idnits 2.17.1 draft-ietf-dnsop-kskroll-sentinel-15.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == There are 1 instance of lines with non-RFC2606-compliant FQDNs in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (July 02, 2018) is 2097 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) No issues found here. Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 DNSOP G. Huston 3 Internet-Draft J. Damas 4 Intended status: Standards Track APNIC 5 Expires: January 3, 2019 W. Kumari 6 Google 7 July 02, 2018 9 A Root Key Trust Anchor Sentinel for DNSSEC 10 draft-ietf-dnsop-kskroll-sentinel-15 12 Abstract 14 The DNS Security Extensions (DNSSEC) were developed to provide origin 15 authentication and integrity protection for DNS data by using digital 16 signatures. These digital signatures can be verified by building a 17 chain of trust starting from a trust anchor and proceeding down to a 18 particular node in the DNS. This document specifies a mechanism that 19 will allow an end user and third parties to determine the trusted key 20 state for the root key of the resolvers that handle that user's DNS 21 queries. Note that this method is only applicable for determining 22 which keys are in the trust store for the root key. 24 [ This document is being collaborated on in Github at: 25 https://github.com/APNIC-Labs/draft-kskroll-sentinel. The most 26 recent version of the document, open issues, etc should all be 27 available here. The authors (gratefully) accept pull requests. RFC 28 Editor, please remove text in square brackets before publication. ] 30 Status of This Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at https://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on January 3, 2019. 47 Copyright Notice 49 Copyright (c) 2018 IETF Trust and the persons identified as the 50 document authors. All rights reserved. 52 This document is subject to BCP 78 and the IETF Trust's Legal 53 Provisions Relating to IETF Documents 54 (https://trustee.ietf.org/license-info) in effect on the date of 55 publication of this document. Please review these documents 56 carefully, as they describe your rights and restrictions with respect 57 to this document. Code Components extracted from this document must 58 include Simplified BSD License text as described in Section 4.e of 59 the Trust Legal Provisions and are provided without warranty as 60 described in the Simplified BSD License. 62 Table of Contents 64 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 65 1.1. Terminology . . . . . . . . . . . . . . . . . . . . . . . 4 66 2. Sentinel Mechanism in Resolvers . . . . . . . . . . . . . . . 4 67 2.1. Preconditions . . . . . . . . . . . . . . . . . . . . . . 4 68 2.2. Special Processing . . . . . . . . . . . . . . . . . . . 5 69 3. Sentinel Tests for a Single DNS Resolver . . . . . . . . . . 6 70 3.1. Forwarders . . . . . . . . . . . . . . . . . . . . . . . 8 71 4. Sentinel Tests from Hosts with More than One Configured 72 Resolve . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 73 4.1. Test Scenario and Objective . . . . . . . . . . . . . . . 9 74 4.2. Test Assumptions . . . . . . . . . . . . . . . . . . . . 10 75 4.3. Test Procedure . . . . . . . . . . . . . . . . . . . . . 10 76 5. Security Considerations . . . . . . . . . . . . . . . . . . . 12 77 6. Privacy Considerations . . . . . . . . . . . . . . . . . . . 12 78 7. Implementation Experience . . . . . . . . . . . . . . . . . . 12 79 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 80 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 13 81 10. Change Log . . . . . . . . . . . . . . . . . . . . . . . . . 14 82 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 17 83 11.1. Normative References . . . . . . . . . . . . . . . . . . 18 84 11.2. Informative References . . . . . . . . . . . . . . . . . 18 85 Appendix A. Protocol Walkthrough Example . . . . . . . . . . . . 18 86 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 21 88 1. Introduction 90 The DNS Security Extensions (DNSSEC) [RFC4033], [RFC4034] and 91 [RFC4035] were developed to provide origin authentication and 92 integrity protection for DNS data by using digital signatures. 93 DNSSEC uses Key Tags to efficiently match signatures to the keys from 94 which they are generated. The Key Tag is a 16-bit value computed 95 from the RDATA of a DNSKEY RR as described in Appendix B of 96 [RFC4034]. RRSIG RRs contain a Key Tag field whose value is equal to 97 the Key Tag of the DNSKEY RR that was used to generate the 98 corresponding signature. 100 This document specifies how security-aware DNS resolvers that perform 101 validation of their responses can respond to certain queries in a 102 manner that allows an agent performing the queries to deduce whether 103 a particular key for the root has been loaded into that resolver's 104 trusted key store. This document also describes a procedure where a 105 collection of resolvers can be tested to determine if at least one of 106 these resolvers has loaded a given key into its trusted key store. 107 These tests can be used to determine whether a certain root zone Key 108 Signing Key (KSK) is ready to be used as a trusted key, within the 109 context of a planned root zone KSK key roll. 111 There are two primary use cases for this mechanism: 113 o Users may wish to ascertain whether their DNS resolution 114 environment resolvers is ready for an upcoming root KSK rollover. 116 o Researchers want to perform Internet-wide studies about the 117 proportion of users who will be negatively impacted an upcoming 118 root KSK rollover. 120 The mechanism described in this document satisfy the requirements of 121 both these use-cases. This mechanism is OPTIONAL to implement and 122 use. If implemented, this mechanism SHOULD be enabled by default to 123 facilitate Internet-wide measurement. Configuration options MAY be 124 provided to disable the mechanism for reasons of local policy. 126 The KSK sentinel tests described in this document use a test 127 comprising of a set of DNS queries to domain names that have special 128 values for the left-most label. The test relies on recursive 129 resolvers supporting a mechanism that recognises this special name 130 pattern in queries, and under certain defined circumstances will 131 return a DNS SERVFAIL response code (RCODE 2), mimicking the response 132 code that is returned by security-aware resolvers when DNSSEC 133 validation fails. 135 If a browser or operating system is configured with multiple 136 resolvers, and those resolvers have different properties (for 137 example, one performs DNSSEC validation and one does not), the 138 sentinel test described in this document can still be used, but it 139 makes a number of assumptions about DNS resolution behaviour that may 140 not necessarily hold in all environments. If these assumptions do 141 not hold (such as, for example, requiring the stub resolver to query 142 the next recursive resolver in the locally configured set upon 143 receipt of a SERVFAIL response code) then this test may produce 144 indeterminate or inconsistent results. In some cases where these 145 assumptions do not hold, repeating the same test query set may 146 generate different results. 148 Note that the measurements facilitated by the mechanism described in 149 this document are different from those of [RFC8145]. RFC 8145 relies 150 on resolvers reporting towards the root servers a list of locally 151 cached trust anchors for the root zone. Those reports can be used to 152 infer how many resolvers may be impacted by a KSK roll, but not what 153 the user impact of the KSK roll will be. 155 1.1. Terminology 157 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 158 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 159 document are to be interpreted as described in RFC 2119. 161 2. Sentinel Mechanism in Resolvers 163 DNSSEC-Validating resolvers that implement this mechanism MUST 164 perform validation of responses in accordance with the DNSSEC 165 response validation specification [RFC4035]. 167 This sentinel mechanism makes use of two special labels: 169 o root-key-sentinel-is-ta- 171 o root-key-sentinel-not-ta- 173 These labels trigger special processing in the validating DNS 174 resolver when responses from authoritative servers are received. 175 Labels containing "root-key-sentinel-is-ta-" is used to 176 answer the question "Is this the Key Tag of a key which the 177 validating DNS resolver is currently trusting as a trust anchor?" 178 Labels containing "root-key-sentinel-not-ta-" is used to 179 answer the question "Is this the Key Tag of a key which the 180 validating DNS resolver is *not* currently trusting as a trust 181 anchor?" 183 2.1. Preconditions 185 All of the following conditions must be met to trigger special 186 processing inside resolver code: 188 o The DNS response is DNSSEC validated. 190 o The result of validation is "Secure". 192 o The Checking Disabled (CD) bit in the query is not set. 194 o The QTYPE is either A or AAAA (Query Type value 1 or 28). 196 o The OPCODE is QUERY. 198 o The leftmost label of the original QNAME (the name sent in the 199 Question Section in the original query) is either "root-key- 200 sentinel-is-ta-" or "root-key-sentinel-not-ta-". 202 If any one of the preconditions is not met, the resolver MUST NOT 203 alter the DNS response based on the mechanism in this document. 205 Note that the is specified in the DNS label as unsigned 206 decimal integer (as described in [RFC4034], section 5.3), but zero- 207 padded to five digits (for example, a Key Tag value of 42 would be 208 represented in the label as 00042). The precise specification of the 209 special labels above should be followed exactly. For example, a 210 label that does not include a Key Tag zero-padded to five digits does 211 not match this specification, and should not be processed as if they 212 did -- in other words, such queries should be handled as any other 213 label and not according to Section 2.2. 215 2.2. Special Processing 217 Responses which fulfil all of the preconditions in Section 2.1 218 require special processing, depending on leftmost label in the QNAME. 220 First, the resolver determines if the numerical value of is 221 equal to any of the Key Tag values of an active root zone KSK which 222 is currently trusted by the local resolver and is stored in its store 223 of trusted keys. An active root zone KSK is one which could 224 currently be used for validation (that is, a key that is not in 225 either the AddPend or Revoked state as described in [RFC5011]). 227 Second, the resolver alters the response being sent to the original 228 query based on both the left-most label and the presence of a key 229 with given Key Tag in the trust anchor store. Two labels and two 230 possible states of the corresponding key generate four possible 231 combinations summarized in the table: 233 Label | Key is trusted | Key is not trusted 234 ------------------------------------------------------------------ 235 is-ta | return original answer | return SERVFAIL 236 not-ta | return SERVFAIL | return original answer 238 Instruction "return SERVFAIL" means that the resolver MUST set 239 RCODE=SERVFAIL (value 2) and the ANSWER section of the DNS response 240 MUST be empty, ignoring all other documents which specify content of 241 the ANSWER section. 243 Instruction "return original answer" means that the resolver MUST 244 process the query without any further special processing; that is, 245 exactly as if the mechanism described in this document was not 246 implemented or disabled. 248 3. Sentinel Tests for a Single DNS Resolver 250 This section describes the use of the sentinel detection mechanism 251 against a single DNS recursive resolver in order to determine whether 252 this resolver is using a particular trust anchor to validate DNSSEC- 253 signed responses. 255 Note that the test in this section applies to a single DNS resolver. 256 The test described in Section 4 applies instead to a collection of 257 DNS resolvers, as might be found in the DNS configuration of an end- 258 user environment. 260 The critical aspect of the DNS names used in this mechanism is that 261 they contain the specified label for either the positive and negative 262 test as the left-most label in the query name. 264 The sentinel detection procedure can test a DNS resolver using three 265 queries: 267 o A query name containing the left-most label "root-key-sentinel-is- 268 ta-". This corresponds to a a validly-signed RRset in 269 the zone, so that responses associated with queried names in this 270 zone can be authenticated by a DNSSEC-validating resolver. Any 271 validly-signed DNS zone can be used for this test. 273 o A query name containing the left-most label "root-key-sentinel- 274 not-ta-". This is also a validly-signed name. Any 275 validly-signed DNS zone can be used for this test. 277 o A query name that is signed with a DNSSEC signature that cannot be 278 validated (described as a "bogus" RRset in Section 5 of [RFC4033], 279 when, for example, an RRset is not signed with a valid RRSIG 280 record). 282 The responses received from queries to resolve each of these names 283 can be evaluated to infer a trust key state of the DNS resolver. 285 An essential assumption here is that this technique relies on 286 security-aware (DNSSEC validating) resolvers responding with a 287 SERVFAIL response code to queries where DNSSEC checking is requested 288 and the response cannot be validated. Note that a slew of other 289 issues can also cause SERVFAIL responses, and so the sentinel 290 processing may sometimes result in incorrect or indeterminate 291 conclusions. 293 To describe this process of classification, DNS resolvers are 294 classified by five distinct behavior types using the labels: "Vnew", 295 "Vold", "Vind", "nonV", and "other". These labels correspond to 296 resolver system behaviour types as follows: 298 Vnew: A DNS resolver that is configured to implement this mechanism 299 and has loaded the nominated key into their local trusted key 300 stores will respond with an A or AAAA RRset response for the 301 associated "root-key-sentinel-is-ta" queries, SERVFAIL for "root- 302 key-sentinel-not-ta" queries and SERVFAIL for the signed name 303 queries that return "bogus" validation status. 305 Vold: A DNS resolver that is configured to implement this mechanism 306 and has not loaded the nominated key into their local trusted key 307 stores will respond with an SERVFAIL for the associated "root-key- 308 sentinel-is-ta" queries, an A or AAAA RRset response for "root- 309 key-sentinel-not-ta" queries and SERVFAIL for the signed name 310 queries that return "bogus" validation status. 312 Vind: A DNS resolver that has is not configured to implement this 313 mechanism will respond with an A or AAAA RRset response for "root- 314 key-sentinel-is-ta", an A or AAAA RRset response for "root-key- 315 sentinel-not-ta" and SERVFAIL for the name that returns "bogus" 316 validation status. This set of responses does not give any 317 information about the trust anchors used by this resolver. 319 nonV: A non-security-aware DNS resolver will respond with an A or 320 AAAA record response for "root-key-sentinel-is-ta", an A record 321 response for "root-key-sentinel-not-ta" and an A or AAAA RRset 322 response for the name that returns "bogus" validation status. 324 other: There is the potential to admit other combinations of 325 responses to these three queries. While this may appear self- 326 contradictory, there are cases where such an outcome is possible. 327 For example, in DNS resolver farms what appears to be a single DNS 328 resolver that responds to queries passed to a single IP address is 329 in fact constructed as a a collection of slave resolvers, and the 330 query is passed to one of these internal resolver engines. If 331 these individual slave resolvers in the farm do not behave 332 identically, then other sets of results can be expected from these 333 three queries. In such a case, no determination about the 334 capabilities of this DNS resolver farm can be made. 336 Note that SERVFAIL might be cached according to Section 7 of 337 [RFC2308] for up to 5 minutes and a positive answer for up to its 338 TTL. 340 If a client directs these three queries to a single resolver, the 341 responses should allow the client to determine the capability of the 342 resolver, and if it supports this sentinel mechanism, whether or not 343 it has a particular key in its trust anchor store, as in the 344 following table: 346 Query 347 +----------+-----------+------------+ 348 | is-ta | not-ta | bogus | 349 +-------+----------+-----------+------------+ 350 | Vnew | A | SERVFAIL | SERVFAIL | 351 | Vold | SERVFAIL | A | SERVFAIL | 352 Type | Vind | A | A | SERVFAIL | 353 | nonV | A | A | A | 354 | other | * | * | * | 355 +-------+----------+-----------+------------+ 357 Vnew: The nominated key is trusted by the resolver. 359 Vold: The nominated key is not yet trusted by the resolver. 361 Vind: There is no information about the trust anchors of the 362 resolver. 364 nonV: The resolver does not perform DNSSEC validation. 366 other: The properties of the resolver cannot be analyzed by this 367 protocol. 369 3.1. Forwarders 371 Some resolvers are configured not to answer queries using the 372 recursive algorithm first described in [RFC1034] section 4.3.2, but 373 instead relay queries to one or more other resolvers. Resolvers 374 configured in this manner are referred to in this document as 375 "forwarders". 377 If the resolver is non-validating, and it has a single forwarder, 378 then the resolver will presumably mirror the capabilities of the 379 forwarder target resolver. 381 If the validating resolver has a forwarding configuration, and uses 382 the CD bit on all forwarded queries, then this resolver is acting in 383 a manner that is identical to a standalone resolver. 385 A more complex case is where all of the following conditions hold: 387 o Both the validating resolver and the forwarder target resolver 388 support this trusted key sentinel mechanism 390 o The local resolver's queries do not have the CD bit set 392 o The trusted key state differs between the forwarding resolver and 393 the forwarder target resolver 395 In such a case, either the outcome is indeterminate validating 396 ("Vind"), or a case of mixed signals such as SERVFAIL in all three 397 responses, ("other") which is similarly an indeterminate response 398 with respect to the trusted key state. 400 4. Sentinel Tests from Hosts with More than One Configured Resolve 402 The description in Section 3 describes a trust anchor test that can 403 be used in the simple situation where the test queries were being 404 passed to a single recursive resolver that directly queries 405 authoritative name servers. 407 However, the common end-user scenario is where a user's local DNS 408 resolution environment is configured to use more than one recursive 409 resolver. The single resolver test technique will not function 410 reliably in such cases, as a a SERVFAIL response from one resolver 411 may cause the local stub resolver to repeat the query against one of 412 the other configured resolvers and the results may be inconclusive. 414 In describing a test procedure that can be used in this environment 415 of a set of DNS resolvers there are some necessary changes to the 416 nature of the question that this test can answer, the assumptions 417 about the behaviour of the DNS resolution environment, and some 418 further observations about potential variability in the test 419 outcomes. 421 4.1. Test Scenario and Objective 423 This test is not intended to expose which trust anchors are used by 424 any single DNS resolver. 426 The test scenario is explicitly restricted to that of the KSK 427 environment where a current active KSK (called "KSK-current") is to 428 be replaced with a new KSK (called "KSK-new"). The test is designed 429 to be run between when KSK-new is introduced into the root zone and 430 when the root zone is signed with KSK-new. 432 The objective of the test is to determine if the user will be 433 negatively impacted by the KSK roll. A "negative impact" for the 434 user is defined such that all the configured resolvers are security- 435 aware resolvers that perform validation of DNSSEC-signed responses, 436 and none of these resolvers have loaded KSK-new into their local 437 trust anchor set. In this situation, it is anticipated that once the 438 KSK is rolled the entire set of the user's resolvers will not be able 439 to validate the contents of the root zone and the user is likely to 440 lose DNS service as a result of this inability to perform successful 441 DNSSEC validation. 443 4.2. Test Assumptions 445 There are a number of assumptions about the DNS environment used in 446 this test. Where these assumptions do not hold, the results of the 447 test will be indeterminate. 449 o When a recursive resolver returns SERVFAIL to the user's stub 450 resolver, the stub resolver will send the same query to the next 451 resolver in the locally configured resolver set. It will continue 452 to do this until it gets a non-SERVFAIL response or until it runs 453 out of resolvers to try. 455 o When the user's stub resolver passes a query to a resolver in the 456 configured resolver set, it will get a consistent answer over the 457 timeframe of the queries. This assumption implies that if the 458 same query is asked by the same stub resolver multiple times in 459 succession to the same recursive resolver, the recursive 460 resolver's response will be the same for each of these queries. 462 o All DNSSEC-validating resolvers have KSK-current in their local 463 trust anchor cache. 465 There is no current published measurement data that indicates to what 466 extent the first two assumptions listed here are valid, and how many 467 end users may be impacted by these assumptions. In particular, the 468 first assumption, that a consistent SERFAIL response will cause the 469 local stub DNS resolution environment to query all of its configured 470 recursive resolvers before concluding that the name cannot be 471 resolved, is a very critical assumption for this test. 473 4.3. Test Procedure 475 The sentinel detection process tests a DNS resolution environment 476 with three query names: 478 o A query name that is signed with a DNSSEC signature that cannot be 479 validated (described as a "bogus" RRset in Section 5 of [RFC4033], 480 when, for example, an RRset is not signed with a valid RRSIG 481 record). 483 o A query name containing the left-most label "root-key-sentinel- 484 not-ta-". This name MUST be a validly- 485 signed. Any validly-signed DNS zone can be used for this test. 487 o A query name containing the left-most label "root-key-sentinel-is- 488 ta-". This name MUST be a validly-signed. 489 Any validly-signed DNS zone can be used for this test. 491 The responses received from queries to resolve each of these names 492 can be evaluated to infer a trust key state of the user's DNS 493 resolution environment. 495 The responses to these queries are described using a simplified 496 notation. Each query will either result in a SERFVAIL response 497 (denoted as "S"), indicating that all of the resolvers in the 498 recursive resolver set returned the SERVFAIL response code, or result 499 in a response with the desire RRset value (denoted as "A"). The 500 queries are ordered by the "invalid" name, the "not-ta" label, then 501 the "is-ta" label, and a triplet notation denotes a particular 502 response. For example, the triplet "(S S A)" denotes a SERVFAIL 503 response to the invalid query, a SERVFAIL response to the "not-ta" 504 query and a RRset response to the "is-ta" query. 506 The set of all possible responses to these three queries are: 508 (A * *): If any resolver returns an "A" response for the query for 509 the invalid name, then the resolver set contains at least one non- 510 validating DNS resolver, and the user will not be impacted by the 511 KSK roll. 513 (S A *): If any of the resolvers returns an "A" response the the 514 "not-ta" query, then at least one of the resolvers does not 515 recognise the sentinel mechanism, and the behaviour of the 516 collection of resolvers during the KSK roll cannot be reliably 517 determined. 519 (S S A): This case implies that all of the resolvers in the set 520 perform DNSSEC-validation, all of the resolvers are aware of the 521 sentinel mechanism, and at least one resolver has loaded KSK-new 522 as a local trust anchor. The user will not be impacted by the KSK 523 roll. 525 (S S S): This case implies that all of the resolvers in the set 526 perform DNSSEC-validation, all of the resolvers are aware of the 527 sentinel mechanism, and none of the resolvers has loaded KSK-new 528 as a local trust anchor. The user will be negatively impacted by 529 the KSK roll. 531 5. Security Considerations 533 This document describes a mechanism to allow users to determine the 534 trust anchor state of root zone key signing keys in the DNS 535 resolution system that they use. If the user executes third party 536 code, then this information may also be available to the third party. 538 The mechanism does not require resolvers to set otherwise 539 unauthenticated responses to be marked as authenticated, and does not 540 alter the security properties of DNSSEC with respect to the 541 interpretation of the authenticity of responses that are so marked. 543 The mechanism does not require any further significant processing of 544 DNS responses, and queries of the form described in this document do 545 not impose any additional load that could be exploited in an attack 546 over the normal DNSSEC validation processing load. 548 6. Privacy Considerations 550 The mechanism in this document enables third parties (with either 551 good or bad intentions) to learn something about the security 552 configuration of recursive DNS resolvers. That is, someone who can 553 cause an Internet user to make specific DNS queries (e.g. via web- 554 based advertisements or javascript in web pages), can, under certain 555 specific circumstances that includes additional knowledge of the 556 resolvers that are invoked by the user, determine which trust anchors 557 are configured in these resolvers. Without this additional 558 knowledge, the third party can infer the aggregate capabilities of 559 the user's DNS resolution environment, but cannot necessarily infer 560 the trust configuration of any recursive name server. 562 7. Implementation Experience 564 [ RFC Editor: Please remove before publication. As this section will 565 be removed, it is more conversational than would appear in a 566 published doc. ] 568 List of known resolver implementations (alphabetical): 570 BIND Ondrej Sury of ISC reported to the DNSOP Working Group in 571 April 2018 that this technique was peer-reviewed and merged into 572 BIND master branch with the intent to backport the feature into 573 older release branches. The merge request: 574 https://gitlab.isc.org/isc-projects/bind9/merge_requests/123 575 Information on configuring this can be found in the BIND 9.13.0 576 Administrator Reference Manual (ARM), available at 577 https://ftp.isc.org/isc/bind9/9.13.0/doc/arm/Bv9ARM.pdf 579 Knot resolver Petr Spacek implemented early versions of this 580 technique into the Knot resolver, identified a number of places 581 where it wasn't clear, and provided very helpful text to address 582 these issues and make the document mode clear. Petr also 583 identified an embarrassingly large number of typos (and similar) 584 in the ksk-test setup. More information is at http://knot- 585 resolver.readthedocs.io/en/stable/modules.html#sentinel-for- 586 detecting-trusted-keys 588 Unbound Benno Overeinder of NLnet Labs reported to the DNSOP Working 589 Group in April 2018 an intention to support this technique in 590 Unbound in the near future. This is now implemented in Unbound 591 version 1.7.1, available from http://unbound.nlnetlabs.nl/ 592 download.html . Configuration information is at 593 http://unbound.nlnetlabs.nl/documentation/unbound.conf.html 595 A (partial) list of "client" / user side implementations (the author 596 was keeping a more complete list of implementations, but has 597 misplaced it - apologies, I'm happy to re-add them if you send me a 598 note.): 600 http://www.ksk-test.net An Javascript implementation of the client 601 side of this protocol is available at: http://www.ksk-test.net 603 http://test.kskroll.dnssec.lab.nic.cl/ Hugo Salgado-Hernandez has 604 created an implementation at 605 http://test.kskroll.dnssec.lab.nic.cl/ 607 http://sentinel.research.icann.org/ The code for this implementation 608 is published at https://github.com/paulehoffman/sentinel-testbed 610 http://www.bellis.me.uk/sentinel/ Ray Bellis client implementation - 611 http://www.bellis.me.uk/sentinel/ 613 8. IANA Considerations 615 This document has no IANA actions. 617 9. Acknowledgements 619 This document has borrowed extensively from [RFC8145] for the 620 introductory text, and the authors would like to acknowledge and 621 thank the authors of that document both for some text excerpts and 622 for the more general stimulation of thoughts about monitoring the 623 progress of a roll of the KSK of the root zone of the DNS. 625 The authors would like to thank Joe Abley, Mehmet Akcin, Mark 626 Andrews, Richard Barnes, Ray Bellis, Stephane Bortzmeyer, David 627 Conrad, Ralph Dolmans, John Dickinson, Steinar Haug, Bob Harold, Wes 628 Hardaker, Paul Hoffman, Matt Larson, Jinmei Tatuya, Edward Lewis, 629 George Michaelson, Benno Overeinder, Matthew Pounsett, Hugo Salgado- 630 Hernandez, Andreas Schulze, Mukund Sivaraman, Petr Spacek, Job 631 Snijders, Andrew Sullivan, Ondrej Sury, Paul Vixie, Duane Wessels and 632 Paul Wouters for their helpful feedback. 634 The authors would like to especially call out Paul Hoffman and Duane 635 Wessels for providing comments in the form of pull requests. Joe 636 Abley also helpfully provided extensive review and OLD / NEW text. 638 Petr Spacek wrote some very early implmentations, and provided 639 significant feedback (including pointing out when the test bed didn't 640 match the document!) 642 10. Change Log 644 RFC Editor: Please remove this section! 646 Note that this document is being worked on in GitHub - see Abstract. 647 The below is mainly large changes, and is not authoritative. 649 From -14 to -15: 651 o Addressed Joe Abley's thorough review, at: 652 https://mailarchive.ietf.org/arch/msg/dnsop/8ZnN1xj55Yimet2cg- 653 LrdoJafEA 655 From -13 to -14: 657 o Addressed nits from Bob Harold - 658 https://mailarchive.ietf.org/arch/msg/dnsop/ 659 j4Serw0z24o470AnlD8ISo8o9k4 661 o Formatting changes (and a bit more text) in the implementation 662 section. 664 o Closes PR #21: Clarify indeterminate and resolution systems, 666 o Closes PR #22: Updates to -13 describing the test procedure for a 667 set of resolvers 669 o Closes PR #23: Fix sundry typos, 671 o Closes PR #24: Editorial and clarifications to the new text 672 o Closes PR #25: Clarified when the test can be run 674 From -12 to -13: 676 o Merged Paul Hoffmans PR#19, PR#20. 678 o Moved toy ksk-test.net to implementation section. 680 o Split the test procedures between the test of a single DNS 681 resolvers and the test of a collection of DNS resolvers as would 682 be found in an end user environment. 684 From -11 to -12: 686 o Moved the Walkthrough Example to the end of the document as an 687 appendix. 689 o Incorporated changes as proposed by Ondrej Sury, relating to a 690 consistent use of Key Tag and a reference to the definition of a 691 Bogus RRset. 693 o Corrected minor typos. 695 o Revised the Privacy Considerations. 697 o In response to a request from DNSOP Working Group chairs, a 698 section on reported Implementation Experience has been added, 699 based on postings to the DNSOP Working Group mailing list. 701 From -10 to -11: 703 o Clarified the preconditions for this mechanism as per Working 704 Group mailing list discussion. 706 o Corrected minor typo. 708 From -09 to -10: 710 o Clarified the precondition list to specify that the resolver had 711 performed DNSSEC-validation by setting the AD bit in the response 713 o Clarified the language referring to the operation of RFC8145 714 signalling. 716 From -08 to -09: 718 o Incorporated Paul Hoffman's PR # 15 (Two issues from the 719 Hackathon) - https://github.com/APNIC-Labs/draft-kskroll-sentinel/ 720 pull/15 722 o Clarifies that the match is on the *original* QNAME. 724 From -08 to -07: 726 o Changed title from "A Sentinel for Detecting Trusted Keys in 727 DNSSEC" to "A Root Key Trust Anchor Sentinel for DNSSEC". 729 o Changed magic string from "kskroll-sentinel-" to "root-key- 730 sentinel-" -- this time for sure, Rocky! 732 From -07 to -06: 734 o Addressed GitHub PR #14: Clarifications regarding caching and 735 SERVFAIL responses 737 o Addressed GitHub PR #12, #13: Clarify situation with multiple 738 resolvers, Fix editorial nits. 740 From -05 to -06: 742 o Paul improved my merging of Petr's text to make it more readable. 743 Minor change, but this is just before the cut-off, so I wanted it 744 maximally readable. 746 From -04 to -05: 748 o Incorporated Duane's #10 750 o Integrated Petr Spacek's Issue - https://github.com/APNIC-Labs/ 751 draft-kskroll-sentinel/issues/9 (note that commit-log incorrectly 752 referred to Duane's PR as number 9, it is actually 10). 754 From -03 to -04: 756 o Addressed GitHub pull requests #4, #5, #6, #7 #8. 758 o Added Duane's privacy concerns 760 o Makes the use cases clearer 762 o Fixed some A/AAAA stuff 764 o Changed the example numbers 765 o Made it clear that names and addresses must be real 767 From -02 to -03: 769 o Integrated / published comments from Paul in GitHub PR #2 - 770 https://github.com/APNIC-Labs/draft-kskroll-sentinel/pull/2 772 o Made the Key Tag be decimal, not hex (thread / consensus in 773 https://mailarchive.ietf.org/arch/msg/dnsop/ 774 Kg7AtDhFRNw31He8n0_bMr9hBuE ) 776 From -01 to 02: 778 o Removed Address Record definition. 780 o Clarified that many things can cause SERVFAIL. 782 o Made examples FQDN. 784 o Fixed a number of typos. 786 o Had accidentally said that Charlie was using a non-validating 787 resolver in example. 789 o [ TODO(WK): Doc says Key Tags are hex, is this really what the WG 790 wants? ] 792 o And active key is one that can be used *now* (not e.g AddPend) 794 From -00 to 01: 796 o Added a conversational description of how the system is intended 797 to work. 799 o Clarification that this is for the root. 801 o Changed the label template from _is-ta- to kskroll- 802 sentinel-is-ta-. This is because BIND (at least) will 803 not allow records which start with an underscore to have address 804 records (CNAMEs, yes, A/AAAA no). Some browsers / operating 805 systems also will not fetch resources from names which start with 806 an underscore. 808 11. References 809 11.1. Normative References 811 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 812 STD 13, RFC 1034, DOI 10.17487/RFC1034, November 1987, 813 . 815 [RFC2308] Andrews, M., "Negative Caching of DNS Queries (DNS 816 NCACHE)", RFC 2308, DOI 10.17487/RFC2308, March 1998, 817 . 819 [RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. 820 Rose, "DNS Security Introduction and Requirements", 821 RFC 4033, DOI 10.17487/RFC4033, March 2005, 822 . 824 [RFC4034] Arends, R., Austein, R., Larson, M., Massey, D., and S. 825 Rose, "Resource Records for the DNS Security Extensions", 826 RFC 4034, DOI 10.17487/RFC4034, March 2005, 827 . 829 [RFC4035] Arends, R., Austein, R., Larson, M., Massey, D., and S. 830 Rose, "Protocol Modifications for the DNS Security 831 Extensions", RFC 4035, DOI 10.17487/RFC4035, March 2005, 832 . 834 [RFC5011] StJohns, M., "Automated Updates of DNS Security (DNSSEC) 835 Trust Anchors", STD 74, RFC 5011, DOI 10.17487/RFC5011, 836 September 2007, . 838 11.2. Informative References 840 [RFC8145] Wessels, D., Kumari, W., and P. Hoffman, "Signaling Trust 841 Anchor Knowledge in DNS Security Extensions (DNSSEC)", 842 RFC 8145, DOI 10.17487/RFC8145, April 2017, 843 . 845 Appendix A. Protocol Walkthrough Example 847 This Appendix provides a non-normative example of how the sentinel 848 mechanism could be used, and what each participant does. It is 849 provided in a conversational tone to be easier to follow. The 850 examples here all assume that each person has just one resolver, or a 851 system of resolvers that have the same properties. 853 Alice is in charge of the DNS root KSK (Key Signing Key), and would 854 like to roll / replace the key with a new one. She publishes the new 855 KSK, but would like to be able to predict / measure what the impact 856 will be before removing/revoking the old key. The current KSK has a 857 Key Tag of 11112, the new KSK has a Key Tag of 02323. Users want to 858 verify that their resolver will not break after Alice rolls the root 859 KSK key (that is, starts signing with just the KSK whose Key Tag is 860 02323). 862 Bob, Charlie, Dave, Ed are all users. They use the DNS recursive 863 resolvers supplied by their ISPs. They would like to confirm that 864 their ISPs have picked up the new KSK. Bob's ISP does not perform 865 validation. Charlie's ISP does validate, but the resolvers have not 866 yet been upgraded to support this mechanism. Dave and Ed's resolvers 867 have been upgraded to support this mechanism; Dave's resolver has the 868 new KSK, Ed's resolver hasn't managed to install the 02323 KSK in its 869 trust store yet. 871 Geoff is a researcher, and would like to both provide a means for 872 Bob, Charlie, Dave and Ed to be able to perform tests, and also would 873 like to be able to perform Internet-wide measurements of what the 874 impact will be (and report this back to Alice). 876 Geoff sets an authoritative DNS server for example.com, and also a 877 webserver (www.example.com). He adds three address records to 878 example.com: 880 bogus.example.com. IN AAAA 2001:db8::1 882 root-key-sentinel-is-ta-02323.example.com. IN AAAA 2001:db8::1 884 root-key-sentinel-not-ta-11112.example.com. IN AAAA 2001:db8::1 886 Note that the use of "example.com" names and the addresses here are 887 examples. In a real deployment, the domain names need to be under 888 control of the researcher, and the addresses must be real, reachable 889 addresses. 891 Geoff then DNSSEC signs the example.com zone, and intentionally makes 892 the bogus.example.com record have bogus validation status (for 893 example, by editing the signed zone and entering garbage for the 894 signature). Geoff also configures his webserver to listen on 895 2001:db8::1 and serve a resource (for example, a 1x1 GIF, 1x1.gif) 896 for all of these names. The webserver also serves a webpage 897 (www.example.com) which contains links to these 3 resources 898 (http://bogus.example.com/1x1.gif, http://root-key-sentinel-is-ta- 899 02323.example.com/1x1.gif, http://root-key-sentinel-not-ta- 900 11112.example.com/1x1.gif). 902 Geoff then asks Bob, Charlie, Dave and Ed to browse to 903 www.example.com. Using the methods described in this document, the 904 users can figure out what their fate will be when the 11112 KSK is 905 removed. 907 Bob is not using a validating resolver. This means that he will be 908 able to resolve bogus.example.com (and fetch the 1x1 GIF) - this 909 tells him that the KSK roll does not affect him, and so he will be 910 OK. 912 Charlie's resolvers are validating, but they have not been upgraded 913 to support the KSK sentinel mechanism. Charlie will not be able to 914 fetch the http://bogus.example.com/1x1.gif resource (the 915 bogus.example.com record is bogus, and none of his resolvers will 916 resolve it). He is able to fetch both of the other resources - from 917 this he knows (see the logic in the body of this document) that he is 918 using validating resolvers, but at least one of these resolvers is 919 not configured to perform sentinel processing. The KSK sentinel 920 method cannot provide him with a definitive answer to the question of 921 whether he will be impacted by the KSK roll. 923 Dave's resolvers implement the sentinel method, and have picked up 924 the new KSK. For the same reason as Charlie, he cannot fetch the 925 "bogus" resource. His resolver resolves the root-key-sentinel-is-ta- 926 02323.example.com name normally (it contacts the example.com 927 authoritative servers, etc); as it supports the sentinel mechanism, 928 just before Dave's recursive resolver sends the reply to Dave's stub, 929 it performs the KSK Sentinel check. The QNAME starts with "root-key- 930 sentinel-is-ta-", and the recursive resolver does indeed have a key 931 with the Key Tag of 02323 in its root trust store. This means that 932 that this part of the KSK Sentinel check passes (it is true that Key 933 Tag 02323 is in the trust anchor store), and the recursive resolver 934 replies normally (with the answer provided by the authoritative 935 server). Dave's recursive resolver then resolves the root-key- 936 sentinel-not-ta-11112.example.com name. Once again, it performs the 937 normal resolution process, but because it implements KSK Sentinel 938 (and the QNAME starts with "root-key-sentinel-not-ta-"), just before 939 sending the reply, it performs the KSK Sentinel check. As it has the 940 key with key-tag 11112 in it's trust anchor store, the answer to "is 941 this *not* a trust anchor" is false, and so the recursive resolver 942 does not reply with the answer from the authoritative server - 943 instead, it replies with a SERVFAIL (note that replying with SERVFAIL 944 instead of the original answer is the only mechanism that KSK 945 Sentinel uses). This means that Dave cannot fetch "bogus", he can 946 fetch "root-key-sentinel-is-ta-02323", but he cannot fetch "root-key- 947 sentinel-not-ta-11112". From this, Dave knows that he is behind an 948 collection of resolvers that all validate, all have the key with key 949 tag 11112 loaded and at least one of these resolvers has loaded the 950 key with key-tag 02323 into its local trust anchor cache, Dave will 951 not be impacted by the KSK roll. 953 Just like Charlie and Dave, Ed cannot fetch the "bogus" record. This 954 tells him that his resolvers are validating. When his (sentinel- 955 aware) resolvers performs the KSK Sentinel check for "root-key- 956 sentinel-is-ta-02323", none of them have loaded the new key with key- 957 tag 02323 in their local trust anchor store. This means check fails, 958 and Ed's recursive resolver converts the (valid) answer into a 959 SERVFAIL error response. It performs the same check for root-key- 960 sentinel-not-ta-11112.example.com, and as all of Ed's resolvers both 961 perform DNSSEC validation and recognise the sentinel label Ed will be 962 unable to fetch the "root-key-sentinel-not-ta-11112" resource. This 963 tells Ed that his resolvers have not installed the new KSK and he 964 will be negatively implacted by the KSK roll.. 966 Geoff would like to do a large scale test and provide the information 967 back to Alice. He uses some mechanism such as causing users to go to 968 a web page to cause a large number of users to attempt to resolve the 969 three resources, and then analyzes the results of the tests to 970 determine what percentage of users will be affected by the KSK 971 rollover event. 973 This description is a simplified example - it is not anticipated that 974 Bob, Charlie, Dave and Ed will actually look for the absence or 975 presence of web resources; instead, the webpage that they load would 976 likely contain JavaScript (or similar) which displays the result of 977 the tests, sends the results to Geoff, or both. This sentinel 978 mechanism does not rely on the web: it can equally be used by trying 979 to resolve the names (for example, using the common "dig" command) 980 and checking which result in a SERVFAIL. 982 Authors' Addresses 984 Geoff Huston 986 Email: gih@apnic.net 987 URI: http://www.apnic.net 989 Joao Silva Damas 991 Email: joao@apnic.net 992 URI: http://www.apnic.net 994 Warren Kumari 996 Email: warren@kumari.net