idnits 2.17.1 draft-ietf-sidr-rpki-rtr-18.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 : ---------------------------------------------------------------------------- ** The abstract seems to contain references ([I-D.ietf-sidr-arch]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (October 14, 2011) is 4578 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-10) exists of draft-ietf-sidr-pfx-validate-02 ** Obsolete normative reference: RFC 5226 (Obsoleted by RFC 8126) ** Obsolete normative reference: RFC 5246 (Obsoleted by RFC 8446) Summary: 3 errors (**), 0 flaws (~~), 2 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group R. Bush 3 Internet-Draft Internet Initiative Japan 4 Intended status: Standards Track R. Austein 5 Expires: April 16, 2012 Dragon Research Labs 6 October 14, 2011 8 The RPKI/Router Protocol 9 draft-ietf-sidr-rpki-rtr-18 11 Abstract 13 In order to formally validate the origin ASs of BGP announcements, 14 routers need a simple but reliable mechanism to receive RPKI 15 [I-D.ietf-sidr-arch] prefix origin data from a trusted cache. This 16 document describes a protocol to deliver validated prefix origin data 17 to routers. 19 Requirements Language 21 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 22 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 23 document are to be interpreted as described in [RFC2119]. 25 Status of this Memo 27 This Internet-Draft is submitted in full conformance with the 28 provisions of BCP 78 and BCP 79. 30 Internet-Drafts are working documents of the Internet Engineering 31 Task Force (IETF). Note that other groups may also distribute 32 working documents as Internet-Drafts. The list of current Internet- 33 Drafts is at http://datatracker.ietf.org/drafts/current/. 35 Internet-Drafts are draft documents valid for a maximum of six months 36 and may be updated, replaced, or obsoleted by other documents at any 37 time. It is inappropriate to use Internet-Drafts as reference 38 material or to cite them other than as "work in progress." 40 This Internet-Draft will expire on April 16, 2012. 42 Copyright Notice 44 Copyright (c) 2011 IETF Trust and the persons identified as the 45 document authors. All rights reserved. 47 This document is subject to BCP 78 and the IETF Trust's Legal 48 Provisions Relating to IETF Documents 49 (http://trustee.ietf.org/license-info) in effect on the date of 50 publication of this document. Please review these documents 51 carefully, as they describe your rights and restrictions with respect 52 to this document. Code Components extracted from this document must 53 include Simplified BSD License text as described in Section 4.e of 54 the Trust Legal Provisions and are provided without warranty as 55 described in the Simplified BSD License. 57 Table of Contents 59 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 60 2. Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 61 3. Deployment Structure . . . . . . . . . . . . . . . . . . . . . 4 62 4. Operational Overview . . . . . . . . . . . . . . . . . . . . . 4 63 5. Protocol Data Units (PDUs) . . . . . . . . . . . . . . . . . . 5 64 5.1. Serial Notify . . . . . . . . . . . . . . . . . . . . . . 6 65 5.2. Serial Query . . . . . . . . . . . . . . . . . . . . . . . 6 66 5.3. Reset Query . . . . . . . . . . . . . . . . . . . . . . . 7 67 5.4. Cache Response . . . . . . . . . . . . . . . . . . . . . . 7 68 5.5. IPv4 Prefix . . . . . . . . . . . . . . . . . . . . . . . 8 69 5.6. IPv6 Prefix . . . . . . . . . . . . . . . . . . . . . . . 9 70 5.7. End of Data . . . . . . . . . . . . . . . . . . . . . . . 9 71 5.8. Cache Reset . . . . . . . . . . . . . . . . . . . . . . . 10 72 5.9. Error Report . . . . . . . . . . . . . . . . . . . . . . . 10 73 5.10. Fields of a PDU . . . . . . . . . . . . . . . . . . . . . 11 74 6. Protocol Sequences . . . . . . . . . . . . . . . . . . . . . . 13 75 6.1. Start or Restart . . . . . . . . . . . . . . . . . . . . . 13 76 6.2. Typical Exchange . . . . . . . . . . . . . . . . . . . . . 14 77 6.3. No Incremental Update Available . . . . . . . . . . . . . 14 78 6.4. Cache has No Data Available . . . . . . . . . . . . . . . 15 79 7. Transport . . . . . . . . . . . . . . . . . . . . . . . . . . 15 80 7.1. SSH Transport . . . . . . . . . . . . . . . . . . . . . . 16 81 7.2. TLS Transport . . . . . . . . . . . . . . . . . . . . . . 17 82 7.3. TCP MD5 Transport . . . . . . . . . . . . . . . . . . . . 17 83 7.4. TCP-AO Transport . . . . . . . . . . . . . . . . . . . . . 18 84 8. Router-Cache Set-Up . . . . . . . . . . . . . . . . . . . . . 18 85 9. Deployment Scenarios . . . . . . . . . . . . . . . . . . . . . 19 86 10. Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . 20 87 11. Security Considerations . . . . . . . . . . . . . . . . . . . 21 88 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 22 89 13. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 23 90 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 23 91 14.1. Normative References . . . . . . . . . . . . . . . . . . . 23 92 14.2. Informative References . . . . . . . . . . . . . . . . . . 24 93 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 24 95 1. Introduction 97 In order to formally validate the origin ASs of BGP announcements, 98 routers need a simple but reliable mechanism to receive RPKI 99 [I-D.ietf-sidr-arch] formally validated prefix origin data from a 100 trusted cache. This document describes a protocol to deliver 101 validated prefix origin data to routers. 103 Section 3 describes the deployment structure and Section 4 then 104 presents an operational overview. The binary payloads of the 105 protocol are formally described in Section 5, and the expected PDU 106 sequences are described in Section 6. The transport protocol options 107 are described in Section 7. Section 8 details how routers and caches 108 are configured to connect and authenticate. Section 9 describes 109 likely deployment scenarios. The traditional security and IANA 110 considerations end the document. 112 The protocol is extensible to support new PDUs with new semantics 113 when and as needed, as indicated by deployment experience. PDUs are 114 versioned should deployment experience call for change. 116 2. Glossary 118 The following terms are used with special meaning: 120 Global RPKI: The authoritative data of the RPKI are published in a 121 distributed set of servers at the IANA, RIRs, NIRs, and ISPs, see 122 [I-D.ietf-sidr-repos-struct]. 124 Cache: A coalesced copy of the RPKI which is periodically fetched/ 125 refreshed directly or indirectly from the global RPKI using the 126 [RFC5781] protocol/tools. Relying party update software such as 127 rcynic is used to gather and validate the distributed data of the 128 RPKI into a cache. Trusting this cache further is a matter 129 between the provider of the cache and a relying party. 131 Serial Number: A 32-bit monotonically increasing ordinal which wraps 132 from 2^32-1 to 0. It denotes the logical version of a cache. A 133 cache increments the value by one when it successfully updates its 134 data from a parent cache or from primary RPKI data. As a cache is 135 receiving, new incoming data and implicit deletes are marked with 136 the new serial but MUST NOT be sent until the fetch is complete. 137 A serial number is not commensurate between caches, nor need it be 138 maintained across resets of the cache server. See [RFC1982] on 139 DNS Serial Number Arithmetic for too much detail on serial number 140 arithmetic. 142 Nonce: When a cache server is started, it generates a nonce to 143 identify the instance of the cache and to bind it to the sequence 144 of Serial Numbers that cache instance will generate. This allows 145 the router to restart a failed session knowing that the Serial 146 Number it is using is commensurate with that of the cache. 148 3. Deployment Structure 150 Deployment of the RPKI to reach routers has a three level structure 151 as follows: 153 Global RPKI: The authoritative data of the RPKI are published in a 154 distributed set of servers, RPKI publication repositories, e.g. 155 the IANA, RIRs, NIRs, and ISPs, see [I-D.ietf-sidr-repos-struct]. 157 Local Caches: A local set of one or more collected and verified 158 caches. A relying party, e.g. router or other client, MUST have a 159 trust relationship with, and a trusted transport channel to, any 160 authoritative cache(s) it uses. 162 Routers: A router fetches data from a local cache using the protocol 163 described in this document. It is said to be a client of the 164 cache. There MAY be mechanisms for the router to assure itself of 165 the authenticity of the cache and to authenticate itself to the 166 cache. 168 4. Operational Overview 170 A router establishes and keeps open a connection to one or more 171 caches with which it has client/server relationships. It is 172 configured with a semi-ordered list of caches, and establishes a 173 connection to the most preferred cache, or set of caches, which 174 accept the connections. 176 The router MUST choose the most preferred, by configuration, cache or 177 set of caches so that the operator may control load on their caches 178 and the Global RPKI. 180 Periodically, the router sends to the cache the serial number of the 181 highest numbered data it has received from that cache, i.e. the 182 router's current serial number. When a router establishes a new 183 connection to a cache, or wishes to reset a current relationship, it 184 sends a Reset Query. 186 The Cache responds with all data records which have serial numbers 187 greater than that in the router's query. This may be the null set, 188 in which case the End of Data PDU is still sent. Note that 'greater' 189 must take wrap-around into account, see [RFC1982]. 191 When the router has received all data records from the cache, it sets 192 its current serial number to that of the serial number in the End of 193 Data PDU. 195 When the cache updates its database, it sends a Notify message to 196 every currently connected router. This is a hint that now would be a 197 good time for the router to poll for an update, but is only a hint. 198 The protocol requires the router to poll for updates periodically in 199 any case. 201 Strictly speaking, a router could track a cache simply by asking for 202 a complete data set every time it updates, but this would be very 203 inefficient. The serial number based incremental update mechanism 204 allows an efficient transfer of just the data records which have 205 changed since last update. As with any update protocol based on 206 incremental transfers, the router must be prepared to fall back to a 207 full transfer if for any reason the cache is unable to provide the 208 necessary incremental data. Unlike some incremental transfer 209 protocols, this protocol requires the router to make an explicit 210 request to start the fallback process; this is deliberate, as the 211 cache has no way of knowing whether the router has also established 212 sessions with other caches that may be able to provide better 213 service. 215 As a cache server must evaluate certificates and ROAs which are time 216 dependent, servers' clocks MUST be correct to a tolerance of 217 approximately an hour. 219 5. Protocol Data Units (PDUs) 221 The exchanges between the cache and the router are sequences of 222 exchanges of the following PDUs according to the rules described in 223 Section 6. 225 5.1. Serial Notify 227 The cache notifies the router that the cache has new data. 229 The Cache Nonce reassures the router that the serial numbers are 230 commensurate, i.e. the cache session has not been changed. 232 Serial Notify is only message that the cache can send that is not in 233 response to a message from the router. 235 0 8 16 24 31 236 .-------------------------------------------. 237 | Protocol | PDU | | 238 | Version | Type | Cache Nonce | 239 | 0 | 0 | | 240 +-------------------------------------------+ 241 | | 242 | Length=12 | 243 | | 244 +-------------------------------------------+ 245 | | 246 | Serial Number | 247 | | 248 `-------------------------------------------' 250 5.2. Serial Query 252 Serial Query: The router sends Serial Query to ask the cache for all 253 payload PDUs which have serial numbers higher than the serial number 254 in the Serial Query. 256 The cache replies to this query with a Cache Response PDU 257 (Section 5.4) if the cache has a, possibly null, record of the 258 changes since the serial number specified by the router. If there 259 have been no changes since the router last queried, the cache then 260 sends an End Of Data PDU. 262 If the cache does not have the data needed to update the router, 263 perhaps because its records do not go back to the Serial Number in 264 the Serial Query, then it responds with a Cache Reset PDU 265 (Section 5.8). 267 The Cache Nonce tells the cache what instance the router expects to 268 ensure that the serial numbers are commensurate, i.e. the cache 269 session has not been changed. 271 0 8 16 24 31 272 .-------------------------------------------. 273 | Protocol | PDU | | 274 | Version | Type | Cache Nonce | 275 | 0 | 1 | | 276 +-------------------------------------------+ 277 | | 278 | Length=12 | 279 | | 280 +-------------------------------------------+ 281 | | 282 | Serial Number | 283 | | 284 `-------------------------------------------' 286 5.3. Reset Query 288 Reset Query: The router tells the cache that it wants to receive the 289 total active, current, non-withdrawn, database. The cache responds 290 with a Cache Response PDU (Section 5.4). 292 0 8 16 24 31 293 .-------------------------------------------. 294 | Protocol | PDU | | 295 | Version | Type | reserved = zero | 296 | 0 | 2 | | 297 +-------------------------------------------+ 298 | | 299 | Length=8 | 300 | | 301 `-------------------------------------------' 303 5.4. Cache Response 305 Cache Response: The cache responds with zero or more payload PDUs. 306 When replying to a Serial Query request (Section 5.2), the cache 307 sends the set of all data records it has with serial numbers greater 308 than that sent by the client router. When replying to a Reset Query, 309 the cache sends the set of all data records it has; in this case the 310 withdraw/announce field in the payload PDUs MUST have the value 1 311 (announce). 313 In response to a Reset Query, the new value of Cache Nonce tells the 314 router the instance of the cache session for future confirmation. In 315 response to a Serial Query, the Cache Nonce being the same reassures 316 the router that the serial numbers are commensurate, i.e. the cache 317 session has not changed. 319 0 8 16 24 31 320 .-------------------------------------------. 321 | Protocol | PDU | | 322 | Version | Type | Cache Nonce | 323 | 0 | 3 | | 324 +-------------------------------------------+ 325 | | 326 | Length=8 | 327 | | 328 `-------------------------------------------' 330 5.5. IPv4 Prefix 332 0 8 16 24 31 333 .-------------------------------------------. 334 | Protocol | PDU | | 335 | Version | Type | reserved = zero | 336 | 0 | 4 | | 337 +-------------------------------------------+ 338 | | 339 | Length=20 | 340 | | 341 +-------------------------------------------+ 342 | | Prefix | Max | | 343 | Flags | Length | Length | zero | 344 | | 0..32 | 0..32 | | 345 +-------------------------------------------+ 346 | | 347 | IPv4 Prefix | 348 | | 349 +-------------------------------------------+ 350 | | 351 | Autonomous System Number | 352 | | 353 `-------------------------------------------' 355 The lowest order bit of the Flags field is 1 for an announcement and 356 0 for a withdrawal. 358 In the RPKI, nothing prevents a signing certificate from issuing two 359 identical ROAs, and nothing prohibits the existence of two identical 360 route: or route6: objects in the IRR. In this case there would be no 361 semantic difference between the objects, merely a process redundancy. 363 In the RPKI, there is also an actual need for what might appear to a 364 router as identical IPvX PDUs. This can occur when an upstream 365 certificate is being reissued or there is an address ownership 366 transfer up the validation chain. The ROA would be identical in the 367 router sense, i.e. have the same {prefix, len, max-len, asn}, but a 368 different validation path in the RPKI. This is important to the 369 RPKI, but not to the router. 371 The cache server MUST ensure that it has told the router client to 372 have one and only one IPvX PDU for a unique {prefix, len, max-len, 373 asn} at any one point in time. Should the router client receive an 374 IPvX PDU with a {prefix, len, max-len, asn} identical to one it 375 already has active, it SHOULD raise a Duplicate Announcement Received 376 error. 378 5.6. IPv6 Prefix 380 0 8 16 24 31 381 .-------------------------------------------. 382 | Protocol | PDU | | 383 | Version | Type | reserved = zero | 384 | 0 | 6 | | 385 +-------------------------------------------+ 386 | | 387 | Length=32 | 388 | | 389 +-------------------------------------------+ 390 | | Prefix | Max | | 391 | Flags | Length | Length | zero | 392 | | 0..128 | 0..128 | | 393 +-------------------------------------------+ 394 | | 395 +--- ---+ 396 | | 397 +--- IPv6 Prefix ---+ 398 | | 399 +--- ---+ 400 | | 401 +-------------------------------------------+ 402 | | 403 | Autonomous System Number | 404 | | 405 `-------------------------------------------' 407 5.7. End of Data 409 End of Data: Cache tells router it has no more data for the request. 411 The Cache Nonce MUST be the same as that of the corresponding Cache 412 Response which began the, possibly null, sequence of data PDUs. 414 0 8 16 24 31 415 .-------------------------------------------. 416 | Protocol | PDU | | 417 | Version | Type | Cache Nonce | 418 | 0 | 7 | | 419 +-------------------------------------------+ 420 | | 421 | Length=12 | 422 | | 423 +-------------------------------------------+ 424 | | 425 | Serial Number | 426 | | 427 `-------------------------------------------' 429 5.8. Cache Reset 431 The cache may respond to a Serial Query informing the router that the 432 cache cannot provide an incremental update starting from the serial 433 number specified by the router. The router must decide whether to 434 issue a Reset Query or switch to a different cache. 436 0 8 16 24 31 437 .-------------------------------------------. 438 | Protocol | PDU | | 439 | Version | Type | reserved = zero | 440 | 0 | 8 | | 441 +-------------------------------------------+ 442 | | 443 | Length=8 | 444 | | 445 `-------------------------------------------' 447 5.9. Error Report 449 This PDU is used by either party to report an error to the other. 451 Error reports are only sent as responses to other PDUs. 453 The Error Code is described in Section 10. 455 If the error is not associated with any particular PDU, the Erroneous 456 PDU field MUST be empty and the Length of Encapsulated PDU field MUST 457 be zero. 459 An Error Report PDU MUST NOT be sent for an Error Report PDU. 461 If the error is associated with a PDU of excessive, or possibly 462 corrupt, length, the Erroneous PDU field MAY be truncated. 464 The diagnostic text is optional, if not present the Length of Error 465 Text field SHOULD be zero. If error text is present, it SHOULD be a 466 string in US-ASCII, for maximum portability; if non-US-ASCII 467 characters are absolutely required, the error text MUST use UTF-8 468 encoding. 470 0 8 16 24 31 471 .-------------------------------------------. 472 | Protocol | PDU | | 473 | Version | Type | Error Code | 474 | 0 | 10 | | 475 +-------------------------------------------+ 476 | | 477 | Length | 478 | | 479 +-------------------------------------------+ 480 | | 481 | Length of Encapsulated PDU | 482 | | 483 +-------------------------------------------+ 484 | | 485 ~ Copy of Erroneous PDU ~ 486 | | 487 +-------------------------------------------+ 488 | | 489 | Length of Error Text | 490 | | 491 +-------------------------------------------+ 492 | | 493 | Arbitrary Text | 494 | of | 495 ~ Error Diagnostic Message ~ 496 | | 497 `-------------------------------------------' 499 5.10. Fields of a PDU 501 PDUs contain the following data elements: 503 Protocol Version: An ordinal, currently 0, denoting the version of 504 this protocol. 506 PDU Type: An ordinal, denoting the type of the PDU, e.g. IPv4 507 Prefix, etc. 509 Serial Number: The serial number of the RPKI Cache when this ROA was 510 received from the cache's up-stream cache server or gathered from 511 the global RPKI. A cache increments its serial number when 512 completing an rigorously validated update from a parent cache, for 513 example via rcynic. See [RFC1982] on DNS Serial Number Arithmetic 514 for too much detail on serial number arithmetic. 516 Cache Nonce: When a cache server is started, it generates a nonce to 517 identify the instance of the cache and to bind it to the sequence 518 of Serial Numbers that cache instance will generate. This allows 519 the router to restart a failed session knowing that the Serial 520 Number it is using is commensurate with that of the cache. If, at 521 any time, either the router or the cache finds the value of the 522 nonces they hold disagree, they MUST completely drop the session 523 and the router MUST flush all data learned from that cache. 525 The nonce might be a pseudo-random, a monotonically increasing 526 value if the cache has reliable storage, etc. An implementation 527 which uses a fine granularity of time for the Serial Number might 528 never change the Cache Nonce. 530 Length: A 32 bit ordinal which has as its value the count of the 531 bytes in the entire PDU, including the eight bytes of header which 532 end with the length field. 534 Flags: The lowest order bit of the Flags field is 1 for an 535 announcement and 0 for a withdrawal, whether this PDU announces a 536 new right to announce the prefix or withdraws a previously 537 announced right. A withdraw effectively deletes one previously 538 announced IPvX Prefix PDU with the exact same Prefix, Length, Max- 539 Len, and ASN. 541 Prefix Length: An ordinal denoting the shortest prefix allowed for 542 the prefix. 544 Max Length: An ordinal denoting the longest prefix allowed by the 545 prefix. This MUST NOT be less than the Prefix Length element. 547 Prefix: The IPv4 or IPv6 prefix of the ROA. 549 Autonomous System Number: ASN allowed to announce this prefix, a 32 550 bit ordinal. 552 Zero: Fields shown as zero or reserved MUST be zero. The value of 553 such a field MUST be ignored on receipt. 555 6. Protocol Sequences 557 The sequences of PDU transmissions fall into three conversations as 558 follows: 560 6.1. Start or Restart 562 Cache Router 563 ~ ~ 564 | <----- Reset Query -------- | R requests data (or Serial Query) 565 | | 566 | ----- Cache Response -----> | C confirms request 567 | ------- IPvX Prefix ------> | C sends zero or more 568 | ------- IPvX Prefix ------> | IPv4 and IPv6 Prefix 569 | ------- IPvX Prefix ------> | Payload PDUs 570 | ------ End of Data ------> | C sends End of Data 571 | | and sends new serial 572 ~ ~ 574 When a transport session is first established, the router MAY send a 575 Reset Query and the cache responds with a data sequence of all data 576 it contains. 578 Alternatively, if the router has significant unexpired data from a 579 broken session with the same cache, it MAY start with a Serial Query 580 containing the Cache Nonce from the previous session to ensure the 581 serial numbers are commensurate. 583 This Reset Query sequence is also used when the router receives a 584 Cache Reset, chooses a new cache, or fears that it has otherwise lost 585 its way. 587 To limit the length of time a cache must keep the data necessary to 588 generate incremental updates, a router MUST send either a Serial 589 Query or a Reset Query no less frequently than once an hour. This 590 also acts as a keep alive at the application layer. 592 As the cache MAY not keep updates for little more than one hour, the 593 router MUST have a polling interval of no greater than once an hour. 595 6.2. Typical Exchange 597 Cache Router 598 ~ ~ 599 | -------- Notify ----------> | (optional) 600 | | 601 | <----- Serial Query ------- | R requests data 602 | | 603 | ----- Cache Response -----> | C confirms request 604 | ------- IPvX Prefix ------> | C sends zero or more 605 | ------- IPvX Prefix ------> | IPv4 and IPv6 Prefix 606 | ------- IPvX Prefix ------> | Payload PDUs 607 | ------ End of Data ------> | C sends End of Data 608 | | and sends new serial 609 ~ ~ 611 The cache server SHOULD send a notify PDU with its current serial 612 number when the cache's serial changes, with the expectation that the 613 router MAY then issue a serial query earlier than it otherwise might. 614 This is analogous to DNS NOTIFY in [RFC1996]. The cache MUST rate 615 limit Serial Notifies to no more frequently than one per minute. 617 When the transport layer is up and either a timer has gone off in the 618 router, or the cache has sent a Notify, the router queries for new 619 data by sending a Serial Query, and the cache sends all data newer 620 than the serial in the Serial Query. 622 To limit the length of time a cache must keep old withdraws, a router 623 MUST send either a Serial Query or a Reset Query no less frequently 624 than once an hour. 626 6.3. No Incremental Update Available 628 Cache Router 629 ~ ~ 630 | <----- Serial Query ------ | R requests data 631 | ------- Cache Reset ------> | C cannot supply update 632 | | from specified serial 633 | <------ Reset Query ------- | R requests new data 634 | ----- Cache Response -----> | C confirms request 635 | ------- IPvX Prefix ------> | C sends zero or more 636 | ------- IPvX Prefix ------> | IPv4 and IPv6 Prefix 637 | ------- IPvX Prefix ------> | Payload PDUs 638 | ------ End of Data ------> | C sends End of Data 639 | | and sends new serial 640 ~ ~ 642 The cache may respond to a Serial Query with a Cache Reset, informing 643 the router that the cache cannot supply an incremental update from 644 the serial number specified by the router. This might be because the 645 cache has lost state, or because the router has waited too long 646 between polls and the cache has cleaned up old data that it no longer 647 believes it needs, or because the cache has run out of storage space 648 and had to expire some old data early. Regardless of how this state 649 arose, the cache replies with a Cache Reset to tell the router that 650 it cannot honor the request. When a router receives this, the router 651 SHOULD attempt to connect to any more preferred caches in its cache 652 list. If there are no more preferred caches it MUST issue a Reset 653 Query and get an entire new load from the cache. 655 6.4. Cache has No Data Available 657 Cache Router 658 ~ ~ 659 | <----- Serial Query ------ | R requests data 660 | ---- Error Report PDU ----> | C No Data Available 661 ~ ~ 663 Cache Router 664 ~ ~ 665 | <----- Reset Query ------- | R requests data 666 | ---- Error Report PDU ----> | C No Data Available 667 ~ ~ 669 The cache may respond to either a Serial Query or a Reset Query 670 informing the router that the cache cannot supply any update at all. 671 The most likely cause is that the cache has lost state, perhaps due 672 to a restart, and has not yet recovered. While it is possible that a 673 cache might go into such a state without dropping any of its active 674 sessions, a router is more likely to see this behavior when it 675 initially connects and issues a Reset Query while the cache is still 676 rebuilding its database. 678 When a router receives this kind of error, the router SHOULD attempt 679 to connect to any other caches in its cache list, in preference 680 order. If no other caches are available, the router MUST issue 681 periodic Reset Queries until it gets a new usable load from the 682 cache. 684 7. Transport 686 The transport layer session between a router and a cache carries the 687 binary Protocol Data Units (PDUs) in a persistent session. 689 To prevent cache spoofing and DoS attacks by illegitimate routers, it 690 is highly desirable that the router and the cache are authenticated 691 to each other. Integrity protection for payloads is also desirable 692 to protect against monkey in the middle attacks. Unfortunately, 693 there is no protocol to do so on all currently used platforms. 694 Therefore, as of this document, there is no mandatory to implement 695 transport which provides authentication and integrity protection. 697 To reduce exposure to dropped but non-terminated sessions, both 698 caches and routers SHOULD enable keep alives when available in the 699 chosen transport protocol. 701 It is expected that, when TCP-AO [RFC5925]is available on all 702 platforms deployed by operators, it will become the mandatory to 703 implement transport. 705 Caches and routers MUST implement unprotected transport over TCP 706 using a port, RPKI-Rtr, to be assigned, see Section 12. Operators 707 SHOULD use procedural means, ACLs, ... to reduce the exposure to 708 authentication issues. 710 If available to the operator, caches and routers SHOULD use one of 711 the following more protected protocols. 713 Caches and routers SHOULD use TCP AO transport [RFC5925] over the 714 RPKI-Rtr port. 716 Caches and routers MAY use SSH transport [RFC4252] using using a the 717 normal SSH port. For an example, see Section 7.1. 719 Caches and routers MAY use TCP MD5 transport [RFC5925] using the 720 RPKI-Rtr port. 722 Caches and routers MAY use IPsec transport [RFC4301] using the RPKI- 723 Rtr port. 725 Caches and routers MAY use TLS transport [RFC5246] using using a 726 port, RPKI-Rtr TLS, to be assigned, see Section 12. 728 7.1. SSH Transport 730 To run over SSH, the client router first establishes an SSH transport 731 connection using the SSH transport protocol, and the client and 732 server exchange keys for message integrity and encryption. The 733 client then invokes the "ssh-userauth" service to authenticate the 734 application, as described in the SSH authentication protocol RFC 4252 735 [RFC4252]. Once the application has been successfully authenticated, 736 the client invokes the "ssh-connection" service, also known as the 737 SSH connection protocol. 739 After the ssh-connection service is established, the client opens a 740 channel of type "session", which results in an SSH session. 742 Once the SSH session has been established, the application invokes 743 the application transport as an SSH subsystem called "rpki-rtr". 744 Subsystem support is a feature of SSH version 2 (SSHv2) and is not 745 included in SSHv1. Running this protocol as an SSH subsystem avoids 746 the need for the application to recognize shell prompts or skip over 747 extraneous information, such as a system message that is sent at 748 shell start-up. 750 It is assumed that the router and cache have exchanged keys out of 751 band by some reasonably secured means. 753 Cache servers supporting SSH transport MUST accept RSA and DSA 754 authentication, and SHOULD accept ECDSA authentication. User 755 authentication MUST be supported; host authentication MAY be 756 supported. Implementations MAY support password authentication. 757 Client routers SHOULD verify the public key of the cache, to avoid 758 monkey-in-the-middle attacks. 760 7.2. TLS Transport 762 Client routers using TLS transport MUST use client-side certificates 763 for authentication. While in principle any type of certificate and 764 certificate authority may be used, in general cache operators will 765 generally wish to create their own small-scale CA and issue 766 certificates to each authorized router. This simplifies credential 767 roll-over; any unrevoked, unexpired certificate from the proper CA 768 may be used. If such certificates are used, the CN field [RFC5280] 769 MUST be used to denote the router's identity. 771 Clients SHOULD verify the cache's certificate as well, to avoid 772 monkey-in-the-middle attacks. 774 7.3. TCP MD5 Transport 776 If TCP-MD5 is used, implementations MUST support key lengths of at 777 least 80 printable ASCII bytes, per section 4.5 of [RFC5925]. 778 Implementations MUST also support hexadecimal sequences of at least 779 32 characters, i.e., 128 bits. 781 Key rollover with TCP-MD5 is problematic. Cache servers SHOULD 782 support [RFC4808]. 784 7.4. TCP-AO Transport 786 Implementations MUST support key lengths of at least 80 printable 787 ASCII bytes. Implementations MUST also support hexadecimal sequences 788 of at least 32 characters, i.e., 128 bits. MAC lengths of at least 789 96 bits MUST be supported, per section 5.3 of [RFC5925]. 791 The cryptographic algorithms and associcated parameters described in 792 [RFC5926] MUST be supported. 794 8. Router-Cache Set-Up 796 A cache has the public authentication data for each router it is 797 configured to support. 799 A router may be configured to peer with a selection of caches, and a 800 cache may be configured to support a selection of routers. Each must 801 have the name of, and authentication data for, each peer. In 802 addition, in a router, this list has a non-unique preference value 803 for each server in order of preference. This preference merely 804 denotes proximity, not trust, preferred belief, etc. The client 805 router attempts to establish a session with each potential serving 806 cache in preference order, and then starts to load data from the most 807 preferred cache to which it can connect and authenticate. The 808 router's list of caches has the following elements: 810 Preference: An ordinal denoting the router's preference to connect 811 to that cache, the lower the value the more preferred. 813 Name: The IP Address or fully qualified domain name of the cache. 815 Key: Any needed public key of the cache. 817 MyKey: Any needed private key or certificate of this client. 819 Due to the distributed nature of the RPKI, caches simply can not be 820 rigorously synchronous. A client may hold data from multiple caches, 821 but MUST keep the data marked as to source, as later updates MUST 822 affect the correct data. 824 Just as there may be more than one covering ROA from a single cache, 825 there may be multiple covering ROAs from multiple caches. The 826 results are as described in [I-D.ietf-sidr-pfx-validate]. 828 If data from multiple caches are held, implementations MUST NOT 829 distinguish between data sources when performing validation. 831 When a more preferred cache becomes available, if resources allow, it 832 would be prudent for the client to start fetching from that cache. 834 The client SHOULD attempt to maintain at least one set of data, 835 regardless of whether it has chosen a different cache or established 836 a new connection to the previous cache. 838 A client MAY drop the data from a particular cache when it is fully 839 in synch with one or more other caches. 841 A client SHOULD delete the data from a cache when it has been unable 842 to refresh from that cache for a configurable timer value. The 843 default for that value is twice the polling period for that cache. 845 If a client loses connectivity to a cache it is using, or otherwise 846 decides to switch to a new cache, it SHOULD retain the data from the 847 previous cache until it has a full set of data from one or more other 848 caches. Note that this may already be true at the point of 849 connection loss if the client has connections to more than one cache. 851 9. Deployment Scenarios 853 For illustration, we present three likely deployment scenarios. 855 Small End Site: The small multi-homed end site may wish to outsource 856 the RPKI cache to one or more of their upstream ISPs. They would 857 exchange authentication material with the ISP using some out of 858 band mechanism, and their router(s) would connect to one or more 859 up-streams' caches. The ISPs would likely deploy caches intended 860 for customer use separately from the caches with which their own 861 BGP speakers peer. 863 Large End Site: A larger multi-homed end site might run one or more 864 caches, arranging them in a hierarchy of client caches, each 865 fetching from a serving cache which is closer to the global RPKI. 866 They might configure fall-back peerings to up-stream ISP caches. 868 ISP Backbone: A large ISP would likely have one or more redundant 869 caches in each major PoP, and these caches would fetch from each 870 other in an ISP-dependent topology so as not to place undue load 871 on the global RPKI publication infrastructure. 873 Experience with large DNS cache deployments has shown that complex 874 topologies are ill-advised as it is easy to make errors in the graph, 875 e.g. not maintaining a loop-free condition. 877 Of course, these are illustrations and there are other possible 878 deployment strategies. It is expected that minimizing load on the 879 global RPKI servers will be a major consideration. 881 To keep load on global RPKI services from unnecessary peaks, it is 882 recommended that primary caches which load from the distributed 883 global RPKI not do so all at the same times, e.g. on the hour. 884 Choose a random time, perhaps the ISP's AS number modulo 60 and 885 jitter the inter-fetch timing. 887 10. Error Codes 889 This section contains a preliminary list of error codes. The authors 890 expect additions to this section during development of the initial 891 implementations. Errors which are considered fatal SHOULD cause the 892 session to be dropped. 894 0: Corrupt Data (fatal): The receiver believes the received PDU to 895 be corrupt in a manner not specified by other error codes. 897 1: Internal Error (fatal): The party reporting the error experienced 898 some kind of internal error unrelated to protocol operation (ran 899 out of memory, a coding assertion failed, et cetera). 901 2: No Data Available: The cache believes itself to be in good 902 working order, but is unable to answer either a Serial Query or a 903 Reset Query because it has no useful data available at this time. 904 This is likely to be a temporary error, and most likely indicates 905 that the cache has not yet completed pulling down an initial 906 current data set from the global RPKI system after some kind of 907 event that invalidated whatever data it might have previously held 908 (reboot, network partition, et cetera). 910 3: Invalid Request (fatal): The cache server believes the client's 911 request to be invalid. 913 4: Unsupported Protocol Version (fatal): The Protocol Version is not 914 known by the receiver of the PDU. 916 5: Unsupported PDU Type (fatal): The PDU Type is not known by the 917 receiver of the PDU. 919 6: Withdrawal of Unknown Record (fatal): The received PDU has Flag=0 920 but a record for the Prefix/PrefixLength/MaxLength triple does not 921 exist in the receiver's database. 923 7: Duplicate Announcement Received (fatal): The received PDU has an 924 identical {prefix, len, max-len, asn} tuple as a PDU which is 925 still active in the router. 927 11. Security Considerations 929 As this document describes a security protocol, many aspects of 930 security interest are described in the relevant sections. This 931 section points out issues which may not be obvious in other sections. 933 Cache Validation: In order for a collection of caches as described 934 in Section 9 to guarantee a consistent view, they need to be given 935 consistent trust anchors to use in their internal validation 936 process. Distribution of a consistent trust anchor is assumed to 937 be out of band. 939 Cache Peer Identification: The router initiates a transport session 940 to a cache, which it identifies by either IP address or fully 941 qualified domain name. Be aware that a DNS or address spoofing 942 attack could make the correct cache unreachable. No session would 943 be established, as the authorization keys would not match. 945 Transport Security: The RPKI relies on object, not server or 946 transport, trust. I.e. the IANA root trust anchor is distributed 947 to all caches through some out of band means, and can then be used 948 by each cache to validate certificates and ROAs all the way down 949 the tree. The inter-cache relationships are based on this object 950 security model, hence the inter-cache transport can be lightly 951 protected. 953 But this protocol document assumes that the routers can not do the 954 validation cryptography. Hence the last link, from cache to 955 router, is secured by server authentication and transport level 956 security. This is dangerous, as server authentication and 957 transport have very different threat models than object security. 959 So the strength of the trust relationship and the transport 960 between the router(s) and the cache(s) are critical. You're 961 betting your routing on this. 963 While we can not say the cache must be on the same LAN, if only 964 due to the issue of an enterprise wanting to off-load the cache 965 task to their upstream ISP(s), locality, trust, and control are 966 very critical issues here. The cache(s) really SHOULD be as 967 close, in the sense of controlled and protected (against DDoS, 968 MITM) transport, to the router(s) as possible. It also SHOULD be 969 topologically close so that a minimum of validated routing data 970 are needed to bootstrap a router's access to a cache. 972 The identity of the cache server MUST be verified and 973 authenticated by the router client, and vice versa, before any 974 data are exchanged. 976 12. IANA Considerations 978 This document requests the IANA to assign 'well known' TCP Port 979 Numbers to the RPKI-Router Protocol for the following, see Section 7: 981 RPKI-Rtr 982 RPKI-Rtr TLS 984 This document requests the IANA to create a registry for tuples of 985 Protocol Version / PDU Type, each of which may range from 0 to 255. 986 The name of the registry should be rpki-rtr-pdu. The policy for 987 adding to the registry is RFC Required per [RFC5226], either 988 standards track or experimental. The initial entries should be as 989 follows: 991 Protocol 992 Version PDU Type 993 -------- ------------------- 994 0 0 - Serial Notify 995 0 1 - Serial Query 996 0 2 - Reset Query 997 0 3 - Cache Response 998 0 4 - IPv4 Prefix 999 0 6 - IPv6 Prefix 1000 0 7 - End of Data 1001 0 8 - Cache Reset 1002 0 10 - Error Report 1003 0 255 - Reserved 1005 This document requests the IANA to create a registry for Error Codes 1006 0 to 255. The name of the registry should be rpki-rtr-error. The 1007 policy for adding to the registry is Expert Review per [RFC5226], 1008 where the responsible IESG area director should appoint the Expert 1009 Reviewer. The initial entries should be as follows: 1011 0 - Corrupt Data 1012 1 - Internal Error 1013 2 - No Data Available 1014 3 - Invalid Request 1015 4 - Unsupported Protocol Version 1016 5 - Unsupported PDU Type 1017 6 - Withdrawal of Unknown Record 1018 7 - Duplicate Announcement Received 1019 255 - Reserved 1021 This document requests the IANA to add an SSH Connection Protocol 1022 Subsystem Name, as defined in [RFC4250], of 'rpki-rtr'. 1024 13. Acknowledgments 1026 The authors wish to thank Steve Bellovin, Rex Fernando, Paul Hoffman, 1027 Russ Housley, Pradosh Mohapatra, Keyur Patel, Sandy Murphy, Robert 1028 Raszuk, John Scudder, Ruediger Volk, and David Ward. Particular 1029 thanks go to Hannes Gredler for showing us the dangers of unnecessary 1030 fields. 1032 14. References 1034 14.1. Normative References 1036 [I-D.ietf-sidr-pfx-validate] 1037 Mohapatra, P., Scudder, J., Ward, D., Bush, R., and R. 1038 Austein, "BGP Prefix Origin Validation", 1039 draft-ietf-sidr-pfx-validate-02 (work in progress), 1040 July 2011. 1042 [RFC1982] Elz, R. and R. Bush, "Serial Number Arithmetic", RFC 1982, 1043 August 1996. 1045 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1046 Requirement Levels", BCP 14, RFC 2119, March 1997. 1048 [RFC4250] Lehtinen, S. and C. Lonvick, "The Secure Shell (SSH) 1049 Protocol Assigned Numbers", RFC 4250, January 2006. 1051 [RFC4252] Ylonen, T. and C. Lonvick, "The Secure Shell (SSH) 1052 Authentication Protocol", RFC 4252, January 2006. 1054 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 1055 Internet Protocol", RFC 4301, December 2005. 1057 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 1058 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 1059 May 2008. 1061 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 1062 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 1064 [RFC5280] Cooper, D., Santesson, S., Farrell, S., Boeyen, S., 1065 Housley, R., and W. Polk, "Internet X.509 Public Key 1066 Infrastructure Certificate and Certificate Revocation List 1067 (CRL) Profile", RFC 5280, May 2008. 1069 [RFC5925] Touch, J., Mankin, A., and R. Bonica, "The TCP 1070 Authentication Option", RFC 5925, June 2010. 1072 [RFC5926] Lebovitz, G. and E. Rescorla, "Cryptographic Algorithms 1073 for the TCP Authentication Option (TCP-AO)", RFC 5926, 1074 June 2010. 1076 14.2. Informative References 1078 [I-D.ietf-sidr-arch] 1079 Lepinski, M. and S. Kent, "An Infrastructure to Support 1080 Secure Internet Routing", draft-ietf-sidr-arch-13 (work in 1081 progress), May 2011. 1083 [I-D.ietf-sidr-repos-struct] 1084 Huston, G., Loomans, R., and G. Michaelson, "A Profile for 1085 Resource Certificate Repository Structure", 1086 draft-ietf-sidr-repos-struct-09 (work in progress), 1087 July 2011. 1089 [RFC1996] Vixie, P., "A Mechanism for Prompt Notification of Zone 1090 Changes (DNS NOTIFY)", RFC 1996, August 1996. 1092 [RFC4808] Bellovin, S., "Key Change Strategies for TCP-MD5", 1093 RFC 4808, March 2007. 1095 [RFC5781] Weiler, S., Ward, D., and R. Housley, "The rsync URI 1096 Scheme", RFC 5781, February 2010. 1098 Authors' Addresses 1100 Randy Bush 1101 Internet Initiative Japan 1102 5147 Crystal Springs 1103 Bainbridge Island, Washington 98110 1104 US 1106 Phone: +1 206 780 0431 x1 1107 Email: randy@psg.com 1109 Rob Austein 1110 Dragon Research Labs 1112 Email: sra@hactrn.net