idnits 2.17.1 draft-ietf-bess-dci-evpn-overlay-10.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There are 16 instances of lines with control characters in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (March 2, 2018) is 2246 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references to lower-maturity documents in RFCs) ** Downref: Normative reference to an Informational RFC: RFC 7041 == Outdated reference: A later version (-22) exists of draft-ietf-idr-tunnel-encaps-08 == Outdated reference: A later version (-12) exists of draft-ietf-bess-evpn-overlay-11 Summary: 2 errors (**), 0 flaws (~~), 3 warnings (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 BESS Workgroup J. Rabadan (Ed.) 3 Internet Draft S. Sathappan 4 Intended status: Standards Track W. Henderickx 5 Nokia 7 A. Sajassi 8 Cisco 10 J. Drake 11 Juniper 13 Expires: September 3, 2018 March 2, 2018 15 Interconnect Solution for EVPN Overlay networks 16 draft-ietf-bess-dci-evpn-overlay-10 18 Abstract 20 This document describes how Network Virtualization Overlays (NVO) can 21 be connected to a Wide Area Network (WAN) in order to extend the 22 layer-2 connectivity required for some tenants. The solution analyzes 23 the interaction between NVO networks running Ethernet Virtual Private 24 Networks (EVPN) and other L2VPN technologies used in the WAN, such as 25 Virtual Private LAN Services (VPLS), VPLS extensions for Provider 26 Backbone Bridging (PBB-VPLS), EVPN or PBB-EVPN. It also describes how 27 the existing technical specifications apply to the Interconnection 28 and extends the EVPN procedures needed in some cases. In particular, 29 this document describes how EVPN routes are processed on Gateways 30 (GWs) that interconnect EVPN-Overlay and EVPN-MPLS networks, as well 31 as the Interconnect Ethernet Segment (I-ES) to provide multi-homing, 32 and the use of the Unknown MAC route to avoid MAC scale issues on 33 Data Center Network Virtualization Edge (NVE) devices. 35 Status of this Memo 37 This Internet-Draft is submitted in full conformance with the 38 provisions of BCP 78 and BCP 79. 40 Internet-Drafts are working documents of the Internet Engineering 41 Task Force (IETF), its areas, and its working groups. Note that 42 other groups may also distribute working documents as Internet- 43 Drafts. 45 Internet-Drafts are draft documents valid for a maximum of six months 46 and may be updated, replaced, or obsoleted by other documents at any 47 time. It is inappropriate to use Internet-Drafts as reference 48 material or to cite them other than as "work in progress." 50 The list of current Internet-Drafts can be accessed at 51 http://www.ietf.org/ietf/1id-abstracts.txt 53 The list of Internet-Draft Shadow Directories can be accessed at 54 http://www.ietf.org/shadow.html 56 This Internet-Draft will expire on September 3, 2018. 58 Copyright Notice 60 Copyright (c) 2018 IETF Trust and the persons identified as the 61 document authors. All rights reserved. 63 This document is subject to BCP 78 and the IETF Trust's Legal 64 Provisions Relating to IETF Documents 65 (http://trustee.ietf.org/license-info) in effect on the date of 66 publication of this document. Please review these documents 67 carefully, as they describe your rights and restrictions with respect 68 to this document. Code Components extracted from this document must 69 include Simplified BSD License text as described in Section 4.e of 70 the Trust Legal Provisions and are provided without warranty as 71 described in the Simplified BSD License. 73 Table of Contents 75 1. Conventions and Terminology . . . . . . . . . . . . . . . . . . 3 76 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 77 3. Decoupled Interconnect solution for EVPN overlay networks . . . 6 78 3.1. Interconnect requirements . . . . . . . . . . . . . . . . . 7 79 3.2. VLAN-based hand-off . . . . . . . . . . . . . . . . . . . . 8 80 3.3. PW-based (Pseudowire-based) hand-off . . . . . . . . . . . 8 81 3.4. Multi-homing solution on the GWs . . . . . . . . . . . . . 9 82 3.5. Gateway Optimizations . . . . . . . . . . . . . . . . . . . 9 83 3.5.1. MAC Address Advertisement Control . . . . . . . . . . . 9 84 3.5.2. ARP/ND flooding control . . . . . . . . . . . . . . . . 10 85 3.5.3. Handling failures between GW and WAN Edge routers . . . 11 86 4. Integrated Interconnect solution for EVPN overlay networks . . 11 87 4.1. Interconnect requirements . . . . . . . . . . . . . . . . . 12 88 4.2. VPLS Interconnect for EVPN-Overlay networks . . . . . . . . 13 89 4.2.1. Control/Data Plane setup procedures on the GWs . . . . 13 90 4.2.2. Multi-homing procedures on the GWs . . . . . . . . . . 14 91 4.3. PBB-VPLS Interconnect for EVPN-Overlay networks . . . . . . 14 92 4.3.1. Control/Data Plane setup procedures on the GWs . . . . 14 93 4.3.2. Multi-homing procedures on the GWs . . . . . . . . . . 15 94 4.4. EVPN-MPLS Interconnect for EVPN-Overlay networks . . . . . 15 95 4.4.1. Control Plane setup procedures on the GWs . . . . . . . 15 96 4.4.2. Data Plane setup procedures on the GWs . . . . . . . . 17 97 4.4.3. Multi-homing procedure extensions on the GWs . . . . . 18 98 4.4.4. Impact on MAC Mobility procedures . . . . . . . . . . . 20 99 4.4.5. Gateway optimizations . . . . . . . . . . . . . . . . . 20 100 4.4.6. Benefits of the EVPN-MPLS Interconnect solution . . . . 21 101 4.5. PBB-EVPN Interconnect for EVPN-Overlay networks . . . . . . 22 102 4.5.1. Control/Data Plane setup procedures on the GWs . . . . 22 103 4.5.2. Multi-homing procedures on the GWs . . . . . . . . . . 22 104 4.5.3. Impact on MAC Mobility procedures . . . . . . . . . . . 23 105 4.5.4. Gateway optimizations . . . . . . . . . . . . . . . . . 23 106 4.6. EVPN-VXLAN Interconnect for EVPN-Overlay networks . . . . . 23 107 4.6.1. Globally unique VNIs in the Interconnect network . . . 24 108 4.6.2. Downstream assigned VNIs in the Interconnect network . 24 109 5. Security Considerations . . . . . . . . . . . . . . . . . . . . 25 110 6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 26 111 7. References . . . . . . . . . . . . . . . . . . . . . . . . . . 26 112 7.1. Normative References . . . . . . . . . . . . . . . . . . . 26 113 7.2. Informative References . . . . . . . . . . . . . . . . . . 27 114 8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . 28 115 9. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 28 116 10. Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 29 118 1. Conventions and Terminology 120 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 121 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 122 "OPTIONAL" in this document are to be interpreted as described in BCP 123 14 [RFC2119] [RFC8174] when, and only when, they appear in all 124 capitals, as shown here. 126 AC: Attachment Circuit. 128 ARP: Address Resolution Protocol. 130 BUM: refers to Broadcast, Unknown unicast and Multicast traffic. 132 CE: Customer Equipment. 134 CFM: Connectivity Fault Management. 136 DC and DCI: Data Center and Data Center Interconnect. 138 DC RR(s) and WAN RR(s): refers to the Data Center and Wide Area 139 Network Route Reflectors, respectively. 141 DF and NDF: Designated Forwarder and Non-Designated Forwarder. 143 EVPN: Ethernet Virtual Private Network, as in [RFC7432]. 145 EVI: EVPN Instance. 147 EVPN Tunnel binding: refers to a tunnel to a remote PE/NVE for a 148 given EVI. Ethernet packets in these bindings are encapsulated with 149 the Overlay or MPLS encapsulation and the EVPN label at the bottom of 150 the stack. 152 ES and vES: Ethernet Segment and virtual Ethernet Segment. 154 ESI: Ethernet Segment Identifier. 156 GW: Gateway or Data Center Gateway. 158 I-ES and I-ESI: Interconnect Ethernet Segment and Interconnect 159 Ethernet Segment Identifier. An I-ES is defined on the GWs for multi- 160 homing to/from the WAN. 162 MAC-VRF: refers to an EVI instance in a particular node. 164 MP2P and LSM tunnels: refer to Multi-Point to Point and Label 165 Switched Multicast tunnels. 167 ND: Neighbor Discovery protocol. 169 NVE: Network Virtualization Edge. 171 NVGRE: Network Virtualization using Generic Routing Encapsulation. 173 NVO: refers to Network Virtualization Overlays. 175 OAM: Operations and Maintenance. 177 PBB: Provider Backbone Bridging. 179 PE: Provider Edge. 181 PW: Pseudowire. 183 RD: Route-Distinguisher. 185 RT: Route-Target. 187 S/C-TAG: refers to a combination of Service Tag and Customer Tag in a 188 802.1Q frame. 190 TOR: Top-Of-Rack switch. 192 UMR: Unknown MAC Route. 194 VNI/VSID: refers to VXLAN/NVGRE virtual identifiers. 196 VPLS: Virtual Private LAN Service. 198 VSI: Virtual Switch Instance or VPLS instance in a particular PE. 200 VXLAN: Virtual eXtensible LAN. 202 2. Introduction 204 [EVPN-Overlays] discusses the use of Ethernet Virtual Private 205 Networks (EVPN) [RFC7432] as the control plane for Network 206 Virtualization Overlays (NVO), where VXLAN [RFC7348], NVGRE [RFC7637] 207 or MPLS over GRE [RFC4023] can be used as possible data plane 208 encapsulation options. 210 While this model provides a scalable and efficient multi-tenant 211 solution within the Data Center, it might not be easily extended to 212 the Wide Area Network (WAN) in some cases due to the requirements and 213 existing deployed technologies. For instance, a Service Provider 214 might have an already deployed Virtual Private LAN Service (VPLS) 215 [RFC4761][RFC4762], VPLS extensions for Provider Backbone Bridging 216 (PBB-VPLS) [RFC7041], EVPN [RFC7432] or PBB-EVPN [RFC7623] network 217 that has to be used to interconnect Data Centers and WAN VPN users. A 218 Gateway (GW) function is required in these cases. In fact, [EVPN- 219 Overlays] discusses two main Data Center Interconnect solution 220 groups: "DCI using GWs" and "DCI using ASBRs". This document 221 specifies the solutions that correspond to the "DCI using GWs" group. 223 It is assumed that the NVO Gateway (GW) and the WAN Edge functions 224 can be decoupled in two separate systems or integrated into the same 225 system. The former option will be referred as "Decoupled Interconnect 226 solution" throughout the document, whereas the latter one will be 227 referred as "Integrated Interconnect solution". 229 The specified procedures are local to the redundant GWs connecting a 230 DC to the WAN. The document does not preclude any combination across 231 different DCs for the same tenant. For instance, a "Decoupled" 232 solution can be used in GW1 and GW2 (for DC1) and an "Integrated" 233 solution can be used in GW3 and GW4 (for DC2). 235 While the Gateways and WAN PEs use existing specifications in some 236 cases, the document also defines extensions that are specific to DCI. 237 In particular, those extensions are: 239 o The Interconnect Ethernet Segment (I-ES), an Ethernet Segment that 240 can be associated to a set of PWs or other tunnels. I-ES defined in 241 this document is not associated with a set of Ethernet links, as 242 per [RFC7432], but rather with a set of virtual tunnels (e.g., a 243 set of PWs). This set of virtual tunnels is referred to as vES 244 [VIRTUAL-ES]. 246 o The use of the Unknown MAC route in a DCI scenario. 248 o The processing of EVPN routes on Gateways with MAC-VRFs connecting 249 EVPN-Overlay and EVPN-MPLS networks, or EVPN-Overlay and EVPN- 250 Overlay networks. 252 3. Decoupled Interconnect solution for EVPN overlay networks 254 This section describes the interconnect solution when the GW and WAN 255 Edge functions are implemented in different systems. Figure 1 depicts 256 the reference model described in this section. Note that, although 257 not shown in Figure 1, GWs may have local ACs (Attachment Circuits). 259 +--+ 260 |CE| 261 +--+ 262 | 263 +----+ 264 +----| PE |----+ 265 +---------+ | +----+ | +---------+ 266 +----+ | +---+ +----+ +----+ +---+ | +----+ 267 |NVE1|--| | | |WAN | |WAN | | | |--|NVE3| 268 +----+ | |GW1|--|Edge| |Edge|--|GW3| | +----+ 269 | +---+ +----+ +----+ +---+ | 270 | NVO-1 | | WAN | | NVO-2 | 271 | +---+ +----+ +----+ +---+ | 272 | | | |WAN | |WAN | | | | 273 +----+ | |GW2|--|Edge| |Edge|--|GW4| | +----+ 274 |NVE2|--| +---+ +----+ +----+ +---+ |--|NVE4| 275 +----+ +---------+ | | +---------+ +----+ 276 +--------------+ 278 |<-EVPN-Overlay-->|<-VLAN->|<-WAN L2VPN->|<--PW-->|<--EVPN-Overlay->| 279 hand-off hand-off 281 Figure 1 Decoupled Interconnect model 283 The following section describes the interconnect requirements for 284 this model. 286 3.1. Interconnect requirements 288 The Decoupled Interconnect architecture is intended to be deployed in 289 networks where the EVPN-Overlay and WAN providers are different 290 entities and a clear demarcation is needed. This solution solves the 291 following requirements: 293 o A simple connectivity hand-off between the EVPN-Overlay network 294 provider and the WAN provider so that QoS and security enforcement 295 is easily accomplished. 297 o Independence of the Layer Two VPN (L2VPN) technology deployed in 298 the WAN. 300 o Multi-homing between GW and WAN Edge routers, including per-service 301 load balancing. Per-flow load balancing is not a strong requirement 302 since a deterministic path per service is usually required for an 303 easy QoS and security enforcement. 305 o Support of Ethernet OAM and Connectivity Fault Management (CFM) 306 [802.1AG][Y.1731] functions between the GW and the WAN Edge router 307 to detect individual AC failures. 309 o Support for the following optimizations at the GW: 310 + Flooding reduction of unknown unicast traffic sourced from the DC 311 Network Virtualization Edge devices (NVEs). 312 + Control of the WAN MAC addresses advertised to the DC. 313 + Address Resolution Protocol (ARP) and Neighbor Discovery (ND) 314 flooding control for the requests coming from the WAN. 316 3.2. VLAN-based hand-off 318 In this option, the hand-off between the GWs and the WAN Edge routers 319 is based on VLANs [802.1Q-2014]. This is illustrated in Figure 1 320 (between the GWs in NVO-1 and the WAN Edge routers). Each MAC-VRF in 321 the GW is connected to a different VSI/MAC-VRF instance in the WAN 322 Edge router by using a different C-TAG VLAN ID or a different 323 combination of S/C-TAG VLAN IDs that matches at both sides. 325 This option provides the best possible demarcation between the DC and 326 WAN providers and it does not require control plane interaction 327 between both providers. The disadvantage of this model is the 328 provisioning overhead since the service has to be mapped to a C-TAG 329 or S/C-TAG VLAN ID combination at both GW and WAN Edge routers. 331 In this model, the GW acts as a regular Network Virtualization Edge 332 (NVE) towards the DC. Its control plane, data plane procedures and 333 interactions are described in [EVPN-Overlays]. 335 The WAN Edge router acts as a (PBB-)VPLS or (PBB-)EVPN PE with 336 attachment circuits (ACs) to the GWs. Its functions are described in 337 [RFC4761], [RFC4762], [RFC6074] or [RFC7432], [RFC7623]. 339 3.3. PW-based (Pseudowire-based) hand-off 341 If MPLS between the GW and the WAN Edge router is an option, a PW- 342 based Interconnect solution can be deployed. In this option the 343 hand-off between both routers is based on FEC128-based PWs [RFC4762] 344 or FEC129-based PWs (for a greater level of network automation) 345 [RFC6074]. Note that this model still provides a clear demarcation 346 boundary between DC and WAN (since there is a single PW between each 347 MAC-VRF and peer VSI), and security/QoS policies may be applied on a 348 per PW basis. This model provides better scalability than a C-TAG 349 based hand-off and less provisioning overhead than a combined C/S-TAG 350 hand-off. The PW-based hand-off interconnect is illustrated in Figure 351 1 (between the NVO-2 GWs and the WAN Edge routers). 353 In this model, besides the usual MPLS procedures between GW and WAN 354 Edge router [RFC3031], the GW MUST support an interworking function 355 in each MAC-VRF that requires extension to the WAN: 357 o If a FEC128-based PW is used between the MAC-VRF (GW) and the VSI 358 (WAN Edge), the corresponding VCID MUST be provisioned on the MAC- 359 VRF and match the VCID used in the peer VSI at the WAN Edge router. 361 o If BGP Auto-discovery [RFC6074] and FEC129-based PWs are used 362 between the GW MAC-VRF and the WAN Edge VSI, the provisioning of 363 the VPLS-ID MUST be supported on the MAC-VRF and MUST match the 364 VPLS-ID used in the WAN Edge VSI. 366 If a PW-based handoff is used, the GW's AC (or point of attachment to 367 the EVPN Instance) uses a combination of a PW label and VLAN IDs. PWs 368 are treated as service interfaces defined in [RFC7432]. 370 3.4. Multi-homing solution on the GWs 372 EVPN single-active multi-homing, i.e. per-service load-balancing 373 multi-homing is required in this type of interconnect. 375 The GWs will be provisioned with a unique ES per WAN interconnect, 376 and the hand-off attachment circuits or PWs between the GW and the 377 WAN Edge router will be assigned an ESI for such ES. The ESI will be 378 administratively configured on the GWs according to the procedures in 379 [RFC7432]. This Interconnect ES will be referred as "I-ES" hereafter, 380 and its identifier will be referred as "I-ESI". [RFC7432] describes 381 different ESI Types. The use of Type 0 for the I-ESI is RECOMMENDED 382 in this document. 384 The solution (on the GWs) MUST follow the single-active multi-homing 385 procedures as described in [EVPN-Overlays] for the provisioned I-ESI, 386 i.e. Ethernet A-D routes per ES and per EVI will be advertised to the 387 DC NVEs for the multi-homing functions, ES routes will be advertised 388 so that ES discovery and Designated Forwarder (DF) procedures can be 389 followed. The MAC addresses learned (in the data plane) on the hand- 390 off links will be advertised with the I-ESI encoded in the ESI field. 392 3.5. Gateway Optimizations 394 The following GW features are optional and optimize the control plane 395 and data plane in the DC. 397 3.5.1. MAC Address Advertisement Control 399 The use of EVPN in NVO networks brings a significant number of 400 benefits as described in [EVPN-Overlays]. However, if multiple DCs 401 are interconnected into a single EVI, each DC will have to import all 402 of the MAC addresses from each of the other DCs. 404 Even if optimized BGP techniques like RT-constraint [RFC4684] are 405 used, the number of MAC addresses to advertise or withdraw (in case 406 of failure) by the GWs of a given DC could overwhelm the NVEs within 407 that DC, particularly when the NVEs reside in the hypervisors. 409 The solution specified in this document uses the 'Unknown MAC Route' 410 (UMR) which is advertised into a given DC by each of the DC's GWs. 411 This route is defined in [RFC7543] and is a regular EVPN MAC/IP 412 Advertisement route in which the MAC Address Length is set to 48, the 413 MAC address is set to 0, and the ESI field is set to the DC GW's I- 414 ESI. 416 An NVE within that DC that understands and process the UMR will send 417 unknown unicast frames to one of the DCs GWs, which will then forward 418 that packet to the correct egress PE. Note that, because the ESI is 419 set to the DC GW's I-ESI, all-active multi-homing can be applied to 420 unknown unicast MAC addresses. An NVE that does not understand the 421 Unknown MAC route will handle unknown unicast as described in 422 [RFC7432]. 424 This document proposes that local policy determines whether MAC 425 addresses and/or the UMR are advertised into a given DC. As an 426 example, when all the DC MAC addresses are learned in the 427 control/management plane, it may be appropriate to advertise only the 428 UMR. Advertising all the DC MAC addresses in the control/management 429 plane is usually the case when the NVEs reside in hypervisors. Refer 430 to [EVPN-Overlays] section 7. 432 It is worth noting that the UMR usage in [RFC7543] and the UMR usage 433 in this document are different. In the former, a Virtual Spoke (V- 434 spoke) does not necessarily learn all the MAC addresses pertaining to 435 hosts in other V-spokes of the same network. The communication 436 between two V-spokes is done through the DMG, until the V-spokes 437 learn each other's MAC addresses. In this document, two leaf switches 438 in the same DC are recommended to learn each other's MAC addresses 439 for the same EVI. The leaf to leaf communication is always direct and 440 does not go through the GW. 442 3.5.2. ARP/ND flooding control 444 Another optimization mechanism, naturally provided by EVPN in the 445 GWs, is the Proxy ARP/ND function. The GWs should build a Proxy 446 ARP/ND cache table as per [RFC7432]. When the active GW receives an 447 ARP/ND request/solicitation coming from the WAN, the GW does a Proxy 448 ARP/ND table lookup and replies as long as the information is 449 available in its table. 451 This mechanism is especially recommended on the GWs, since it 452 protects the DC network from external ARP/ND-flooding storms. 454 3.5.3. Handling failures between GW and WAN Edge routers 456 Link/PE failures are handled on the GWs as specified in [RFC7432]. 457 The GW detecting the failure will withdraw the EVPN routes as per 458 [RFC7432]. 460 Individual AC/PW failures may be detected by OAM mechanisms. For 461 instance: 463 o If the Interconnect solution is based on a VLAN hand-off, Ethernet- 464 CFM [802.1AG][Y.1731] may be used to detect individual AC failures 465 on both, the GW and WAN Edge router. An individual AC failure will 466 trigger the withdrawal of the corresponding A-D per EVI route as 467 well as the MACs learned on that AC. 469 o If the Interconnect solution is based on a PW hand-off, the Label 470 Distribution Protocol (LDP) PW Status bits TLV [RFC6870] may be 471 used to detect individual PW failures on both, the GW and WAN Edge 472 router. 474 4. Integrated Interconnect solution for EVPN overlay networks 476 When the DC and the WAN are operated by the same administrative 477 entity, the Service Provider can decide to integrate the GW and WAN 478 Edge PE functions in the same router for obvious CAPEX and OPEX 479 saving reasons. This is illustrated in Figure 2. Note that this model 480 does not provide an explicit demarcation link between DC and WAN 481 anymore. Although not shown in Figure 2, note that the GWs may have 482 local ACs. 484 +--+ 485 |CE| 486 +--+ 487 | 488 +----+ 489 +----| PE |----+ 490 +---------+ | +----+ | +---------+ 491 +----+ | +---+ +---+ | +----+ 492 |NVE1|--| | | | | |--|NVE3| 493 +----+ | |GW1| |GW3| | +----+ 494 | +---+ +---+ | 495 | NVO-1 | WAN | NVO-2 | 496 | +---+ +---+ | 497 | | | | | | 498 +----+ | |GW2| |GW4| | +----+ 499 |NVE2|--| +---+ +---+ |--|NVE4| 500 +----+ +---------+ | | +---------+ +----+ 501 +--------------+ 503 |<--EVPN-Overlay--->|<-----VPLS--->|<---EVPN-Overlay-->| 504 |<--PBB-VPLS-->| 505 Interconnect -> |<-EVPN-MPLS-->| 506 options |<--EVPN-Ovl-->|* 507 |<--PBB-EVPN-->| 509 Figure 2 Integrated Interconnect model 511 * EVPN-Ovl stands for EVPN-Overlay (and it's an Interconnect option). 513 4.1. Interconnect requirements 515 The Integrated Interconnect solution meets the following 516 requirements: 518 o Control plane and data plane interworking between the EVPN-overlay 519 network and the L2VPN technology supported in the WAN, irrespective 520 of the technology choice, i.e. (PBB-)VPLS or (PBB-)EVPN, as 521 depicted in Figure 2. 523 o Multi-homing, including single-active multi-homing with per-service 524 load balancing or all-active multi-homing, i.e. per-flow load- 525 balancing, as long as the technology deployed in the WAN supports 526 it. 528 o Support for end-to-end MAC Mobility, Static MAC protection and 529 other procedures (e.g. proxy-arp) described in [RFC7432] as long as 530 EVPN-MPLS is the technology of choice in the WAN. 532 o Independent inclusive multicast trees in the WAN and in the DC. 533 That is, the inclusive multicast tree type defined in the WAN does 534 not need to be the same as in the DC. 536 4.2. VPLS Interconnect for EVPN-Overlay networks 538 4.2.1. Control/Data Plane setup procedures on the GWs 540 Regular MPLS tunnels and TLDP/BGP sessions will be setup to the WAN 541 PEs and RRs as per [RFC4761], [RFC4762], [RFC6074] and overlay 542 tunnels and EVPN will be setup as per [EVPN-Overlays]. Note that 543 different route-targets for the DC and for the WAN are normally 544 required (unless [RFC4762] is used in the WAN, in which case no WAN 545 route-target is needed). A single type-1 RD per service may be used. 547 In order to support multi-homing, the GWs will be provisioned with an 548 I-ESI (see section 3.4), that will be unique per interconnection. The 549 I-ES in this case will represent the group of PWs to the WAN PEs and 550 GWs. All the [RFC7432] procedures are still followed for the I-ES, 551 e.g. any MAC address learned from the WAN will be advertised to the 552 DC with the I-ESI in the ESI field. 554 A MAC-VRF per EVI will be created in each GW. The MAC-VRF will have 555 two different types of tunnel bindings instantiated in two different 556 split-horizon-groups: 558 o VPLS PWs will be instantiated in the "WAN split-horizon-group". 560 o Overlay tunnel bindings (e.g. VXLAN, NVGRE) will be instantiated 561 in the "DC split-horizon-group". 563 Attachment circuits are also supported on the same MAC-VRF (although 564 not shown in Figure 2), but they will not be part of any of the above 565 split-horizon-groups. 567 Traffic received in a given split-horizon-group will never be 568 forwarded to a member of the same split-horizon-group. 570 As far as BUM flooding is concerned, a flooding list will be composed 571 of the sub-list created by the inclusive multicast routes and the 572 sub-list created for VPLS in the WAN. BUM frames received from a 573 local Attachment Circuit (AC) will be forwarded to the flooding list. 574 BUM frames received from the DC or the WAN will be forwarded to the 575 flooding list observing the split-horizon-group rule described above. 577 Note that the GWs are not allowed to have an EVPN binding and a PW to 578 the same far-end within the same MAC-VRF, so that loops and packet 579 duplication are avoided. In case a GW can successfully establish 580 both, an EVPN binding and a PW to the same far-end PE, the EVPN 581 binding will prevail and the PW will be brought operationally down. 583 The optimizations procedures described in section 3.5 can also be 584 applied to this model. 586 4.2.2. Multi-homing procedures on the GWs 588 This model supports single-active multi-homing on the GWs. All-active 589 multi-homing is not supported by VPLS, therefore it cannot be used on 590 the GWs. 592 In this case, for a given EVI, all the PWs in the WAN split-horizon- 593 group are assigned to I-ES. All the single-active multi-homing 594 procedures as described by [EVPN-Overlays] will be followed for the 595 I-ES. 597 The non-DF GW for the I-ES will block the transmission and reception 598 of all the PWs in the "WAN split-horizon-group" for BUM and unicast 599 traffic. 601 4.3. PBB-VPLS Interconnect for EVPN-Overlay networks 603 4.3.1. Control/Data Plane setup procedures on the GWs 605 In this case, there is no impact on the procedures described in 606 [RFC7041] for the B-component. However the I-component instances 607 become EVI instances with EVPN-Overlay bindings and potentially local 608 attachment circuits. A number of MAC-VRF instances can be multiplexed 609 into the same B-component instance. This option provides significant 610 savings in terms of PWs to be maintained in the WAN. 612 The I-ESI concept described in section 4.2.1 will also be used for 613 the PBB-VPLS-based Interconnect. 615 B-component PWs and I-component EVPN-overlay bindings established to 616 the same far-end will be compared. The following rules will be 617 observed: 619 o Attempts to setup a PW between the two GWs within the B- 620 component context will never be blocked. 622 o If a PW exists between two GWs for the B-component and an 623 attempt is made to setup an EVPN binding on an I-component linked 624 to that B-component, the EVPN binding will be kept operationally 625 down. Note that the BGP EVPN routes will still be valid but not 626 used. 628 o The EVPN binding will only be up and used as long as there is no 629 PW to the same far-end in the corresponding B-component. The EVPN 630 bindings in the I-components will be brought down before the PW in 631 the B-component is brought up. 633 The optimizations procedures described in section 3.5 can also be 634 applied to this Interconnect option. 636 4.3.2. Multi-homing procedures on the GWs 638 This model supports single-active multi-homing on the GWs. All-active 639 multi-homing is not supported by this scenario. 641 The single-active multi-homing procedures as described by [EVPN- 642 Overlays] will be followed for the I-ES for each EVI instance 643 connected to the B-component. Note that in this case, for a given 644 EVI, all the EVPN bindings in the I-component are assigned to the I- 645 ES. The non-DF GW for the I-ES will block the transmission and 646 reception of all the I-component EVPN bindings for BUM and unicast 647 traffic. When learning MACs from the WAN, the non-DF MUST NOT 648 advertise EVPN MAC/IP routes for those MACs. 650 4.4. EVPN-MPLS Interconnect for EVPN-Overlay networks 652 If EVPN for MPLS tunnels, EVPN-MPLS hereafter, is supported in the 653 WAN, an end-to-end EVPN solution can be deployed. The following 654 sections describe the proposed solution as well as the impact 655 required on the [RFC7432] procedures. 657 4.4.1. Control Plane setup procedures on the GWs 659 The GWs MUST establish separate BGP sessions for sending/receiving 660 EVPN routes to/from the DC and to/from the WAN. Normally each GW will 661 setup one BGP EVPN session to the DC RR (or two BGP EVPN sessions if 662 there are redundant DC RRs) and one session to the WAN RR (or two 663 sessions if there are redundant WAN RRs). 665 In order to facilitate separate BGP processes for DC and WAN, EVPN 666 routes sent to the WAN SHOULD carry a different route-distinguisher 667 (RD) than the EVPN routes sent to the DC. In addition, although 668 reusing the same value is possible, different route-targets are 669 expected to be handled for the same EVI in the WAN and the DC. Note 670 that the EVPN service routes sent to the DC RRs will normally include 671 a [TUNNEL-ENCAP] BGP encapsulation extended community with a 672 different tunnel type than the one sent to the WAN RRs. 674 As in the other discussed options, an I-ES and its assigned I-ESI 675 will be configured on the GWs for multi-homing. This I-ES represents 676 the WAN EVPN-MPLS PEs to the DC but also the DC EVPN-Overlay NVEs to 677 the WAN. Optionally, different I-ESI values are configured for 678 representing the WAN and the DC. If different EVPN-Overlay networks 679 are connected to the same group of GWs, each EVPN-Overlay network 680 MUST get assigned a different I-ESI. 682 Received EVPN routes will never be reflected on the GWs but consumed 683 and re-advertised (if needed): 685 o Ethernet A-D routes, ES routes and Inclusive Multicast routes 686 are consumed by the GWs and processed locally for the 687 corresponding [RFC7432] procedures. 689 o MAC/IP advertisement routes will be received, imported and if 690 they become active in the MAC-VRF, the information will be re- 691 advertised as new routes with the following fields: 693 + The RD will be the GW's RD for the MAC-VRF. 695 + The ESI will be set to the I-ESI. 697 + The Ethernet-tag value will be kept from the received NLRI. 699 + The MAC length, MAC address, IP Length and IP address values 700 will be kept from the received NLRI. 702 + The MPLS label will be a local 20-bit value (when sent to the 703 WAN) or a DC-global 24-bit value (when sent to the DC for 704 encapsulations using a VNI). 706 + The appropriate Route-Targets (RTs) and [TUNNEL-ENCAP] BGP 707 Encapsulation extended community will be used according to 708 [EVPN-Overlays]. 710 The GWs will also generate the following local EVPN routes that will 711 be sent to the DC and WAN, with their corresponding RTs and [TUNNEL- 712 ENCAP] BGP Encapsulation extended community values: 714 o ES route(s) for the I-ESI(s). 716 o Ethernet A-D routes per ES and EVI for the I-ESI(s). The A-D 717 per-EVI routes sent to the WAN and the DC will have consistent 718 Ethernet-Tag values. 720 o Inclusive Multicast routes with independent tunnel type value 721 for the WAN and DC. E.g. a P2MP LSP may be used in the WAN 722 whereas ingress replication may be used in the DC. The routes 723 sent to the WAN and the DC will have a consistent Ethernet-Tag. 725 o MAC/IP advertisement routes for MAC addresses learned in local 726 attachment circuits. Note that these routes will not include the 727 I-ESI, but ESI=0 or different from 0 for local multi-homed 728 Ethernet Segments (ES). The routes sent to the WAN and the DC 729 will have a consistent Ethernet-Tag. 731 Assuming GW1 and GW2 are peer GWs of the same DC, each GW will 732 generate two sets of the above local service routes: Set-DC will be 733 sent to the DC RRs and will include A-D per EVI, Inclusive Multicast 734 and MAC/IP routes for the DC encapsulation and RT. Set-WAN will be 735 sent to the WAN RRs and will include the same routes but using the 736 WAN RT and encapsulation. GW1 and GW2 will receive each other's set- 737 DC and set-WAN. This is the expected behavior on GW1 and GW2 for 738 locally generated routes: 740 o Inclusive multicast routes: when setting up the flooding lists 741 for a given MAC-VRF, each GW will include its DC peer GW only in 742 the EVPN-MPLS flooding list (by default) and not the EVPN- 743 Overlay flooding list. That is, GW2 will import two Inclusive 744 Multicast routes from GW1 (from set-DC and set-WAN) but will 745 only consider one of the two, having the set-WAN route higher 746 priority. An administrative option MAY change this preference so 747 that the set-DC route is selected first. 749 o MAC/IP advertisement routes for local attachment circuits: as 750 above, the GW will select only one, having the route from the 751 set-WAN a higher priority. As with the Inclusive multicast 752 routes, an administrative option MAY change this priority. 754 4.4.2. Data Plane setup procedures on the GWs 756 The procedure explained at the end of the previous section will make 757 sure there are no loops or packet duplication between the GWs of the 758 same EVPN-Overlay network (for frames generated from local ACs) since 759 only one EVPN binding per EVI (or per Ethernet Tag in case of VLAN- 760 aware bundle services) will be setup in the data plane between the 761 two nodes. That binding will by default be added to the EVPN-MPLS 762 flooding list. 764 As for the rest of the EVPN tunnel bindings, they will be added to 765 one of the two flooding lists that each GW sets up for the same MAC- 766 VRF: 768 o EVPN-overlay flooding list (composed of bindings to the remote 769 NVEs or multicast tunnel to the NVEs). 771 o EVPN-MPLS flooding list (composed of MP2P or LSM tunnel to the 772 remote PEs) 774 Each flooding list will be part of a separate split-horizon-group: 775 the WAN split-horizon-group or the DC split-horizon-group. Traffic 776 generated from a local AC can be flooded to both 777 split-horizon-groups. Traffic from a binding of a split-horizon-group 778 can be flooded to the other split-horizon-group and local ACs, but 779 never to a member of its own split-horizon-group. 781 When either GW1 or GW2 receive a BUM frame on an MPLS tunnel 782 including an ESI label at the bottom of the stack, they will perform 783 an ESI label lookup and split-horizon filtering as per [RFC7432] in 784 case the ESI label identifies a local ESI (I-ESI or any other non- 785 zero ESI). 787 4.4.3. Multi-homing procedure extensions on the GWs 789 This model supports single-active as well as all-active multi-homing. 791 All the [RFC7432] multi-homing procedures for the DF election on I- 792 ES(s) as well as the backup-path (single-active) and aliasing (all- 793 active) procedures will be followed on the GWs. Remote PEs in the 794 EVPN-MPLS network will follow regular [RFC7432] aliasing or backup- 795 path procedures for MAC/IP routes received from the GWs for the same 796 I-ESI. So will NVEs in the EVPN-Overlay network for MAC/IP routes 797 received with the same I-ESI. 799 As far as the forwarding plane is concerned, by default, the EVPN- 800 Overlay network will have an analogous behavior to the access ACs in 801 [RFC7432] multi-homed Ethernet Segments. 803 The forwarding behavior on the GWs is described below: 805 o Single-active multi-homing; assuming a WAN split-horizon-group 806 (comprised of EVPN-MPLS bindings), a DC split-horizon-group 807 (comprised of EVPN-Overlay bindings) and local ACs on the GWs: 809 + Forwarding behavior on the non-DF: the non-DF MUST block 810 ingress and egress forwarding on the EVPN-Overlay bindings 811 associated to the I-ES. The EVPN-MPLS network is considered to 812 be the core network and the EVPN-MPLS bindings to the remote 813 PEs and GWs will be active. 815 + Forwarding behavior on the DF: the DF MUST NOT forward BUM or 816 unicast traffic received from a given split-horizon-group to a 817 member of his own split-horizon group. Forwarding to other 818 split-horizon-groups and local ACs is allowed (as long as the 819 ACs are not part of an ES for which the node is non-DF). As 820 per [RFC7432] and for split-horizon purposes, when receiving 821 BUM traffic on the EVPN-Overlay bindings associated to an I- 822 ES, the DF GW SHOULD add the I-ESI label when forwarding to 823 the peer GW over EVPN-MPLS. 825 + When receiving EVPN MAC/IP routes from the WAN, the non-DF 826 MUST NOT re-originate the EVPN routes and advertise them to 827 the DC peers. In the same way, EVPN MAC/IP routes received 828 from the DC MUST NOT be advertised to the WAN peers. This is 829 consistent with [RFC7432] and allows the remote PE/NVEs know 830 who the primary GW is, based on the reception of the MAC/IP 831 routes. 833 o All-active multi-homing; assuming a WAN split-horizon-group 834 (comprised of EVPN-MPLS bindings), a DC split-horizon-group 835 (comprised of EVPN-Overlay bindings) and local ACs on the GWs: 837 + Forwarding behavior on the non-DF: the non-DF follows the same 838 behavior as the non-DF in the single-active case but only for 839 BUM traffic. Unicast traffic received from a split-horizon- 840 group MUST NOT be forwarded to a member of its own split- 841 horizon-group but can be forwarded normally to the other 842 split-horizon-groups and local ACs. If a known unicast packet 843 is identified as a "flooded" packet, the procedures for BUM 844 traffic MUST be followed. 846 + Forwarding behavior on the DF: the DF follows the same 847 behavior as the DF in the single-active case but only for BUM 848 traffic. Unicast traffic received from a split-horizon-group 849 MUST NOT be forwarded to a member of its own split-horizon- 850 group but can be forwarded normally to the other split- 851 horizon-group and local ACs. If a known unicast packet is 852 identified as a "flooded" packet, the procedures for BUM 853 traffic MUST be followed. As per [RFC7432] and for split- 854 horizon purposes, when receiving BUM traffic on the EVPN- 855 Overlay bindings associated to an I-ES, the DF GW MUST add the 856 I-ESI label when forwarding to the peer GW over EVPN-MPLS. 858 + Contrary to the single-active multi-homing case, both DF and 859 non-DF re-originate and advertise MAC/IP routes received from 860 the WAN/DC peers, adding the corresponding I-ESI so that the 861 remote PE/NVEs can perform regular aliasing as per [RFC7432]. 863 The example in Figure 3 illustrates the forwarding of BUM traffic 864 originated from an NVE on a pair of all-active multi-homing GWs. 866 |<--EVPN-Overlay--->|<--EVPN-MPLS-->| 868 +---------+ +--------------+ 869 +----+ BUM +---+ | 870 |NVE1+----+----> | +-+-----+ | 871 +----+ | | DF |GW1| | | | 872 | | +-+-+ | | ++--+ 873 | | | | +--> |PE1| 874 | +--->X +-+-+ | ++--+ 875 | NDF| | | | 876 +----+ | |GW2<-+ | 877 |NVE2+--+ +-+-+ | 878 +----+ +--------+ | +------------+ 879 v 880 +--+ 881 |CE| 882 +--+ 884 Figure 3 Multi-homing BUM forwarding 886 GW2 is the non-DF for the I-ES and blocks the BUM forwarding. GW1 is 887 the DF and forwards the traffic to PE1 and GW2. Packets sent to GW2 888 will include the ESI-label for the I-ES. Based on the ESI-label, GW2 889 identifies the packets as I-ES-generated packets and will only 890 forward them to local ACs (CE in the example) and not back to the 891 EVPN-Overlay network. 893 4.4.4. Impact on MAC Mobility procedures 895 MAC Mobility procedures described in [RFC7432] are not modified by 896 this document. 898 Note that an intra-DC MAC move still leaves the MAC attached to the 899 same I-ES, so under the rules of [RFC7432] this is not considered a 900 MAC mobility event. Only when the MAC moves from the WAN domain to 901 the DC domain (or from one DC to another) the MAC will be learned 902 from a different ES and the MAC Mobility procedures will kick in. 904 The sticky bit indication in the MAC Mobility extended community MUST 905 be propagated between domains. 907 4.4.5. Gateway optimizations 909 All the Gateway optimizations described in section 3.5 MAY be applied 910 to the GWs when the Interconnect is based on EVPN-MPLS. 912 In particular, the use of the Unknown MAC Route, as described in 913 section 3.5.1, solves some transient packet duplication issues in 914 cases of all-active multi-homing, as explained below. 916 Consider the diagram in Figure 2 for EVPN-MPLS Interconnect and all- 917 active multi-homing, and the following sequence: 919 a) MAC Address M1 is advertised from NVE3 in EVI-1. 921 b) GW3 and GW4 learn M1 for EVI-1 and re-advertise M1 to the WAN 922 with I-ESI-2 in the ESI field. 924 c) GW1 and GW2 learn M1 and install GW3/GW4 as next-hops following 925 the EVPN aliasing procedures. 927 d) Before NVE1 learns M1, a packet arrives at NVE1 with 928 destination M1. If the Unknown MAC Route had not been 929 advertised into the DC, NVE1 would have flooded the packet 930 throughout the DC, in particular to both GW1 and GW2. If the 931 same VNI/VSID is used for both known unicast and BUM traffic, 932 as is typically the case, there is no indication in the packet 933 that it is a BUM packet and both GW1 and GW2 would have 934 forwarded it, creating packet duplication. However, because the 935 Unknown MAC Route had been advertised into the DC, NVE1 will 936 unicast the packet to either GW1 or GW2. 938 e) Since both GW1 and GW2 know M1, the GW receiving the packet 939 will forward it to either GW3 or GW4. 941 4.4.6. Benefits of the EVPN-MPLS Interconnect solution 943 The [EVPN-Overlays] "DCI using ASBRs" solution and the GW solution 944 with EVPN-MPLS Interconnect may be seen similar since they both 945 retain the EVPN attributes between Data Centers and throughout the 946 WAN. However the EVPN-MPLS Interconnect solution on the GWs has 947 significant benefits compared to the "DCI using ASBRs" solution: 949 o As in any of the described GW models, this solution supports the 950 connectivity of local attachment circuits on the GWs. This is 951 not possible in a "DCI using ASBRs" solution. 953 o Different data plane encapsulations can be supported in the DC 954 and the WAN, while a uniform encapsulation is needed in the "DCI 955 using ASBRs" solution. 957 o Optimized multicast solution, with independent inclusive 958 multicast trees in DC and WAN. 960 o MPLS Label aggregation: for the case where MPLS labels are 961 signaled from the NVEs for MAC/IP Advertisement routes, this 962 solution provides label aggregation. A remote PE MAY receive a 963 single label per GW MAC-VRF as opposed to a label per NVE/MAC- 964 VRF connected to the GW MAC-VRF. For instance, in Figure 2, PE 965 would receive only one label for all the routes advertised for a 966 given MAC-VRF from GW1, as opposed to a label per NVE/MAC-VRF. 968 o The GW will not propagate MAC mobility for the MACs moving 969 within a DC. Mobility intra-DC is solved by all the NVEs in the 970 DC. The MAC Mobility procedures on the GWs are only required in 971 case of mobility across DCs. 973 o Proxy-ARP/ND function on the DC GWs can be leveraged to reduce 974 ARP/ND flooding in the DC or/and in the WAN. 976 4.5. PBB-EVPN Interconnect for EVPN-Overlay networks 978 PBB-EVPN [RFC7623] is yet another Interconnect option. It requires 979 the use of GWs where I-components and associated B-components are 980 part of EVI instances. 982 4.5.1. Control/Data Plane setup procedures on the GWs 984 EVPN will run independently in both components, the I-component MAC- 985 VRF and B-component MAC-VRF. Compared to [RFC7623], the DC C-MACs are 986 no longer learned in the data plane on the GW but in the control 987 plane through EVPN running on the I-component. Remote C-MACs coming 988 from remote PEs are still learned in the data plane. B-MACs in the B- 989 component will be assigned and advertised following the procedures 990 described in [RFC7623]. 992 An I-ES will be configured on the GWs for multi-homing, but its I-ESI 993 will only be used in the EVPN control plane for the I-component EVI. 994 No non-reserved ESIs will be used in the control plane of the B- 995 component EVI as per [RFC7623], that is, the I-ES will be represented 996 to the WAN PBB-EVPN PEs using shared or dedicated B-MACs. 998 The rest of the control plane procedures will follow [RFC7432] for 999 the I-component EVI and [RFC7623] for the B-component EVI. 1001 From the data plane perspective, the I-component and B-component EVPN 1002 bindings established to the same far-end will be compared and the I- 1003 component EVPN-overlay binding will be kept down following the rules 1004 described in section 4.3.1. 1006 4.5.2. Multi-homing procedures on the GWs 1007 This model supports single-active as well as all-active multi-homing. 1009 The forwarding behavior of the DF and non-DF will be changed based on 1010 the description outlined in section 4.4.3, only replacing the "WAN 1011 split-horizon-group" for the B-component, and using [RFC7623] 1012 procedures for the traffic sent or received on the B-component. 1014 4.5.3. Impact on MAC Mobility procedures 1016 C-MACs learned from the B-component will be advertised in EVPN within 1017 the I-component EVI scope. If the C-MAC was previously known in the 1018 I-component database, EVPN would advertise the C-MAC with a higher 1019 sequence number, as per [RFC7432]. From a Mobility perspective and 1020 the related procedures described in [RFC7432], the C-MACs learned 1021 from the B-component are considered local. 1023 4.5.4. Gateway optimizations 1025 All the considerations explained in section 4.4.5 are applicable to 1026 the PBB-EVPN Interconnect option. 1028 4.6. EVPN-VXLAN Interconnect for EVPN-Overlay networks 1030 If EVPN for Overlay tunnels is supported in the WAN and a GW function 1031 is required, an end-to-end EVPN solution can be deployed. While 1032 multiple Overlay tunnel combinations at the WAN and the DC are 1033 possible (MPLSoGRE, nvGRE, etc.), VXLAN is described here, given its 1034 popularity in the industry. This section focuses on the specific case 1035 of EVPN for VXLAN (EVPN-VXLAN hereafter) and the impact on the 1036 [RFC7432] procedures. 1038 The procedures described in section 4.4 apply to this section too, 1039 only replacing EVPN-MPLS for EVPN-VXLAN control plane specifics and 1040 using [EVPN-Overlays] "Local Bias" procedures instead of section 1041 4.4.3. Since there are no ESI-labels in VXLAN, GWs need to rely on 1042 "Local Bias" to apply split-horizon on packets generated from the I- 1043 ES and sent to the peer GW. 1045 This use-case assumes that NVEs need to use the VNIs or VSIDs as a 1046 globally unique identifiers within a data center, and a Gateway needs 1047 to be employed at the edge of the data center network to translate 1048 the VNI or VSID when crossing the network boundaries. This GW 1049 function provides VNI and tunnel IP address translation. The use-case 1050 in which local downstream assigned VNIs or VSIDs can be used (like 1051 MPLS labels) is described by [EVPN-Overlays]. 1053 While VNIs are globally significant within each DC, there are two 1054 possibilities in the Interconnect network: 1056 a) Globally unique VNIs in the Interconnect network: 1057 In this case, the GWs and PEs in the Interconnect network will 1058 agree on a common VNI for a given EVI. The RT to be used in the 1059 Interconnect network can be auto-derived from the agreed 1060 Interconnect VNI. The VNI used inside each DC MAY be the same 1061 as the Interconnect VNI. 1063 b) Downstream assigned VNIs in the Interconnect network. 1064 In this case, the GWs and PEs MUST use the proper RTs to 1065 import/export the EVPN routes. Note that even if the VNI is 1066 downstream assigned in the Interconnect network, and unlike 1067 option (a), it only identifies the pair and 1068 not the pair. The VNI used inside 1069 each DC MAY be the same as the Interconnect VNI. GWs SHOULD 1070 support multiple VNI spaces per EVI (one per Interconnect 1071 network they are connected to). 1073 In both options, NVEs inside a DC only have to be aware of a single 1074 VNI space, and only GWs will handle the complexity of managing 1075 multiple VNI spaces. In addition to VNI translation above, the GWs 1076 will provide translation of the tunnel source IP for the packets 1077 generated from the NVEs, using their own IP address. GWs will use 1078 that IP address as the BGP next-hop in all the EVPN updates to the 1079 Interconnect network. 1081 The following sections provide more details about these two options. 1083 4.6.1. Globally unique VNIs in the Interconnect network 1085 Considering Figure 2, if a host H1 in NVO-1 needs to communicate with 1086 a host H2 in NVO-2, and assuming that different VNIs are used in each 1087 DC for the same EVI, e.g. VNI-10 in NVO-1 and VNI-20 in NVO-2, then 1088 the VNIs MUST be translated to a common Interconnect VNI (e.g. VNI- 1089 100) on the GWs. Each GW is provisioned with a VNI translation 1090 mapping so that it can translate the VNI in the control plane when 1091 sending BGP EVPN route updates to the Interconnect network. In other 1092 words, GW1 and GW2 MUST be configured to map VNI-10 to VNI-100 in the 1093 BGP update messages for H1's MAC route. This mapping is also used to 1094 translate the VNI in the data plane in both directions, that is, VNI- 1095 10 to VNI-100 when the packet is received from NVO-1 and the reverse 1096 mapping from VNI-100 to VNI-10 when the packet is received from the 1097 remote NVO-2 network and needs to be forwarded to NVO-1. 1099 The procedures described in section 4.4 will be followed, considering 1100 that the VNIs advertised/received by the GWs will be translated 1101 accordingly. 1103 4.6.2. Downstream assigned VNIs in the Interconnect network 1104 In this case, if a host H1 in NVO-1 needs to communicate with a host 1105 H2 in NVO-2, and assuming that different VNIs are used in each DC for 1106 the same EVI, e.g. VNI-10 in NVO-1 and VNI-20 in NVO-2, then the VNIs 1107 MUST be translated as in section 4.6.1. However, in this case, there 1108 is no need to translate to a common Interconnect VNI on the GWs. Each 1109 GW can translate the VNI received in an EVPN update to a locally 1110 assigned VNI advertised to the Interconnect network. Each GW can use 1111 a different Interconnect VNI, hence this VNI does not need to be 1112 agreed on all the GWs and PEs of the Interconnect network. 1114 The procedures described in section 4.4 will be followed, taking the 1115 considerations above for the VNI translation. 1117 5. Security Considerations 1119 This document applies existing specifications to a number of 1120 Interconnect models. The Security Considerations included in those 1121 documents, such as [RFC7432], [EVPN-Overlays], [RFC7623], [RFC4761] 1122 and [RFC4762] apply to this document whenever those technologies are 1123 used. 1125 As discussed, [EVPN-Overlays] discusses two main DCI solution groups: 1126 "DCI using GWs" and "DCI using ASBRs". This document specifies the 1127 solutions that correspond to the "DCI using GWs" group. It is 1128 important to note that the use of GWs provide a superior level of 1129 security on a per tenant basis, compared to the use of ASBRs. This is 1130 due to the fact that GWs need to perform a MAC lookup on the frames 1131 being received from the WAN, and they apply security procedures, such 1132 as filtering of undesired frames, filtering of frames with a source 1133 MAC that matches a protected MAC in the DC or application of MAC 1134 duplication procedures defined in [RFC7432]. On ASBRs though, traffic 1135 is forwarded based on a label or VNI swap and there is usually no 1136 visibility of the encapsulated frames, which can carry malicious 1137 traffic. 1139 In addition, the GW optimizations specified in this document, provide 1140 additional protection of the DC Tenant Systems. For instance, the MAC 1141 address advertisement control and Unknown MAC Route defined in 1142 section 3.5.1 protect the DC NVEs from being overwhelmed with an 1143 excessive number MAC/IP routes being learned on the GWs from the WAN. 1144 The ARP/ND flooding control described in 3.5.2 can reduce/suppress 1145 broadcast storms being injected from the WAN. 1147 Finally, the reader should be aware of the potential security 1148 implications of designing a DCI with the Decoupled Interconnect 1149 solution (section 3) or the Integrated Interconnect solution (section 1150 4). In the Decoupled Interconnect solution the DC is typically easier 1151 to protect from the WAN, since each GW has a single logical link to 1152 one WAN PE, whereas in the Integrated solution, the GW has logical 1153 links to all the WAN PEs that are attached to the tenant. In either 1154 model, proper control plane and data plane policies should be put in 1155 place in the GWs in order to protect the DC from potential attacks 1156 coming from the WAN. 1158 6. IANA Considerations 1160 This document has no IANA actions. 1162 7. References 1164 7.1. Normative References 1166 [RFC4761] Kompella, K., Ed., and Y. Rekhter, Ed., "Virtual Private 1167 LAN Service (VPLS) Using BGP for Auto-Discovery and Signaling", 1168 RFC 4761, DOI 10.17487/RFC4761, January 2007, . 1171 [RFC4762] Lasserre, M., Ed., and V. Kompella, Ed., "Virtual Private 1172 LAN Service (VPLS) Using Label Distribution Protocol (LDP) 1173 Signaling", RFC 4762, DOI 10.17487/RFC4762, January 2007, 1174 . 1176 [RFC6074] Rosen, E., Davie, B., Radoaca, V., and W. Luo, 1177 "Provisioning, Auto-Discovery, and Signaling in Layer 2 Virtual 1178 Private Networks (L2VPNs)", RFC 6074, DOI 10.17487/RFC6074, January 1179 2011, . 1181 [RFC7041] Balus, F., Ed., Sajassi, A., Ed., and N. Bitar, Ed., 1182 "Extensions to the Virtual Private LAN Service (VPLS) Provider Edge 1183 (PE) Model for Provider Backbone Bridging", RFC 7041, DOI 1184 10.17487/RFC7041, November 2013, . 1187 [RFC7432] Sajassi, A., Ed., Aggarwal, R., Bitar, N., Isaac, A., 1188 Uttaro, J., Drake, J., and W. Henderickx, "BGP MPLS-Based Ethernet 1189 VPN", RFC 7432, DOI 10.17487/RFC7432, February 2015, . 1192 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1193 Requirement Levels", BCP 14, RFC 2119, DOI 10.17487/RFC2119, March 1194 1997, . 1196 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1197 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, May 2017, 1198 . 1200 [TUNNEL-ENCAP] Rosen et al., "The BGP Tunnel Encapsulation 1201 Attribute", draft-ietf-idr-tunnel-encaps-08, work in progress, 1202 January 11, 2018. 1204 [RFC7623] Sajassi et al., "Provider Backbone Bridging Combined with 1205 Ethernet VPN (PBB-EVPN)", RFC 7623, September, 2015, . 1208 [EVPN-Overlays] Sajassi-Drake et al., "A Network Virtualization 1209 Overlay Solution using EVPN", draft-ietf-bess-evpn-overlay-11.txt, 1210 work in progress, January 2018. 1212 [RFC7543] Jeng, H., Jalil, L., Bonica, R., Patel, K., and L. Yong, 1213 "Covering Prefixes Outbound Route Filter for BGP-4", RFC 7543, DOI 1214 10.17487/RFC7543, May 2015, . 1217 7.2. Informative References 1219 [RFC4684] Marques, P., Bonica, R., Fang, L., Martini, L., Raszuk, 1220 R., Patel, K., and J. Guichard, "Constrained Route Distribution for 1221 Border Gateway Protocol/MultiProtocol Label Switching (BGP/MPLS) 1222 Internet Protocol (IP) Virtual Private Networks (VPNs)", RFC 4684, 1223 DOI 10.17487/RFC4684, November 2006, . 1226 [RFC7348] Mahalingam, M., Dutt, D., Duda, K., Agarwal, P., Kreeger, 1227 L., Sridhar, T., Bursell, M., and C. Wright, "Virtual eXtensible 1228 Local Area Network (VXLAN): A Framework for Overlaying Virtualized 1229 Layer 2 Networks over Layer 3 Networks", RFC 7348, DOI 1230 10.17487/RFC7348, August 2014, . 1233 [RFC7637] Garg, P., et al., "NVGRE: Network Virtualization using 1234 Generic Routing Encapsulation", RFC 7637, September, 2015 1236 [RFC4023] Worster, T., Rekhter, Y., and E. Rosen, Ed., 1237 "Encapsulating MPLS in IP or Generic Routing Encapsulation (GRE)", 1238 RFC 4023, DOI 10.17487/RFC4023, March 2005, . 1241 [Y.1731] ITU-T Recommendation Y.1731, "OAM functions and mechanisms 1242 for Ethernet based networks", July 2011. 1244 [802.1AG] IEEE 802.1AG_2007, "IEEE Standard for Local and 1245 Metropolitan Area Networks - Virtual Bridged Local Area Networks 1246 Amendment 5: Connectivity Fault Management", January 2008. 1248 [802.1Q-2014] IEEE 802.1Q-2014, "IEEE Standard for Local and 1249 metropolitan area networks--Bridges and Bridged Networks", December 1250 2014. 1252 [RFC6870] Muley, P., Ed., and M. Aissaoui, Ed., "Pseudowire 1253 Preferential Forwarding Status Bit", RFC 6870, DOI 10.17487/RFC6870, 1254 February 2013, . 1256 [RFC3031] Rosen, E., Viswanathan, A., and R. Callon, "Multiprotocol 1257 Label Switching Architecture", RFC 3031, DOI 10.17487/RFC3031, 1258 January 2001, . 1260 [VIRTUAL-ES] Sajassi et al., "EVPN Virtual Ethernet Segment", draft- 1261 sajassi-bess-evpn-virtual-eth-segment-03, work in progress, February 1262 2018. 1264 8. Acknowledgments 1266 The authors would like to thank Neil Hart, Vinod Prabhu and Kiran 1267 Nagaraj for their valuable comments and feedback. We would also like 1268 to thank Martin Vigoureux and Alvaro Retana for his detailed review 1269 and comments. 1271 9. Contributors 1273 In addition to the authors listed on the front page, the following 1274 co-authors have also contributed to this document: 1276 Ravi Shekhar 1277 Anil Lohiya 1278 Wen Lin 1279 Juniper Networks 1281 Florin Balus 1282 Patrice Brissette 1283 Cisco 1285 Senad Palislamovic 1286 Nokia 1288 Dennis Cai 1289 Alibaba 1291 10. Authors' Addresses 1293 Jorge Rabadan 1294 Nokia 1295 777 E. Middlefield Road 1296 Mountain View, CA 94043 USA 1297 Email: jorge.rabadan@nokia.com 1299 Senthil Sathappan 1300 Nokia 1301 Email: senthil.sathappan@nokia.com 1303 Wim Henderickx 1304 Nokia 1305 Email: wim.henderickx@nokia.com 1307 Ali Sajassi 1308 Cisco 1309 Email: sajassi@cisco.com 1311 John Drake 1312 Juniper 1313 Email: jdrake@juniper.net