idnits 2.17.1 draft-ietf-pals-ethernet-cw-05.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year (Using the creation date from RFC4448, updated by this document, for RFC5378 checks: 2002-09-04) -- 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 (April 16, 2018) is 2202 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) No issues found here. Summary: 0 errors (**), 0 flaws (~~), 1 warning (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 PALS Working Group S. Bryant 3 Internet-Draft A. Malis 4 Updates: 4448 (if approved) Huawei 5 Intended status: Standards Track I. Bagdonas 6 Expires: October 18, 2018 Equinix 7 April 16, 2018 9 Use of Ethernet Control Word RECOMMENDED 10 draft-ietf-pals-ethernet-cw-05 12 Abstract 14 The pseudowire (PW) encapsulation of Ethernet, as defined in RFC 15 4448, specifies that the use of the control word (CW) is optional. 16 In the absence of the CW an Ethernet pseudowire packet can be 17 misidentified as an IP packet by a label switching router (LSR). 18 This in turn may lead to the selection of the wrong equal-cost-multi- 19 path (ECMP) path for the packet, leading in turn to the misordering 20 of packets. This problem has become more serious due to the 21 deployment of equipment with Ethernet MAC addresses that start with 22 0x4 or 0x6. The use of the Ethernet PW CW addresses this problem. 23 This document recommends the use of the Ethernet pseudowire control 24 word in all but exceptional circumstances. 26 This document updates RFC 4448. 28 Status of This Memo 30 This Internet-Draft is submitted in full conformance with the 31 provisions of BCP 78 and BCP 79. 33 Internet-Drafts are working documents of the Internet Engineering 34 Task Force (IETF). Note that other groups may also distribute 35 working documents as Internet-Drafts. The list of current Internet- 36 Drafts is at http://datatracker.ietf.org/drafts/current/. 38 Internet-Drafts are draft documents valid for a maximum of six months 39 and may be updated, replaced, or obsoleted by other documents at any 40 time. It is inappropriate to use Internet-Drafts as reference 41 material or to cite them other than as "work in progress." 43 This Internet-Draft will expire on October 18, 2018. 45 Copyright Notice 47 Copyright (c) 2018 IETF Trust and the persons identified as the 48 document authors. All rights reserved. 50 This document is subject to BCP 78 and the IETF Trust's Legal 51 Provisions Relating to IETF Documents 52 (http://trustee.ietf.org/license-info) in effect on the date of 53 publication of this document. Please review these documents 54 carefully, as they describe your rights and restrictions with respect 55 to this document. Code Components extracted from this document must 56 include Simplified BSD License text as described in Section 4.e of 57 the Trust Legal Provisions and are provided without warranty as 58 described in the Simplified BSD License. 60 Table of Contents 62 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 63 2. Specification of Requirements . . . . . . . . . . . . . . . . 3 64 3. Background . . . . . . . . . . . . . . . . . . . . . . . . . 3 65 4. Recommendation . . . . . . . . . . . . . . . . . . . . . . . 5 66 5. Equal Cost Multi-path (ECMP) . . . . . . . . . . . . . . . . 5 67 6. Mitigations . . . . . . . . . . . . . . . . . . . . . . . . . 6 68 7. Operational Considerations . . . . . . . . . . . . . . . . . 6 69 8. Security Considerations . . . . . . . . . . . . . . . . . . . 7 70 9. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 7 71 10. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 7 72 11. References . . . . . . . . . . . . . . . . . . . . . . . . . 7 73 11.1. Normative References . . . . . . . . . . . . . . . . . . 7 74 11.2. Informative References . . . . . . . . . . . . . . . . . 8 75 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 8 77 1. Introduction 79 The pseudowire(PW) encapsulation of Ethernet, as defined in 80 [RFC4448], specifies that the use of the control word (CW) is 81 optional. It is common for label switching routers (LSRs) to search 82 past the end of the label stack to determine whether the payload is 83 an IP packet, and if the payload is an IP packet, to select the next 84 hop based on the so called "five-tuple" (IP source address, IP 85 destination address, protocol/next-header, transport layer source 86 port and transport layer destination port). In the absence of a PW 87 CW an Ethernet pseudowire packet can be misidentified as an IP packet 88 by a label switching router (LSR) selecting the equal-cost-multi-path 89 (ECMP) path based on the five-tuple. This in turn may lead to the 90 selection of the wrong ECMP path for the packet, leading in turn to 91 the misordering of packets. Further discussion of this topic is 92 published in [RFC4928]. 94 Flow misordering can also happen in a single path scenario when 95 traffic classification and differential forwarding treatment 96 mechanisms are in use. This occurs when a forwarder incorrectly 97 assumes that the packet is IP and applies forwarding policy based on 98 fields in the PW payload. 100 This problem has recently become more serious for a number of 101 reasons. Firstly, due to the deployment of equipment with Ethernet 102 MAC addresses that start with 0x4 or 0x6 assigned by the IEEE 103 Registration Authority Committee (RAC). Secondly, concerns over 104 privacy have led to the use of MAC address randomization which 105 assigns local MAC addresses randomly for privacy. Random assignment 106 results in addresses starting with one of these two values one time 107 in eight. 109 The use of the Ethernet PW CW addresses this problem. 111 This document recommends the use of the Ethernet pseudowire control 112 word in all but exceptional circumstances. 114 2. Specification of Requirements 116 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 117 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 118 document are to be interpreted as described in RFC 2119 [RFC2119]. 120 3. Background 122 Ethernet pseudowire encapsulation is specified in [RFC4448]. In 123 particular the reader is drawn to section 4.6, part of which is 124 quoted below for the convenience of the reader: 126 "The control word defined in this section is based on the Generic 127 PW MPLS Control Word as defined in [RFC4385]. It provides the 128 ability to sequence individual frames on the PW, avoidance of 129 equal-cost multiple-path load-balancing (ECMP) [RFC2992], and 130 Operations and Management (OAM) mechanisms including VCCV 131 [RFC5085]. 133 "[RFC4385] states, "If a PW is sensitive to packet misordering 134 and is being carried over an MPLS PSN that uses the contents 135 of the MPLS payload to select the ECMP path, it MUST employ a 136 mechanism which prevents packet misordering." This is necessary 137 because ECMP implementations may examine the first nibble after 138 the MPLS label stack to determine whether the labeled packet 139 is IP or not. Thus, if the source MAC address of an Ethernet 140 frame carried over the PW without a control word present begins 141 with 0x4 or 0x6, it could be mistaken for an IPv4 or IPv6 142 packet. This could, depending on the configuration and 143 topology of the MPLS network, lead to a situation where all 144 packets for a given PW do not follow the same path. This may 145 increase out-of-order frames on a given PW, or cause OAM packets 146 to follow a different path than actual traffic (see 147 Section 4.4.3, "Frame Ordering"). 149 "The features that the control word provides may not be needed 150 for a given Ethernet PW. For example, ECMP may not be present 151 or active on a given MPLS network, strict frame sequencing may 152 not be required, etc. If this is the case, the control word 153 provides little value and is therefore optional. Early Ethernet 154 PW implementations have been deployed that do not include a 155 control word or the ability to process one if present. To 156 aid in backwards compatibility, future implementations MUST 157 be able to send and receive frames without the control word 158 present." 160 At the time when pseudowires were first deployed, some equipment of 161 commercial significance was unable to process the Ethernet Control 162 Word. In addition, at that time it was considered that no Ethernet 163 MAC address had been issued by the IEEE Registration Authority 164 Committee (RAC) that starts with 0x4 or 0x6, and thus it was thought 165 to be safe to deploy Ethernet PWs without the CW. 167 Since that time the RAC has issued Ethernet MAC addresses start with 168 0x4 or 0x6 and thus the assumption that in practical networks there 169 would be no confusion between an Ethernet PW packet without the CW 170 and an IP packet is no longer correct. 172 Possibly through the use of unauthorized Ethernet MAC addresses, this 173 assumption has been unsafe for a while, leading some equipment 174 vendors to implement more complex, proprietary, methods to 175 discriminate between Ethernet PW packets and IP packets. Such 176 mechanisms rely on the heuristics of examining the transit packets in 177 trying to find out the exact payload type of the packet and cannot be 178 reliable due to the random nature of the payload carried within such 179 packets. 181 A recent posting on the Nanog email list has highlighted this 182 problem: 184 https://mailman.nanog.org/pipermail/nanog/2016-December/089395.html 186 RFC EDITOR Please delete this paragraph. 187 Kramdown does not include references when they are only found in 188 literal text so I include them here: [RFC4385] [RFC2992] [RFC5085] as 189 a fixup. 191 4. Recommendation 193 The ambiguity between an MPLS payload that is an Ethernet PW and one 194 that is an IP packet is resolved when the Ethernet PW control word is 195 used. This document updates RFC4448 [RFC4448] to state that where 196 both the ingress PE and the egress PE support the Ethernet pseudowire 197 control word, then the CW MUST be used. 199 Where the application of ECMP to an Ethernet PW traffic is required, 200 and where both the ingress and the egress PEs support RFC6790 201 [RFC6790] (ELI) or both the ingress and the egress PEs support 202 RFC6391 [RFC6391] (FAT PW), then either method may be used. The use 203 of both methods on the same PW is not normally necessary and should 204 be avoided unless circumstances require it. In the case of multi- 205 segment PWs, if ELI/EL is used then it SHOULD be used on every 206 segment of the PW. The method by which usage of ELI/EL on every 207 segment is guaranteed is out of scope of this document. 209 5. Equal Cost Multi-path (ECMP) 211 Where the volume of traffic on an Ethernet PW is such that ECMP is 212 required then one of two methods may be used: 214 o Flow-Aware Transport (FAT) of Pseudowires over an MPLS Packet 215 Switched Network specified in [RFC6391], or 217 o LSP entropy labels specified [RFC6790] 219 RFC6391 works by increasing the entropy of the bottom of stack label. 220 It requires that both the ingress and egress provider edge (PE)s 221 support this feature. It also requires that sufficient LSRs on the 222 LSP between the ingress and egress PE be able to select an 223 ECMP path on an MPLS packet with the resultant stack depth. 225 RFC6790 works by including an entropy value in the LSP part of the 226 label stack. This requires that the Ingress and Egress PEs support 227 the insertion and removal of the entropy label (EL) and the entropy 228 label indicator, and that sufficient LSRs on the LSP are able to 229 preform ECMP based on the EL. 231 In both cases there are considerations in getting Operations, 232 Administration, and Maintenance (OAM) packets to follow the same path 233 as a data packet. This is described in detail section 7 of 234 [RFC6391], and section 6 of RFC6790. However in both cases the 235 situation is improved compared to the ECMP behavior in the case where 236 the Ethernet PW CW was not used, since there is currently no known 237 method of getting a PW OAM packet to follow the same path as a PW 238 data packet subjected to ECMP based on the five tuple of the IP 239 payload. 241 The PW label is pushed before the LSP label. As the EL/ELI labels 242 are part of the LSP layer rather than part of the PW layer, they are 243 pushed after the PW label has been pushed. 245 6. Mitigations 247 Where it is not possible to use the Ethernet PW CW, the effects of 248 ECMP can be disabled by carrying the PW over a traffic engineered 249 path that does not subject the payload to load balancing (for example 250 [RFC3209]). However such paths may be subjected to link bundle load 251 balancing and of course the single LSP has to carry the full PW load. 253 7. Operational Considerations 255 CW presence on the PW is controlled by the configuration and may be 256 subject to default operational mode of not being enabled. Care needs 257 to be taken to ensure that software that implements this 258 recommendation does not depend on existing configuration setting that 259 prevents the use of control word. It is recommended that platform 260 software emits a rate limited message indicating that CW can be used 261 but is disabled due to existing configuration. 263 Instead of including a payload type in the packet, MPLS relies on the 264 control plane to signal the payload type that follows the bottom of 265 the label stack. Some LSRs attempt to deduce the packet type by MPLS 266 payload inspection, in some cases looking past the PW CW. If the 267 payload appears to be IP or IP carried in an Ethernet header they 268 perform an ECMP calculation based on what they assume to be the five 269 tuple fields. However deduction of the payload type in this way is 270 not an exact science, and where a packet that is not IP is mistaken 271 for an IP packet the result can be packets delivered out of order. 272 Misordering of this type can be difficult for an operator to 273 diagnose. 274 Operators should be aware when enabling capability that allows 275 information gleaned from packet inspection past the PW CW to be used 276 in any ECMP calculation, that this may cause Ethernet frames to be 277 delivered out of order despite the presence of the CW. 279 8. Security Considerations 281 This document expresses a preference for one existing and widely 282 deployed Ethernet PW encapsulation over another. These methods have 283 identical security considerations, which are discussed in [RFC4448]. 284 This document introduces no additional security issues. 286 9. IANA Considerations 288 This document makes no IANA requests. 290 10. Acknowledgments 292 The authors thank Job Snijders for drawing attention to this problem. 293 The authors also thank Pat Thaler for clarifying the matter of local 294 MAC address assignment. We thank Sasha Vainshtein for his valuable 295 review comments. 297 11. References 299 11.1. Normative References 301 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 302 Requirement Levels", BCP 14, RFC 2119, 303 DOI 10.17487/RFC2119, March 1997, . 306 [RFC4385] Bryant, S., Swallow, G., Martini, L., and D. McPherson, 307 "Pseudowire Emulation Edge-to-Edge (PWE3) Control Word for 308 Use over an MPLS PSN", RFC 4385, DOI 10.17487/RFC4385, 309 February 2006, . 311 [RFC4448] Martini, L., Ed., Rosen, E., El-Aawar, N., and G. Heron, 312 "Encapsulation Methods for Transport of Ethernet over MPLS 313 Networks", RFC 4448, DOI 10.17487/RFC4448, April 2006, 314 . 316 [RFC4928] Swallow, G., Bryant, S., and L. Andersson, "Avoiding Equal 317 Cost Multipath Treatment in MPLS Networks", BCP 128, 318 RFC 4928, DOI 10.17487/RFC4928, June 2007, 319 . 321 [RFC6391] Bryant, S., Ed., Filsfils, C., Drafz, U., Kompella, V., 322 Regan, J., and S. Amante, "Flow-Aware Transport of 323 Pseudowires over an MPLS Packet Switched Network", 324 RFC 6391, DOI 10.17487/RFC6391, November 2011, 325 . 327 [RFC6790] Kompella, K., Drake, J., Amante, S., Henderickx, W., and 328 L. Yong, "The Use of Entropy Labels in MPLS Forwarding", 329 RFC 6790, DOI 10.17487/RFC6790, November 2012, 330 . 332 11.2. Informative References 334 [RFC2992] Hopps, C., "Analysis of an Equal-Cost Multi-Path 335 Algorithm", RFC 2992, DOI 10.17487/RFC2992, November 2000, 336 . 338 [RFC3209] Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V., 339 and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP 340 Tunnels", RFC 3209, DOI 10.17487/RFC3209, December 2001, 341 . 343 [RFC5085] Nadeau, T., Ed. and C. Pignataro, Ed., "Pseudowire Virtual 344 Circuit Connectivity Verification (VCCV): A Control 345 Channel for Pseudowires", RFC 5085, DOI 10.17487/RFC5085, 346 December 2007, . 348 Authors' Addresses 350 Stewart Bryant 351 Huawei 353 Email: stewart.bryant@gmail.com 355 Andrew G Malis 356 Huawei 358 Email: agmalis@gmail.com 359 Ignas Bagdonas 360 Equinix 362 Email: ibagdona.ietf@gmail.com>