idnits 2.17.1 draft-ietf-dnsop-no-response-issue-17.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 1048 has weird spacing: '...r zones for ...' -- The document date (March 11, 2020) is 1505 days in the past. Is this intentional? Checking references for intended status: Best Current Practice ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Obsolete informational reference (is this intentional?): RFC 2671 (Obsoleted by RFC 6891) Summary: 0 errors (**), 0 flaws (~~), 2 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group M. Andrews 3 Internet-Draft R. Bellis 4 Intended status: Best Current Practice ISC 5 Expires: September 12, 2020 March 11, 2020 7 A Common Operational Problem in DNS Servers - Failure To Communicate. 8 draft-ietf-dnsop-no-response-issue-17 10 Abstract 12 The DNS is a query / response protocol. Failing to respond to 13 queries, or responding incorrectly, causes both immediate operational 14 problems and long term problems with protocol development. 16 This document identifies a number of common kinds of queries to which 17 some servers either fail to respond or else respond incorrectly. 18 This document also suggests procedures for zone operators to apply to 19 identify and remediate the problem. 21 The document does not look at the DNS data itself, just the structure 22 of the responses. 24 Status of This Memo 26 This Internet-Draft is submitted in full conformance with the 27 provisions of BCP 78 and BCP 79. 29 Internet-Drafts are working documents of the Internet Engineering 30 Task Force (IETF). Note that other groups may also distribute 31 working documents as Internet-Drafts. The list of current Internet- 32 Drafts is at https://datatracker.ietf.org/drafts/current/. 34 Internet-Drafts are draft documents valid for a maximum of six months 35 and may be updated, replaced, or obsoleted by other documents at any 36 time. It is inappropriate to use Internet-Drafts as reference 37 material or to cite them other than as "work in progress." 39 This Internet-Draft will expire on September 12, 2020. 41 Copyright Notice 43 Copyright (c) 2020 IETF Trust and the persons identified as the 44 document authors. All rights reserved. 46 This document is subject to BCP 78 and the IETF Trust's Legal 47 Provisions Relating to IETF Documents 48 (https://trustee.ietf.org/license-info) in effect on the date of 49 publication of this document. Please review these documents 50 carefully, as they describe your rights and restrictions with respect 51 to this document. Code Components extracted from this document must 52 include Simplified BSD License text as described in Section 4.e of 53 the Trust Legal Provisions and are provided without warranty as 54 described in the Simplified BSD License. 56 Table of Contents 58 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 59 2. Consequences . . . . . . . . . . . . . . . . . . . . . . . . 4 60 3. Common kinds of queries that result in no or bad responses. . 5 61 3.1. Basic DNS Queries . . . . . . . . . . . . . . . . . . . . 5 62 3.1.1. Zone Existence . . . . . . . . . . . . . . . . . . . 5 63 3.1.2. Unknown / Unsupported Type Queries . . . . . . . . . 5 64 3.1.3. DNS Flags . . . . . . . . . . . . . . . . . . . . . . 6 65 3.1.4. Unknown DNS opcodes . . . . . . . . . . . . . . . . . 6 66 3.1.5. TCP Queries . . . . . . . . . . . . . . . . . . . . . 6 67 3.2. EDNS Queries . . . . . . . . . . . . . . . . . . . . . . 6 68 3.2.1. EDNS Queries - Version Independent . . . . . . . . . 7 69 3.2.2. EDNS Queries - Version Specific . . . . . . . . . . . 7 70 3.2.3. EDNS Options . . . . . . . . . . . . . . . . . . . . 7 71 3.2.4. EDNS Flags . . . . . . . . . . . . . . . . . . . . . 7 72 3.2.5. Truncated EDNS Responses . . . . . . . . . . . . . . 8 73 3.2.6. DO=1 Handling . . . . . . . . . . . . . . . . . . . . 8 74 3.2.7. EDNS over TCP . . . . . . . . . . . . . . . . . . . . 8 75 4. Firewalls and Load Balancers . . . . . . . . . . . . . . . . 8 76 5. Scrubbing Services . . . . . . . . . . . . . . . . . . . . . 9 77 6. Whole Answer Caches . . . . . . . . . . . . . . . . . . . . . 10 78 7. Response Code Selection . . . . . . . . . . . . . . . . . . . 10 79 8. Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 80 8.1. Testing - Basic DNS . . . . . . . . . . . . . . . . . . . 11 81 8.1.1. Is The Server Configured For The Zone? . . . . . . . 11 82 8.1.2. Testing Unknown Types . . . . . . . . . . . . . . . . 12 83 8.1.3. Testing Header Bits . . . . . . . . . . . . . . . . . 13 84 8.1.4. Testing Unknown Opcodes . . . . . . . . . . . . . . . 15 85 8.1.5. Testing TCP . . . . . . . . . . . . . . . . . . . . . 15 86 8.2. Testing - Extended DNS . . . . . . . . . . . . . . . . . 16 87 8.2.1. Testing Minimal EDNS . . . . . . . . . . . . . . . . 16 88 8.2.2. Testing EDNS Version Negotiation . . . . . . . . . . 17 89 8.2.3. Testing Unknown EDNS Options . . . . . . . . . . . . 17 90 8.2.4. Testing Unknown EDNS Flags . . . . . . . . . . . . . 18 91 8.2.5. Testing EDNS Version Negotiation With Unknown EDNS 92 Flags . . . . . . . . . . . . . . . . . . . . . . . . 18 93 8.2.6. Testing EDNS Version Negotiation With Unknown EDNS 94 Options . . . . . . . . . . . . . . . . . . . . . . . 19 95 8.2.7. Testing Truncated Responses . . . . . . . . . . . . . 20 96 8.2.8. Testing DO=1 Handling . . . . . . . . . . . . . . . . 20 97 8.2.9. Testing EDNS Version Negotiation With DO=1 . . . . . 21 98 8.2.10. Testing With Multiple Defined EDNS Options . . . . . 22 99 8.3. When EDNS Is Not Supported . . . . . . . . . . . . . . . 22 100 9. Remediation . . . . . . . . . . . . . . . . . . . . . . . . . 22 101 10. Security Considerations . . . . . . . . . . . . . . . . . . . 24 102 11. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 24 103 12. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 24 104 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 24 105 13.1. Normative References . . . . . . . . . . . . . . . . . . 24 106 13.2. Informative References . . . . . . . . . . . . . . . . . 25 107 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 26 109 1. Introduction 111 The DNS [RFC1034], [RFC1035] is a query / response protocol. Failing 112 to respond to queries, or responding incorrectly, causes both 113 immediate operational problems and long term problems with protocol 114 development. 116 Failure to respond to a query is indistinguishable from packet loss 117 without doing an analysis of query-response patterns. Additionally 118 failure to respond results in unnecessary queries being made by DNS 119 clients, and introduces delays to the resolution process. 121 Due to the inability to distinguish between packet loss and 122 nameservers dropping EDNS [RFC6891] queries, packet loss is sometimes 123 misclassified as lack of EDNS support which can lead to DNSSEC 124 validation failures. 126 The existence of servers which fail to respond to queries results in 127 developers being hesitant to deploy new standards. Such servers need 128 to be identified and remediated. 130 The DNS has response codes that cover almost any conceivable query 131 response. A nameserver should be able to respond to any conceivable 132 query using them. There should be no need to drop queries because a 133 nameserver does not understand them. 135 Unless a nameserver is under attack, it should respond to all DNS 136 requests directed to it. When a nameserver is under attack it may 137 wish to drop packets. A common attack is to use a nameserver as an 138 amplifier by sending spoofed packets. This is done because response 139 packets are bigger than the queries and large amplification factors 140 are available especially if EDNS is supported. Limiting the rate of 141 responses is reasonable when this is occurring and the client should 142 retry. This however only works if legitimate clients are not being 143 forced to guess whether EDNS queries are accepted or not. While 144 there is still a pool of servers that don't respond to EDNS requests, 145 clients have no way to know if the lack of response is due to packet 146 loss, or EDNS packets not being supported, or rate limiting due to 147 the server being under attack. Misclassification of server behaviour 148 is unavoidable when rate limiting is used until the population of 149 servers which fail to respond to well-formed queries drops to near 150 zero. 152 Nameservers should respond to queries even if the queried name is not 153 for any name the server is configured to answer for. Misconfigured 154 nameservers are a common occurrence in the DNS and receiving queries 155 for zones that the server is not configured for is not necessarily an 156 indication that the server is under attack. Parent zone operators 157 are advised to regularly check that the delegating NS records are 158 consistent with those of the delegated zone and to correct them when 159 they are not [RFC1034]. Doing this regularly should reduce the 160 instances of broken delegations. 162 This document does not try to identify all possible errors nor does 163 it supply an exhaustive list of tests. 165 2. Consequences 167 Failure to follow the relevant DNS RFCs has multiple adverse 168 consequences. Some are caused directly from the non-compliant 169 behaviour and others as a result of work-arounds forced on recursive 170 servers. Addressing known issues now will reduce future 171 interoperability issues as the DNS protocol continues to evolve and 172 clients make use of newly-introduced DNS features. In particular the 173 base DNS specification [RFC1034], [RFC1035] and the EDNS 174 specification [RFC6891], when implemented, need to be followed. 176 Some examples of known consequences include: 178 o The AD flag bit in a response cannot be trusted to mean anything 179 as some servers incorrectly copy the flag bit from the request to 180 the response [RFC1035], [RFC4035]. 182 o Widespread non-response to EDNS queries has lead to recursive 183 servers having to assume that EDNS is not supported and that 184 fallback to plain DNS is required, potentially causing DNSSEC 185 validation failures. 187 o Widespread non-response to EDNS options, requires recursive 188 servers to have to decide whether to probe to see if it is the 189 EDNS option or just EDNS that is causing the non response. In the 190 limited amount of time required to resolve a query before the 191 client times out this is not possible. 193 o Incorrectly returning FORMERR to an EDNS option being present, 194 leads to the recursive server not being able to determine if the 195 server is just broken in the handling of the EDNS option or 196 doesn't support EDNS at all. 198 o Mishandling of unknown query types has contributed to the 199 abandonment of the transition of the SPF type. 201 o Mishandling of unknown query types has slowed up the development 202 of DANE and resulted in additional rules being specified to reduce 203 the probability of interacting with a broken server when making 204 TLSA queries. 206 The consequences of servers not following the RFCs will only grow if 207 measures are not put in place to remove non compliant servers from 208 the ecosystem. Working around issues due to non-compliance with RFCs 209 is not sustainable. 211 Most (if not all) of these consequences could have been avoided if 212 action had been taken to remove non-compliant servers as soon as 213 people were aware of them, i.e. to actively seek out broken 214 implementations and servers and inform their developers and operators 215 that they need to fix their servers. 217 3. Common kinds of queries that result in no or bad responses. 219 This section is broken down into Basic DNS requests and EDNS 220 requests. 222 3.1. Basic DNS Queries 224 3.1.1. Zone Existence 226 If a zone is delegated to a server, that server should respond to an 227 SOA query for that zone with an SOA record. Failing to respond at 228 all is always incorrect, regardless of the configuration of the 229 server. Responding with anything other than an SOA record in the 230 Answer section indicates a bad delegation. 232 3.1.2. Unknown / Unsupported Type Queries 234 Some servers fail to respond to unknown or unsupported types. If a 235 server receives a query for a type that it doesn't recognise, or 236 doesn't implement, it is expected to return the appropriate response 237 as if it did recognise the type but does not have any data for that 238 type: either NOERROR, or NXDOMAIN. The exception to this are queries 239 for Meta-RR types which may return NOTIMP. 241 3.1.3. DNS Flags 243 Some servers fail to respond to DNS queries with various DNS flags 244 set, regardless of whether they are defined or still reserved. At 245 the time of writing there are servers that fail to respond to queries 246 with the AD bit set to 1 and servers that fail to respond to queries 247 with the last reserved flag bit set. 249 Servers should respond to such queries. If the server does not know 250 the meaning of a flag bit it must not copy it to the response 251 [RFC1035] Section 4.1.1. If the server does not understand the 252 meaning of a request it should reply with a FORMERR response with 253 unknown flags set to zero. 255 3.1.3.1. Recursive Queries 257 A non-recursive server is supposed to respond to recursive queries as 258 if the RD bit is not set [RFC1034]. 260 3.1.4. Unknown DNS opcodes 262 The use of previously undefined opcodes is to be expected. Since the 263 DNS was first defined two new opcodes have been added, UPDATE and 264 NOTIFY. 266 NOTIMP is the expected rcode to an unknown or unimplemented opcode. 268 Note: while new opcodes will most probably use the current layout 269 structure for the rest of the message there is no requirement that 270 anything other than the DNS header match. 272 3.1.5. TCP Queries 274 All DNS servers are supposed to respond to queries over TCP 275 [RFC7766]. While firewalls should not block TCP connection attempts 276 if they do they should cleanly terminate the connection by sending 277 TCP RESET or sending ICMP/ICMPv6 Administratively Prohibited 278 messages. Dropping TCP connections introduces excessive delays to 279 the resolution process. 281 3.2. EDNS Queries 283 EDNS queries are specified in [RFC6891]. 285 3.2.1. EDNS Queries - Version Independent 287 Identifying servers that fail to respond to EDNS queries can be done 288 by first confirming that the server responds to regular DNS queries, 289 followed by a series of otherwise identical queries using EDNS, then 290 making the original query again. A series of EDNS queries is needed 291 as at least one DNS implementation responds to the first EDNS query 292 with FORMERR but fails to respond to subsequent queries from the same 293 address for a period until a regular DNS query is made. The EDNS 294 query should specify a UDP buffer size of 512 bytes to avoid false 295 classification of not supporting EDNS due to response packet size. 297 If the server responds to the first and last queries but fails to 298 respond to most or all of the EDNS queries, it is probably faulty. 299 The test should be repeated a number of times to eliminate the 300 likelihood of a false positive due to packet loss. 302 Firewalls may also block larger EDNS responses but there is no easy 303 way to check authoritative servers to see if the firewall is mis- 304 configured. 306 3.2.2. EDNS Queries - Version Specific 308 Some servers respond correctly to EDNS version 0 queries but fail to 309 respond to EDNS queries with version numbers that are higher than 310 zero. Servers should respond with BADVERS to EDNS queries with 311 version numbers that they do not support. 313 Some servers respond correctly to EDNS version 0 queries but fail to 314 set QR=1 when responding to EDNS versions they do not support. Such 315 answers responses may be discarded as invalid (as QR is not 1) or 316 treated as requests (when the source port of the original request was 317 port 53). 319 3.2.3. EDNS Options 321 Some servers fail to respond to EDNS queries with EDNS options set. 322 The original EDNS specification left this behaviour undefined 323 [RFC2671], but the correct behaviour was clarified in [RFC6891]. 324 Unknown EDNS options are supposed to be ignored by the server. 326 3.2.4. EDNS Flags 328 Some servers fail to respond to EDNS queries with EDNS flags set. 329 Servers should ignore EDNS flags they do not understand and must not 330 add them to the response [RFC6891]. 332 3.2.5. Truncated EDNS Responses 334 Some EDNS aware servers fail to include an OPT record when a 335 truncated response is sent. An OPT record is supposed to be included 336 in a truncated response [RFC6891]. 338 Some EDNS aware servers fail to honour the advertised EDNS UDP buffer 339 size and send over-sized responses [RFC6891]. Servers must send UDP 340 responses no larger than the advertised EDNS UDP buffer size. 342 3.2.6. DO=1 Handling 344 Some nameservers incorrectly only return an EDNS response when the DO 345 bit [RFC3225] is 1 in the query. Servers that support EDNS should 346 always respond to EDNS requests with EDNS responses. 348 Some nameservers fail to copy the DO bit to the response despite 349 clearly supporting DNSSEC by returning an RRSIG records to EDNS 350 queries with DO=1. 352 3.2.7. EDNS over TCP 354 Some EDNS aware servers incorrectly limit the TCP response sizes to 355 the advertised UDP response size. This breaks DNS resolution to 356 clients where the response sizes exceed the advertised UDP response 357 size despite the server and the client being capable of sending and 358 receiving larger TCP responses respectively. It effectively defeats 359 setting TC=1 in UDP responses. 361 4. Firewalls and Load Balancers 363 Firewalls and load balancers can affect the externally visible 364 behaviour of a nameserver. Tests for conformance should to be done 365 from outside of any firewall so that the system is tested as a whole. 367 Firewalls and load balancers should not drop DNS packets that they 368 don't understand. They should either pass the packets or generate an 369 appropriate error response. 371 Requests for unknown query types are normal client behaviour and 372 should not be construed as an attack. Nameservers have always been 373 expected to be able to handle such queries. 375 Requests for unknown query classes are normal client behaviour and 376 should not be construed as an attack. Nameservers have always been 377 expected to be able to handle such queries. 379 Requests with unknown opcodes are normal client behaviour and should 380 not be construed as an attack. Nameservers have always been expected 381 to be able to handle such queries. 383 Requests with unassigned flags set (DNS or EDNS) are expected client 384 behaviour and should not be construed as an attack. The behaviour 385 for unassigned flags is to ignore them in the request and to not set 386 them in the response. Dropping DNS / EDNS packets with unassigned 387 flags makes it difficult to deploy extensions that make use of them 388 due to the need to reconfigure and update firewalls. 390 Requests with unknown EDNS options are expected client behaviour and 391 should not be construed as an attack. The correct behaviour for 392 unknown EDNS options is to ignore their presence when constructing a 393 reply. 395 Requests with unknown EDNS versions are expected client behaviour and 396 should not be construed as an attack. The correct behaviour for 397 unknown EDNS versions is to return BADVERS along with the highest 398 EDNS version the server supports. Dropping EDNS packets breaks EDNS 399 version negotiation. 401 Firewalls should not assume that there will only be a single response 402 message to a request. There have been proposals to use EDNS to 403 signal that multiple DNS messages be returned rather than a single 404 UDP message that is fragmented at the IP layer. 406 DNS, and EDNS in particular, are designed to allow clients to be able 407 to use new features against older servers without having to validate 408 every option. Indiscriminate blocking of messages breaks that 409 design. 411 However, there may be times when a nameserver mishandles messages 412 with a particular flag, EDNS option, EDNS version field, opcode, type 413 or class field or combination thereof to the point where the 414 integrity of the nameserver is compromised. Firewalls should offer 415 the ability to selectively reject messages using an appropriately 416 constructed response based on all these fields while awaiting a fix 417 from the nameserver vendor. 419 5. Scrubbing Services 421 Scrubbing services can affect the externally visible behaviour of a 422 nameserver in a similar way to firewalls. If an operator uses a 423 scrubbing service, they should check that legitimate queries are not 424 being blocked. 426 Scrubbing services, unlike firewalls, are also turned on and off in 427 response to denial of service attacks. One needs to take care when 428 choosing a scrubbing service. 430 Ideally, Operators should run these tests against a scrubbing service 431 to ensure that these tests are not seen as attack vectors. 433 6. Whole Answer Caches 435 Whole answer caches take a previously constructed answer and return 436 it to a subsequent query for the same question. However, they can 437 return the wrong response if they do not take all of the relevant 438 attributes of the query into account. 440 In addition to the standard tuple of a non- 441 exhaustive set of attributes that must be considered include: RD, AD, 442 CD, OPT record, DO, EDNS buffer size, EDNS version, EDNS options, and 443 transport. 445 7. Response Code Selection 447 Choosing the correct response code when responding to DNS queries is 448 important. Response codes should be chosen considering how clients 449 will handle them. 451 For unimplemented opcodes NOTIMP is the expected response code. 452 Note: Newly implemented opcodes may change the message format by 453 extending the header, changing the structure of the records, etc. 454 Servers are not expected to be able to parse these, and should 455 respond with a response code of NOTIMP rather than FORMERR (which 456 would be expected if there was a parse error with an known opcode). 458 For unimplemented type codes, and in the absence of other errors, the 459 only valid response is NoError if the qname exists, and NameError 460 (NXDOMAIN) otherwise. For Meta-RRs NOTIMP may be returned instead. 462 If a zone cannot be loaded because it contains unimplemented type 463 codes that are not encoded as unknown record types according to 464 [RFC3597] then the expected response is SERVFAIL as the whole zone 465 should be rejected Section 5.2 [RFC1035]. If a zone loads then 466 Section 4.3.2 [RFC1034] applies. 468 If the server supports EDNS and receives a query with an unsupported 469 EDNS version, the correct response is BADVERS [RFC6891]. 471 If the server does not support EDNS at all, FORMERR is the expected 472 error code. That said a minimal EDNS server implementation requires 473 parsing the OPT records and responding with an empty OPT record in 474 the additional section in most cases. There is no need to interpret 475 any EDNS options present in the request as unsupported EDNS options 476 are expected to be ignored [RFC6891]. Additionally EDNS flags can be 477 ignored. The only part of the OPT record that needs to be examined 478 is the version field to determine if BADVERS needs to be sent or not. 480 8. Testing 482 Testing is divided into two sections. "Basic DNS", which all servers 483 should meet, and "Extended DNS", which should be met by all servers 484 that support EDNS (a server is deemed to support EDNS if it gives a 485 valid EDNS response to any EDNS query). If a server does not support 486 EDNS it should still respond to all the tests. 488 These tests query for records at the apex of a zone that the server 489 is nominally configured to serve. All tests should use the same 490 zone. 492 It is advisable to run all of the tests below in parallel so as to 493 minimise the delays due to multiple timeouts when the servers do not 494 respond. There are 16 queries directed to each nameserver (assuming 495 no packet loss) testing different aspects of Basic DNS and Extended 496 DNS. 498 The tests below use dig from BIND 9.11.0. 500 When testing recursive servers set RD=1 and choose a zone name that 501 is know to exist and is not being served by the recursive server. 502 The root zone (".") is often a good candidate as it is DNSSEC signed. 503 RD=1, rather than RD=0, should be present in the responses for all 504 test involving the opcode QUERY. Non-authoritative answers (AA=0) 505 are expected when talking to a recursive server. AD=1 is only 506 expected if the server is validating responses and one or both AD=1 507 or DO=1 is set in the request otherwise AD=0 is expected. 509 8.1. Testing - Basic DNS 511 This first set of tests cover basic DNS server behaviour and all 512 servers should pass these tests. 514 8.1.1. Is The Server Configured For The Zone? 516 Ask for the SOA record of the configured zone. This query is made 517 with no DNS flag bits set and without EDNS. 519 We expect the SOA record for the zone to be returned in the answer 520 section, the rcode to be set to NOERROR, and the AA and QR bits to be 521 set in the header; RA may also be set [RFC1034]. We do not expect an 522 OPT record to be returned [RFC6891]. 524 Verify the server is configured for the zone: 526 dig +noedns +noad +norec soa $zone @$server 528 expect: status: NOERROR 529 expect: the SOA record to be present in the answer section 530 expect: flag: aa to be present 531 expect: flag: rd to NOT be present 532 expect: flag: ad to NOT be present 533 expect: the OPT record to NOT be present 535 8.1.2. Testing Unknown Types 537 Identifying servers that fail to respond to unknown or unsupported 538 types can be done by making an initial DNS query for an A record, 539 making a number of queries for an unallocated type, then making a 540 query for an A record again. IANA maintains a registry of allocated 541 types. 543 If the server responds to the first and last queries but fails to 544 respond to the queries for the unallocated type, it is probably 545 faulty. The test should be repeated a number of times to eliminate 546 the likelihood of a false positive due to packet loss. 548 Ask for the TYPE1000 RRset at the configured zone's name. This query 549 is made with no DNS flag bits set and without EDNS. TYPE1000 has 550 been chosen for this purpose as IANA is unlikely to allocate this 551 type in the near future and it is not in a range reserved for private 552 use [RFC6895]. Any unallocated type code could be chosen for this 553 test. 555 We expect no records to be returned in the answer section, the rcode 556 to be set to NOERROR, and the AA and QR bits to be set in the header; 557 RA may also be set [RFC1034]. We do not expect an OPT record to be 558 returned [RFC6891]. 560 Check that queries for an unknown type work: 562 dig +noedns +noad +norec type1000 $zone @$server 564 expect: status: NOERROR 565 expect: an empty answer section. 566 expect: flag: aa to be present 567 expect: flag: rd to NOT be present 568 expect: flag: ad to NOT be present 569 expect: the OPT record to NOT be present 571 8.1.3. Testing Header Bits 573 8.1.3.1. Testing CD=1 Queries 575 Ask for the SOA record of the configured zone. This query is made 576 with only the CD DNS flag bit set, all other DNS bits clear, and 577 without EDNS. 579 We expect the SOA record for the zone to be returned in the answer 580 section, the rcode to be set to NOERROR, and the AA and QR bits to be 581 set in the header. We do not expect an OPT record to be returned. 583 If the server supports DNSSEC, CD should be set in the response 584 [RFC4035] otherwise CD should be clear [RFC1034]. 586 Check that queries with CD=1 work: 588 dig +noedns +noad +norec +cd soa $zone @$server 590 expect: status: NOERROR 591 expect: the SOA record to be present in the answer section 592 expect: flag: aa to be present 593 expect: flag: rd to NOT be present 594 expect: flag: ad to NOT be present 595 expect: the OPT record to NOT be present 597 8.1.3.2. Testing AD=1 Queries 599 Ask for the SOA record of the configured zone. This query is made 600 with only the AD DNS flag bit set and all other DNS bits clear and 601 without EDNS. 603 We expect the SOA record for the zone to be returned in the answer 604 section, the rcode to be set to NOERROR, and the AA and QR bits to be 605 set in the header. We do not expect an OPT record to be returned. 606 The purpose of this query is to detect blocking of queries with the 607 AD bit present, not the specific value of AD in the response. 609 Check that queries with AD=1 work: 611 dig +noedns +norec +ad soa $zone @$server 613 expect: status: NOERROR 614 expect: the SOA record to be present in the answer section 615 expect: flag: aa to be present 616 expect: flag: rd to NOT be present 617 expect: the OPT record to NOT be present 619 AD use in queries is defined in [RFC6840]. 621 8.1.3.3. Testing Reserved Bit 623 Ask for the SOA record of the configured zone. This query is made 624 with only the final reserved DNS flag bit set and all other DNS bits 625 clear and without EDNS. 627 We expect the SOA record for the zone to be returned in the answer 628 section, the rcode to be set to NOERROR, and the AA and QR bits to be 629 set in the header; RA may be set. The final reserved bit must not be 630 set [RFC1034]. We do not expect an OPT record to be returned 631 [RFC6891]. 633 Check that queries with the last unassigned DNS header flag work and 634 that the flag bit is not copied to the response: 636 dig +noedns +noad +norec +zflag soa $zone @$server 638 expect: status: NOERROR 639 expect: the SOA record to be present in the answer section 640 expect: MBZ to NOT be in the response (see below) 641 expect: flag: aa to be present 642 expect: flag: rd to NOT be present 643 expect: flag: ad to NOT be present 644 expect: the OPT record to NOT be present 646 MBZ (Must Be Zero) is a dig-specific indication that the flag bit has 647 been incorrectly copied. See Section 4.1.1, [RFC1035] "Z Reserved 648 for future use. Must be zero in all queries and responses." 650 8.1.3.4. Testing Recursive Queries 652 Ask for the SOA record of the configured zone. This query is made 653 with only the RD DNS flag bit set and without EDNS. 655 We expect the SOA record for the zone to be returned in the answer 656 section, the rcode to be set to NOERROR, and the AA, QR and RD bits 657 to be set in the header; RA may also be set [RFC1034]. We do not 658 expect an OPT record to be returned [RFC6891]. 660 Check that recursive queries work: 662 dig +noedns +noad +rec soa $zone @$server 664 expect: status: NOERROR 665 expect: the SOA record to be present in the answer section 666 expect: flag: aa to be present 667 expect: flag: rd to be present 668 expect: flag: ad to NOT be present 669 expect: the OPT record to NOT be present 671 8.1.4. Testing Unknown Opcodes 673 Construct a DNS message that consists of only a DNS header with 674 opcode set to 15 (currently not allocated), no DNS header bits set 675 and empty question, answer, authority and additional sections. 677 Check that new opcodes are handled: 679 dig +noedns +noad +opcode=15 +norec +header-only @$server 681 expect: status: NOTIMP 682 expect: opcode: 15 683 expect: all sections to be empty 684 expect: flag: aa to NOT be present 685 expect: flag: rd to NOT be present 686 expect: flag: ad to NOT be present 687 expect: the OPT record to NOT be present 689 8.1.5. Testing TCP 691 Whether a server accepts TCP connections can be tested by first 692 checking that it responds to UDP queries to confirm that it is up and 693 operating, then attempting the same query over TCP. An additional 694 query should be made over UDP if the TCP connection attempt fails to 695 confirm that the server under test is still operating. 697 Ask for the SOA record of the configured zone. This query is made 698 with no DNS flag bits set and without EDNS. This query is to be sent 699 using TCP. 701 We expect the SOA record for the zone to be returned in the answer 702 section, the rcode to be set to NOERROR, and the AA and QR bits to be 703 set in the header; RA may also be set [RFC1034]. We do not expect an 704 OPT record to be returned [RFC6891]. 706 Check that TCP queries work: 708 dig +noedns +noad +norec +tcp soa $zone @$server 710 expect: status: NOERROR 711 expect: the SOA record to be present in the answer section 712 expect: flag: aa to be present 713 expect: flag: rd to NOT be present 714 expect: flag: ad to NOT be present 715 expect: the OPT record to NOT be present 717 The requirement that TCP be supported is defined in [RFC7766]. 719 8.2. Testing - Extended DNS 721 The next set of tests cover various aspects of EDNS behaviour. If 722 any of these tests succeed (indicating at least some EDNS support) 723 then all of them should succeed. There are servers that support EDNS 724 but fail to handle plain EDNS queries correctly so a plain EDNS query 725 is not a good indicator of lack of EDNS support. 727 8.2.1. Testing Minimal EDNS 729 Ask for the SOA record of the configured zone. This query is made 730 with no DNS flag bits set. EDNS version 0 is used without any EDNS 731 options or EDNS flags set. 733 We expect the SOA record for the zone to be returned in the answer 734 section, the rcode to be set to NOERROR, and the AA and QR bits to be 735 set in the header; RA may also be set [RFC1034]. We expect an OPT 736 record to be returned. There should be no EDNS flags present in the 737 response. The EDNS version field should be 0 and there should be no 738 EDNS options present [RFC6891]. 740 Check that plain EDNS queries work: 742 dig +nocookie +edns=0 +noad +norec soa $zone @$server 744 expect: status: NOERROR 745 expect: the SOA record to be present in the answer section 746 expect: an OPT record to be present in the additional section 747 expect: EDNS Version 0 in response 748 expect: flag: aa to be present 749 expect: flag: ad to NOT be present 751 +nocookie disables sending a EDNS COOKIE option which is otherwise 752 enabled by default in BIND 9.11.0 (and later). 754 8.2.2. Testing EDNS Version Negotiation 756 Ask for the SOA record of a zone the server is nominally configured 757 to serve. This query is made with no DNS flag bits set. EDNS 758 version 1 is used without any EDNS options or EDNS flags set. 760 We expect the SOA record for the zone to NOT be returned in the 761 answer section with the extended rcode set to BADVERS and the QR bit 762 to be set in the header; RA may also be set [RFC1034]. We expect an 763 OPT record to be returned. There should be no EDNS flags present in 764 the response. The EDNS version field should be 0 in the response as 765 no other EDNS version has as yet been specified [RFC6891]. 767 Check that EDNS version 1 queries work (EDNS supported): 769 dig +nocookie +edns=1 +noednsneg +noad +norec soa $zone @$server 771 expect: status: BADVERS 772 expect: the SOA record to NOT be present in the answer section 773 expect: an OPT record to be present in the additional section 774 expect: EDNS Version 0 in response 775 expect: flag: aa to NOT be present 776 expect: flag: ad to NOT be present 778 +noednsneg has been set as dig supports EDNS version negotiation and 779 we want to see only the response to the initial EDNS version 1 query. 781 8.2.3. Testing Unknown EDNS Options 783 Ask for the SOA record of the configured zone. This query is made 784 with no DNS flag bits set. EDNS version 0 is used without any EDNS 785 flags. An EDNS option is present with a value that has not yet been 786 assigned by IANA. We have picked an unassigned code of 100 for the 787 example below. Any unassigned EDNS option code could have been 788 choosen for this test. 790 We expect the SOA record for the zone to be returned in the answer 791 section, the rcode to be set to NOERROR, and the AA and QR bits to be 792 set in the header; RA may also be set [RFC1034]. We expect an OPT 793 record to be returned. There should be no EDNS flags present in the 794 response. The EDNS version field should be 0 as EDNS versions other 795 than 0 are yet to be specified and there should be no EDNS options 796 present as unknown EDNS options are supposed to be ignored by the 797 server [RFC6891] Section 6.1.2. 799 Check that EDNS queries with an unknown option work (EDNS supported): 801 dig +nocookie +edns=0 +noad +norec +ednsopt=100 soa $zone @$server 803 expect: status: NOERROR 804 expect: the SOA record to be present in the answer section 805 expect: an OPT record to be present in the additional section 806 expect: OPT=100 to NOT be present 807 expect: EDNS Version 0 in response 808 expect: flag: aa to be present 809 expect: flag: ad to NOT be present 811 8.2.4. Testing Unknown EDNS Flags 813 Ask for the SOA record of the configured zone. This query is made 814 with no DNS flag bits set. EDNS version 0 is used without any EDNS 815 options. An unassigned EDNS flag bit is set (0x40 in this case). 817 We expect the SOA record for the zone to be returned in the answer 818 section, the rcode to be set to NOERROR, and the AA and QR bits to be 819 set in the header; RA may also be set [RFC1034]. We expect an OPT 820 record to be returned. There should be no EDNS flags present in the 821 response as unknown EDNS flags are supposed to be ignored. The EDNS 822 version field should be 0 and there should be no EDNS options present 823 [RFC6891]. 825 Check that EDNS queries with unknown flags work (EDNS supported): 827 dig +nocookie +edns=0 +noad +norec +ednsflags=0x40 soa $zone @$server 829 expect: status: NOERROR 830 expect: the SOA record to be present in the answer section 831 expect: an OPT record to be present in the additional section 832 expect: MBZ not to be present 833 expect: EDNS Version 0 in response 834 expect: flag: aa to be present 835 expect: flag: ad to NOT be present 837 MBZ (Must Be Zero) is a dig-specific indication that a flag bit has 838 been incorrectly copied as per Section 6.1.4, [RFC6891]. 840 8.2.5. Testing EDNS Version Negotiation With Unknown EDNS Flags 842 Ask for the SOA record of the configured zone. This query is made 843 with no DNS flag bits set. EDNS version 1 is used without any EDNS 844 options. An unassigned EDNS flag bit is set (0x40 in this case). 846 We expect the SOA record for the zone to NOT be returned in the 847 answer section with the extended rcode set to BADVERS and the QR bit 848 to be set in the header; RA may also be set [RFC1034]. We expect an 849 OPT record to be returned. There should be no EDNS flags present in 850 the response as unknown EDNS flags are supposed to be ignored. The 851 EDNS version field should be 0 as EDNS versions other than 0 are yet 852 to be specified and there should be no EDNS options present 853 [RFC6891]. 855 Check that EDNS version 1 queries with unknown flags work (EDNS 856 supported): 858 dig +nocookie +edns=1 +noednsneg +noad +norec +ednsflags=0x40 soa \ 859 $zone @$server 861 expect: status: BADVERS 862 expect: SOA record to NOT be present 863 expect: an OPT record to be present in the additional section 864 expect: MBZ not to be present 865 expect: EDNS Version 0 in response 866 expect: flag: aa to NOT be present 867 expect: flag: ad to NOT be present 869 8.2.6. Testing EDNS Version Negotiation With Unknown EDNS Options 871 Ask for the SOA record of the configured zone. This query is made 872 with no DNS flag bits set. EDNS version 1 is used. An unknown EDNS 873 option is present. We have picked an unassigned code of 100 for the 874 example below. Any unassigned EDNS option code could have been 875 chosen for this test. 877 We expect the SOA record for the zone to NOT be returned in the 878 answer section with the extended rcode set to BADVERS and the QR bit 879 to be set in the header; RA may also be set [RFC1034]. We expect an 880 OPT record to be returned. There should be no EDNS flags present in 881 the response. The EDNS version field should be 0 as EDNS versions 882 other than 0 are yet to be specified and there should be no EDNS 883 options present [RFC6891]. 885 Check that EDNS version 1 queries with unknown options work (EDNS 886 supported): 888 dig +nocookie +edns=1 +noednsneg +noad +norec +ednsopt=100 soa \ 889 $zone @$server 891 expect: status: BADVERS 892 expect: SOA record to NOT be present 893 expect: an OPT record to be present in the additional section 894 expect: OPT=100 to NOT be present 895 expect: EDNS Version 0 in response 896 expect: flag: aa to NOT be present 897 expect: flag: ad to NOT be present 899 8.2.7. Testing Truncated Responses 901 Ask for the DNSKEY records of the configured zone, which must be a 902 DNSSEC signed zone. This query is made with no DNS flag bits set. 903 EDNS version 0 is used without any EDNS options. The only EDNS flag 904 set is DO. The EDNS UDP buffer size is set to 512. The intention of 905 this query is to elicit a truncated response from the server. Most 906 signed DNSKEY responses are bigger than 512 bytes. This test will 907 not give a valid result if the zone is not signed. 909 We expect a response, the rcode to be set to NOERROR, and the AA and 910 QR bits to be set, AD may be set in the response if the server 911 supports DNSSEC otherwise it should be clear; TC and RA may also be 912 set [RFC1035] [RFC4035]. We expect an OPT record to be present in 913 the response. There should be no EDNS flags other than DO present in 914 the response. The EDNS version field should be 0 and there should be 915 no EDNS options present [RFC6891]. 917 If TC is not set it is not possible to confirm that the server 918 correctly adds the OPT record to the truncated responses or not. 920 dig +norec +dnssec +bufsize=512 +ignore dnskey $zone @$server 921 expect: NOERROR 922 expect: OPT record with version set to 0 924 8.2.8. Testing DO=1 Handling 926 Ask for the SOA record of the configured zone, which does not need to 927 be DNSSEC signed. This query is made with no DNS flag bits set. 928 EDNS version 0 is used without any EDNS options. The only EDNS flag 929 set is DO. 931 We expect the SOA record for the zone to be returned in the answer 932 section, the rcode to be set to NOERROR, and the AA and QR bits to be 933 set in the response, AD may be set in the response if the server 934 supports DNSSEC otherwise it should be clear; RA may also be set 935 [RFC1034]. We expect an OPT record to be returned. There should be 936 no EDNS flags other than DO present in the response which should be 937 present if the server supports DNSSEC. The EDNS version field should 938 be 0 and there should be no EDNS options present [RFC6891]. 940 Check that DO=1 queries work (EDNS supported): 942 dig +nocookie +edns=0 +noad +norec +dnssec soa $zone @$server 944 expect: status: NOERROR 945 expect: the SOA record to be present in the answer section 946 expect: an OPT record to be present in the additional section 947 expect: DO=1 to be present if an RRSIG is in the response 948 expect: EDNS Version 0 in response 949 expect: flag: aa to be present 951 8.2.9. Testing EDNS Version Negotiation With DO=1 953 Ask for the SOA record of the configured zone, which does not need to 954 be DNSSEC signed. This query is made with no DNS flag bits set. 955 EDNS version 1 is used without any EDNS options. The only EDNS flag 956 set is DO. 958 We expect the SOA record for the zone to NOT be returned in the 959 answer section, the rcode to be set to NOERROR, ; the QR bit and 960 possibly the RA bit to be set [RFC1034]. We expect an OPT record to 961 be returned. There should be no EDNS flags other than DO present in 962 the response which should be there if the server supports DNSSEC. 963 The EDNS version field should be 0 and there should be no EDNS 964 options present [RFC6891]. 966 Check that EDNS version 1, DO=1 queries work (EDNS supported): 968 dig +nocookie +edns=1 +noednsneg +noad +norec +dnssec soa \ 969 $zone @$server 971 expect: status: BADVERS 972 expect: SOA record to NOT be present 973 expect: an OPT record to be present in the additional section 974 expect: DO=1 to be present if the EDNS version 0 DNSSEC query test 975 returned DO=1 976 expect: EDNS Version 0 in response 977 expect: flag: aa to NOT be present 979 8.2.10. Testing With Multiple Defined EDNS Options 981 Ask for the SOA record of the configured zone. This query is made 982 with no DNS flag bits set. EDNS version 0 is used. A number of 983 defined EDNS options are present (NSID [RFC5001], DNS COOKIE 984 [RFC7873], EDNS Client Subnet [RFC7871] and EDNS Expire [RFC7314]). 986 We expect the SOA record for the zone to be returned in the answer 987 section, the rcode to be set to NOERROR, and the AA and QR bits to be 988 set in the header; RA may also be set [RFC1034]. We expect an OPT 989 record to be returned. There should be no EDNS flags present in the 990 response. The EDNS version field should be 0. Any of the requested 991 EDNS options supported by the server and permitted server 992 configuration may be returned [RFC6891]. 994 Check that EDNS queries with multiple defined EDNS options work: 996 dig +edns=0 +noad +norec +cookie +nsid +expire +subnet=0.0.0.0/0 \ 997 soa $zone @$server 999 expect: status: NOERROR 1000 expect: the SOA record to be present in the answer section 1001 expect: an OPT record to be present in the additional section 1002 expect: EDNS Version 0 in response 1003 expect: flag: aa to be present 1004 expect: flag: ad to NOT be present 1006 8.3. When EDNS Is Not Supported 1008 If EDNS is not supported by the nameserver, we expect a response to 1009 each of the above queries. That response may be a FORMERR error 1010 response or the OPT record may just be ignored. 1012 Some nameservers only return a EDNS response when a particular EDNS 1013 option or flag (e.g. DO=1) is present in the request. This 1014 behaviour is not compliant behaviour and may hide other incorrect 1015 behaviour from the above tests. Re-testing with the triggering 1016 option / flag present will expose this misbehaviour. 1018 9. Remediation 1020 Name server operators are generally expected to test their own 1021 infrastructure for compliance to standards. The above tests should 1022 be run when new systems are brought online, and should be repeated 1023 periodically to ensure continued interoperability. 1025 Domain registrants who do not maintain their own DNS infrastructure 1026 are entitled to a DNS service that conforms to standards and 1027 interoperates well. Registrants who become aware that their DNS 1028 operator does not have a well maintained or compliant infrastructure 1029 should insist that their service provider correct issues, and switch 1030 providers if they do not. 1032 In the event that an operator experiences problems due to the 1033 behaviour of name servers outside their control, the above tests will 1034 help in narrowing down the precise issue(s) which can then be 1035 reported to the relevant party. 1037 If contact information for the operator of a misbehaving name server 1038 is not already known, the following methods of communication could be 1039 considered: 1041 o the RNAME of the zone authoritative for the name of the 1042 misbehaving server 1044 o the RNAME of zones for which the offending server is authoritative 1046 o administrative or technical contacts listed in the registration 1047 information for the parent domain of the name of the misbehaving 1048 server, or for zones for which the name server is authoritative 1050 o the registrar or registry for such zones 1052 o DNS-specific operational fora (e.g. mailing lists) 1054 Operators of parent zones may wish to regularly test the 1055 authoritative name servers of their child zones. However, parent 1056 operators can have widely varying capabilities in terms of 1057 notification or remediation depending on whether they have a direct 1058 relationship with the child operator. Many TLD registries, for 1059 example, cannot directly contact their registrants and may instead 1060 need to communicate through the relevant registrar. In such cases 1061 it may be most efficient for registrars to take on the responsibility 1062 for testing the name servers of their registrants, since they have a 1063 direct relationship. 1065 When notification is not effective at correcting problems with a 1066 misbehaving name server, parent operators can choose to remove NS 1067 record sets (and glue records below) that refer to the faulty server 1068 until the servers are fixed. This should only be done as a last 1069 resort and with due consideration, as removal of a delegation can 1070 have unanticipated side effects. For example, other parts of the DNS 1071 tree may depend on names below the removed zone cut, and the parent 1072 operator may find themselves responsible for causing new DNS failures 1073 to occur. 1075 10. Security Considerations 1077 Testing protocol compliance can potentially result in false reports 1078 of attempts to break services from Intrusion Detection Services and 1079 firewalls. All of the tests are well-formed (though not necessarily 1080 common) DNS queries. None the tests listed above should cause any 1081 harm to a protocol-compliant server. 1083 Relaxing firewall settings to ensure EDNS compliance could 1084 potentially expose a critical implementation flaw in the nameserver. 1085 Nameservers should be tested for conformance before relaxing firewall 1086 settings. 1088 When removing delegations for non-compliant servers there can be a 1089 knock on effect on other zones that require these zones to be 1090 operational for the nameservers addresses to be resolved. 1092 11. IANA Considerations 1094 There are no actions for IANA. 1096 12. Acknowledgements 1098 The contributions of the following are gratefully acknowledged: 1100 Matthew Pounsett, Tim Wicinski. 1102 13. References 1104 13.1. Normative References 1106 [RFC1034] Mockapetris, P., "Domain names - concepts and facilities", 1107 STD 13, RFC 1034, DOI 10.17487/RFC1034, November 1987, 1108 . 1110 [RFC1035] Mockapetris, P., "Domain names - implementation and 1111 specification", STD 13, RFC 1035, DOI 10.17487/RFC1035, 1112 November 1987, . 1114 [RFC3225] Conrad, D., "Indicating Resolver Support of DNSSEC", 1115 RFC 3225, DOI 10.17487/RFC3225, December 2001, 1116 . 1118 [RFC4035] Arends, R., Austein, R., Larson, M., Massey, D., and S. 1119 Rose, "Protocol Modifications for the DNS Security 1120 Extensions", RFC 4035, DOI 10.17487/RFC4035, March 2005, 1121 . 1123 [RFC6840] Weiler, S., Ed. and D. Blacka, Ed., "Clarifications and 1124 Implementation Notes for DNS Security (DNSSEC)", RFC 6840, 1125 DOI 10.17487/RFC6840, February 2013, 1126 . 1128 [RFC6891] Damas, J., Graff, M., and P. Vixie, "Extension Mechanisms 1129 for DNS (EDNS(0))", STD 75, RFC 6891, 1130 DOI 10.17487/RFC6891, April 2013, 1131 . 1133 [RFC6895] Eastlake 3rd, D., "Domain Name System (DNS) IANA 1134 Considerations", BCP 42, RFC 6895, DOI 10.17487/RFC6895, 1135 April 2013, . 1137 [RFC7766] Dickinson, J., Dickinson, S., Bellis, R., Mankin, A., and 1138 D. Wessels, "DNS Transport over TCP - Implementation 1139 Requirements", RFC 7766, DOI 10.17487/RFC7766, March 2016, 1140 . 1142 13.2. Informative References 1144 [RFC2671] Vixie, P., "Extension Mechanisms for DNS (EDNS0)", 1145 RFC 2671, DOI 10.17487/RFC2671, August 1999, 1146 . 1148 [RFC3597] Gustafsson, A., "Handling of Unknown DNS Resource Record 1149 (RR) Types", RFC 3597, DOI 10.17487/RFC3597, September 1150 2003, . 1152 [RFC5001] Austein, R., "DNS Name Server Identifier (NSID) Option", 1153 RFC 5001, DOI 10.17487/RFC5001, August 2007, 1154 . 1156 [RFC7314] Andrews, M., "Extension Mechanisms for DNS (EDNS) EXPIRE 1157 Option", RFC 7314, DOI 10.17487/RFC7314, July 2014, 1158 . 1160 [RFC7871] Contavalli, C., van der Gaast, W., Lawrence, D., and W. 1161 Kumari, "Client Subnet in DNS Queries", RFC 7871, 1162 DOI 10.17487/RFC7871, May 2016, 1163 . 1165 [RFC7873] Eastlake 3rd, D. and M. Andrews, "Domain Name System (DNS) 1166 Cookies", RFC 7873, DOI 10.17487/RFC7873, May 2016, 1167 . 1169 Authors' Addresses 1171 M. Andrews 1172 Internet Systems Consortium 1173 950 Charter Street 1174 Redwood City, CA 94063 1175 US 1177 Email: marka@isc.org 1179 Ray Bellis 1180 Internet Systems Consortium 1181 950 Charter Street 1182 Redwood City, CA 94063 1183 US 1185 Email: ray@isc.org