idnits 2.17.1 draft-ietf-bier-mld-01.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 -- The document date (June 29, 2018) is 2127 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-16) exists of draft-ietf-bier-path-mtu-discovery-04 == Outdated reference: A later version (-02) exists of draft-venaas-bier-mtud-01 ** Downref: Normative reference to an Experimental draft: draft-venaas-bier-mtud (ref. 'I-D.venaas-bier-mtud') Summary: 1 error (**), 0 flaws (~~), 3 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group P. Pfister 3 Internet-Draft IJ. Wijnands 4 Intended status: Standards Track S. Venaas 5 Expires: December 31, 2018 Cisco Systems 6 C. Wang 8 Z. Zhang 9 ZTE Corporation 10 M. Stenberg 11 June 29, 2018 13 BIER Ingress Multicast Flow Overlay using Multicast Listener Discovery 14 Protocols 15 draft-ietf-bier-mld-01 17 Abstract 19 This document specifies the ingress part of a multicast flow overlay 20 for BIER networks. Using existing multicast listener discovery 21 protocols, it enables multicast membership information sharing from 22 egress routers, acting as listeners, toward ingress routers, acting 23 as queriers. Ingress routers keep per-egress-router state, used to 24 construct the BIER bit mask associated with IP multicast packets 25 entering the BIER domain. 27 Status of This Memo 29 This Internet-Draft is submitted in full conformance with the 30 provisions of BCP 78 and BCP 79. 32 Internet-Drafts are working documents of the Internet Engineering 33 Task Force (IETF). Note that other groups may also distribute 34 working documents as Internet-Drafts. The list of current Internet- 35 Drafts is at https://datatracker.ietf.org/drafts/current/. 37 Internet-Drafts are draft documents valid for a maximum of six months 38 and may be updated, replaced, or obsoleted by other documents at any 39 time. It is inappropriate to use Internet-Drafts as reference 40 material or to cite them other than as "work in progress." 42 This Internet-Draft will expire on December 31, 2018. 44 Copyright Notice 46 Copyright (c) 2018 IETF Trust and the persons identified as the 47 document authors. All rights reserved. 49 This document is subject to BCP 78 and the IETF Trust's Legal 50 Provisions Relating to IETF Documents 51 (https://trustee.ietf.org/license-info) in effect on the date of 52 publication of this document. Please review these documents 53 carefully, as they describe your rights and restrictions with respect 54 to this document. Code Components extracted from this document must 55 include Simplified BSD License text as described in Section 4.e of 56 the Trust Legal Provisions and are provided without warranty as 57 described in the Simplified BSD License. 59 Table of Contents 61 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 62 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 63 3. Overview . . . . . . . . . . . . . . . . . . . . . . . . . . 3 64 4. Applicability Statement . . . . . . . . . . . . . . . . . . . 4 65 5. Querier and Listener Specifications . . . . . . . . . . . . . 4 66 5.1. Configuration Parameters . . . . . . . . . . . . . . . . 5 67 5.2. MLDv2 instances. . . . . . . . . . . . . . . . . . . . . 5 68 5.2.1. Sending Queries . . . . . . . . . . . . . . . . . . . 6 69 5.2.2. Sending Reports . . . . . . . . . . . . . . . . . . . 6 70 5.2.3. Receiving Queries . . . . . . . . . . . . . . . . . . 7 71 5.2.4. Receiving Reports . . . . . . . . . . . . . . . . . . 7 72 5.3. Packet Forwarding . . . . . . . . . . . . . . . . . . . . 8 73 6. Security Considerations . . . . . . . . . . . . . . . . . . . 8 74 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 8 75 8. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 8 76 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 8 77 9.1. Normative References . . . . . . . . . . . . . . . . . . 8 78 9.2. Informative References . . . . . . . . . . . . . . . . . 9 79 Appendix A. BIER Use Case in Data Centers . . . . . . . . . . . 10 80 A.1. Convention and Terminology . . . . . . . . . . . . . . . 12 81 A.2. BIER in data centers . . . . . . . . . . . . . . . . . . 12 82 A.3. A BIER MLD solution for Virtual Network information . . . 13 83 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 14 85 1. Introduction 87 The Bit Index Explicit Replication (BIER - [RFC8279]) forwarding 88 technique enables IP multicast transport across a BIER domain. When 89 receiving or originating a packet, ingress routers have to construct 90 a bit mask indicating which BIER egress routers located within the 91 same BIER domain will receive the packet. A stateless approach would 92 consist of forwarding all incoming packets toward all egress routers, 93 which would in turn make a forwarding decision based on local 94 information. But any more efficient approach would require ingress 95 routers to keep some state about egress routers multicast membership 96 information, hence requiring state sharing from egress routers toward 97 ingress routers. 99 This document specifies how to use the Multicast Listener Discovery 100 protocol version 2 [RFC3810] (resp. the Internet Group Management 101 protocol version 3 [RFC3376]) as the ingress part of a BIER multicast 102 flow overlay (BIER layering is described in [RFC8279]) for IPv6 103 (resp. IPv4). It enables multicast membership information sharing 104 from egress routers, acting as listeners, toward ingress routers, 105 acting as queriers. Ingress routers keep per-egress-router state, 106 used to construct the BIER bit mask associated with IP multicast 107 packets entering the BIER domain. 109 This specification is applicable to both IP version 4 and version 6. 110 It therefore specifies two separate mechanisms operating 111 independently. For the sake of simplicity, the rest of this document 112 uses IPv6 terminology. It can be applied to IPv4 by replacing 113 'MLDv2' with 'IGMPv3', and following specific requirements when 114 explicitly stated. 116 2. Terminology 118 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 119 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 120 "OPTIONAL" in this document are to be interpreted as described in BCP 121 14 [RFC2119] [RFC8174] when, and only when, they appear in all 122 capitals, as shown here. 124 The terms "Bit-Forwarding Router" (BFR), "Bit-Forwarding Egress 125 Router" (BFER), "Bit-Forwarding Ingress Router" (BFIR), "BFR-id" and 126 "BFR-Prefix" are to be interpreted as described in [RFC8279]. 128 Additionally, the following definitions are used: 130 BIER Multicast Listener Discovery (BMLD): The modified version of 131 MLD specified in this document. 133 BMLD Querier: A BFR implementing the Querier part of this 134 specification. A BMLD Node MAY be both a Querier and a Listener. 136 BMLD Listener: A BFR implementing the Listener part of this 137 specification. A BMLD Node MAY be both a Querier and a Listener. 139 3. Overview 141 This document proposes to use the mechanisms described in MLDv2 in 142 order to enable multicast membership information sharing from BFERs 143 toward BFIRs within a given BIER domain. BMLD queries (resp. 145 reports) are sent over BIER toward all BMLD Nodes (resp. BMLD 146 Queriers) using modified MLDv2 messages which IP destination is set 147 to a configured 'all BMLD Nodes' (resp. 'all BMLD Queriers') IP 148 multicast address. 150 By running MLDv2 instances with per-listener explicit tracking, BMLD 151 Queriers are able to map BMLD Listeners with MLDv2 membership states. 152 This state is then used to construct the set of BFERs associated with 153 each incoming IP multicast data packet. 155 4. Applicability Statement 157 BMLD runs on top of a BIER Layer and provides the ingress part of a 158 BIER multicast flow overlay, i.e, it specifies how BFIRs construct 159 the set of BFERs for each ingress IP multicast data packet. The BFER 160 part of the Multicast Flow Overlay is out of scope of this document. 162 The BIER Layer MUST be able to transport BMLD messages toward all 163 BMLD Queriers and Listeners. Such packets are IP multicast packets 164 with a BFR-Prefix as source address, a multicast destination address, 165 and containing a MLDv2 message. 167 BMLD only requires state to be kept by Queriers, and is therefore 168 more scalable than PIMv2 [RFC7761] in terms of overall state, but is 169 also likely to be less scalable than PIMv2 in terms of the amount of 170 control traffic and the size of the state that is kept by individual 171 routers. 173 This specification is applicable to both IP version 4 and version 6. 174 It therefore specifies two separate mechanisms operating 175 independently. For the sake of simplicity, this document uses IPv6 176 terminology. It can be applied to IPv4 by replacing 'MLDv2' with 177 'IGMPv3', and following specific requirements when explicitly stated. 179 5. Querier and Listener Specifications 181 Routers desiring to receive IP multicast traffic (e.g., for their own 182 use, or for forwarding) MUST behave as BMLD Listeners. Routers 183 receiving IP multicast traffic from outside the BIER domain, or 184 originating multicast traffic, MUST behave as BMLD Queriers. 186 BMLD Queriers (resp. BMLD Listeners) MUST act as MLDv2 Queriers 187 (resp. MLDv2 Listeners) as specified in [RFC3810] unless stated 188 otherwise in this section. 190 5.1. Configuration Parameters 192 Both Queriers and Listeners MUST operate as BFIRs and BFERs within 193 the BIER domain in order to send and receive BMLD messages. They 194 MUST therefore be configured accordingly, as specified in [RFC8279]. 196 All Listeners MUST be configured with an 'all BMLD Queriers' 197 multicast address and the BFR-ids of all the BMLD Queriers. This is 198 used by Listeners to send BMLD reports over BIER toward all Queriers. 199 All Queriers MUST be configured to accept BMLD reports sent to this 200 address. 202 All Queriers MUST be configured with an 'all BMLD Nodes' multicast 203 address and the BFR-ids of all the Queriers and Listeners. This 204 information is used by Queriers to send BMLD queries over BIER toward 205 all BMLD Nodes. All BMLD Nodes MUST be configured to accept BMLD 206 queries sent to this address. 208 It may be cumbersone to configure the exact set of BFR-ids for 209 Queriers and Listeners. One MAY configure the set of BFR-ids to 210 contain any potentially used BFR-id, perhaps having all bit positions 211 set. There is no harm in configuring unused BFR-ids. Configuring 212 the BFR-ids of additional routers would in most cases cause no harm, 213 as a router would drop the BMLD message unless it is configured as a 214 Querier or a Listener. 216 Note that BMLD (unlike MLDv2) makes use of per-instance configured 217 multicast group addresses rather than well-known addresses so that 218 multiple instances of BMLD (using different group addresses) can be 219 run simultaneously within the same BIER domain. Configured group 220 addresses MAY be obtained from allocated IP prefixes using [RFC3306]. 221 One MAY choose to use the well-known MLDv2 addresses in one instance, 222 but different instances MUST use different addresses. 224 IP packets coming from outside of the BIER domain and having a 225 destination address set to the configured 'all BMLD Queriers' or the 226 'all BMLD Nodes' group address MUST be dropped. It is RECOMMENDED 227 that these configured addresses have a limited scope, enforcing this 228 behavior by scope-based filtering on BIER domain's egress interfaces. 230 5.2. MLDv2 instances. 232 BMLD Queriers MUST run a MLDv2 Querier instance with per-host 233 tracking, which means they keep track of the MLDv2 state associated 234 with each BMLD Listener. For that purpose, Listeners are identified 235 by their respective BFR-Prefix, used as IP source address in all BMLD 236 reports. 238 BMLD Listeners MUST run a MLDv2 Listener instance expressing their 239 interest in the multicast traffic they are supposed to receive for 240 local use or forwarding. 242 BMLD Listeners and Queriers MUST NOT run the MLDv1 (IGMPv2 and IGMPv1 243 for IPv4) backward compatibility procedures. 245 5.2.1. Sending Queries 247 BMLD Queries are IP packets sent over BIER by BMLD Queriers: 249 o Toward all BMLD Nodes (i.e., providing to the BIER Layer the BFR- 250 ids of all BMLD Nodes). 252 o Without the IPv6 router alert option [RFC2711] in the hop-by-hop 253 extension header [RFC8200] (or the IPv4 router alert option 254 [RFC2113] for IPv4). 256 o With the IP destination address set to the 'all BMLD Nodes' group 257 address. 259 o With the IP source address set to the BFR-Prefix of the sender. 261 o With a TTL value large enough such that the packet can be received 262 by all BMLD Nodes, depending on the underlying BIER layer (whether 263 it decrements the IP TTL or not) and the size of the network. The 264 default value is 64. 266 5.2.2. Sending Reports 268 BMLD Reports are IP packets sent over BIER by BMLD Listeners: 270 o Toward all BMLD Queriers (i.e., providing to the BIER layer the 271 BFR-ids of all BMLD Queriers). 273 o Without the IPv6 router alert option [RFC2711] in the hop-by-hop 274 extension header [RFC8200] (or the IPv4 router alert option 275 [RFC2113] for IPv4). 277 o With the IP destination address set to the 'all BMLD Queriers' 278 group address. 280 o With the IP source address set to the BFR-Prefix of the sender. 282 o With a TTL value large enough such that the packet can be received 283 by all BMLD Queriers, depending on the underlying BIER layer 284 (whether it decrements the IP TTL or not) and the size of the 285 network. The default value is 64. 287 Since the reports may contain a large number of records, they may 288 become larger than the maximum BIER payload that can be delivered to 289 all the BMLD Queriers. Hence an implementation will need to either 290 use a small default maximum size, allow configuration of a maximum 291 size, or rely on MTU discovery. MTU discovery may be done for a sub- 292 domain using BIER MTU Discovery [I-D.venaas-bier-mtud]) or for the 293 set of BMLD Queriers using Path MTU Discovery 294 [I-D.ietf-bier-path-mtu-discovery]). 296 5.2.3. Receiving Queries 298 BMLD Queriers and Listeners MUST check the destination address of all 299 the IP packets that are received or forwarded over BIER whenever 300 their own BIER bit is set in the packet. If the destination address 301 is equal to the 'all BMLD Nodes' group address the packet is 302 processed as specified in this section. 304 If the IPv6 (resp. IPv4) packet contains an ICMPv6 (resp. IGMP) 305 message of type 'Multicast Listener Query' (resp. of type 'Membership 306 Query'), it is processed by the MLDv2 (resp. IGMPv3) instance run by 307 the BMLD Querier. It MUST be dropped otherwise. 309 During the MLDv2 processing, the packet MUST NOT be checked against 310 the MLDv2 consistency conditions (i.e., the presence of the router 311 alert option, the TTL equaling 1 and, for IPv6 only, the source 312 address being link-local). 314 5.2.4. Receiving Reports 316 BMLD Queriers MUST check the destination address of all the IP 317 packets that are received or forwarded over BIER whenever their own 318 BIER bit is set. If the destination address is equal to the 'all 319 BMLD Queriers' the packet is processed as specified in this section. 321 If the IPv6 (resp. IPv4) packet contains an ICMPv6 (resp. IGMP) 322 message of type 'Multicast Listener Report Message v2' (resp. 323 'Version 3 Membership Report'), it is processed by the MLDv2 (resp. 324 IGMPv3) instance run by the BMLD Querier. It MUST be dropped 325 otherwise. 327 During the MLDv2 processing, the packet MUST NOT be checked against 328 the MLDv2 consistency conditions (i.e., the presence of the router 329 alert option, the TTL equaling 1 and, for IPv6 only, the source 330 address being link-local). 332 5.3. Packet Forwarding 334 BMLD Queriers configure the BIER Layer using the information obtained 335 using BMLD, which associates BMLD Listeners (identified by their BFR- 336 Prefixes) with their respective MLDv2 membership state. 338 More specifically, the MLDv2 state associated with each BMLD Listener 339 is provided to the BIER layer such that whenever a multicast packet 340 enters the BIER domain, if that packet matches the membership 341 information from a BMLD Listener, its BFR-id is added to the set of 342 BFR-ids the packet should be forwarded to by the BIER-Layer. 344 6. Security Considerations 346 BMLD makes use of IP MLDv2 messages transported over BIER in order to 347 configure the BIER Layer of BFIRs. BMLD messages MUST be secured, 348 either by relying on physical or link-layer security, by securing the 349 IP packets (e.g., using IPSec [RFC4301]), or by relying on security 350 features provided by the BIER Layer. 352 Whenever an attacker would be able to spoof the identity of a router, 353 it could: 355 o Redirect undesired traffic toward the spoofed router by 356 subscribing to undesired multicast traffic. 358 o Prevent desired multicast traffic from reaching the spoofed router 359 by unsubscribing to some desired multicast traffic. 361 7. IANA Considerations 363 This specification does not require any action from IANA. 365 8. Acknowledgements 367 Comments concerning this document are very welcome. 369 9. References 371 9.1. Normative References 373 [I-D.ietf-bier-path-mtu-discovery] 374 Mirsky, G., Przygienda, T., and A. Dolganow, "Path Maximum 375 Transmission Unit Discovery (PMTUD) for Bit Index Explicit 376 Replication (BIER) Layer", draft-ietf-bier-path-mtu- 377 discovery-04 (work in progress), June 2018. 379 [I-D.venaas-bier-mtud] 380 Venaas, S., Wijnands, I., Ginsberg, L., and M. Sivakumar, 381 "BIER MTU Discovery", draft-venaas-bier-mtud-01 (work in 382 progress), June 2018. 384 [RFC2113] Katz, D., "IP Router Alert Option", RFC 2113, 385 DOI 10.17487/RFC2113, February 1997, 386 . 388 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 389 Requirement Levels", BCP 14, RFC 2119, 390 DOI 10.17487/RFC2119, March 1997, 391 . 393 [RFC3376] Cain, B., Deering, S., Kouvelas, I., Fenner, B., and A. 394 Thyagarajan, "Internet Group Management Protocol, Version 395 3", RFC 3376, DOI 10.17487/RFC3376, October 2002, 396 . 398 [RFC3810] Vida, R., Ed. and L. Costa, Ed., "Multicast Listener 399 Discovery Version 2 (MLDv2) for IPv6", RFC 3810, 400 DOI 10.17487/RFC3810, June 2004, 401 . 403 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 404 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 405 May 2017, . 407 [RFC8279] Wijnands, IJ., Ed., Rosen, E., Ed., Dolganow, A., 408 Przygienda, T., and S. Aldrin, "Multicast Using Bit Index 409 Explicit Replication (BIER)", RFC 8279, 410 DOI 10.17487/RFC8279, November 2017, 411 . 413 9.2. Informative References 415 [RFC2711] Partridge, C. and A. Jackson, "IPv6 Router Alert Option", 416 RFC 2711, DOI 10.17487/RFC2711, October 1999, 417 . 419 [RFC3306] Haberman, B. and D. Thaler, "Unicast-Prefix-based IPv6 420 Multicast Addresses", RFC 3306, DOI 10.17487/RFC3306, 421 August 2002, . 423 [RFC4301] Kent, S. and K. Seo, "Security Architecture for the 424 Internet Protocol", RFC 4301, DOI 10.17487/RFC4301, 425 December 2005, . 427 [RFC5015] Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano, 428 "Bidirectional Protocol Independent Multicast (BIDIR- 429 PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007, 430 . 432 [RFC7348] Mahalingam, M., Dutt, D., Duda, K., Agarwal, P., Kreeger, 433 L., Sridhar, T., Bursell, M., and C. Wright, "Virtual 434 eXtensible Local Area Network (VXLAN): A Framework for 435 Overlaying Virtualized Layer 2 Networks over Layer 3 436 Networks", RFC 7348, DOI 10.17487/RFC7348, August 2014, 437 . 439 [RFC7365] Lasserre, M., Balus, F., Morin, T., Bitar, N., and Y. 440 Rekhter, "Framework for Data Center (DC) Network 441 Virtualization", RFC 7365, DOI 10.17487/RFC7365, October 442 2014, . 444 [RFC7761] Fenner, B., Handley, M., Holbrook, H., Kouvelas, I., 445 Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent 446 Multicast - Sparse Mode (PIM-SM): Protocol Specification 447 (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March 448 2016, . 450 [RFC8200] Deering, S. and R. Hinden, "Internet Protocol, Version 6 451 (IPv6) Specification", STD 86, RFC 8200, 452 DOI 10.17487/RFC8200, July 2017, 453 . 455 Appendix A. BIER Use Case in Data Centers 457 In current data center virtualization, virtual eXtensible Local Area 458 Network (VXLAN) [RFC7348] is a kind of network virtualization overlay 459 technology which is overlaid between NVEs and is intended for multi- 460 tenancy data center networks, whose reference architecture is 461 illustrated as per Figure 1. 463 +--------+ +--------+ 464 | Tenant +--+ +----| Tenant | 465 | System | | (') | System | 466 +--------+ | ................ ( ) +--------+ 467 | +-+--+ . . +--+-+ (_) 468 | | NVE|--. .--| NVE| | 469 +--| | . . | |---+ 470 +-+--+ . . +--+-+ 471 / . . 472 / . L3 Overlay . +--+-++--------+ 473 +--------+ / . Network . | NVE|| Tenant | 474 | Tenant +--+ . .--| || System | 475 | System | . . +--+-++--------+ 476 +--------+ ................ 478 Figure 1: NVO3 Architecture 480 And there are two kinds of most common methods about how to forward 481 BUM packets in this virtualization overlay network. One is using PIM 482 as underlay multicast routing protocol to build explicit multicast 483 distribution tree, such as PIM-SM [RFC7761] or PIM-BIDIR [RFC5015] 484 multicast routing protocol. Then, when BUM packets arrive at NVE, it 485 requires NVE to have a mapping between the VXLAN Network Identifier 486 and the IP multicast group. According to the mapping, NVE can 487 encapsulate BUM packets in a multicast packet which group address is 488 the mapping IP multicast group address and steer them through 489 explicit multicast distribution tree to the destination NVEs. This 490 method has two serious drawbacks. It need the underlay network 491 supports complicated multicast routing protocol and maintains 492 multicast related per-flow state in every transit nodes. What is 493 more, how to configure the ratio of the mapping between VNI and IP 494 multicast group is also an issue. If the ratio is 1:1, there should 495 be 16M multicast groups in the underlay network at maximum to map to 496 the 16M VNIs, which is really a significant challenge for the data 497 center devices. If the ratio is n:1, it would result in inefficiency 498 bandwidth utilization which is not optimal in data center networks. 500 The other method is using ingress replication to require each NVE to 501 create a mapping between the VXLAN Network Identifier and the remote 502 addresses of NVEs which belong to the same virtual network. When NVE 503 receives BUM traffic from the attached tenant, NVE can encapsulate 504 these BUM packets in unicast packets and replicate them and tunnel 505 them to different remote NVEs respectively. Although this method can 506 eliminate the burden of running multicast protocol in the underlay 507 network, it has a significant disadvantage: large waste of bandwidth, 508 especially in big-sized data center where there are many receivers. 510 BIER [RFC8279] is an architecture that provides optimal multicast 511 forwarding through a "BIER domain" without requiring intermediate 512 routers to maintain any multicast related per-flow state. BIER also 513 does not require any explicit tree-building protocol for its 514 operation. A multicast data packet enters a BIER domain at a "Bit- 515 Forwarding Ingress Router" (BFIR), and leaves the BIER domain at one 516 or more "Bit-Forwarding Egress Routers" (BFERs). The BFIR router 517 adds a BIER header to the packet. The BIER header contains a bit- 518 string in which each bit represents exactly one BFER to forward the 519 packet to. The set of BFERs to which the multicast packet needs to 520 be forwarded is expressed by setting the bits that correspond to 521 those routers in the BIER header. Specifically, for BIER-TE, the 522 BIER header may also contain a bit-string in which each bit indicates 523 the link the flow passes through. 525 The following sub-sections try to propose how to take full advantage 526 of overlay multicast protocol to carry virtual network information, 527 and create a mapping between the virtual network information and the 528 bit-string to implement BUM services in data centers. 530 A.1. Convention and Terminology 532 The terms about NVO3 are defined in [RFC7365]. The most common 533 terminology used in this appendix is listed below. 535 NVE: Network Virtualization Edge, which is the entity that 536 implements the overlay functionality. An NVE resides at the 537 boundary between a Tenant System and the overlay network. 539 VXLAN: Virtual eXtensible Local Area Network 541 VNI: VXLAN Network Identifier 543 Virtal Network Context Identifier: Field in an overlay encapsulation 544 header that identifies the specific VN the packet belongs to. 546 A.2. BIER in data centers 548 This section tries to describe how to use BIER as an optimal scheme 549 to forward the broadcast, unknown and multicast (BUM) packets when 550 they arrive at the ingress NVE in data centers. 552 The principle of using BIER to forward BUM traffic is that: firstly, 553 it requires each ingress NVE to have a mapping between the Virtual 554 Network Context Identifier and the bit-string in which each bit 555 represents exactly one egress NVE to forward the packet to. And 556 then, when receiving the BUM traffic, the BFIR/Ingree NVE maps the 557 receiving BUM traffic to the mapping bit-string, encapsulates the 558 BIER header, and forwards the encapsulated BUM traffic into the BIER 559 domain to the other BFERs/Egress NVEs indicated by the bit-string. 561 Furthermore, as for how each ingress NVE knows the other egress NVEs 562 that belong to the same virtual network and creates the mapping is 563 the main issue discussed below. Basically, BIER Multicast Listener 564 Discovery is an overlay solution to support ingress routers to keep 565 per-egress-router state to construct the BIER bit-string associated 566 with IP multicast packets entering the BIER domain. The following 567 section tries to extend BIER MLD to carry virtual network 568 information(such as Virtual Network Context identifier), and 569 advertise them between NVEs. When each NVE receive these 570 information, they create the mapping between the virtual network 571 information and the bit-string representing the other NVEs belonged 572 to the same virtual network. 574 A.3. A BIER MLD solution for Virtual Network information 576 The BIER MLD solution allows having multiple MLD instances by having 577 unique pairs of BMLD Nodes and BMLD Querier addresses for each 578 instance. Assume for now that we have a unique instance per VNI and 579 that all BMLD routers are using the same mapping between VNIs and 580 BMLD address pairs. Also for each VNI there is a multicast group 581 used for encapsulation of BUM traffic over BIER. This group may 582 potentially be shared by some or all of the VNIs. 584 Each NVE acquires the Virtual Network information, and advertises 585 this Virtual Network information to other NVEs through the MLD 586 messages. For a given VNI it sends BMLD reports to the BMLD nodes 587 address used for that VNI, for the group used for delivering BUM 588 traffic for that VNI. This allows all NVE routers to know which 589 other NVE routers have interest in BUM traffic for a particular VNI. 590 If one attached virtual network is migrated, the NVE will withdraw 591 the Virtual Network information by sending an unsolicited BMLD 592 report. Note that NVEs also respond to periodic queries to BMLD 593 Nodes addresses corresponding to VNIs for which they have interest. 595 When ingress NVE receives the Virtual Network information 596 advertisement message, it builds a mapping between the receiving 597 Virtual Network Context Identifier in this message and the bit-string 598 in which each bit represents one egress NVE who sends the same 599 Virtual Network information. Subsequently, once this ingress NVE 600 receives some other MLD advertisements which include the same Virtual 601 Network information from some other NVEs , it updates the bit-string 602 in the mapping and adds the corresponding sending NVE to the updated 603 bit-string. Once the ingress NVE removes one virtual network, it 604 will delete the mapping corresponding to this virtual network as well 605 as send withdraw message to other NVEs. 607 After finishing the above interaction of MLD messages, each ingress 608 NVE knows where the other egress NVEs are in the same virtual 609 network. When receiving BUM traffic from the attached virtual 610 network, each ingress NVE knows exactly how to encapsulate this 611 traffic and where to forward them to. 613 This can be used in both IPv4 network and IPv6 network. In IPv4, 614 IGMP protocol does the similar extension for carrying Virtual Network 615 information TLV in Version 2 membership report message. 617 Note that it is possible to have multiple VNIs map to the same pair 618 of BMLD addresses. Provided VNIs that map to the same BMLD address 619 uses different multicast groups for encapsulation, this is not a 620 problem, because each instance is tracking interest for each 621 multicast group separately. If multiple VNIs map to the same pair 622 and the multicast group used is not unique, some NVEs may receive BUM 623 traffic for which they are not interested. An NVE would drop packets 624 for an unknown VNI, but it means wasting some bandwidth and 625 processing. This is similar to the non-BIER case where there is not 626 a unique multicast group for encapsulation. The improvement offered 627 by using BMLD is by using multiple instance, hence reducing the 628 problems caused by using the same transport group for multiple VNIs. 630 Authors' Addresses 632 Pierre Pfister 633 Cisco Systems 634 Paris 635 France 637 Email: pierre.pfister@darou.fr 639 IJsbrand Wijnands 640 Cisco Systems 641 De Kleetlaan 6a 642 Diegem 1831 643 Belgium 645 Email: ice@cisco.com 646 Stig Venaas 647 Cisco Systems 648 Tasman Drive 649 San Jose, CA 95134 650 USA 652 Email: stig@cisco.com 654 Cui(Linda) Wang 656 Email: lindawangjoy@gmail.com 658 Zheng(Sandy) Zhang 659 ZTE Corporation 660 No.50 Software Avenue, Yuhuatai District 661 Nanjing, CA 662 China 664 Email: zhang.zheng@zte.com.cn 666 Markus Stenberg 667 Helsinki 00930 668 Finland 670 Email: markus.stenberg@iki.fi