idnits 2.17.1 draft-templin-intarea-seal-44.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 (June 23, 2012) is 4315 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Unused Reference: 'RFC3971' is defined on line 1471, but no explicit reference was found in the text == Unused Reference: 'RFC4861' is defined on line 1478, but no explicit reference was found in the text == Unused Reference: 'I-D.templin-aero' is defined on line 1505, but no explicit reference was found in the text == Unused Reference: 'RFC1146' is defined on line 1536, but no explicit reference was found in the text == Unused Reference: 'RFC2675' is defined on line 1557, but no explicit reference was found in the text == Unused Reference: 'RFC4191' is defined on line 1587, but no explicit reference was found in the text == Unused Reference: 'RFC4987' is defined on line 1608, but no explicit reference was found in the text == Unused Reference: 'RFC5445' is defined on line 1621, but no explicit reference was found in the text ** Obsolete normative reference: RFC 2460 (Obsoleted by RFC 8200) == Outdated reference: A later version (-07) exists of draft-ietf-intarea-ipv4-id-update-05 == Outdated reference: A later version (-12) exists of draft-templin-aero-08 == Outdated reference: A later version (-40) exists of draft-templin-intarea-vet-33 == Outdated reference: A later version (-16) exists of draft-templin-ironbis-10 -- Obsolete informational reference (is this intentional?): RFC 1063 (Obsoleted by RFC 1191) -- Obsolete informational reference (is this intentional?): RFC 1146 (Obsoleted by RFC 6247) -- Obsolete informational reference (is this intentional?): RFC 1981 (Obsoleted by RFC 8201) -- Obsolete informational reference (is this intentional?): RFC 5226 (Obsoleted by RFC 8126) -- Obsolete informational reference (is this intentional?): RFC 5246 (Obsoleted by RFC 8446) Summary: 1 error (**), 0 flaws (~~), 14 warnings (==), 6 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: Informational June 23, 2012 5 Expires: December 25, 2012 7 The Subnetwork Encapsulation and Adaptation Layer (SEAL) 8 draft-templin-intarea-seal-44.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, packet reordering, source address spoofing and 18 traversal of links with diverse Maximum Transmission Units (MTUs). 19 This document specifies a Subnetwork Encapsulation and Adaptation 20 Layer (SEAL) that addresses these issues. 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 December 25, 2012. 39 Copyright Notice 41 Copyright (c) 2012 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 . . . . . . . . . . . . . . . . . 7 60 3. Applicability Statement . . . . . . . . . . . . . . . . . . . 9 61 4. SEAL Specification . . . . . . . . . . . . . . . . . . . . . . 10 62 4.1. VET Interface Model . . . . . . . . . . . . . . . . . . . 10 63 4.2. SEAL Model of Operation . . . . . . . . . . . . . . . . . 11 64 4.3. SEAL Header and Trailer Format . . . . . . . . . . . . . . 12 65 4.4. ITE Specification . . . . . . . . . . . . . . . . . . . . 14 66 4.4.1. IP Protocol Constants . . . . . . . . . . . . . . . . 14 67 4.4.2. Tunnel Interface MTU . . . . . . . . . . . . . . . . . 14 68 4.4.3. Tunnel Neighbor Soft State . . . . . . . . . . . . . . 15 69 4.4.4. Pre-Encapsulation . . . . . . . . . . . . . . . . . . 16 70 4.4.5. SEAL Encapsulation . . . . . . . . . . . . . . . . . . 17 71 4.4.6. Outer Encapsulation . . . . . . . . . . . . . . . . . 18 72 4.4.7. Path Probing and ETE Reachability Verification . . . . 19 73 4.4.8. Processing ICMP Messages . . . . . . . . . . . . . . . 19 74 4.4.9. IPv4 Middlebox Reassembly Testing . . . . . . . . . . 20 75 4.4.10. Stateful MTU Determination . . . . . . . . . . . . . . 21 76 4.4.11. Detecting Path MTU Changes . . . . . . . . . . . . . . 22 77 4.5. ETE Specification . . . . . . . . . . . . . . . . . . . . 22 78 4.5.1. Tunnel Neighbor Soft State . . . . . . . . . . . . . . 22 79 4.5.2. IP-Layer Reassembly . . . . . . . . . . . . . . . . . 22 80 4.5.3. Decapsulation and Re-Encapsulation . . . . . . . . . . 23 81 4.6. The SEAL Control Message Protocol (SCMP) . . . . . . . . . 24 82 4.6.1. Generating SCMP Error Messages . . . . . . . . . . . . 25 83 4.6.2. Processing SCMP Error Messages . . . . . . . . . . . . 27 84 5. Link Requirements . . . . . . . . . . . . . . . . . . . . . . 29 85 6. End System Requirements . . . . . . . . . . . . . . . . . . . 29 86 7. Router Requirements . . . . . . . . . . . . . . . . . . . . . 29 87 8. Nested Encapsulation Considerations . . . . . . . . . . . . . 29 88 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 30 89 10. Security Considerations . . . . . . . . . . . . . . . . . . . 30 90 11. Related Work . . . . . . . . . . . . . . . . . . . . . . . . . 31 91 12. Implementation Status . . . . . . . . . . . . . . . . . . . . 31 92 13. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 32 93 14. References . . . . . . . . . . . . . . . . . . . . . . . . . . 32 94 14.1. Normative References . . . . . . . . . . . . . . . . . . . 32 95 14.2. Informative References . . . . . . . . . . . . . . . . . . 33 96 Appendix A. Reliability . . . . . . . . . . . . . . . . . . . . . 36 97 Appendix B. Integrity . . . . . . . . . . . . . . . . . . . . . . 37 98 Appendix C. Transport Mode . . . . . . . . . . . . . . . . . . . 37 99 Appendix D. Historic Evolution of PMTUD . . . . . . . . . . . . . 38 100 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 39 102 1. Introduction 104 As Internet technology and communication has grown and matured, many 105 techniques have developed that use virtual topologies (including 106 tunnels of one form or another) over an actual network that supports 107 the Internet Protocol (IP) [RFC0791][RFC2460]. Those virtual 108 topologies have elements that appear as one hop in the virtual 109 topology, but are actually multiple IP or sub-IP layer hops. These 110 multiple hops often have quite diverse properties that are often not 111 even visible to the endpoints of the virtual hop. This introduces 112 failure modes that are not dealt with well in current approaches. 114 The use of IP encapsulation (also known as "tunneling") has long been 115 considered as the means for creating such virtual topologies. 116 However, the encapsulation headers often include insufficiently 117 provisioned per-packet identification values. This can present 118 issues for duplicate packet detection and detection of packet 119 reordering within the subnetwork. IP encapsulation also allows an 120 attacker to produce encapsulated packets with spoofed source 121 addresses even if the source address in the encapsulating header 122 cannot be spoofed. A denial-of-service vector that is not possible 123 in non-tunneled subnetworks is therefore presented. 125 Additionally, the insertion of an outer IP header reduces the 126 effective path MTU visible to the inner network layer. When IPv6 is 127 used as the encapsulation protocol, original sources expect to be 128 informed of the MTU limitation through IPv6 Path MTU discovery 129 (PMTUD) [RFC1981]. When IPv4 is used, this reduced MTU can be 130 accommodated through the use of IPv4 fragmentation, but unmitigated 131 in-the-network fragmentation has been found to be harmful through 132 operational experience and studies conducted over the course of many 133 years [FRAG][FOLK][RFC4963]. Additionally, classical IPv4 PMTUD 134 [RFC1191] has known operational issues that are exacerbated by in- 135 the-network tunnels [RFC2923][RFC4459]. 137 The following subsections present further details on the motivation 138 and approach for addressing these issues. 140 1.1. Motivation 142 Before discussing the approach, it is necessary to first understand 143 the problems. In both the Internet and private-use networks today, 144 IP is ubiquitously deployed as the Layer 3 protocol. The primary 145 functions of IP are to provide for routing, addressing, and a 146 fragmentation and reassembly capability used to accommodate links 147 with diverse MTUs. While it is well known that the IP address space 148 is rapidly becoming depleted, there is a lesser-known but growing 149 consensus that other IP protocol limitations have already or may soon 150 become problematic. 152 First, the Internet historically provided no means for discerning 153 whether the source addresses of IP packets are authentic. This 154 shortcoming is being addressed more and more through the deployment 155 of site border router ingress filters [RFC2827], however the use of 156 encapsulation provides a vector for an attacker to circumvent 157 filtering for the encapsulated packet even if filtering is correctly 158 applied to the encapsulation header. Secondly, the IP header does 159 not include a well-behaved identification value unless the source has 160 included a fragment header for IPv6 or unless the source permits 161 fragmentation for IPv4. These limitations preclude an efficient 162 means for routers to detect duplicate packets and packets that have 163 been re-ordered within the subnetwork. 165 For IPv4 encapsulation, when fragmentation is permitted the header 166 includes a 16-bit Identification field, meaning that at most 2^16 167 unique packets with the same (source, destination, protocol)-tuple 168 can be active in the Internet at the same time 169 [I-D.ietf-intarea-ipv4-id-update]. (When middleboxes such as Network 170 Address Translators (NATs) re-write the Identification field to 171 random values, the number of unique packets is even further reduced.) 172 Due to the escalating deployment of high-speed links, however, these 173 numbers have become too small by several orders of magnitude for high 174 data rate packet sources such as tunnel endpoints [RFC4963]. 176 Furthermore, there are many well-known limitations pertaining to IPv4 177 fragmentation and reassembly - even to the point that it has been 178 deemed "harmful" in both classic and modern-day studies (see above). 179 In particular, IPv4 fragmentation raises issues ranging from minor 180 annoyances (e.g., in-the-network router fragmentation [RFC1981]) to 181 the potential for major integrity issues (e.g., mis-association of 182 the fragments of multiple IP packets during reassembly [RFC4963]). 184 As a result of these perceived limitations, a fragmentation-avoiding 185 technique for discovering the MTU of the forward path from a source 186 to a destination node was devised through the deliberations of the 187 Path MTU Discovery Working Group (PMTUDWG) during the late 1980's 188 through early 1990's (see Appendix D). In this method, the source 189 node provides explicit instructions to routers in the path to discard 190 the packet and return an ICMP error message if an MTU restriction is 191 encountered. However, this approach has several serious shortcomings 192 that lead to an overall "brittleness" [RFC2923]. 194 In particular, site border routers in the Internet have been known to 195 discard ICMP error messages coming from the outside world. This is 196 due in large part to the fact that malicious spoofing of error 197 messages in the Internet is trivial since there is no way to 198 authenticate the source of the messages [RFC5927]. Furthermore, when 199 a source node that requires ICMP error message feedback when a packet 200 is dropped due to an MTU restriction does not receive the messages, a 201 path MTU-related black hole occurs. This means that the source will 202 continue to send packets that are too large and never receive an 203 indication from the network that they are being discarded. This 204 behavior has been confirmed through documented studies showing clear 205 evidence of PMTUD failures for both IPv4 and IPv6 in the Internet 206 today [TBIT][WAND][SIGCOMM]. 208 The issues with both IP fragmentation and this "classical" PMTUD 209 method are exacerbated further when IP tunneling is used [RFC4459]. 210 For example, an ingress tunnel endpoint (ITE) may be required to 211 forward encapsulated packets into the subnetwork on behalf of 212 hundreds, thousands, or even more original sources. If the ITE 213 allows IP fragmentation on the encapsulated packets, persistent 214 fragmentation could lead to undetected data corruption due to 215 Identification field wrapping and/or reassembly congestion at the 216 ETE. If the ITE instead uses classical IP PMTUD it must rely on ICMP 217 error messages coming from the subnetwork that may be suspect, 218 subject to loss due to filtering middleboxes, or insufficiently 219 provisioned for translation into error messages to be returned to the 220 original sources. 222 Although recent works have led to the development of a robust end-to- 223 end MTU determination scheme [RFC4821], they do not excuse tunnels 224 from delivering PMTUD feedback when packets are lost due to size 225 restrictions. Moreover, in current practice existing tunneling 226 protocols mask the MTU issues by selecting a "lowest common 227 denominator" MTU that may be much smaller than necessary for most 228 paths and difficult to change at a later date. Therefore, a new 229 approach to accommodate tunnels over links with diverse MTUs is 230 necessary. 232 1.2. Approach 234 For the purpose of this document, a subnetwork is defined as a 235 virtual topology configured over a connected network routing region 236 and bounded by encapsulating border nodes. Example connected network 237 routing regions include Mobile Ad hoc Networks (MANETs), enterprise 238 networks and the global public Internet itself. Subnetwork border 239 nodes forward unicast and multicast packets over the virtual topology 240 across multiple IP and/or sub-IP layer forwarding hops that may 241 introduce packet duplication and/or traverse links with diverse 242 Maximum Transmission Units (MTUs). 244 This document introduces a Subnetwork Encapsulation and Adaptation 245 Layer (SEAL) for tunneling inner network layer protocol packets over 246 IP subnetworks that connect Ingress and Egress Tunnel Endpoints 247 (ITEs/ETEs) of border nodes. It provides a modular specification 248 designed to be tailored to specific associated tunneling protocols. 249 A transport-mode of operation is also possible, and described in 250 Appendix C. 252 SEAL provides a mid-layer encapsulation that accommodates links with 253 diverse MTUs, and allows routers in the subnetwork to perform 254 efficient duplicate packet and packet reordering detection. The 255 encapsulation further ensures data origin authentication, packet 256 header integrity and anti-replay in environments in which these 257 functions are necessary. 259 SEAL treats tunnels that traverse the subnetwork as ordinary links 260 that must support network layer services. Moreover, SEAL provides 261 dynamic mechanisms to ensure a maximal path MTU over the tunnel. 262 This is in contrast to static approaches which avoid MTU issues by 263 selecting a lowest common denominator MTU value that may be overly 264 conservative for the vast majority of tunnel paths and difficult to 265 change even when larger MTUs become available. 267 The following sections provide the SEAL normative specifications, 268 while the appendices present non-normative additional considerations. 270 2. Terminology and Requirements 272 The following terms are defined within the scope of this document: 274 subnetwork 275 a virtual topology configured over a connected network routing 276 region and bounded by encapsulating border nodes. 278 IP 279 used to generically refer to either Internet Protocol (IP) 280 version, i.e., IPv4 or IPv6. 282 Ingress Tunnel Endpoint (ITE) 283 a virtual interface over which an encapsulating border node (host 284 or router) sends encapsulated packets into the subnetwork. 286 Egress Tunnel Endpoint (ETE) 287 a virtual interface over which an encapsulating border node (host 288 or router) receives encapsulated packets from the subnetwork. 290 ETE Link Path 291 a subnetwork path from an ITE to an ETE beginning with an 292 underlying link of the ITE as the first hop. Note that, if the 293 ITE's interface connection to the underlying link assigns multiple 294 IP addresses, each address represents a separate ETE link path. 296 inner packet 297 an unencapsulated network layer protocol packet (e.g., IPv4 298 [RFC0791], OSI/CLNP [RFC0994], IPv6 [RFC2460], etc.) before any 299 outer encapsulations are added. Internet protocol numbers that 300 identify inner packets are found in the IANA Internet Protocol 301 registry [RFC3232]. SEAL protocol packets that incur an 302 additional layer of SEAL encapsulation are also considered inner 303 packets. 305 outer IP packet 306 a packet resulting from adding an outer IP header (and possibly 307 other outer headers) to a SEAL-encapsulated inner packet. 309 packet-in-error 310 the leading portion of an invoking data packet encapsulated in the 311 body of an error control message (e.g., an ICMPv4 [RFC0792] error 312 message, an ICMPv6 [RFC4443] error message, etc.). 314 Packet Too Big (PTB) message 315 a control plane message indicating an MTU restriction (e.g., an 316 ICMPv6 "Packet Too Big" message [RFC4443], an ICMPv4 317 "Fragmentation Needed" message [RFC0792], etc.). 319 Don't Fragment (DF) bit 320 a bit that indicates whether the packet may be fragmented by the 321 network.The DF bit is explicitly included in the IPv4 header 322 [RFC0791] and may be set to '0' to allow fragmentation or '1' to 323 disallow fragmentation. The bit is absent from the IPv6 header 324 [RFC2460], but implicitly set to '1'. 326 The following abbreviations correspond to terms used within this 327 document and/or elsewhere in common Internetworking nomenclature: 329 ETE - Egress Tunnel Endpoint 331 HLEN - the length of the SEAL header plus outer headers 333 ICV - Integrity Check Vector 335 ITE - Ingress Tunnel Endpoint 336 MTU - Maximum Transmission Unit 338 SCMP - the SEAL Control Message Protocol 340 SDU - SCMP Destination Unreachable message 342 SPP - SCMP Parameter Problem message 344 SPTB - SCMP Packet Too Big message 346 SEAL - Subnetwork Encapsulation and Adaptation Layer 348 TE - Tunnel Endpoint (i.e., either ingress or egress) 350 VET - Virtual Enterprise Traversal 352 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 353 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 354 document are to be interpreted as described in [RFC2119]. When used 355 in lower case (e.g., must, must not, etc.), these words MUST NOT be 356 interpreted as described in [RFC2119], but are rather interpreted as 357 they would be in common English. 359 3. Applicability Statement 361 SEAL was originally motivated by the specific case of subnetwork 362 abstraction for Mobile Ad hoc Networks (MANETs), however the domain 363 of applicability also extends to subnetwork abstractions over 364 enterprise networks, ISP networks, SOHO networks, the global public 365 Internet itself, and any other connected network routing region. 366 SEAL, along with the Virtual Enterprise Traversal (VET) 367 [I-D.templin-intarea-vet] tunnel virtual interface abstraction, are 368 the functional building blocks for the Internet Routing Overlay 369 Network (IRON) [I-D.templin-ironbis] and Routing and Addressing in 370 Networks with Global Enterprise Recursion (RANGER) [RFC5720][RFC6139] 371 architectures. 373 SEAL provides a network sublayer for encapsulation of an inner 374 network layer packet within outer encapsulating headers. SEAL can 375 also be used as a sublayer within a transport layer protocol data 376 payload, where transport layer encapsulation is typically used for 377 Network Address Translator (NAT) traversal as well as operation over 378 subnetworks that give preferential treatment to certain "core" 379 Internet protocols (e.g., TCP, UDP, etc.). The SEAL header is 380 processed the same as for IPv6 extension headers, i.e., it is not 381 part of the outer IP header but rather allows for the creation of an 382 arbitrarily extensible chain of headers in the same way that IPv6 383 does. 385 To accommodate MTU diversity, the Egress Tunnel Endpoint (ETE) acts 386 as a passive observer that simply informs the Ingress Tunnel Endpoint 387 (ITE) of any packet size limitations. This allows the ITE to return 388 appropriate PMTUD feedback even if the network path between the ITE 389 and ETE filters ICMP messages. 391 SEAL further provides mechanisms to ensure data origin 392 authentication, packet header integrity, and anti-replay. The SEAL 393 framework is therefore similar to the IP Security (IPsec) 394 Authentication Header (AH) [RFC4301][RFC4302], however it provides 395 only minimal hop-by-hop authenticating services while leaving full 396 data integrity, authentication and confidentiality services as an 397 end-to-end consideration. While SEAL performs data origin 398 authentication, the origin site must also perform the necessary 399 ingress filtering in order to provide full source address 400 verification [I-D.ietf-savi-framework]. 402 In many aspects, SEAL also very closely resembles the Generic Routing 403 Encapsulation (GRE) framework [RFC1701]. SEAL can therefore be 404 applied in the same use cases that are traditionally addressed by 405 GRE, and can also provide additional capabilities as described in 406 this document. 408 4. SEAL Specification 410 The following sections specify the operation of SEAL: 412 4.1. VET Interface Model 414 SEAL is an encapsulation sublayer used within VET non-broadcast, 415 multiple access (NBMA) tunnel virtual interfaces. Each VET interface 416 is configured over one or more underlying interfaces attached to 417 subnetwork links. The VET interface connects an ITE to one or more 418 ETE "neighbors" via tunneling across an underlying subnetwork, where 419 the tunnel neighbor relationship may be either unidirectional or 420 bidirectional. 422 A unidirectional tunnel neighbor relationship allows the near end ITE 423 to send data packets forward to the far end ETE, while the ETE only 424 returns control messages when necessary. A bidirectional tunnel 425 neighbor relationship is one over which both TEs can exchange both 426 data and control messages. 428 Implications of the VET unidirectional and bidirectional models are 429 discussed in [I-D.templin-intarea-vet]. 431 4.2. SEAL Model of Operation 433 SEAL-enabled ITEs encapsulate each inner packet in a SEAL header and 434 any outer header encapsulations as shown in Figure 1: 436 +--------------------+ 437 ~ outer IP header ~ 438 +--------------------+ 439 ~ other outer hdrs ~ 440 +--------------------+ 441 ~ SEAL Header ~ 442 +--------------------+ +--------------------+ 443 | | --> | | 444 ~ Inner ~ --> ~ Inner ~ 445 ~ Packet ~ --> ~ Packet ~ 446 | | --> | | 447 +--------------------+ +----------+---------+ 449 Figure 1: SEAL Encapsulation 451 The ITE inserts the SEAL header according to the specific tunneling 452 protocol. For simple encapsulation of an inner network layer packet 453 within an outer IP header, the ITE inserts the SEAL header following 454 the outer IP header and before the inner packet as: IP/SEAL/{inner 455 packet}. 457 For encapsulations over transports such as UDP, the ITE inserts the 458 SEAL header following the outer transport layer header and before the 459 inner packet, e.g., as IP/UDP/SEAL/{inner packet}. In that case, the 460 UDP header is seen as an "other outer header" as depicted in 461 Figure 1. 463 SEAL supports both "nested" tunneling and "re-encapsulating" 464 tunneling. Nested tunneling occurs when a first tunnel is 465 encapsulated within a second tunnel, which may then further be 466 encapsulated within additional tunnels. Nested tunneling can be 467 useful, and stands in contrast to "recursive" tunneling which is an 468 anomalous condition incurred due to misconfiguration or a routing 469 loop. Considerations for nested tunneling are discussed in Section 4 470 of [RFC2473]. 472 Re-encapsulating tunneling occurs when a packet arrives at a first 473 ETE, which then acts as an ITE to re-encapsulate and forward the 474 packet to a second ETE connected to the same subnetwork. In that 475 case each ITE/ETE transition represents a segment of a bridged path 476 between the ITE nearest the source and the ETE nearest the 477 destination. Combinations of nested and re-encapsulating tunneling 478 are also naturally supported by SEAL. 480 The SEAL ITE considers each underlying interface as the ingress 481 attachment point to a subnetwork link path to the ETE. The ITE 482 therefore may experience different path MTUs on different ETE link 483 paths. 485 Finally, the SEAL ITE ensures that the inner network layer protocol 486 will see a minimum MTU of 1500 bytes over each ETE link path 487 regardless of the outer network layer protocol version, i.e., even if 488 a small amount of fragmentation and reassembly are necessary. This 489 is necessary to avoid path MTU "black holes" for the minimum MTU 490 configured by the vast majority of links in the Internet. 492 4.3. SEAL Header and Trailer Format 494 The SEAL header is formatted as follows: 496 0 1 2 3 497 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 1 498 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 499 |VER|C|A|R|L|I|V|SEG| Reserved | NEXTHDR | LINK_ID |LEVEL| 500 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 501 | Identification (optional) | 502 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 503 | Integrity Check Vector (ICV) (optional) | 504 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 506 Figure 2: SEAL Header Format 508 VER (2) 509 a 2-bit version field. This document specifies Version 0 of the 510 SEAL protocol, i.e., the VER field encodes the value 0. 512 C (1) 513 the "Control/Data" bit. Set to 1 by the ITE in SEAL Control 514 Message Protocol (SCMP) control messages, and set to 0 in ordinary 515 data packets. 517 A (1) 518 the "Acknowledgement Requested" bit. Set to 1 by the ITE in SEAL 519 data packets for which it wishes to receive an explicit 520 acknowledgement from the ETE. 522 R (1) 523 the "Redirects Permitted" bit. For data packets, set to 1 by the 524 ITE to inform the ETE that the source is accepting Redirects (see: 525 [I-D.templin-intarea-vet]). 527 L (1) 528 the "Rate Limit" bit for IPv4 ETE link paths. Reserved for future 529 use for IPv6 ETE link paths. 531 I (1) 532 the "Identification Included" bit. 534 V (1) 535 the "ICV included" bit. 537 SEG (2) a 2-bit segment number bit. Set to 0 if this is the only 538 segment of a single-segment SEAL packet. Set to 1, 2, or 3 if 539 this is the first, second or third segment (respectively) of a 540 multi-segment SEAL packet. 542 Reserved (6) a 6-bit Reserved field. Must be set to 0 for this 543 version of the specification. 545 NEXTHDR (8) an 8-bit field that encodes the next header Internet 546 Protocol number the same as for the IPv4 protocol and IPv6 next 547 header fields. 549 LINK_ID (5) 550 a 5-bit link identification value, set to a unique value by the 551 ITE for each link path over which it will send encapsulated 552 packets to the ETE (up to 32 link paths per ETE are therefore 553 supported). Note that, if the ITE's interface connection to the 554 underlying link assigns multiple IP addresses, each address 555 represents a separate ETE link path that must be assigned a 556 separate LINK_ID. 558 LEVEL (3) 559 a 3-bit nesting level; use to limit the number of tunnel nesting 560 levels. Set to an integer value up to 7 in the innermost SEAL 561 encapsulation, and decremented by 1 for each successive additional 562 SEAL encapsulation nesting level. Up to 8 levels of nesting are 563 therefore supported. 565 Identification (32) 566 an optional 32-bit per-packet identification field; present when 567 I==1. Set to a monotonically-incrementing 32-bit value for each 568 SEAL packet transmitted to this ETE, beginning with 0. 570 Integrity Check Vector (ICV) (32) 571 an optional 32-bit header integrity check value; present when 572 V==1. Covers the leading 128 bytes of the packet beginning with 573 the SEAL header. The value 128 is chosen so that at least the 574 SEAL header as well as the inner packet network and transport 575 layer headers are covered by the integrity check. 577 4.4. ITE Specification 579 4.4.1. IP Protocol Constants 581 SEAL obesrves the IP protocol constants for minimum link MTU (MINMTU) 582 and minimum reassembly buffer (MINMRU) sizes. 584 For IPv6, the ITE sets MINMTU=1280 and MINMRU=1500 (see: [RFC2460]). 586 For IPv4, the ITE sets both MINMTU=576 and MINMRU=576 (see: 587 [RFC1122]) i.e., even though the true MINMTU for IPv4 is only 68 588 bytes (see: [RFC0791]). 590 4.4.2. Tunnel Interface MTU 592 The tunnel interface must present a constant MTU value to the inner 593 network layer as the size for admission of inner packets into the 594 interface. Since VET NBMA tunnel virtual interfaces may support a 595 large set of ETE link paths that accept widely varying maximum packet 596 sizes, however, a number of factors should be taken into 597 consideration when selecting a tunnel interface MTU. 599 Due to the ubiquitous deployment of standard Ethernet and similar 600 networking gear, the nominal Internet cell size has become 1500 601 bytes; this is the de facto size that end systems have come to expect 602 will either be delivered by the network without loss due to an MTU 603 restriction on the path or a suitable ICMP Packet Too Big (PTB) 604 message returned. When large packets sent by end systems incur 605 additional encapsulation at an ITE, however, they may be dropped 606 silently within the tunnel since the network may not always deliver 607 the necessary PTBs [RFC2923]. The ITE should therefore set a tunnel 608 interface MTU of at least 1500 bytes. 610 The inner network layer protocol consults the tunnel interface MTU 611 when admitting a packet into the interface. For non-SEAL inner IPv4 612 packets with the IPv4 Don't Fragment (DF) bit set to 0, if the packet 613 is larger than the tunnel interface MTU the inner IPv4 layer uses 614 IPv4 fragmentation to break the packet into fragments no larger than 615 the tunnel interface MTU. The ITE then admits each fragment into the 616 interface as an independent packet. 618 For all other inner packets, the inner network layer admits the 619 packet if it is no larger than the tunnel interface MTU; otherwise, 620 it drops the packet and sends a PTB error message to the source with 621 the MTU value set to the tunnel interface MTU. The message contains 622 as much of the invoking packet as possible without the entire message 623 exceeding the network layer MINMTU size. 625 The ITE can alternatively set an indefinite MTU on the tunnel 626 interface such that all inner packets are admitted into the interface 627 regardless of their size. For ITEs that host applications that use 628 the tunnel interface directly, this option must be carefully 629 coordinated with protocol stack upper layers since some upper layer 630 protocols (e.g., TCP) derive their packet sizing parameters from the 631 MTU of the outgoing interface and as such may select too large an 632 initial size. This is not a problem for upper layers that use 633 conservative initial maximum segment size estimates and/or when the 634 tunnel interface can reduce the upper layer's maximum segment size, 635 e.g., by reducing the size advertised in the MSS option of outgoing 636 TCP messages (sometimes known as "MSS clamping"). 638 In light of the above considerations, the ITE should configure an 639 indefinite MTU on tunnel *router* interfaces so that subnetwork 640 adaptation is handled from within the interface. The ITE can instead 641 set a smaller MTU on tunnel *host* interfaces (e.g., the smallest MTU 642 among all of the underlying links minus the size of the encapsulation 643 headers) but should not set an MTU smaller than 1500 bytes. 645 4.4.3. Tunnel Neighbor Soft State 647 The tunnel virtual interface maintains a number of soft state 648 variables for each ETE and for each ETE link path. 650 When per-packet identification is required, the ITE maintains a per 651 ETE window of Identification values for the packets it has recently 652 sent to this ETE. The ITE then sets a variable "USE_ID" to TRUE, and 653 includes an Identification in each packet it sends to this ETE; 654 otherwise, it sets USE_ID to FALSE. 656 When data origin authentication and integrity checking is required, 657 the ITE also maintains a per ETE integrity check vector (ICV) 658 calculation algorithm and a symmetric secret key to calculate the ICV 659 in each packet it will send to this ETE. The ITE then sets a 660 variable "USE_ICV" to TRUE, and includes an ICV in each packet it 661 sends to this ETE; otherwise, it sets USE_ICV to FALSE. 663 For IPv4 ETE link paths, the ITE further maintains a variable 664 "RATE_LIMIT" initialized to FALSE. If the link path subsequently 665 exhibits unavoidable IPv4 fragmentation the ETE sets RATE_LIMIT to 666 TRUE. 668 For each ETE link path, the ITE must also account for encapsulation 669 header lengths. The ITE therefore maintains the per ETE link path 670 constant values "SHLEN" set to the length of the SEAL header, "THLEN" 671 set to the length of the outer encapsulating transport layer headers 672 (or 0 if outer transport layer encapsulation is not used), "IHLEN" 673 set to the length of the outer IP layer header, and "HLEN" set to 674 (SHLEN+THLEN+IHLEN). (The ITE must include the length of the 675 uncompressed headers even if header compression is enabled when 676 calculating these lengths.) In addition, the ITE maintains a per ETE 677 link path variable "PATH_MTU" initialized to the maximum of 1500 678 bytes and the MTU of the underlying link minus HLEN. (Thereafter, 679 the ITE must not reduce PATH_MTU to a value smaller than 1500 bytes.) 681 The ITE may instead maintain the packet sizing variables and 682 constants as per ETE (rather than per ETE link path) values. In that 683 case, the values reflect the lowest-common-denominator size across 684 all of the ETE's link paths. 686 4.4.4. Pre-Encapsulation 688 For each inner packet admitted into the tunnel interface, if the 689 packet is itself a SEAL packet (i.e., one with the port number for 690 SEAL in the transport layer header or one with the protocol number 691 for SEAL in the IP layer header) and the LEVEL field of the SEAL 692 header contains the value 0, the ITE silently discards the packet. 694 Otherwise, for non-atomic inner packets (i.e., a non-SEAL packet with 695 (DF==0 || MF!=0 || Offset !=0) in the IP header), if the packet is 696 larger than (MINMTU-HLEN) bytes the ITE fragments the packet into N 697 roughly equal-length pieces, where N is minimized and each fragment 698 is no larger than (MINMTU-HLEN). The ITE then submits each inner 699 fragment for SEAL encapsulation as specified in Section 4.4.5. 701 For atomic inner packets (i.e., a SEAL packet and/or a packet with 702 (DF==1 && MF==0 && Offset==0)), if the packet is larger than (MINMTU- 703 HLEN) but no larger than 1500 bytes the ITE must ensure that it will 704 traverse the tunnel using SEAL segmentation as specified in Section 705 4.4.5. For non-SEAL IPv6 packets, the ITE also sends a PTB error 706 message toward the source address of the inner packet. The ITE 707 writes (MINMTU-HLEN) in the MTU field of the PTB message, and the 708 original source will include an IPv6 fragment header in subsequent 709 packets that it sends. 711 For atomic inner packets larger than 1500 bytes, if the packet is no 712 larger than PATH_MTU for the corresponding ETE link path, the ITE 713 submits it for SEAL encapsulation. Otherwise, the ITE sends a PTB 714 error message toward the source address of the inner packet. To send 715 the PTB message, the ITE first checks its forwarding tables to 716 discover the previous hop toward the source address of the inner 717 packet. If the previous hop is reached via the same tunnel 718 interface, the ITE sends an SCMP PTB (SPTB) message to the previous 719 hop (see: Section 4.6.1.1) with the MTU field set to PATH_MTU. 720 Otherwise, the ITE sends an ordinary PTB message appropriate to the 721 inner protocol version with the MTU field set to PATH_MTU. (For IPv4 722 SEAL packets with DF==0, the ITE should set DF=1 and re-calculate the 723 IPv4 header checksum before generating the PTB message in order to 724 avoid bogon filters.) After sending the (S)PTB message, the ITE 725 discards the inner packet. 727 4.4.5. SEAL Encapsulation 729 For each inner packet/fragment submitted for SEAL encapsulation, the 730 ITE next encapsulates the packet in a SEAL header formatted as 731 specified in Section 4.3. The SEAL header includes an Identification 732 field when USE_ID is TRUE, followed by an ICV field when USE_ICV is 733 TRUE. 735 The ITE next sets C=0 in the SEAL header. The ITE also sets A=1 if 736 necessary for ETE reachability determination (see: Section 4.4.6) or 737 for stateful MTU determination (see Section 4.4.9). Otherwise, the 738 ITE sets A=0. Next, when RATE_LIMIT is TRUE the ITE sets L=1; 739 otherwise, it sets L=0. 741 The ITE then sets R=1 if redirects are permitted (see: 742 [I-D.templin-intarea-vet]). (Note that if this process is entered 743 via re-encapsulation (see: Section 4.5.4), R is instead copied from 744 the SEAL header of the re-encapsulated packet. This implies that the 745 R value is propagated across a re-encapsulating chain of ITE/ETEs.) 747 The ITE then sets LINK_ID to the value assigned to the underlying ETE 748 link path, and sets NEXTHDR to the protocol number corresponding to 749 the address family of the encapsulated inner packet. For example, 750 the ITE sets NEXTHDR to the value '4' for encapsulated IPv4 packets 751 [RFC2003], '41' for encapsulated IPv6 packets [RFC2473][RFC4213], 752 '80' for encapsulated OSI/CLNP packets [RFC1070], etc. 754 Next, if the inner packet is not itself a SEAL packet the ITE sets 755 LEVEL to an integer value between 0 and 7 as a specification of the 756 number of additional layers of nested SEAL encapsulations permitted. 757 If the inner packet is a SEAL packet that is undergoing nested 758 encapsulation, the ITE instead sets LEVEL to the value that appears 759 in the inner packet's SEAL header minus 1. If the inner packet is 760 undergoing SEAL re-encapsulation, the ITE instead copies the LEVEL 761 value from the SEAL header of the packet to be re-encapsulated. 763 Next, if the inner packet is no larger than (MINMTU-HLEN) or larger 764 than 1500, the ITE sets SEG=0. Otherwise, the ITE breaks the packet 765 into N roughly equal-length pieces that are no larger than (MINMTU- 766 HLEN) and appends a clone of the SEAL header from the first piece 767 onto the head of each additional piece. The ITE then sets SEG=1 in 768 the first piece, sets SEG=2 in the second piece, and sets SEG=3 in 769 the third piece (if needed). (Note that since MINMTU for IPv4 is 576 770 bytes this will result in at most 3 pieces.) 772 When USE_ID is FALSE, the ITE next sets I=0. Otherwise, the ITE sets 773 I=1 and writes a monotonically-increasing integer value for this ETE 774 in the Identification field beginning with 0 in the first packet 775 transmitted. (For SEAL packets that have been split into multiple 776 pieces, the ITE writes the same Identification value in each piece.) 778 When USE_ICV is FALSE, the ITE next sets V=0. Otherwise, the ITE 779 sets V=1 and calculates the packet header ICV value using an 780 algorithm agreed on by the ITE and ETE. When data origin 781 authentication is required, the algorithm uses a symmetric secret key 782 so that the ETE can verify that the ICV was generated by the ITE. 783 Beginning with the SEAL header, the ITE calculates the ICV over the 784 leading 128 bytes of the packet (or up to the end of the packet if 785 there are fewer than 128 bytes) and places result in the ICV field. 786 (For SEAL packets that have been split into two pieces, each piece 787 caclulates its own ICV value.) 789 The ITE then adds the outer encapsulating headers as specified in 790 Section 4.4.6. 792 4.4.6. Outer Encapsulation 794 Following SEAL encapsulation, the ITE next encapsulates each packet 795 in the requisite outer transport (when necessary) and IP layer 796 headers. When a transport layer header is included, the ITE writes 797 the port number for SEAL in the transport destination service port 798 field and writes the protocol number of the transport protocol in the 799 outer IP header protocol field. Otherwise, the ITE writes the 800 protocol number for SEAL in the outer IP header protocol field. 802 The ITE then sets the other fields of the outer transport and IP 803 layer headers as specified in Sections 5.5.4 and 5.5.5 804 of[I-D.templin-intarea-vet]. If this process is entered via re- 805 encapsulation (see: Section 4.5.4), the ITE instead follows the re- 806 encapsulation procedures specified in Section 5.5.6 of 807 [I-D.templin-intarea-vet]. 809 For IPv4 ETE link paths, the ITE sets DF=0 in the IPv4 header to 810 allow the packet to be fragmented if it encounters a restricting 811 link. (For IPv6 link paths, the DF bit is implicitly set to 1.) 813 The ITE then sends each outer packet/fragment via the underlying link 814 corresponding to LINK_ID. For IPv4 ETE link paths with 815 RATE_LIMIT=TRUE, the ITE sends the packet subject to rate limiting so 816 that the IPv4 Identification value is not repeated within the IPv4 817 Maximum Segment Lifetime (i.e., 120 seconds) [RFC1122]. 819 4.4.7. Path Probing and ETE Reachability Verification 821 All SEAL data packets sent by the ITE are considered implicit probes. 822 SEAL data packets will elicit an SCMP message from the ETE if it 823 needs to acknowledge a probe and/or report an error condition. SEAL 824 data packets may also be dropped by either the ETE or a router on the 825 path, which will return an ICMP message. 827 The ITE can also send an SCMP Router/Neighbor Solicitation message to 828 elicit an SCMP Router/Neighbor Advertisement response (see: 829 [I-D.templin-intarea-vet]) as verification that the ETE is still 830 reachable via a specific link path. 832 The ITE processes ICMP messages as specified in Section 4.4.7. 834 The ITE processes SCMP messages as specified in Section 4.6.2. 836 4.4.8. Processing ICMP Messages 838 When the ITE sends SEAL packets, it may receive ICMP error messages 839 [RFC0792][RFC4443] from an ordinary router within the subnetwork or 840 from another ITE on the path to the ETE (i.e., in case of nested 841 encapsulations). Each ICMP message includes an outer IP header, 842 followed by an ICMP header, followed by a portion of the SEAL data 843 packet that generated the error (also known as the "packet-in-error") 844 beginning with the outer IP header. 846 The ITE should process ICMPv4 Protocol Unreachable messages and 847 ICMPv6 Parameter Problem messages with Code "Unrecognized Next Header 848 type encountered" as a hint that the ETE does not implement the SEAL 849 protocol. The ITE can also process other ICMP messages that do not 850 include sufficient information in the packet-in-error as a hint that 851 the ETE link path may be failing. Specific actions that the ITE may 852 take in these cases are out of scope. 854 For other ICMP messages, the ITE should use any outer header 855 information available as a first-pass authentication filter (e.g., to 856 determine if the source of the message is within the same 857 administrative domain as the ITE) and discards the message if first 858 pass filtering fails. 860 Next, the ITE examines the packet-in-error beginning with the SEAL 861 header. If the value in the Identification field (if present) is not 862 within the window of packets the ITE has recently sent to this ETE, 863 or if the value in the SEAL header ICV field (if present) is 864 incorrect, the ITE discards the message. 866 Next, if the received ICMP message is a PTB the ITE sets the 867 temporary variable "PMTU" for this ETE link path to the MTU value in 868 the PTB message. If PMTU==0, the ITE consults a plateau table (e.g., 869 as described in [RFC1191]) to determine PMTU based on the length 870 field in the outer IP header of the packet-in-error. For example, if 871 the ITE receives a PTB message with MTU==0 and length 4KB, it can set 872 PMTU=2KB. If the ITE subsequently receives a PTB message with MTU==0 873 and length 2KB, it can set PMTU=1792, etc. to a minimum value of 874 PMTU=(1500+HLEN). If the ITE is performing stateful MTU 875 determination for this ETE link path (see Section 4.4.9), the ITE 876 next sets PATH_MTU=MAX((PMTU-HLEN), 1500). 878 If the ICMP message was not discarded, the ITE then transcribes it 879 into a message to return to the previous hop. If the previous hop 880 toward the inner source address within the packet-in-error is reached 881 via the same tunnel interface the SEAL data packet was sent on, the 882 ITE transcribes the ICMP message into an SCMP message. Otherwise, 883 the ITE transcribes the ICMP message into a message appropriate for 884 the inner protocol version. 886 To transcribe the message, the ITE extracts the inner packet from 887 within the ICMP message packet-in-error field and uses it to generate 888 a new message corresponding to the type of the received ICMP message. 889 For SCMP messages, the ITE generates the message the same as 890 described for ETE generation of SCMP messages in Section 4.6.1. For 891 (S)PTB messages, the ITE writes (PMTU-HLEN) in the MTU field. 893 The ITE finally forwards the transcribed message to the previous hop 894 toward the inner source address. 896 4.4.9. IPv4 Middlebox Reassembly Testing 898 The ITE can perform a qualification exchange to ensure that the 899 subnetwork correctly delivers fragments to the ETE. This procedure 900 can be used, e.g., to determine whether there are middleboxes on the 901 path that violate the [RFC1812], Section 5.2.6 requirement that: "A 902 router MUST NOT reassemble any datagram before forwarding it". 904 The ITE should use knowledge of its topological arrangement as an aid 905 in determining when middlebox reassembly testing is necessary. For 906 example, if the ITE is aware that the ETE is located somewhere in the 907 public Internet, middlebox reassembly testing should not be 908 necessary. If the ITE is aware that the ETE is located behind a NAT 909 or a firewall, however, then middlebox reassembly testing is 910 recommended. 912 The ITE can perform a middlebox reassembly test by selecting a data 913 packet to be used as a probe. While performing the test with real 914 data packets, the ITE should select only inner packets that are no 915 larger than (1500-HLEN) bytes for testing purposes. The ITE can also 916 construct a dummy probe packet instead of using ordinary SEAL data 917 packets. 919 To generate a dummy probe packet, the ITE creates a packet buffer 920 beginning with the same outer headers, SEAL header and inner network 921 layer header that would appear in an ordinary data packet, then pads 922 the packet with random data to a length that is at least 128 bytes 923 but no longer than (1500-HLEN) bytes. The ITE then writes the value 924 '0' in the inner network layer TTL (for IPv4) or Hop Limit (for IPv6) 925 field. 927 The ITE then sets (C=0; R=0) in the SEAL header of the probe packet 928 and sets the NEXTHDR field to the inner network layer protocol type. 929 (The ITE may also set A=1 if it requires a positive acknowledgement; 930 otherwise, it sets A=0.) Next, the ITE sets LINK_ID and LEVEL to the 931 appropriate values for this ETE link path, sets Identification and 932 I=1 (when USE_ID is TRUE), then finally calculates the ICV and sets 933 V=1(when USE_ICV is TRUE). 935 The ITE then encapsulates the probe packet in the appropriate outer 936 headers, splits it into two outer IPv4 fragments, then sends both 937 fragments over the same ETE link path. 939 The ITE should send a series of probe packets (e.g., 3-5 probes with 940 1sec intervals between tests) instead of a single isolated probe in 941 case of packet loss. If the ETE returns an SCMP PTB message with MTU 942 != 0, then the ETE link path correctly supports fragmentation; 943 otherwise, the ITE enables stateful MTU determination for this ETE 944 link path as specified in Section 4.4.9. 946 (Examples of middleboxes that may perform reassembly include stateful 947 NATs and firewalls. Such devices could still allow for stateless MTU 948 determination if they gather the fragments of a fragmented IPv4 SEAL 949 data packet for packet analysis purposes but then forward the 950 fragments on to the final destination rather than forwarding the 951 reassembled packet.) 953 4.4.10. Stateful MTU Determination 955 SEAL supports a stateless MTU determination capability, however the 956 ITE may in some instances wish to impose a stateful MTU limit on a 957 particular ETE link path. For example, when the ETE is situated 958 behind a middlebox that performs IPv4 reassembly (see: Section 4.4.8) 959 it is imperative that fragmentation be avoided. In other instances 960 (e.g., when the ETE link path includes performance-constrained 961 links), the ITE may deem it necessary to cache a conservative static 962 MTU in order to avoid sending large packets that would only be 963 dropped due to an MTU restriction somewhere on the path. 965 To determine a static MTU value, the ITE can send a series of dummy 966 probe packets of various sizes to the ETE with A=1 in the SEAL header 967 and DF=1 in the outer IP header. The ITE can then cache the size 'S' 968 of the largest packet for which it receives a probe reply from the 969 ETE by setting PATH_MTU=MAX((S-HLEN), 1500) for this ETE link path. 971 For example, the ITE could send probe packets of 4KB, followed by 972 2KB, followed by 1792 bytes, etc. While probing, the ITE processes 973 any ICMP PTB message it receives as a potential indication of probe 974 failure then discards the message. 976 4.4.11. Detecting Path MTU Changes 978 When stateful determination is used, the ITE can periodically reset 979 PATH_MTU and/or re-probe the path to determine whether PATH_MTU has 980 increased. If the path still has a too-small MTU, the ITE will 981 receive a PTB message that reports a smaller size. 983 For IPv4 ETE link paths, when the path correctly implements 984 fragmentation and RATE_LIMIT is TRUE, the ITE can periodically reset 985 RATE_LIMIT=FALSE to determine whether the path still requires rate 986 limiting. If the ITE receives an SPTB message it should again set 987 RATE_LIMIT=TRUE. 989 4.5. ETE Specification 991 4.5.1. Tunnel Neighbor Soft State 993 When data origin authentication and integrity checking is required, 994 the ETE maintains a per-ITE ICV calculation algorithm and a symmetric 995 secret key to verify the ICV. When per-packet identification is 996 required, the ETE also maintains a window of Identification values 997 for the packets it has recently received from this ITE. 999 When the tunnel neighbor relationship is bidirectional, the ETE 1000 further maintains a per ETE link path mapping of outer IP and 1001 transport layer addresses to the LINK_ID that appears in packets 1002 received from the ITE. 1004 4.5.2. IP-Layer Reassembly 1006 The ETE should maintain conservative reassembly cache high- and low- 1007 water marks. When the size of the reassembly cache exceeds this 1008 high-water mark, the ETE should actively discard stale incomplete 1009 reassemblies (e.g., using an Active Queue Management (AQM) strategy) 1010 until the size falls below the low-water mark. The ETE should also 1011 actively discard any pending reassemblies that clearly have no 1012 opportunity for completion, e.g., when a considerable number of new 1013 fragments have arrived before a fragment that completes a pending 1014 reassembly arrives. 1016 The ETE processes non-SEAL IP packets as specified in the normative 1017 references, i.e., it performs any necessary IP reassembly then 1018 discards the packet if it is larger than the reassembly buffer size 1019 or delivers the (fully-reassembled) packet to the appropriate upper 1020 layer protocol module. 1022 For SEAL packets, the ITE performs any necessary IP reassembly then 1023 submits the packet for SEAL decapsulation as specified in Section 1024 4.5.3. (Note that if the packet is larger than the reassembly buffer 1025 size, the ITE still returns the leading portion of the (partially) 1026 reassembled packet.) 1028 4.5.3. Decapsulation and Re-Encapsulation 1030 For each SEAL packet accepted for decapsulation, when I==1 the ETE 1031 first examines the Identification field. If the Identification is 1032 not within the window of acceptable values for this ITE, the ETE 1033 silently discards the packet. 1035 Next, if V==1 the ETE verifies the ICV value (with the ICV field 1036 itself reset to 0) and silently discards the packet if the value is 1037 incorrect. 1039 Next, if the packet arrived as multiple IPv4 fragments and L ==0, the 1040 ETE sends an SPTB message back to the ITE with MTU set to the size of 1041 the largest fragment received minus HLEN (see: Section 4.6.1.1). 1043 Next, if the packet arrived as multiple IP fragments and the inner 1044 packet is larger than 1500 bytes, the ETE silently discards the 1045 packet; otherwise, it continues to process the packet. 1047 Next, if there is an incorrect value in a SEAL header field (e.g., an 1048 incorrect "VER" field value), the ETE discards the packet. If the 1049 SEAL header has C==0, the ETE also returns an SCMP "Parameter 1050 Problem" (SPP) message (see Section 4.6.1.2). 1052 Next, if the SEAL header has C==1, the ETE processes the packet as an 1053 SCMP packet as specified in Section 4.6.2. Otherwise, the ETE 1054 continues to process the packet as a SEAL data packet. 1056 Next, if the SEAL header has SEG != 0 the ETE checks to see if the 1057 other segments of this already-segmented SEAL packet have arrived, 1058 i.e., by looking for additional segments that have the same outer IP 1059 source address, source transport port number (if present) and SEAL 1060 Identification value. If the other segments have already arrived, 1061 the ETE discards the SEAL header and other outer headers from the 1062 other segments and appends the non-initial segments onto the end of 1063 the first segment. Otherwise, the ETE caches the segment for at most 1064 60 seconds while awaiting the arrival of its partners. To support 1065 this process, the ETE must be able to buffer segmented SEAL packets 1066 up to (1500+HLEN) bytes in length. 1068 Next, if the SEAL header in the (reassembled) packet has A==1, the 1069 ETE sends an SPTB message back to the ITE with MTU=0 (see: Section 1070 4.6.1.1). 1072 Finally, the ETE discards the outer headers and processes the inner 1073 packet according to the header type indicated in the SEAL NEXTHDR 1074 field. If the inner (TTL / Hop Limit) field encodes the value 0, the 1075 ETE silently discards the packet. Otherwise, if the next hop toward 1076 the inner destination address is via a different interface than the 1077 SEAL packet arrived on, the ETE discards the SEAL header and delivers 1078 the inner packet either to the local host or to the next hop 1079 interface if the packet is not destined to 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 and without 1084 decrementing the value in the inner (TTL / Hop Limit) field. In this 1085 process, the packet remains within the tunnel (i.e., it does not exit 1086 and then re-enter the tunnel); hence, the packet is not discarded if 1087 the LEVEL field in the SEAL header contains the 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 32-bit 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 3: 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 3: 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 4: 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 invoking SEAL data packet as possible | 1133 ~ (beginning with the SEAL header) without the SCMP ~ 1134 | packet exceeding MINMTU bytes (*) | 1136 (*) also known as the "packet-in-error" 1138 Figure 4: SCMP Error Message Format 1140 The error message includes the 32-bit SCMP message header, followed 1141 by a 32-bit Type-Specific Data field, followed by the leading portion 1142 of the invoking SEAL data packet beginning with the SEAL header as 1143 the "packet-in-error". The packet-in-error includes as much of the 1144 invoking packet as possible extending to a length that would not 1145 cause the entire SCMP packet following outer encapsulation to exceed 1146 MINMTU bytes. 1148 When the ETE processes a SEAL data packet for which the 1149 Identification and ICV values are correct but an error must be 1150 returned, it prepares an SCMP error message as shown in Figure 4. 1151 The ETE sets the Type and Code fields to the same values that would 1152 appear in the corresponding ICMPv6 message [RFC4443], but calculates 1153 the Checksum beginning with the SCMP message header using the 1154 algorithm specified for ICMPv4 in [RFC0792]. 1156 The ETE next encapsulates the SCMP message in the requisite SEAL and 1157 outer headers as shown in Figure 3. During encapsulation, the ETE 1158 sets the outer destination address/port numbers of the SCMP packet to 1159 the values associated with the ITE and sets the outer source address/ 1160 port numbers to its own outer address/port numbers. 1162 The ETE then sets (C=1; A=0; R=0; L=0; SEG=0) in the SEAL header, 1163 then sets I, V, NEXTHDR and LEVEL to the same values that appeared in 1164 the SEAL header of the data packet. If the neighbor relationship 1165 between the ITE and ETE is unidirectional, the ETE next sets the 1166 LINK_ID field to the same value that appeared in the SEAL header of 1167 the data packet. Otherwise, the ETE sets the LINK_ID field to the 1168 value it would use in sending a SEAL packet to this ITE. 1170 When I==1, the ETE next sets the Identification field to an 1171 appropriate value for the ITE. If the neighbor relationship between 1172 the ITE and ETE is unidirectional, the ETE sets the Identification 1173 field to the same value that appeared in the SEAL header of the data 1174 packet. Otherwise, the ETE sets the Identification field to the 1175 value it would use in sending the next SEAL packet to this ITE. 1177 When V==1, the ETE then calculates and sets the ICV field the same as 1178 specified for SEAL data packet encapsulation in Section 4.4.4. 1180 Finally, the ETE sends the resulting SCMP packet to the ITE the same 1181 as specified for SEAL data packets in Section 4.4.5. 1183 The following sections describe additional considerations for various 1184 SCMP error messages: 1186 4.6.1.1. Generating SCMP Packet Too Big (SPTB) Messages 1188 An ETE generates an SCMP "Packet Too Big" (SPTB) message when it 1189 receives a SEAL data packet that arrived as multiple outer IPv4 1190 fragments and for which L==0. The ETE prepares the SPTB message the 1191 same as for the corresponding ICMPv6 PTB message, and writes the 1192 length of the largest outer IP fragment received minus HLEN in the 1193 MTU field of the message. 1195 The ETE also generates an SPTB message when it accepts a SEAL 1196 protocol data packet with A==1 in the SEAL header. The ETE prepares 1197 the SPTB message the same as above, except that it writes the value 0 1198 in the MTU field. 1200 4.6.1.2. Generating Other SCMP Error Messages 1202 An ETE generates an SCMP "Destination Unreachable" (SDU) message 1203 under the same circumstances that an IPv6 system would generate an 1204 ICMPv6 Destination Unreachable message. 1206 An ETE generates an SCMP "Parameter Problem" (SPP) message when it 1207 receives a SEAL packet with an incorrect value in the SEAL 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 with C==1 in the SEAL header after 1217 sending packets to an ETE. The ITE first verifies that the outer 1218 addresses of the SCMP packet are correct, and (when I==1) that the 1219 Identification field contains an acceptable value. The ITE next 1220 verifies that the SEAL header fields are set correctly as specified 1221 in Section 4.6.1. When V==1, the ITE then verifies the ICV value. 1222 The ITE next verifies the Checksum value in the SCMP message header. 1223 If any of these values are incorrect, the ITE silently discards the 1224 message; otherwise, it processes the message as follows: 1226 4.6.2.1. Processing SCMP PTB Messages 1228 After an ITE sends a SEAL data packet to an ETE, it may receive an 1229 SPTB message with a packet-in-error containing the leading portion of 1230 the packet (see: Section 4.6.1.1). For IP SPTB messages with MTU==0, 1231 the ITE processes the message as confirmation that the ETE received a 1232 SEAL data packet with A==1 in the SEAL header. The ITE then discards 1233 the message. 1235 For IPv4 SPTB messages with MTU != 0, the ITE instead processes the 1236 message as an indication of a packet size limitation as follows. If 1237 the SEAL header of the packet-in-error has (SEG==0 || SEG ==1), and 1238 the inner packet length is less than 1500, the ITE sets 1239 RATE_LIMIT=TRUE. Otherwise, the ITE instead examines the SPTB 1240 message MTU field. If the MTU value is not substantially less than 1241 1280 bytes, the value is likely to reflect the true MTU of the 1242 restricting link on the path to the ETE; otherwise, a router on the 1243 path may be generating runt fragments. 1245 In that case, the ITE can consult a plateau table (e.g., as described 1246 in [RFC1191]) to rewrite the MTU value to a reduced size. For 1247 example, if the ITE receives an IPv4 SPTB message with MTU==256 and 1248 inner packet length 4KB, it can rewrite the MTU to 2KB. If the ITE 1249 subsequently receives an IPv4 SPTB message with MTU==256 and inner 1250 packet length 2KB, it can rewrite the MTU to 1792, etc If the ITE is 1251 performing stateful MTU determination for this ETE link path, it then 1252 writes the new MTU value minus HLEN in PATH_MTU. 1254 The ITE then checks its forwarding tables to discover the previous 1255 hop toward the source address of the inner packet. If the previous 1256 hop is reached via the same tunnel interface the SPTB message arrived 1257 on, the ITE relays the message to the previous hop. In order to 1258 relay the message, the first writes zero in the Identification and 1259 ICV fields of the SEAL header within the packet-in-error. The ITE 1260 next rewrites the outer SEAL header fields with values corresponding 1261 to the previous hop and recalculates the ICV using the ICV 1262 calculation parameters associated with the previous hop. Next, the 1263 ITE replaces the SPTB's outer headers with headers of the appropriate 1264 protocol version and fills in the header fields as specified in 1265 Sections 5.5.4-5.5.6 of [I-D.templin-intarea-vet], where the 1266 destination address/port correspond to the previous hop and the 1267 source address/port correspond to the ITE. The ITE then sends the 1268 message to the previous hop the same as if it were issuing a new SPTB 1269 message. (Note that, in this process, the values within the SEAL 1270 header of the packet-in-error are meaningless to the previous hop and 1271 therefore cannot be used by the previous hop for authentication 1272 purposes.) 1274 If the previous hop is not reached via the same tunnel interface, the 1275 ITE instead transcribes the message into a format appropriate for the 1276 inner packet (i.e., the same as described for transcribing ICMP 1277 messages in Section 4.4.7) and sends the resulting transcribed 1278 message to the original source. (NB: if the inner packet within the 1279 SPTB message is an IPv4 SEAL packet with DF==0, the ITE should set 1280 DF=1 and re-calculate the IPv4 header checksum while transcribing the 1281 message in order to avoid bogon filters.) The ITE then discards the 1282 SPTB message. 1284 4.6.2.2. Processing Other SCMP Error Messages 1286 An ITE may receive an SDU message with an appropriate code under the 1287 same circumstances that an IPv6 node would receive an ICMPv6 1288 Destination Unreachable message. The ITE either transcribes or 1289 relays the message toward the source address of the inner packet 1290 within the packet-in-error the same as specified for SPTB messages in 1291 Section 4.6.2.1. 1293 An ITE may receive an SPP message when the ETE receives a SEAL packet 1294 with an incorrect value in the SEAL header. The ITE should examine 1295 the SEAL header within the packet-in-error to determine whether a 1296 different setting should be used in subsequent packets, but does not 1297 relay the message further. 1299 TEs process other SCMP message types using methods and procedures 1300 specified in other documents. For example, SCMP message types used 1301 for tunnel neighbor coordinations are specified in VET 1302 [I-D.templin-intarea-vet]. 1304 5. Link Requirements 1306 Subnetwork designers are expected to follow the recommendations in 1307 Section 2 of [RFC3819] when configuring link MTUs. 1309 6. End System Requirements 1311 End systems are encouraged to implement end-to-end MTU assurance 1312 (e.g., using Packetization Layer PMTUD per [RFC4821]) even if the 1313 subnetwork is using SEAL. 1315 7. Router Requirements 1317 Routers within the subnetwork are expected to observe the router 1318 requirements found in the normative references, including the 1319 implementation of IP fragmentation and reassembly [RFC1812][RFC2460] 1320 as well as the generation of ICMP messages [RFC0792][RFC4443]. 1322 8. Nested Encapsulation Considerations 1324 SEAL supports nested tunneling for up to 8 layers of encapsulation. 1325 In this model, the SEAL ITE has a tunnel neighbor relationship only 1326 with ETEs at its own nesting level, i.e., it does not have a tunnel 1327 neighbor relationship with other ITEs, nor with ETEs at other nesting 1328 levels. 1330 Therefore, when an ITE 'A' within an inner nesting level needs to 1331 return an error message to an ITE 'B' within an outer nesting level, 1332 it generates an ordinary ICMP error message the same as if it were an 1333 ordinary router within the subnetwork. 'B' can then perform message 1334 validation as specified in Section 4.4.7, but full message origin 1335 authentication is not possible. 1337 Since ordinary ICMP messages are used for coordinations between ITEs 1338 at different nesting levels, nested SEAL encapsulations should only 1339 be used when the ITEs are within a common administrative domain 1340 and/or when there is no ICMP filtering middlebox such as a firewall 1341 or NAT between them. An example would be a recursive nesting of 1342 mobile networks, where the first network receives service from an 1343 ISP, the second network receives service from the first network, the 1344 third network receives service from the second network, etc. 1346 NB: As an alternative, the SCMP protocol could be extended to allow 1347 ITE 'A' to return an SCMP message to ITE 'B' rather than return an 1348 ICMP message. This would conceptually allow the control messages to 1349 pass through firewalls and NATs, however it would give no more 1350 message origin authentication assurance than for ordinary ICMP 1351 messages. It was therefore determined that the complexity of 1352 extending the SCMP protocol was of little value within the context of 1353 the anticipated use cases for nested encapsulations. 1355 9. IANA Considerations 1357 The IANA is instructed to allocate a System Port number for "SEAL" in 1358 the 'port-numbers' registry for the TCP, UDP, DCCP and SCTP 1359 protocols. 1361 The IANA is further instructed to allocate an IP protocol number for 1362 "SEAL" in the "protocol-numbers" registry. 1364 Considerations for port and protocol number assignments appear in 1365 [RFC2780][RFC5226][RFC6335]. 1367 10. Security Considerations 1369 SEAL provides a segment-by-segment data origin authentication and 1370 anti-replay service across the (potentially) multiple segments of a 1371 re-encapsulating tunnel. It further provides a segment-by-segment 1372 integrity check of the headers of encapsulated packets, but does not 1373 verify the integrity of the rest of the packet beyond the headers 1374 unless fragmentation is unavoidable. SEAL therefore considers full 1375 message integrity checking, authentication and confidentiality as 1376 end-to-end considerations in a manner that is compatible with 1377 securing mechanisms such as TLS/SSL [RFC5246]. 1379 An amplification/reflection/buffer overflow attack is possible when 1380 an attacker sends IP fragments with spoofed source addresses to an 1381 ETE in an attempt to clog the ETE's reassembly buffer and/or cause 1382 the ETE to generate a stream of SCMP messages returned to a victim 1383 ITE. The SCMP message ICV, Identification, as well as the inner 1384 headers of the packet-in-error, provide mitigation for the ETE to 1385 detect and discard SEAL segments with spoofed source addresses. 1387 The SEAL header is sent in-the-clear the same as for the outer IP and 1388 other outer headers. In this respect, the threat model is no 1389 different than for IPv6 extension headers. Unlike IPv6 extension 1390 headers, however, the SEAL header can be protected by an integrity 1391 check that also covers the inner packet headers. 1393 Security issues that apply to tunneling in general are discussed in 1394 [RFC6169]. 1396 11. Related Work 1398 Section 3.1.7 of [RFC2764] provides a high-level sketch for 1399 supporting large tunnel MTUs via a tunnel-level segmentation and 1400 reassembly capability to avoid IP level fragmentation. 1402 Section 3 of [RFC4459] describes inner and outer fragmentation at the 1403 tunnel endpoints as alternatives for accommodating the tunnel MTU. 1405 Section 4 of [RFC2460] specifies a method for inserting and 1406 processing extension headers between the base IPv6 header and 1407 transport layer protocol data. The SEAL header is inserted and 1408 processed in exactly the same manner. 1410 IPsec/AH is [RFC4301][RFC4301] is used for full message integrity 1411 verification between tunnel endpoints, whereas SEAL only ensures 1412 integrity for the inner packet headers. The AYIYA proposal 1413 [I-D.massar-v6ops-ayiya] uses similar means for providing message 1414 authentication and integrity. 1416 The concepts of path MTU determination through the report of 1417 fragmentation and extending the IPv4 Identification field were first 1418 proposed in deliberations of the TCP-IP mailing list and the Path MTU 1419 Discovery Working Group (MTUDWG) during the late 1980's and early 1420 1990's. An historical analysis of the evolution of these concepts, 1421 as well as the development of the eventual PMTUD mechanism, appears 1422 in Appendix D of this document. 1424 12. Implementation Status 1426 An early implementation of the first revision of SEAL [RFC5320] is 1427 available at: http://isatap.com/seal/pre-rfc5320.txt 1429 13. Acknowledgments 1431 The following individuals are acknowledged for helpful comments and 1432 suggestions: Jari Arkko, Fred Baker, Iljitsch van Beijnum, Oliver 1433 Bonaventure, Teco Boot, Bob Braden, Brian Carpenter, Steve Casner, 1434 Ian Chakeres, Noel Chiappa, Remi Denis-Courmont, Remi Despres, Ralph 1435 Droms, Aurnaud Ebalard, Gorry Fairhurst, Washam Fan, Dino Farinacci, 1436 Joel Halpern, Sam Hartman, John Heffner, Thomas Henderson, Bob 1437 Hinden, Christian Huitema, Eliot Lear, Darrel Lewis, Joe Macker, Matt 1438 Mathis, Erik Nordmark, Dan Romascanu, Dave Thaler, Joe Touch, Mark 1439 Townsley, Ole Troan, Margaret Wasserman, Magnus Westerlund, Robin 1440 Whittle, James Woodyatt, and members of the Boeing Research & 1441 Technology NST DC&NT group. 1443 Discussions with colleagues following the publication of [RFC5320] 1444 have provided useful insights that have resulted in significant 1445 improvements to this, the Second Edition of SEAL. 1447 Path MTU determination through the report of fragmentation was first 1448 proposed by Charles Lynn on the TCP-IP mailing list in 1987. 1449 Extending the IP identification field was first proposed by Steve 1450 Deering on the MTUDWG mailing list in 1989. 1452 14. References 1454 14.1. Normative References 1456 [RFC0791] Postel, J., "Internet Protocol", STD 5, RFC 791, 1457 September 1981. 1459 [RFC0792] Postel, J., "Internet Control Message Protocol", STD 5, 1460 RFC 792, September 1981. 1462 [RFC1122] Braden, R., "Requirements for Internet Hosts - 1463 Communication Layers", STD 3, RFC 1122, October 1989. 1465 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1466 Requirement Levels", BCP 14, RFC 2119, March 1997. 1468 [RFC2460] Deering, S. and R. Hinden, "Internet Protocol, Version 6 1469 (IPv6) Specification", RFC 2460, December 1998. 1471 [RFC3971] Arkko, J., Kempf, J., Zill, B., and P. Nikander, "SEcure 1472 Neighbor Discovery (SEND)", RFC 3971, March 2005. 1474 [RFC4443] Conta, A., Deering, S., and M. Gupta, "Internet Control 1475 Message Protocol (ICMPv6) for the Internet Protocol 1476 Version 6 (IPv6) Specification", RFC 4443, March 2006. 1478 [RFC4861] Narten, T., Nordmark, E., Simpson, W., and H. Soliman, 1479 "Neighbor Discovery for IP version 6 (IPv6)", RFC 4861, 1480 September 2007. 1482 14.2. Informative References 1484 [FOLK] Shannon, C., Moore, D., and k. claffy, "Beyond Folklore: 1485 Observations on Fragmented Traffic", December 2002. 1487 [FRAG] Kent, C. and J. Mogul, "Fragmentation Considered Harmful", 1488 October 1987. 1490 [I-D.ietf-intarea-ipv4-id-update] 1491 Touch, J., "Updated Specification of the IPv4 ID Field", 1492 draft-ietf-intarea-ipv4-id-update-05 (work in progress), 1493 May 2012. 1495 [I-D.ietf-savi-framework] 1496 Wu, J., Bi, J., Bagnulo, M., Baker, F., and C. Vogt, 1497 "Source Address Validation Improvement Framework", 1498 draft-ietf-savi-framework-06 (work in progress), 1499 January 2012. 1501 [I-D.massar-v6ops-ayiya] 1502 Massar, J., "AYIYA: Anything In Anything", 1503 draft-massar-v6ops-ayiya-02 (work in progress), July 2004. 1505 [I-D.templin-aero] 1506 Templin, F., "Asymmetric Extended Route Optimization 1507 (AERO)", draft-templin-aero-08 (work in progress), 1508 February 2012. 1510 [I-D.templin-intarea-vet] 1511 Templin, F., "Virtual Enterprise Traversal (VET)", 1512 draft-templin-intarea-vet-33 (work in progress), 1513 December 2011. 1515 [I-D.templin-ironbis] 1516 Templin, F., "The Internet Routing Overlay Network 1517 (IRON)", draft-templin-ironbis-10 (work in progress), 1518 December 2011. 1520 [MTUDWG] "IETF MTU Discovery Working Group mailing list, 1521 gatekeeper.dec.com/pub/DEC/WRL/mogul/mtudwg-log, November 1522 1989 - February 1995.". 1524 [RFC0994] International Organization for Standardization (ISO) and 1525 American National Standards Institute (ANSI), "Final text 1526 of DIS 8473, Protocol for Providing the Connectionless- 1527 mode Network Service", RFC 994, March 1986. 1529 [RFC1063] Mogul, J., Kent, C., Partridge, C., and K. McCloghrie, "IP 1530 MTU discovery options", RFC 1063, July 1988. 1532 [RFC1070] Hagens, R., Hall, N., and M. Rose, "Use of the Internet as 1533 a subnetwork for experimentation with the OSI network 1534 layer", RFC 1070, February 1989. 1536 [RFC1146] Zweig, J. and C. Partridge, "TCP alternate checksum 1537 options", RFC 1146, March 1990. 1539 [RFC1191] Mogul, J. and S. Deering, "Path MTU discovery", RFC 1191, 1540 November 1990. 1542 [RFC1701] Hanks, S., Li, T., Farinacci, D., and P. Traina, "Generic 1543 Routing Encapsulation (GRE)", RFC 1701, October 1994. 1545 [RFC1812] Baker, F., "Requirements for IP Version 4 Routers", 1546 RFC 1812, June 1995. 1548 [RFC1981] McCann, J., Deering, S., and J. Mogul, "Path MTU Discovery 1549 for IP version 6", RFC 1981, August 1996. 1551 [RFC2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, 1552 October 1996. 1554 [RFC2473] Conta, A. and S. Deering, "Generic Packet Tunneling in 1555 IPv6 Specification", RFC 2473, December 1998. 1557 [RFC2675] Borman, D., Deering, S., and R. Hinden, "IPv6 Jumbograms", 1558 RFC 2675, August 1999. 1560 [RFC2764] Gleeson, B., Heinanen, J., Lin, A., Armitage, G., and A. 1561 Malis, "A Framework for IP Based Virtual Private 1562 Networks", RFC 2764, February 2000. 1564 [RFC2780] Bradner, S. and V. Paxson, "IANA Allocation Guidelines For 1565 Values In the Internet Protocol and Related Headers", 1566 BCP 37, RFC 2780, March 2000. 1568 [RFC2827] Ferguson, P. and D. Senie, "Network Ingress Filtering: 1569 Defeating Denial of Service Attacks which employ IP Source 1570 Address Spoofing", BCP 38, RFC 2827, May 2000. 1572 [RFC2923] Lahey, K., "TCP Problems with Path MTU Discovery", 1573 RFC 2923, September 2000. 1575 [RFC3232] Reynolds, J., "Assigned Numbers: RFC 1700 is Replaced by 1576 an On-line Database", RFC 3232, January 2002. 1578 [RFC3366] Fairhurst, G. and L. Wood, "Advice to link designers on 1579 link Automatic Repeat reQuest (ARQ)", BCP 62, RFC 3366, 1580 August 2002. 1582 [RFC3819] Karn, P., Bormann, C., Fairhurst, G., Grossman, D., 1583 Ludwig, R., Mahdavi, J., Montenegro, G., Touch, J., and L. 1584 Wood, "Advice for Internet Subnetwork Designers", BCP 89, 1585 RFC 3819, July 2004. 1587 [RFC4191] Draves, R. and D. Thaler, "Default Router Preferences and 1588 More-Specific Routes", RFC 4191, November 2005. 1590 [RFC4213] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms 1591 for IPv6 Hosts and Routers", RFC 4213, October 2005. 1593 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 1594 Internet Protocol", RFC 4301, December 2005. 1596 [RFC4302] Kent, S., "IP Authentication Header", RFC 4302, 1597 December 2005. 1599 [RFC4459] Savola, P., "MTU and Fragmentation Issues with In-the- 1600 Network Tunneling", RFC 4459, April 2006. 1602 [RFC4821] Mathis, M. and J. Heffner, "Packetization Layer Path MTU 1603 Discovery", RFC 4821, March 2007. 1605 [RFC4963] Heffner, J., Mathis, M., and B. Chandler, "IPv4 Reassembly 1606 Errors at High Data Rates", RFC 4963, July 2007. 1608 [RFC4987] Eddy, W., "TCP SYN Flooding Attacks and Common 1609 Mitigations", RFC 4987, August 2007. 1611 [RFC5226] Narten, T. and H. Alvestrand, "Guidelines for Writing an 1612 IANA Considerations Section in RFCs", BCP 26, RFC 5226, 1613 May 2008. 1615 [RFC5246] Dierks, T. and E. Rescorla, "The Transport Layer Security 1616 (TLS) Protocol Version 1.2", RFC 5246, August 2008. 1618 [RFC5320] Templin, F., "The Subnetwork Encapsulation and Adaptation 1619 Layer (SEAL)", RFC 5320, February 2010. 1621 [RFC5445] Watson, M., "Basic Forward Error Correction (FEC) 1622 Schemes", RFC 5445, March 2009. 1624 [RFC5720] Templin, F., "Routing and Addressing in Networks with 1625 Global Enterprise Recursion (RANGER)", RFC 5720, 1626 February 2010. 1628 [RFC5927] Gont, F., "ICMP Attacks against TCP", RFC 5927, July 2010. 1630 [RFC6139] Russert, S., Fleischman, E., and F. Templin, "Routing and 1631 Addressing in Networks with Global Enterprise Recursion 1632 (RANGER) Scenarios", RFC 6139, February 2011. 1634 [RFC6169] Krishnan, S., Thaler, D., and J. Hoagland, "Security 1635 Concerns with IP Tunneling", RFC 6169, April 2011. 1637 [RFC6335] Cotton, M., Eggert, L., Touch, J., Westerlund, M., and S. 1638 Cheshire, "Internet Assigned Numbers Authority (IANA) 1639 Procedures for the Management of the Service Name and 1640 Transport Protocol Port Number Registry", BCP 165, 1641 RFC 6335, August 2011. 1643 [SIGCOMM] Luckie, M. and B. Stasiewicz, "Measuring Path MTU 1644 Discovery Behavior", November 2010. 1646 [TBIT] Medina, A., Allman, M., and S. Floyd, "Measuring 1647 Interactions Between Transport Protocols and Middleboxes", 1648 October 2004. 1650 [TCP-IP] "Archive/Hypermail of Early TCP-IP Mail List, 1651 http://www-mice.cs.ucl.ac.uk/multimedia/misc/tcp_ip/, May 1652 1987 - May 1990.". 1654 [WAND] Luckie, M., Cho, K., and B. Owens, "Inferring and 1655 Debugging Path MTU Discovery Failures", October 2005. 1657 Appendix A. Reliability 1659 Although a SEAL tunnel may span an arbitrarily-large subnetwork 1660 expanse, the IP layer sees the tunnel as a simple link that supports 1661 the IP service model. Links with high bit error rates (BERs) (e.g., 1662 IEEE 802.11) use Automatic Repeat-ReQuest (ARQ) mechanisms [RFC3366] 1663 to increase packet delivery ratios, while links with much lower BERs 1664 typically omit such mechanisms. Since SEAL tunnels may traverse 1665 arbitrarily-long paths over links of various types that are already 1666 either performing or omitting ARQ as appropriate, it would therefore 1667 be inefficient to require the tunnel endpoints to also perform ARQ. 1669 Appendix B. Integrity 1671 The SEAL header includes an integrity check field that covers the 1672 SEAL header and at least the inner packet headers. This provides for 1673 header integrity verification on a segment-by-segment basis for a 1674 segmented re-encapsulating tunnel path. 1676 Fragmentation and reassembly schemes must also consider packet- 1677 splicing errors, e.g., when two fragments from the same packet are 1678 concatenated incorrectly, when a fragment from packet X is 1679 reassembled with fragments from packet Y, etc. The primary sources 1680 of such errors include implementation bugs and wrapping IPv4 ID 1681 fields. 1683 In particular, the IPv4 16-bit ID field can wrap with only 64K 1684 packets with the same (src, dst, protocol)-tuple alive in the system 1685 at a given time [RFC4963]. When the IPv4 ID field is re-written by a 1686 middlebox such as a NAT or Firewall, ID field wrapping can occur with 1687 even fewer packets alive in the system. 1689 When outer IPv4 fragmentation is unavoidable, SEAL institutes rate 1690 limiting so that the number of packets admitted into the tunnel by 1691 the ITE does not exceed the number of unique packets that may be 1692 alive within the Internet. 1694 Appendix C. Transport Mode 1696 SEAL can also be used in "transport-mode", e.g., when the inner layer 1697 comprises upper-layer protocol data rather than an encapsulated IP 1698 packet. For instance, TCP peers can negotiate the use of SEAL (e.g., 1699 by inserting an unspecified 'SEAL_OPTION' TCP option during 1700 connection establishment) for the carriage of protocol data 1701 encapsulated as IP/SEAL/TCP. In this sense, the "subnetwork" becomes 1702 the entire end-to-end path between the TCP peers and may potentially 1703 span the entire Internet. 1705 If both TCPs agree on the use of SEAL, their protocol messages will 1706 be carried as IP/SEAL/TCP and the connection will be serviced by the 1707 SEAL protocol using TCP (instead of an encapsulating tunnel endpoint) 1708 as the transport layer protocol. The SEAL protocol for transport 1709 mode otherwise observes the same specifications as for Section 4. 1711 Appendix D. Historic Evolution of PMTUD 1713 The topic of Path MTU discovery (PMTUD) saw a flurry of discussion 1714 and numerous proposals in the late 1980's through early 1990. The 1715 initial problem was posed by Art Berggreen on May 22, 1987 in a 1716 message to the TCP-IP discussion group [TCP-IP]. The discussion that 1717 followed provided significant reference material for [FRAG]. An IETF 1718 Path MTU Discovery Working Group [MTUDWG] was formed in late 1989 1719 with charter to produce an RFC. Several variations on a very few 1720 basic proposals were entertained, including: 1722 1. Routers record the PMTUD estimate in ICMP-like path probe 1723 messages (proposed in [FRAG] and later [RFC1063]) 1725 2. The destination reports any fragmentation that occurs for packets 1726 received with the "RF" (Report Fragmentation) bit set (Steve 1727 Deering's 1989 adaptation of Charles Lynn's Nov. 1987 proposal) 1729 3. A hybrid combination of 1) and Charles Lynn's Nov. 1987 (straw 1730 RFC draft by McCloughrie, Fox and Mogul on Jan 12, 1990) 1732 4. Combination of the Lynn proposal with TCP (Fred Bohle, Jan 30, 1733 1990) 1735 5. Fragmentation avoidance by setting "IP_DF" flag on all packets 1736 and retransmitting if ICMPv4 "fragmentation needed" messages 1737 occur (Geof Cooper's 1987 proposal; later adapted into [RFC1191] 1738 by Mogul and Deering). 1740 Option 1) seemed attractive to the group at the time, since it was 1741 believed that routers would migrate more quickly than hosts. Option 1742 2) was a strong contender, but repeated attempts to secure an "RF" 1743 bit in the IPv4 header from the IESG failed and the proponents became 1744 discouraged. 3) was abandoned because it was perceived as too 1745 complicated, and 4) never received any apparent serious 1746 consideration. Proposal 5) was a late entry into the discussion from 1747 Steve Deering on Feb. 24th, 1990. The discussion group soon 1748 thereafter seemingly lost track of all other proposals and adopted 1749 5), which eventually evolved into [RFC1191] and later [RFC1981]. 1751 In retrospect, the "RF" bit postulated in 2) is not needed if a 1752 "contract" is first established between the peers, as in proposal 4) 1753 and a message to the MTUDWG mailing list from jrd@PTT.LCS.MIT.EDU on 1754 Feb 19. 1990. These proposals saw little discussion or rebuttal, and 1755 were dismissed based on the following the assertions: 1757 o routers upgrade their software faster than hosts 1759 o PCs could not reassemble fragmented packets 1761 o Proteon and Wellfleet routers did not reproduce the "RF" bit 1762 properly in fragmented packets 1764 o Ethernet-FDDI bridges would need to perform fragmentation (i.e., 1765 "translucent" not "transparent" bridging) 1767 o the 16-bit IP_ID field could wrap around and disrupt reassembly at 1768 high packet arrival rates 1770 The first four assertions, although perhaps valid at the time, have 1771 been overcome by historical events. The final assertion is addressed 1772 by the mechanisms specified in SEAL. 1774 Author's Address 1776 Fred L. Templin (editor) 1777 Boeing Research & Technology 1778 P.O. Box 3707 1779 Seattle, WA 98124 1780 USA 1782 Email: fltemplin@acm.org