 16 466 bits as 0x0001. A Binding response has class=0b10 (success response) 467 and method=0b000000000001, and is encoded into the first 16 bits as 468 0x0101. 470 Note: This unfortunate encoding is due to assignment of values in 471 [RFC3489] which did not consider encoding Indications, Success, 472 and Errors using bit fields. 474 The magic cookie field MUST contain the fixed value 0x2112A442 in 475 network byte order. In RFC 3489 [RFC3489], this field was part of 476 the transaction ID; placing the magic cookie in this location allows 477 a server to detect if the client will understand certain attributes 478 that were added in this revised specification. In addition, it aids 479 in distinguishing STUN packets from packets of other protocols when 480 STUN is multiplexed with those other protocols on the same port. 482 The transaction ID is a 96 bit identifier, used to uniquely identify 483 STUN transactions. For request/response transactions, the 484 transaction ID is chosen by the STUN client for the request and 485 echoed by the server in the response. For indications, it is chosen 486 by the agent sending the indication. It primarily serves to 487 correlate requests with responses, though it also plays a small role 488 in helping to prevent certain types of attacks. The server also uses 489 the transaction ID as a key to identify each transaction uniquely 490 across all clients. As such, the transaction ID MUST be uniformly 491 and randomly chosen from the interval 0 .. 2**96-1, and SHOULD be 492 cryptographically random. Resends of the same request reuse the same 493 transaction ID, but the client MUST choose a new transaction ID for 494 new transactions unless the new request is bit-wise identical to the 495 previous request and sent from the same transport address to the same 496 IP address. Success and error responses MUST carry the same 497 transaction ID as their corresponding request. When an agent is 498 acting as a STUN server and STUN client on the same port, the 499 transaction IDs in requests sent by the agent have no relationship to 500 the transaction IDs in requests received by the agent. 502 The message length MUST contain the size, in bytes, of the message 503 not including the 20 byte STUN header. Since all STUN attributes are 504 padded to a multiple of four bytes, the last two bits of this field 505 are always zero. This provides another way to distinguish STUN 506 packets from packets of other protocols. 508 Following the STUN fixed portion of the header are zero or more 509 attributes. Each attribute is TLV (type-length-value) encoded. The 510 details of the encoding, and of the attributes themselves is given in 511 Section 15. 513 7. Base Protocol Procedures 515 This section defines the base procedures of the STUN protocol. It 516 describes how messages are formed, how they are sent, and how they 517 are processed when they are received. It also defines the detailed 518 processing of the Binding method. Other sections in this document 519 describe optional procedures that a usage may elect to use in certain 520 situations. Other documents may define other extensions to STUN, by 521 adding new methods, new attributes, or new error response codes. 523 7.1. Forming a Request or an Indication 525 When formulating a request or indication message, the agent MUST 526 follow the rules in Section 6 when creating the header. In addition, 527 the message class MUST be either "Request" or "Indication" (as 528 appropriate), and the method must be either Binding or some method 529 defined in another document. 531 The agent then adds any attributes specified by the method or the 532 usage. For example, some usages may specify that the agent use an 533 authentication method (Section 10) or the FINGERPRINT attribute 534 (Section 8). 536 If the agent sending a request, it SHOULD add a SOFTWARE attribute to 537 the request. Agents MAY include a SOFTWARE attribute in indications, 538 depending on the method. Extensions to STUN should discuss whether 539 SOFTWARE is useful in new indications. 541 For the Binding method with no authentication, no attributes are 542 required unless the usage specifies otherwise. 544 All STUN messages sent over UDP SHOULD be less than the path MTU, if 545 known. If the path MTU is unknown, messages SHOULD be the smaller of 546 576 bytes and the first-hop MTU for IPv4 [RFC1122] and 1280 bytes for 547 IPv6 [RFC2460]. This value corresponds to the overall size of the IP 548 packet. Consequently, for IPv4, the actual STUN message would need 549 to be less than 548 bytes (576 minus 20 bytes IP header, minus 8 byte 550 UDP header, assuming no IP options are used). STUN provides no 551 ability to handle the case where the request is under the MTU but the 552 response would be larger than the MTU. It is not envisioned that 553 this limitation will be an issue for STUN. The MTU limitation is a 554 SHOULD, and not a MUST, to account for cases where STUN itself is 555 being used to probe for MTU characteristics 556 [I-D.ietf-behave-nat-behavior-discovery]. Outside of this or similar 557 applications, the MTU constraint MUST be followed. 559 7.2. Sending the Request or Indication 561 The agent then sends the request or indication. This document 562 specifies how to send STUN messages over UDP, TCP, or TLS-over-TCP; 563 other transport protocols may be added in the future. The STUN usage 564 must specify which transport protocol is used, and how the agent 565 determines the IP address and port of the recipient. Section 9 566 describes a DNS-based method of determining the IP address and port 567 of a server which a usage may elect to use. STUN may be used with 568 anycast addresses, but only with UDP and in usages where 569 authentication is not used. 571 At any time, a client MAY have multiple outstanding STUN requests 572 with the same STUN server (that is, multiple transactions in 573 progress, with different transaction ids). Absent other limits to 574 the rate of new transactions (such as those specified by ICE for 575 connectivity checks or when STUN is run over TCP), a client SHOULD 576 space new transactions to a server by RTO and SHOULD limit itself to 577 ten outstanding transactions to the same server. 579 7.2.1. Sending over UDP 581 When running STUN over UDP it is possible that the STUN message might 582 be dropped by the network. Reliability of STUN request/response 583 transactions is accomplished through retransmissions of the request 584 message by the client application itself. STUN indications are not 585 retransmitted; thus indication transactions over UDP are not 586 reliable. 588 A client SHOULD retransmit a STUN request message starting with an 589 interval of RTO ("Retransmission TimeOut"), doubling after each 590 retransmission. The RTO is an estimate of the round-trip-time, and 591 is computed as described in RFC 2988 [RFC2988], with two exceptions. 592 First, the initial value for RTO SHOULD be configurable (rather than 593 the 3s recommended in RFC 2988) and SHOULD be greater than 500ms. 594 The exception cases for this SHOULD are when other mechanisms are 595 used to derive congestion thresholds (such as the ones defined in ICE 596 for fixed rate streams), or when STUN is used in non-Internet 597 environments with known network capacities. In fixed-line access 598 links, a value of 500ms is RECOMMENDED. Secondly, the value of RTO 599 SHOULD NOT be rounded up to the nearest second. Rather, a 1ms 600 accuracy SHOULD be maintained. As with TCP, the usage of Karn's 601 algorithm is RECOMMENDED [KARN87]. When applied to STUN, it means 602 that RTT estimates SHOULD NOT be computed from STUN transactions 603 which result in the retransmission of a request. 605 The value for RTO SHOULD be cached by a client after the completion 606 of the transaction, and used as the starting value for RTO for the 607 next transaction to the same server (based on equality of IP 608 address). The value SHOULD be considered stale and discarded after 609 10 minutes. 611 Retransmissions continue until a response is received, or until a 612 total of Rc requests have been sent. Rc SHOULD be configurable and 613 SHOULD have a default of 7. If, after the last request, a duration 614 equal to Rm times the RTO has passed without a response (providing 615 ample time to get a response if only this final request actually 616 succeeds), the client SHOULD consider the transaction to have failed. 617 Rm SHOULD be configurable and SHOULD have a default of 16. A STUN 618 transaction over UDP is also considered failed if there has been a 619 hard ICMP error [RFC1122]. For example, assuming an RTO of 500ms, 620 requests would be sent at times 0ms, 500ms, 1500ms, 3500ms, 7500ms, 621 15500ms, and 31500ms. If the client has not received a response 622 after 39500ms, the client will consider the transaction to have timed 623 out. 625 7.2.2. Sending over TCP or TLS-over-TCP 627 For TCP and TLS-over-TCP, the client opens a TCP connection to the 628 server. 630 In some usages of STUN, STUN is sent as the only protocol over the 631 TCP connection. In this case, it can be sent without the aid of any 632 additional framing or demultiplexing. In other usages, or with other 633 extensions, it may be multiplexed with other data over a TCP 634 connection. In that case, STUN MUST be run on top of some kind of 635 framing protocol, specified by the usage or extension, which allows 636 for the agent to extract complete STUN messages and complete 637 application layer messages. The STUN service running on the well 638 known port or ports discovered through the the DNS procedures in 639 Section 9 is for STUN alone, and not for STUN multiplexed with other 640 data. Consequently, no framing protocols are used in connections to 641 those servers. When additional framing is utilized, the usage will 642 specify how the client knows to apply it and what port to connect to. 643 For example, in the case of ICE connectivity checks, this information 644 is learned through out-of-band negotiation between client and server. 646 When STUN is run by itself over TLS-over-TCP, the 647 TLS_RSA_WITH_AES_128_CBC_SHA ciphersuite MUST be implemented at a 648 minimum. Implementations MAY also support any other ciphersuite. 649 When it receives the TLS Certificate message, the client SHOULD 650 verify the certificate and inspect the site identified by the 651 certificate. If the certificate is invalid, revoked, or if it does 652 not identify the appropriate party, the client MUST NOT send the STUN 653 message or otherwise proceed with the STUN transaction. The client 654 MUST verify the identity of the server. To do that, it follows the 655 identification procedures defined in Section 3.1 of RFC 2818 656 [RFC2818]. Those procedures assume the client is dereferencing a 657 URI. For purposes of usage with this specification, the client 658 treats the domain name or IP address used in Section 8.1 as the host 659 portion of the URI that has been dereferenced. Alternatively, a 660 client MAY be configured with a set of domains or IP addresses that 661 are trusted; if a certificate is received that identifies one of 662 those domains or IP addresses, the client considers the identity of 663 the server to be verified. 665 When STUN is run multiplexed with other protocols over a TLS-over-TCP 666 connection, the mandatory ciphersuites and TLS handling procedures 667 operate as defined by those protocols. 669 Reliability of STUN over TCP and TLS-over-TCP is handled by TCP 670 itself, and there are no retransmissions at the STUN protocol level. 671 However, for a request/response transaction, if the client has not 672 received a response by Ti seconds after it sent the SYN to establish 673 the connection, it considers the transaction to have timed out. Ti 674 SHOULD be configurable and SHOULD have a default of 39.5s. This 675 value has been chosen to equalize the TCP and UDP timeouts for the 676 default initial RTO. 678 In addition, if the client is unable to establish the TCP connection, 679 or the TCP connection is reset or fails before a response is 680 received, any request/response transaction in progress is considered 681 to have failed 683 The client MAY send multiple transactions over a single TCP (or TLS- 684 over-TCP) connection, and it MAY send another request before 685 receiving a response to the previous. The client SHOULD keep the 686 connection open until it 688 o has no further STUN requests or indications to send over that 689 connection, and; 691 o has no plans to use any resources (such as a mapped address 692 (MAPPED-ADDRESS or XOR-MAPPED-ADDRESS) or relayed address 693 [I-D.ietf-behave-turn]) that were learned though STUN requests 694 sent over that connection, and; 696 o if multiplexing other application protocols over that port, has 697 finished using that other application, and; 699 o if using that learned port with a remote peer, has established 700 communications with that remote peer, as is required by some TCP 701 NAT traversal techniques (e.g., [I-D.ietf-mmusic-ice-tcp]). 703 At the server end, the server SHOULD keep the connection open, and 704 let the client close it, unless the server has determined that the 705 connection has timed out (for example, due to the client 706 disconnecting from the network). Bindings learned by the client will 707 remain valid in intervening NATs only while the connection remains 708 open. Only the client knows how long it needs the binding. The 709 server SHOULD NOT close a connection if a request was received over 710 that connection for which a response was not sent. A server MUST NOT 711 ever open a connection back towards the client in order to send a 712 response. Servers SHOULD follow best practices regarding connection 713 management in cases of overload. 715 7.3. Receiving a STUN Message 717 This section specifies the processing of a STUN message. The 718 processing specified here is for STUN messages as defined in this 719 specification; additional rules for backwards compatibility are 720 defined in in Section 12. Those additional procedures are optional, 721 and usages can elect to utilize them. First, a set of processing 722 operations are applied that are independent of the class. This is 723 followed by class-specific processing, described in the subsections 724 which follow. 726 When a STUN agent receives a STUN message, it first checks that the 727 message obeys the rules of Section 6. It checks that the first two 728 bits are 0, that the magic cookie field has the correct value, that 729 the message length is sensible, and that the method value is a 730 supported method. If the message-class is Success Response or Error 731 Response, the agent checks that the transaction ID matches a 732 transaction that is still in progress. If the FINGERPRINT extension 733 is being used, the agent checks that the FINGERPRINT attribute is 734 present and contains the correct value. If any errors are detected, 735 the message is silently discarded. In the case when STUN is being 736 multiplexed with another protocol, an error may indicate that this is 737 not really a STUN message; in this case, the agent should try to 738 parse the message as a different protocol. 740 The STUN agent then does any checks that are required by a 741 authentication mechanism that the usage has specified (see 742 Section 10. 744 Once the authentication checks are done, the STUN agent checks for 745 unknown attributes and known-but-unexpected attributes in the 746 message. Unknown comprehension-optional attributes MUST be ignored 747 by the agent. Known-but-unexpected attributes SHOULD be ignored by 748 the agent. Unknown comprehension-required attributes cause 749 processing that depends on the message-class and is described below. 751 At this point, further processing depends on the message class of the 752 request. 754 7.3.1. Processing a Request 756 If the request contains one or more unknown comprehension-required 757 attributes, the server replies with an error response with an error 758 code of 420 (Unknown Attribute), and includes an UNKNOWN-ATTRIBUTES 759 attribute in the response that lists the unknown comprehension- 760 required attributes. 762 The server then does any additional checking that the method or the 763 specific usage requires. If all the checks succeed, the server 764 formulates a success response as described below. 766 If the request uses UDP transport and is a retransmission of a 767 request for which the server has already generated a success response 768 within the last 40 seconds, the server MUST retransmit the same 769 success response, and for an error response with the last 40 seconds, 770 the server SHOULD retransmit the same error response. One way for a 771 server to do this is to remember all transaction IDs received over 772 UDP and their corresponding responses in the last 40 seconds. 773 Another way is to reprocess the request and recompute the response. 774 The latter technique MUST only be applied to requests which are 775 idempotent (a request is considered idempotent when the same request 776 can be safely repeated without impacting the overall state of the 777 system) and result in the same success response for the same request. 778 The SHOULD strength requirement for error responses allows servers to 779 avoid storing transaction state for failed non-idempotent requests. 780 The Binding method is considered to be idempotent. Note that, there 781 are certain rare network events that could cause the reflexive 782 transport address value to change, resulting in a different mapped 783 address in different success responses. Extensions to STUN MUST 784 discuss the implications of request retransmissions on servers which 785 do not store transaction state. 787 7.3.1.1. Forming a Success or Error Response 789 When forming the response (success or error), the server follows the 790 rules of section 6. The method of the response is the same as that 791 of the request, and the message class is either "Success Response" or 792 "Error Response". 794 For an error response, the server MUST add an ERROR-CODE attribute 795 containing the error code specified in the processing above. The 796 reason phrase is not fixed, but SHOULD be something suitable for the 797 error code. For certain errors, additional attributes are added to 798 the message. These attributes are spelled out in the description 799 where the error code is specified. For example, for an error code of 800 420 (Unknown Attribute), the server MUST include an UNKNOWN- 801 ATTRIBUTES attribute. Certain authentication errors also cause 802 attributes to be added (see Section 10). Extensions may define other 803 errors and/or additional attributes to add in error cases. 805 If the server authenticated the request using an authentication 806 mechanism, then the server SHOULD add the appropriate authentication 807 attributes to the response (see Section 10). 809 The server also adds any attributes required by the specific method 810 or usage. In addition, the server SHOULD add a SOFTWARE attribute to 811 the message. 813 For the Binding method, no additional checking is required unless the 814 usage specifies otherwise. When forming the success response, the 815 server adds a XOR-MAPPED-ADDRESS attribute to the response, where the 816 contents of the attribute are the source transport address of the 817 request message. For UDP, this is the source IP address and source 818 UDP port of the request message. For TCP and TLS-over-TCP, this is 819 the source IP address and source TCP port of the TCP connection as 820 seen by the server. 822 7.3.1.2. Sending the Success or Error Response 824 The response (success or error) is sent over the same transport as 825 the request was received on. If the request was received over UDP, 826 the destination IP address and port of the response is the source IP 827 address and port of the received request message, and the source IP 828 address and port of the response is equal to the destination IP 829 address and port of the received request message. If the request was 830 received over TCP or TLS-over-TCP, the response is sent back on the 831 same TCP connection as the request was received on. 833 7.3.2. Processing an Indication 835 If the indication contains unknown comprehension-required attributes, 836 the indication is discarded and processing ceases. 838 The agent then does any additional checking that the method or the 839 specific usage requires. If all the checks succeed, the agent then 840 processes the indication. No response is generated for an 841 indication. 843 For the Binding method, no additional checking or processing is 844 required, unless the usage specifies otherwise. The mere receipt of 845 the message by the agent has refreshed the "bindings" in the 846 intervening NATs. 848 Since indications are not re-transmitted over UDP (unlike requests), 849 there is no need to handle re-transmissions of indications at the 850 sending agent. 852 7.3.3. Processing a Success Response 854 If the success response contains unknown comprehension-required 855 attributes, the response is discarded and the transaction is 856 considered to have failed. 858 The client then does any additional checking that the method or the 859 specific usage requires. If all the checks succeed, the client then 860 processes the success response. 862 For the Binding method, the client checks that the XOR-MAPPED-ADDRESS 863 attribute is present in the response. The client checks the address 864 family specified. If it is an unsupported address family, the 865 attribute SHOULD be ignored. If it is an unexpected but supported 866 address family (for example, the Binding transaction was sent over 867 IPv4, but the address family specified is IPv6), then the client MAY 868 accept and use the value. 870 7.3.4. Processing an Error Response 872 If the error response contains unknown comprehension-required 873 attributes, or if the error response does not contain an ERROR-CODE 874 attribute, then the transaction is simply considered to have failed. 876 The client then does any processing specified by the authentication 877 mechanism (see Section 10). This may result in a new transaction 878 attempt. 880 The processing at this point depends on the error-code, the method, 881 and the usage; the following are the default rules: 883 o If the error code is 300 through 399, the client SHOULD consider 884 the transaction as failed unless the ALTERNATE-SERVER extension is 885 being used. See Section 11. 887 o If the error code is 400 through 499, the client declares the 888 transaction failed; in the case of 420 (Unknown Attribute), the 889 response should contain a UNKNOWN-ATTRIBUTES attribute that gives 890 additional information. 892 o If the error code is 500 through 599, the client MAY resend the 893 request; clients that do so MUST limit the number of times they do 894 this. 896 Any other error code causes the client to consider the transaction 897 failed. 899 8. FINGERPRINT Mechanism 901 This section describes an optional mechanism for STUN that aids in 902 distinguishing STUN messages from packets of other protocols when the 903 two are multiplexed on the same transport address. This mechanism is 904 optional, and a STUN usage must describe if and when it is used. The 905 FINGERPRINT mechanism is not backwards compatible with RFC3489, and 906 cannot be used in environments where such compatibility is required. 908 In some usages, STUN messages are multiplexed on the same transport 909 address as other protocols, such as RTP. In order to apply the 910 processing described in Section 7, STUN messages must first be 911 separated from the application packets. Section 6 describes three 912 fixed fields in the STUN header that can be used for this purpose. 913 However, in some cases, these three fixed fields may not be 914 sufficient. 916 When the FINGERPRINT extension is used, an agent includes the 917 FINGERPRINT attribute in messages it sends to another agent. 918 Section 15.5 describes the placement and value of this attribute. 919 When the agent receives what it believes is a STUN message, then, in 920 addition to other basic checks, the agent also checks that the 921 message contains a FINGERPRINT attribute and that the attribute 922 contains the correct value. Section 7.3 describes when in the 923 overall processing of a STUN message the FINGERPRINT check is 924 performed. This additional check helps the agent detect messages of 925 other protocols that might otherwise seem to be STUN messages. 927 9. DNS Discovery of a Server 929 This section describes an optional procedure for STUN that allows a 930 client to use DNS to determine the IP address and port of a server. 931 A STUN usage must describe if and when this extension is used. To 932 use this procedure, the client must know a server's domain name and a 933 service name; the usage must also describe how the client obtains 934 these. Hard-coding the domain-name of the server into software is 935 NOT RECOMMENDED in case the domain name is lost or needs to change 936 for legal or other reasons. 938 When a client wishes to locate a STUN server in the public Internet 939 that accepts Binding Request/Response transactions, the SRV service 940 name is "stun". When it wishes to locate a STUN server which accepts 941 Binding Request/Response transactions over a TLS session, the SRV 942 service name is "stuns". STUN usages MAY define additional DNS SRV 943 service names. 945 The domain name is resolved to a transport address using the SRV 946 procedures specified in [RFC2782]. The DNS SRV service name is the 947 service name provided as input to this procedure. The protocol in 948 the SRV lookup is the transport protocol the client will run STUN 949 over: "udp" for UDP and "tcp" for TCP. Note that only "tcp" is 950 defined with "stuns" at this time. 952 The procedures of RFC 2782 are followed to determine the server to 953 contact. RFC 2782 spells out the details of how a set of SRV records 954 are sorted and then tried. However, RFC2782 only states that the 955 client should "try to connect to the (protocol, address, service)" 956 without giving any details on what happens in the event of failure. 957 When following these procedures, if the STUN transaction times out 958 without receipt of a response, the client SHOULD retry the request to 959 the next server in the ordered defined by RFC 2782. Such a retry is 960 only possible for request/response transmissions, since indication 961 transactions generate no response or timeout. 963 The default port for STUN requests is 3478, for both TCP and UDP. 965 Administrators of STUN servers SHOULD use this port in their SRV 966 records for UDP and TCP. In all cases, the port in DNS MUST reflect 967 the one the server is listening on. The default port for STUN over 968 TLS is XXXX [[NOTE TO RFC EDITOR: Replace with IANA registered port 969 number for stuns]]. Servers can run STUN over TLS on the same port 970 as STUN over TCP if the server software supports determining whether 971 the initial message is a TLS or STUN message. 973 If no SRV records were found, the client performs an A or AAAA record 974 lookup of the domain name. The result will be a list of IP 975 addresses, each of which can be contacted at the default port using 976 UDP or TCP, independent of the STUN usage. For usages that require 977 TLS, the client connects to one of the IP addresses using the default 978 STUN over TLS port. 980 10. Authentication and Message-Integrity Mechanisms 982 This section defines two mechanisms for STUN that a client and server 983 can use to provide authentication and message-integrity; these two 984 mechanisms are known as the short-term credential mechanism and the 985 long-term credential mechanism. These two mechanisms are optional, 986 and each usage must specify if and when these mechanisms are used. 987 Consequently, both clients and servers will know which mechanism (if 988 any) to follow based on knowledge of which usage applies. For 989 example, a STUN server on the public Internet supporting ICE would 990 have no authentication, whereas the STUN server functionality in an 991 agent supporting connectivity checks would utilize short term 992 credentials. An overview of these two mechanisms is given in 993 Section 3. 995 Each mechanism specifies the additional processing required to use 996 that mechanism, extending the processing specified in Section 7. The 997 additional processing occurs in three different places: when forming 998 a message; when receiving a message immediately after the basic 999 checks have been performed; and when doing the detailed processing of 1000 error responses. 1002 10.1. Short-Term Credential Mechanism 1004 The short-term credential mechanism assumes that, prior to the STUN 1005 transaction, the client and server have used some other protocol to 1006 exchange a credential in the form of a username and password. This 1007 credential is time-limited. The time-limit is defined by the usage. 1008 As an example, in the ICE usage [I-D.ietf-mmusic-ice], the two 1009 endpoints use out-of-band signaling to agree on a username and 1010 password, and this username and password is applicable for the 1011 duration of the media session. 1013 This credential is used to form a message integrity check in each 1014 request and in many responses. There is no challenge and response as 1015 in the long term mechanism; consequently, replay is prevented by 1016 virtue of the time-limited nature of the credential. 1018 10.1.1. Forming a Request or Indication 1020 For a request or indication message, the agent MUST include the 1021 USERNAME and MESSAGE-INTEGRITY attributes in the message. The HMAC 1022 for the MESSAGE-INTEGRITY attribute is computed as described in 1023 Section 15.4. Note that the password is never included in the 1024 request or indication. 1026 10.1.2. Receiving a Request or Indication 1028 After the agent has done the basic processing of a message, the agent 1029 performs the checks listed below in order specified: 1031 o If the message does not contain both a MESSAGE-INTEGRITY and a 1032 USERNAME attribute: 1034 * If the message is a request, the server MUST reject the request 1035 with an error response. This response MUST use an error code 1036 of 400 (Bad Request). 1038 * If the message is an indication, the agent MUST silently 1039 discard the indication. 1041 o If the USERNAME does not contain a username value currently valid 1042 within the server: 1044 * If the message is a request, the server MUST reject the request 1045 with an error response. This response MUST use an error code 1046 of 401 (Unauthorized). 1048 * If the message is an indication, the agent MUST silently 1049 discard the indication. 1051 o Using the password associated with the username, compute the value 1052 for the message-integrity as described in Section 15.4. If the 1053 resulting value does not match the contents of the MESSAGE- 1054 INTEGRITY attribute: 1056 * If the message is a request, the server MUST reject the request 1057 with an error response. This response MUST use an error code 1058 of 401 (Unauthorized). 1060 * If the message is an indication, the agent MUST silently 1061 discard the indication. 1063 If these checks pass, the agent continues to process the request or 1064 indication. Any response generated by a server MUST include the 1065 MESSAGE-INTEGRITY attribute, computed using the password utilized to 1066 authenticate the request. The response MUST NOT contain the USERNAME 1067 attribute. 1069 If any of the checks fail, a server MUST NOT include a MESSAGE- 1070 INTEGRITY or USERNAME attribute in the error response. This is 1071 because, in these failure cases, the server cannot determine the 1072 shared secret necessary to compute MESSAGE-INTEGRITY. 1074 10.1.3. Receiving a Response 1076 The client looks for the MESSAGE-INTEGRITY attribute in the response. 1077 If present, the client computes the message integrity over the 1078 response as defined in Section 15.4, using the same password it 1079 utilized for the request. If the resulting value matches the 1080 contents of the MESSAGE-INTEGRITY attribute, the response is 1081 considered authenticated. If the value does not match, or if 1082 MESSAGE-INTEGRITY was absent, the response MUST be discarded, as if 1083 it was never received. This means that retransmits, if applicable, 1084 will continue. 1086 10.2. Long-term Credential Mechanism 1088 The long-term credential mechanism relies on a long term credential, 1089 in the form of a username and password, that are shared between 1090 client and server. The credential is considered long-term since it 1091 is assumed that it is provisioned for a user, and remains in effect 1092 until the user is no longer a subscriber of the system, or is 1093 changed. This is basically a traditional "log-in" username and 1094 password given to users. 1096 Because these usernames and passwords are expected to be valid for 1097 extended periods of time, replay prevention is provided in the form 1098 of a digest challenge. In this mechanism, the client initially sends 1099 a request, without offering any credentials or any integrity checks. 1100 The server rejects this request, providing the user a realm (used to 1101 guide the user or agent in selection of a username and password) and 1102 a nonce. The nonce provides the replay protection. It is a cookie, 1103 selected by the server, and encoded in such a way as to indicate a 1104 duration of validity or client identity from which it is valid. The 1105 client retries the request, this time including its username, the 1106 realm, and echoing the nonce provided by the server. The client also 1107 includes a message-integrity, which provides an HMAC over the entire 1108 request, including the nonce. The server validates the nonce, and 1109 checks the message-integrity. If they match, the request is 1110 authenticated. If the nonce is no longer valid, it is considered 1111 "stale", and the server rejects the request, providing a new nonce. 1113 In subsequent requests to the same server, the client reuses the 1114 nonce, username, realm and password it used previously. In this way, 1115 subsequent requests are not rejected until the nonce becomes invalid 1116 by the server, in which case the rejection provides a new nonce to 1117 the client. 1119 Note that the long-term credential mechanism cannot be used to 1120 protect indications, since indications cannot be challenged. Usages 1121 utilizing indications must either use a short-term credential, or 1122 omit authentication and message integrity for them. 1124 Since the long-term credential mechanism is susceptible to offline 1125 dictionary attacks, deployments SHOULD utilize passwords which are 1126 difficult to guess. In cases where the credentials are not entered 1127 by the user, but are rather placed on a client device during device 1128 provisioning, the password SHOULD have at least 128 bits of 1129 randomness. In cases where the credentials are entered by the user, 1130 they should follow best current practices around password structure. 1132 10.2.1. Forming a Request 1134 There are two cases when forming a request. In the first case, this 1135 is the first request from the client to the server (as identified by 1136 its IP address and port). In the second case, the client is 1137 submitting a subsequent request once a previous request/response 1138 transaction has completed successfully. Forming a request as a 1139 consequence of a 401 or 438 error response is covered in 1140 Section 10.2.3 and is not considered a "subsequent request" and thus 1141 does not utilize the rules described in Section 10.2.1.2. 1143 10.2.1.1. First Request 1145 If the client has not completed a successful request/response 1146 transaction with the server (as identified by hostname, if the DNS 1147 procedures of Section 9 are used, else IP address if not), it SHOULD 1148 omit the USERNAME, MESSAGE-INTEGRITY, REALM, and NONCE attributes. 1149 In other words, the very first request is sent as if there were no 1150 authentication or message integrity applied. The exception to this 1151 rule are requests sent to another server as a consequence of the 1152 ALTERNATE-SERVER mechanism described in Section 11. Those requests 1153 do include the USERNAME, REALM and NONCE from the original request, 1154 along with a newly computed MESSAGE-INTEGRITY based on them. 1156 10.2.1.2. Subsequent Requests 1158 Once a request/response transaction has completed successfully, the 1159 client will have been been presented a realm and nonce by the server, 1160 and selected a username and password with which it authenticated. 1161 The client SHOULD cache the username, password, realm, and nonce for 1162 subsequent communications with the server. When the client sends a 1163 subsequent request, it SHOULD include the USERNAME, REALM, and NONCE 1164 attributes with these cached values. It SHOULD include a MESSAGE- 1165 INTEGRITY attribute, computed as described in Section 15.4 using the 1166 cached password. 1168 10.2.2. Receiving a Request 1170 After the server has done the basic processing of a request, it 1171 performs the checks listed below in the order specified: 1173 o If the message does not contain a MESSAGE-INTEGRITY attribute, the 1174 server MUST generate an error response with an error code of 401 1175 (Unauthorized). This response MUST include a REALM value. It is 1176 RECOMMENDED that the REALM value be the domain name of the 1177 provider of the STUN server. The response MUST include a NONCE, 1178 selected by the server. The response SHOULD NOT contain a 1179 USERNAME or MESSAGE-INTEGRITY attribute. 1181 o If the message contains a MESSAGE-INTEGRITY attribute, but is 1182 missing the USERNAME, REALM or NONCE attributes, the server MUST 1183 generate an error response with an error code of 400 (Bad 1184 Request). This response SHOULD NOT include a USERNAME, NONCE, 1185 REALM or MESSAGE-INTEGRITY attribute. 1187 o If the NONCE is no longer valid, the server MUST generate an error 1188 response with an error code of 438 (Stale Nonce). This response 1189 MUST include a NONCE and REALM attribute and SHOULD NOT incude the 1190 USERNAME or MESSAGE-INTEGRITY attribute. Servers can invalidate 1191 nonces in order to provide additional security. See Section 4.3 1192 of [RFC2617] for guidelines. 1194 o If the username in the USERNAME attribute is not valid, the server 1195 MUST generate an error response with an error code of 401 1196 (Unauthorized). This response MUST include a REALM value. It is 1197 RECOMMENDED that the REALM value be the domain name of the 1198 provider of the STUN server. The response MUST include a NONCE, 1199 selected by the server. The response SHOULD NOT contain a 1200 USERNAME or MESSAGE-INTEGRITY attribute. 1202 o Using the password associated with the username in the USERNAME 1203 attribute, compute the value for the message-integrity as 1204 described in Section 15.4. If the resulting value does not match 1205 the contents of the MESSAGE-INTEGRITY attribute, the server MUST 1206 reject the request with an error response. This response MUST use 1207 an error code of 401 (Unauthorized). It MUST include a REALM and 1208 NONCE attribute and SHOULD NOT include the USERNAME or MESSAGE- 1209 INTEGRITY attribute. 1211 If these checks pass, the server continues to process the request. 1212 Any response generated by the server (excepting the cases described 1213 above) MUST include the MESSAGE-INTEGRITY attribute, computed using 1214 the username and password utilized to authenticate the request. The 1215 REALM, NONCE, and USERNAME attributes SHOULD NOT be included. 1217 10.2.3. Receiving a Response 1219 If the response is an error response, with an error code of 401 1220 (Unauthorized), the client SHOULD retry the request with a new 1221 transaction. This request MUST contain a USERNAME, determined by the 1222 client as the appropriate username for the REALM from the error 1223 response. The request MUST contain the REALM, copied from the error 1224 response. The request MUST contain the NONCE, copied from the error 1225 response. The request MUST contain the MESSAGE-INTEGRITY attribute, 1226 computed using the password associated with the username in the 1227 USERNAME attribute. The client MUST NOT perform this retry if it is 1228 not changing the USERNAME or REALM or its associated password, from 1229 the previous attempt. 1231 If the response is an error response with an error code of 438 (Stale 1232 Nonce), the client MUST retry the request, using the new NONCE 1233 supplied in the 438 (Stale Nonce) response. This retry MUST also 1234 include the USERNAME, REALM and MESSAGE-INTEGRITY. 1236 The client looks for the MESSAGE-INTEGRITY attribute in the response 1237 (either success or failure). If present, the client computes the 1238 message integrity over the response as defined in Section 15.4, using 1239 the same password it utilized for the request. If the resulting 1240 value matches the contents of the MESSAGE-INTEGRITY attribute, the 1241 response is considered authenticated. If the value does not match, 1242 or if MESSAGE-INTEGRITY was absent, the response MUST be discarded, 1243 as if it was never received. This means that retransmits, if 1244 applicable, will continue. 1246 11. ALTERNATE-SERVER Mechanism 1248 This section describes a mechanism in STUN that allows a server to 1249 redirect a client to another server. This extension is optional, and 1250 a usage must define if and when this extension is used. To prevent 1251 denial-of-service attacks, this extension MUST only be used in 1252 situations where the client and server are using an authentication 1253 and message-integrity mechanism. 1255 A server using this extension redirects a client to another server by 1256 replying to a request message with an error response message with an 1257 error code of 300 (Try Alternate). The server MUST include a 1258 ALTERNATE-SERVER attribute in the error response. The error response 1259 message MUST be authenticated, which in practice means the request 1260 message must have passed the authentication checks. 1262 A client using this extension handles a 300 (Try Alternate) error 1263 code as follows. If the error response has passed the authentication 1264 checks, then the client looks for a ALTERNATE-SERVER attribute in the 1265 error response. If one is found, then the client considers the 1266 current transaction as failed, and re-attempts the request with the 1267 server specified in the attribute, using the same transport protocol 1268 used for the previous request. The client SHOULD reuse any 1269 authentication credentials from the old request in the new 1270 transaction. If the client has been redirected to a server on which 1271 it has already tried this request within the last five minutes, it 1272 MUST ignore the redirection and consider the transaction to have 1273 failed. This prevents infinite ping-ponging between servers in case 1274 of redirection loops. 1276 12. Backwards Compatibility with RFC 3489 1278 This section defines procedures that allow a degree of backwards 1279 compatible with the original protocol defined in RFC 3489 [RFC3489]. 1280 This mechanism is optional, meant to be utilized only in cases where 1281 a new client can connect to an old server, or vice-a-versa. A usage 1282 must define if and when this procedure is used. 1284 Section 19 lists all the changes between this specification and RFC 1285 3489 [RFC3489]. However, not all of these differences are important, 1286 because "classic STUN" was only used in a few specific ways. For the 1287 purposes of this extension, the important changes are the following. 1288 In RFC 3489: 1290 o UDP was the only supported transport; 1292 o The field that is now the Magic Cookie field was a part of the 1293 transaction id field, and transaction ids were 128 bits long; 1295 o The XOR-MAPPED-ADDRESS attribute did not exist, and the Binding 1296 method used the MAPPED-ADDRESS attribute instead; 1298 o There were three comprehension-required attributes, RESPONSE- 1299 ADDRESS, CHANGE-REQUEST, and CHANGED-ADDRESS that have been 1300 removed from this specification; 1302 * These attributes are now part of the NAT Behavior Discovery 1303 usage. [I-D.ietf-behave-nat-behavior-discovery] 1305 12.1. Changes to Client Processing 1307 A client that wants to interoperate with a [RFC3489] server SHOULD 1308 send a request message that uses the Binding method, contains no 1309 attributes, and uses UDP as the transport protocol to the server. If 1310 successful, the success response received from the server will 1311 contain a MAPPED-ADDRESS attribute rather than an XOR-MAPPED-ADDRESS 1312 attribute. A client seeking to interoperate with an older server 1313 MUST be prepared to receive either. Furthermore, the client MUST 1314 ignore any Reserved comprehension-required attributes which might 1315 appear in the response. Of the Reserved attributes in in 1316 Section 18.2, 0x0002,0x0004,0x0005 and 0x000B may appear in Binding 1317 Responses from a server compliant to RFC 3489. Other than this 1318 change, the processing of the response is identical to the procedures 1319 described above. 1321 12.2. Changes to Server Processing 1323 A STUN server can detect when a given Binding Request message was 1324 sent from an RFC 3489 [RFC3489] client by the absence of the correct 1325 value in the Magic Cookie field. When the server detects an RFC 3489 1326 client, it SHOULD copy the value seen in the Magic Cookie field in 1327 the Binding Request to the Magic Cookie field in the Binding Response 1328 message, and insert a MAPPED-ADDRESS attribute instead of an XOR- 1329 MAPPED-ADDRESS attribute. 1331 The client might, in rare situations, include either the RESPONSE- 1332 ADDRESS or CHANGE-REQUEST attributes. In these situations, the 1333 server will view these as unknown comprehension-required attributes 1334 and reply with an error response. Since the mechanisms utilizing 1335 those attributes are no longer supported, this behavior is 1336 acceptable. 1338 The RFC 3489 version of STUN lacks both the Magic Cookie and the 1339 FINGERPRINT attribute that allows for a very high probablility of 1340 correctly identifying STUN messages when multiplexed with other 1341 protocols. Therefore, STUN implementations that are backwards 1342 compatible with RFC 3489 SHOULD NOT be used in cases where STUN will 1343 be multiplexed with another protocol. However, that should not be an 1344 issues as such multiplexing was not available in RFC 3489. 1346 13. Basic Server Behavior 1348 This section defines the behavior of a basic, standalone STUN server. 1349 A basic STUN server provides clients with server reflexive transport 1350 addresses by receiving and replying to STUN Binding Requests. 1352 The STUN server MUST support the Binding method. It SHOULD NOT 1353 utilize the short term or long term credential mechanism. This is 1354 because the work involved in authenticating the request is more than 1355 the work in simply processing it. It SHOULD NOT utilize the 1356 ALTERNATE-SERVER mechanism for the same reason. It MUST support UDP 1357 and TCP. It MAY support STUN over TCP/TLS, however TLS provides 1358 minimal security benefits in this basic mode of operation. It MAY 1359 utilize the FINGERPRINT mechanism but MUST NOT require it. Since the 1360 standalove server only runs STUN, FINGERPRINT provides no benefit. 1361 Requiring it would break compatibility with RFC 3489, and such 1362 compatibility is desirable in a standalone server. Standalone STUN 1363 servers SHOULD support backwards compatibility with [RFC3489] 1364 clients, as described in Section 12. 1366 It is RECOMMENDED that administrators of STUN servers provide DNS 1367 entries for those servers as described in Section 9. 1369 A basic STUN server is not a solution for NAT traversal by itself. 1370 However, it can be utilized as part of a solution through STUN 1371 usages. This is discussed further in Section 14. 1373 14. STUN Usages 1375 STUN by itself is not a solution to the NAT traversal problem. 1376 Rather, STUN defines a tool that can be used inside a larger 1377 solution. The term "STUN Usage" is used for any solution that uses 1378 STUN as a component. 1380 At the time of writing, three STUN usages are defined: Interactive 1381 Connectivity Establishment (ICE) [I-D.ietf-mmusic-ice], Client- 1382 initiated connections for SIP [I-D.ietf-sip-outbound], and NAT 1383 Behavior Discovery [I-D.ietf-behave-nat-behavior-discovery]. Other 1384 STUN usages may be defined in the future. 1386 A STUN usage defines how STUN is actually utilized - when to send 1387 requests, what to do with the responses, and which optional 1388 procedures defined here (or in an extension to STUN) are to be used. 1389 A usage would also define: 1391 o Which STUN methods are used; 1392 o What authentication and message integrity mechanisms are used; 1394 o The considerations around manual vs. automatic key derivation for 1395 the integrity mechanism, as discussed in [RFC4107]; 1397 o What mechanisms are used to distinguish STUN messages from other 1398 messages. When STUN is run over TCP, a framing mechanism may be 1399 required; 1401 o How a STUN client determines the IP address and port of the STUN 1402 server; 1404 o Whether backwards compatibility to RFC 3489 is required; 1406 o What optional attributes defined here (such as FINGERPRINT and 1407 ALTERNATE-SERVER) or in other extensions are required. 1409 In addition, any STUN usage must consider the security implications 1410 of using STUN in that usage. A number of attacks against STUN are 1411 known (see the Security Considerations section in this document) and 1412 any usage must consider how these attacks can be thwarted or 1413 mitigated. 1415 Finally, a usage must consider whether its usage of STUN is an 1416 example of the Unilateral Self-Address Fixing approach to NAT 1417 traversal, and if so, address the questions raised in RFC 3424. 1418 [RFC3424] 1420 15. STUN Attributes 1422 After the STUN header are zero or more attributes. Each attribute 1423 MUST be TLV encoded, with a 16 bit type, 16 bit length, and value. 1424 Each STUN attribute MUST end on a 32 bit boundary. As mentioned 1425 above, all fields in an attribute are transmitted most significant 1426 bit first. 1428 0 1 2 3 1429 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 1430 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1431 | Type | Length | 1432 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1433 | Value (variable) .... 1434 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1436 Figure 4: Format of STUN Attributes 1438 The value in the Length field MUST contain the length of the Value 1439 part of the attribute, prior to padding, measured in bytes. Since 1440 STUN aligns attributes on 32 bit boundaries, attributes whose content 1441 is not a multiple of 4 bytes are padded with 1, 2 or 3 bytes of 1442 padding so that its value contains a multiple of 4 bytes. The 1443 padding bits are ignored, and may be any value. 1445 Any attribute type MAY appear more than once in a STUN message. 1446 Unless specified otherwise, the order of appearance is significant: 1447 only the first occurance needs to be processed by a receiver, and any 1448 duplicates MAY be ignored by a receiver. 1450 To allow future revisions of this specification to add new attributes 1451 if needed, the attribute space is divided into two ranges. 1452 Attributes with type values between 0x0000 and 0x7FFF are 1453 comprehension-required attributes, which means that the STUN agent 1454 cannot successfully process the message unless it understands the 1455 attribute. Attributes with type values between 0x8000 and 0xFFFF are 1456 comprehension-optional attributes, which means that those attributes 1457 can be ignored by the STUN agent if it does not understand them. 1459 The set of STUN attribute types is maintained by IANA. The initial 1460 set defined by this specification is found in Section 18.2. 1462 The rest of this section describes the format of the various 1463 attributes defined in this specification. 1465 15.1. MAPPED-ADDRESS 1467 The MAPPED-ADDRESS attribute indicates a reflexive transport address 1468 of the client. It consists of an eight bit address family, and a 1469 sixteen bit port, followed by a fixed length value representing the 1470 IP address. If the address family is IPv4, the address MUST be 32 1471 bits. If the address family is IPv6, the address MUST be 128 bits. 1472 All fields must be in network byte order. 1474 The format of the MAPPED-ADDRESS attribute is: 1476 0 1 2 3 1477 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 1478 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1479 |0 0 0 0 0 0 0 0| Family | Port | 1480 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1481 | | 1482 | Address (32 bits or 128 bits) | 1483 | | 1484 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1486 Figure 5: Format of MAPPED-ADDRESS attribute 1488 The address family can take on the following values: 1490 0x01:IPv4 1491 0x02:IPv6 1493 The first 8 bits of the MAPPED-ADDRESS MUST be set to 0 and MUST be 1494 ignored by receivers. These bits are present for aligning parameters 1495 on natural 32 bit boundaries. 1497 This attribute is used only by servers for achieving backwards 1498 compatibility with RFC 3489 [RFC3489] clients. 1500 15.2. XOR-MAPPED-ADDRESS 1502 The XOR-MAPPED-ADDRESS attribute is identical to the MAPPED-ADDRESS 1503 attribute, except that the reflexive transport address is obfuscated 1504 through the XOR function. 1506 The format of the XOR-MAPPED-ADDRESS is: 1508 0 1 2 3 1509 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 1510 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1511 |x x x x x x x x| Family | X-Port | 1512 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1513 | X-Address (Variable) 1514 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1516 Figure 7: Format of XOR-MAPPED-ADDRESS Attribute 1518 The Family represents the IP address family, and is encoded 1519 identically to the Family in MAPPED-ADDRESS. 1521 X-Port is computed by taking the mapped port in host byte order, 1522 XOR'ing it with the most significant 16 bits of the magic cookie, and 1523 then the converting the result to network byte order. If the IP 1524 address family is IPv4, X-Address is computed by taking the mapped IP 1525 address in host byte order, XOR'ing it with the magic cookie, and 1526 converting the result to network byte order. If the IP address 1527 family is IPv6, X-Address is computed by taking the mapped IP address 1528 in host byte order, XOR'ing it with the concatenation of the magic 1529 cookie and the 96-bit transaction ID, and converting the result to 1530 network byte order. 1532 The rules for encoding and processing the first 8 bits of the 1533 attribute's value, the rules for handling multiple occurrences of the 1534 attribute, and the rules for processing addresses families are the 1535 same as for MAPPED-ADDRESS. 1537 NOTE: XOR-MAPPED-ADDRESS and MAPPED-ADDRESS differ only in their 1538 encoding of the transport address. The former encodes the transport 1539 address by exclusive-or'ing it with the magic cookie. The latter 1540 encodes it directly in binary. RFC 3489 originally specified only 1541 MAPPED-ADDRESS. However, deployment experience found that some NATs 1542 rewrite the 32-bit binary payloads containing the NAT's public IP 1543 address, such as STUN's MAPPED-ADDRESS attribute, in the well-meaning 1544 but misguided attempt at providing a generic ALG function. Such 1545 behavior interferes with the operation of STUN and also causes 1546 failure of STUN's message integrity checking. 1548 15.3. USERNAME 1550 The USERNAME attribute is used for message integrity. It identifies 1551 the username and password combination used in the message integrity 1552 check. 1554 The value of USERNAME is a variable length value. It MUST contain a 1555 UTF-8 [RFC3629] encoded sequence of less than 513 bytes, and MUST 1556 have been processed using SASLPrep [RFC4013]. 1558 15.4. MESSAGE-INTEGRITY 1560 The MESSAGE-INTEGRITY attribute contains an HMAC-SHA1 [RFC2104] of 1561 the STUN message. The MESSAGE-INTEGRITY attribute can be present in 1562 any STUN message type. Since it uses the SHA1 hash, the HMAC will be 1563 20 bytes. The text used as input to HMAC is the STUN message, 1564 including the header, up to and including the attribute preceding the 1565 MESSAGE-INTEGRITY attribute. With the exception of the FINGERPRINT 1566 attribute, which appears after MESSAGE-INTEGRITY, agents MUST ignore 1567 all other attributes that follow MESSAGE-INTEGRITY. 1569 The key for the HMAC depends on whether long term or short term 1570 credentials are in use. For long term credentials, the key is 16 1571 bytes: 1573 key = MD5(username ":" realm ":" SASLPrep(password)) 1575 That is, the 16 byte key is formed by taking the MD5 hash of the 1576 result of concatenating the following five fields: (1) The username, 1577 with any quotes and trailing nulls removed, as taken from the 1578 USERNAME attribute (in which case SASLPrep has already been applied) 1579 (2) A single colon, (3) The realm, with any quotes and trailing nulls 1580 removed, (4) A single colon, and (5) the password, with any trailing 1581 nulls removed and after processing using SASLPrep. For example, if 1582 the username was 'user', the realm was 'realm', and the password was 1583 'pass', then the 16-byte HMAC key would be the result of performing 1584 an MD5 hash on the string 'user:realm:pass', the resulting hash being 1585 0x8493fbc53ba582fb4c044c456bdc40eb. 1587 For short term credentials: 1589 key = SASLPrep(password) 1591 Where MD5 is defined in RFC 1321 [RFC1321] and SASLPrep() is defined 1592 in [RFC4013]. 1594 The structure of the key when used with long term credentials 1595 facilitates deployment in systems that also utilize SIP. Typically, 1596 SIP systems utilizing SIP's digest authentication mechanism do not 1597 actually store the password in the database. Rather, they store a 1598 value called H(A1), which is equal to the key defined above. 1600 Based on the rules above, the hash includes the length field from the 1601 STUN message header. Prior to performing the hash, the MESSAGE- 1602 INTEGRITY attribute MUST be inserted into the message (with dummy 1603 content). The length MUST then be set to point to the length of the 1604 message up to, and including, the MESSAGE-INTEGRITY attribute itself, 1605 but excluding any attributes after it. Once the computation is 1606 performed, the value of the MESSAGE-INTEGRITY attribute can be filled 1607 in, and the value of the length in the STUN header can be set to its 1608 correct value - the length of the entire message. Similarly, when 1609 validating the MESSAGE-INTEGRITY, the length field should be adjusted 1610 to point to the end of the MESSAGE-INTEGRITY attribute prior to 1611 calculating the HMAC. Such adjustment is necessary when attributes, 1612 such as FINGERPRINT, appear after MESSAGE-INTEGRITY. 1614 15.5. FINGERPRINT 1616 The FINGERPRINT attribute MAY be present in all STUN messages. The 1617 value of the attribute is computed as the CRC-32 of the STUN message 1618 up to (but excluding) the FINGERPRINT attribute itself, xor-d with 1619 the 32 bit value 0x5354554e (the XOR helps in cases where an 1620 application packet is also using CRC-32 in it). The 32 bit CRC is 1621 the one defined in ITU V.42 [ITU.V42.2002], which has a generator 1622 polynomial of x32+x26+x23+x22+x16+x12+x11+x10+x8+x7+x5+x4+x2+x+1. 1623 When present, the FINGERPRINT attribute MUST be the last attribute in 1624 the message, and thus will appear after MESSAGE-INTEGRITY. 1626 The FINGERPRINT attribute can aid in distinguishing STUN packets from 1627 packets of other protocols. See Section 8. 1629 As with MESSAGE-INTEGRITY, the CRC used in the FINGERPRINT attribute 1630 covers the length field from the STUN message header. Therefore, 1631 this value must be correct, and include the CRC attribute as part of 1632 the message length, prior to computation of the CRC. When using the 1633 FINGERPRINT attribute in a message, the attribute is first placed 1634 into the message with a dummy value, then the CRC is computed, and 1635 then the value of the attribute is updated. If the MESSAGE-INTEGRITY 1636 attribute is also present, then it must be present with the correct 1637 message-integrity value before the CRC is computed, since the CRC is 1638 done over the value of the MESSAGE-INTEGRITY attribute as well. 1640 15.6. ERROR-CODE 1642 The ERROR-CODE attribute is used in Error Response messages. It 1643 contains a numeric error code value in the range of 300 to 699 plus a 1644 textual reason phrase encoded in UTF-8 [RFC3629], and is consistent 1645 in its code assignments and semantics with SIP [RFC3261] and HTTP 1646 [RFC2616]. The reason phrase is meant for user consumption, and can 1647 be anything appropriate for the error code. Recommended reason 1648 phrases for the defined error codes are presented below. The reason 1649 phrase MUST be a UTF-8 [RFC3629] encoded sequence of less than 128 1650 characters (which can be as long as 763 bytes). 1652 0 1 2 3 1653 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 1654 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1655 | Reserved, should be 0 |Class| Number | 1656 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1657 | Reason Phrase (variable) .. 1658 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1660 Figure 10: ERROR-CODE Attribute 1662 To facilitate processing, the class of the error code (the hundreds 1663 digit) is encoded separately from the rest of the code, as shown in 1664 Figure 10. 1666 The Reserved bits SHOULD be 0, and are for alignment on 32-bit 1667 boundaries. Receivers MUST ignore these bits. The Class represents 1668 the hundreds digit of the error code. The value MUST be between 3 1669 and 6. The number represents the error code modulo 100, and its 1670 value MUST be between 0 and 99. 1672 The following error codes, along with their recommended reason 1673 phrases are defined: 1675 300 Try Alternate: The client should contact an alternate server for 1676 this request. This error response MUST only be sent if the 1677 request included a USERNAME attribute and a valid MESSAGE- 1678 INTEGRITY attribute; otherwise it MUST NOT be sent and error 1679 code 400 (Bad Request) is suggested. This error response MUST 1680 be protected with the MESSAGE-INTEGRITY attribute, and receivers 1681 MUST validate the MESSAGE-INTEGRITY of this response before 1682 redirecting themselves to an alternate server. 1684 Note: failure to generate and validate message-integrity 1685 for a 300 response allows an on-path attacker to falsify a 1686 300 response thus causing subsequent STUN messages to be 1687 sent to a victim. 1689 400 Bad Request: The request was malformed. The client SHOULD NOT 1690 retry the request without modification from the previous 1691 attempt. The server may not be able to generate a valid 1692 MESSAGE-INTEGRITY for this error, so the client MUST NOT expect 1693 a valid MESSAGE-INTEGRITY attribute on this response. 1695 401 Unauthorized: The request did not contain the correct 1696 credentials to proceed. The client should retry the request 1697 with proper credentials. 1699 420 Unknown Attribute: The server received STUN packet containing a 1700 comprehension-required attribute which it did not understand. 1701 The server MUST put this unknown attribute in the UNKNOWN- 1702 ATTRIBUTE attribute of its error response. 1704 438 Stale Nonce: The NONCE used by the client was no longer valid. 1705 The client should retry, using the NONCE provided in the 1706 response. 1708 500 Server Error: The server has suffered a temporary error. The 1709 client should try again. 1711 15.7. REALM 1713 The REALM attribute may be present in requests and responses. It 1714 contains text which meets the grammar for "realm-value" as described 1715 in RFC 3261 [RFC3261] but without the double quotes and their 1716 surrounding whitespace. That is, it is an unquoted realm-value (and 1717 is therefore a sequence of qdtext or quoted-pair). It MUST be a 1718 UTF-8 [RFC3629] encoded sequence of less than 128 characters (which 1719 can be as long as 763 bytes), and MUST have been processed using 1720 SASLPrep [RFC4013]. 1722 Presence of the REALM attribute in a request indicates that long-term 1723 credentials are being used for authentication. Presence in certain 1724 error responses indicates that the server wishes the client to use a 1725 long-term credential for authentication. 1727 15.8. NONCE 1729 The NONCE attribute may be present in requests and responses. It 1730 contains a sequence of qdtext or quoted-pair, which are defined in 1731 RFC 3261 [RFC3261]. Note that this means that the NONCE attribute 1732 will not contain actual quote characters. See RFC 2617 [RFC2617], 1733 Section 4.3, for guidance on selection of nonce values in a server. 1734 It MUST be less than 128 characters (which can be as long as 763 1735 bytes). 1737 15.9. UNKNOWN-ATTRIBUTES 1739 The UNKNOWN-ATTRIBUTES attribute is present only in an error response 1740 when the response code in the ERROR-CODE attribute is 420. 1742 The attribute contains a list of 16 bit values, each of which 1743 represents an attribute type that was not understood by the server. 1745 0 1 2 3 1746 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 1747 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1748 | Attribute 1 Type | Attribute 2 Type | 1749 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1750 | Attribute 3 Type | Attribute 4 Type ... 1751 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1753 Figure 11: Format of UNKNOWN-ATTRIBUTES attribute 1755 Note: In [RFC3489], this field was padded to 32 by duplicating the 1756 last attribute. In this version of the specification, the normal 1757 padding rules for attributes are used instead. 1759 15.10. SOFTWARE 1761 The SOFTWARE attribute contains a textual description of the software 1762 being used by the agent sending the message. It is used by clients 1763 and servers. Its value SHOULD include manufacturer and version 1764 number. The attribute has no impact on operation of the protocol, 1765 and serves only as a tool for diagnostic and debugging purposes. The 1766 value of SOFTWARE is variable length. It MUST be a UTF-8 [RFC3629] 1767 encoded sequence of less than 128 characters (which can be as long as 1768 763 bytes). 1770 15.11. ALTERNATE-SERVER 1772 The alternate server represents an alternate transport address 1773 identifying a different STUN server which the STUN client should try. 1775 It is encoded in the same way as MAPPED-ADDRESS, and thus refers to a 1776 single server by IP address. The IP address family MUST be identical 1777 to that of the source IP address of the request. 1779 This attribute MUST only appear in an error response that contains a 1780 MESSAGE-INTEGRITY attribute. This prevents it from being used in 1781 denial-of-service attacks. 1783 16. Security Considerations 1785 16.1. Attacks against the Protocol 1787 16.1.1. Outside Attacks 1789 An attacker can try to modify STUN messages in transit, in order to 1790 cause a failure in STUN operation. These attacks are detected for 1791 both requests and responses through the message integrity mechanism, 1792 using either a short term or long term credential. Of course, once 1793 detected, the manipulated packets will be dropped, causing the STUN 1794 transaction to effectively fail. This attack is possible only by an 1795 on-path attacker. 1797 An attacker that can observe, but not modify STUN messages in-transit 1798 (for example, an attacker present on a shared access medium, such as 1799 Wi-Fi), can see a STUN request, and then immediately send a STUN 1800 response, typically an error response, in order to disrupt STUN 1801 processing. This attack is also prevented for messages that utilize 1802 MESSAGE-INTEGRITY. However, some error responses, those related to 1803 authentication in particular, cannot be protected by MESSAGE- 1804 INTEGRITY. When STUN itself is run over a secure transport protocol 1805 (e.g., TLS), these attacks are completely mitigated. 1807 Depending on the STUN usage, these attacks may be of minimal 1808 consequence and thus do not require message integrity to mitigate. 1809 For example, when STUN is used to a basic STUN server to discover a 1810 server reflexive candidate for usage with ICE, authentication and 1811 message integrity are not required since these attacks are detected 1812 during the connectivity check phase. The connectivity checks 1813 themselves, however, require protection for proper operation of ICE 1814 overall. As described in Section 14, STUN usages describe when 1815 authentication and message integrity are needed. 1817 Since STUN uses the HMAC of a shared secret for authentication and 1818 integrity protection, it is subject to offline dictionary attacks. 1819 When authentication is utilized, it SHOULD be with a strong password 1820 that is not readily subject to offline dictionary attacks. 1821 Protection of the channel itself, using TLS, mitigates these attacks. 1822 However, STUN is most often run over UDP and in those cases, strong 1823 passwords are the only way to protect against these attacks. 1825 16.1.2. Inside Attacks 1827 A rogue client may try to launch a DoS attack against a server by 1828 sending it a large number of STUN requests. Fortunately, STUN 1829 requests can be processed statelessly by a server, making such 1830 attacks hard to launch. 1832 A rogue client may use a STUN server as a reflector, sending it 1833 requests with a falsified source IP address and port. In such a 1834 case, the response would be delivered to that source IP and port. 1835 There is no amplification of the number of packets with this attack 1836 (the STUN server sends one packet for each packet sent by the 1837 client), though there is a small increase in the amount of data, 1838 since STUN responses are typically larger than requests. This attack 1839 is mitigated by ingress source address filtering. 1841 Revealing the specific software version of the agent through the 1842 SOFTWARE attribute might allow them to become more vulnerable to 1843 attacks against software that is known to contain security holes. 1844 Implementers SHOULD make usage of the SOFTWARE attribute a 1845 configurable option. 1847 16.2. Attacks Affecting the Usage 1849 This section lists attacks that might be launched against a usage of 1850 STUN. Each STUN usage must consider whether these attacks are 1851 applicable to it, and if so, discuss counter-measures. 1853 Most of the attacks in this section revolve around an attacker 1854 modifying the reflexive address learned by a STUN client through a 1855 Binding Request/Binding Response transaction. Since the usage of the 1856 reflexive address is a function of the usage, the applicability and 1857 remediation of these attacks is usage-specific. In common 1858 situations, modification of the reflexive address by an on-path 1859 attacker is easy to do. Consider, for example, the common situation 1860 where STUN is run directly over UDP. In this case, an on-path 1861 attacker can modify the source IP address of the Binding Request 1862 before it arrives at the STUN server. The STUN server will then 1863 return this IP address in the XOR-MAPPED-ADDRESS attribute to the 1864 client, and send the response back to that (falsified) IP address and 1865 port. If the attacker can also intercept this response, it can 1866 direct it back towards the client. Protecting against this attack by 1867 using a message-integrity check is impossible, since a message- 1868 integrity value cannot cover the source IP address, since the 1869 intervening NAT must be able to modify this value. Instead, one 1870 solution to preventing the attacks listed below is for the client to 1871 verify the reflexive address learned, as is done in ICE 1872 [I-D.ietf-mmusic-ice]. Other usages may use other means to prevent 1873 these attacks. 1875 16.2.1. Attack I: DDoS Against a Target 1877 In this attack, the attacker provides one or more clients with the 1878 same faked reflexive address that points to the intended target. 1879 This will trick the STUN clients into thinking that their reflexive 1880 addresses are equal to that of the target. If the clients hand out 1881 that reflexive address in order to receive traffic on it (for 1882 example, in SIP messages), the traffic will instead be sent to the 1883 target. This attack can provide substantial amplification, 1884 especially when used with clients that are using STUN to enable 1885 multimedia applications. However, it can only be launched against 1886 targets for which packets from the STUN server to the target pass 1887 through the attacker, limiting the cases in which it is possible 1889 16.2.2. Attack II: Silencing a Client 1891 In this attack, the attacker provides a STUN client with a faked 1892 reflexive address. The reflexive address it provides is a transport 1893 address that routes to nowhere. As a result, the client won't 1894 receive any of the packets it expects to receive when it hands out 1895 the reflexive address. This exploitation is not very interesting for 1896 the attacker. It impacts a single client, which is frequently not 1897 the desired target. Moreover, any attacker that can mount the attack 1898 could also deny service to the client by other means, such as 1899 preventing the client from receiving any response from the STUN 1900 server, or even a DHCP server. As with the attack in Section 16.2.1, 1901 this attack is only possible when the attacker is on path for packets 1902 sent from the STUN server towards this unused IP address. 1904 16.2.3. Attack III: Assuming the Identity of a Client 1906 This attack is similar to attack II. However, the faked reflexive 1907 address points to the attacker itself. This allows the attacker to 1908 receive traffic which was destined for the client. 1910 16.2.4. Attack IV: Eavesdropping 1912 In this attack, the attacker forces the client to use a reflexive 1913 address that routes to itself. It then forwards any packets it 1914 receives to the client. This attack would allow the attacker to 1915 observe all packets sent to the client. However, in order to launch 1916 the attack, the attacker must have already been able to observe 1917 packets from the client to the STUN server. In most cases (such as 1918 when the attack is launched from an access network), this means that 1919 the attacker could already observe packets sent to the client. This 1920 attack is, as a result, only useful for observing traffic by 1921 attackers on the path from the client to the STUN server, but not 1922 generally on the path of packets being routed towards the client. 1924 16.3. Hash Agility Plan 1926 This specification uses HMAC-SHA-1 for computation of the message 1927 integrity. If, at a later time, HMAC-SHA-1 is found to be 1928 compromised, the following is the remedy that will be applied. 1930 We will define a STUN extension which introduces a new message 1931 integrity attribute, computed using a new hash. Clients would be 1932 required to include both the new and old message integrity attributes 1933 in their requests or indications. A new server will utilize the new 1934 message integrity attribute, and an old one, the old. After a 1935 transition period where mixed implementations are in deployment, the 1936 old message-integrity attribute will be deprecated by another 1937 specification, and clients will cease including it in requests. 1939 It is also important to note that the HMAC is done using a key which 1940 is itself computed using an MD5 of the user's password. The choice 1941 of the MD5 hash was made because of the existence of legacy databases 1942 which store passwords in that form. If future work finds that an 1943 HMAC of an MD5 input is not secure, and a different hash is needed, 1944 it can also be changed using this plan. However, this would require 1945 administrators to repopulate their databases. 1947 17. IAB Considerations 1949 The IAB has studied the problem of "Unilateral Self Address Fixing" 1950 (UNSAF), which is the general process by which a client attempts to 1951 determine its address in another realm on the other side of a NAT 1952 through a collaborative protocol reflection mechanism (RFC3424 1953 [RFC3424]). STUN can be used to perform this function using a 1954 Binding Request/Response transaction if one agent is behind a NAT and 1955 the other is on the public side of the NAT. 1957 The IAB has mandated that protocols developed for this purpose 1958 document a specific set of considerations. Because some STUN usages 1959 provide UNSAF functions (such as ICE [I-D.ietf-mmusic-ice] ), and 1960 others do not (such as SIP Outbound [I-D.ietf-sip-outbound]), answers 1961 to these considerations need to be addressed by the usages 1962 themselves. 1964 18. IANA Considerations 1966 IANA is hereby requested to create three new registries: a "STUN 1967 Methods Registry", a "STUN Attributes Registry", and a "STUN Error 1968 Codes registry". IANA is also requested to change the name of the 1969 assigned IANA port for STUN from "nat-stun-port" to "stun". 1971 18.1. STUN Methods Registry 1973 A STUN method is a hex number in the range 0x000 - 0xFFF. The 1974 encoding of STUN method into a STUN message is described in 1975 Section 6. 1977 The initial STUN methods are: 1979 0x000: (Reserved) 1980 0x001: Binding 1981 0x002: (Reserved; was SharedSecret) 1983 STUN methods in the range 0x000 - 0x7FF are assigned by IETF Review 1984 [RFC5226]. STUN methods in the range 0x800 - 0xFFF are assigned by 1985 Designated Expert [RFC5226] 1987 18.2. STUN Attribute Registry 1989 A STUN Attribute type is a hex number in the range 0x0000 - 0xFFFF. 1990 STUN attribute types in the range 0x0000 - 0x7FFF are considered 1991 comprehension-required; STUN attribute types in the range 0x8000 - 1992 0xFFFF are considered comprehension-optional. A STUN agent handles 1993 unknown comprehension-required and comprehension-optional attributes 1994 differently. 1996 The initial STUN Attributes types are: 1998 Comprehension-required range (0x0000-0x7FFF): 1999 0x0000: (Reserved) 2000 0x0001: MAPPED-ADDRESS 2001 0x0002: (Reserved; was RESPONSE-ADDRESS) 2002 0x0004: (Reserved; was SOURCE-ADDRESS) 2003 0x0005: (Reserved; was CHANGED-ADDRESS) 2004 0x0006: USERNAME 2005 0x0007: (Reserved; was PASSWORD) 2006 0x0008: MESSAGE-INTEGRITY 2007 0x0009: ERROR-CODE 2008 0x000A: UNKNOWN-ATTRIBUTES 2009 0x000B: (Reserved; was REFLECTED-FROM) 2010 0x0014: REALM 2011 0x0015: NONCE 2012 0x0020: XOR-MAPPED-ADDRESS 2014 Comprehension-optional range (0x8000-0xFFFF) 2015 0x8022: SOFTWARE 2016 0x8023: ALTERNATE-SERVER 2017 0x8028: FINGERPRINT 2019 STUN Attribute types in the first half of the comprehension-required 2020 range (0x0000 - 0x3FFF) and in the first half of the comprehension- 2021 optional range (0x8000 - 0xBFFF) are assigned by IETF Review 2022 [RFC5226]. STUN Attribute types in the second half of the 2023 comprehension-required range (0x4000 - 0x7FFF) and in the second half 2024 of the comprehension-optional range (0xC000 - 0xFFFF) are assigned by 2025 Designated Expert [RFC5226]. The responsibility of the expert is to 2026 verify that the selected codepoint(s) are not in use, and that the 2027 request is not for an abnormally large number of codepoints. 2028 Technical review of the extension itself is outside the scope of the 2029 designated expert responsibility. 2031 18.3. STUN Error Code Registry 2033 A STUN Error code is a number in the range 0 - 699. STUN error codes 2034 are accompanied by a textual reason phrase in UTF-8 [RFC3629] which 2035 is intended only for human consumption and can be anything 2036 appropriate; this document proposes only suggested values. 2038 STUN error codes are consistent in codepoint assignments and 2039 semantics with SIP [RFC3261] and HTTP [RFC2616]. 2041 The initial values in this registry are given in Section 15.6. 2043 New STUN error codes are assigned based on IETF Review [RFC5226]. 2045 The specification must carefully consider how clients that do not 2046 understand this error code will process it before granting the 2047 request. See the rules in Section 7.3.4. 2049 18.4. STUN UDP and TCP Port Numbers 2051 IANA has previously assigned port 3478 for STUN. This port appears 2052 in the IANA registry under the moniker "nat-stun-port". In order to 2053 align the DNS SRV procedures with the registered protocol service, 2054 IANA is requested to change the name of protocol assigned to port 2055 3478 from "nat-stun-port" to "stun", and the textual name from 2056 "Simple Traversal of UDP Through NAT (STUN)" to "Session Traversal 2057 Utilities for NAT", so that the IANA port registry would read: 2059 stun 3478/tcp Session Traversal Utilities for NAT (STUN) port 2060 stun 3478/udp Session Traversal Utilities for NAT (STUN) port 2062 In addition, IANA is requested to assign port numbers for the "stuns" 2063 service, defined over TCP and UDP. The UDP port is not currently 2064 defined however is reserved for future use. 2066 19. Changes Since RFC 3489 2068 This specification obsoletes RFC3489 [RFC3489]. This specification 2069 differs from RFC3489 in the following ways: 2071 o Removed the notion that STUN is a complete NAT traversal solution. 2072 STUN is now a tool that can be used to produce a NAT traversal 2073 solution. As a consequence, changed the name of the protocol to 2074 Session Traversal Utilities for NAT. 2076 o Introduced the concept of STUN usages, and described what a usage 2077 of STUN must document. 2079 o Removed the usage of STUN for NAT type detection and binding 2080 lifetime discovery. These techniques have proven overly brittle 2081 due to wider variations in the types of NAT devices than described 2082 in this document. Removed the RESPONSE-ADDRESS, CHANGED-ADDRESS, 2083 CHANGE-REQUEST, SOURCE-ADDRESS, and REFLECTED-FROM attributes. 2085 o Added a fixed 32-bit magic cookie and reduced length of 2086 transaction ID by 32 bits. The magic cookie begins at the same 2087 offset as the original transaction ID. 2089 o Added the XOR-MAPPED-ADDRESS attribute, which is included in 2090 Binding Responses if the magic cookie is present in the request. 2091 Otherwise the RFC3489 behavior is retained (that is, Binding 2092 Response includes MAPPED-ADDRESS). See discussion in XOR-MAPPED- 2093 ADDRESS regarding this change. 2095 o Introduced formal structure into the Message Type header field, 2096 with an explicit pair of bits for indication of request, response, 2097 error response or indication. Consequently, the message type 2098 field is split into the class (one of the previous four) and 2099 method. 2101 o Explicitly point out that the most significant two bits of STUN 2102 are 0b00, allowing easy differentiation with RTP packets when used 2103 with ICE. 2105 o Added the FINGERPRINT attribute to provide a method of definitely 2106 detecting the difference between STUN and another protocol when 2107 the two protocols are multiplexed together. 2109 o Added support for IPv6. Made it clear that an IPv4 client could 2110 get a v6 mapped address, and vice-a-versa. 2112 o Added long-term credential-based authentication. 2114 o Added the SOFTWARE, REALM, NONCE, and ALTERNATE-SERVER attributes. 2116 o Removed the SharedSecret method, and thus the PASSWORD attribute. 2117 This method was almost never implemented and is not needed with 2118 current usages. 2120 o Removed recommendation to continue listening for STUN Responses 2121 for 10 seconds in an attempt to recognize an attack. 2123 o Changed transaction timers to be more TCP friendly. 2125 o Removed the STUN example that centered around the separation of 2126 the control and media planes. Instead, provided more information 2127 on using STUN with protocols. 2129 o Defined a generic padding mechanism that changes the 2130 interpretation of the length attribute. This would, in theory, 2131 break backwards compatibility. However, the mechanism in RFC 3489 2132 never worked for the few attributes that weren't aligned naturally 2133 on 32 bit boundaries. 2135 o REALM, SERVER, reason phrases and NONCE limited to 127 characters. 2136 USERNAME to 513 bytes. 2138 o Changed the DNS SRV procedures for TCP and TLS. UDP remains the 2139 same as before. 2141 20. Contributors 2143 Christian Huitema and Joel Weinberger were original co-authors of RFC 2144 3489. 2146 21. Acknowledgements 2148 The authors would like to thank Cedric Aoun, Pete Cordell, Cullen 2149 Jennings, Bob Penfield, Xavier Marjou, Magnus Westerlund, Miguel 2150 Garcia, Bruce Lowekamp and Chris Sullivan for their comments, and 2151 Baruch Sterman and Alan Hawrylyshen for initial implementations. 2152 Thanks for Leslie Daigle, Allison Mankin, Eric Rescorla, and Henning 2153 Schulzrinne for IESG and IAB input on this work. 2155 22. References 2157 22.1. Normative References 2159 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2160 Requirement Levels", BCP 14, RFC 2119, March 1997. 2162 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 2163 September 1981. 2165 [RFC2782] Gulbrandsen, A., Vixie, P., and L. Esibov, "A DNS RR for 2166 specifying the location of services (DNS SRV)", RFC 2782, 2167 February 2000. 2169 [RFC2818] Rescorla, E., "HTTP Over TLS", RFC 2818, May 2000. 2171 [RFC1122] Braden, R., "Requirements for Internet Hosts - 2172 Communication Layers", STD 3, RFC 1122, October 1989. 2174 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 2175 (IPv6) Specification", RFC 2460, December 1998. 2177 [RFC2617] Franks, J., Hallam-Baker, P., Hostetler, J., Lawrence, S., 2178 Leach, P., Luotonen, A., and L. Stewart, "HTTP 2179 Authentication: Basic and Digest Access Authentication", 2180 RFC 2617, June 1999. 2182 [RFC2988] Paxson, V. and M. Allman, "Computing TCP's Retransmission 2183 Timer", RFC 2988, November 2000. 2185 [RFC2104] Krawczyk, H., Bellare, M., and R. Canetti, "HMAC: Keyed- 2186 Hashing for Message Authentication", RFC 2104, 2187 February 1997. 2189 [ITU.V42.2002] 2190 International Telecommunications Union, "Error-correcting 2191 Procedures for DCEs Using Asynchronous-to-Synchronous 2192 Conversion", ITU-T Recommendation V.42, March 2002. 2194 [RFC3629] Yergeau, F., "UTF-8, a transformation format of ISO 2195 10646", STD 63, RFC 3629, November 2003. 2197 [RFC1321] Rivest, R., "The MD5 Message-Digest Algorithm", RFC 1321, 2198 April 1992. 2200 [RFC4013] Zeilenga, K., "SASLprep: Stringprep Profile for User Names 2201 and Passwords", RFC 4013, February 2005. 2203 22.2. Informational References 2205 [RFC3261] Rosenberg, J., Schulzrinne, H., Camarillo, G., Johnston, 2206 A., Peterson, J., Sparks, R., Handley, M., and E. 2207 Schooler, "SIP: Session Initiation Protocol", RFC 3261, 2208 June 2002. 2210 [RFC2616] Fielding, R., Gettys, J., Mogul, J., Frystyk, H., 2211 Masinter, L., Leach, P., and T. Berners-Lee, "Hypertext 2212 Transfer Protocol -- HTTP/1.1", RFC 2616, June 1999. 2214 [RFC4107] Bellovin, S. and R. Housley, "Guidelines for Cryptographic 2215 Key Management", BCP 107, RFC 4107, June 2005. 2217 [I-D.ietf-mmusic-ice] 2218 Rosenberg, J., "Interactive Connectivity Establishment 2219 (ICE): A Protocol for Network Address Translator (NAT) 2220 Traversal for Offer/Answer Protocols", 2221 draft-ietf-mmusic-ice-19 (work in progress), October 2007. 2223 [RFC3489] Rosenberg, J., Weinberger, J., Huitema, C., and R. Mahy, 2224 "STUN - Simple Traversal of User Datagram Protocol (UDP) 2225 Through Network Address Translators (NATs)", RFC 3489, 2226 March 2003. 2228 [I-D.ietf-behave-turn] 2229 Rosenberg, J., Mahy, R., and P. Matthews, "Traversal Using 2230 Relays around NAT (TURN): Relay Extensions to Session 2231 Traversal Utilities for NAT (STUN)", 2232 draft-ietf-behave-turn-08 (work in progress), June 2008. 2234 [I-D.ietf-sip-outbound] 2235 Jennings, C. and R. Mahy, "Managing Client Initiated 2236 Connections in the Session Initiation Protocol (SIP)", 2237 draft-ietf-sip-outbound-15 (work in progress), June 2008. 2239 [I-D.ietf-behave-nat-behavior-discovery] 2240 MacDonald, D. and B. Lowekamp, "NAT Behavior Discovery 2241 Using STUN", draft-ietf-behave-nat-behavior-discovery-03 2242 (work in progress), February 2008. 2244 [I-D.ietf-mmusic-ice-tcp] 2245 Rosenberg, J., "TCP Candidates with Interactive 2246 Connectivity Establishment (ICE)", 2247 draft-ietf-mmusic-ice-tcp-06 (work in progress), 2248 February 2008. 2250 [RFC3264] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model 2251 with Session Description Protocol (SDP)", RFC 3264, 2252 June 2002. 2254 [RFC3424] Daigle, L. and IAB, "IAB Considerations for UNilateral 2255 Self-Address Fixing (UNSAF) Across Network Address 2256 Translation", RFC 3424, November 2002. 2258 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 2259 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 2260 May 2008. 2262 [KARN87] Karn, P. and C. Partridge, "Improving Round-Trip Time 2263 Estimates in Reliable Transport Protocols", SIGCOMM 1987, 2264 August 1987. 2266 Appendix A. C Snippet to Determine STUN Message Types 2268 Given an 16-bit STUN message type value in host byte order in 2269 msg_type parameter, below are C macros to determine the STUN message 2270 types: 2272 #define IS_REQUEST(msg_type) (((msg_type) & 0x0110) == 0x0000) 2273 #define IS_INDICATION(msg_type) (((msg_type) & 0x0110) == 0x0010) 2274 #define IS_SUCCESS_RESP(msg_type) (((msg_type) & 0x0110) == 0x0100) 2275 #define IS_ERR_RESP(msg_type) (((msg_type) & 0x0110) == 0x0110) 2277 Authors' Addresses 2279 Jonathan Rosenberg 2280 Cisco 2281 Edison, NJ 2282 US 2284 Email: jdrosen@cisco.com 2285 URI: http://www.jdrosen.net 2287 Rohan Mahy 2288 Plantronics 2289 345 Encinal Street 2290 Santa Cruz, CA 95060 2291 US 2293 Email: rohan@ekabal.com 2295 Philip Matthews 2296 Avaya 2297 1135 Innovation Drive 2298 Ottawa, Ontario K2K 3G7 2299 Canada 2301 Phone: +1 613 592 4343 x224 2302 Fax: 2303 Email: philip_matthews@magma.ca 2304 URI: 2306 Dan Wing 2307 Cisco 2308 771 Alder Drive 2309 San Jose, CA 95035 2310 US 2312 Email: dwing@cisco.com 2314 Full Copyright Statement 2316 Copyright (C) The IETF Trust (2008). 2318 This document is subject to the rights, licenses and restrictions 2319 contained in BCP 78, and except as set forth therein, the authors 2320 retain all their rights. 2322 This document and the information contained herein are provided on an 2323 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 2324 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 2325 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 2326 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 2327 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 2328 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 2330 Intellectual Property 2332 The IETF takes no position regarding the validity or scope of any 2333 Intellectual Property Rights or other rights that might be claimed to 2334 pertain to the implementation or use of the technology described in 2335 this document or the extent to which any license under such rights 2336 might or might not be available; nor does it represent that it has 2337 made any independent effort to identify any such rights. Information 2338 on the procedures with respect to rights in RFC documents can be 2339 found in BCP 78 and BCP 79. 2341 Copies of IPR disclosures made to the IETF Secretariat and any 2342 assurances of licenses to be made available, or the result of an 2343 attempt made to obtain a general license or permission for the use of 2344 such proprietary rights by implementers or users of this 2345 specification can be obtained from the IETF on-line IPR repository at 2346 http://www.ietf.org/ipr. 2348 The IETF invites any interested party to bring to its attention any 2349 copyrights, patents or patent applications, or other proprietary 2350 rights that may cover technology that may be required to implement 2351 this standard. Please address the information to the IETF at 2352 ietf-ipr@ietf.org. 2354 Acknowledgment 2356 Funding for the RFC Editor function is provided by the IETF 2357 Administrative Support Activity (IASA).