1 Network Working Group R. Austein 2 Request for Comments: 5001 ISC 3 Category: Standards Track August 2007 4 5 6 DNS Name Server Identifier (NSID) Option 7 8 Status of This Memo 9 10 This document specifies an Internet standards track protocol for the 11 Internet community, and requests discussion and suggestions for 12 improvements. Please refer to the current edition of the "Internet 13 Official Protocol Standards" (STD 1) for the standardization state 14 and status of this protocol. Distribution of this memo is unlimited. 15 16 Copyright Notice 17 18 Copyright (C) The IETF Trust (2007). 19 20 Abstract 21 22 With the increased use of DNS anycast, load balancing, and other 23 mechanisms allowing more than one DNS name server to share a single 24 IP address, it is sometimes difficult to tell which of a pool of name 25 servers has answered a particular query. While existing ad-hoc 26 mechanisms allow an operator to send follow-up queries when it is 27 necessary to debug such a configuration, the only completely reliable 28 way to obtain the identity of the name server that responded is to 29 have the name server include this information in the response itself. 30 This note defines a protocol extension to support this functionality. 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Austein Standards Track [Page 1] 53 RFC 5001 DNS NSID August 2007 54 55 56 Table of Contents 57 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 2 59 1.1. Reserved Words . . . . . . . . . . . . . . . . . . . . . . 3 60 2. Protocol . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 61 2.1. Resolver Behavior . . . . . . . . . . . . . . . . . . . . 3 62 2.2. Name Server Behavior . . . . . . . . . . . . . . . . . . . 3 63 2.3. The NSID Option . . . . . . . . . . . . . . . . . . . . . 4 64 2.4. Presentation Format . . . . . . . . . . . . . . . . . . . 4 65 3. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 4 66 3.1. The NSID Payload . . . . . . . . . . . . . . . . . . . . . 4 67 3.2. NSID Is Not Transitive . . . . . . . . . . . . . . . . . . 7 68 3.3. User Interface Issues . . . . . . . . . . . . . . . . . . 7 69 3.4. Truncation . . . . . . . . . . . . . . . . . . . . . . . . 8 70 4. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 8 71 5. Security Considerations . . . . . . . . . . . . . . . . . . . 9 72 6. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 9 73 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 9 74 7.1. Normative References . . . . . . . . . . . . . . . . . . . 9 75 7.2. Informative References . . . . . . . . . . . . . . . . . . 10 76 77 1. Introduction 78 79 With the increased use of DNS anycast, load balancing, and other 80 mechanisms allowing more than one DNS name server to share a single 81 IP address, it is sometimes difficult to tell which of a pool of name 82 servers has answered a particular query. 83 84 Existing ad-hoc mechanisms allow an operator to send follow-up 85 queries when it is necessary to debug such a configuration, but there 86 are situations in which this is not a totally satisfactory solution, 87 since anycast routing may have changed, or the server pool in 88 question may be behind some kind of extremely dynamic load balancing 89 hardware. Thus, while these ad-hoc mechanisms are certainly better 90 than nothing (and have the advantage of already being deployed), a 91 better solution seems desirable. 92 93 Given that a DNS query is an idempotent operation with no retained 94 state, it would appear that the only completely reliable way to 95 obtain the identity of the name server that responded to a particular 96 query is to have that name server include identifying information in 97 the response itself. This note defines a protocol enhancement to 98 achieve this. 99 100 101 102 103 104 105 106 107 Austein Standards Track [Page 2] 108 RFC 5001 DNS NSID August 2007 109 110 111 1.1. Reserved Words 112 113 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 114 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 115 document are to be interpreted as described in [RFC2119]. 116 117 2. Protocol 118 119 This note uses an EDNS [RFC2671] option to signal the resolver's 120 desire for information identifying the name server and to hold the 121 name server's response, if any. 122 123 2.1. Resolver Behavior 124 125 A resolver signals its desire for information identifying a name 126 server by sending an empty NSID option (Section 2.3) in an EDNS OPT 127 pseudo-RR in the query message. 128 129 The resolver MUST NOT include any NSID payload data in the query 130 message. 131 132 The semantics of an NSID request are not transitive. That is: the 133 presence of an NSID option in a query is a request that the name 134 server which receives the query identify itself. If the name server 135 side of a recursive name server receives an NSID request, the client 136 is asking the recursive name server to identify itself; if the 137 resolver side of the recursive name server wishes to receive 138 identifying information, it is free to add NSID requests in its own 139 queries, but that is a separate matter. 140 141 2.2. Name Server Behavior 142 143 A name server that understands the NSID option and chooses to honor a 144 particular NSID request responds by including identifying information 145 in a NSID option (Section 2.3) in an EDNS OPT pseudo-RR in the 146 response message. 147 148 The name server MUST ignore any NSID payload data that might be 149 present in the query message. 150 151 The NSID option is not transitive. A name server MUST NOT send an 152 NSID option back to a resolver which did not request it. In 153 particular, while a recursive name server may choose to add an NSID 154 option when sending a query, this has no effect on the presence or 155 absence of the NSID option in the recursive name server's response to 156 the original client. 157 158 159 160 161 162 Austein Standards Track [Page 3] 163 RFC 5001 DNS NSID August 2007 164 165 166 As stated in Section 2.1, this mechanism is not restricted to 167 authoritative name servers; the semantics are intended to be equally 168 applicable to recursive name servers. 169 170 2.3. The NSID Option 171 172 The OPTION-CODE for the NSID option is 3. 173 174 The OPTION-DATA for the NSID option is an opaque byte string, the 175 semantics of which are deliberately left outside the protocol. See 176 Section 3.1 for discussion. 177 178 2.4. Presentation Format 179 180 User interfaces MUST read and write the contents of the NSID option 181 as a sequence of hexadecimal digits, two digits per payload octet. 182 183 The NSID payload is binary data. Any comparison between NSID 184 payloads MUST be a comparison of the raw binary data. Copy 185 operations MUST NOT assume that the raw NSID payload is null- 186 terminated. Any resemblance between raw NSID payload data and any 187 form of text is purely a convenience, and does not change the 188 underlying nature of the payload data. 189 190 See Section 3.3 for discussion. 191 192 3. Discussion 193 194 This section discusses certain aspects of the protocol and explains 195 considerations that led to the chosen design. 196 197 3.1. The NSID Payload 198 199 The syntax and semantics of the content of the NSID option are 200 deliberately left outside the scope of this specification. 201 202 Choosing the NSID content is a prerogative of the server 203 administrator. The server administrator might choose to encode the 204 NSID content in such a way that the server operator (or clients 205 authorized by the server operator) can decode the NSID content to 206 obtain more information than other clients can. Alternatively, the 207 server operator might choose unencoded NSID content that is equally 208 meaningful to any client. 209 210 This section describes some of the kinds of data that server 211 administrators might choose to provide as the content of the NSID 212 option, and explains the reasoning behind specifying a simple opaque 213 byte string in Section 2.3. 214 215 216 217 Austein Standards Track [Page 4] 218 RFC 5001 DNS NSID August 2007 219 220 221 There are several possibilities for the payload of the NSID option: 222 223 o It could be the "real" name of the specific name server within the 224 name server pool. 225 226 o It could be the "real" IP address (IPv4 or IPv6) of the name 227 server within the name server pool. 228 229 o It could be some sort of pseudo-random number generated in a 230 predictable fashion somehow using the server's IP address or name 231 as a seed value. 232 233 o It could be some sort of probabilistically unique identifier 234 initially derived from some sort of random number generator then 235 preserved across reboots of the name server. 236 237 o It could be some sort of dynamically generated identifier so that 238 only the name server operator could tell whether or not any two 239 queries had been answered by the same server. 240 241 o It could be a blob of signed data, with a corresponding key which 242 might (or might not) be available via DNS lookups. 243 244 o It could be a blob of encrypted data, the key for which could be 245 restricted to parties with a need to know (in the opinion of the 246 server operator). 247 248 o It could be an arbitrary string of octets chosen at the discretion 249 of the name server operator. 250 251 Each of these options has advantages and disadvantages: 252 253 o Using the "real" name is simple, but the name server may not have 254 a "real" name. 255 256 o Using the "real" address is also simple, and the name server 257 almost certainly does have at least one non-anycast IP address for 258 maintenance operations, but the operator of the name server may 259 not be willing to divulge its non-anycast address. 260 261 o Given that one common reason for using anycast DNS techniques is 262 an attempt to harden a critical name server against denial of 263 service attacks, some name server operators are likely to want an 264 identifier other than the "real" name or "real" address of the 265 name server instance. 266 267 o Using a hash or pseudo-random number can provide a fixed length 268 value that the resolver can use to tell two name servers apart 269 270 271 272 Austein Standards Track [Page 5] 273 RFC 5001 DNS NSID August 2007 274 275 276 without necessarily being able to tell where either one of them 277 "really" is, but makes debugging more difficult if one happens to 278 be in a friendly open environment. Furthermore, hashing might not 279 add much value, since a hash based on an IPv4 address still only 280 involves a 32-bit search space, and DNS names used for servers 281 that operators might have to debug at 4am tend not to be very 282 random. 283 284 o Probabilistically unique identifiers have properties similar to 285 hashed identifiers, but (given a sufficiently good random number 286 generator) are immune to the search space issues. However, the 287 strength of this approach is also its weakness: there is no 288 algorithmic transformation by which even the server operator can 289 associate name server instances with identifiers while debugging, 290 which might be annoying. This approach also requires the name 291 server instance to preserve the probabilistically unique 292 identifier across reboots, but this does not appear to be a 293 serious restriction, since authoritative nameservers almost always 294 have some form of non-volatile storage. In the rare case of a 295 name server that does not have any way to store such an 296 identifier, nothing terrible will happen if the name server 297 generates a new identifier every time it reboots. 298 299 o Using an arbitrary octet string gives name server operators yet 300 another setting to configure, or mis-configure, or forget to 301 configure. Having all the nodes in an anycast name server 302 constellation identify themselves as "My Name Server" would not be 303 particularly useful. 304 305 o A signed blob is not particularly useful as an NSID payload unless 306 the signed data is dynamic and includes some kind of replay 307 protection, such as a timestamp or some kind of data identifying 308 the requestor. Signed blobs that meet these criteria could 309 conceivably be useful in some situations but would require 310 detailed security analysis beyond the scope of this document. 311 312 o A static encrypted blob would not be particularly useful, as it 313 would be subject to replay attacks and would, in effect, just be a 314 random number to any party that does not possess the decryption 315 key. Dynamic encrypted blobs could conceivably be useful in some 316 situations but, as with signed blobs, dynamic encrypted blobs 317 would require detailed security analysis beyond the scope of this 318 document. 319 320 Given all of the issues listed above, there does not appear to be a 321 single solution that will meet all needs. Section 2.3 therefore 322 defines the NSID payload to be an opaque byte string and leaves the 323 choice of payload up to the implementor and name server operator. 324 325 326 327 Austein Standards Track [Page 6] 328 RFC 5001 DNS NSID August 2007 329 330 331 The following guidelines may be useful to implementors and server 332 operators: 333 334 o Operators for whom divulging the unicast address is an issue could 335 use the raw binary representation of a probabilistically unique 336 random number. This should probably be the default implementation 337 behavior. 338 339 o Operators for whom divulging the unicast address is not an issue 340 could just use the raw binary representation of a unicast address 341 for simplicity. This should only be done via an explicit 342 configuration choice by the operator. 343 344 o Operators who really need or want the ability to set the NSID 345 payload to an arbitrary value could do so, but this should only be 346 done via an explicit configuration choice by the operator. 347 348 This approach appears to provide enough information for useful 349 debugging without unintentionally leaking the maintenance addresses 350 of anycast name servers to nogoodniks, while also allowing name 351 server operators who do not find such leakage threatening to provide 352 more information at their own discretion. 353 354 3.2. NSID Is Not Transitive 355 356 As specified in Section 2.1 and Section 2.2, the NSID option is not 357 transitive. This is strictly a hop-by-hop mechanism. 358 359 Most of the discussion of name server identification to date has 360 focused on identifying authoritative name servers, since the best 361 known cases of anycast name servers are a subset of the name servers 362 for the root zone. However, given that anycast DNS techniques are 363 also applicable to recursive name servers, the mechanism may also be 364 useful with recursive name servers. The hop-by-hop semantics support 365 this. 366 367 While there might be some utility in having a transitive variant of 368 this mechanism (so that, for example, a stub resolver could ask a 369 recursive server to tell it which authoritative name server provided 370 a particular answer to the recursive name server), the semantics of 371 such a variant would be more complicated, and are left for future 372 work. 373 374 3.3. User Interface Issues 375 376 Given the range of possible payload contents described in 377 Section 3.1, it is not possible to define a single presentation 378 format for the NSID payload that is efficient, convenient, 379 380 381 382 Austein Standards Track [Page 7] 383 RFC 5001 DNS NSID August 2007 384 385 386 unambiguous, and aesthetically pleasing. In particular, while it is 387 tempting to use a presentation format that uses some form of textual 388 strings, attempting to support this would significantly complicate 389 what's intended to be a very simple debugging mechanism. 390 391 In some cases the content of the NSID payload may be binary data 392 meaningful only to the name server operator, and may not be 393 meaningful to the user or application, but the user or application 394 must be able to capture the entire content anyway in order for it to 395 be useful. Thus, the presentation format must support arbitrary 396 binary data. 397 398 In cases where the name server operator derives the NSID payload from 399 textual data, a textual form such as US-ASCII or UTF-8 strings might 400 at first glance seem easier for a user to deal with. There are, 401 however, a number of complex issues involving internationalized text 402 which, if fully addressed here, would require a set of rules 403 significantly longer than the rest of this specification. See 404 [RFC2277] for an overview of some of these issues. 405 406 It is much more important for the NSID payload data to be passed 407 unambiguously from server administrator to user and back again than 408 it is for the payload data to be pretty while in transit. In 409 particular, it's critical that it be straightforward for a user to 410 cut and paste an exact copy of the NSID payload output by a debugging 411 tool into other formats such as email messages or web forms without 412 distortion. Hexadecimal strings, while ugly, are also robust. 413 414 3.4. Truncation 415 416 In some cases, adding the NSID option to a response message may 417 trigger message truncation. This specification does not change the 418 rules for DNS message truncation in any way, but implementors will 419 need to pay attention to this issue. 420 421 Including the NSID option in a response is always optional, so this 422 specification never requires name servers to truncate response 423 messages. 424 425 By definition, a resolver that requests NSID responses also supports 426 EDNS, so a resolver that requests NSID responses can also use the 427 "sender's UDP payload size" field of the OPT pseudo-RR to signal a 428 receive buffer size large enough to make truncation unlikely. 429 430 4. IANA Considerations 431 432 IANA has allocated EDNS option code 3 for the NSID option 433 (Section 2.3). 434 435 436 437 Austein Standards Track [Page 8] 438 RFC 5001 DNS NSID August 2007 439 440 441 5. Security Considerations 442 443 This document describes a channel signaling mechanism intended 444 primarily for debugging. Channel signaling mechanisms are outside 445 the scope of DNSSEC, per se. Applications that require integrity 446 protection for the data being signaled will need to use a channel 447 security mechanism such as TSIG [RFC2845]. 448 449 Section 3.1 discusses a number of different kinds of information that 450 a name server operator might choose to provide as the value of the 451 NSID option. Some of these kinds of information are security 452 sensitive in some environments. This specification deliberately 453 leaves the syntax and semantics of the NSID option content up to the 454 implementation and the name server operator. 455 456 Two of the possible kinds of payload data discussed in Section 3.1 457 involve a digital signature and encryption, respectively. While this 458 specification discusses some of the pitfalls that might lurk for 459 careless users of these kinds of payload data, full analysis of the 460 issues that would be involved in these kinds of payload data would 461 require knowledge of the content to be signed or encrypted, 462 algorithms to be used, and so forth, which is beyond the scope of 463 this document. Implementors should seek competent advice before 464 attempting to use these kinds of NSID payloads. 465 466 6. Acknowledgements 467 468 Thanks to: Joe Abley, Harald Alvestrand, Dean Anderson, Mark Andrews, 469 Roy Arends, Steve Bellovin, Alex Bligh, Randy Bush, David Conrad, 470 John Dickinson, Alfred Hoenes, Johan Ihren, Daniel Karrenberg, Peter 471 Koch, William Leibzon, Ed Lewis, Thomas Narten, Mike Patton, Geoffrey 472 Sisson, Andrew Sullivan, Mike StJohns, Tom Taylor, Paul Vixie, Sam 473 Weiler, and Suzanne Woolf, none of whom are responsible for what the 474 author did with their comments and suggestions. Apologies to anyone 475 inadvertently omitted from the above list. 476 477 7. References 478 479 7.1. Normative References 480 481 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 482 Requirement Levels", RFC 2119, BCP 14, March 1997. 483 484 [RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", 485 RFC 2671, August 1999. 486 487 488 489 490 491 492 Austein Standards Track [Page 9] 493 RFC 5001 DNS NSID August 2007 494 495 496 [RFC2845] Vixie, P., Gudmundsson, O., Eastlake 3rd, D., and B. 497 Wellington, "Secret Key Transaction Authentication for DNS 498 (TSIG)", RFC 2845, May 2000. 499 500 7.2. Informative References 501 502 [RFC2277] Alvestrand, H., "IETF Policy on Character Sets and 503 Languages", RFC 2277, BCP 18, January 1998. 504 505 Author's Address 506 507 Rob Austein 508 ISC 509 950 Charter Street 510 Redwood City, CA 94063 511 USA 512 513 EMail: email@example.com 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 Austein Standards Track [Page 10] 548 RFC 5001 DNS NSID August 2007 549 550 551 Full Copyright Statement 552 553 Copyright (C) The IETF Trust (2007). 554 555 This document is subject to the rights, licenses and restrictions 556 contained in BCP 78, and except as set forth therein, the authors 557 retain all their rights. 558 559 This document and the information contained herein are provided on an 560 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 561 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND 562 THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS 563 OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF 564 THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 565 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 566 567 Intellectual Property 568 569 The IETF takes no position regarding the validity or scope of any 570 Intellectual Property Rights or other rights that might be claimed to 571 pertain to the implementation or use of the technology described in 572 this document or the extent to which any license under such rights 573 might or might not be available; nor does it represent that it has 574 made any independent effort to identify any such rights. Information 575 on the procedures with respect to rights in RFC documents can be 576 found in BCP 78 and BCP 79. 577 578 Copies of IPR disclosures made to the IETF Secretariat and any 579 assurances of licenses to be made available, or the result of an 580 attempt made to obtain a general license or permission for the use of 581 such proprietary rights by implementers or users of this 582 specification can be obtained from the IETF on-line IPR repository at 583 http://www.ietf.org/ipr. 584 585 The IETF invites any interested party to bring to its attention any 586 copyrights, patents or patent applications, or other proprietary 587 rights that may cover technology that may be required to implement 588 this standard. Please address the information to the IETF at 589 firstname.lastname@example.org. 590 591 Acknowledgement 592 593 Funding for the RFC Editor function is currently provided by the 594 Internet Society. 595 596 597 598 599 600 601 602 Austein Standards Track [Page 11] 603
The IETF is responsible for the creation and maintenance of the DNS RFCs. The ICANN DNS RFC annotation project provides a forum for collecting community annotations on these RFCs as an aid to understanding for implementers and any interested parties. The annotations displayed here are not the result of the IETF consensus process.
This RFC is included in the DNS RFCs annotation project whose home page is here.
This RFC is implemented in BIND 9.18 (all versions).