idnits 2.17.1 draft-templin-intarea-seal-38.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == There are 1 instance of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (November 17, 2011) is 4515 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) == Unused Reference: 'RFC3971' is defined on line 1448, but no explicit reference was found in the text == Unused Reference: 'RFC4861' is defined on line 1455, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-aero' is defined on line 1482, but no explicit reference was found in the text == Unused Reference: 'RFC2675' is defined on line 1523, but no explicit reference was found in the text == Unused Reference: 'RFC4191' is defined on line 1545, but no explicit reference was found in the text == Unused Reference: 'RFC4987' is defined on line 1570, but no explicit reference was found in the text == Unused Reference: 'RFC5445' is defined on line 1576, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) == Outdated reference: A later version (-07) exists of draft-ietf-intarea-ipv4-id-update-04 == Outdated reference: A later version (-06) exists of draft-ietf-savi-framework-05 == Outdated reference: A later version (-12) exists of draft-templin-aero-04 == Outdated reference: A later version (-40) exists of draft-templin-intarea-vet-29 == Outdated reference: A later version (-16) exists of draft-templin-ironbis-07 -- Obsolete informational reference (is this intentional?): RFC 1063 (Obsoleted by RFC 1191) -- Obsolete informational reference (is this intentional?): RFC 1981 (Obsoleted by RFC 8201) -- Obsolete informational reference (is this intentional?): RFC 5246 (Obsoleted by RFC 8446) Summary: 1 error (**), 0 flaws (~~), 14 warnings (==), 4 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group F. Templin, Ed. 3 Internet-Draft Boeing Research & Technology 4 Intended status: Standards Track November 17, 2011 5 Expires: May 20, 2012 7 The Subnetwork Encapsulation and Adaptation Layer (SEAL) 8 draft-templin-intarea-seal-38.txt 10 Abstract 12 For the purpose of this document, a subnetwork is defined as a 13 virtual topology configured over a connected IP network routing 14 region and bounded by encapsulating border nodes. These virtual 15 topologies are manifested by tunnels that may span multiple IP and/or 16 sub-IP layer forwarding hops, and can introduce failure modes due to 17 packet duplication and/or links with diverse Maximum Transmission 18 Units (MTUs). This document specifies a Subnetwork Encapsulation and 19 Adaptation Layer (SEAL) that accommodates such virtual topologies 20 over diverse underlying link technologies. 22 Status of this Memo 24 This Internet-Draft is submitted in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF). Note that other groups may also distribute 29 working documents as Internet-Drafts. The list of current Internet- 30 Drafts is at http://datatracker.ietf.org/drafts/current/. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 This Internet-Draft will expire on May 20, 2012. 39 Copyright Notice 41 Copyright (c) 2011 IETF Trust and the persons identified as the 42 document authors. All rights reserved. 44 This document is subject to BCP 78 and the IETF Trust's Legal 45 Provisions Relating to IETF Documents 46 (http://trustee.ietf.org/license-info) in effect on the date of 47 publication of this document. Please review these documents 48 carefully, as they describe your rights and restrictions with respect 49 to this document. Code Components extracted from this document must 50 include Simplified BSD License text as described in Section 4.e of 51 the Trust Legal Provisions and are provided without warranty as 52 described in the Simplified BSD License. 54 Table of Contents 56 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 57 1.1. Motivation . . . . . . . . . . . . . . . . . . . . . . . . 4 58 1.2. Approach . . . . . . . . . . . . . . . . . . . . . . . . . 6 59 2. Terminology and Requirements . . . . . . . . . . . . . . . . . 6 60 3. Applicability Statement . . . . . . . . . . . . . . . . . . . 8 61 4. SEAL Specification . . . . . . . . . . . . . . . . . . . . . . 9 62 4.1. VET Interface Model . . . . . . . . . . . . . . . . . . . 9 63 4.2. SEAL Model of Operation . . . . . . . . . . . . . . . . . 10 64 4.3. SEAL Header and Trailer Format . . . . . . . . . . . . . . 11 65 4.4. ITE Specification . . . . . . . . . . . . . . . . . . . . 14 66 4.4.1. Tunnel Interface MTU . . . . . . . . . . . . . . . . . 14 67 4.4.2. Tunnel Neighbor Soft State . . . . . . . . . . . . . . 15 68 4.4.3. Pre-Encapsulation . . . . . . . . . . . . . . . . . . 16 69 4.4.4. SEAL Encapsulation . . . . . . . . . . . . . . . . . . 17 70 4.4.5. Outer Encapsulation . . . . . . . . . . . . . . . . . 18 71 4.4.6. Path Probing and ETE Reachability Verification . . . . 19 72 4.4.7. Processing ICMP Messages . . . . . . . . . . . . . . . 19 73 4.4.8. IPv4 Middlebox Reassembly Testing . . . . . . . . . . 20 74 4.4.9. Stateful MTU Determination . . . . . . . . . . . . . . 22 75 4.4.10. Detecting Path MTU Changes . . . . . . . . . . . . . . 22 76 4.5. ETE Specification . . . . . . . . . . . . . . . . . . . . 23 77 4.5.1. Tunnel Neighbor Soft State . . . . . . . . . . . . . . 23 78 4.5.2. IP-Layer Reassembly . . . . . . . . . . . . . . . . . 23 79 4.5.3. Decapsulation and Re-Encapsulation . . . . . . . . . . 23 80 4.6. The SEAL Control Message Protocol (SCMP) . . . . . . . . . 25 81 4.6.1. Generating SCMP Error Messages . . . . . . . . . . . . 25 82 4.6.2. Processing SCMP Error Messages . . . . . . . . . . . . 27 83 5. Link Requirements . . . . . . . . . . . . . . . . . . . . . . 29 84 6. End System Requirements . . . . . . . . . . . . . . . . . . . 29 85 7. Router Requirements . . . . . . . . . . . . . . . . . . . . . 29 86 8. Nested Encapsulation Considerations . . . . . . . . . . . . . 30 87 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 30 88 10. Security Considerations . . . . . . . . . . . . . . . . . . . 30 89 11. Related Work . . . . . . . . . . . . . . . . . . . . . . . . . 31 90 12. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 32 91 13. References . . . . . . . . . . . . . . . . . . . . . . . . . . 32 92 13.1. Normative References . . . . . . . . . . . . . . . . . . . 32 93 13.2. Informative References . . . . . . . . . . . . . . . . . . 33 94 Appendix A. Reliability . . . . . . . . . . . . . . . . . . . . . 36 95 Appendix B. Integrity . . . . . . . . . . . . . . . . . . . . . . 36 96 Appendix C. Transport Mode . . . . . . . . . . . . . . . . . . . 37 97 Appendix D. Historic Evolution of PMTUD . . . . . . . . . . . . . 37 98 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 38 100 1. Introduction 102 As Internet technology and communication has grown and matured, many 103 techniques have developed that use virtual topologies (including 104 tunnels of one form or another) over an actual network that supports 105 the Internet Protocol (IP) [RFC0791][RFC2460]. Those virtual 106 topologies have elements that appear as one hop in the virtual 107 topology, but are actually multiple IP or sub-IP layer hops. These 108 multiple hops often have quite diverse properties that are often not 109 even visible to the endpoints of the virtual hop. This introduces 110 failure modes that are not dealt with well in current approaches. 112 The use of IP encapsulation (also known as "tunneling") has long been 113 considered as the means for creating such virtual topologies. 114 However, the insertion of an outer IP header reduces the effective 115 path MTU visible to the inner network layer. When IPv4 is used, this 116 reduced MTU can be accommodated through the use of IPv4 117 fragmentation, but 119 unmitigated in-the-network fragmentation has been found to be harmful 120 through operational experience and studies conducted over the course 121 of many years [FRAG][FOLK][RFC4963]. Additionally, classical path 122 MTU discovery [RFC1191] has known operational issues that are 123 exacerbated by in-the-network tunnels [RFC2923][RFC4459]. The 124 following subsections present further details on the motivation and 125 approach for addressing these issues. 127 1.1. Motivation 129 Before discussing the approach, it is necessary to first understand 130 the problems. In both the Internet and private-use networks today, 131 IPv4 is ubiquitously deployed as the Layer 3 protocol. The primary 132 functions of IPv4 are to provide for routing, addressing, and a 133 fragmentation and reassembly capability used to accommodate links 134 with diverse MTUs. While it is well known that the IPv4 address 135 space is rapidly becoming depleted, there is a lesser-known but 136 growing consensus that other IPv4 protocol limitations have already 137 or may soon become problematic. 139 First, the IPv4 header Identification field is only 16 bits in 140 length, meaning that at most 2^16 unique packets with the same 141 (source, destination, protocol)-tuple may be active in the Internet 142 at a given time [I-D.ietf-intarea-ipv4-id-update]. Due to the 143 escalating deployment of high-speed links, however, this number has 144 become too small by several orders of magnitude for high data rate 145 packet sources such as tunnel endpoints [RFC4963]. Furthermore, 146 there are many well-known limitations pertaining to IPv4 147 fragmentation and reassembly - even to the point that it has been 148 deemed "harmful" in both classic and modern-day studies (see above). 149 In particular, IPv4 fragmentation raises issues ranging from minor 150 annoyances (e.g., in-the-network router fragmentation [RFC1981]) to 151 the potential for major integrity issues (e.g., mis-association of 152 the fragments of multiple IP packets during reassembly [RFC4963]). 154 As a result of these perceived limitations, a fragmentation-avoiding 155 technique for discovering the MTU of the forward path from a source 156 to a destination node was devised through the deliberations of the 157 Path MTU Discovery Working Group (PMTUDWG) during the late 1980's 158 through early 1990's (see Appendix D). In this method, the source 159 node provides explicit instructions to routers in the path to discard 160 the packet and return an ICMP error message if an MTU restriction is 161 encountered. However, this approach has several serious shortcomings 162 that lead to an overall "brittleness" [RFC2923]. 164 In particular, site border routers in the Internet have been known to 165 discard ICMP error messages coming from the outside world. This is 166 due in large part to the fact that malicious spoofing of error 167 messages in the Internet is trivial since there is no way to 168 authenticate the source of the messages [RFC5927]. Furthermore, when 169 a source node that requires ICMP error message feedback when a packet 170 is dropped due to an MTU restriction does not receive the messages, a 171 path MTU-related black hole occurs. This means that the source will 172 continue to send packets that are too large and never receive an 173 indication from the network that they are being discarded. This 174 behavior has been confirmed through documented studies showing clear 175 evidence of path MTU discovery failures in the Internet today 176 [TBIT][WAND][SIGCOMM]. 178 The issues with both IPv4 fragmentation and this "classical" method 179 of path MTU discovery are exacerbated further when IP tunneling is 180 used [RFC4459]. For example, an ingress tunnel endpoint (ITE) may be 181 required to forward encapsulated packets into the subnetwork on 182 behalf of hundreds, thousands, or even more original sources within 183 the end site that it serves. If the ITE allows IPv4 fragmentation on 184 the encapsulated packets, persistent fragmentation could lead to 185 undetected data corruption due to Identification field wrapping. If 186 the ITE instead uses classical IPv4 path MTU discovery, it must rely 187 on ICMP error messages coming from the subnetwork that may be 188 suspect, subject to loss due to filtering middleboxes, or 189 insufficiently provisioned for translation into error messages to be 190 returned to the original sources. 192 Although recent works have led to the development of a robust end-to- 193 end MTU determination scheme [RFC4821], they do not excuse tunnels 194 from delivering path MTU discovery feedback when packets are lost due 195 to size restrictions. Moreover, in current practice existing 196 tunneling protocols mask the MTU issues by selecting a "lowest common 197 denominator" MTU that may be much smaller than necessary for most 198 paths and difficult to change at a later date. Therefore, a new 199 approach to accommodate tunnels over links with diverse MTUs is 200 necessary. 202 1.2. Approach 204 For the purpose of this document, a subnetwork is defined as a 205 virtual topology configured over a connected network routing region 206 and bounded by encapsulating border nodes. Example connected network 207 routing regions include Mobile Ad hoc Networks (MANETs), enterprise 208 networks and the global public Internet itself. Subnetwork border 209 nodes forward unicast and multicast packets over the virtual topology 210 across multiple IP and/or sub-IP layer forwarding hops that may 211 introduce packet duplication and/or traverse links with diverse 212 Maximum Transmission Units (MTUs). 214 This document introduces a Subnetwork Encapsulation and Adaptation 215 Layer (SEAL) for tunneling network layer protocols (e.g., IPv4, IPv6, 216 OSI, etc.) over IP subnetworks that connect Ingress and Egress Tunnel 217 Endpoints (ITEs/ETEs) of border nodes. It provides a modular 218 specification designed to be tailored to specific associated 219 tunneling protocols. A transport-mode of operation is also possible, 220 and described in Appendix C. 222 SEAL provides a mid-layer encapsulation that accommodates links with 223 diverse MTUs and allows routers in the subnetwork to perform 224 efficient duplicate packet detection. The encapsulation further 225 ensures data origin authentication, packet header integrity and anti- 226 replay. 228 SEAL treats tunnels that traverse the subnetwork as ordinary links 229 that must support network layer services. Moreover, SEAL provides 230 dynamic mechanisms to ensure a maximal path MTU over the tunnel. 231 This is in contrast to static approaches which avoid MTU issues by 232 selecting a lowest common denominator MTU value that may be overly 233 conservative for the vast majority of tunnel paths and difficult to 234 change even when larger MTUs become available. 236 The following sections provide the SEAL normative specifications, 237 while the appendices present non-normative additional considerations. 239 2. Terminology and Requirements 241 The following terms are defined within the scope of this document: 243 subnetwork 244 a virtual topology configured over a connected network routing 245 region and bounded by encapsulating border nodes. 247 Ingress Tunnel Endpoint (ITE) 248 a virtual interface over which an encapsulating border node (host 249 or router) sends encapsulated packets into the subnetwork. 251 Egress Tunnel Endpoint (ETE) 252 a virtual interface over which an encapsulating border node (host 253 or router) receives encapsulated packets from the subnetwork. 255 ETE Link Path 256 a subnetwork path from an ITE to an ETE beginning with an 257 underlying link of the ITE as the first hop. 259 inner packet 260 an unencapsulated network layer protocol packet (e.g., IPv6 261 [RFC2460], IPv4 [RFC0791], OSI/CLNP [RFC1070], etc.) before any 262 outer encapsulations are added. Internet protocol numbers that 263 identify inner packets are found in the IANA Internet Protocol 264 registry [RFC3232]. SEAL protocol packets that incur an 265 additional layer of SEAL encapsulation are also considered inner 266 packets. 268 outer IP packet 269 a packet resulting from adding an outer IP header (and possibly 270 other outer headers) to a SEAL-encapsulated inner packet. 272 packet-in-error 273 the leading portion of an invoking data packet encapsulated in the 274 body of an error control message (e.g., an ICMPv4 [RFC0792] error 275 message, an ICMPv6 [RFC4443] error message, etc.). 277 Packet Too Big (PTB) 278 a control plane message indicating an MTU restriction (e.g., an 279 ICMPv6 "Packet Too Big" message [RFC4443], an ICMPv4 280 "Fragmentation Needed" message [RFC0792], etc.). 282 IP 283 used to generically refer to either Internet Protocol (IP) 284 version, i.e., IPv4 or IPv6. 286 The following abbreviations correspond to terms used within this 287 document and/or elsewhere in common Internetworking nomenclature: 289 DF - the IPv4 header "Don't Fragment" flag [RFC0791] 291 ETE - Egress Tunnel Endpoint 293 HLEN - the length of the SEAL header plus outer headers 295 ICV - Integrity Check Vector 297 ITE - Ingress Tunnel Endpoint 299 MTU - Maximum Transmission Unit 301 SCMP - the SEAL Control Message Protocol 303 SDU - SCMP Destination Unreachable message 305 SNA - SCMP Neighbor Advertisement message 307 SNS - SCMP Neighbor Solicitation message 309 SPP - SCMP Parameter Problem message 311 SPTB - SCMP Packet Too Big message 313 SEAL - Subnetwork Encapsulation and Adaptation Layer 315 SEAL_PORT - a transport-layer service port number used for SEAL 317 SEAL_PROTO - an IP protocol number used for SEAL 319 TE - Tunnel Endpoint (i.e., either ingress or egress) 321 VET - Virtual Enterprise Traversal 323 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 324 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 325 document are to be interpreted as described in [RFC2119]. When used 326 in lower case (e.g., must, must not, etc.), these words MUST NOT be 327 interpreted as described in [RFC2119], but are rather interpreted as 328 they would be in common English. 330 3. Applicability Statement 332 SEAL was originally motivated by the specific case of subnetwork 333 abstraction for Mobile Ad hoc Networks (MANETs), however the domain 334 of applicability also extends to subnetwork abstractions over 335 enterprise networks, ISP networks, SOHO networks, the global public 336 Internet itself, and any other connected network routing region. 337 SEAL, along with the Virtual Enterprise Traversal (VET) 338 [I-D.templin-intarea-vet] tunnel virtual interface abstraction, are 339 the functional building blocks for the Internet Routing Overlay 340 Network (IRON) [I-D.templin-ironbis] and Routing and Addressing in 341 Networks with Global Enterprise Recursion (RANGER) [RFC5720][RFC6139] 342 architectures. 344 SEAL provides a network sublayer for encapsulation of an inner 345 network layer packet within outer encapsulating headers. SEAL can 346 also be used as a sublayer within a transport layer protocol data 347 payload, where transport layer encapsulation is typically used for 348 Network Address Translator (NAT) traversal as well as operation over 349 subnetworks that give preferential treatment to certain "core" 350 Internet protocols (e.g., TCP, UDP, etc.). The SEAL header is 351 processed the same as for IPv6 extension headers, i.e., it is not 352 part of the outer IP header but rather allows for the creation of an 353 arbitrarily extensible chain of headers in the same way that IPv6 354 does. 356 To accommodate MTU diversity, the Egress Tunnel Endpoint (ETE) acts 357 as a passive observer that simply informs the Ingress Tunnel Endpoint 358 (ITE) of any packet size limitations. This allows the ITE to return 359 appropriate path MTU discovery feedback even if the network path 360 between the ITE and ETE filters ICMP messages. 362 SEAL further ensures data origin authentication, packet header 363 integrity, and anti-replay. The SEAL framework is therefore similar 364 to the IP Security (IPsec) Authentication Header (AH) 365 [RFC4301][RFC4302], however it provides only minimal hop-by-hop 366 authenticating services along a path while leaving full data 367 integrity, authentication and confidentiality services as an end-to- 368 end consideration. While SEAL performs data origin authentication, 369 the origin site must also perform the necessary ingress filtering in 370 order to provide full source address verification 371 [I-D.ietf-savi-framework]. 373 4. SEAL Specification 375 The following sections specify the operation of SEAL: 377 4.1. VET Interface Model 379 SEAL is an encapsulation sublayer used within VET non-broadcast, 380 multiple access (NBMA) tunnel virtual interfaces. Each VET interface 381 is configured over one or more underlying interfaces attached to 382 subnetwork links. The VET interface connects an ITE to one or more 383 ETE "neighbors" via tunneling across an underlying subnetwork, where 384 tunnel neighbor relationship may be either unidirectional or 385 bidirectional. 387 A unidirectional tunnel neighbor relationship allows the near end ITE 388 to send data packets forward to the far end ETE, while the ETE only 389 returns control messages when necessary. A bidirectional tunnel 390 neighbor relationship is one over which both TEs can exchange both 391 data and control messages. 393 Implications of the VET unidirectional and bidirectional models are 394 discussed in [I-D.templin-intarea-vet]. 396 4.2. SEAL Model of Operation 398 SEAL-enabled ITEs encapsulate each inner packet in a SEAL header, any 399 outer header encapsulations, and in some instances a SEAL trailer as 400 shown in Figure 1: 402 +--------------------+ 403 ~ outer IP header ~ 404 +--------------------+ 405 ~ other outer hdrs ~ 406 +--------------------+ 407 ~ SEAL Header ~ 408 +--------------------+ +--------------------+ 409 | | --> | | 410 ~ Inner ~ --> ~ Inner ~ 411 ~ Packet ~ --> ~ Packet ~ 412 | | --> | | 413 +--------------------+ +--------------------+ 414 | SEAL Trailer | 415 +--------------------+ 417 Figure 1: SEAL Encapsulation 419 The ITE inserts the SEAL header according to the specific tunneling 420 protocol. For simple encapsulation of an inner network layer packet 421 within an outer IP header (e.g., 422 [RFC1070][RFC2003][RFC2473][RFC4213], etc.), the ITE inserts the SEAL 423 header between the inner packet and outer IP headers as: IP/SEAL/ 424 {inner packet}. 426 For encapsulations over transports such as UDP, the ITE inserts the 427 SEAL header between the outer transport layer header and the inner 428 packet, e.g., as IP/UDP/SEAL/{inner packet} (similar to [RFC4380]). 429 In that case, the UDP header is seen as an "other outer header" as 430 depicted in Figure 1. 432 When necessary, the ITE also appends a SEAL trailer at the end of the 433 SEAL packet. In that case, the trailer is added after the final byte 434 of the encapsulated packet. 436 SEAL supports both "nested" tunneling and "re-encapsulating" 437 tunneling. Nested tunneling occurs when a first tunnel is 438 encapsulated within a second tunnel, which may then further be 439 encapsulated within additional tunnels. Nested tunneling can be 440 useful, and stands in contrast to "recursive" tunneling which is an 441 anomalous condition incurred due to misconfiguration or a routing 442 loop. Considerations for nested tunneling are discussed in Section 4 443 of [RFC2473]. 445 Re-encapsulating tunneling occurs when a packet arrives at a first 446 ETE, which then acts as an ITE to re-encapsulate and forward the 447 packet to a second ETE connected to the same subnetwork. In that 448 case each ITE/ETE transition represents a segment of a bridged path 449 between the ITE nearest the source and the ETE nearest the 450 destination. Combinations of nested and re-encapsulating tunneling 451 are also naturally supported by SEAL. 453 The SEAL ITE considers each {underlying interface, IP address} pair 454 as the ingress attachment point to a subnetwork link path to the ETE. 455 The ITE therefore maintains path MTU state on a per ETE link path 456 basis, although it may instead maintain only the lowest-common- 457 denominator values for all of the ETE's link paths in order to reduce 458 state. 460 Finally, the SEAL ITE ensures that the inner network layer protocol 461 will see a minimum MTU of 1280 bytes over each ETE link path 462 regardless of the outer network layer protocol version, i.e., even if 463 a small amount of fragmentation and reassembly are necessary. 465 4.3. SEAL Header and Trailer Format 467 The SEAL header is formatted as follows: 469 0 1 2 3 470 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 471 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 472 |VER|C|P|R|T|U|Z| NEXTHDR | PREFLEN | LINK_ID |LEVEL| 473 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 474 | PKT_ID | 475 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 476 | ICV1 | 477 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 478 ~ PREFIX (when present) ~ 479 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 481 Figure 2: SEAL Header Format 483 VER (2) 484 a 2-bit version field. This document specifies Version 0 of the 485 SEAL protocol, i.e., the VER field encodes the value 0. 487 C (1) 488 the "Control/Data" bit. Set to 1 by the ITE in SEAL Control 489 Message Protocol (SCMP) control messages, and set to 0 in ordinary 490 data packets. 492 P (1) 493 the "Prefix Included" bit. Set to 1 if the header includes a 494 Prefix Field. Used for SCMP messages that do not include a 495 packet-in-error (see: [I-D.templin-intarea-vet]), and for NULL 496 SEAL data packets used as probes (see: Section 4.4.6). 498 R (1) 499 the "Redirects Permitted" bit. For data packets, set to 1 by the 500 ITE to inform the ETE that the source is accepting Redirects (see: 501 [I-D.templin-intarea-vet]). 503 T (1) 504 the "Trailer Included" bit. Set to 1 if the ITE was obliged to 505 include a trailer. 507 U (1) 508 the "Unfragmented Packet" bit. Set to 1 by the ITE in SEAL data 509 packets for which it wishes to receive an explicit acknowledgement 510 from the ETE if the packet arrives unfragmented. 512 Z (1) 513 the "Reserved" bit. Must be set to 0 for this version of the SEAL 514 specification. 516 NEXTHDR (8) an 8-bit field that encodes the next header Internet 517 Protocol number the same as for the IPv4 protocol and IPv6 next 518 header fields. 520 PREFLEN (8) an 8-bit field that encodes the length of the prefix to 521 be applied to the source address of the inner packets (when P==0) 522 or the prefix included in the PREFIX field (when P==1). 524 LINK_ID (5) 525 a 5-bit link identification value, set to a unique value by the 526 ITE for each underlying link as the first hop of a path over which 527 it will send encapsulated packets to ETEs. Up to 32 ETE link 528 paths are therefore supported for each ETE. 530 LEVEL (3) 531 a 3-bit nesting level; use to limit the number of tunnel nesting 532 levels. Set to an integer value up to 7 in the innermost SEAL 533 encapsulation, and decremented by 1 for each successive additional 534 SEAL encapsulation nesting level. Up to 8 levels of nesting are 535 therefore supported. 537 PKT_ID (32) 538 a 32-bit per-packet identification field. Set to a monotonically- 539 incrementing 32-bit value for each SEAL packet transmitted to this 540 ETE, beginning with 0. 542 ICV1 (32) 543 a 32-bit header integrity check value that covers the leading 128 544 bytes of the packet beginning with the SEAL header. The value 128 545 is chosen so that at least the SEAL header as well as the inner 546 packet network and transport layer headers are covered by the 547 integrity check. 549 PREFIX (variable) 550 a variable-length string of bytes; present only when P==1. The 551 field length is determined by calculating Len=(Ceiling(PREFLEN / 552 32) * 4). For example, if PREFLEN==63, the field is 8 bytes in 553 length and encodes the leading 63 bits of the inner network layer 554 prefix beginning with the most significant bit. 556 When T==1, SEAL encapsulation also includes a trailer formatted as 557 follows: 559 0 1 2 3 560 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 561 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 562 | ICV2 | 563 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 565 Figure 3: SEAL Trailer Format 567 ICV2 (32) 568 a 32-bit packet integrity check value. Present only when T==1, 569 and covers the remaining length of the encapsulated packet beyond 570 the leading 128 bytes (i.e., the remaining portion that was not 571 covered by ICV1). Added as a trailing 32 bit field following the 572 final byte of the encapsulated SEAL packet and used to detect 573 reassembly misassociations. Need not be aligned on an even byte 574 boundary. 576 4.4. ITE Specification 578 4.4.1. Tunnel Interface MTU 580 The tunnel interface must present a constant MTU value to the inner 581 network layer as the size for admission of inner packets into the 582 interface. Since VET NBMA tunnel virtual interfaces may support a 583 large set of ETE link paths that accept widely varying maximum packet 584 sizes, however, a number of factors should be taken into 585 consideration when selecting a tunnel interface MTU. 587 Due to the ubiquitous deployment of standard Ethernet and similar 588 networking gear, the nominal Internet cell size has become 1500 589 bytes; this is the de facto size that end systems have come to expect 590 will either be delivered by the network without loss due to an MTU 591 restriction on the path or a suitable ICMP Packet Too Big (PTB) 592 message returned. When large packets sent by end systems incur 593 additional encapsulation at an ITE, however, they may be dropped 594 silently within the tunnel since the network may not always deliver 595 the necessary PTBs [RFC2923]. 597 The ITE should therefore set a tunnel interface MTU of at least 1500 598 bytes plus extra room to accommodate any additional encapsulations 599 that may occur on the path from the original source. The ITE can 600 also set smaller MTU values; however, care must be taken not to set 601 so small a value that original sources would experience an MTU 602 underflow. In particular, IPv6 sources must see a minimum path MTU 603 of 1280 bytes, and IPv4 sources should see a minimum path MTU of 576 604 bytes. 606 The inner network layer protocol consults the tunnel interface MTU 607 when admitting a packet into the interface. For non-SEAL inner IPv4 608 packets with the IPv4 Don't Fragment (DF) bit set to 0, if the packet 609 is larger than the tunnel interface MTU the inner IPv4 layer uses 610 IPv4 fragmentation to break the packet into fragments no larger than 611 the tunnel interface MTU. The ITE then admits each fragment into the 612 interface as an independent packet. 614 For all other inner packets, the inner network layer admits the 615 packet if it is no larger than the tunnel interface MTU; otherwise, 616 it drops the packet and sends a PTB error message to the source with 617 the MTU value set to the tunnel interface MTU. The message contains 618 as much of the invoking packet as possible without the entire message 619 exceeding the network layer minimum MTU (e.g., 1280 bytes for IPv6, 620 576 bytes for IPv4, etc.). 622 The ITE can alternatively set an indefinite MTU on the tunnel 623 interface such that all inner packets are admitted into the interface 624 regardless of their size. For ITEs that host applications that use 625 the tunnel interface directly, this option must be carefully 626 coordinated with protocol stack upper layers since some upper layer 627 protocols (e.g., TCP) derive their packet sizing parameters from the 628 MTU of the outgoing interface and as such may select too large an 629 initial size. This is not a problem for upper layers that use 630 conservative initial maximum segment size estimates and/or when the 631 tunnel interface can reduce the upper layer's maximum segment size, 632 e.g., by reducing the size advertised in the MSS option of outgoing 633 TCP messages (sometimes known as "MSS clamping"). 635 In light of the above considerations, the ITE SHOULD configure an 636 indefinite MTU on tunnel *router* interfaces so that subnetwork 637 adaptation is handled from within the interface. The ITE MAY instead 638 set a finite MTU on tunnel *host* interfaces. 640 4.4.2. Tunnel Neighbor Soft State 642 Within the tunnel virtual interface, the ITE maintains a per tunnel 643 neighbor (i.e., a per-ETE) integrity check vector (ICV) calculation 644 algorithm and (when data origin authentication is required) a 645 symmetric secret key to calculate the ICV(s) in packets it will send 646 to this ETE. The ITE also maintains a window of PKT_ID values for 647 the packets it has recently sent to this ETE. 649 For each ETE link path, the ITE must account for the lengths of the 650 headers to be used for encapsulation. The ITE therefore maintains 651 the per ETE link path constant values "SHLEN" set to length of the 652 SEAL header, "UHLEN" set to the length of the UDP encapsulating 653 header (or 0 if UDP encapsulation is not used), "IHLEN" set to the 654 length of the outer IP layer header, and "HLEN" set to (SHLEN+UHLEN+ 655 IHLEN). (The ITE must include the length of the uncompressed headers 656 even if header compression is enabled when calculating these 657 lengths.) In addition, the ETE maintains a constant value "MIN_MTU" 658 set to 1280+HLEN as well as a variable "PATH_MTU" initialized to the 659 MTU of the underlying link. 661 For IPv4, the ITE also maintains the per ETE link path boolean 662 variables "USE_DF" (initialized to "FALSE") and "USE_TRAILER" 663 (initialized to "TRUE" if PATH_MTU is less than MIN_MTU; otherwise 664 initialized to "FALSE") . 666 The ITE may instead maintain *HLEN, MIN_MTU, PATH_MTU, USE_DF, and 667 USE_TRAILER as per ETE (rather than per ETE link path) values. In 668 that case, the values reflect the lowest-common-denominator MTU 669 across all of the ETE's link paths. 671 4.4.3. Pre-Encapsulation 673 For each inner packet admitted into the tunnel interface, if the 674 packet is itself a SEAL packet (i.e., one with either SEAL_PROTO in 675 the IP protocol/next-header field, or with SEAL_PORT in the transport 676 layer destination port field) and the LEVEL field of the SEAL header 677 contains the value 0, the ITE silently discards the packet. 679 Otherwise, for IPv4 inner packets with DF==0 in the IPv4 header, if 680 the packet is larger than 512 bytes and is not the first fragment of 681 a SEAL packet (i.e., not a packet that includes a SEAL header) the 682 ITE fragments the packet into inner fragments no larger than 512 683 bytes. The ITE then submits each inner fragment for SEAL 684 encapsulation as specified in Section 4.4.4. 686 For all other packets, if the packet is no larger than (MAX(PATH_MTU, 687 MIN_MTU)-HLEN) for the corresponding ETE link path, the ITE submits 688 it for SEAL encapsulation as specified in Section 4.4.4. Otherwise, 689 the ITE sends a PTB error message toward the source address of the 690 inner packet. 692 To send the PTB message, the ITE first checks its forwarding tables 693 to discover the previous hop toward the source address of the inner 694 packet. If the previous hop is reached via the same tunnel 695 interface, the ITE sends an SCMP PTB (SPTB) message to the previous 696 hop (see: Section 4.6.1.1) with the MTU field set to (MAX(PATH_MTU, 697 MIN_MTU)-HLEN). Otherwise, the ITE sends an ordinary PTB message 698 appropriate to the inner protocol version with the MTU field set to 699 (MAX(PATH_MTU, MIN_MTU)-HLEN). 701 After sending the (S)PTB message, the ITE discards the inner packet. 703 4.4.4. SEAL Encapsulation 705 The ITE next encapsulates the inner packet in a SEAL header formatted 706 as specified in Section 4.3. The ITE sets NEXTHDR to the protocol 707 number corresponding to the address family of the encapsulated inner 708 packet. For example, the ITE sets NEXTHDR to the value '4' for 709 encapsulated IPv4 packets [RFC2003], '41' for encapsulated IPv6 710 packets [RFC2473][RFC4213], '80' for encapsulated OSI/CLNP packets 711 [RFC1070], etc. 713 The ITE then sets R=1 if redirects are permitted (see: 714 [I-D.templin-intarea-vet]) and sets PREFLEN to the length of the 715 prefix to be applied to the inner source address. The ITE's claimed 716 PREFLEN is subject to verification by the ETE; hence, the ITE MUST 717 set PREFLEN to the exact prefix length that it is authorized to use. 718 (Note that if this process is entered via re-encapsulation (see: 719 Section 4.5.4), PREFLEN and R are instead copied from the SEAL header 720 of the re-encapsulated packet. This implies that the PREFLEN and R 721 values are propagated across a re-encapsulating chain of ITE/ETEs 722 that must all be authorized to represent the prefix.) 724 Next, the ITE sets (C=0; P=0; Z=0), then sets LINK_ID to the value 725 assigned to the underlying ETE link path and sets PKT_ID to a 726 monotonically-increasing integer value for this ETE, beginning with 0 727 in the first packet transmitted. The ITE also sets U=1 if it needs 728 to determine whether the ETE will receive the packet without 729 fragmentation, e.g., for ETE reachability determination (see: Section 730 4.4.6), to test whether a middlebox on the path is reassembling 731 fragmented packets before they arrive at the ETE (see: Section 732 4.4.8), for stateful MTU determination (see Section 4.4.9), etc. 733 Otherwise, the ITE sets U=0. 735 Next, if the inner packet is not itself a SEAL packet the ITE sets 736 LEVEL to an integer value between 0 and 7 as a specification of the 737 number of additional layers of nested SEAL encapsulations permitted. 738 If the inner packet is a SEAL packet that is undergoing nested 739 encapsulation, the ITE instead sets LEVEL to the value that appears 740 in the inner packet's SEAL header minus 1. If the inner packet is 741 undergoing SEAL re-encapsulation, the ITE instead copies the LEVEL 742 value from the SEAL header of the packet to be re-encapsulated. 744 Next, if this is an IPv4 ETE link path with USE_TRAILER==TRUE, and 745 the inner packet is larger than (128-SHLEN-UHLEN) bytes but no larger 746 than 1280 bytes, the ITE sets T=1. Otherwise, the ITE sets T=0. The 747 ITE then adds the outer encapsulating headers, calculates the ICV(s) 748 and performs any necessary outer fragmentation as specified in 749 Section 4.4.5. 751 4.4.5. Outer Encapsulation 753 Following SEAL encapsulation, the ITE next encapsulates the packet in 754 the requisite outer headers according to the specific encapsulation 755 format (e.g., [RFC1070], [RFC2003], [RFC2473], [RFC4213], etc.), 756 except that it writes 'SEAL_PROTO' in the protocol field of the outer 757 IP header (when simple IP encapsulation is used) or writes 758 'SEAL_PORT' in the outer destination transport service port field 759 (e.g., when IP/UDP encapsulation is used). 761 When UDP encapsulation is used, the ITE sets the UDP header fields as 762 specified in Section 5.5.4 of [I-D.templin-intarea-vet] (where the 763 UDP header length field includes the length of the SEAL trailer, if 764 present). The ITE then performs outer IP header encapsulation as 765 specified in Section 5.5.5 of [I-D.templin-intarea-vet]. If this 766 process is entered via re-encapsulation (see: Section 4.5.4), the ITE 767 instead follows the outer IP/UDP re-encapsulation procedures 768 specified in Section 5.5.6 of [I-D.templin-intarea-vet]. 770 When IPv4 is used as the outer encapsulation layer, if USE_DF==FALSE 771 the ITE sets DF=0 in the IPv4 header to allow the packet to be 772 fragmented within the subnetwork if it encounters a restricting link. 773 Otherwise, the ITE sets DF=1. 775 When IPv6 is used as the outer encapsulation layer, the "DF" flag is 776 absent but implicitly set to 1. The packet therefore will not be 777 fragmented within the subnetwork, since IPv6 deprecates in-the- 778 network fragmentation. 780 The ITE next sets ICV1=0 in the SEAL header and calculates the packet 781 ICVs. The ICVs are calculated using an algorithm agreed on by the 782 ITE and ETE. When data origin authentication is required, the 783 algorithm uses a symmetric secret key so that the ETE can verify that 784 the ICVs were generated by the ITE. 786 The ITE first calculates the ICV over the leading 128 bytes of the 787 packet (or up to the end of the packet if there are fewer than 128 788 bytes) beginning with the UDP header (if present) then places result 789 in the ICV1 field in the header. If T==1, the ITE next calculates 790 the ICV over the remainder of the packet and places the result in the 791 ICV2 field in the SEAL trailer. The ITE then submits the packet for 792 outer encapsulation. 794 Next, the ITE uses IP fragmentation if necessary to fragment the 795 encapsulated packet into outer IP fragments that are no larger than 796 PATH_MTU. By virtue of the pre-encapsulation packet size 797 calculations specified in Section 4.4.3, fragmentation will therefore 798 only occur for outer packets that are larger than PATH_MTU but no 799 larger than MIN_MTU. (Note that, for IPv6, fragmentation must be 800 performed by the ITE itself, while for IPv4 the fragmentation could 801 instead be performed by a router in the ETE link path.) 803 The ITE then sends each outer packet/fragment via the underlying link 804 corresponding to LINK_ID. 806 4.4.6. Path Probing and ETE Reachability Verification 808 All SEAL data packets sent by the ITE are considered implicit probes. 809 SEAL data packets will elicit an SCMP message from the ETE if it 810 needs to acknowledge a probe and/or report an error condition. SEAL 811 data packets may also be dropped by either the ETE or a router on the 812 path, which will return an ICMP message. 814 The ITE can also send an SCMP Router/Neighbor Solicitation message to 815 elicit an SCMP Router/Neighbor Advertisement response (see: 816 [I-D.templin-intarea-vet]) as verification that the ETE is still 817 reachable via a specific link path. 819 The ITE processes ICMP messages as specified in Section 4.4.7. 821 The ITE processes SCMP messages as specified in Section 4.6.2. 823 4.4.7. Processing ICMP Messages 825 When the ITE sends SEAL packets, it may receive ICMP error 826 messages[RFC0792][RFC4443] from another ITE on the path to the ETE 827 (i.e., in case of nested encapsulations) or from an ordinary router 828 within the subnetwork. Each ICMP message includes an outer IP 829 header, followed by an ICMP header, followed by a portion of the SEAL 830 data packet that generated the error (also known as the "packet-in- 831 error") beginning with the outer IP header. 833 The ITE should process ICMPv4 Protocol Unreachable messages and 834 ICMPv6 Parameter Problem messages with Code "Unrecognized Next Header 835 type encountered" as a hint that the ETE does not implement the SEAL 836 protocol. The ITE can also process other ICMP messages that do not 837 include sufficient information in the packet-in-error as a hint that 838 the ETE link path may be failing. Specific actions that the ITE may 839 take in these cases are out of scope. 841 For other ICMP messages, the should use any outer header information 842 available as a first-pass authentication filter (e.g., to determine 843 if the source of the message is within the same administrative domain 844 as the ITE) and discards the message if first pass filtering fails. 846 Next, the ITE examines the packet-in-error beginning with the SEAL 847 header. If the value in the PKT_ID field is not within the window of 848 packets the ITE has recently sent to this ETE, or if the value in the 849 SEAL header ICV1 field is incorrect, the ITE discards the message. 851 Next, if the received ICMP message is a PTB the ITE sets the 852 temporary variable "PMTU" for this ETE link path to the MTU value in 853 the PTB message. If PMTU==0, the ITE consults a plateau table (e.g., 854 as described in [RFC1191]) to determine PMTU based on the length 855 field in the outer IP header of the packet-in-error. (For example, 856 if the ITE receives a PTB message with MTU==0 and length 1500, it can 857 set PMTU=1450. If the ITE subsequently receives a PTB message with 858 MTU==0 and length 1450, it can set PMTU=1400, etc.) If the ITE is 859 performing stateful MTU determination for this ETE link path (see 860 Section 4.4.9), the ITE next sets PATH_MTU=PMTU. If PMTU is less 861 than MIN_MTU, the ITE sets PATH_MTU=PMTU (and for IPv4 also sets 862 USE_TRAILER=TRUE), then discards the message. 864 If the ICMP message was not discarded, the ITE then transcribes it 865 into a message to return to the previous hop. If the previous hop 866 toward the inner source address within the packet-in-error is reached 867 via the same tunnel interface the SEAL data packet was sent on, the 868 ITE transcribes the ICMP message into an SCMP message. Otherwise, 869 the ITE transcribes the ICMP message into a message appropriate for 870 the inner protocol version. 872 To transcribe the message, the ITE extracts the inner packet from 873 within the ICMP message packet-in-error field and uses it to generate 874 a new message corresponding to the type of the received ICMP message. 875 For SCMP messages, the ITE generates the message the same as 876 described for ETE generation of SCMP messages in Section 4.6.1. For 877 (S)PTB messages, the ITE writes (PMTU-HLEN) in the MTU field. 879 The ITE finally forwards the transcribed message to the previous hop 880 toward the inner source address. 882 4.4.8. IPv4 Middlebox Reassembly Testing 884 For IPv4, the ITE can perform a qualification exchange over an ETE 885 link path to ensure that the subnetwork correctly delivers fragments 886 to the ETE. This procedure can be used, e.g., to determine whether 887 there are middleboxes on the path that violate the [RFC1812], Section 888 5.2.6 requirement that: "A router MUST NOT reassemble any datagram 889 before forwarding it". 891 When possible, the ITE should use knowledge of its topological 892 arrangement as an aid in determining when middlebox reassembly 893 testing is necessary. For example, if the ITE is aware that the ETE 894 is located somewhere in the public Internet, middlebox reassembly 895 testing is unnecessary. If the ITE is aware that the ETE is located 896 behind a NAT or a firewall, however, then middlebox reassembly 897 testing is recommended. 899 The ITE can perform a middlebox reassembly test by setting U=1 in the 900 header of a SEAL data packet to be used as a probe. Next, the ITE 901 encapsulates the packet in the appropriate outer headers, splits it 902 into two outer IPv4 fragments, then sends both fragments over the 903 same ETE link path. 905 While performing the test, the ITE should select only inner packets 906 that are no larger than 1280 bytes for testing purposes in order to 907 avoid reassembly buffer overruns. The ITE can also construct a NULL 908 test packet instead of using ordinary SEAL data packets for testing. 910 To create the NULL packet, the ITE prepares a data packet with (C=0; 911 P=1; R=0; T=0; U=1; Z=0) in the SEAL header, writes the length of the 912 ITE's claimed prefix in the PREFLEN field, and writes the ITE's 913 claimed prefix in the PREFIX field. The ITE then sets NEXTHDR 914 according to the address family of the PREFIX, i.e., it sets NEXTHDR 915 to the value '4' for an IPv4 prefix, '41' for an IPv6 prefix , '80' 916 for an OSI/CLNP prefix, etc. 918 The ITE can further add padding following the PREFIX field to a 919 length that would not cause the size of the NULL packet to exceed 920 1280 bytes before encapsulation. The ITE then sets LINK_ID, LEVEL 921 and PKT_ID to the appropriate values for this ETE link path and 922 calculates ICV1 the same as for an ordinary SEAL data packet. 924 The ITE should send a series of test packets (e.g., 3-5 tests with 925 1sec intervals between tests) instead of a single isolated test in 926 case of packet loss, and will eventually receive an SPTB message from 927 the ITE (see: Section 4.6.2.1). If the ETE returns an SCMP PTB 928 message with MTU != 0, then the ETE link path correctly supports 929 fragmentation. 931 If the ETE returns an SCMP PTB message with MTU==0, however, then a 932 middlebox in the subnetwork is reassembling the fragments before 933 forwarding them to the ETE. In that case, the ITE sets 934 PATH_MTU=MIN_MTU and sets (USE_TRAILER=TRUE; USE_DF=FALSE). The ITE 935 may instead enable stateful MTU determination for this ETE link path 936 as specified in Section 4.4.9 to attempt to discover larger MTUs. 938 NB: Examples of middleboxes that may perform reassembly include 939 stateful NATs and firewalls. Such devices could still allow for 940 stateless MTU determination if they gather the fragments of a 941 fragmented IPv4 SEAL data packet for packet analysis purposes but 942 then forward the fragments on to the final destination rather than 943 forwarding the reassembled packet. 945 4.4.9. Stateful MTU Determination 947 SEAL supports a stateless MTU determination capability, however the 948 ITE may in some instances wish to impose a stateful MTU limit on a 949 particular ETE link path. For example, when the ETE is situated 950 behind a middlebox that performs IPv4 reassembly (see: Section 4.4.8) 951 it is imperative that fragmentation of large packets be avoided on 952 the path to the middlebox. In other instances (e.g., when the ETE 953 link path includes performance-constrained links), the ITE may deem 954 it necessary to cache a conservative static MTU in order to avoid 955 sending large packets that would only be dropped due to an MTU 956 restriction somewhere on the path. 958 To determine a static MTU value, the ITE can send a series of probe 959 packets of various sizes to the ETE with U=1 in the SEAL header and 960 DF=1 in the outer IP header. The ITE can then cache the size of the 961 largest packet for which it receives a probe reply from the ETE as 962 the PATH_MTU value this ETE link path. 964 For example, the ITE could send NULL probe packets of 1500 bytes, 965 followed by 1450 bytes, followed by 1400 bytes, etc. then set 966 PATH_MTU for this ETE link path to the size of the largest probe 967 packet for which it receives an SPTB reply message. While probing 968 with NULL probe packets, the ITE processes any ICMP PTB message it 969 receives as a potential indication of probe failure then discards the 970 message. 972 For IPv4, if the largest successful probe is larger than MIN_MTU the 973 ITE then sets (USE_TRAILER=FALSE; USE_DF=TRUE) for this ETE link 974 path; otherwise, the ITE sets (USE_TRAILER=TRUE; USE_DF=FALSE). 976 4.4.10. Detecting Path MTU Changes 978 For IPv6, the ITE can periodically reset PATH_MTU to the MTU of the 979 underlying link to determine whether the ETE link path now supports 980 larger packet sizes. If the path still has a too-small MTU, the ITE 981 will receive a PTB message that reports a smaller size. 983 For IPv4, when USE_TRAILER==TRUE and PATH_MTU is larger than MIN_MTU 984 the ITE can periodically reset USE_TRAILER=FALSE to determine whether 985 the ETE link path still requires trailers. If the ITE receives an 986 SPTB message for an inner packet that is no larger than 1280 bytes 987 (see: Section 4.6.1.1), the ITE should again set USE_TRAILER=TRUE. 989 When stateful MTU determination is used, the ITE should periodically 990 re-probe the path as described in Section 4.4.9 to determine whether 991 routing changes have resulted in a reduced or increased PATH_MTU. 993 4.5. ETE Specification 995 4.5.1. Tunnel Neighbor Soft State 997 The ETE maintains a per-ITE ICV calculation algorithm and (when data 998 origin authentication is required) a symmetric secret key to verify 999 the ICV(s) in the SEAL header and trailer. The ETE also maintains a 1000 window of PKT_ID values for the packets it has recently received from 1001 this ITE. 1003 4.5.2. IP-Layer Reassembly 1005 The ETE must maintain a minimum IP-layer reassembly buffer size of 1006 1500 bytes for both IPv4 [RFC0791] and IPv6 [RFC2460]. 1008 The ETE should maintain conservative reassembly cache high- and low- 1009 water marks. When the size of the reassembly cache exceeds this 1010 high-water mark, the ETE should actively discard stale incomplete 1011 reassemblies (e.g., using an Active Queue Management (AQM) strategy) 1012 until the size falls below the low-water mark. The ETE should also 1013 actively discard any pending reassemblies that clearly have no 1014 opportunity for completion, e.g., when a considerable number of new 1015 fragments have arrived before a fragment that completes a pending 1016 reassembly arrives. 1018 The ETE processes non-SEAL IP packets as specified in the normative 1019 references, i.e., it performs any necessary IP reassembly then 1020 discards the packet if it is larger than the reassembly buffer size 1021 or delivers the (fully-reassembled) packet to the appropriate upper 1022 layer protocol module. 1024 For SEAL packets, the ITE performs any necessary IP reassembly until 1025 it has received at least the first 1280 bytes beyond the SEAL header 1026 or up to the end of the packet. For IPv4, the ETE then submits the 1027 (fully- or partially-reassembled) packet for decapsulation as 1028 specified in Section 4.5.3. For IPv6, the ETE only submits the 1029 packet if it was fully-reassembled and no larger than the reassembly 1030 buffer size. 1032 4.5.3. Decapsulation and Re-Encapsulation 1034 For each SEAL packet submitted for decapsulation, the ETE first 1035 examines the PKT_ID and ICV1 fields. If the PKT_ID is not within the 1036 window of acceptable values for this ITE, or if the ICV1 field 1037 includes an incorrect value, the ETE silently discards the packet. 1039 Next, if the SEAL header has T==1 and the inner packet is larger than 1040 1280 bytes the ETE silently discards the packet. If the SEAL header 1041 has T==1 and the inner packet is no larger than 1280 bytes, the ETE 1042 instead verifies the ICV2 value and silently discards the packet if 1043 the value is incorrect. 1045 Next, if the SEAL header has C==0 and there is an incorrect value in 1046 a SEAL header field (e.g., an incorrect "VER" field value), the ETE 1047 returns an SCMP "Parameter Problem" (SPP) message (see Section 1048 4.6.1.2) and discards the packet. 1050 Next, if the packet arrived as multiple IPv4 fragments and the inner 1051 packet is larger than 1280 bytes, the ETE sends an SPTB message back 1052 to the ITE with MTU set to the size of the largest fragment received 1053 minus HLEN (see: Section 4.6.1.1) then discards the packet. If the 1054 packet arrived as multiple IPv6 fragments and the inner packet is 1055 larger than 1280 bytes, the ETE instead silently discards the packet. 1057 Next, if the packet arrived as multiple IPv4 fragments, the SEAL 1058 header has (C==0; T==0), and the inner packet is larger than (128- 1059 SHLEN-UHLEN) bytes, the ETE sends an SPTB message back to the ITE 1060 with MTU set to the size of the largest fragment received minus HLEN 1061 (see: Section 4.6.1.1) then continues to process the packet. 1063 Next, if the SEAL header has C==1, the ETE processes the packet as an 1064 SCMP packet as specified in Section 4.6.2. Otherwise, the ETE 1065 continues to process the packet as a SEAL data packet. 1067 Next, if the packet arrived unfragmented and the SEAL header has 1068 U==1, the ETE sends an SPTB message back to the ITE with MTU=0 (see: 1069 Section 4.6.1.1). 1071 Next, if the SEAL header has P==1 the ETE discards the (NULL) packet. 1073 Finally, the ETE discards the outer headers and processes the inner 1074 packet according to the header type indicated in the SEAL NEXTHDR 1075 field. If the next hop toward the inner destination address is via a 1076 different interface than the SEAL packet arrived on, the ETE discards 1077 the SEAL header and delivers the inner packet either to the local 1078 host or to the next hop interface if the packet is not destined to 1079 the local host. 1081 If the next hop is on the same interface the SEAL packet arrived on, 1082 however, the ETE submits the packet for SEAL re-encapsulation 1083 beginning with the specification in Section 4.4.3 above. In this 1084 process, the packet remains within the tunnel interface (i.e., it 1085 does not exit and then re-enter the interface); hence, the packet is 1086 not discarded if the LEVEL field in the SEAL header contains the 1087 value 0. 1089 4.6. The SEAL Control Message Protocol (SCMP) 1091 SEAL provides a companion SEAL Control Message Protocol (SCMP) that 1092 uses the same message types and formats as for the Internet Control 1093 Message Protocol for IPv6 (ICMPv6) [RFC4443]. As for ICMPv6, each 1094 SCMP message includes a 4-byte header and a variable-length body. 1095 The TE encapsulates the SCMP message in a SEAL header and outer 1096 headers as shown in Figure 4: 1098 +--------------------+ 1099 ~ outer IP header ~ 1100 +--------------------+ 1101 ~ other outer hdrs ~ 1102 +--------------------+ 1103 ~ SEAL Header ~ 1104 +--------------------+ +--------------------+ 1105 | SCMP message header| --> | SCMP message header| 1106 +--------------------+ +--------------------+ 1107 | | --> | | 1108 ~ SCMP message body ~ --> ~ SCMP message body ~ 1109 | | --> | | 1110 +--------------------+ +--------------------+ 1112 SCMP Message SCMP Packet 1113 before encapsulation after encapsulation 1115 Figure 4: SCMP Message Encapsulation 1117 The following sections specify the generation, processing and 1118 relaying of SCMP messages. 1120 4.6.1. Generating SCMP Error Messages 1122 ETEs generate SCMP error messages in response to receiving certain 1123 SEAL data packets using the format shown in Figure 5: 1125 0 1 2 3 1126 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1127 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1128 | Type | Code | Checksum | 1129 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1130 | Type-Specific Data | 1131 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1132 | As much of the inner packet within the invoking | 1133 ~ SEAL data packet as possible without the SCMP ~ 1134 | packet exceeding 576 bytes (*) | 1136 (*) also known as the "packet-in-error" 1138 Figure 5: SCMP Error Message Format 1140 The error message includes the 4 byte SCMP message header, followed 1141 by a 4 byte Type-Specific Data field, followed by the leading portion 1142 of the inner packet within the invoking SEAL data packet (i.e., 1143 beginning immediately after the SEAL header) as the "packet-in- 1144 error". The packet-in-error includes as much of the inner packet as 1145 possible extending to a length that would not cause the entire SCMP 1146 packet following outer encapsulation to exceed 576 bytes. 1148 When the ETE processes a SEAL data packet for which the ICVs are 1149 correct but an error must be returned, it prepares an SCMP error 1150 message as shown in Figure 5. The ETE sets the Type and Code fields 1151 to the same values that would appear in the corresponding ICMPv6 1152 message and calculates the Checksum beginning with the SCMP message 1153 header and continuing to the end of the message. (When calculating 1154 the Checksum, the TE sets the Checksum field itself to 0.) 1156 The ETE next encapsulates the SCMP message in the requisite SEAL 1157 header, outer headers and SEAL trailer as shown in Figure 4. During 1158 encapsulation, the ETE sets the outer destination address/port 1159 numbers of the SCMP packet to the outer source address/port numbers 1160 of the original SEAL data packet and sets the outer source address/ 1161 port numbers to its own outer address/port numbers. 1163 The ETE then sets (C=1; R=0; T=0; U=0; Z=0) in the SEAL header, then 1164 sets NEXTHDR, PREFLEN, LINK_ID, LEVEL, and PKT_ID to the same values 1165 that appeared in the SEAL header of the data packet. If the SEAL 1166 data packet header had P==1, the ETE also copies the PREFIX field 1167 from the data packet into the SEAL header and sets P=1; otherwise, it 1168 sets P=0. 1170 The ETE then calculates and sets the ICV1 field the same as specified 1171 for SEAL data packet encapsulation in Section 4.4.4. Next, the ETE 1172 encapsulates the SCMP message in the requisite outer encapsulations 1173 and sends the resulting SCMP packet to the ITE the same as specified 1174 for SEAL data packets in Section 4.4.5. 1176 The following sections describe additional considerations for various 1177 SCMP error messages: 1179 4.6.1.1. Generating SCMP Packet Too Big (SPTB) Messages 1181 An ETE generates an SCMP "Packet Too Big" (SPTB) message when it 1182 receives a SEAL data packet that arrived as multiple outer IPv4 1183 fragments and for which the reassembled inner packet would be larger 1184 than 1280 bytes. The ETE also generates an SPTB when it receives the 1185 fragments of a fragmented IPv4-encapsulated SEAL data packet with 1186 T==0 in the SEAL header but that following reassembly would be larger 1187 than (128-SHLEN-UHLEN) bytes but no larger than 1280 bytes. The ETE 1188 prepares the SPTB message the same as for the corresponding ICMPv6 1189 PTB message, and writes the length of the largest outer IP fragment 1190 received minus HLEN in the MTU field of the message. 1192 The ETE also generates an SPTB message when it accepts a SEAL 1193 protocol data packet which did not undergo IP fragmentation and with 1194 U==1 in the SEAL header. The ETE prepares the SPTB message the same 1195 as above, except that it writes the value 0 in the MTU field. 1197 4.6.1.2. Generating Other SCMP Error Messages 1199 An ETE generates an SCMP "Destination Unreachable" (SDU) message 1200 under the same circumstances that an IPv6 system would generate an 1201 ICMPv6 Destination Unreachable message. 1203 An ETE generates an SCMP "Parameter Problem" (SPP) message when it 1204 receives a SEAL packet with an incorrect value in the SEAL header. 1205 IN THIS CASE ALONE, the ETE prepares the packet-in-error beginning 1206 with the SEAL header instead of beginning immediately after the SEAL 1207 header. 1209 TEs generate other SCMP message types using methods and procedures 1210 specified in other documents. For example, SCMP message types used 1211 for tunnel neighbor coordinations are specified in VET 1212 [I-D.templin-intarea-vet]. 1214 4.6.2. Processing SCMP Error Messages 1216 An ITE may receive SCMP messages after sending packets to an ETE. 1217 The ITE first verifies that the outer addresses of the SCMP packet 1218 are correct, and that the PKT_ID is within its window of values for 1219 this ETE. The ITE next verifies that the SEAL header fields are set 1220 correctly as specified in Section 4.6.1. The ITE then verifies the 1221 ICV1 value. If the outer addresses, SEAL header information and/or 1222 ICV1 value are incorrect, the ITE silently discards the message; 1223 otherwise, it processes the message as follows: 1225 4.6.2.1. Processing SCMP PTB Messages 1227 After an ITE sends a SEAL data packet to an ETE, it may receive an 1228 SPTB message with a packet-in-error containing the leading portion of 1229 the inner packet (see: Section 4.6.1.1). For IP SPTB messages with 1230 MTU==0, the ITE processes the message as confirmation that the ETE 1231 received an unfragmented SEAL data packet with U==1 in the SEAL 1232 header. The ITE then discards the message. 1234 For IPv4 SPTB messages with MTU != 0, the ITE instead processes the 1235 message as an indication of a packet size limitation as follows. The 1236 ITE first determines the inner packet length by subtracting SHLEN 1237 from the length field in the UDP header within the packet-in-error 1238 (and also subtracting the length of the SEAL trailer when T=1). If 1239 the inner packet is no larger than 1280 bytes, the ITE sets 1240 USE_TRAILER=TRUE. If the inner packet is larger than 1280 bytes, the 1241 ITE instead examines the SPTB message MTU field. If the MTU value is 1242 not substantially less than (1500-HLEN), the value is likely to 1243 reflect the true MTU of the restricting link on the path to the ETE; 1244 otherwise, a router on the path may be generating runt fragments. 1246 In that case, the ITE can consult a plateau table (e.g., as described 1247 in [RFC1191]) to rewrite the MTU value to a reduced size. For 1248 example, if the ITE receives an IPv4 SPTB message with MTU==256 and 1249 inner packet length 1500, it can rewrite the MTU to 1450. If the ITE 1250 subsequently receives an IPv4 SPTB message with MTU==256 and inner 1251 packet length 1450, it can rewrite the MTU to 1400, etc. If the ITE 1252 is performing stateful MTU determination for this ETE link path, it 1253 then writes the new MTU value in PATH_MTU. 1255 The ITE then checks its forwarding tables to discover the previous 1256 hop toward the source address of the inner packet. If the previous 1257 hop is reached via the same tunnel interface the SPTB message arrived 1258 on, the ITE relays the message to the previous hop. In order to 1259 relay the message, the ITE rewrites the SEAL header fields with 1260 values corresponding to the previous hop and recalculates the ICV1 1261 values using the ICV calculation parameters associated with the 1262 previous hop. Next, the ITE replaces the SPTB's outer headers with 1263 headers of the appropriate protocol version and fills in the header 1264 fields as specified in Sections 5.5.4-5.5.6 of 1265 [I-D.templin-intarea-vet], where the destination address/port 1266 correspond to the previous hop and the source address/port correspond 1267 to the ITE. The ITE then sends the message to the previous hop the 1268 same as if it were issuing a new SPTB message. 1270 If the previous hop is not reached via the same tunnel interface, the 1271 ITE instead transcribes the message into a format appropriate for the 1272 inner packet (i.e., the same as described for transcribing ICMP 1273 messages in Section 4.4.7) and sends the resulting transcribed 1274 message to the original source. The ITE then discards the SPTB 1275 message. 1277 4.6.2.2. Processing Other SCMP Error Messages 1279 An ITE may receive an SDU message with an appropriate code under the 1280 same circumstances that an IPv6 node would receive an ICMPv6 1281 Destination Unreachable message. The ITE either transcribes or 1282 relays the message toward the source address of the inner packet 1283 within the packet-in-error the same as specified for SPTB messages in 1284 Section 4.6.2.1. 1286 An ITE may receive an SPP message when the ETE receives a SEAL packet 1287 with an incorrect value in the SEAL header. The ITE should examine 1288 the SEAL header within the packet-in-error to determine whether a 1289 different setting should be used in subsequent packets, but does not 1290 relay the message further. 1292 TEs process other SCMP message types using methods and procedures 1293 specified in other documents. For example, SCMP message types used 1294 for tunnel neighbor coordinations are specified in VET 1295 [I-D.templin-intarea-vet]. 1297 5. Link Requirements 1299 Subnetwork designers are expected to follow the recommendations in 1300 Section 2 of [RFC3819] when configuring link MTUs. 1302 6. End System Requirements 1304 End systems are encouraged to implement end-to-end MTU assurance 1305 (e.g., using Packetization Layer Path MTU Discovery per [RFC4821]) 1306 even if the subnetwork is using SEAL. 1308 7. Router Requirements 1310 Routers within the subnetwork are expected to observe the router 1311 requirements found in the normative references, including the 1312 implementation of IP fragmentation and reassembly [RFC1812][RFC2460] 1313 as well as the generation of ICMP messages [RFC0792][RFC4443]. 1315 8. Nested Encapsulation Considerations 1317 SEAL supports nested tunneling for up to 8 layers of encapsulation. 1318 In this model, the SEAL ITE has a tunnel neighbor relationship only 1319 with ETEs at its own nesting level, i.e., it does not have a tunnel 1320 neighbor relationship with any ITEs/ETEs at other nesting levels. 1322 Therefore, when an ITE 'A' within an inner nesting level needs to 1323 return an error message to an ITE 'B' within an outer nesting level, 1324 it generates an ordinary ICMP error message the same as if it were an 1325 ordinary router within the subnetwork. 'B' can then perform message 1326 validation as specified in Section 4.4.7, but full message origin 1327 authentication is not possible. 1329 Since ordinary ICMP messages are used for coordinations between ITEs 1330 at different nesting levels, nested SEAL encapsulations should only 1331 be used when the ITEs are within a common administrative domain 1332 and/or when there is no ICMP filtering middlebox such as a firewall 1333 or NAT between them. An example would be a recursive nesting of 1334 mobile networks, where the first network receives service from an 1335 ISP, the second network receives service from the first network, the 1336 third network receives service from the second network, etc. 1338 9. IANA Considerations 1340 The IANA is instructed to allocate an IP protocol number for 1341 'SEAL_PROTO' in the 'protocol-numbers' registry. 1343 The IANA is instructed to allocate a Well-Known Port number for 1344 'SEAL_PORT' in the 'port-numbers' registry. 1346 The IANA is instructed to establish a "SEAL Protocol" registry to 1347 record SEAL Version values. This registry should be initialized to 1348 include the initial SEAL Version number, i.e., Version 0. 1350 10. Security Considerations 1352 SEAL provides a segment-by-segment data origin authentication and 1353 anti-replay service across the (potentially) multiple segments of a 1354 re-encapsulating tunnel. It further provides a segment-by-segment 1355 integrity check of the headers of encapsulated packets, but does not 1356 verify the integrity of the rest of the packet beyond the headers 1357 unless fragmentation is unavoidable. SEAL therefore considers full 1358 message integrity checking, authentication and confidentiality as 1359 end-to-end considerations in a manner that is compatible with 1360 securing mechanisms such as TLS/SSL [RFC5246]. 1362 An amplification/reflection/buffer overflow attack is possible when 1363 an attacker sends IP fragments with spoofed source addresses to an 1364 ETE in an attempt to clog the ETE's reassembly buffer and/or cause 1365 the ETE to generate a stream of SCMP messages returned to a victim 1366 ITE. The SCMP message ICVs, PKT_ID, as well as the inner headers of 1367 the packet-in-error, provide mitigation for the ETE to detect and 1368 discard SEAL segments with spoofed source addresses. 1370 The SEAL header is sent in-the-clear the same as for the outer IP and 1371 other outer headers. In this respect, the threat model is no 1372 different than for IPv6 extension headers. Unlike IPv6 extension 1373 headers, however, the SEAL header is protected by an integrity check 1374 that also covers the inner packet headers. 1376 Security issues that apply to tunneling in general are discussed in 1377 [RFC6169]. 1379 11. Related Work 1381 Section 3.1.7 of [RFC2764] provides a high-level sketch for 1382 supporting large tunnel MTUs via a tunnel-level segmentation and 1383 reassembly capability to avoid IP level fragmentation. This 1384 capability was implemented in the first edition of SEAL, but is now 1385 deprecated. 1387 Section 3 of [RFC4459] describes inner and outer fragmentation at the 1388 tunnel endpoints as alternatives for accommodating the tunnel MTU. 1390 Section 4 of [RFC2460] specifies a method for inserting and 1391 processing extension headers between the base IPv6 header and 1392 transport layer protocol data. The SEAL header is inserted and 1393 processed in exactly the same manner. 1395 IPsec/AH is [RFC4301][RFC4301] is used for full message integrity 1396 verification between tunnel endpoints, whereas SEAL only ensures 1397 integrity for the inner packet headers. The AYIYA proposal 1398 [I-D.massar-v6ops-ayiya] uses similar means for providing full 1399 message authentication and integrity. 1401 The concepts of path MTU determination through the report of 1402 fragmentation and extending the IPv4 Identification field were first 1403 proposed in deliberations of the TCP-IP mailing list and the Path MTU 1404 Discovery Working Group (MTUDWG) during the late 1980's and early 1405 1990's. An historical analysis of the evolution of these concepts, 1406 as well as the development of the eventual path MTU discovery 1407 mechanism, appears in Appendix D of this document. 1409 12. Acknowledgments 1411 The following individuals are acknowledged for helpful comments and 1412 suggestions: Jari Arkko, Fred Baker, Iljitsch van Beijnum, Oliver 1413 Bonaventure, Teco Boot, Bob Braden, Brian Carpenter, Steve Casner, 1414 Ian Chakeres, Noel Chiappa, Remi Denis-Courmont, Remi Despres, Ralph 1415 Droms, Aurnaud Ebalard, Gorry Fairhurst, Washam Fan, Dino Farinacci, 1416 Joel Halpern, Sam Hartman, John Heffner, Thomas Henderson, Bob 1417 Hinden, Christian Huitema, Eliot Lear, Darrel Lewis, Joe Macker, Matt 1418 Mathis, Erik Nordmark, Dan Romascanu, Dave Thaler, Joe Touch, Mark 1419 Townsley, Ole Troan, Margaret Wasserman, Magnus Westerlund, Robin 1420 Whittle, James Woodyatt, and members of the Boeing Research & 1421 Technology NST DC&NT group. 1423 Discussions with colleagues following the publication of RFC5320 have 1424 provided useful insights that have resulted in significant 1425 improvements to this, the Second Edition of SEAL. 1427 Path MTU determination through the report of fragmentation was first 1428 proposed by Charles Lynn on the TCP-IP mailing list in 1987. 1429 Extending the IP identification field was first proposed by Steve 1430 Deering on the MTUDWG mailing list in 1989. 1432 13. References 1434 13.1. Normative References 1436 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 1437 September 1981. 1439 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 1440 RFC 792, September 1981. 1442 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1443 Requirement Levels", BCP 14, RFC 2119, March 1997. 1445 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1446 (IPv6) Specification", RFC 2460, December 1998. 1448 [RFC3971] Arkko, J., Kempf, J., Zill, B., and P. Nikander, "SEcure 1449 Neighbor Discovery (SEND)", RFC 3971, March 2005. 1451 [RFC4443] Conta, A., Deering, S., and M. Gupta, "Internet Control 1452 Message Protocol (ICMPv6) for the Internet Protocol 1453 Version 6 (IPv6) Specification", RFC 4443, March 2006. 1455 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 1456 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 1457 September 2007. 1459 13.2. Informative References 1461 [FOLK] Shannon, C., Moore, D., and k. claffy, "Beyond Folklore: 1462 Observations on Fragmented Traffic", December 2002. 1464 [FRAG] Kent, C. and J. Mogul, "Fragmentation Considered Harmful", 1465 October 1987. 1467 [I-D.ietf-intarea-ipv4-id-update] 1468 Touch, J., "Updated Specification of the IPv4 ID Field", 1469 draft-ietf-intarea-ipv4-id-update-04 (work in progress), 1470 September 2011. 1472 [I-D.ietf-savi-framework] 1473 Wu, J., Bi, J., Bagnulo, M., Baker, F., and C. Vogt, 1474 "Source Address Validation Improvement Framework", 1475 draft-ietf-savi-framework-05 (work in progress), 1476 July 2011. 1478 [I-D.massar-v6ops-ayiya] 1479 Massar, J., "AYIYA: Anything In Anything", 1480 draft-massar-v6ops-ayiya-02 (work in progress), July 2004. 1482 [I-D.templin-aero] 1483 Templin, F., "Asymmetric Extended Route Optimization 1484 (AERO)", draft-templin-aero-04 (work in progress), 1485 October 2011. 1487 [I-D.templin-intarea-vet] 1488 Templin, F., "Virtual Enterprise Traversal (VET)", 1489 draft-templin-intarea-vet-29 (work in progress), 1490 November 2011. 1492 [I-D.templin-ironbis] 1493 Templin, F., "The Internet Routing Overlay Network 1494 (IRON)", draft-templin-ironbis-07 (work in progress), 1495 October 2011. 1497 [MTUDWG] "IETF MTU Discovery Working Group mailing list, 1498 gatekeeper.dec.com/pub/DEC/WRL/mogul/mtudwg-log, November 1499 1989 - February 1995.". 1501 [RFC1063] Mogul, J., Kent, C., Partridge, C., and K. McCloghrie, "IP 1502 MTU discovery options", RFC 1063, July 1988. 1504 [RFC1070] Hagens, R., Hall, N., and M. Rose, "Use of the Internet as 1505 a subnetwork for experimentation with the OSI network 1506 layer", RFC 1070, February 1989. 1508 [RFC1191] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191, 1509 November 1990. 1511 [RFC1812] Baker, F., "Requirements for IP Version 4 Routers", 1512 RFC 1812, June 1995. 1514 [RFC1981] McCann, J., Deering, S., and J. Mogul, "Path MTU Discovery 1515 for IP version 6", RFC 1981, August 1996. 1517 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 1518 October 1996. 1520 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 1521 IPv6 Specification", RFC 2473, December 1998. 1523 [RFC2675] Borman, D., Deering, S., and R. Hinden, "IPv6 Jumbograms", 1524 RFC 2675, August 1999. 1526 [RFC2764] Gleeson, B., Heinanen, J., Lin, A., Armitage, G., and A. 1527 Malis, "A Framework for IP Based Virtual Private 1528 Networks", RFC 2764, February 2000. 1530 [RFC2923] Lahey, K., "TCP Problems with Path MTU Discovery", 1531 RFC 2923, September 2000. 1533 [RFC3232] Reynolds, J., "Assigned Numbers: RFC 1700 is Replaced by 1534 an On-line Database", RFC 3232, January 2002. 1536 [RFC3366] Fairhurst, G. and L. Wood, "Advice to link designers on 1537 link Automatic Repeat reQuest (ARQ)", BCP 62, RFC 3366, 1538 August 2002. 1540 [RFC3819] Karn, P., Bormann, C., Fairhurst, G., Grossman, D., 1541 Ludwig, R., Mahdavi, J., Montenegro, G., Touch, J., and L. 1542 Wood, "Advice for Internet Subnetwork Designers", BCP 89, 1543 RFC 3819, July 2004. 1545 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 1546 More-Specific Routes", RFC 4191, November 2005. 1548 [RFC4213] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms 1549 for IPv6 Hosts and Routers", RFC 4213, October 2005. 1551 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 1552 Internet Protocol", RFC 4301, December 2005. 1554 [RFC4302] Kent, S., "IP Authentication Header", RFC 4302, 1555 December 2005. 1557 [RFC4380] Huitema, C., "Teredo: Tunneling IPv6 over UDP through 1558 Network Address Translations (NATs)", RFC 4380, 1559 February 2006. 1561 [RFC4459] Savola, P., "MTU and Fragmentation Issues with In-the- 1562 Network Tunneling", RFC 4459, April 2006. 1564 [RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU 1565 Discovery", RFC 4821, March 2007. 1567 [RFC4963] Heffner, J., Mathis, M., and B. Chandler, "IPv4 Reassembly 1568 Errors at High Data Rates", RFC 4963, July 2007. 1570 [RFC4987] Eddy, W., "TCP SYN Flooding Attacks and Common 1571 Mitigations", RFC 4987, August 2007. 1573 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 1574 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 1576 [RFC5445] Watson, M., "Basic Forward Error Correction (FEC) 1577 Schemes", RFC 5445, March 2009. 1579 [RFC5720] Templin, F., "Routing and Addressing in Networks with 1580 Global Enterprise Recursion (RANGER)", RFC 5720, 1581 February 2010. 1583 [RFC5927] Gont, F., "ICMP Attacks against TCP", RFC 5927, July 2010. 1585 [RFC6139] Russert, S., Fleischman, E., and F. Templin, "Routing and 1586 Addressing in Networks with Global Enterprise Recursion 1587 (RANGER) Scenarios", RFC 6139, February 2011. 1589 [RFC6169] Krishnan, S., Thaler, D., and J. Hoagland, "Security 1590 Concerns with IP Tunneling", RFC 6169, April 2011. 1592 [SIGCOMM] Luckie, M. and B. Stasiewicz, "Measuring Path MTU 1593 Discovery Behavior", November 2010. 1595 [TBIT] Medina, A., Allman, M., and S. Floyd, "Measuring 1596 Interactions Between Transport Protocols and Middleboxes", 1597 October 2004. 1599 [TCP-IP] "Archive/Hypermail of Early TCP-IP Mail List, 1600 http://www-mice.cs.ucl.ac.uk/multimedia/misc/tcp_ip/, May 1601 1987 - May 1990.". 1603 [WAND] Luckie, M., Cho, K., and B. Owens, "Inferring and 1604 Debugging Path MTU Discovery Failures", October 2005. 1606 Appendix A. Reliability 1608 Although a SEAL tunnel may span an arbitrarily-large subnetwork 1609 expanse, the IP layer sees the tunnel as a simple link that supports 1610 the IP service model. Links with high bit error rates (BERs) (e.g., 1611 IEEE 802.11) use Automatic Repeat-ReQuest (ARQ) mechanisms [RFC3366] 1612 to increase packet delivery ratios, while links with much lower BERs 1613 typically omit such mechanisms. Since SEAL tunnels may traverse 1614 arbitrarily-long paths over links of various types that are already 1615 either performing or omitting ARQ as appropriate, it would therefore 1616 often be inefficient to also require the tunnel endpoints to also 1617 perform ARQ. 1619 Appendix B. Integrity 1621 The SEAL header includes an ICV field that covers the SEAL header and 1622 at least the inner packet headers. This provides for header 1623 integrity verification on a segment-by-segment basis for a segmented 1624 re-encapsulating tunnel path. When IPv4 fragmentation is needed, the 1625 SEAL packet also contains a trailer with a secondary ICV that covers 1626 the remainder of the packet. 1628 Fragmentation and reassembly schemes must consider packet-splicing 1629 errors, e.g., when two fragments from the same packet are 1630 concatenated incorrectly, when a fragment from packet X is 1631 reassembled with fragments from packet Y, etc. The primary sources 1632 of such errors include implementation bugs and wrapping IPv4 ID 1633 fields. 1635 In terms of wrapping ID fields, the IPv4 16-bit ID field can wrap 1636 with only 64K packets with the same (src, dst, protocol)-tuple alive 1637 in the system at a given time [RFC4963] increasing the likelihood of 1638 reassembly mis-associations 1640 When reassembly is unavoidable, SEAL provides an extended ICV to 1641 detect reassembly mis-associations for packets no larger than 1280 1642 bytes and also discards any reassembled packets larger than 1280 1643 bytes. 1645 Appendix C. Transport Mode 1647 SEAL can also be used in "transport-mode", e.g., when the inner layer 1648 comprises upper-layer protocol data rather than an encapsulated IP 1649 packet. For instance, TCP peers can negotiate the use of SEAL (e.g., 1650 by inserting a 'SEAL_OPTION' TCP option during connection 1651 establishment) for the carriage of protocol data encapsulated as IP/ 1652 SEAL/TCP. In this sense, the "subnetwork" becomes the entire end-to- 1653 end path between the TCP peers and may potentially span the entire 1654 Internet. 1656 If both TCPs agree on the use of SEAL, their protocol messages will 1657 be carried as IP/SEAL/TCP and the connection will be serviced by the 1658 SEAL protocol using TCP (instead of an encapsulating tunnel endpoint) 1659 as the transport layer protocol. The SEAL protocol for transport 1660 mode otherwise observes the same specifications as for Section 4. 1662 Appendix D. Historic Evolution of PMTUD 1664 The topic of Path MTU discovery (PMTUD) saw a flurry of discussion 1665 and numerous proposals in the late 1980's through early 1990. The 1666 initial problem was posed by Art Berggreen on May 22, 1987 in a 1667 message to the TCP-IP discussion group [TCP-IP]. The discussion that 1668 followed provided significant reference material for [FRAG]. An IETF 1669 Path MTU Discovery Working Group [MTUDWG] was formed in late 1989 1670 with charter to produce an RFC. Several variations on a very few 1671 basic proposals were entertained, including: 1673 1. Routers record the PMTUD estimate in ICMP-like path probe 1674 messages (proposed in [FRAG] and later [RFC1063]) 1676 2. The destination reports any fragmentation that occurs for packets 1677 received with the "RF" (Report Fragmentation) bit set (Steve 1678 Deering's 1989 adaptation of Charles Lynn's Nov. 1987 proposal) 1680 3. A hybrid combination of 1) and Charles Lynn's Nov. 1987 (straw 1681 RFC draft by McCloughrie, Fox and Mogul on Jan 12, 1990) 1683 4. Combination of the Lynn proposal with TCP (Fred Bohle, Jan 30, 1684 1990) 1686 5. Fragmentation avoidance by setting "IP_DF" flag on all packets 1687 and retransmitting if ICMPv4 "fragmentation needed" messages 1688 occur (Geof Cooper's 1987 proposal; later adapted into [RFC1191] 1689 by Mogul and Deering). 1691 Option 1) seemed attractive to the group at the time, since it was 1692 believed that routers would migrate more quickly than hosts. Option 1693 2) was a strong contender, but repeated attempts to secure an "RF" 1694 bit in the IPv4 header from the IESG failed and the proponents became 1695 discouraged. 3) was abandoned because it was perceived as too 1696 complicated, and 4) never received any apparent serious 1697 consideration. Proposal 5) was a late entry into the discussion from 1698 Steve Deering on Feb. 24th, 1990. The discussion group soon 1699 thereafter seemingly lost track of all other proposals and adopted 1700 5), which eventually evolved into [RFC1191] and later [RFC1981]. 1702 In retrospect, the "RF" bit postulated in 2) is not needed if a 1703 "contract" is first established between the peers, as in proposal 4) 1704 and a message to the MTUDWG mailing list from jrd@PTT.LCS.MIT.EDU on 1705 Feb 19. 1990. These proposals saw little discussion or rebuttal, and 1706 were dismissed based on the following the assertions: 1708 o routers upgrade their software faster than hosts 1710 o PCs could not reassemble fragmented packets 1712 o Proteon and Wellfleet routers did not reproduce the "RF" bit 1713 properly in fragmented packets 1715 o Ethernet-FDDI bridges would need to perform fragmentation (i.e., 1716 "translucent" not "transparent" bridging) 1718 o the 16-bit IP_ID field could wrap around and disrupt reassembly at 1719 high packet arrival rates 1721 The first four assertions, although perhaps valid at the time, have 1722 been overcome by historical events. The final assertion is addressed 1723 by the mechanisms specified in SEAL. 1725 Author's Address 1727 Fred L. Templin (editor) 1728 Boeing Research & Technology 1729 P.O. Box 3707 1730 Seattle, WA 98124 1731 USA 1733 Email: fltemplin@acm.org