idnits 2.17.1 draft-ietf-dnsop-kskroll-sentinel-14.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 (June 11, 2018) is 2136 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: December 13, 2018 W. Kumari 6 Google 7 June 11, 2018 9 A Root Key Trust Anchor Sentinel for DNSSEC 10 draft-ietf-dnsop-kskroll-sentinel-14 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 December 13, 2018. 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 . . . . . . . . . . . . . . . . . . . . . . 5 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 for a Set of Resolvers . . . . . . . . . . . . 9 72 4.1. Test Scenario and Objective . . . . . . . . . . . . . . . 9 73 4.2. Test Assumptions . . . . . . . . . . . . . . . . . . . . 10 74 4.3. Test Procedure . . . . . . . . . . . . . . . . . . . . . 10 75 5. Security Considerations . . . . . . . . . . . . . . . . . . . 12 76 6. Privacy Considerations . . . . . . . . . . . . . . . . . . . 12 77 7. Implementation Experience . . . . . . . . . . . . . . . . . . 12 78 8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 79 9. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 13 80 10. Change Log . . . . . . . . . . . . . . . . . . . . . . . . . 14 81 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 17 82 11.1. Normative References . . . . . . . . . . . . . . . . . . 17 83 11.2. Informative References . . . . . . . . . . . . . . . . . 18 84 Appendix A. Protocol Walkthrough Example . . . . . . . . . . . . 18 85 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 21 87 1. Introduction 89 The DNS Security Extensions (DNSSEC) [RFC4033], [RFC4034] and 90 [RFC4035] were developed to provide origin authentication and 91 integrity protection for DNS data by using digital signatures. 92 DNSSEC uses Key Tags to efficiently match signatures to the keys from 93 which they are generated. The Key Tag is a 16-bit value computed 94 from the RDATA portion of a DNSKEY RR using a formula found in "Key 95 Tag Calculation" (Appendix B of "Resource Records for the DNS 96 Security Extensions" [RFC4034]), a formula similar to a ones- 97 complement checksum. RRSIG RRs contain a Key Tag field whose value 98 is equal to the Key Tag of the DNSKEY RR that validates the 99 signature. 101 This document specifies how security-aware DNS resolvers that perform 102 validation of their responses can respond to certain queries in a 103 manner that allows an agent performing the queries to deduce whether 104 a particular key for the root has been loaded into that resolver's 105 trusted key store. This document also describes a procedure where a 106 collection of resolvers can be tested to determine if at least one of 107 these resolvers has loaded a given key into its trusted key store. 108 These tests can be used to determine whether a certain root zone KSK 109 is ready to be used as a trusted key, within the context of a planned 110 root zone KSK key roll. 112 There are two primary use cases for this mechanism: 114 o Users may wish to ascertain whether their DNS resolution 115 environment resolvers is ready for an upcoming root KSK rollover. 117 o Researchers want to perform Internet-wide studies about the 118 proportion of users who will be negatively impacted an upcoming 119 root KSK rollover. 121 The mechanism described in this document meets both of these use 122 cases. This new mechanism is OPTIONAL to implement and use, although 123 for reasons of supporting broad-based measurement techniques, it is 124 strongly preferred that configurations of DNSSEC-validating resolvers 125 enabled this mechanism by default, allowing for local configuration 126 directives to disable this mechanism if desired. 128 The KSK sentinel tests described in this document use a test 129 comprising of a set of DNS queries to domain names that have special 130 values for the left-most label. The test relies on recursive 131 resolvers supporting a mechanism that recognises this special name 132 pattern in queries, and under certain defined circumstances will 133 return a DNS SERVFAIL response code (RCODE 2), mimicking the response 134 code that is returned by security-aware resolvers when DNSSEC 135 validation fails. 137 If a browser or operating system is configured with multiple 138 resolvers, and those resolvers have different properties (for 139 example, one performs DNSSEC validation and one does not), the 140 sentinel test described in this document can still be used, but it 141 makes a number of assumptions about DNS resolution behaviour that may 142 not necessarily hold in all environments. If these assumptions do 143 not hold (such as, for example, requiring the stub resolver to query 144 the next recursive resolver in the locally configured set upon 145 receipt of a SERVFAIL response code) then this test may produce 146 indeterminate or inconsistent results. In some cases where these 147 assumptions do not hold, repeating the same test query set may 148 generate different results. 150 Note that the sentinel mechanism described here measures a very 151 different (and likely more useful) metric than [RFC8145]. RFC 8145 152 relies on resolvers reporting towards the root servers a list of 153 locally cached trust anchors for the root zone. Those reports can be 154 used to infer how many resolvers may be impacted by a KSK roll, but 155 not what the user impact of the KSK roll will be. 157 1.1. Terminology 159 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 160 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 161 document are to be interpreted as described in RFC 2119. 163 2. Sentinel Mechanism in Resolvers 165 DNSSEC-Validating resolvers that implement this mechanism MUST 166 perform validation of responses in accordance with the DNSSEC 167 response validation specification [RFC4035]. 169 This sentinel mechanism makes use of two special labels: 171 o root-key-sentinel-is-ta- 173 o root-key-sentinel-not-ta- 175 Note that the is specified in the DNS label as unsigned 176 decimal integer (as described in [RFC4034], section 5.3), but zero- 177 padded to five digits (for example, a Key Tag value of 42 would be 178 represented in the label as 00042). 180 These labels trigger special processing in the validating DNS 181 resolver when responses from authoritative servers are received. 182 Labels containing "root-key-sentinel-is-ta-" is used to 183 answer the question "Is this the Key Tag of a key which the 184 validating DNS resolver is currently trusting as a trust anchor?" 185 Labels containing "root-key-sentinel-not-ta-" is used to 186 answer the question "Is this the Key Tag of a key which the 187 validating DNS resolver is *not* currently trusting as a trust 188 anchor?" 190 2.1. Preconditions 192 All of the following conditions must be met to trigger special 193 processing inside resolver code: 195 o The DNS response is DNSSEC validated. 197 o The result of validation is "Secure". 199 o The Checking Disabled (CD) bit in the query is not set. 201 o The QTYPE is either A or AAAA (Query Type value 1 or 28). 203 o The OPCODE is QUERY. 205 o The leftmost label of the original QNAME (the name sent in the 206 Question Section in the original query) is either "root-key- 207 sentinel-is-ta-" or "root-key-sentinel-not-ta-". 209 If any one of the preconditions is not met, the resolver MUST NOT 210 alter the DNS response based on the mechanism in this document. 212 2.2. Special Processing 214 Responses which fulfil all of the preconditions in Section 2.1 215 require special processing, depending on leftmost label in the QNAME. 217 First, the resolver determines if the numerical value of is 218 equal to any of the Key Tag values of an active root zone KSK which 219 is currently trusted by the local resolver and is stored in its store 220 of trusted keys. An active root zone KSK is one which could 221 currently be used for validation (that is, a key that is not in 222 either the AddPend or Revoked state as described in [RFC5011]). 224 Second, the resolver alters the response being sent to the original 225 query based on both the left-most label and the presence of a key 226 with given Key Tag in the trust anchor store. Two labels and two 227 possible states of the corresponding key generate four possible 228 combinations summarized in the table: 230 Label | Key is trusted | Key is not trusted 231 ------------------------------------------------------------------ 232 is-ta | return original answer | return SERVFAIL 233 not-ta | return SERVFAIL | return original answer 235 Instruction "return SERVFAIL" means that the resolver MUST set 236 RCODE=SERVFAIL (value 2) and the ANSWER section of the DNS response 237 MUST be empty, ignoring all other documents which specify content of 238 the ANSWER section. 240 3. Sentinel Tests for a Single DNS Resolver 242 This section describes the use of the sentinel detection mechanism 243 against a single DNS recursive resolver in order to determine whether 244 this resolver is using a particular trust anchor to validate DNSSEC- 245 signed responses. 247 Note that the test in this section applies to a single DNS resolver. 248 The test described in Section 4 applies instead to a collection of 249 DNS resolvers, as might be found in the DNS configuration of an end- 250 user environment. 252 The critical aspect of the DNS names used in this mechanism is that 253 they contain the specified label for either the positive and negative 254 test as the left-most label in the query name. 256 The sentinel detection procedure can test a DNS resolver using three 257 queries: 259 o A query name containing the left-most label "root-key-sentinel-is- 260 ta-". This corresponds to a a validly-signed RRset in 261 the zone, so that responses associated with queried names in this 262 zone can be authenticated by a DNSSEC-validating resolver. Any 263 validly-signed DNS zone can be used for this test. 265 o A query name containing the left-most label "root-key-sentinel- 266 not-ta-". This is also a validly-signed name. Any 267 validly-signed DNS zone can be used for this test. 269 o A query name that is signed with a DNSSEC signature that cannot be 270 validated (described as a "bogus" RRset in Section 5 of [RFC4033], 271 when, for example, an RRset is not signed with a valid RRSIG 272 record). 274 The responses received from queries to resolve each of these names 275 can be evaluated to infer a trust key state of the DNS resolver. 277 An essential assumption here is that this technique relies on 278 security-aware (DNSSEC validating) resolvers responding with a 279 SERVFAIL response code to queries where DNSSEC checking is requested 280 and the response cannot be validated. Note that a slew of other 281 issues can also cause SERVFAIL responses, and so the sentinel 282 processing may sometimes result in incorrect or indeterminate 283 conclusions. 285 To describe this process of classification, DNS resolvers are 286 classified by five distinct behavior types using the labels: "Vnew", 287 "Vold", "Vind", "nonV", and "other". These labels correspond to 288 resolver system behaviour types as follows: 290 Vnew: A DNS resolver that is configured to implement this mechanism 291 and has loaded the nominated key into their local trusted key 292 stores will respond with an A or AAAA RRset response for the 293 associated "root-key-sentinel-is-ta" queries, SERVFAIL for "root- 294 key-sentinel-not-ta" queries and SERVFAIL for the signed name 295 queries that return "bogus" validation status. 297 Vold: A DNS resolver that is configured to implement this mechanism 298 and has not loaded the nominated key into their local trusted key 299 stores will respond with an SERVFAIL for the associated "root-key- 300 sentinel-is-ta" queries, an A or AAAA RRset response for "root- 301 key-sentinel-not-ta" queries and SERVFAIL for the signed name 302 queries that return "bogus" validation status. 304 Vind: A DNS resolver that has is not configured to implement this 305 mechanism will respond with an A or AAAA RRset response for "root- 306 key-sentinel-is-ta", an A or AAAA RRset response for "root-key- 307 sentinel-not-ta" and SERVFAIL for the name that returns "bogus" 308 validation status. This set of responses does not give any 309 information about the trust anchors used by this resolver. 311 nonV: A non-security-aware DNS resolver will respond with an A or 312 AAAA record response for "root-key-sentinel-is-ta", an A record 313 response for "root-key-sentinel-not-ta" and an A or AAAA RRset 314 response for the name that returns "bogus" validation status. 316 other: There is the potential to admit other combinations of 317 responses to these three queries. While this may appear self- 318 contradictory, there are cases where such an outcome is possible. 319 For example, in DNS resolver farms what appears to be a single DNS 320 resolver that responds to queries passed to a single IP address is 321 in fact constructed as a a collection of slave resolvers, and the 322 query is passed to one of these internal resolver engines. If 323 these individual slave resolvers in the farm do not behave 324 identically, then other sets of results can be expected from these 325 three queries. In such a case, no determination about the 326 capabilities of this DNS resolver farm can be made. 328 Note that SERVFAIL might be cached according to Section 7 of 329 [RFC2308] for up to 5 minutes and a positive answer for up to its 330 TTL. 332 If a client directs these three queries to a single resolver, the 333 responses should allow the client to determine the capability of the 334 resolver, and if it supports this sentinel mechanism, whether or not 335 it has a particular key in its trust anchor store, as in the 336 following table: 338 Query 339 +----------+-----------+------------+ 340 | is-ta | not-ta | bogus | 341 +-------+----------+-----------+------------+ 342 | Vnew | A | SERVFAIL | SERVFAIL | 343 | Vold | SERVFAIL | A | SERVFAIL | 344 Type | Vind | A | A | SERVFAIL | 345 | nonV | A | A | A | 346 | other | * | * | * | 347 +-------+----------+-----------+------------+ 349 Vnew: The nominated key is trusted by the resolver. 351 Vold: The nominated key is not yet trusted by the resolver. 353 Vind: There is no information about the trust anchors of the 354 resolver. 356 nonV: The resolver does not perform DNSSEC validation. 358 other: The properties of the resolver cannot be analyzed by this 359 protocol. 361 3.1. Forwarders 363 There is also the common case of a recursive resolver using a 364 forwarder. 366 If the resolver is non-validating, and it has a single forwarder, 367 then the resolver will presumably mirror the capabilities of the 368 forwarder target resolver. 370 If the validating resolver has a forwarding configuration, and uses 371 the CD bit on all forwarded queries, then this resolver is acting in 372 a manner that is identical to a standalone resolver. 374 A more complex case is where all of the following conditions hold: 376 o Both the validating resolver and the forwarder target resolver 377 support this trusted key sentinel mechanism 379 o The local resolver's queries do not have the CD bit set 380 o The trusted key state differs between the forwarding resolver and 381 the forwarder target resolver 383 In such a case, either the outcome is indeterminate validating 384 ("Vind"), or a case of mixed signals such as SERVFAIL in all three 385 responses, ("other") which is similarly an indeterminate response 386 with respect to the trusted key state. 388 4. Sentinel Tests for a Set of Resolvers 390 The description in Section 3 describes a trust anchor test that can 391 be used in the simple situation where the test queries were being 392 passed to a single recursive resolver that directly queries 393 authoritative name servers. 395 However, the common end user scenario is where a user's local DNS 396 resolution environment is configured to use a set of recursive 397 resolvers. The single resolver test technique will not function 398 reliably in such cases, as a a SERVFAIL response from one resolver 399 may cause the local stub resolver to repeat the query against one of 400 the other configured resolvers and the results may be inconclusive. 402 In describing a test procedure that can be used in this environment 403 of a set of DNS resolvers there are some necessary changes to the 404 nature of the question that this test can answer, the assumptions 405 about the behaviour of the DNS resolution environment, and some 406 further observations about potential variability in the test 407 outcomes. 409 4.1. Test Scenario and Objective 411 This test is not intended to expose which trust anchors are used by 412 any single DNS resolver. 414 The test scenario is explicitly restricted to that of the KSK 415 environment where a current active KSK (called "KSK-current") is to 416 be replaced with a new KSK (called "KSK-new"). The test is designed 417 to be run between when KSK-new is introduced into the root zone and 418 when the root zone is signed with KSK-new. 420 The objective of the test is to determine if the user will be 421 negatively impacted by the KSK roll. A "negative impact" for the 422 user is defined such that all the configured resolvers are security- 423 aware resolvers that perform validation of DNSSEC-signed responses, 424 and none of these resolvers have loaded KSK-new into their local 425 trust anchor set. In this situation, it is anticipated that once the 426 KSK is rolled the entire set of the user's resolvers will not be able 427 to validate the contents of the root zone and the user is likely to 428 lose DNS service as a result of this inability to perform successful 429 DNSSEC validation. 431 4.2. Test Assumptions 433 There are a number of assumptions about the DNS environment used in 434 this test. Where these assumptions do not hold, the results of the 435 test will be indeterminate. 437 o When a recursive resolver returns SERVFAIL to the user's stub 438 resolver, the stub resolver will send the same query to the next 439 resolver in the locally configured resolver set. It will continue 440 to do this until it gets a non-SERVFAIL response or until it runs 441 out of resolvers to try. 443 o When the user's stub resolver passes a query to a resolver in the 444 configured resolver set, it will get a consistent answer over the 445 timeframe of the queries. This assumption implies that if the 446 same query is asked by the same stub resolver multiple times in 447 succession to the same recursive resolver, the recursive 448 resolver's response will be the same for each of these queries. 450 o All DNSSEC-validating resolvers have KSK-current in their local 451 trust anchor cache. 453 There is no current published measurement data that indicates to what 454 extent the first two assumptions listed here are valid, and how many 455 end users may be impacted by these assumptions. In particular, the 456 first assumption, that a consistent SERFAIL response will cause the 457 local stub DNS resolution environment to query all of its configured 458 recursive resolvers before concluding that the name cannot be 459 resolved, is a very critical assumption for this test. 461 4.3. Test Procedure 463 The sentinel detection process tests a DNS resolution environment 464 with three query names: 466 o A query name that is signed with a DNSSEC signature that cannot be 467 validated (described as a "bogus" RRset in Section 5 of [RFC4033], 468 when, for example, an RRset is not signed with a valid RRSIG 469 record). 471 o A query name containing the left-most label "root-key-sentinel- 472 not-ta-". This name MUST be a validly- 473 signed. Any validly-signed DNS zone can be used for this test. 475 o A query name containing the left-most label "root-key-sentinel-is- 476 ta-". This name MUST be a validly-signed. 477 Any validly-signed DNS zone can be used for this test. 479 The responses received from queries to resolve each of these names 480 can be evaluated to infer a trust key state of the user's DNS 481 resolution environment. 483 The responses to these queries are described using a simplified 484 notation. Each query will either result in a SERFVAIL response 485 (denoted as "S"), indicating that all of the resolvers in the 486 recursive resolver set returned the SERVFAIL response code, or result 487 in a response with the desire RRset value (denoted as "A"). The 488 queries are ordered by the "invalid" name, the "not-ta" label, then 489 the "is-ta" label, and a triplet notation denotes a particular 490 response. For example, the triplet "(S S A)" denotes a SERVFAIL 491 response to the invalid query, a SERVFAIL response to the "not-ta" 492 query and a RRset response to the "is-ta" query. 494 The set of all possible responses to these three queries are: 496 (A * *): If any resolver returns an "A" response for the query for 497 the invalid name, then the resolver set contains at least one non- 498 validating DNS resolver, and the user will not be impacted by the 499 KSK roll. 501 (S A *): If any of the resolvers returns an "A" response the the 502 "not-ta" query, then at least one of the resolvers does not 503 recognise the sentinel mechanism, and the behaviour of the 504 collection of resolvers during the KSK roll cannot be reliably 505 determined. 507 (S S A): This case implies that all of the resolvers in the set 508 perform DNSSEC-validation, all of the resolvers are aware of the 509 sentinel mechanism, and at least one resolver has loaded KSK-new 510 as a local trust anchor. The user will not be impacted by the KSK 511 roll. 513 (S S S): This case implies that all of the resolvers in the set 514 perform DNSSEC-validation, all of the resolvers are aware of the 515 sentinel mechanism, and none of the resolvers has loaded KSK-new 516 as a local trust anchor. The user will be negatively impacted by 517 the KSK roll. 519 5. Security Considerations 521 This document describes a mechanism to allow users to determine the 522 trust anchor state of root zone key signing keys in the DNS 523 resolution system that they use. If the user executes third party 524 code, then this information may also be available to the third party. 526 The mechanism does not require resolvers to set otherwise 527 unauthenticated responses to be marked as authenticated, and does not 528 alter the security properties of DNSSEC with respect to the 529 interpretation of the authenticity of responses that are so marked. 531 The mechanism does not require any further significant processing of 532 DNS responses, and queries of the form described in this document do 533 not impose any additional load that could be exploited in an attack 534 over the normal DNSSEC validation processing load. 536 6. Privacy Considerations 538 The mechanism in this document enables third parties (with either 539 good or bad intentions) to learn something about the security 540 configuration of recursive DNS resolvers. That is, someone who can 541 cause an Internet user to make specific DNS queries (e.g. via web- 542 based advertisements or javascript in web pages), can, under certain 543 specific circumstances that includes additional knowledge of the 544 resolvers that are invoked by the user, determine which trust anchors 545 are configured in these resolvers. Without this additional 546 knowledge, the third party can infer the aggregate capabilities of 547 the user's DNS resolution environment, but cannot necessarily infer 548 the trust configuration of any recursive name server. 550 7. Implementation Experience 552 [ RFC Editor: Please remove before publication. As this section will 553 be removed, it is more conversational than would appear in a 554 published doc. ] 556 List of known resolver implementations (alphabetical): 558 BIND Ondrej Sury of ISC reported to the DNSOP Working Group in 559 April 2018 that this technique was peer-reviewed and merged into 560 BIND master branch with the intent to backport the feature into 561 older release branches. The merge request: 562 https://gitlab.isc.org/isc-projects/bind9/merge_requests/123 563 Information on configuring this can be found in the BIND 9.13.0 564 Administrator Reference Manual (ARM), available at 565 https://ftp.isc.org/isc/bind9/9.13.0/doc/arm/Bv9ARM.pdf 567 Knot resolver Petr Spacek implemented early versions of this 568 technique into the Knot resolver, identified a number of places 569 where it wasn't clear, and provided very helpful text to address 570 these issues and make the document mode clear. Petr also 571 identified an embarrassingly large number of typos (and similar) 572 in the ksk-test setup. More information is at http://knot- 573 resolver.readthedocs.io/en/stable/modules.html#sentinel-for- 574 detecting-trusted-keys 576 Unbound Benno Overeinder of NLnet Labs reported to the DNSOP Working 577 Group in April 2018 an intention to support this technique in 578 Unbound in the near future. This is now implemented in Unbound 579 version 1.7.1, available from http://unbound.nlnetlabs.nl/ 580 download.html . Configuration information is at 581 http://unbound.nlnetlabs.nl/documentation/unbound.conf.html 583 A (partial) list of "client" / user side implementations (the author 584 was keeping a more complete list of implementations, but has 585 misplaced it - apologies, I'm happy to re-add them if you send me a 586 note.): 588 http://www.ksk-test.net An Javascript implementation of the client 589 side of this protocol is available at: http://www.ksk-test.net 591 http://test.kskroll.dnssec.lab.nic.cl/ Hugo Salgado-Hernandez has 592 created an implementation at 593 http://test.kskroll.dnssec.lab.nic.cl/ 595 http://sentinel.research.icann.org/ The code for this implementation 596 is published at https://github.com/paulehoffman/sentinel-testbed 598 8. IANA Considerations 600 [Note to IANA, to be removed prior to publication: there are no IANA 601 considerations stated in this version of the document.] 603 9. Acknowledgements 605 This document has borrowed extensively from [RFC8145] for the 606 introductory text, and the authors would like to acknowledge and 607 thank the authors of that document both for some text excerpts and 608 for the more general stimulation of thoughts about monitoring the 609 progress of a roll of the KSK of the root zone of the DNS. 611 The authors would like to thank Joe Abley, Mehmet Akcin, Mark 612 Andrews, Richard Barnes, Ray Bellis, Stephane Bortzmeyer, David 613 Conrad, Ralph Dolmans, John Dickinson, Steinar Haug, Bob Harold, Wes 614 Hardaker, Paul Hoffman, Matt Larson, Jinmei Tatuya, Edward Lewis, 615 George Michaelson, Benno Overeinder, Matthew Pounsett, Hugo Salgado- 616 Hernandez, Andreas Schulze, Mukund Sivaraman, Petr Spacek, Job 617 Snijders, Andrew Sullivan, Ondrej Sury, Paul Vixie, Duane Wessels and 618 Paul Wouters for their helpful feedback. 620 The authors would like to especially call out Paul Hoffman and Duane 621 Wessels for providing comments in the form of a pull request. 623 10. Change Log 625 RFC Editor: Please remove this section! 627 Note that this document is being worked on in GitHub - see Abstract. 628 The below is mainly large changes, and is not authoritative. 630 From -13 to -14: 632 o Addressed nits from Bob Harold - 633 https://mailarchive.ietf.org/arch/msg/dnsop/ 634 j4Serw0z24o470AnlD8ISo8o9k4 636 o Formatting changes (and a bit more text) in the implementation 637 section. 639 o Closes PR #21: Clarify indeterminate and resolution systems, 641 o Closes PR #22: Updates to -13 describing the test procedure for a 642 set of resolvers 644 o Closes PR #23: Fix sundry typos, 646 o Closes PR #24: Editorial and clarifications to the new text 648 o Closes PR #25: Clarified when the test can be run 650 From -12 to -13: 652 o Merged Paul Hoffmans PR#19, PR#20. 654 o Moved toy ksk-test.net to implementation section. 656 o Split the test procedures between the test of a single DNS 657 resolvers and the test of a collection of DNS resolvers as would 658 be found in an end user environment. 660 From -11 to -12: 662 o Moved the Walkthrough Example to the end of the document as an 663 appendix. 665 o Incorporated changes as proposed by Ondrej Sury, relating to a 666 consistent use of Key Tag and a reference to the definition of a 667 Bogus RRset. 669 o Corrected minor typos. 671 o Revised the Privacy Considerations. 673 o In response to a request from DNSOP Working Group chairs, a 674 section on reported Implementation Experience has been added, 675 based on postings to the DNSOP Working Group mailing list. 677 From -10 to -11: 679 o Clarified the preconditions for this mechanism as per Working 680 Group mailing list discussion. 682 o Corrected minor typo. 684 From -09 to -10: 686 o Clarified the precondition list to specify that the resolver had 687 performed DNSSEC-validation by setting the AD bit in the response 689 o Clarified the language referring to the operation of RFC8145 690 signalling. 692 From -08 to -09: 694 o Incorporated Paul Hoffman's PR # 15 (Two issues from the 695 Hackathon) - https://github.com/APNIC-Labs/draft-kskroll-sentinel/ 696 pull/15 698 o Clarifies that the match is on the *original* QNAME. 700 From -08 to -07: 702 o Changed title from "A Sentinel for Detecting Trusted Keys in 703 DNSSEC" to "A Root Key Trust Anchor Sentinel for DNSSEC". 705 o Changed magic string from "kskroll-sentinel-" to "root-key- 706 sentinel-" -- this time for sure, Rocky! 708 From -07 to -06: 710 o Addressed GitHub PR #14: Clarifications regarding caching and 711 SERVFAIL responses 713 o Addressed GitHub PR #12, #13: Clarify situation with multiple 714 resolvers, Fix editorial nits. 716 From -05 to -06: 718 o Paul improved my merging of Petr's text to make it more readable. 719 Minor change, but this is just before the cut-off, so I wanted it 720 maximally readable. 722 From -04 to -05: 724 o Incorporated Duane's #10 726 o Integrated Petr Spacek's Issue - https://github.com/APNIC-Labs/ 727 draft-kskroll-sentinel/issues/9 (note that commit-log incorrectly 728 referred to Duane's PR as number 9, it is actually 10). 730 From -03 to -04: 732 o Addressed GitHub pull requests #4, #5, #6, #7 #8. 734 o Added Duane's privacy concerns 736 o Makes the use cases clearer 738 o Fixed some A/AAAA stuff 740 o Changed the example numbers 742 o Made it clear that names and addresses must be real 744 From -02 to -03: 746 o Integrated / published comments from Paul in GitHub PR #2 - 747 https://github.com/APNIC-Labs/draft-kskroll-sentinel/pull/2 749 o Made the Key Tag be decimal, not hex (thread / consensus in 750 https://mailarchive.ietf.org/arch/msg/dnsop/ 751 Kg7AtDhFRNw31He8n0_bMr9hBuE ) 753 From -01 to 02: 755 o Removed Address Record definition. 757 o Clarified that many things can cause SERVFAIL. 759 o Made examples FQDN. 761 o Fixed a number of typos. 763 o Had accidentally said that Charlie was using a non-validating 764 resolver in example. 766 o [ TODO(WK): Doc says Key Tags are hex, is this really what the WG 767 wants? ] 769 o And active key is one that can be used *now* (not e.g AddPend) 771 From -00 to 01: 773 o Added a conversational description of how the system is intended 774 to work. 776 o Clarification that this is for the root. 778 o Changed the label template from _is-ta- to kskroll- 779 sentinel-is-ta-. This is because BIND (at least) will 780 not allow records which start with an underscore to have address 781 records (CNAMEs, yes, A/AAAA no). Some browsers / operating 782 systems also will not fetch resources from names which start with 783 an underscore. 785 11. References 787 11.1. Normative References 789 [RFC2308] Andrews, M., "Negative Caching of DNS Queries (DNS 790 NCACHE)", RFC 2308, DOI 10.17487/RFC2308, March 1998, 791 . 793 [RFC4033] Arends, R., Austein, R., Larson, M., Massey, D., and S. 794 Rose, "DNS Security Introduction and Requirements", 795 RFC 4033, DOI 10.17487/RFC4033, March 2005, 796 . 798 [RFC4034] Arends, R., Austein, R., Larson, M., Massey, D., and S. 799 Rose, "Resource Records for the DNS Security Extensions", 800 RFC 4034, DOI 10.17487/RFC4034, March 2005, 801 . 803 [RFC4035] Arends, R., Austein, R., Larson, M., Massey, D., and S. 804 Rose, "Protocol Modifications for the DNS Security 805 Extensions", RFC 4035, DOI 10.17487/RFC4035, March 2005, 806 . 808 [RFC5011] StJohns, M., "Automated Updates of DNS Security (DNSSEC) 809 Trust Anchors", STD 74, RFC 5011, DOI 10.17487/RFC5011, 810 September 2007, . 812 11.2. Informative References 814 [RFC8145] Wessels, D., Kumari, W., and P. Hoffman, "Signaling Trust 815 Anchor Knowledge in DNS Security Extensions (DNSSEC)", 816 RFC 8145, DOI 10.17487/RFC8145, April 2017, 817 . 819 Appendix A. Protocol Walkthrough Example 821 This Appendix provides a non-normative example of how the sentinel 822 mechanism could be used, and what each participant does. It is 823 provided in a conversational tone to be easier to follow. The 824 examples here all assume that each person has just one resolver, or a 825 system of resolvers that have the same properties. 827 Alice is in charge of the DNS root KSK (Key Signing Key), and would 828 like to roll / replace the key with a new one. She publishes the new 829 KSK, but would like to be able to predict / measure what the impact 830 will be before removing/revoking the old key. The current KSK has a 831 Key Tag of 11112, the new KSK has a Key Tag of 02323. Users want to 832 verify that their resolver will not break after Alice rolls the root 833 KSK key (that is, starts signing with just the KSK whose Key Tag is 834 02323). 836 Bob, Charlie, Dave, Ed are all users. They use the DNS recursive 837 resolvers supplied by their ISPs. They would like to confirm that 838 their ISPs have picked up the new KSK. Bob's ISP does not perform 839 validation. Charlie's ISP does validate, but the resolvers have not 840 yet been upgraded to support this mechanism. Dave and Ed's resolvers 841 have been upgraded to support this mechanism; Dave's resolver has the 842 new KSK, Ed's resolver hasn't managed to install the 02323 KSK in its 843 trust store yet. 845 Geoff is a researcher, and would like to both provide a means for 846 Bob, Charlie, Dave and Ed to be able to perform tests, and also would 847 like to be able to perform Internet-wide measurements of what the 848 impact will be (and report this back to Alice). 850 Geoff sets an authoritative DNS server for example.com, and also a 851 webserver (www.example.com). He adds three address records to 852 example.com: 854 bogus.example.com. IN AAAA 2001:db8::1 855 root-key-sentinel-is-ta-02323.example.com. IN AAAA 2001:db8::1 857 root-key-sentinel-not-ta-11112.example.com. IN AAAA 2001:db8::1 859 Note that the use of "example.com" names and the addresses here are 860 examples. In a real deployment, the domain names need to be under 861 control of the researcher, and the addresses must be real, reachable 862 addresses. 864 Geoff then DNSSEC signs the example.com zone, and intentionally makes 865 the bogus.example.com record have bogus validation status (for 866 example, by editing the signed zone and entering garbage for the 867 signature). Geoff also configures his webserver to listen on 868 2001:db8::1 and serve a resource (for example, a 1x1 GIF, 1x1.gif) 869 for all of these names. The webserver also serves a webpage 870 (www.example.com) which contains links to these 3 resources 871 (http://bogus.example.com/1x1.gif, http://root-key-sentinel-is-ta- 872 02323.example.com/1x1.gif, http://root-key-sentinel-not-ta- 873 11112.example.com/1x1.gif). 875 Geoff then asks Bob, Charlie, Dave and Ed to browse to 876 www.example.com. Using the methods described in this document, the 877 users can figure out what their fate will be when the 11112 KSK is 878 removed. 880 Bob is not using a validating resolver. This means that he will be 881 able to resolve bogus.example.com (and fetch the 1x1 GIF) - this 882 tells him that the KSK roll does not affect him, and so he will be 883 OK. 885 Charlie's resolvers are validating, but they have not been upgraded 886 to support the KSK sentinel mechanism. Charlie will not be able to 887 fetch the http://bogus.example.com/1x1.gif resource (the 888 bogus.example.com record is bogus, and none of his resolvers will 889 resolve it). He is able to fetch both of the other resources - from 890 this he knows (see the logic in the body of this document) that he is 891 using validating resolvers, but at least one of these resolvers is 892 not configured to perform sentinel processing. The KSK sentinel 893 method cannot provide him with a definitive answer to the question of 894 whether he will be impacted by the KSK roll. 896 Dave's resolvers implement the sentinel method, and have picked up 897 the new KSK. For the same reason as Charlie, he cannot fetch the 898 "bogus" resource. His resolver resolves the root-key-sentinel-is-ta- 899 02323.example.com name normally (it contacts the example.com 900 authoritative servers, etc); as it supports the sentinel mechanism, 901 just before Dave's recursive resolver sends the reply to Dave's stub, 902 it performs the KSK Sentinel check. The QNAME starts with "root-key- 903 sentinel-is-ta-", and the recursive resolver does indeed have a key 904 with the Key Tag of 02323 in its root trust store. This means that 905 that this part of the KSK Sentinel check passes (it is true that Key 906 Tag 02323 is in the trust anchor store), and the recursive resolver 907 replies normally (with the answer provided by the authoritative 908 server). Dave's recursive resolver then resolves the root-key- 909 sentinel-not-ta-11112.example.com name. Once again, it performs the 910 normal resolution process, but because it implements KSK Sentinel 911 (and the QNAME starts with "root-key-sentinel-not-ta-"), just before 912 sending the reply, it performs the KSK Sentinel check. As it has the 913 key with key-tag 11112 in it's trust anchor store, the answer to "is 914 this *not* a trust anchor" is false, and so the recursive resolver 915 does not reply with the answer from the authoritative server - 916 instead, it replies with a SERVFAIL (note that replying with SERVFAIL 917 instead of the original answer is the only mechanism that KSK 918 Sentinel uses). This means that Dave cannot fetch "bogus", he can 919 fetch "root-key-sentinel-is-ta-02323", but he cannot fetch "root-key- 920 sentinel-not-ta-11112". From this, Dave knows that he is behind an 921 collection of resolvers that all validate, all have the key with key 922 tag 11112 loaded and at least one of these resolvers has loaded the 923 key with key-tag 02323 into its local trust anchor cache, Dave will 924 not be impacted by the KSK roll. 926 Just like Charlie and Dave, Ed cannot fetch the "bogus" record. This 927 tells him that his resolvers are validating. When his (sentinel- 928 aware) resolvers performs the KSK Sentinel check for "root-key- 929 sentinel-is-ta-02323", none of them have loaded the new key with key- 930 tag 02323 in their local trust anchor store. This means check fails, 931 and Ed's recursive resolver converts the (valid) answer into a 932 SERVFAIL error response. It performs the same check for root-key- 933 sentinel-not-ta-11112.example.com, and as all of Ed's resolvers both 934 perform DNSSEC validation and recognise the sentinel label Ed will be 935 unable to fetch the "root-key-sentinel-not-ta-11112" resource. This 936 tells Ed that his resolvers have not installed the new KSK and he 937 will be negatively implacted by the KSK roll.. 939 Geoff would like to do a large scale test and provide the information 940 back to Alice. He uses some mechanism such as causing users to go to 941 a web page to cause a large number of users to attempt to resolve the 942 three resources, and then analyzes the results of the tests to 943 determine what percentage of users will be affected by the KSK 944 rollover event. 946 This description is a simplified example - it is not anticipated that 947 Bob, Charlie, Dave and Ed will actually look for the absence or 948 presence of web resources; instead, the webpage that they load would 949 likely contain JavaScript (or similar) which displays the result of 950 the tests, sends the results to Geoff, or both. This sentinel 951 mechanism does not rely on the web: it can equally be used by trying 952 to resolve the names (for example, using the common "dig" command) 953 and checking which result in a SERVFAIL. 955 Authors' Addresses 957 Geoff Huston 959 Email: gih@apnic.net 960 URI: http://www.apnic.net 962 Joao Silva Damas 964 Email: joao@apnic.net 965 URI: http://www.apnic.net 967 Warren Kumari 969 Email: warren@kumari.net