idnits 2.17.1 draft-iab-link-encaps-04.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 13. -- Found old boilerplate from RFC 3978, Section 5.5 on line 1077. -- Found old boilerplate from RFC 3979, Section 5, paragraph 1 on line 1054. -- Found old boilerplate from RFC 3979, Section 5, paragraph 2 on line 1061. -- Found old boilerplate from RFC 3979, Section 5, paragraph 3 on line 1067. ** Found boilerplate matching RFC 3978, Section 5.4, paragraph 1 (on line 1083), which is fine, but *also* found old RFC 2026, Section 10.4C, paragraph 1 text on line 35. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. 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 2 instances of too long lines in the document, the longest one being 1 character in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == The document seems to lack the recommended RFC 2119 boilerplate, even if it appears to use RFC 2119 keywords. (The document does seem to have the reference to RFC 2119 which the ID-Checklist requires). -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (5 October 2006) is 6412 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- == Missing Reference: 'IEEE-802-1A.190' is mentioned on line 85, but not defined == Missing Reference: 'RFC1122' is mentioned on line 307, but not defined == Missing Reference: 'IEEE-802.1Q' is mentioned on line 508, but not defined == Unused Reference: 'IEEE.802-1D.1998' is defined on line 903, but no explicit reference was found in the text == Unused Reference: 'IEEE.802-3.1985' is defined on line 910, but no explicit reference was found in the text == Unused Reference: 'RFC2119' is defined on line 946, but no explicit reference was found in the text -- Obsolete informational reference (is this intentional?): RFC 1972 (Obsoleted by RFC 2464) -- Obsolete informational reference (is this intentional?): RFC 2472 (Obsoleted by RFC 5072, RFC 5172) Summary: 5 errors (**), 0 flaws (~~), 8 warnings (==), 9 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group B. Aboba, Ed. 3 INTERNET-DRAFT Elwyn Davies 4 Category: Informational D. Thaler 5 Internet Architecture Board 6 5 October 2006 8 Multiple Encapsulation Methods Considered Harmful 10 By submitting this Internet-Draft, each author represents that any 11 applicable patent or other IPR claims of which he or she is aware 12 have been or will be disclosed, and any of which he or she becomes 13 aware will be disclosed, in accordance with Section 6 of BCP 79. 15 Internet-Drafts are working documents of the Internet Engineering 16 Task Force (IETF), its areas, and its working groups. Note that 17 other groups may also distribute working documents as Internet- 18 Drafts. 20 Internet-Drafts are draft documents valid for a maximum of six months 21 and may be updated, replaced, or obsoleted by other documents at any 22 time. It is inappropriate to use Internet-Drafts as reference 23 material or to cite them other than as "work in progress." 25 The list of current Internet-Drafts can be accessed at 26 http://www.ietf.org/ietf/1id-abstracts.txt. 28 The list of Internet-Draft Shadow Directories can be accessed at 29 http://www.ietf.org/shadow.html. 31 This Internet-Draft will expire on April 1, 2007. 33 Copyright Notice 35 Copyright (C) The Internet Society (2006). All Rights Reserved. 37 Abstract 39 This document describes architectural and operational issues that 40 arise from link layer protocols supporting multiple Internet Protocol 41 encapsulation methods. 43 Table of Contents 45 1. Introduction .......................................... 3 46 1.1 Terminology .................................... 3 47 1.2 Ethernet Experience ............................ 4 48 1.3 PPP Experience ................................. 7 49 1.4 Potential Mitigations .......................... 8 50 2. Evaluation of Arguments for Multiple Encapsulations ... 9 51 2.1 Efficiency ..................................... 9 52 2.2 Multicast/Broadcast ............................ 10 53 2.3 Multiple Uses .................................. 11 54 3. Additional Issues ..................................... 12 55 3.1 Generality ..................................... 13 56 3.2 Layer Interdependence .......................... 14 57 3.3 Inspection of Payload Contents ................. 14 58 3.4 Interoperability Guidance ...................... 15 59 3.5 Service Consistency ............................ 16 60 3.6 Implementation Complexity ...................... 16 61 3.7 Negotiation .................................... 17 62 3.8 Roaming ........................................ 17 63 4. Security Considerations ............................... 18 64 5. IANA Considerations ................................... 18 65 6. Conclusion ............................................ 18 66 7. References ............................................ 19 67 7.1 Informative References .......................... 19 68 Acknowledgments .............................................. 22 69 Appendix A - IAB Members ..................................... 22 70 Authors' Addresses ........................................... 23 71 Intellectual Property Statement .............................. 24 72 Disclaimer of Validity ....................................... 24 73 Copyright Statement .......................................... 24 74 1. Introduction 76 This document describes architectural and operational issues arising 77 from use of multiple ways of encapsulating IP packets on the same 78 link. 80 While typically a link layer protocol supports only a single Internet 81 Protocol (IP) encapsulation method, this is not always the case. For 82 example, on the same cable it is possible to encapsulate an IPv4 83 packet using Ethernet [DIX] encapsulation as defined in "A Standard 84 for the Transmission of IP Datagrams over Ethernet Networks" [RFC894] 85 or IEEE 802 [IEEE-802-1A.190] encapsulation as defined in "A Standard 86 for the Transmission of IP Datagrams over IEEE 802 Networks" 87 [RFC1042]. Historically, a further encapsulation method was used on 88 some Ethernet systems as specified in "Trailer Encapsulations" 89 [RFC893]. 91 Similarly, both Point-to-Point Protocol (PPP) [RFC1661] and IEEE 92 802.16 [IEEE-802.16] support multiple encapsulation mechanisms. 94 1.1. Terminology 96 Broadcast domain 97 The set of all endpoints that receive broadcast frames sent by an 98 endpoint in the set. 100 Classification 101 As defined in [IEEE-802.16e], the process by which a Medium Access 102 Control (MAC) Service Data Unit (SDU) is mapped into a particular 103 transport connection for transmission between MAC peers. 105 Connection Identifier (CID) 106 In [IEEE-802.16e] the connection identifier is a 16-bit value that 107 identifies a transport connection or a uplink (UL)/downlink (DL) 108 pair of associated management connections. A connection is a 109 unidirectional mapping between base station (BS) and subscriber 110 station (SS) MAC peers. Each transport connection has a particular 111 set of associated parameters indicating characteristics such as the 112 ciphersuite and quality-of-service. 114 Link A communication facility or medium over which nodes can communicate 115 at the link layer, i.e., the layer immediately below IP. 117 Link Layer 118 The conceptual layer of control or processing logic that is 119 responsible for maintaining control of the link. The link layer 120 functions provide an interface between the higher-layer logic and 121 the link. The link layer is the layer immediately below IP. 123 1.2. Ethernet Experience 125 The fundamental issues with multiple encapsulation methods on the 126 same link are described in [RFC1042] and "Requirements for Internet 127 Hosts -- Communication Layers" [RFC1122]. This section summarizes 128 the concerns articulated in those documents and also describes the 129 limitations of approaches suggested to mitigate the problems, 130 including encapsulation negotiation and use of routers. 132 [RFC1042] described the potential issues resulting from 133 contemporaneous use of Ethernet and IEEE 802.3 encapsulation on the 134 same physical cable: 136 Interoperation with Ethernet 138 It is possible to use the Ethernet link level protocol [DIX] on 139 the same physical cable with the IEEE 802.3 link level 140 protocol. A computer interfaced to a physical cable used in 141 this way could potentially read both Ethernet and 802.3 packets 142 from the network. If a computer does read both types of 143 packets, it must keep track of which link protocol was used 144 with each other computer on the network and use the proper link 145 protocol when sending packets. 147 One should note that in such an environment, link level 148 broadcast packets will not reach all the computers attached to 149 the network, but only those using the link level protocol used 150 for the broadcast. 152 Since it must be assumed that most computers will read and send 153 using only one type of link protocol, it is recommended that if 154 such an environment (a network with both link protocols) is 155 necessary, an IP gateway be used as if there were two distinct 156 networks. 158 Note that the MTU for the Ethernet allows a 1500 octet IP 159 datagram, with the MTU for the 802.3 network allows only a 1492 160 octet IP datagram. 162 When multiple IP encapsulation methods were supported on a given 163 link, all hosts could not be assumed to support the same set of 164 encapsulation methods. This in turn implied that the broadcast 165 domain might not include all hosts on the link. Where a single 166 encapsulation does not reach all hosts on the link, a host needs to 167 determine the appropriate encapsulation prior to sending. While a 168 host supporting reception of multiple encapsulations could keep track 169 of the encapsulations it receives, this does not enable initiation of 170 communication; supporting initiation requires a host to support 171 sending of multiple encapsulations in order to determine which one to 172 use. However, requiring hosts to send and receive multiple 173 encapsulations is a potentially onerous requirement. 175 [RFC1122], Section 2.3.3 notes the difficulties with this approach: 177 Furthermore, it is not useful or even possible for a dual-format 178 host to discover automatically which format to send, because of 179 the problem of link-layer broadcasts. 181 To enable hosts that only support sending and receiving of a single 182 encapsulation to communicate with each other, a router can be 183 utilized to segregate the hosts by encapsulation. Here only the 184 router needs to support sending and receiving of multiple 185 encapsulations. This requires assigning a separate unicast prefix to 186 each encapsulation, or else all hosts in the broadcast domain would 187 not be reachable with a single encapsulation. 189 [RFC1122] Section 2.3.3 provided guidance on encapsulation support: 191 Every Internet host connected to a 10Mbps Ethernet cable: 193 o MUST be able to send and receive packets using RFC-894 194 encapsulation; 196 o SHOULD be able to receive RFC-1042 packets, intermixed 197 with RFC-894 packets; and 199 o MAY be able to send packets using RFC-1042 encapsulation. 201 An Internet host that implements sending both the RFC-894 and 202 the RFC-1042 encapsulation MUST provide a configuration switch 203 to select which is sent, and this switch MUST default to RFC- 204 894. 206 By making Ethernet encapsulation mandatory to implement for both send 207 and receive, and also the default for sending, [RFC1122] recognized 208 Ethernet as the predominant encapsulation, heading off potential 209 interoperability problems. 211 1.2.1. Trailer Encapsulation 213 As noted in "Trailer Encapsulations" [RFC893], trailer encapsulation 214 was an optimization developed to minimize memory-to-memory copies on 215 reception. By placing variable length IP and transport headers at 216 the end of the packet, page alignment of data could be more easily 217 maintained. Trailers were implemented in 4.2 Berkeley System 218 Distribution (BSD) (among others). While in theory trailer 219 encapsulation could have been applied to both Ethernet and IEEE 802 220 encapsulations (creating four potential encapsulations of IP!), in 221 practice trailer encapsulation was only supported for Ethernet. A 222 separate Ethertype was utilized in order to enable IP packets in 223 trailer encapsulation to be distinguished from [RFC894] 224 encapsulation. 226 [RFC1122] Section 2.3.1 described the issues with trailer 227 encapsulation: 229 DISCUSSION 231 The trailer protocol is a link-layer encapsulation technique 232 that rearranges the data contents of packets sent on the 233 physical network. In some cases, trailers improve the 234 throughput of higher layer protocols by reducing the amount of 235 data copying within the operating system. Higher layer 236 protocols are unaware of trailer use, but both the sending and 237 receiving host MUST understand the protocol if it is used. 238 Improper use of trailers can result in very confusing symptoms. 239 Only packets with specific size attributes are encapsulated 240 using trailers, and typically only a small fraction of the 241 packets being exchanged have these attributes. Thus, if a 242 system using trailers exchanges packets with a system that does 243 not, some packets disappear into a black hole while others are 244 delivered successfully. 246 IMPLEMENTATION: 248 On an Ethernet, packets encapsulated with trailers use a 249 distinct Ethernet type [RFC893], and trailer negotiation is 250 performed at the time that ARP is used to discover the link- 251 layer address of a destination system. 253 Specifically, the ARP exchange is completed in the usual manner 254 using the normal IP protocol type, but a host that wants to 255 speak trailers will send an additional "trailer ARP reply" 256 packet, i.e., an ARP reply that specifies the trailer 257 encapsulation protocol type but otherwise has the format of a 258 normal ARP reply. If a host configured to use trailers 259 receives a trailer ARP reply message from a remote machine, it 260 can add that machine to the list of machines that understand 261 trailers, e.g., by marking the corresponding entry in the ARP 262 cache. 264 Hosts wishing to receive trailers send trailer ARP replies 265 whenever they complete exchanges of normal ARP messages for IP. 266 Thus, a host that received an ARP request for its IP protocol 267 address would send a trailer ARP reply in addition to the 268 normal IP ARP reply; a host that sent the IP ARP request would 269 send a trailer ARP reply when it received the corresponding IP 270 ARP reply. In this way, either the requesting or responding 271 host in an IP ARP exchange may request that it receive 272 trailers. 274 This scheme, using extra trailer ARP reply packets rather than 275 sending an ARP request for the trailer protocol type, was 276 designed to avoid a continuous exchange of ARP packets with a 277 misbehaving host that, contrary to any specification or common 278 sense, responded to an ARP reply for trailers with another ARP 279 reply for IP. This problem is avoided by sending a trailer ARP 280 reply in response to an IP ARP reply only when the IP ARP reply 281 answers an outstanding request; this is true when the hardware 282 address for the host is still unknown when the IP ARP reply is 283 received. A trailer ARP reply may always be sent along with an 284 IP ARP reply responding to an IP ARP request. 286 Since trailer encapsulation negotiation depends on ARP, it can only 287 be used where all hosts on the link are within the same broadcast 288 domain. It was assumed that all hosts supported sending and 289 receiving ARP packets in standard Ethernet encapsulation [RFC894], so 290 that negotiation between Ethernet and IEEE 802 encapsulation was not 291 required, only negotiation between standard Ethernet [RFC894] and 292 trailer [RFC893] encapsulation. Had hosts supporting trailer 293 encapsulation also supported IEEE 802 framing, the negotiation would 294 have been complicated still further. 296 [RFC1122] Section 2.3.1 provided the following guidance for use of 297 trailer encapsulation: 299 The trailer protocol for link-layer encapsulation MAY be used, but 300 only when it has been verified that both systems (host or gateway) 301 involved in the link-layer communication implement trailers. If 302 the system does not dynamically negotiate use of the trailer 303 protocol on a per-destination basis, the default configuration 304 MUST disable the protocol. 306 4.2BSD did not support dynamic negotiation, only configuration of 307 trailer encapsulation at boot time, and therefore [RFC1122] required 308 that the trailer encapsulation be disabled by default on those 309 systems. 311 1.3. PPP Experience 313 PPP can support both encapsulation of IEEE 802 frames as defined in 314 [RFC3518], as well as IPv4 and IPv6 [RFC2472] packets. Multiple 315 compression schemes are also supported. 317 In addition to PPP Data Link Layer (DLL) protocol numbers allocated 318 for IPv4 (0x0021) and IPv6 (0x0057), Bridging PDU (0x0031), two 319 codepoints have been assigned for RObust Header Compression (ROHC) 320 [RFC3095] (ROHC small-CID (0x0003), ROHC large-CIDE (0x0005)), two 321 for Van Jacobson compression [RFC1144] (Compressed TCP/IP (0x002d), 322 Uncompressed TCP/IP (002f)), one for IPv6 Header Compression 323 [RFC2507] (0x004f) and nine codepoints for RTP IP Header Compression 324 [RFC3544], (Full Header (0x0061), Compressed TCP (0x0063), Compressed 325 Non TCP (0x0065), UDP 8 (0x0067), RTP 8 (0x0069), Compressed TCP No 326 Delta (0x2063), Context State (0x2065), UDP 16 (0x2067), RTP 16 327 (0x2069)). 329 Although PPP can encapsulate IP packets in multiple ways, typically 330 multiple encapsulation schemes are not operational on the same link 331 and therefore the issues described in this document rarely arise. 332 For example, while PPP can support both encapsulation of IEEE 802 333 frames as defined in [RFC3518], as well as IPv4 and IPv6 [RFC2472] 334 packets, in practice multiple encapsulation mechanisms are not 335 operational on the same link. Similarly, only a single compression 336 scheme is typically negotiated for use on a link. 338 1.4. Potential Mitigations 340 In order to mitigate problems arising from multiple encapsulation 341 methods, it may be possible to use switches or routers, or to attempt 342 to negotiate the encapsulation method to be used. As described 343 below, neither approach may be completely satisfactory. 345 The use of switches or routers to enable communication between hosts 346 utilizing multiple encapsulation methods is not a panacea. If 347 separate unicast prefixes are used for each encapsulation, then the 348 choice of encapsulation can be determined from the routing table. If 349 the same unicast prefix is used for each encapsulation method, it is 350 necessary to keep state for each destination host. However, this may 351 not work in situations where hosts using different encapsulations 352 respond to the same anycast address. 354 In situations where multiple encapsulation methods are enabled on a 355 single link, negotiation may be supported to allow hosts to determine 356 how to encapsulate a packet for a particular destination host. 358 Negotiating the encapsulation above the link layer is potentially 359 problematic since the negotiation itself may need to be carried out 360 using multiple encapsulations. In theory it is possible to negotiate 361 an encapsulation method by sending negotiation packets over all 362 encapsulation methods supported, and keeping state for each 363 destination host. However, if the encapsulation method must be 364 dynamically negotiated for each new on-link destination, 365 communication to new destinations may be delayed. If most 366 communication is short, and the negotiation requires an extra round 367 trip beyond link-layer address resolution, this can become a 368 noticeable factor in performance. Also, the negotiation may result 369 in consumption of additional bandwidth. 371 2. Evaluation of Arguments for Multiple Encapsulations 373 There are several reasons often given in support of multiple 374 encapsulation methods. We discuss each in turn, below. 376 2.1. Efficiency 378 Claim: Multiple encapsulation methods allow for greater efficiency. 379 For example, it has been argued that IEEE 802 or Ethernet 380 encapsulation of IP results in excessive overhead due to the size of 381 the data frame headers, and that this can adversely affect 382 performance on wireless networks, particularly in situations where 383 support of Voice over IP (VOIP) is required. 385 Discussion: Even where these performance concerns are valid, 386 solutions exist that do not require defining multiple IP 387 encapsulation methods. For example, links may support Ethernet frame 388 compression so that Ethernet Source and Destination Address fields 389 are not sent with every packet. 391 It is possible for link layers to negotiate compression without 392 requiring higher layer awareness; the Point-to-Point Protocol (PPP) 393 [RFC1661] is an example. "The PPP Compression Control Protocol 394 (CCP)" [RFC1962] enables negotiation of data compression mechanisms, 395 and "Robust Header Compression (ROHC) over PPP" [RFC3241] and "IP 396 Header Compression over PPP" [RFC3544] enable negotiation of header 397 compression, without Internet layer awareness. Any frame can be 398 "decompressed" based on the content of the frame, and prior state 399 based on previous control messages or data frames. Use of 400 compression is a good way to solve the efficiency problem without 401 introducing problems at higher layers. 403 There are also situations in which use of multiple encapsulations can 404 degrade performance or result in packet loss. The use of multiple 405 encapsulation methods with differing Maximum Transfer Units (MTUs) 406 can result in differing MTUs for on-link destinations. If the link- 407 layer protocol does not provide per-destination MTUs to the IP layer, 408 it will need to use a default MTU; to avoid fragmentation this must 409 be less than or equal to the minimum MTU of on-link destinations. If 410 the default MTU is too low, the full bandwidth may not be achievable. 412 If the default MTU is too high, packet loss will result unless or 413 until IP Path MTU Discovery is used to discover the correct MTU. 415 Recommendation: Where encapsulation is an efficiency issue, use 416 header compression. Where the encapsulation method, or the use of 417 compression, must be negotiated, negotiation should either occur as 418 part of bringing up the link, or be piggybacked in the link-layer 419 address resolution exchange; only a single compression scheme should 420 be negotiated on a link. Where the MTU may vary among destinations 421 on the same link, the link layer protocol should provide a per 422 destination MTU to IP. 424 2.2. Multicast/Broadcast 426 Claim: Support for Ethernet encapsulation requires layer 2 support 427 for distribution of IP multicast/broadcast packets. In situations 428 where this is difficult, support for Ethernet is problematic and 429 other encapsulations are necessary. 431 Discussion: Irrespective of the encapsulation used, IP packets sent 432 to multicast (IPv4/IPv6) or broadcast addresses (IPv4) need to reach 433 all potential on-link receivers. Use of alternative encapsulations 434 cannot remove this requirement, although there is considerable 435 flexibility in how it can be met. Non-Broadcast Multiple Access 436 (NBMA) networks can still support the broadcast/multicast service via 437 replication of unicast frames. 439 Techniques are also available for improving the efficiency of IP 440 multicast/broadcast delivery in wireless networks. In order to be 441 receivable by any host within listening range, an IP 442 multicast/broadcast packet sent as link layer multicast/broadcast 443 over a wireless link needs to be sent at the lowest rate supported by 444 listeners. If the sender does not keep track of the rates negotiated 445 by group listeners, by default multicast/broadcast traffic is sent 446 at the lowest supported rate, resulting in increased overhead. 447 However, a sender can also deliver an IP multicast/broadcast packet 448 using unicast frame(s) where this would be more efficient. For 449 example, in IEEE 802.11, multicast/broadcast traffic sent from the 450 Station (STA) to the Access Point (AP) is always sent as unicast, and 451 the AP tracks the negotiated rate for each STA, so that it can send 452 unicast frames at a rate appropriate for each station. 454 In order to limit the propagation of link-scope multicast or 455 broadcast traffic, it is possible to assign a separate prefix to each 456 host. 458 Unlike broadcasts, which are received by all hosts on the link 459 regardless of the protocol they are running, multicasts only need be 460 received by those hosts belonging to the multicast group. In wired 461 networks, it is possible to avoid forwarding multicast traffic on 462 switch ports without group members, by snooping of Internet Group 463 Management Protocol (IGMP) and Multicast Listener Discovery (MLD) 464 traffic as described in "Considerations for IGMP and MLD Snooping 465 Switches" [RFC4541]. 467 In wireless media where data rates to specific destinations are 468 negotiated and may vary over a wide range, it may be more efficient 469 to send multiple frames via link layer unicast than to send a single 470 multicast/broadcast frame. For example, in [IEEE-802.11] 471 multicast/broadcast traffic from the client station (STA) to the 472 Access Point (AP) is sent via link layer unicast. 474 Recommendation: Where support for link layer multicast/broadcast is 475 problematic, limit the propagation of link-scope multicast and 476 broadcast traffic by assignment of separate prefixes to hosts. In 477 some circumstances, it may be more efficient to distribute 478 multicast/broadcast traffic as multiple link-layer unicast frames. 480 2.3. Multiple Uses 482 Claim: No single encapsulation is optimal for all purposes. 483 Therefore where a link layer is utilized in disparate scenarios (such 484 as both fixed and mobile deployments), multiple encapsulations are a 485 practical requirement. 487 Discussion: "Architectural Principles of the Internet" [RFC1958] 488 point 3.2 states: 490 If there are several ways of doing the same thing, choose one. If 491 a previous design, in the Internet context or elsewhere, has 492 successfully solved the same problem, choose the same solution 493 unless there is a good technical reason not to. Duplication of 494 the same protocol functionality should be avoided as far as 495 possible, without of course using this argument to reject 496 improvements. 498 Existing encapsulations have proven themselves capable of supporting 499 disparate usage scenarios. For example, the Point-to-Point Protocol 500 (PPP) has been utilized by wireless link layers such as GPRS, as well 501 as in wired networks in applications such as "PPP over SONET/SDH" 502 [RFC2615]. PPP can even support bridging, as described in "Point-to- 503 Point Protocol (PPP) Bridging Control Protocol (BCP)" [RFC3518]. 505 Similarly, Ethernet encapsulation has been used in wired networks as 506 well as Wireless Local Area Networks (LANs) such as IEEE 802.11 507 [IEEE-802.11]. Ethernet can also support Virtual LANs (VLANs) and 508 Quality of Service (QoS) [IEEE-802.1Q]. 510 Therefore disparate usage scenarios can be addressed by choice of a 511 single encapsulation, rather than multiple encapsulations. Where an 512 existing encapsulation is suitable, this is preferable to creating a 513 new encapsulation. 515 Where encapsulations other than IP over Point-to-Point Protocol (PPP) 516 [RFC1661], Ethernet or IEEE 802 are supported, difficulties in 517 operating system integration can lead to interoperability problems. 519 In order to take advantage of operating system support for IP 520 encapsulation over PPP, Ethernet or IEEE 802, it may be tempting for 521 a driver supporting an alternative encapsulation to emulate PPP, 522 Ethernet or IEEE 802 support. Typically, PPP emulation requires that 523 the driver implement PPP, enabling translation of PPP control and 524 data frames to the equivalent native facilities. Similarly, Ethernet 525 or IEEE 802 emulation typically requires that the driver implement 526 Dynamic Host Configuration Protocol (DHCP)v4 or v6, Router 527 Solicitation/Router Advertisement (RS/RA), Address Resolution 528 Protocol (ARP) or IPv6 Neighbor Discovery (ND) in order to enable 529 translation of these frames to and from native facilities. 531 Where drivers are implemented in kernel mode, the work required to 532 provide faithful emulation may be substantial. This creates the 533 temptation to cut corners, potentially resulting in interoperability 534 problems. 536 For example, it might be tempting for driver implementations to 537 neglect IPv6 support. A driver emulating PPP might support only 538 IPCP, but not IPCPv6; a driver emulating Ethernet or IEEE 802 might 539 support only DHCPv4 and ARP, but not DHCPv6, RS/RA or ND. As a 540 result, an IPv6 host connecting to a network supporting IPv6 might 541 find itself unable to use IPv6 due to lack of driver support. 543 Recommendation: Support a single existing encapsulation where 544 possible. Emulation of PPP, Ethernet or IEEE 802 on top of 545 alternative encapsulations should be avoided. 547 3. Additional Issues 549 There are a number of additional issues arising from use of multiple 550 encapsulation methods, as hinted at in section 1. We discuss each of 551 these below. 553 3.1. Generality 555 Link layer protocols such as [IEEE.802-1A.1990] and [DIX] inherently 556 support the ability to add support for a new packet type without 557 modification to the link layer protocol. 559 IEEE 802.16 [IEEE-802.16] splits the Media Access Control (MAC) layer 560 into a number of sublayers. For the uppermost of these, the standard 561 defines the concept of a service-specific Convergence Sublayer (CS). 562 The two underlying sublayers (the MAC Common Part Sublayer and the 563 Security Sublayer) provide common services for all instantiations of 564 the CS. While [IEEE-802.16] defined support for the Asynchronous 565 Transfer Mode (ATM) CS and the Packet CS, [IEEE-802.16e] added 566 support for four new Convergence Sublayers. As a result, 567 [IEEE-802.16e] defines the ATM CS, Packet CS, Ethernet CS, IPv4 CS 568 and IPv6 CS as well as eight other Convergence Sublayers. 570 As noted in [Generic], [IEEE-802.16] appears to imply that the 571 standard will need to be modified to support new packet types: 573 We are concerned that the 802.16 protocol cannot easily be 574 extendable to transport new protocols over the 802.16 air 575 interface. It would appear that a Convergence Sublayer is needed 576 for every type of protocol transported over the 802.16 MAC. Every 577 time a new protocol type needs to be transported over the 802.16 578 air interface, the 802.16 standard needs to be modified to define 579 a new CS type. We need to have a generic Packet Convergence 580 Sublayer that can support multi-protocols and which does not 581 require further modification to the 802.16 standard to support new 582 protocols. We believe that this was the original intention of the 583 Packet CS. Furthermore, we believe it is difficult for the 584 industry to agree on a set of CSes that all devices must implement 585 to claim "compliance". 587 The use of IP and/or upper layer protocol specific encapsulation 588 methods, rather than a 'neutral' general purpose encapsulation may 589 give rise to a number of undesirable effects explored in the 590 following subsections. 592 If the link layer does not provide a general purpose encapsulation 593 method, deployment of new IP and/or upper layer protocols will be 594 dependent on deployment of the corresponding new encapsulation 595 support in the link layer. 597 Even if a single encapsulation method is used, problems can still 598 occur if de-multiplexing of ARP, IPv4, IPv6, and any other protocols 599 in use, is not supported at the link layer. While is possible to 600 demultiplex such packets based on the Version field (first four bits 601 on the packet), this assumes that IPv4-only implementations will be 602 able to properly handle IPv6 packets. As a result, a more robust 603 design is to demultiplex protocols in the link layer, such as by 604 assigning a different protocol type, as is done in IEEE 802 media 605 where a Type of 0x0800 is used for IPv4, and 0x86DD for IPv6. 607 Recommendations: Link layer protocols should enable network packets 608 (IPv4, IPv6, ARP, etc.) to be de-multiplexed in the link layer. 610 3.2. Layer Interdependence 612 Within IEEE 802.16, the process by which frames are selected for 613 transmission on a connection identifier (CID) is known as 614 "classification". Fields in the Ethernet, IP and UDP/TCP headers can 615 be used for classification; for a particular CS a defined subset of 616 header fields may be applied for that purpose. 618 Utilizing IP and/or upper layer headers in link layer classification 619 will almost inevitably lead to interdependencies between link layer 620 and upper layer specifications. Although this might appear to be 621 desirable in terms of providing a highly specific (and hence 622 interoperable) mapping between the capabilities provided by the link 623 layer (e.g., quality of service support) and those that are needed by 624 upper layers, this sort of capability is probably better provided by 625 a more comprehensive service interface (Application Programming 626 Interface) in conjunction with a single encapsulation mechanism. 628 IPv6, in particular, provides an extensible header system. An upper 629 layer specific classification scheme would still have to provide a 630 degree of generality in order to cope with future extensions of IPv6 631 that might wish to make use of some of the link layer services 632 already provided. 634 Recommendations: Upper layer specific classification schemes should 635 be avoided. 637 3.3. Inspection of Payload Contents 639 If a classification scheme utilizing higher layer headers proposes to 640 inspect the contents of the packet being encapsulated (e.g., 802.16 641 IP CS mechanisms for determining the connection identifier (CID) to 642 use to transmit a packet), the fields available for inspection may be 643 limited if the packet is compressed or encrypted before passing to 644 the link layer. This may prevent the link layer from utilizing 645 existing compression mechanisms, such as Van Jacobson Compression 646 [RFC1144], ROHC [RFC3095][RFC3759], Compressed RTP (CRTP) [RFC2508], 647 Enhanced Compressed RTP (ECRTP) [RFC3545] or IP Header Compression 648 [RFC2507]. 650 Recommendations: Link layer classification schemes should not rely on 651 the contents of higher layer headers. 653 3.4. Interoperability Guidance 655 In situations where multiple CSes are operational and capable of 656 carrying IP traffic, interoperability problems are possible in the 657 absence of clear implementation guidelines. For example, there is no 658 guarantee that other hosts on the link will support the same set of 659 CSes, or that if they do, that their routing tables will result in 660 identical preferences. 662 In 802.16 the Subscriber Station (SS) indicates the Convergence 663 Sublayers it supports to the Base Station (BS), which selects from 664 the list one or more that it will support on the link. Therefore it 665 is possible for multiple CSes to be operational. 667 Note that IEEE 802.16 does not provide multiple encapsulation methods 668 for the same kind of data payload; it defines exactly one 669 encapsulation scheme for each data payload. For example, there is 670 one way to encapsulate a raw IPv4 packet into an IEEE 802.16 MAC 671 frame, one encapsulation scheme for a raw IPv6 packet, etc. There is 672 also one way to encapsulate an Ethernet frame, even when there are 673 multiple possibilities for classifying an Ethernet frame for 674 forwarding over a connection identifier (CID). Since support for 675 multiple CSes enables IEEE 802.16 to encapsulate layer 2 frames as 676 well as layer 3 packets, IP packets may be directly encapsulated in 677 IEEE802.16 MAC frames as well as framed with Ethernet headers in 678 IEEE802.16 MAC frames. Where CSes supporting both layer 2 frames as 679 well as layer 3 packets are operational on the same link, a number of 680 issues may arise, including: 682 Use of Address Resolution Protocol (ARP) 683 Where both IPv4 CS and Ethernet CS are operational, it may not be 684 obvious how address resolution should be implemented. For example, 685 should an ARP frame be encapsulated over the Ethernet CS, or should 686 alternative mechanisms be used for address resolution, utilizing 687 the IPv4 CS? 689 Data Frame Encapsulation 690 When sending an IP packet, which CS should be used? Where multiple 691 encapsulations are operational, multiple connection identifiers 692 (CIDs) will also be present. The issue can therefore be treated as 693 a multi-homing problem, with each CID constituting its own 694 interface. Since a given CID may have associated bandwidth or 695 quality of service constraints, routing metrics could be adjusted 696 to take this into account, allowing the routing layer to choose 697 based on which CID (and encapsulation) appears more attractive. 699 This could lead to interoperability problems or routing asymmetry. 700 For example, consider the effects on IPv6 Neighbor Discovery: 702 [a] If hosts choose to send IPv6 Neighbor Discovery traffic on 703 different CSes, it is possible that a host sending an IPv6 Neighbor 704 Discovery packet will not receive a reply, even though the target 705 host is reachable over another CS. 707 [b] Where hosts all support the same set of CSes, but have different 708 routing preferences, it is possible for a host to send an IPv6 709 Neighbor Discovery packet over one CS and receive a reply over 710 another CS. 712 Recommendations: Given these issues, it is strongly recommended that 713 only a single CS supporting a single encapsulation method be usable 714 in a given circumstance. 716 3.5. Service Consistency 718 If a link layer protocol provides multiple encapsulation methods, the 719 services offered to the IP and upper layer protocols may differ 720 qualitatively between the different encapsulation methods. For 721 example, the 802.16 [IEEE-802.16] link layer protocol offers both 722 'native' encapsulation for raw IPv4 and IPv6 packets, and emulated 723 Ethernet encapsulation. In the raw case, the IP layer has direct 724 access to the quality of service (QoS) capabilities of the 802.16 725 transmission channels, whereas using the Ethernet encapsulation the 726 IP QoS has first to be mapped through the rather more limited 727 capabilities of Ethernet QoS. Consequently, the service offered to 728 an application depends on the encapsulation method employed and may 729 be inconsistent between sessions. This may be confusing for the user 730 and the application. 732 Recommendations: If multiple encapsulation methods for IP packets on 733 a single link layer technology are deemed to be necessary, care 734 should be taken to match the services available between encapsulation 735 methods as closely as possible. 737 3.6. Implementation Complexity 739 Support of multiple encapsulation methods results in additional 740 implementation complexity. Lack of uniform encapsulation support 741 also results in potential interoperability problems. To avoid 742 interoperability issues, devices with limited resources may be 743 required to implement multiple encapsulation mechanisms, which may 744 not be practical. 746 When encapsulation methods require hardware support, implementations 747 may choose to support different encapsulation sets, resulting in 748 market fragmentation. This can prevent users from benefiting from 749 economies of scale, precluding some uses of the technology entirely. 751 Recommendations: Choose a single mandatory to implement 752 encapsulation mechanism for both sending and receiving, and make that 753 encapsulation mechanism the default for sending. 755 3.7. Negotiation 757 The complexity of negotiation within ARP or IP can be reduced by 758 performing encapsulation negotiation within the link layer. 760 However, unless the link layer allows the negotiation of the 761 encapsulation between any two hosts, then interoperability problems 762 can still result if more than one encapsulation is possible on a 763 given link. In general, a host cannot assume that all other hosts on 764 a link support the same set of encapsulation methods, so that unless 765 a link layer protocol only supports point-to-point communication, 766 negotiation of multiple potential encapsulation methods will be 767 problematic. To avoid this problem, it is desirable for link layer 768 encapsulation negotiation to determine a single IP encapsulation, not 769 merely to indicate which encapsulation methods are possible. 771 Recommendations: Encapsulation negotiation is best handled in the 772 link layer. In order to avoid dependencies on the data frame 773 encapsulation mechanism, it is preferable for the negotiation to be 774 carried out using management frames, if they are supported. If 775 multiple encapsulations are required and negotiation is provided, 776 then the negotiation should result in a single encapsulation method 777 being negotiated on the link. 779 3.8. Roaming 781 Where a mobile node roams between base stations or to a fixed 782 infrastructure and the base stations and fixed infrastructure do not 783 all support the same set of encapsulations, then it may be necessary 784 to alter the encapsulation method, potentially in mid-conversation. 785 Even if the change can be handled seamlessly at the link and IP layer 786 so that applications are not affected, unless the services offered 787 over the different encapsulations are equivalent (see Section 3.5) 788 the service experienced by the application may change as the mobile 789 node crosses boundaries. If the service is significantly different, 790 it might even require 'in-flight' renegotiation which most 791 applications are not equipped to manage. 793 Recommendations: Ensure uniformity of the encapsulation set 794 (preferably only a single encapsulation) within a given mobile 795 domain, between mobile domains, and between mobile domains and fixed 796 infrastructure. If a link layer protocol offers multiple 797 encapsulation methods for IP packets, it is strongly recommended that 798 only one of these encapsulation methods should be in use on any given 799 link or within a single wireless transmission domain. 801 4. Security Considerations 803 The use of multiple encapsulation methods does not appear to have 804 significant security implications. 806 An attacker might be able to utilize an encapsulation method which 807 was not in normal use on a link to cause a Denial of Service attack 808 which would exhaust the processing resources of interfaces if packets 809 utilizing this encapsulation were passed up the stack to any 810 significant degree before being discarded. 812 An attacker might be able to force a more cumbersome encapsulation 813 method between two endpoints, even when a lighter weight one is 814 available, hence forcing higher resource consumption on the link and 815 within those endpoints, or causing fragmentation. Since IP fragments 816 are more difficult to classify than non-fragments, this may result in 817 packet loss or may even expose security vulnerabilities [WEP]. 819 If different methods have different security properties, an attacker 820 might be able to force a less secure method as an elevation path to 821 get access to some other resource or data. Similarly, if one method 822 is rarely used, that method is potentially more likely to have 823 exploitable implementation bugs. 825 Since lower layer classification methods may need to inspect fields 826 in the packet being encapsulated, this might deter the deployment of 827 end-to-end security which is undesireable. Where encryption of 828 upper layer headers (e.g. IPsec tunnel mode) is required, this may 829 obscure headers required for classification. As a result, it may be 830 necessary for all encrypted traffic to flow over a single connection. 832 5. IANA Considerations 834 This document has no actions for IANA. 836 6. Conclusion 838 The use of multiple encapsulation methods on the same link is 839 problematic, as discussed above. 841 Although multiple IP encapsulation methods were defined on Ethernet 842 cabling, recent implementations support only the Ethernet 843 encapsulation of IPv4 defined in [RFC894]. In order to avoid a 844 repeat of the experience with IPv4, for operation of IPv6 on IEEE 845 802.3 media, only the Ethernet encapsulation was defined in "A Method 846 for the Transmission of IPv6 Packets over Ethernet Networks" 847 [RFC1972], later updated in [RFC2464]. 849 In addition to the recommendations given earlier, we give the 850 following general recommendations to avoid problems resulting from 851 use of multiple IP encapsulation methods: 853 When developing standards for encapsulating IP packets on a link 854 layer technology, it is desirable that only a single encapsulation 855 method should be standardized for each link layer technology; 857 If a link layer protocol offers multiple encapsulation methods for 858 IP packets, it is strongly recommended that only one of these 859 encapsulation methods should be in use within any given link or 860 wireless transmission domain; 862 Where multiple encapsulation methods are supported on a link, a 863 single encapsulation should be mandatory to implement for send and 864 receive. 866 7. References 868 7.1. Informative References 870 [DIX] Digital Equipment Corporation, Intel Corporation, and 871 Xerox Corporation, "The Ethernet -- A Local Area Network: 872 Data Link Layer and Physical Layer (Version 2.0)", 873 November 1982. 875 [Generic] Wang, L. et al, "A Generic Packet Convergence Sublayer 876 (GPCS) for Supporting Multiple Protocols over 802.16 Air 877 Interface", Submission to IEEE 802.16g: 878 CB0216g_05_025r4.pdf, November 2005, . 881 [IEEE-802.16] Institute of Electrical and Electronics Engineers, 882 "Information technology - Telecommunications and 883 information exchange between systems - Local and 884 metropolitan area networks, Part 16: Air Interface for 885 Fixed Broadband Wireless Access Systems", IEEE Standard 886 802.16-2004, October 2004. 888 [IEEE-802.16e] Institute of Electrical and Electronics Engineers, 889 "Information technology - Telecommunications and 890 information exchange between systems - Local and 891 Metropolitan Area Networks - Part 16: Air Interface for 892 Fixed and Mobile Broadband Wireless Access Systems, 893 Amendment for Physical and Medium Access Control Layers 894 for Combined Fixed and Mobile Operation in Licensed 895 Bands", IEEE P802.16e, September 2005. 897 [IEEE.802-1A.1990] 898 Institute of Electrical and Electronics Engineers, "Local 899 Area Networks and Metropolitan Area Networks: Overview 900 and Architecture of Network Standards", IEEE Standard 901 802.1A, 1990. 903 [IEEE.802-1D.1998] 904 Institute of Electrical and Electronics Engineers, 905 "Information technology - Telecommunications and 906 information exchange between systems - Local area 907 networks - Media access control (MAC) bridges", IEEE 908 Standard 802.1D, 1998. 910 [IEEE.802-3.1985] 911 Institute of Electrical and Electronics Engineers, 912 "Carrier Sense Multiple Access with Collision Detection 913 (CSMA/CD) Access Method and Physical Layer 914 Specifications", IEEE Standard 802.3, 1985. 916 [IEEE-802.11] Institute of Electrical and Electronics Engineers, 917 "Wireless LAN Medium Access Control (MAC) and Physical 918 Layer (PHY) Specifications", IEEE Standard 802.11, 2003. 920 [RFC893] Leffler, S. and M. Karels, "Trailer encapsulations", RFC 921 893, April 1984. 923 [RFC894] Hornig, C., "Standard for the transmission of IP 924 datagrams over Ethernet networks", STD 41, RFC 894, April 925 1984. 927 [RFC1042] Postel, J. and J. Reynolds, "Standard for the 928 transmission of IP datagrams over IEEE 802 networks", STD 929 43, RFC 1042, February 1988. 931 [RFC1144] Jacobson, V., "Compressing TCP/IP Headers for Low-Speed 932 Serial Links", RFC 1144, February 1990. 934 [RFC1661] Simpson, W., "The Point-to-Point Protocol (PPP)", STD 51, 935 RFC 1661, July 1994. 937 [RFC1958] Carpenter, B., "Architectural Principles of the 938 Internet", RFC 1958, June 1996. 940 [RFC1962] Rand, D., "The PPP Compression Control Protocol (CCP)", 941 RFC 1962, June 1996. 943 [RFC1972] Crawford, M., "A Method for the Transmission of IPv6 944 Packets over Ethernet Networks", RFC 1972, August 1996. 946 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 947 Requirement Levels", BCP 14, RFC 2119, March 1997. 949 [RFC2472] Haskin, D. and E. Allen, "IP Version 6 over PPP", RFC 950 2472, December 1998. 952 [RFC2464] Crawford, M., "Transmission of IPv6 Packets over Ethernet 953 Networks", RFC 2464, December 1998. 955 [RFC2507] Degermark, M., Nordgren, B., and S. Pink, "IP Header 956 Compression", RFC 2507, February 1999. 958 [RFC2508] Casner, S. and V. Jacobson, "Compressing IP/UDP/RTP 959 Headers for Low-Speed Serial Links", RFC 2508, February 960 1999. 962 [RFC2615] Malis, A. and W. Simpson, "PPP over SONET/SDH", RFC 2615, 963 June 1999. 965 [RFC3095] Bormann, C., et. al, "RObust Header Compression (ROHC): 966 Framework and four profiles: RTP, UDP, ESP and 967 uncompressed", RFC 3095, July 2001. 969 [RFC3241] Bormann, C., "Robust Header Compression (ROHC) over PPP", 970 RFC 3241, April 2002. 972 [RFC3518] Higashiyama, M., Baker, F. and T. Liao, "Point-to-Point 973 Protocol (PPP) Bridging Control Protocol (BCP)", RFC 974 3518, April 2003. 976 [RFC3544] Koren, T., Casner, S. and C. Bormann, "IP Header 977 Compression over PPP", RFC 3544, July 2003. 979 [RFC3545] Koren, T., Casner, S., Geevarghese, J., Thompson, B., and 980 P. Ruddy, "Enhanced Compressed RTP (CRTP) for Links with 981 High Delay, Packet Loss and Reordering", RFC 3545, July 982 2003. 984 [RFC3759] Jonsson, L-E., "RObust Header Compression (ROHC): 985 Terminology and Channel Mapping Examples", RFC 3759, 986 April 2004. 988 [RFC4541] Christensen, M., Kimball, K. and F. Solensky, 989 "Considerations for Internet Group Management Protocol 990 (IGMP) and Multicast Listener Discovery (MLD) Snooping 991 Switches", RFC 4541, May 2006. 993 [WEP] Bittau, A., Handley, M. and J. Lackey, "The Final Nail in 994 WEP's Coffin", Proceedings of the 2006 IEEE Symposium on 995 Security and Privacy, pp. 386-400. 997 Acknowledgments 999 The authors would like to acknowledge Jeff Mandin, Bob Hinden, Jari 1000 Arkko, and Phil Roberts for contributions to this document. 1002 Appendix A - IAB Members at the time of this writing 1004 Bernard Aboba 1005 Loa Andersson 1006 Brian Carpenter 1007 Leslie Daigle 1008 Elwyn Davies 1009 Kevin Fall 1010 Olaf Kolkman 1011 Kurtis Lindqvist 1012 David Meyer 1013 David Oran 1014 Eric Rescorla 1015 Dave Thaler 1016 Lixia Zhang 1018 Authors' Addresses 1020 Bernard Aboba 1021 Microsoft Corporation 1022 One Microsoft Way 1023 Redmond, WA 98052 1025 EMail: bernarda@microsoft.com 1026 Phone: +1 425 706 6605 1027 Fax: +1 425 936 7329 1029 Elwyn B. Davies 1030 Consultant 1031 Soham, Cambs 1032 UK 1034 EMail: elwynd@dial.pipex.com 1035 Phone: +44 7889 488 335 1037 Dave Thaler 1038 Microsoft Corporation 1039 One Microsoft Way 1040 Redmond, WA 98052 1042 EMail: dthaler@microsoft.com 1043 Phone: +1 425 703 8835 1045 Intellectual Property Statement 1047 The IETF takes no position regarding the validity or scope of any 1048 Intellectual Property Rights or other rights that might be claimed to 1049 pertain to the implementation or use of the technology described in 1050 this document or the extent to which any license under such rights 1051 might or might not be available; nor does it represent that it has 1052 made any independent effort to identify any such rights. Information 1053 on the procedures with respect to rights in RFC documents can be 1054 found in BCP 78 and BCP 79. 1056 Copies of IPR disclosures made to the IETF Secretariat and any 1057 assurances of licenses to be made available, or the result of an 1058 attempt made to obtain a general license or permission for the use of 1059 such proprietary rights by implementers or users of this 1060 specification can be obtained from the IETF on-line IPR repository at 1061 http://www.ietf.org/ipr. 1063 The IETF invites any interested party to bring to its attention any 1064 copyrights, patents or patent applications, or other proprietary 1065 rights that may cover technology that may be required to implement 1066 this standard. Please address the information to the IETF at 1067 ietf-ipr@ietf.org. 1069 Disclaimer of Validity 1071 This document and the information contained herein are provided on an 1072 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 1073 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 1074 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 1075 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 1076 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 1077 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 1079 Copyright Statement 1081 Copyright (C) The Internet Society (2006). This document is subject 1082 to the rights, licenses and restrictions contained in BCP 78, and 1083 except as set forth therein, the authors retain all their rights. 1085 Acknowledgment 1087 Funding for the RFC Editor function is currently provided by the 1088 Internet Society.