1 Network Working Group R. Elz 2 Request for Comments: 2181 University of Melbourne 3 Updates: 1034, 1035, 1123 R. Bush 4 Category: Standards Track RGnet, Inc. 5 July 1997 6 7 8 Clarifications to the DNS Specification 9 10 Status of this Memo 11 12 This document specifies an Internet standards track protocol for the 13 Internet community, and requests discussion and suggestions for 14 improvements. Please refer to the current edition of the "Internet 15 Official Protocol Standards" (STD 1) for the standardization state 16 and status of this protocol. Distribution of this memo is unlimited. 17 18 1. Abstract 19 20 This document considers some areas that have been identified as 21 problems with the specification of the Domain Name System, and 22 proposes remedies for the defects identified. Eight separate issues 23 are considered: 24 25 + IP packet header address usage from multi-homed servers, 26 + TTLs in sets of records with the same name, class, and type, 27 + correct handling of zone cuts, 28 + three minor issues concerning SOA records and their use, 29 + the precise definition of the Time to Live (TTL) 30 + Use of the TC (truncated) header bit 31 + the issue of what is an authoritative, or canonical, name, 32 + and the issue of what makes a valid DNS label. 33 34 The first six of these are areas where the correct behaviour has been 35 somewhat unclear, we seek to rectify that. The other two are already 36 adequately specified, however the specifications seem to be sometimes 37 ignored. We seek to reinforce the existing specifications. 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Elz & Bush Standards Track [Page 1] 53 RFC 2181 Clarifications to the DNS Specification July 1997 54 55 56 57 58 Contents 59 60 1 Abstract ................................................... 1 61 2 Introduction ............................................... 2 62 3 Terminology ................................................ 3 63 4 Server Reply Source Address Selection ...................... 3 64 5 Resource Record Sets ....................................... 4 65 6 Zone Cuts .................................................. 8 66 7 SOA RRs .................................................... 10 67 8 Time to Live (TTL) ......................................... 10 68 9 The TC (truncated) header bit .............................. 11 69 10 Naming issues .............................................. 11 70 11 Name syntax ................................................ 13 71 12 Security Considerations .................................... 14 72 13 References ................................................. 14 73 14 Acknowledgements ........................................... 15 74 15 Authors' Addresses ......................................... 15 75 76 77 78 79 2. Introduction 80 81 Several problem areas in the Domain Name System specification 82 [RFC1034, RFC1035] have been noted through the years [RFC1123]. This 83 document addresses several additional problem areas. The issues here 84 are independent. Those issues are the question of which source 85 address a multi-homed DNS server should use when replying to a query, 86 the issue of differing TTLs for DNS records with the same label, 87 class and type, and the issue of canonical names, what they are, how 88 CNAME records relate, what names are legal in what parts of the DNS, 89 and what is the valid syntax of a DNS name. 90 91 Clarifications to the DNS specification to avoid these problems are 92 made in this memo. A minor ambiguity in RFC1034 concerned with SOA 93 records is also corrected, as is one in the definition of the TTL 94 (Time To Live) and some possible confusion in use of the TC bit. 95 96 97 98 99 100 101 102 103 104 105 106 107 Elz & Bush Standards Track [Page 2] 108 RFC 2181 Clarifications to the DNS Specification July 1997 109 110 111 3. Terminology 112 113 This memo does not use the oft used expressions MUST, SHOULD, MAY, or 114 their negative forms. In some sections it may seem that a 115 specification is worded mildly, and hence some may infer that the 116 specification is optional. That is not correct. Anywhere that this 117 memo suggests that some action should be carried out, or must be 118 carried out, or that some behaviour is acceptable, or not, that is to 119 be considered as a fundamental aspect of this specification, 120 regardless of the specific words used. If some behaviour or action 121 is truly optional, that will be clearly specified by the text. 122 123 4. Server Reply Source Address Selection 124 125 Most, if not all, DNS clients, expect the address from which a reply 126 is received to be the same address as that to which the query 127 eliciting the reply was sent. This is true for servers acting as 128 clients for the purposes of recursive query resolution, as well as 129 simple resolver clients. The address, along with the identifier (ID) 130 in the reply is used for disambiguating replies, and filtering 131 spurious responses. This may, or may not, have been intended when 132 the DNS was designed, but is now a fact of life. 133 134 Some multi-homed hosts running DNS servers generate a reply using a 135 source address that is not the same as the destination address from 136 the client's request packet. Such replies will be discarded by the 137 client because the source address of the reply does not match that of 138 a host to which the client sent the original request. That is, it 139 appears to be an unsolicited response. 140 141 4.1. UDP Source Address Selection 142 143 To avoid these problems, servers when responding to queries using UDP 144 must cause the reply to be sent with the source address field in the 145 IP header set to the address that was in the destination address 146 field of the IP header of the packet containing the query causing the 147 response. If this would cause the response to be sent from an IP 148 address that is not permitted for this purpose, then the response may 149 be sent from any legal IP address allocated to the server. That 150 address should be chosen to maximise the possibility that the client 151 will be able to use it for further queries. Servers configured in 152 such a way that not all their addresses are equally reachable from 153 all potential clients need take particular care when responding to 154 queries sent to anycast, multicast, or similar, addresses. 155 156 157 158 159 160 161 162 Elz & Bush Standards Track [Page 3] 163 RFC 2181 Clarifications to the DNS Specification July 1997 164 165 166 4.2. Port Number Selection 167 168 Replies to all queries must be directed to the port from which they 169 were sent. When queries are received via TCP this is an inherent 170 part of the transport protocol. For queries received by UDP the 171 server must take note of the source port and use that as the 172 destination port in the response. Replies should always be sent from 173 the port to which they were directed. Except in extraordinary 174 circumstances, this will be the well known port assigned for DNS 175 queries [RFC1700]. 176 177 5. Resource Record Sets 178 179 Each DNS Resource Record (RR) has a label, class, type, and data. It 180 is meaningless for two records to ever have label, class, type and 181 data all equal - servers should suppress such duplicates if 182 encountered. It is however possible for most record types to exist 183 with the same label, class and type, but with different data. Such a 184 group of records is hereby defined to be a Resource Record Set 185 (RRSet). 186 187 5.1. Sending RRs from an RRSet 188 189 A query for a specific (or non-specific) label, class, and type, will 190 always return all records in the associated RRSet - whether that be 191 one or more RRs. The response must be marked as "truncated" if the 192 entire RRSet will not fit in the response. 193 194 5.2. TTLs of RRs in an RRSet 195 196 Resource Records also have a time to live (TTL). It is possible for 197 the RRs in an RRSet to have different TTLs. No uses for this have 198 been found that cannot be better accomplished in other ways. This 199 can, however, cause partial replies (not marked "truncated") from a 200 caching server, where the TTLs for some but not all the RRs in the 201 RRSet have expired. 202 203 Consequently the use of differing TTLs in an RRSet is hereby 204 deprecated, the TTLs of all RRs in an RRSet must be the same. 205 206 Should a client receive a response containing RRs from an RRSet with 207 differing TTLs, it should treat this as an error. If the RRSet 208 concerned is from a non-authoritative source for this data, the 209 client should simply ignore the RRSet, and if the values were 210 required, seek to acquire them from an authoritative source. Clients 211 that are configured to send all queries to one, or more, particular 212 servers should treat those servers as authoritative for this purpose. 213 Should an authoritative source send such a malformed RRSet, the 214 215 216 217 Elz & Bush Standards Track [Page 4] 218 RFC 2181 Clarifications to the DNS Specification July 1997 219 220 221 client should treat the RRs for all purposes as if all TTLs in the 222 RRSet had been set to the value of the lowest TTL in the RRSet. In 223 no case may a server send an RRSet with TTLs not all equal. 224 225 5.3. DNSSEC Special Cases 226 227 Two of the record types added by DNS Security (DNSSEC) [RFC2065] 228 require special attention when considering the formation of Resource 229 Record Sets. Those are the SIG and NXT records. It should be noted 230 that DNS Security is still very new, and there is, as yet, little 231 experience with it. Readers should be prepared for the information 232 related to DNSSEC contained in this document to become outdated as 233 the DNS Security specification matures. 234 235 5.3.1. SIG records and RRSets 236 237 A SIG record provides signature (validation) data for another RRSet 238 in the DNS. Where a zone has been signed, every RRSet in the zone 239 will have had a SIG record associated with it. The data type of the 240 RRSet is included in the data of the SIG RR, to indicate with which 241 particular RRSet this SIG record is associated. Were the rules above 242 applied, whenever a SIG record was included with a response to 243 validate that response, the SIG records for all other RRSets 244 associated with the appropriate node would also need to be included. 245 In some cases, this could be a very large number of records, not 246 helped by their being rather large RRs. 247 248 Thus, it is specifically permitted for the authority section to 249 contain only those SIG RRs with the "type covered" field equal to the 250 type field of an answer being returned. However, where SIG records 251 are being returned in the answer section, in response to a query for 252 SIG records, or a query for all records associated with a name 253 (type=ANY) the entire SIG RRSet must be included, as for any other RR 254 type. 255 256 Servers that receive responses containing SIG records in the 257 authority section, or (probably incorrectly) as additional data, must 258 understand that the entire RRSet has almost certainly not been 259 included. Thus, they must not cache that SIG record in a way that 260 would permit it to be returned should a query for SIG records be 261 received at that server. RFC2065 actually requires that SIG queries 262 be directed only to authoritative servers to avoid the problems that 263 could be caused here, and while servers exist that do not understand 264 the special properties of SIG records, this will remain necessary. 265 However, careful design of SIG record processing in new 266 implementations should permit this restriction to be relaxed in the 267 future, so resolvers do not need to treat SIG record queries 268 specially. 269 270 271 272 Elz & Bush Standards Track [Page 5] 273 RFC 2181 Clarifications to the DNS Specification July 1997 274 275 276 It has been occasionally stated that a received request for a SIG 277 record should be forwarded to an authoritative server, rather than 278 being answered from data in the cache. This is not necessary - a 279 server that has the knowledge of SIG as a special case for processing 280 this way would be better to correctly cache SIG records, taking into 281 account their characteristics. Then the server can determine when it 282 is safe to reply from the cache, and when the answer is not available 283 and the query must be forwarded. 284 285 5.3.2. NXT RRs 286 287 Next Resource Records (NXT) are even more peculiar. There will only 288 ever be one NXT record in a zone for a particular label, so 289 superficially, the RRSet problem is trivial. However, at a zone cut, 290 both the parent zone, and the child zone (superzone and subzone in 291 RFC2065 terminology) will have NXT records for the same name. Those 292 two NXT records do not form an RRSet, even where both zones are 293 housed at the same server. NXT RRSets always contain just a single 294 RR. Where both NXT records are visible, two RRSets exist. However, 295 servers are not required to treat this as a special case when 296 receiving NXT records in a response. They may elect to notice the 297 existence of two different NXT RRSets, and treat that as they would 298 two different RRSets of any other type. That is, cache one, and 299 ignore the other. Security aware servers will need to correctly 300 process the NXT record in the received response though. 301 302 5.4. Receiving RRSets 303 304 Servers must never merge RRs from a response with RRs in their cache 305 to form an RRSet. If a response contains data that would form an 306 RRSet with data in a server's cache the server must either ignore the 307 RRs in the response, or discard the entire RRSet currently in the 308 cache, as appropriate. Consequently the issue of TTLs varying 309 between the cache and a response does not cause concern, one will be 310 ignored. That is, one of the data sets is always incorrect if the 311 data from an answer differs from the data in the cache. The 312 challenge for the server is to determine which of the data sets is 313 correct, if one is, and retain that, while ignoring the other. Note 314 that if a server receives an answer containing an RRSet that is 315 identical to that in its cache, with the possible exception of the 316 TTL value, it may, optionally, update the TTL in its cache with the 317 TTL of the received answer. It should do this if the received answer 318 would be considered more authoritative (as discussed in the next 319 section) than the previously cached answer. 320 321 322 323 324 325 326 327 Elz & Bush Standards Track [Page 6] 328 RFC 2181 Clarifications to the DNS Specification July 1997 329 330 331 5.4.1. Ranking data 332 333 When considering whether to accept an RRSet in a reply, or retain an 334 RRSet already in its cache instead, a server should consider the 335 relative likely trustworthiness of the various data. An 336 authoritative answer from a reply should replace cached data that had 337 been obtained from additional information in an earlier reply. 338 However additional information from a reply will be ignored if the 339 cache contains data from an authoritative answer or a zone file. 340 341 The accuracy of data available is assumed from its source. 342 Trustworthiness shall be, in order from most to least: 343 344 + Data from a primary zone file, other than glue data, 345 + Data from a zone transfer, other than glue, 346 + The authoritative data included in the answer section of an 347 authoritative reply. 348 + Data from the authority section of an authoritative answer, 349 + Glue from a primary zone, or glue from a zone transfer, 350 + Data from the answer section of a non-authoritative answer, and 351 non-authoritative data from the answer section of authoritative 352 answers, 353 + Additional information from an authoritative answer, 354 Data from the authority section of a non-authoritative answer, 355 Additional information from non-authoritative answers. 356 357 Note that the answer section of an authoritative answer normally 358 contains only authoritative data. However when the name sought is an 359 alias (see section 10.1.1) only the record describing that alias is 360 necessarily authoritative. Clients should assume that other records 361 may have come from the server's cache. Where authoritative answers 362 are required, the client should query again, using the canonical name 363 associated with the alias. 364 365 Unauthenticated RRs received and cached from the least trustworthy of 366 those groupings, that is data from the additional data section, and 367 data from the authority section of a non-authoritative answer, should 368 not be cached in such a way that they would ever be returned as 369 answers to a received query. They may be returned as additional 370 information where appropriate. Ignoring this would allow the 371 trustworthiness of relatively untrustworthy data to be increased 372 without cause or excuse. 373 374 When DNS security [RFC2065] is in use, and an authenticated reply has 375 been received and verified, the data thus authenticated shall be 376 considered more trustworthy than unauthenticated data of the same 377 type. Note that throughout this document, "authoritative" means a 378 reply with the AA bit set. DNSSEC uses trusted chains of SIG and KEY 379 380 381 382 Elz & Bush Standards Track [Page 7] 383 RFC 2181 Clarifications to the DNS Specification July 1997 384 385 386 records to determine the authenticity of data, the AA bit is almost 387 irrelevant. However DNSSEC aware servers must still correctly set 388 the AA bit in responses to enable correct operation with servers that 389 are not security aware (almost all currently). 390 391 Note that, glue excluded, it is impossible for data from two 392 correctly configured primary zone files, two correctly configured 393 secondary zones (data from zone transfers) or data from correctly 394 configured primary and secondary zones to ever conflict. Where glue 395 for the same name exists in multiple zones, and differs in value, the 396 nameserver should select data from a primary zone file in preference 397 to secondary, but otherwise may choose any single set of such data. 398 Choosing that which appears to come from a source nearer the 399 authoritative data source may make sense where that can be 400 determined. Choosing primary data over secondary allows the source 401 of incorrect glue data to be discovered more readily, when a problem 402 with such data exists. Where a server can detect from two zone files 403 that one or more are incorrectly configured, so as to create 404 conflicts, it should refuse to load the zones determined to be 405 erroneous, and issue suitable diagnostics. 406 407 "Glue" above includes any record in a zone file that is not properly 408 part of that zone, including nameserver records of delegated sub- 409 zones (NS records), address records that accompany those NS records 410 (A, AAAA, etc), and any other stray data that might appear. 411 412 5.5. Sending RRSets (reprise) 413 414 A Resource Record Set should only be included once in any DNS reply. 415 It may occur in any of the Answer, Authority, or Additional 416 Information sections, as required. However it should not be repeated 417 in the same, or any other, section, except where explicitly required 418 by a specification. For example, an AXFR response requires the SOA 419 record (always an RRSet containing a single RR) be both the first and 420 last record of the reply. Where duplicates are required this way, 421 the TTL transmitted in each case must be the same. 422 423 6. Zone Cuts 424 425 The DNS tree is divided into "zones", which are collections of 426 domains that are treated as a unit for certain management purposes. 427 Zones are delimited by "zone cuts". Each zone cut separates a 428 "child" zone (below the cut) from a "parent" zone (above the cut). 429 The domain name that appears at the top of a zone (just below the cut 430 that separates the zone from its parent) is called the zone's 431 "origin". The name of the zone is the same as the name of the domain 432 at the zone's origin. Each zone comprises that subset of the DNS 433 tree that is at or below the zone's origin, and that is above the 434 435 436 437 Elz & Bush Standards Track [Page 8] 438 RFC 2181 Clarifications to the DNS Specification July 1997 439 440 441 cuts that separate the zone from its children (if any). The 442 existence of a zone cut is indicated in the parent zone by the 443 existence of NS records specifying the origin of the child zone. A 444 child zone does not contain any explicit reference to its parent. 445 446 6.1. Zone authority 447 448 The authoritative servers for a zone are enumerated in the NS records 449 for the origin of the zone, which, along with a Start of Authority 450 (SOA) record are the mandatory records in every zone. Such a server 451 is authoritative for all resource records in a zone that are not in 452 another zone. The NS records that indicate a zone cut are the 453 property of the child zone created, as are any other records for the 454 origin of that child zone, or any sub-domains of it. A server for a 455 zone should not return authoritative answers for queries related to 456 names in another zone, which includes the NS, and perhaps A, records 457 at a zone cut, unless it also happens to be a server for the other 458 zone. 459 460 Other than the DNSSEC cases mentioned immediately below, servers 461 should ignore data other than NS records, and necessary A records to 462 locate the servers listed in the NS records, that may happen to be 463 configured in a zone at a zone cut. 464 465 6.2. DNSSEC issues 466 467 The DNS security mechanisms [RFC2065] complicate this somewhat, as 468 some of the new resource record types added are very unusual when 469 compared with other DNS RRs. In particular the NXT ("next") RR type 470 contains information about which names exist in a zone, and hence 471 which do not, and thus must necessarily relate to the zone in which 472 it exists. The same domain name may have different NXT records in 473 the parent zone and the child zone, and both are valid, and are not 474 an RRSet. See also section 5.3.2. 475 476 Since NXT records are intended to be automatically generated, rather 477 than configured by DNS operators, servers may, but are not required 478 to, retain all differing NXT records they receive regardless of the 479 rules in section 5.4. 480 481 For a secure parent zone to securely indicate that a subzone is 482 insecure, DNSSEC requires that a KEY RR indicating that the subzone 483 is insecure, and the parent zone's authenticating SIG RR(s) be 484 present in the parent zone, as they by definition cannot be in the 485 subzone. Where a subzone is secure, the KEY and SIG records will be 486 present, and authoritative, in that zone, but should also always be 487 present in the parent zone (if secure). 488 489 490 491 492 Elz & Bush Standards Track [Page 9] 493 RFC 2181 Clarifications to the DNS Specification July 1997 494 495 496 Note that in none of these cases should a server for the parent zone, 497 not also being a server for the subzone, set the AA bit in any 498 response for a label at a zone cut. 499 500 7. SOA RRs 501 502 Three minor issues concerning the Start of Zone of Authority (SOA) 503 Resource Record need some clarification. 504 505 7.1. Placement of SOA RRs in authoritative answers 506 507 RFC1034, in section 3.7, indicates that the authority section of an 508 authoritative answer may contain the SOA record for the zone from 509 which the answer was obtained. When discussing negative caching, 510 RFC1034 section 4.3.4 refers to this technique but mentions the 511 additional section of the response. The former is correct, as is 512 implied by the example shown in section 6.2.5 of RFC1034. SOA 513 records, if added, are to be placed in the authority section. 514 515 7.2. TTLs on SOA RRs 516 517 It may be observed that in section 3.2.1 of RFC1035, which defines 518 the format of a Resource Record, that the definition of the TTL field 519 contains a throw away line which states that the TTL of an SOA record 520 should always be sent as zero to prevent caching. This is mentioned 521 nowhere else, and has not generally been implemented. 522 Implementations should not assume that SOA records will have a TTL of 523 zero, nor are they required to send SOA records with a TTL of zero. 524 525 7.3. The SOA.MNAME field 526 527 It is quite clear in the specifications, yet seems to have been 528 widely ignored, that the MNAME field of the SOA record should contain 529 the name of the primary (master) server for the zone identified by 530 the SOA. It should not contain the name of the zone itself. That 531 information would be useless, as to discover it, one needs to start 532 with the domain name of the SOA record - that is the name of the 533 zone. 534
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).
535 8. Time to Live (TTL) 536 537 The definition of values appropriate to the TTL field in STD 13 is 538 not as clear as it could be, with respect to how many significant 539 bits exist, and whether the value is signed or unsigned. It is 540 hereby specified that a TTL value is an unsigned number, with a 541 minimum value of 0, and a maximum value of 2147483647. That is, a 542 maximum of 2^31 - 1. When transmitted, this value shall be encoded 543 in the less significant 31 bits of the 32 bit TTL field, with the 544 545 546 547 Elz & Bush Standards Track [Page 10] 548 RFC 2181 Clarifications to the DNS Specification July 1997 549 550 551 most significant, or sign, bit set to zero. 552 553 Implementations should treat TTL values received with the most 554 significant bit set as if the entire value received was zero. 555 556 Implementations are always free to place an upper bound on any TTL 557 received, and treat any larger values as if they were that upper 558 bound. The TTL specifies a maximum time to live, not a mandatory 559 time to live. 560 561 9. The TC (truncated) header bit 562 563 The TC bit should be set in responses only when an RRSet is required 564 as a part of the response, but could not be included in its entirety. 565 The TC bit should not be set merely because some extra information 566 could have been included, but there was insufficient room. This 567 includes the results of additional section processing. In such cases 568 the entire RRSet that will not fit in the response should be omitted, 569 and the reply sent as is, with the TC bit clear. If the recipient of 570 the reply needs the omitted data, it can construct a query for that 571 data and send that separately. 572 573 Where TC is set, the partial RRSet that would not completely fit may 574 be left in the response. When a DNS client receives a reply with TC 575 set, it should ignore that response, and query again, using a 576 mechanism, such as a TCP connection, that will permit larger replies. 577 578 10. Naming issues 579 580 It has sometimes been inferred from some sections of the DNS 581 specification [RFC1034, RFC1035] that a host, or perhaps an interface 582 of a host, is permitted exactly one authoritative, or official, name, 583 called the canonical name. There is no such requirement in the DNS. 584 585 10.1. CNAME resource records 586 587 The DNS CNAME ("canonical name") record exists to provide the 588 canonical name associated with an alias name. There may be only one 589 such canonical name for any one alias. That name should generally be 590 a name that exists elsewhere in the DNS, though there are some rare 591 applications for aliases with the accompanying canonical name 592 undefined in the DNS. An alias name (label of a CNAME record) may, 593 if DNSSEC is in use, have SIG, NXT, and KEY RRs, but may have no 594 other data. That is, for any label in the DNS (any domain name) 595 exactly one of the following is true: 596 597 598 599 600 601 602 Elz & Bush Standards Track [Page 11] 603 RFC 2181 Clarifications to the DNS Specification July 1997 604 605 606 + one CNAME record exists, optionally accompanied by SIG, NXT, and 607 KEY RRs, 608 + one or more records exist, none being CNAME records, 609 + the name exists, but has no associated RRs of any type, 610 + the name does not exist at all. 611 612 10.1.1. CNAME terminology 613 614 It has been traditional to refer to the label of a CNAME record as "a 615 CNAME". This is unfortunate, as "CNAME" is an abbreviation of 616 "canonical name", and the label of a CNAME record is most certainly 617 not a canonical name. It is, however, an entrenched usage. Care 618 must therefore be taken to be very clear whether the label, or the 619 value (the canonical name) of a CNAME resource record is intended. 620 In this document, the label of a CNAME resource record will always be 621 referred to as an alias. 622 623 10.2. PTR records 624 625 Confusion about canonical names has lead to a belief that a PTR 626 record should have exactly one RR in its RRSet. This is incorrect, 627 the relevant section of RFC1034 (section 3.6.2) indicates that the 628 value of a PTR record should be a canonical name. That is, it should 629 not be an alias. There is no implication in that section that only 630 one PTR record is permitted for a name. No such restriction should 631 be inferred. 632 633 Note that while the value of a PTR record must not be an alias, there 634 is no requirement that the process of resolving a PTR record not 635 encounter any aliases. The label that is being looked up for a PTR 636 value might have a CNAME record. That is, it might be an alias. The 637 value of that CNAME RR, if not another alias, which it should not be, 638 will give the location where the PTR record is found. That record 639 gives the result of the PTR type lookup. This final result, the 640 value of the PTR RR, is the label which must not be an alias. 641 642 10.3. MX and NS records 643 644 The domain name used as the value of a NS resource record, or part of 645 the value of a MX resource record must not be an alias. Not only is 646 the specification clear on this point, but using an alias in either 647 of these positions neither works as well as might be hoped, nor well 648 fulfills the ambition that may have led to this approach. This 649 domain name must have as its value one or more address records. 650 Currently those will be A records, however in the future other record 651 types giving addressing information may be acceptable. It can also 652 have other RRs, but never a CNAME RR. 653 654 655 656 657 Elz & Bush Standards Track [Page 12] 658 RFC 2181 Clarifications to the DNS Specification July 1997 659 660 661 Searching for either NS or MX records causes "additional section 662 processing" in which address records associated with the value of the 663 record sought are appended to the answer. This helps avoid needless 664 extra queries that are easily anticipated when the first was made. 665 666 Additional section processing does not include CNAME records, let 667 alone the address records that may be associated with the canonical 668 name derived from the alias. Thus, if an alias is used as the value 669 of an NS or MX record, no address will be returned with the NS or MX 670 value. This can cause extra queries, and extra network burden, on 671 every query. It is trivial for the DNS administrator to avoid this 672 by resolving the alias and placing the canonical name directly in the 673 affected record just once when it is updated or installed. In some 674 particular hard cases the lack of the additional section address 675 records in the results of a NS lookup can cause the request to fail. 676 677 11. Name syntax 678 679 Occasionally it is assumed that the Domain Name System serves only 680 the purpose of mapping Internet host names to data, and mapping 681 Internet addresses to host names. This is not correct, the DNS is a 682 general (if somewhat limited) hierarchical database, and can store 683 almost any kind of data, for almost any purpose. 684 685 The DNS itself places only one restriction on the particular labels 686 that can be used to identify resource records. That one restriction 687 relates to the length of the label and the full name. The length of 688 any one label is limited to between 1 and 63 octets. A full domain 689 name is limited to 255 octets (including the separators). The zero 690 length full name is defined as representing the root of the DNS tree, 691 and is typically written and displayed as ".". Those restrictions 692 aside, any binary string whatever can be used as the label of any 693 resource record. Similarly, any binary string can serve as the value 694 of any record that includes a domain name as some or all of its value 695 (SOA, NS, MX, PTR, CNAME, and any others that may be added). 696 Implementations of the DNS protocols must not place any restrictions 697 on the labels that can be used. In particular, DNS servers must not 698 refuse to serve a zone because it contains labels that might not be 699 acceptable to some DNS client programs. A DNS server may be 700 configurable to issue warnings when loading, or even to refuse to 701 load, a primary zone containing labels that might be considered 702 questionable, however this should not happen by default. 703 704 Note however, that the various applications that make use of DNS data 705 can have restrictions imposed on what particular values are 706 acceptable in their environment. For example, that any binary label 707 can have an MX record does not imply that any binary name can be used 708 as the host part of an e-mail address. Clients of the DNS can impose 709 710 711 712 Elz & Bush Standards Track [Page 13] 713 RFC 2181 Clarifications to the DNS Specification July 1997 714 715 716 whatever restrictions are appropriate to their circumstances on the 717 values they use as keys for DNS lookup requests, and on the values 718 returned by the DNS. If the client has such restrictions, it is 719 solely responsible for validating the data from the DNS to ensure 720 that it conforms before it makes any use of that data. 721 722 See also [RFC1123] section 188.8.131.52. 723 724 12. Security Considerations 725 726 This document does not consider security. 727 728 In particular, nothing in section 4 is any way related to, or useful 729 for, any security related purposes. 730 731 Section 5.4.1 is also not related to security. Security of DNS data 732 will be obtained by the Secure DNS [RFC2065], which is mostly 733 orthogonal to this memo. 734 735 It is not believed that anything in this document adds to any 736 security issues that may exist with the DNS, nor does it do anything 737 to that will necessarily lessen them. Correct implementation of the 738 clarifications in this document might play some small part in 739 limiting the spread of non-malicious bad data in the DNS, but only 740 DNSSEC can help with deliberate attempts to subvert DNS data. 741 742 13. References 743 744 [RFC1034] Mockapetris, P., "Domain Names - Concepts and Facilities", 745 STD 13, RFC 1034, November 1987. 746 747 [RFC1035] Mockapetris, P., "Domain Names - Implementation and 748 Specification", STD 13, RFC 1035, November 1987. 749 750 [RFC1123] Braden, R., "Requirements for Internet Hosts - application 751 and support", STD 3, RFC 1123, January 1989. 752 753 [RFC1700] Reynolds, J., Postel, J., "Assigned Numbers", 754 STD 2, RFC 1700, October 1994. 755 756 [RFC2065] Eastlake, D., Kaufman, C., "Domain Name System Security 757 Extensions", RFC 2065, January 1997. 758 759 760 761 762 763 764 765 766 767 Elz & Bush Standards Track [Page 14] 768 RFC 2181 Clarifications to the DNS Specification July 1997 769 770 771 14. Acknowledgements 772 773 This memo arose from discussions in the DNSIND working group of the 774 IETF in 1995 and 1996, the members of that working group are largely 775 responsible for the ideas captured herein. Particular thanks to 776 Donald E. Eastlake, 3rd, and Olafur Gudmundsson, for help with the 777 DNSSEC issues in this document, and to John Gilmore for pointing out 778 where the clarifications were not necessarily clarifying. Bob Halley 779 suggested clarifying the placement of SOA records in authoritative 780 answers, and provided the references. Michael Patton, as usual, and 781 Mark Andrews, Alan Barrett and Stan Barber provided much assistance 782 with many details. Josh Littlefield helped make sure that the 783 clarifications didn't cause problems in some irritating corner cases. 784 785 15. Authors' Addresses 786 787 Robert Elz 788 Computer Science 789 University of Melbourne 790 Parkville, Victoria, 3052 791 Australia. 792 793 EMail: kre@munnari.OZ.AU 794 795 796 Randy Bush 797 RGnet, Inc. 798 5147 Crystal Springs Drive NE 799 Bainbridge Island, Washington, 98110 800 United States. 801 802 EMail: email@example.com 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 Elz & Bush Standards Track [Page 15] 823
All of RFC8767 touches on aspects of how TTLs are calculated and used by resolvers. In particular, RFC 8767 says that it "updates RFC 2181 by interpreting values with the high-order bit set as being positive, rather than 0, and suggests a cap of 7 days".
The last paragraph of Section 3 in RFC4034 updates
the TTL rules, specifically for RRSIG records. It says:
The TTL value of an RRSIG RR MUST match the TTL value of the RRset it covers. This is an exception to the [RFC2181] rules for TTL values of individual RRs within a RRset: individual RRSIG RRs with the same owner name will have different TTL values if the RRsets they cover have different TTL values.
This is also mentioned briefly in RFC4035.