idnits 2.17.1 draft-ietf-l2vpn-evpn-09.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 53 instances of lines with control characters in the document. ** The abstract seems to contain references ([EVPN-REQ]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 1403 has weird spacing: '...ntinues forwa...' -- The document date (October 2, 2014) is 3493 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) == Missing Reference: 'EVPN-REQ' is mentioned on line 432, but not defined == Missing Reference: 'VPLS-MCAST' is mentioned on line 2082, but not defined == Missing Reference: 'RFC5925' is mentioned on line 2212, but not defined == Unused Reference: 'RFC7209' is defined on line 2300, but no explicit reference was found in the text == Unused Reference: 'RFC7117' is defined on line 2303, but no explicit reference was found in the text == Outdated reference: A later version (-07) exists of draft-ietf-l2vpn-evpn-req-04 == Outdated reference: A later version (-16) exists of draft-ietf-l2vpn-vpls-mcast-14 -- Obsolete informational reference (is this intentional?): RFC 5226 (Obsoleted by RFC 8126) Summary: 2 errors (**), 0 flaws (~~), 9 warnings (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group A. Sajassi, Ed. 3 INTERNET-DRAFT Cisco 4 Category: Standards Track 5 R. Aggarwal 6 J. Drake Arktan 7 Juniper Networks 8 N. Bitar 9 W. Henderickx Verizon 10 Alcatel-Lucent 11 Aldrin Isaac 12 Bloomberg 14 J. Uttaro 15 AT&T 17 Expires: April 2, 2015 October 2, 2014 19 BGP MPLS Based Ethernet VPN 20 draft-ietf-l2vpn-evpn-09 22 Status of this Memo 24 This Internet-Draft is submitted to IETF in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF), its areas, and its working groups. Note that 29 other groups may also distribute working documents as 30 Internet-Drafts. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 The list of current Internet-Drafts can be accessed at 38 http://www.ietf.org/1id-abstracts.html 40 The list of Internet-Draft Shadow Directories can be accessed at 41 http://www.ietf.org/shadow.html 43 Copyright and License Notice 45 Copyright (c) 2013 IETF Trust and the persons identified as the 46 document authors. All rights reserved. 48 This document is subject to BCP 78 and the IETF Trust's Legal 49 Provisions Relating to IETF Documents 50 (http://trustee.ietf.org/license-info) in effect on the date of 51 publication of this document. Please review these documents 52 carefully, as they describe your rights and restrictions with respect 53 to this document. Code Components extracted from this document must 54 include Simplified BSD License text as described in Section 4.e of 55 the Trust Legal Provisions and are provided without warranty as 56 described in the Simplified BSD License. 58 Abstract 60 This document describes procedures for BGP MPLS based Ethernet VPNs 61 (EVPN). The procedures described here are intended to meet the 62 requirements specified in [EVPN-REQ]. 64 Table of Contents 66 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 5 67 2. Specification of requirements . . . . . . . . . . . . . . . . . 5 68 3. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . 5 69 4. BGP MPLS Based EVPN Overview . . . . . . . . . . . . . . . . . 6 70 5. Ethernet Segment . . . . . . . . . . . . . . . . . . . . . . . 7 71 6. Ethernet Tag ID . . . . . . . . . . . . . . . . . . . . . . . . 10 72 6.1 VLAN Based Service Interface . . . . . . . . . . . . . . . . 11 73 6.2 VLAN Bundle Service Interface . . . . . . . . . . . . . . . 11 74 6.2.1 Port Based Service Interface . . . . . . . . . . . . . . 11 75 6.3 VLAN Aware Bundle Service Interface . . . . . . . . . . . . 11 76 6.3.1 Port Based VLAN Aware Service Interface . . . . . . . . 12 77 7. BGP EVPN NLRI . . . . . . . . . . . . . . . . . . . . . . . . . 12 78 7.1. Ethernet Auto-Discovery Route . . . . . . . . . . . . . . . 13 79 7.2. MAC/IP Advertisement Route . . . . . . . . . . . . . . . . 13 80 7.3. Inclusive Multicast Ethernet Tag Route . . . . . . . . . . 14 81 7.4 Ethernet Segment Route . . . . . . . . . . . . . . . . . . . 15 82 7.5 ESI Label Extended Community . . . . . . . . . . . . . . . . 15 83 7.6 ES-Import Route Target . . . . . . . . . . . . . . . . . . . 16 84 7.7 MAC Mobility Extended Community . . . . . . . . . . . . . . 16 85 7.8 Default Gateway Extended Community . . . . . . . . . . . . . 17 86 7.9 Route Distinguisher Assignment per EVI . . . . . . . . . . . 17 87 7.10 Route Targets . . . . . . . . . . . . . . . . . . . . . . . 17 88 7.10.1 Auto-Derivation from the Ethernet Tag ID . . . . . . . 17 89 8. Multi-homing Functions . . . . . . . . . . . . . . . . . . . . 18 90 8.1 Multi-homed Ethernet Segment Auto-Discovery . . . . . . . . 18 91 8.1.1 Constructing the Ethernet Segment Route . . . . . . . . 18 92 8.2 Fast Convergence . . . . . . . . . . . . . . . . . . . . . . 18 93 8.2.1 Constructing the Ethernet A-D per Ethernet Segment 94 (ES) Route . . . . . . . . . . . . . . . . . . . . . . . 19 95 8.2.1.1. Ethernet A-D Route Targets . . . . . . . . . . . . 20 97 8.3 Split Horizon . . . . . . . . . . . . . . . . . . . . . . . 20 98 8.3.1 ESI Label Assignment . . . . . . . . . . . . . . . . . . 21 99 8.3.1.1 Ingress Replication . . . . . . . . . . . . . . . . 21 100 8.3.1.2. P2MP MPLS LSPs . . . . . . . . . . . . . . . . . . 22 101 8.4 Aliasing and Backup-Path . . . . . . . . . . . . . . . . . . 23 102 8.4.1 Constructing the Ethernet A-D per EVPN Instance (EVI) 103 Route . . . . . . . . . . . . . . . . . . . . . . . . . 24 104 8.5 Designated Forwarder Election . . . . . . . . . . . . . . . 25 105 8.6. Interoperability with Single-homing PEs . . . . . . . . . . 27 106 9. Determining Reachability to Unicast MAC Addresses . . . . . . . 27 107 9.1. Local Learning . . . . . . . . . . . . . . . . . . . . . . 28 108 9.2. Remote learning . . . . . . . . . . . . . . . . . . . . . . 28 109 9.2.1. Constructing the BGP EVPN MAC/IP Address 110 Advertisement . . . . . . . . . . . . . . . . . . . . . 28 111 9.2.2 Route Resolution . . . . . . . . . . . . . . . . . . . . 30 112 10. ARP and ND . . . . . . . . . . . . . . . . . . . . . . . . . . 31 113 10.1 Default Gateway . . . . . . . . . . . . . . . . . . . . . . 32 114 11. Handling of Multi-Destination Traffic . . . . . . . . . . . . 33 115 11.1. Construction of the Inclusive Multicast Ethernet Tag 116 Route . . . . . . . . . . . . . . . . . . . . . . . . . . 34 117 11.2. P-Tunnel Identification . . . . . . . . . . . . . . . . . 34 118 12. Processing of Unknown Unicast Packets . . . . . . . . . . . . 35 119 12.1. Ingress Replication . . . . . . . . . . . . . . . . . . . 36 120 12.2. P2MP MPLS LSPs . . . . . . . . . . . . . . . . . . . . . . 36 121 13. Forwarding Unicast Packets . . . . . . . . . . . . . . . . . . 37 122 13.1. Forwarding packets received from a CE . . . . . . . . . . 37 123 13.2. Forwarding packets received from a remote PE . . . . . . . 38 124 13.2.1. Unknown Unicast Forwarding . . . . . . . . . . . . . . 38 125 13.2.2. Known Unicast Forwarding . . . . . . . . . . . . . . . 38 126 14. Load Balancing of Unicast Frames . . . . . . . . . . . . . . . 38 127 14.1. Load balancing of traffic from a PE to remote CEs . . . . 39 128 14.1.1 Single-Active Redundancy Mode . . . . . . . . . . . . . 39 129 14.1.2 All-Active Redundancy Mode . . . . . . . . . . . . . . 39 130 14.2. Load balancing of traffic between a PE and a local CE . . 41 131 14.2.1. Data plane learning . . . . . . . . . . . . . . . . . 41 132 14.2.2. Control plane learning . . . . . . . . . . . . . . . . 41 133 15. MAC Mobility . . . . . . . . . . . . . . . . . . . . . . . . . 42 134 15.1. MAC Duplication Issue . . . . . . . . . . . . . . . . . . 43 135 15.2. Sticky MAC addresses . . . . . . . . . . . . . . . . . . . 44 136 16. Multicast & Broadcast . . . . . . . . . . . . . . . . . . . . 44 137 16.1. Ingress Replication . . . . . . . . . . . . . . . . . . . 44 138 16.2. P2MP LSPs . . . . . . . . . . . . . . . . . . . . . . . . 44 139 16.2.1. Inclusive Trees . . . . . . . . . . . . . . . . . . . 45 140 17. Convergence . . . . . . . . . . . . . . . . . . . . . . . . . 45 141 17.1. Transit Link and Node Failures between PEs . . . . . . . . 45 142 17.2. PE Failures . . . . . . . . . . . . . . . . . . . . . . . 46 143 17.3. PE to CE Network Failures . . . . . . . . . . . . . . . . 46 144 18. Frame Ordering . . . . . . . . . . . . . . . . . . . . . . . . 46 145 19. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 47 146 20. Security Considerations . . . . . . . . . . . . . . . . . . . 47 147 21. Contributors . . . . . . . . . . . . . . . . . . . . . . . . . 49 148 22. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 49 149 23. References . . . . . . . . . . . . . . . . . . . . . . . . . . 49 150 23.1 Normative References . . . . . . . . . . . . . . . . . . . 49 151 23.2 Informative References . . . . . . . . . . . . . . . . . . 50 152 24. Author's Address . . . . . . . . . . . . . . . . . . . . . . . 50 154 1. Introduction 156 This document describes procedures for BGP MPLS based Ethernet VPNs 157 (EVPN). The procedures described here are intended to meet the 158 requirements specified in [EVPN-REQ]. Please refer to [EVPN-REQ] for 159 the detailed requirements and motivation. EVPN requires extensions to 160 existing IP/MPLS protocols as described in this document. In addition 161 to these extensions EVPN uses several building blocks from existing 162 MPLS technologies. 164 2. Specification of requirements 166 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 167 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 168 document are to be interpreted as described in [RFC2119]. 170 3. Terminology 172 Broadcast Domain: in a bridged network, it corresponds to a Virtual 173 LAN (VLAN); where a VLAN is typically represented by a single VLAN ID 174 (VID), but can be represented by several VIDs. 176 Bridge Domain: An instantiation of a broadcast domain on a bridge 177 node 179 CE: Customer Edge device e.g., host or router or switch 181 EVI: An EVPN instance spanning across the PEs participating in that 182 EVPN 184 MAC-VRF: A Virtual Routing and Forwarding table for MAC addresses on 185 a PE for an EVI 187 Ethernet Segment Identifier (ESI): If a CE is multi-homed to two or 188 more PEs, the set of Ethernet links that attaches the CE to the PEs 189 is an 'Ethernet segment'. Ethernet segments MUST have a unique non- 190 zero identifier, the 'Ethernet Segment Identifier'. 192 Ethernet Tag: An Ethernet Tag identifies a particular broadcast 193 domain, e.g., a VLAN. An EVPN instance consists of one or more 194 broadcast domains. Ethernet tag(s) are assigned to the broadcast 195 domains of a given EVPN instance by the provider of that EVPN, and 196 each PE in that EVPN instance performs a mapping between broadcast 197 domain identifier(s) understood by each of its attached CEs and the 198 corresponding Ethernet tag. 200 LACP: Link Aggregation Control Protocol 202 MP2MP: Multipoint to Multipoint 204 P2MP: Point to Multipoint 206 P2P: Point to Point 208 Single-Active Redundancy Mode: When only a single PE, among a group 209 of PEs attached to an Ethernet segment, is allowed to forward traffic 210 to/from that Ethernet Segment, then the Ethernet segment is defined 211 to be operating in Single-Active redundancy mode. 213 All-Active Redundancy Mode: When all PEs attached to an Ethernet 214 segment are allowed to forward traffic to/from that Ethernet Segment, 215 then the Ethernet segment is defined to be operating in All-Active 216 redundancy mode. 218 4. BGP MPLS Based EVPN Overview 220 This section provides an overview of EVPN. An EVPN instance comprises 221 CEs that are connected to PEs that form the edge of the MPLS 222 infrastructure. A CE may be a host, a router or a switch. The PEs 223 provide virtual Layer 2 bridged connectivity between the CEs. There 224 may be multiple EVPN instances in the provider's network. 226 The PEs may be connected by an MPLS LSP infrastructure which provides 227 the benefits of MPLS technology such as fast-reroute, resiliency, 228 etc. The PEs may also be connected by an IP infrastructure in which 229 case IP/GRE tunneling or other IP tunneling can be used between the 230 PEs. The detailed procedures in this version of this document are 231 specified only for MPLS LSPs as the tunneling technology. However 232 these procedures are designed to be extensible to IP tunneling as the 233 Packet Switched Network (PSN) tunneling technology. 235 In an EVPN, MAC learning between PEs occurs not in the data plane (as 236 happens with traditional bridging in VPLS [RFC4761] or [RFC4762]) but 237 in the control plane. Control plane learning offers greater control 238 over the MAC learning process, such as restricting who learns what, 239 and the ability to apply policies. Furthermore, the control plane 240 chosen for advertising MAC reachability information is multi-protocol 241 (MP) BGP (similar to IP VPNs (RFC 4364)). This provides flexibility 242 and the ability to preserve the "virtualization" or isolation of 243 groups of interacting agents (hosts, servers, virtual machines) from 244 each other. In EVPN, PEs advertise the MAC addresses learned from the 245 CEs that are connected to them, along with an MPLS label, to other 246 PEs in the control plane using MP-BGP. Control plane learning enables 247 load balancing of traffic to and from CEs that are multi-homed to 248 multiple PEs. This is in addition to load balancing across the MPLS 249 core via multiple LSPs between the same pair of PEs. In other words 250 it allows CEs to connect to multiple active points of attachment. It 251 also improves convergence times in the event of certain network 252 failures. 254 However, learning between PEs and CEs is done by the method best 255 suited to the CE: data plane learning, IEEE 802.1x, LLDP, 802.1aq, 256 ARP, management plane or other protocols. 258 It is a local decision as to whether the Layer 2 forwarding table on 259 a PE is populated with all the MAC destination addresses known to the 260 control plane, or whether the PE implements a cache based scheme. For 261 instance the MAC forwarding table may be populated only with the MAC 262 destinations of the active flows transiting a specific PE. 264 The policy attributes of EVPN are very similar to those of IP-VPN. A 265 EVPN instance requires a Route Distinguisher (RD) which is unique per 266 PE and one or more globally unique Route-Targets (RTs). A CE attaches 267 to a MAC-VRF on a PE, on an Ethernet interface which may be 268 configured for one or more Ethernet Tags, e.g., VLAN IDs. Some 269 deployment scenarios guarantee uniqueness of VLAN IDs across EVPN 270 instances: all points of attachment for a given EVPN instance use the 271 same VLAN ID, and no other EVPN instance uses this VLAN ID. This 272 document refers to this case as a "Unique VLAN EVPN" and describes 273 simplified procedures to optimize for it. 275 5. Ethernet Segment 277 If a CE is multi-homed to two or more PEs, the set of Ethernet links 278 constitutes an "Ethernet Segment". An Ethernet segment may appear to 279 the CE as a Link Aggregation Group (LAG). Ethernet segments have an 280 identifier, called the "Ethernet Segment Identifier" (ESI) which is 281 encoded as a ten octets integer in line format with the most 282 significant octet sent first. The following two ESI values are 283 reserved: 285 - ESI 0 denotes a single-homed CE. 287 - ESI {0xFF} (repeated 10 times) is known as MAX-ESI and is 288 reserved. 290 In general, an Ethernet segment SHOULD have a non-reserved ESI that 291 is unique network wide (i.e., across all EVPN instances on all the 292 PEs). If the CE(s) constituting an Ethernet Segment is (are) managed 293 by the network operator, then ESI uniqueness should be guaranteed; 294 however, if the CE(s) is (are) not managed, then the operator MUST 295 configure a network-wide unique ESI for that Ethernet Segment. This 296 is required to enable auto-discovery of Ethernet Segments and DF 297 election. 299 In a network with managed and not-managed CEs, the ESI has the 300 following format: 302 +---+---+---+---+---+---+---+---+---+---+ 303 | T | ESI Value | 304 +---+---+---+---+---+---+---+---+---+---+ 306 Where: 308 T (ESI Type) is a 1-octet field (most significant octet) that 309 specifies the format of the remaining nine octets (ESI Value). The 310 following 6 ESI types can be used: 312 - Type 0 (T=0x00) - This type indicates an arbitrary nine-octet ESI 313 value, which is managed and configured by the operator. 315 - Type 1 (T=0x01) - When IEEE 802.1AX LACP is used between the PEs 316 and CEs, this ESI type indicates an auto-generated ESI value 317 determined from LACP by concatenating the following parameters: 319 + CE LACP six octets System MAC address. The CE LACP System MAC 320 address MUST be encoded in the high order six octets of the ESI 321 Value field. 323 + CE LACP two octets Port Key. The CE LACP port key MUST be 324 encoded in the two octets next to the System MAC address. 326 + The remaining octet will be set to 0x00. 328 As far as the CE is concerned, it would treat the multiple PEs 329 that it is connected to as the same switch. This allows the CE 330 to aggregate links that are attached to different PEs in the 331 same bundle. 333 This mechanism could be used only if it produces ESIs that satisfy 334 the uniqueness requirement specified above. 336 - Type 2 (T=0x02) - This type is used in the case of indirectly 337 connected hosts via a bridged LAN between the CEs and the PEs. The 338 ESI Value is auto-generated and determined based on the Layer 2 339 bridge protocol as follows: If MST is used in the bridged LAN then 340 the value of the ESI is derived by listening to BPDUs on the Ethernet 341 segment. To achieve this the PE is not required to run MST. However 342 the PE must learn the Root Bridge MAC address and Bridge Priority of 343 the root of the Internal Spanning Tree (IST) by listening to the 344 BPDUs. The ESI Value is constructed as follows: 346 + Root Bridge six octets MAC address. The Root Bridge MAC 347 address MUST be encoded in the high order six octets of the 348 ESI Value field. 350 + Root Bridge two octets Priority. The CE Root Bridge Priority 351 MUST be encoded in the two octets next to the Root Bridge 352 MAC address. 354 + The remaining octet will be set to 0x00. 356 This mechanism could be used only if it produces ESIs that 357 satisfy the uniqueness requirement specified above. 359 - Type 3 (T=0x03) - This type indicates a MAC-based ESI Value that 360 can be auto-generated or configured by the operator. The ESI Value is 361 constructed as follows: 363 + System MAC address (six octets). The PE MAC address MUST 364 be encoded in the high order six octets of the ESI Value field. 366 + Local Discriminator value (three octets). The Local 367 Discriminator MUST be encoded in the low order three octets 368 of the ESI Value. 370 This mechanism could be used only if it produces ESIs that 371 satisfy the uniqueness requirement specified above. 373 - Type 4 (T=0x04) - This type indicates a router-ID ESI Value that 374 can be auto-generated or configured by the operator. The ESI Value is 375 constructed as follows: 377 + Router ID (four octets). The system router ID MUST be encoded in 378 the high order four octets of the ESI Value field. 380 + Local Discriminator value (four octets). The Local 381 Discriminator MUST be encoded in the four octets next to the 382 IP address. 384 + The low order octet of the ESI Value will be set to 0x00. 386 This mechanism could be used only if it produces ESIs that 387 satisfy the uniqueness requirement specified above. 389 - Type 5 (T=0x05) - This type indicates an AS-based ESI Value that 390 can be auto-generated or configured by the operator. The ESI Value is 391 constructed as follows: 393 + AS number (four octets). This is an AS number owned by the 394 system and MUST be encoded in the high order four octets of the 395 ESI Value field. If a two-octet AS number is used, the high order 396 extra two octets will be 0x0000. 398 + Local Discriminator value (four octets). The Local Discriminator 399 MUST be encoded in the four octets next to the AS number. 401 + The low order octet of the ESI Value will be set to 0x00. 403 This mechanism could be used only if it produces ESIs that satisfy 404 the uniqueness requirement specified above. 406 6. Ethernet Tag ID 408 An Ethernet Tag ID is a 32-bit field containing either a 12-bit or a 409 24-bit identifier that identifies a particular broadcast domain 410 (e.g., a VLAN) in an EVPN Instance. The 12-bit identifier is called 411 VLAN ID (VID). An EVPN Instance consists of one or more broadcast 412 domains (one or more VLANs). VLANs are assigned to a given EVPN 413 Instance by the provider of the EVPN service. A given VLAN can itself 414 be represented by multiple VLAN IDs (VIDs). In such cases, the PEs 415 participating in that VLAN for a given EVPN instance are responsible 416 for performing VLAN ID translation to/from locally attached CE 417 devices. 419 If a VLAN is represented by a single VID across all PE devices 420 participating in that VLAN for that EVPN instance, then there is no 421 need for VID translation at the PEs. Furthermore, some deployment 422 scenarios guarantee uniqueness of VIDs across all EVPN instances; 423 all points of attachment for a given EVPN instance use the same VID 424 and no other EVPN instances use that VID. This allows the RT(s) for 425 each EVPN instance to be derived automatically from the corresponding 426 VID, as described in section 7.10.1. 428 The following subsections discuss the relationship between broadcast 429 domains (e.g., VLANs), Ethernet Tag IDs (e.g., VIDs), and MAC-VRFs as 430 well as the setting of the Ethernet Tag ID, in the various EVPN BGP 431 routes (defined in section 8), for the different types of service 432 interfaces described in [EVPN-REQ]. 434 The following value of Ethernet Tag ID is reserved: 436 - Ethernet Tag ID {0xFFFFFFFF} is known as MAX-ET 438 6.1 VLAN Based Service Interface 440 With this service interface, an EVPN instance consists of only a 441 single broadcast domain (e.g., a single VLAN). Therefore, there is a 442 one to one mapping between a VID on this interface and a MAC-VRF. 443 Since a MAC-VRF corresponds to a single VLAN, it consists of a single 444 bridge domain corresponding to that VLAN. If the VLAN is represented 445 by multiple VIDs (e.g., a different VID per Ethernet Segment per PE), 446 then each PE needs to perform VID translation for frames destined to 447 its Ethernet Segment(s). In such scenarios, the Ethernet frames 448 transported over MPLS/IP network SHOULD remain tagged with the 449 originating VID and a VID translation MUST be supported in the data 450 path and MUST be performed on the disposition PE. The Ethernet Tag ID 451 in all EVPN routes MUST be set to 0. 453 6.2 VLAN Bundle Service Interface 455 With this service interface, an EVPN instance corresponds to several 456 broadcast domains (e.g., several VLANs); however, only a single 457 bridge domain is maintained per MAC-VRF which means multiple VLANs 458 share the same bridge domain. This implies MAC addresses MUST be 459 unique across different VLANs for this service to work. In other 460 words, there is a many-to-one mapping between VLANs and a MAC-VRF, 461 and the MAC-VRF consists of a single bridge domain. Furthermore, a 462 single VLAN must be represented by a single VID - e.g., no VID 463 translation is allowed for this service interface type. The MPLS 464 encapsulated frames MUST remain tagged with the originating VID. Tag 465 translation is NOT permitted. The Ethernet Tag ID in all EVPN routes 466 MUST be set to 0. 468 6.2.1 Port Based Service Interface 470 This service interface is a special case of the VLAN Bundle service 471 interface, where all of the VLANs on the port are part of the same 472 service and map to the same bundle. The procedures are identical to 473 those described in section 6.2. 475 6.3 VLAN Aware Bundle Service Interface 477 With this service interface, an EVPN instance consists of several 478 broadcast domains (e.g., several VLANs) with each VLAN having its own 479 bridge domain - i.e., multiple bridge domains (one per VLAN) is 480 maintained by a single MAC-VRF corresponding to the EVPN instance. In 481 the case where a single VLAN is represented by different VIDs on 482 different CEs and thus VID translation is required, a normalized 483 Ethernet Tag ID (VID) MUST be carried in the MPLS encapsulated frames 484 and a Ethernet Tag ID translation function MUST be supported in the 485 data path. This translation MUST be performed in data path on both 486 the imposition as well as the disposition PEs (translating to 487 normalized Ethernet Tag ID on imposition PE and translating to local 488 Ethernet Tag ID on disposition PE). The Ethernet Tag ID in all EVPN 489 routes MUST be set to the normalized value assigned by the EVPN 490 provider. 492 6.3.1 Port Based VLAN Aware Service Interface 494 This service interface is a special case of the VLAN Aware Bundle 495 service interface, where all of the VLANs on the port are part of the 496 same service and are mapped to a single bundle but without any VID 497 translation. The procedures are subset of those described in section 498 6.3. 500 7. BGP EVPN NLRI 502 This document defines a new BGP NLRI, called the EVPN NLRI. 504 Following is the format of the EVPN NLRI: 506 +-----------------------------------+ 507 | Route Type (1 octet) | 508 +-----------------------------------+ 509 | Length (1 octet) | 510 +-----------------------------------+ 511 | Route Type specific (variable) | 512 +-----------------------------------+ 514 The Route Type field defines encoding of the rest of the EVPN NLRI 515 (Route Type specific EVPN NLRI). 517 The Length field indicates the length in octets of the Route Type 518 specific field of EVPN NLRI. 520 This document defines the following Route Types: 522 + 1 - Ethernet Auto-Discovery (A-D) route 523 + 2 - MAC/IP advertisement route 524 + 3 - Inclusive Multicast Ethernet Tag Route 525 + 4 - Ethernet Segment Route 527 The detailed encoding and procedures for these route types are 528 described in subsequent sections. 530 The EVPN NLRI is carried in BGP [RFC4271] using BGP Multiprotocol 531 Extensions [RFC4760] with an AFI of 25 (L2VPN) and a SAFI of 70 532 (EVPN). The NLRI field in the MP_REACH_NLRI/MP_UNREACH_NLRI attribute 533 contains the EVPN NLRI (encoded as specified above). 535 In order for two BGP speakers to exchange labeled EVPN NLRI, they 536 must use BGP Capabilities Advertisement to ensure that they both are 537 capable of properly processing such NLRI. This is done as specified 538 in [RFC4760], by using capability code 1 (multiprotocol BGP) with an 539 AFI of 25 (L2VPN) and a SAFI of 70 (EVPN). 541 7.1. Ethernet Auto-Discovery Route 543 A Ethernet A-D route type specific EVPN NLRI consists of the 544 following: 546 +---------------------------------------+ 547 | Route Distinguisher (RD) (8 octets) | 548 +---------------------------------------+ 549 |Ethernet Segment Identifier (10 octets)| 550 +---------------------------------------+ 551 | Ethernet Tag ID (4 octets) | 552 +---------------------------------------+ 553 | MPLS Label (3 octets) | 554 +---------------------------------------+ 556 For the purpose of BGP route key processing, only the Ethernet 557 Segment Identifier and the Ethernet Tag ID are considered to be part 558 of the prefix in the NLRI. The MPLS Label field is to be treated as 559 a route attribute as opposed to being part of the route. 561 For procedures and usage of this route please see section 8.2 "Fast 562 Convergence" and section 8.4 "Aliasing". 564 7.2. MAC/IP Advertisement Route 566 A MAC/IP advertisement route type specific EVPN NLRI consists of the 567 following: 569 +---------------------------------------+ 570 | RD (8 octets) | 571 +---------------------------------------+ 572 |Ethernet Segment Identifier (10 octets)| 573 +---------------------------------------+ 574 | Ethernet Tag ID (4 octets) | 575 +---------------------------------------+ 576 | MAC Address Length (1 octet) | 577 +---------------------------------------+ 578 | MAC Address (6 octets) | 579 +---------------------------------------+ 580 | IP Address Length (1 octet) | 581 +---------------------------------------+ 582 | IP Address (0 or 4 or 16 octets) | 583 +---------------------------------------+ 584 | MPLS Label1 (3 octets) | 585 +---------------------------------------+ 586 | MPLS Label2 (0 or 3 octets) | 587 +---------------------------------------+ 589 For the purpose of BGP route key processing, only the Ethernet Tag 590 ID, MAC Address Length, MAC Address, IP Address Length, and IP 591 Address Address fields are considered to be part of the prefix in the 592 NLRI. The Ethernet Segment Identifier and MPLS Label1 and MPLS Label2 593 fields are to be treated as route attributes as opposed to being part 594 of the "route". The IP address length is in bits. 596 For procedures and usage of this route please see section 9 597 "Determining Reachability to Unicast MAC Addresses" and section 14 598 "Load Balancing of Unicast Packets". 600 7.3. Inclusive Multicast Ethernet Tag Route 602 An Inclusive Multicast Ethernet Tag route type specific EVPN NLRI 603 consists of the following: 605 +---------------------------------------+ 606 | RD (8 octets) | 607 +---------------------------------------+ 608 | Ethernet Tag ID (4 octets) | 609 +---------------------------------------+ 610 | IP Address Length (1 octet) | 611 +---------------------------------------+ 612 | Originating Router's IP Addr | 613 | (4 or 16 octets) | 614 +---------------------------------------+ 616 For procedures and usage of this route please see section 11 617 "Handling of Multi-Destination Traffic", section 13 "Processing of 618 Unknown Unicast Traffic" and section 16 "Multicast". The IP address 619 length is in bits. For the purpose of BGP route key processing, only 620 the Ethernet Tag ID, IP Address Length, and Originating Router's IP 621 Address fields are considered to be part of the prefix in the NLRI. 623 7.4 Ethernet Segment Route 625 An Ethernet Segment route type specific EVPN NLRI consists of the 626 following: 628 +---------------------------------------+ 629 | RD (8 octets) | 630 +---------------------------------------+ 631 |Ethernet Segment Identifier (10 octets)| 632 +---------------------------------------+ 633 | IP Address Length (1 octet) | 634 +---------------------------------------+ 635 | Originating Router's IP Addr | 636 | (4 or 16 octets) | 637 +---------------------------------------+ 639 For procedures and usage of this route please see section 8.5 640 "Designated Forwarder Election". The IP address length is in bits. 641 For the purpose of BGP route key processing, only the Ethernet 642 Segment ID, IP Address Length, and Originating Router's IP Address 643 fields are considered to be part of the prefix in the NLRI. 645 7.5 ESI Label Extended Community 647 This extended community is a new transitive extended community with 648 the Type field is 0x06, and the Sub-Type of 0x01. It may be 649 advertised along with Ethernet Auto-Discovery routes and it enables 650 split-horizon procedures for multi-homed sites as described in 651 section 8.3 "Split Horizon". ESI Label represents an ES by the 652 advertising PE and it is used in split-horizon filtering by other PEs 653 that are connected to the same multi-homed Ethernet Segment. 655 Each ESI Label Extended Community is encoded as a 8-octet value as 656 follows: 658 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 659 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 660 | Type=0x06 | Sub-Type=0x01 | Flags(1 Octet)| Reserved=0 | 661 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 662 | Reserved = 0 | ESI Label | 663 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 665 The low order bit of the flags octet is defined as the "Single- 666 Active" bit. A value of 0 means that the multi-homed site is 667 operating in All-Active redundancy mode and a value of 1 means that 668 the multi-homed site is operating in Single-Active redundancy mode. 670 7.6 ES-Import Route Target 672 This is a new transitive Route Target extended community carried with 673 the Ethernet Segment route. When used, it enables all the PEs 674 connected to the same multi-homed site to import the Ethernet Segment 675 routes. The value is derived automatically from the ESI by encoding 676 the high order 6-octet portion of the 9-octet ESI Value in the ES- 677 Import Route Target. The high order 6-octet of the ESI incorporates 678 MAC address of ESI (for type 1, 2, and 3) which when encoded in this 679 RT and used in the RT constrain feature, it enables proper route- 680 target filtering. The format of this extended community is as 681 follows: 683 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 684 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 685 | Type=0x06 | Sub-Type=0x02 | ES-Import | 686 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 687 | ES-Import Cont'd | 688 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 690 This document expands the definition of the Route Target extended 691 community to allow the value of high order octet (Type field) to be 692 0x06 (in addition to the values specified in rfc4360). The value of 693 low order octet (Sub-Type field) of 0x02 indicates that this extended 694 community is of type "Route Target". The new value for Type field of 695 0x06 indicates that the structure of this RT is a six-octet value 696 (e.g., a MAC address). A BGP speaker that implements RT-Constrain 697 [RFC4684] MUST apply the RT Constraint procedures to the ES-import RT 698 as well. 700 For procedures and usage of this attribute, please see section 8.1 701 "Multi-homed Ethernet Segment Auto-Discovery". 703 7.7 MAC Mobility Extended Community 705 This extended community is a new transitive extended community with 706 the Type field of 0x06 and the Sub-Type of 0x00. It may be advertised 707 along with MAC Advertisement routes. The procedures for using this 708 Extended Community are described in section 15 "MAC Mobility". 710 The MAC Mobility Extended Community is encoded as an 8-octet value as 711 follows: 713 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 714 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 715 | Type=0x06 | Sub-Type=0x00 |Flags(1 octet)| Reserved=0 | 716 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 717 | Sequence Number | 718 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 720 The low order bit of the flags octet is defined as the 721 "Sticky/static" flag and may be set to 1. A value of 1 means that the 722 MAC address is static and cannot move. The sequence number is used to 723 ensure that PEs retain the correct MAC advertisement route when 724 multiple updates occur for the same MAC address. 726 7.8 Default Gateway Extended Community 728 The Default Gateway community is an Extended Community of an Opaque 729 Type (see 3.3 of rfc4360). It is a transitive community, which means 730 that the first octet is 0x03. The value of the second octet (Sub- 731 Type) is 0x0d (Default Gateway) as assigned by IANA. The Value field 732 of this community is reserved (set to 0 by the senders, ignored by 733 the receivers). 735 7.9 Route Distinguisher Assignment per EVI 737 Route Distinguisher (RD) MUST be set to the RD of the EVI that is 738 advertising the NLRI. An RD MUST be assigned for a given EVI on a PE. 739 This RD MUST be unique across all EVIs on a PE. It is RECOMMENDED to 740 use the Type 1 RD [RFC4364]. The value field comprises an IP address 741 of the PE (typically, the loopback address) followed by a number 742 unique to the PE. This number may be generated by the PE. Or in the 743 Unique VLAN EVPN case, the low order 12 bits may be the 12 bit VLAN 744 ID, with the remaining high order 4 bits set to 0. 746 7.10 Route Targets 748 The EVPN route MAY carry one or more Route Target (RT) attributes. 749 RTs may be configured (as in IP VPNs), or may be derived 750 automatically. 752 If a PE uses RT-Constrain, the PE advertises all such RTs using RT 753 Constraints per [RFC4684]. The use of RT Constrains allows each 754 Ethernet A-D route to reach only those PEs that are configured to 755 import at least one RT from the set of RTs carried in the EVPN route. 757 7.10.1 Auto-Derivation from the Ethernet Tag ID 759 For the "Unique VLAN EVPN" scenario, it is highly desirable to auto- 760 derive the RT from the Ethernet Tag ID (VLAN ID) for that EVPN 761 instance. The following is the procedure for performing such auto- 762 derivation. 764 + The Global Administrator field of the RT MUST be set 765 to the Autonomous System (AS) number that the PE is 766 associated with. 768 + The 12-bit VLAN ID MUST be encoded in the lowest 12 bits of 769 the Local Administrator field. 771 8. Multi-homing Functions 773 This section discusses the functions, procedures and associated BGP 774 routes used to support multi-homing in EVPN. This covers both multi- 775 homed device (MHD) as well as multi-homed network (MHN) scenarios. 777 8.1 Multi-homed Ethernet Segment Auto-Discovery 779 PEs connected to the same Ethernet segment can automatically discover 780 each other with minimal to no configuration through the exchange of 781 the Ethernet Segment route. 783 8.1.1 Constructing the Ethernet Segment Route 785 The Route-Distinguisher (RD) MUST be a Type 1 RD [RFC4364]. The value 786 field comprises an IP address of the PE (typically, the loopback 787 address) followed by 0's. 789 The Ethernet Segment Identifier (ESI) MUST be set to the ten octet 790 value described in section 5. 792 The BGP advertisement that advertises the Ethernet Segment route MUST 793 also carry an ES-Import route target, as defined in section 7.6. 795 The Ethernet Segment Route filtering MUST be done such that the 796 Ethernet Segment Route is imported only by the PEs that are multi- 797 homed to the same Ethernet Segment. To that end, each PE that is 798 connected to a particular Ethernet segment constructs an import 799 filtering rule to import a route that carries the ES-Import extended 800 community, constructed from the ESI. 802 8.2 Fast Convergence 804 In EVPN, MAC address reachability is learnt via the BGP control-plane 805 over the MPLS network. As such, in the absence of any fast protection 806 mechanism, the network convergence time is a function of the number 807 of MAC Advertisement routes that must be withdrawn by the PE 808 encountering a failure. For highly scaled environments, this scheme 809 yields slow convergence. 811 To alleviate this, EVPN defines a mechanism to efficiently and 812 quickly signal, to remote PE nodes, the need to update their 813 forwarding tables upon the occurrence of a failure in connectivity to 814 an Ethernet segment. This is done by having each PE advertise a set 815 of one or more Ethernet A-D per Ethernet segment (per ES) routes for 816 each locally attached Ethernet segment (refer to section 8.2.1 below 817 for details on how these routes are constructed). The reason that a 818 PE may need to advertise more than one Ethernet A-D per ES route for 819 a given ES is that the ES may be in a multiplicity of EVIs and the 820 RTs for all of these EVIs may not fit into a single route. 821 Advertising a set of Ethernet A-D per ES routes for the ES allows 822 each route to contain a subset of the complete set of RTs. 824 Upon a failure in connectivity to the attached segment, the PE 825 withdraws the corresponding set of Ethernet A-D Per ES routes. This 826 triggers all PEs that receive the withdrawal to update their next-hop 827 adjacencies for all MAC addresses associated with the Ethernet 828 segment in question. If no other PE had advertised an Ethernet A-D 829 route for the same segment, then the PE that received the withdrawal 830 simply invalidates the MAC entries for that segment. Otherwise, the 831 PE updates the next-hop adjacencies to point to the backup PE(s). 833 8.2.1 Constructing the Ethernet A-D per Ethernet Segment (ES) Route 835 This section describes the procedures used to construct the Ethernet 836 A-D per ES route, which is used for fast convergence (as discussed 837 above) and for advertising the ESI label used for split-horizon 838 filtering (as discussed in section 8.3). Support of this route is 839 REQUIRED. 841 The Route-Distinguisher (RD) MUST be a Type 1 RD [RFC4364]. The value 842 field comprises an IP address of the PE (typically, the loopback 843 address) followed by a number unique to the PE. 845 The Ethernet Segment Identifier MUST be a ten octet entity as 846 described in section "Ethernet Segment". The Ethernet A-D route is 847 not needed when the Segment Identifier is set to 0 (e.g., single- 848 homed scenarios). 850 The Ethernet Tag ID MUST be set to MAX-ET. 852 The MPLS label in the NLRI MUST be set to 0. 854 The "ESI Label Extended Community" MUST be included in the route. If 855 All-Active redundancy mode is desired, then the "Single-Active" bit 856 in the flags of the ESI Label Extended Community MUST be set to 0 and 857 the MPLS label in that extended community MUST be set to a valid MPLS 858 label value. The MPLS label in this Extended Community is referred to 859 as the ESI label and MUST have the same value in each Ethernet A-D 860 per ES route advertised for the ES. This label MUST be a downstream 861 assigned MPLS label if the advertising PE is using ingress 862 replication for receiving multicast, broadcast or unknown unicast 863 traffic from other PEs. If the advertising PE is using P2MP MPLS LSPs 864 for sending multicast, broadcast or unknown unicast traffic, then 865 this label MUST be an upstream assigned MPLS label. The usage of this 866 label is described in section 8.3. 868 If Single-Active redundancy mode is desired, then the "Single-Active" 869 bit in the flags of the ESI Label Extended Community MUST be set to 1 870 and the ESI label SHOULD be set to a valid MPLS label value. 872 8.2.1.1. Ethernet A-D Route Targets 874 Each Ethernet A-D per ES route MUST carry one or more Route Target 875 (RT) attributes. The set of Ethernet A-D routes per ES MUST carry the 876 entire set of RTs for all the EVPN instances to which the Ethernet 877 Segment belongs. 879 8.3 Split Horizon 881 Consider a CE that is multi-homed to two or more PEs on an Ethernet 882 segment ES1 operating in All-Active redundancy mode. If the CE sends 883 a broadcast, unknown unicast, or multicast (BUM) packet to one of the 884 non-Designated Forwarder (non-DF) PEs, say PE1, then PE1 will forward 885 that packet to all or subset of the other PEs in that EVPN instance 886 including the Designated Forwarder (DF) PE for that Ethernet segment. 887 In this case the DF PE that the CE is multi-homed to MUST drop the 888 packet and not forward back to the CE. This filtering is referred to 889 as "split horizon" filtering in this document. 891 When a set of PEs operating in Single-Active redundancy mode, the use 892 of this split-horizon filtering mechanism is highly recommended 893 because it prevents transient loop at the time of failure or recovery 894 impacting the Ethernet Segment - e.g., when two PEs thinks that both 895 are DFs for that segment before DF election procedure settles down. 897 In order to achieve this split horizon function, every BUM packet 898 originating from a non-DF PE is encapsulated with an MPLS label that 899 identifies the Ethernet segment of origin (i.e. the segment from 900 which the frame entered the EVPN network). This label is referred to 901 as the ESI label, and MUST be distributed by all PEs when operating 902 in All-Active redundancy mode using a set of Ethernet A-D per ES 903 routes per section 8.2.1 above. The ESI label SHOULD be distributed 904 by all PEs when operating in Single-Active redundancy mode using a 905 set of Ethernet A-D per ES route. This route is imported by the PEs 906 connected to the Ethernet Segment and also by the PEs that have at 907 least one EVPN instance in common with the Ethernet Segment in the 908 route. As described in section 8.1.1, the route MUST carry an ESI 909 Label Extended Community with a valid ESI label. The disposition PE 910 rely on the value of the ESI label to determine whether or not a BUM 911 frame is allowed to egress a specific Ethernet segment. 913 8.3.1 ESI Label Assignment 915 The following subsections describe the assignment procedures for the 916 ESI label, which differ depending on the type of tunnels being used 917 to deliver multi-destination packets in the EVPN network. 919 8.3.1.1 Ingress Replication 921 Each PE attached to a given ES that is operating in All-Active or 922 Single-Active redundancy mode and that uses ingress replication to 923 receive BUM traffic advertises a downstream assigned ESI label in the 924 set of Ethernet A-D per ES routes for that ES. This label MUST be 925 programmed in the platform label space by the advertising PE and the 926 forwarding entry for this label must result in NOT forwarding packets 927 received with this label onto the Ethernet segment for which the 928 label was distributed. 930 The rules for the inclusion of the ESI label in a BUM packet by the 931 ingress PE operating in All-Active redundancy mode are as follows: 933 A non-DF ingress PE MUST include the ESI label distributed by the DF 934 egress PE in the copy of a BUM packet sent to it. 936 An ingress PE (DF or non-DF) SHOULD include the ESI label distributed 937 by each non-DF egress PE in the copy of a BUM packet sent to it. 939 The rules for the inclusion of the ESI label in a BUM packet by the 940 ingress PE operating in Single-Active redundancy mode are as follows: 942 An ingress DF PE SHOULD include the ESI label distributed by the 943 egress PE in the copy of a BUM packet sent to it. 945 In both All-Active and Single-Active redundancy mode, an ingress PE 946 MUST NOT include an ESI label in the copy of a BUM packet sent to an 947 egress PE that is not attached to the ES through which the BUM packet 948 entered the EVI. 950 As an example, consider PE1 and PE2 that are multi-homed to CE1 on 951 ES1 and operating in All-Active multi-homing mode. Further consider 952 that PE1 is using P2P or MP2P LSPs to send packets to PE2. Consider 953 that PE1 is the non-DF for VLAN1 and PE2 is the DF for VLAN1, and PE1 954 receives a BUM packet from CE1 on VLAN1 on ES1. In this scenario, PE2 955 distributes an Inclusive Multicast Ethernet Tag route for VLAN1 956 corresponding to an EVPN instance. So, when PE1 sends a BUM packet, 957 that it receives from CE1, it MUST first push onto the MPLS label 958 stack the ESI label that PE2 has distributed for ES1. It MUST then 959 push on the MPLS label distributed by PE2 in the Inclusive Multicast 960 Ethernet Tag route for VLAN1. The resulting packet is further 961 encapsulated in the P2P or MP2P LSP label stack required to transmit 962 the packet to PE2. When PE2 receives this packet, it determines the 963 set of ESIs to replicate the packet to from the top MPLS label, after 964 any P2P or MP2P LSP labels have been removed. If the next label is 965 the ESI label assigned by PE2 for ES1, then PE2 MUST NOT forward the 966 packet onto ES1. If the next label is an ESI label which has not been 967 assigned by PE2, then PE2 MUST drop the packet. It should be noted 968 that in this scenario, if PE2 receives a BUM packet for VLAN1 from 969 CE1, then it SHOULD encapsulate the packet with an ESI label received 970 from PE1 when sending it to PE1 in order to avoid any transient loop 971 during a failure scenario impacting ES1 (e.g., port or link failure). 973 8.3.1.2. P2MP MPLS LSPs 975 The non-DF PEs attached to a given ES that is operating in All-Active 976 redundancy mode and that use P2MP LSPs to send BUM traffic advertise 977 an upstream assigned ESI label in the set of Ethernet A-D per ES 978 routes for that ES. This label is upstream assigned by the PE that 979 advertises the route. This label MUST be programmed by the other PEs, 980 that are connected to the ESI advertised in the route, in the context 981 label space for the advertising PE. Further the forwarding entry for 982 this label must result in NOT forwarding packets received with this 983 label onto the Ethernet segment that the label was distributed for. 984 This label MUST also be programmed by the other PEs, that import the 985 route but are not connected to the ESI advertised in the route, in 986 the context label space for the advertising PE. Further the 987 forwarding entry for this label must be a POP with no other 988 associated action. 990 The DF PE attached to a given ES that is operating in Single-Active 991 redundancy mode and that use P2MP LSPs to send BUM traffic should 992 advertise an upstream assigned ESI label in the set of Ethernet A-D 993 per ES routes for that ES just as above paragraph. 995 As an example, consider PE1 and PE2 that are multi-homed to CE1 on 996 ES1 and operating in All-Active multi-homing mode. Also consider PE3 997 belongs to one of the EVPN instances of ES1. Further, assume that 998 PE1 which is the non-DF, using P2MP MPLS LSPs to send BUM packets. 999 When PE1 sends a BUM packet, that it receives from CE1, it MUST first 1000 push onto the MPLS label stack the ESI label that it has assigned for 1001 the ESI that the packet was received on. The resulting packet is 1002 further encapsulated in the P2MP MPLS label stack necessary to 1003 transmit the packet to the other PEs. Penultimate hop popping MUST be 1004 disabled on the P2MP LSPs used in the MPLS transport infrastructure 1005 for EVPN. When PE2 receives this packet, it de-capsulates the top 1006 MPLS label and forwards the packet using the context label space 1007 determined by the top label. If the next label is the ESI label 1008 assigned by PE1 to ES1, then PE2 MUST NOT forward the packet onto 1009 ES1. When PE3 receives this packet, it de-capsulates the top MPLS 1010 label and forwards the packet using the context label space 1011 determined by the top label. If the next label is the ESI label 1012 assigned by PE1 to ES1 and PE3 is not connected to ES1, then PE3 MUST 1013 pop the label and flood the packet over all local ESIs in that EVPN 1014 instance. It should be noted that when PE2 sends a BUM frame over a 1015 P2MP LSP, it should encapsulate the frame with an ESI label even 1016 though it is the DF for that VLAN in order to avoid any transient 1017 loop during a failure scenario impacting ES1 (e.g., port or link 1018 failure). 1020 8.4 Aliasing and Backup-Path 1022 In the case where a CE is multi-homed to multiple PE nodes, using a 1023 LAG with All-Active redundancy, it is possible that only a single PE 1024 learns a set of the MAC addresses associated with traffic transmitted 1025 by the CE. This leads to a situation where remote PE nodes receive 1026 MAC advertisement routes, for these addresses, from a single PE even 1027 though multiple PEs are connected to the multi-homed segment. As a 1028 result, the remote PEs are not able to effectively load-balance 1029 traffic among the PE nodes connected to the multi-homed Ethernet 1030 segment. This could be the case, for e.g. when the PEs perform data- 1031 plane learning on the access, and the load-balancing function on the 1032 CE hashes traffic from a given source MAC address to a single PE. 1033 Another scenario where this occurs is when the PEs rely on control 1034 plane learning on the access (e.g. using ARP), since ARP traffic will 1035 be hashed to a single link in the LAG. 1037 To address this issue, EVPN introduces the concept of 'Aliasing' 1038 which is the ability of a PE to signal that it has reachability to an 1039 EVPN instance on a given ES even when it has learnt no MAC addresses 1040 from that EVI/ES. The Ethernet A-D per EVI route is used for this 1041 purpose. A remote PE that receives a MAC advertisement route with 1042 non-reserved ESI SHOULD consider the advertised MAC address to be 1043 reachable via all PEs that have advertised reachability to that MAC 1044 address' EVI/ES via the combination of an Ethernet A-D per EVI route 1045 for that EVI/ES (and Ethernet Tag if applicable) AND Ethernet A-D per 1046 ES routes for that ES with the 'Single-Active' bit in the flags of 1047 the ESI Label Extended Community set to 0. 1049 Note that the Ethernet A-D per EVI route may be received by a remote 1050 PE before it receives the set of Ethernet A-D per ES routes. 1051 Therefore, in order to handle corner cases and race conditions, the 1052 Ethernet A-D per EVI route MUST NOT be used for traffic forwarding by 1053 a remote PE until it also receives the associated set of Ethernet A-D 1054 per ES routes. 1056 Backup-path is a closely related function, but it is used in Single- 1057 Active redundancy mode. In this case a PE also advertises that it 1058 has reachability to a give EVI/ES using same combination of Ethernet 1059 A-D per EVI route and Ethernet A-D per ES route as above, but with 1060 the 'Single-Active' bit in the flags of the ESI Label Extended 1061 Community set to 1. A remote PE that receives a MAC advertisement 1062 route with non-reserved ESI SHOULD consider the advertised MAC 1063 address to be reachable via any PE that has advertised this 1064 combination of Ethernet A-D routes and it SHOULD install a backup- 1065 path for that MAC address. 1067 8.4.1 Constructing the Ethernet A-D per EVPN Instance (EVI) Route 1069 This section describes the procedures used to construct the Ethernet 1070 A-D per EVPN Instance (EVI) route, which is used for aliasing (as 1071 discussed above). Support of this route is OPTIONAL. 1073 Route-Distinguisher (RD) MUST be set to the RD of the EVI that is 1074 advertising the NLRI per section 7.9. 1076 The Ethernet Segment Identifier MUST be a ten octet entity as 1077 described in section "Ethernet Segment Identifier". The Ethernet A-D 1078 route is not needed when the Segment Identifier is set to 0. 1080 The Ethernet Tag ID is the identifier of an Ethernet Tag on the 1081 Ethernet segment. This value may be a 12 bit VLAN ID, in which case 1082 the low order 12 bits are set to the VLAN ID and the high order 20 1083 bits are set to 0. Or it may be another Ethernet Tag used by the 1084 EVPN. It MAY be set to the default Ethernet Tag on the Ethernet 1085 segment or to the value 0. 1087 Note that the above allows the Ethernet A-D route to be advertised 1088 with one of the following granularities: 1090 + One Ethernet A-D route for a given tuple 1091 per EVI. This is applicable when the PE uses MPLS-based 1092 disposition. 1094 + One Ethernet A-D route per (where the Ethernet 1095 Tag ID is set to 0). This is applicable when the PE uses 1096 MAC-based disposition, or when the PE uses MPLS-based 1097 disposition when no VLAN translation is required. 1099 The usage of the MPLS label is described in the section on "Load 1100 Balancing of Unicast Packets". 1102 The Next Hop field of the MP_REACH_NLRI attribute of the route MUST 1103 be set to the IPv4 or IPv6 address of the advertising PE. 1105 The Ethernet A-D route MUST carry one or more Route Target (RT) 1106 attributes per section 7.10. 1108 8.5 Designated Forwarder Election 1110 Consider a CE that is a host or a router that is multi-homed directly 1111 to more than one PE in an EVPN instance on a given Ethernet segment. 1112 One or more Ethernet Tags may be configured on the Ethernet segment. 1113 In this scenario only one of the PEs, referred to as the Designated 1114 Forwarder (DF), is responsible for certain actions: 1116 - Sending multicast and broadcast traffic, on a given Ethernet 1117 Tag on a particular Ethernet segment, to the CE. 1119 - Flooding unknown unicast traffic (i.e. traffic for 1120 which a PE does not know the destination MAC address), 1121 on a given Ethernet Tag on a particular Ethernet segment 1122 to the CE, if the environment requires flooding of 1123 unknown unicast traffic. 1125 Note that this behavior, which allows selecting a DF at the 1126 granularity of for multicast, broadcast and unknown 1127 unicast traffic, is the default behavior in this specification. 1129 Note that a CE always sends packets belonging to a specific flow 1130 using a single link towards a PE. For instance, if the CE is a host 1131 then, as mentioned earlier, the host treats the multiple links that 1132 it uses to reach the PEs as a Link Aggregation Group (LAG). The CE 1133 employs a local hashing function to map traffic flows onto links in 1134 the LAG. 1136 If a bridged network is multi-homed to more than one PE in an EVPN 1137 network via switches, then the support of All-Active redundancy mode 1138 requires the bridged network to be connected to two or more PEs using 1139 a LAG. 1141 If a bridged network does not connect to the PEs using LAG, then only 1142 one of the links between the switched bridged network and the PEs 1143 must be the active link for a given EVPN instance. In this case, the 1144 set of Ethernet A-D per ES routes advertised by each PE MUST have the 1145 'Single-Active' bit in the flags of the ESI Label Extended Community 1146 set to 1. 1148 The default procedure for DF election at the granularity of is referred to as "service carving". With service carving, it is 1150 possible to elect multiple DFs per Ethernet Segment (one per EVI) in 1151 order to perform load-balancing of multi-destination traffic destined 1152 to a given Segment. The load-balancing procedures carve up the EVI 1153 space among the PE nodes evenly, in such a way that every PE is the 1154 DF for a disjoint set of EVIs. The procedure for service carving is 1155 as follows: 1157 1. When a PE discovers the ESI of the attached Ethernet Segment, it 1158 advertises an Ethernet Segment route with the associated ES-Import 1159 extended community attribute. 1161 2. The PE then starts a timer (default value = 3 seconds) to allow 1162 the reception of Ethernet Segment routes from other PE nodes 1163 connected to the same Ethernet Segment. This timer value MUST be same 1164 across all PEs connected to the same Ethernet Segment. 1166 3. When the timer expires, each PE builds an ordered list of the IP 1167 addresses of all the PE nodes connected to the Ethernet Segment 1168 (including itself), in increasing numeric value. Each IP address in 1169 this list is extracted from the "Originator Router's IP address" 1170 field of the advertised Ethernet Segment route. Every PE is then 1171 given an ordinal indicating its position in the ordered list, 1172 starting with 0 as the ordinal for the PE with the numerically lowest 1173 IP address. The ordinals are used to determine which PE node will be 1174 the DF for a given EVPN instance on the Ethernet Segment using the 1175 following rule: Assuming a redundancy group of N PE nodes, the PE 1176 with ordinal i is the DF for an EVPN instance with an associated 1177 Ethernet Tag value V when (V mod N) = i. In the case where multiple 1178 Ethernet Tags are associated with a single EVPN instance, then the 1179 numerically lowest Ethernet Tag value in that EVPN instance MUST be 1180 used in the modulo function. 1182 It should be noted that using "Originator Router's IP address" field 1183 in the Ethernet Segment route to get the PE IP address needed for the 1184 ordered list, allows for a CE to be multi-homed across different ASes 1185 if such need ever arises. 1187 4. The PE that is elected as a DF for a given EVPN instance will 1188 unblock traffic for the Ethernet Tags associated with that EVPN 1189 instance. Note that the DF PE unblocks multi-destination traffic in 1190 the egress direction towards the Segment. All non-DF PEs continue to 1191 drop multi-destination traffic (for the associated EVPN instances) in 1192 the egress direction towards the Segment. 1194 In the case of link or port failure, the affected PE withdraws its 1195 Ethernet Segment route. This will re-trigger the service carving 1196 procedures on all the PEs in the RG. For PE node failure, or upon PE 1197 commissioning or decommissioning, the PEs re-trigger the service 1198 carving. In case of a Single-Active multi-homing, when a service 1199 moves from one PE in the RG to another PE as a result of re-carving, 1200 the PE, which ends up being the elected DF for the service, SHOULD 1201 trigger a MAC address flush notification towards the associated 1202 Ethernet Segment. This can be done, for e.g. using IEEE 802.1ak MVRP 1203 'new' declaration. 1205 8.6. Interoperability with Single-homing PEs 1207 Let's refer to PEs that only support single-homed CE devices as 1208 single-homing PEs. For single-homing PEs, all the above multi-homing 1209 procedures can be omitted; however, to allow for single-homing PEs to 1210 fully inter-operate with multi-homing PEs, some of the multi-homing 1211 procedures described above SHOULD be supported even by single-homing 1212 PEs: 1214 - procedures related to processing Ethernet A-D route for the purpose 1215 of Fast Convergence (8.2 Fast Convergence), to let single-homing PEs 1216 benefit from fast convergence 1218 - procedures related to processing Ethernet A-D route for the purpose 1219 of Aliasing (8.4 Aliasing and Backup-path), to let single-homing PEs 1220 benefit from load balancing 1222 - procedures related to processing Ethernet A-D route for the purpose 1223 of Backup-path (8.4 Aliasing and Backup-path), to let single-homing 1224 PEs to benefit from the corresponding convergence improvement 1226 9. Determining Reachability to Unicast MAC Addresses 1228 PEs forward packets that they receive based on the destination MAC 1229 address. This implies that PEs must be able to learn how to reach a 1230 given destination unicast MAC address. 1232 There are two components to MAC address learning, "local learning" 1233 and "remote learning": 1235 9.1. Local Learning 1237 A particular PE must be able to learn the MAC addresses from the CEs 1238 that are connected to it. This is referred to as local learning. 1240 The PEs in a particular EVPN instance MUST support local data plane 1241 learning using standard IEEE Ethernet learning procedures. A PE must 1242 be capable of learning MAC addresses in the data plane when it 1243 receives packets such as the following from the CE network: 1245 - DHCP requests 1247 - ARP request for its own MAC. 1249 - ARP request for a peer. 1251 Alternatively PEs MAY learn the MAC addresses of the CEs in the 1252 control plane or via management plane integration between the PEs and 1253 the CEs. 1255 There are applications where a MAC address that is reachable via a 1256 given PE on a locally attached Segment (e.g. with ESI X) may move 1257 such that it becomes reachable via another PE on another Segment 1258 (e.g. with ESI Y). This is referred to as a "MAC Mobility". 1259 Procedures to support this are described in section "MAC Mobility". 1261 9.2. Remote learning 1263 A particular PE must be able to determine how to send traffic to MAC 1264 addresses that belong to or are behind CEs connected to other PEs 1265 i.e. to remote CEs or hosts behind remote CEs. We call such MAC 1266 addresses "remote" MAC addresses. 1268 This document requires a PE to learn remote MAC addresses in the 1269 control plane. In order to achieve this, each PE advertises the MAC 1270 addresses it learns from its locally attached CEs in the control 1271 plane, to all the other PEs in that EVPN instance, using MP-BGP and 1272 specifically the MAC Advertisement route. 1274 9.2.1. Constructing the BGP EVPN MAC/IP Address Advertisement 1276 BGP is extended to advertise these MAC addresses using the MAC/IP 1277 Advertisement route type in the EVPN NLRI. 1279 The RD MUST be the RD of the EVI that is advertising the NLRI. The 1280 procedures for setting the RD for a given EVI are described in 1281 section 7.9. 1283 The Ethernet Segment Identifier is set to the ten octet ESI described 1284 in section "Ethernet Segment". 1286 The Ethernet Tag ID may be zero or may represent a valid Ethernet Tag 1287 ID. This field may be non-zero when there are multiple bridge 1288 domains in the MAC-VRF (i.e., the PE needs to perform qualified 1289 learning for the VLANs in that MAC-VRF). 1291 When the the Ethernet Tag ID in the NLRI is set to a non-zero value, 1292 for a particular bridge domain, then this Ethernet Tag ID may either 1293 be the CE's Ethernet tag value (e.g., CE VLAN ID) or the EVPN 1294 provider's Ethernet tag value (e.g., provider VLAN ID). The latter 1295 would be the case if the CE Ethernet tags (e.g., CE VLAN ID) for a 1296 particular bridge domain are different on different CEs. 1298 The MAC address length field is in bits and it is set to 48. The MAC 1299 address length values other than 48 bits, are outside the scope of 1300 this document. The encoding of a MAC address MUST be the 6-octet MAC 1301 address specified by [802.1D-ORIG] [802.1D-REV]. 1303 The IP Address Field is optional. By default, the IP Address Length 1304 field is set to 0 and the IP address field is omitted from the route. 1305 When a valid IP address needs to be advertised, it is then encoded in 1306 this route. When an IP address is present, the IP Address Length 1307 field is in bits and it is set to 32 or 128 bits. Other IP Address 1308 Length values are outside the scope of this document. The encoding of 1309 an IP address MUST be either 4 octets for IPv4 or 16 octets for IPv6. 1310 The length field of EVPN NLRI (which is in octets and is described in 1311 section 7) is sufficient to determine whether an IP address is 1312 encoded in this route and if so, whether the encoded IP address is 1313 IPV4 or IPv6. 1315 The MPLS label1 field is encoded as 3 octets, where the high-order 20 1316 bits contain the label value. The MPLS label1 MUST be downstream 1317 assigned and it is associated with the MAC address being advertised 1318 by the advertising PE. The advertising PE uses this label when it 1319 receives an MPLS-encapsulated packet to perform forwarding based on 1320 the destination MAC address toward the CE. The forwarding procedures 1321 are specified in sections 13 and 14. 1323 A PE may advertise the same single EVPN label for all MAC addresses 1324 in a given EVI. This label assignment is referred to as a per EVI 1325 label assignment. Alternatively, a PE may advertise a unique EVPN 1326 label per combination. This label assignment is 1327 referred to as a per label assignment. As a third 1328 option, a PE may advertise a unique EVPN label per MAC address. This 1329 label assignment is referred to as a per MAC label assignment. All of 1330 these label assignment methods have their tradeoffs. The choice of a 1331 particular label assignment methodology is purely local to the PE 1332 that originates the route. 1334 Per EVI label assignment requires the least number of EVPN labels, 1335 but requires a MAC lookup in addition to an MPLS lookup on an egress 1336 PE for forwarding. On the other hand, a unique label per or a unique label per MAC allows an egress PE to 1338 forward a packet that it receives from another PE, to the connected 1339 CE, after looking up only the MPLS labels without having to perform a 1340 MAC lookup. This includes the capability to perform appropriate VLAN 1341 ID translation on egress to the CE. 1343 The MPLS label2 field is an optional field and if it is present, then 1344 it is encoded as 3 octets, where the high-order 20 bits contain the 1345 label value. 1347 The Next Hop field of the MP_REACH_NLRI attribute of the route MUST 1348 be set to the IPv4 or IPv6 address of the advertising PE. 1350 The BGP advertisement for the MAC advertisement route MUST also carry 1351 one or more Route Target (RT) attributes. RTs may be configured (as 1352 in IP VPNs), or may be derived automatically from the Ethernet Tag 1353 ID, in the Unique VLAN case, as described in section 7.10.1. 1355 It is to be noted that this document does not require PEs to create 1356 forwarding state for remote MACs when they are learnt in the control 1357 plane. When this forwarding state is actually created is a local 1358 implementation matter. 1360 9.2.2 Route Resolution 1362 If the Ethernet Segment Identifier field in a received MAC 1363 Advertisement route is set to the reserved ESI value of 0 or MAX-ESI, 1364 then if the receiving PE decides to install forwarding state for the 1365 associated MAC address, it MUST be based on the MAC Advertisement 1366 route alone. 1368 If the Ethernet Segment Identifier field in a received MAC 1369 Advertisement route is set to a non-reserved ESI, and the receiving 1370 PE is locally attached to the same ESI, then the PE does not alter 1371 its forwarding state based on the received route. This ensures that 1372 local routes are preferred to remote routes. 1374 If the Ethernet Segment Identifier field in a received MAC 1375 Advertisement route is set to a non-reserved ESI, then if the 1376 receiving PE decides to install forwarding state for the associated 1377 MAC address, it MUST be when both the MAC Advertisement route AND the 1378 associated set of Ethernet A-D per ES routes have been received. The 1379 dependency of MAC routes installation on Ethernet A-D per ES routes, 1380 is to ensure that MAC routes don't get accidentally installed during 1381 mass withdraw period. 1383 To illustrate this with an example, consider two PEs (PE1 and PE2) 1384 connected to a multi-homed Ethernet Segment ES1. All-Active 1385 redundancy mode is assumed. A given MAC address M1 is learnt by PE1 1386 but not PE2. On PE3, the following states may arise: 1388 T1- When the MAC Advertisement Route from PE1 and the set of Ethernet 1389 A-D per ES routes and Ethernet A-D per EVI routes from PE1 and PE2 1390 are received, PE3 can forward traffic destined to M1 to both PE1 and 1391 PE2. 1393 T2- If after T1, PE1 withdraws its set of Ethernet A-D per ES routes, 1394 then PE3 forwards traffic destined to M1 to PE2 only. 1396 T2'- If after T1, PE2 withdraws its set of Ethernet A-D per ES 1397 routes, then PE3 forwards traffic destined to M1 to PE1 only. 1399 T2''- If after T1, PE1 withdraws its MAC Advertisement route, then 1400 PE3 treats traffic to M1 as unknown unicast. 1402 T3- PE2 also advertises a MAC route for M1 and then PE1 withdraws its 1403 MAC route for M1. PE3 continues forwarding traffic destined to M1 1404 to both PE1 and PE2. In other words, despite M1 withdrawal by PE1, 1405 PE3 forwards the traffic destined to M1 to both PE1 and PE2. This is 1406 because a flow from the CE, resulting in M1 traffic getting hashed to 1407 PE1, can get terminated resulting in M1 to aged out in PE1; however, 1408 M1 can be reachable by both PE1 and PE2. 1410 10. ARP and ND 1412 The IP address field in the MAC advertisement route may optionally 1413 carry one of the IP addresses associated with the MAC address. This 1414 provides an option which can be used to minimize the flooding of ARP 1415 or Neighbor Discovery (ND) messages over the MPLS network and to 1416 remote CEs. This option also minimizes ARP (or ND) message processing 1417 on end-stations/hosts connected to the EVPN network. A PE may learn 1418 the IP address associated with a MAC address in the control or 1419 management plane between the CE and the PE. Or, it may learn this 1420 binding by snooping certain messages to or from a CE. When a PE 1421 learns the IP address associated with a MAC address, of a locally 1422 connected CE, it may advertise this address to other PEs by including 1423 it in the MAC Advertisement route. The IP Address may be an IPv4 1424 address encoded using four octets, or an IPv6 address encoded using 1425 sixteen octets. For ARP and ND purposes, the IP Address length field 1426 MUST be set to 32 for an IPv4 address or to 128 for an IPv6 address. 1428 If there are multiple IP addresses associated with a MAC address, 1429 then multiple MAC advertisement routes MUST be generated, one for 1430 each IP address. For instance, this may be the case when there are 1431 both an IPv4 and an IPv6 address associated with the same MAC address 1432 for dual-IP stack scenarios. When the IP address is dissociated with 1433 the MAC address, then the MAC advertisement route with that 1434 particular IP address MUST be withdrawn. 1436 Note that a MAC-only route can be advertised along with but 1437 independent from MAC/IP route for scenarios where the MAC learning 1438 over access network/node is done in data-plane and independent from 1439 ARP snooping that generates MAC/IP route. In such scenarios when the 1440 ARP entry times out and causes the MAC/IP to be withdrawn, then the 1441 MAC information will not be lost. In scenarios where host MAC/IP is 1442 learned via management or control plane, then the sender PE may only 1443 generates and advertises MAC/IP route. If the receiving PE receives 1444 both the MAC-only route and the MAC/IP route, then when it receives a 1445 withdraw message for the MAC/IP route, it MUST delete the 1446 corresponding entry from the ARP table but not the MAC entry from the 1447 MAC-VRF table unless it receives a withdraw message for MAC-only 1448 route. 1450 When a PE receives an ARP request for an IP address from a CE, and if 1451 the PE has the MAC address binding for that IP address, the PE SHOULD 1452 perform ARP proxy by responding to the ARP request. 1454 10.1 Default Gateway 1456 When a PE needs to perform inter-subnet forwarding where each subnet 1457 is represented by a different broadcast domain (e.g., different VLAN) 1458 the inter-subnet forwarding is performed at layer 3 and the PE that 1459 performs such function is called the default gateway for the EVPN 1460 instance. In this case when the PE receives an ARP Request for the IP 1461 address configured as the default gateway address, the PE originates 1462 an ARP Reply. 1464 Each PE that acts as a default gateway for a given EVPN instance MAY 1465 advertise in the EVPN control plane its default gateway MAC address 1466 using the MAC/IP advertisement route, and indicates that such route 1467 is associated with the default gateway. This is accomplished by 1468 requiring the route to carry the Default Gateway extended community 1469 defined in [Section 7.8 Default Gateway Extended Community]. The ESI 1470 field is set to zero when advertising the MAC route with the Default 1471 Gateway extended community. 1473 The IP address field of the MAC/IP advertisement route is set to the 1474 default GW IP address for that subnet (e.g., EVPN instance). For a 1475 given subnet (e.g., VLAN or EVPN instance), the default GW IP address 1476 is the same across all the participant PEs. The inclusion of this IP 1477 address enables the receiving PE to check its configured default GW 1478 IP address against the one received in the MAC/IP advertisement route 1479 for that subnet (or EVPN instance) and if there is a discrepancy, 1480 then the PE SHOULD notify the operator and log an error message. 1482 Unless it is known a priori (by means outside of this document) that 1483 all PEs of a given EVPN instance act as a default gateway for that 1484 EVPN instance, the MPLS label MUST be set to a valid downstream 1485 assigned label. 1487 Furthermore, even if all PEs of a given EVPN instance do act as a 1488 default gateway for that EVPN instance, but only some, but not all, 1489 of these PEs have sufficient (routing) information to provide inter- 1490 subnet routing for all the inter-subnet traffic originated within the 1491 subnet associated with the EVPN instance, then when such PE 1492 advertises in the EVPN control plane its default gateway MAC address 1493 using the MAC advertisement route, and indicates that such route is 1494 associated with the default gateway, the route MUST carry a valid 1495 downstream assigned label. 1497 If all PEs of a given EVPN instance act as a default gateway for that 1498 EVPN instance, and the same default gateway MAC address is used 1499 across all gateway devices, then no such advertisement is needed. 1500 However, if each default gateway uses a different MAC address, then 1501 each default gateway needs to be aware of other gateways' MAC 1502 addresses and thus the need for such advertisement. This is called 1503 MAC address aliasing since a single default GW can be represented by 1504 multiple MAC addresses. 1506 Each PE that receives this route and imports it as per procedures 1507 specified in this document follows the procedures in this section 1508 when replying to ARP Requests that it receives. 1510 Each PE that acts as a default gateway for a given EVPN instance that 1511 receives this route and imports it as per procedures specified in 1512 this document MUST create MAC forwarding state that enables it to 1513 apply IP forwarding to the packets destined to the MAC address 1514 carried in the route. 1516 11. Handling of Multi-Destination Traffic 1518 Procedures are required for a given PE to send broadcast or multicast 1519 traffic, received from a CE encapsulated in a given Ethernet Tag 1520 (VLAN) in an EVPN instance, to all the other PEs that span that 1521 Ethernet Tag (VLAN) in that EVPN instance. In certain scenarios, 1522 described in section "Processing of Unknown Unicast Packets", a given 1523 PE may also need to flood unknown unicast traffic to other PEs. 1525 The PEs in a particular EVPN instance may use ingress replication, 1526 P2MP LSPs or MP2MP LSPs to send unknown unicast, broadcast or 1527 multicast traffic to other PEs. 1529 Each PE MUST advertise an "Inclusive Multicast Ethernet Tag Route" to 1530 enable the above. The following subsection provides the procedures to 1531 construct the Inclusive Multicast Ethernet Tag route. Subsequent 1532 subsections describe in further detail its usage. 1534 11.1. Construction of the Inclusive Multicast Ethernet Tag Route 1536 The RD MUST be the RD of the EVI that is advertising the NLRI. The 1537 procedures for setting the RD for a given EVPN instance on a PE are 1538 described in section 7.9. 1540 The Ethernet Tag ID is the identifier of the Ethernet Tag. It may be 1541 set to 0 or to a valid Ethernet Tag value. 1543 The Originating Router's IP address MUST be set to an IP address of 1544 the PE. This address SHOULD be common for all the EVIs on the PE 1545 (e.,g., this address may be PE's loopback address). The IP Address 1546 Length field is in bits. 1548 The Next Hop field of the MP_REACH_NLRI attribute of the route MUST 1549 be set to the same IP address as the one carried in the Originating 1550 Router's IP Address field. 1552 The BGP advertisement for the Inclusive Multicast Ethernet Tag route 1553 MUST also carry one or more Route Target (RT) attributes. The 1554 assignment of RTs described in the section 7.10 MUST be followed. 1556 11.2. P-Tunnel Identification 1558 In order to identify the P-Tunnel used for sending broadcast, unknown 1559 unicast or multicast traffic, the Inclusive Multicast Ethernet Tag 1560 route MUST carry a "PMSI Tunnel Attribute" as specified in [BGP 1561 MVPN]. 1563 Depending on the technology used for the P-tunnel for the EVPN 1564 instance on the PE, the PMSI Tunnel attribute of the Inclusive 1565 Multicast Ethernet Tag route is constructed as follows. 1567 + If the PE that originates the advertisement uses a 1568 P-Multicast tree for the P-tunnel for EVPN, the PMSI 1569 Tunnel attribute MUST contain the identity of the tree 1570 (note that the PE could create the identity of the 1571 tree prior to the actual instantiation of the tree). 1573 + A PE that uses a P-Multicast tree for the P-tunnel MAY 1574 aggregate two or more EVPN instances (EVIs) present 1575 on the PE onto the same tree. In this case, in addition 1576 to carrying the identity of the tree, the PMSI Tunnel 1577 attribute MUST carry an MPLS upstream assigned label which 1578 the PE has bound uniquely to the EVI associated with this 1579 update (as determined by its RTs). 1581 If the PE has already advertised Inclusive Multicast 1582 Ethernet Tag routes for two or more EVIs that it now 1583 desires to aggregate, then the PE MUST re-advertise 1584 those routes. The re-advertised routes MUST be the same 1585 as the original ones, except for the PMSI Tunnel attribute 1586 and the label carried in that attribute. 1588 + If the PE that originates the advertisement uses ingress 1589 replication for the P-tunnel for EVPN, the route MUST 1590 include the PMSI Tunnel attribute with the Tunnel Type set to 1591 Ingress Replication and Tunnel Identifier set to a routable 1592 address of the PE. The PMSI Tunnel attribute MUST carry a 1593 downstream assigned MPLS label. This label is used to 1594 demultiplex the broadcast, multicast or unknown unicast EVPN 1595 traffic received over a MP2P tunnel by the PE. 1597 + The Leaf Information Required flag of the PMSI Tunnel 1598 attribute MUST be set to zero, and MUST be ignored on receipt. 1600 12. Processing of Unknown Unicast Packets 1602 The procedures in this document do not require the PEs to flood 1603 unknown unicast traffic to other PEs. If PEs learn CE MAC addresses 1604 via a control plane protocol, the PEs can then distribute MAC 1605 addresses via BGP, and all unicast MAC addresses will be learnt prior 1606 to traffic to those destinations. 1608 However, if a destination MAC address of a received packet is not 1609 known by the PE, the PE may have to flood the packet. When flooding, 1610 one must take into account "split horizon forwarding" as follows: The 1611 principles behind the following procedures are borrowed from the 1612 split horizon forwarding rules in VPLS solutions [RFC4761] and 1613 [RFC4762]. When a PE capable of flooding (say PEx) receives an 1614 unknown destination MAC address, it floods the frame. If the frame 1615 arrived from an attached CE, PEx must send a copy of that frame on 1616 every Ethernet Segment (belonging to that EVI) for which it is the 1617 DF, other than the Ethernet Segment on which it received the frame. 1618 In addition, the PE must flood the frame to all other PEs 1619 participating in that EVPN instance. If, on the other hand, the frame 1620 arrived from another PE (say PEy), PEx must send a copy of the packet 1621 on each Ethernet Segment (belonging to that EVI) for which it is the 1622 DF. PEx MUST NOT send the frame to other PEs, since PEy would have 1623 already done so. Split horizon forwarding rules apply to unknown MAC 1624 addresses. 1626 Whether or not to flood packets to unknown destination MAC addresses 1627 should be an administrative choice, depending on how learning happens 1628 between CEs and PEs. 1630 The PEs in a particular EVPN instance may use ingress replication 1631 using RSVP-TE P2P LSPs or LDP MP2P LSPs for sending unknown unicast 1632 traffic to other PEs. Or they may use RSVP-TE P2MP or LDP P2MP for 1633 sending such traffic to other PEs. 1635 12.1. Ingress Replication 1637 If ingress replication is in use, the P-Tunnel attribute, carried in 1638 the Inclusive Multicast Ethernet Tag routes for the EVPN instance, 1639 specifies the downstream label that the other PEs can use to send 1640 unknown unicast, multicast or broadcast traffic for that EVPN 1641 instance to this particular PE. 1643 The PE that receives a packet with this particular MPLS label MUST 1644 treat the packet as a broadcast, multicast or unknown unicast packet. 1645 Further if the MAC address is a unicast MAC address, the PE MUST 1646 treat the packet as an unknown unicast packet. 1648 12.2. P2MP MPLS LSPs 1650 The procedures for using P2MP LSPs are very similar to VPLS 1651 procedures [VPLS-MCAST]. The P-Tunnel attribute used by a PE for 1652 sending unknown unicast, broadcast or multicast traffic for a 1653 particular EVPN instance is advertised in the Inclusive Ethernet Tag 1654 Multicast route as described in section "Handling of Multi- 1655 Destination Traffic". 1657 The P-Tunnel attribute specifies the P2MP LSP identifier. This is the 1658 equivalent of an Inclusive tree in [VPLS-MCAST]. Note that multiple 1659 Ethernet Tags, which may be in different EVPN instances, may use the 1660 same P2MP LSP, using upstream labels [VPLS-MCAST]. This is the 1661 equivalent of an Aggregate Inclusive tree in [VPLS-MCAST]. When P2MP 1662 LSPs are used for flooding unknown unicast traffic, packet re- 1663 ordering is possible. 1665 The PE that receives a packet on the P2MP LSP specified in the PMSI 1666 Tunnel Attribute MUST treat the packet as a broadcast, multicast or 1667 unknown unicast packet. Further if the MAC address is a unicast MAC 1668 address, the PE MUST treat the packet as an unknown unicast packet. 1670 13. Forwarding Unicast Packets 1672 This section describes procedures for forwarding unicast packets by 1673 PEs, where such packets are received from either directly connected 1674 CEs, or from some other PEs. 1676 13.1. Forwarding packets received from a CE 1678 When a PE receives a packet from a CE, on a given Ethernet Tag ID, it 1679 must first look up the source MAC address of the packet. In certain 1680 environments that enable MAC security, the source MAC address MAY be 1681 used to validate the host identity and determine that traffic from 1682 the host can be allowed into the network. Source MAC lookup MAY also 1683 be used for local MAC address learning. 1685 If the PE decides to forward the packet, the destination MAC address 1686 of the packet must be looked up. If the PE has received MAC address 1687 advertisements for this destination MAC address from one or more 1688 other PEs or learned it from locally connected CEs, it is considered 1689 as a known MAC address. Otherwise, the MAC address is considered as 1690 an unknown MAC address. 1692 For known MAC addresses the PE forwards this packet to one of the 1693 remote PEs or to a locally attached CE. When forwarding to a remote 1694 PE, the packet is encapsulated in the EVPN MPLS label advertised by 1695 the remote PE, for that MAC address, and in the MPLS LSP label stack 1696 to reach the remote PE. 1698 If the MAC address is unknown and if the administrative policy on the 1699 PE requires flooding of unknown unicast traffic then: 1701 - The PE MUST flood the packet to other PEs. The PE MUST first 1702 encapsulate the packet in the ESI MPLS label as described in section 1703 8.3. If ingress replication is used, the packet MUST be replicated to 1704 each remote PE with the VPN label being an MPLS label determined as 1705 follows: This is the MPLS label advertised by the remote PE in a PMSI 1706 Tunnel Attribute in the Inclusive Multicast Ethernet Tag route for an 1707 combination. The Ethernet Tag in the 1708 route may be the same as the Ethernet Tag associated with the 1709 interface on which the ingress PE receives the packet. If P2MP LSPs 1710 are being used the packet MUST be sent on the P2MP LSP that the PE is 1711 the root of for the Ethernet Tag in the EVPN instance. If the same 1712 P2MP LSP is used for all Ethernet Tags, then all the PEs in the EVPN 1713 instance MUST be the leaves of the P2MP LSP. If a distinct P2MP LSP 1714 is used for a given Ethernet Tag in the EVPN instance, then only the 1715 PEs in the Ethernet Tag MUST be the leaves of the P2MP LSP. The 1716 packet MUST be encapsulated in the P2MP LSP label stack. 1718 If the MAC address is unknown then, if the administrative policy on 1719 the PE does not allow flooding of unknown unicast traffic: 1721 - The PE MUST drop the packet. 1723 13.2. Forwarding packets received from a remote PE 1725 This section described the procedures for forwarding known and 1726 unknown unicast packets received from a remote PE. 1728 13.2.1. Unknown Unicast Forwarding 1730 When a PE receives an MPLS packet from a remote PE then, after 1731 processing the MPLS label stack, if the top MPLS label ends up being 1732 a P2MP LSP label associated with an EVPN instance or in case of 1733 ingress replication the downstream label advertised in the P-Tunnel 1734 attribute, and after performing the split horizon procedures 1735 described in section 8.3: 1737 - If the PE is the designated forwarder of BUM traffic on a 1738 particular set of ESIs for the Ethernet Tag, the default behavior is 1739 for the PE to flood the packet on these ESIs. In other words, the 1740 default behavior is for the PE to assume that for BUM traffic, it is 1741 not required to perform a destination MAC address lookup. As an 1742 option, the PE may perform a destination MAC lookup to flood the 1743 packet to only a subset of the CE interfaces in the Ethernet Tag. For 1744 instance the PE may decide to not flood an BUM packet on certain 1745 Ethernet segments even if it is the DF on the Ethernet segment, based 1746 on administrative policy. 1748 - If the PE is not the designated forwarder on any of the ESIs for 1749 the Ethernet Tag, the default behavior is for it to drop the packet. 1751 13.2.2. Known Unicast Forwarding 1753 If the top MPLS label ends up being an EVPN label that was advertised 1754 in the unicast MAC advertisements, then the PE either forwards the 1755 packet based on CE next-hop forwarding information associated with 1756 the label or does a destination MAC address lookup to forward the 1757 packet to a CE. 1759 14. Load Balancing of Unicast Frames 1761 This section specifies the load balancing procedures for sending 1762 known unicast frames to a multi-homed CE. 1764 14.1. Load balancing of traffic from a PE to remote CEs 1766 Whenever a remote PE imports a MAC advertisement for a given in an EVI, it MUST examine all imported Ethernet A-D 1768 routes for that ESI in order to determine the load-balancing 1769 characteristics of the Ethernet segment. 1771 14.1.1 Single-Active Redundancy Mode 1773 For a given ES, if the remote PE has imported the set of Ethernet A-D 1774 per ES routes from at least one PE, where the "Single-Active" flag in 1775 the ESI Label Extended Community is set, then the remote PE MUST 1776 deduce that the ES is operating in Single-Active redundancy mode. As 1777 such, the MAC address will be reachable only via the PE announcing 1778 the associated MAC Advertisement route - this is referred to as the 1779 primary PE. The other PEs advertising the set of Ethernet A-D per ES 1780 routes for the same ES provide backup paths for that ES, in case the 1781 primary PE encounters a failure, and are referred to as backup PEs. 1782 It should be noted that the primary PE for a given is the 1783 DF for that . 1785 If the primary PE encounters a failure, it MAY withdraw its set of 1786 Ethernet A-D per ES routes for the affected ES prior to withdrawing 1787 it set of MAC Advertisement routes. 1789 If there is only one backup PE for a given ES, the remote PE MAY use 1790 the primary PE's withdrawal of its set of Ethernet A-D per ES routes 1791 as a trigger to update its forwarding entries, for the associated MAC 1792 addresses, to point towards the backup PE. As the backup PE starts 1793 learning the MAC addresses over its attached ES, it will start 1794 sending MAC Advertisement routes while the failed PE withdraws its 1795 routes. This mechanism minimizes the flooding of traffic during fail- 1796 over events. 1798 If there is more than one backup PE for a given ES, the remote PE 1799 MUST use the primary PE's withdrawal of its set of Ethernet A-D per 1800 ES routes as a trigger to start flooding traffic for the associated 1801 MAC addresses (as long as flooding of unknown unicast is 1802 administratively allowed), as it is not possible to select a single 1803 backup PE. 1805 14.1.2 All-Active Redundancy Mode 1807 For a given ES, if the remote PE has imported the set of Ethernet A-D 1808 per ES routes from one or more PEs and none of them have the "Single- 1809 Active" flag in the ESI Label Extended Community set, then the remote 1810 PE MUST deduce that the ES is operating in All-Active redundancy 1811 mode. A remote PE that receives a MAC advertisement route with non- 1812 reserved ESI SHOULD consider the advertised MAC address to be 1813 reachable via all PEs that have advertised reachability to that MAC 1814 address' EVI/ES via the combination of an Ethernet A-D per EVI route 1815 for that EVI/ES (and Ethernet Tag if applicable) AND an Ethernet A-D 1816 per ES route for that ES. The remote PE MUST use received MAC 1817 Advertisement routes and Ethernet A-D per EVI/per ES routes to 1818 construct the set of next-hops for the advertised MAC address. 1820 Each next-hop comprises an MPLS label stack that is to be used by the 1821 egress PE to forward the packet. This label stack is determined as 1822 follows: 1824 -If the next-hop is constructed as a result of a MAC route then this 1825 label stack MUST be used. However, if the MAC route doesn't exist for 1826 that PE, then the next-hop and MPLS label stack is constructed as a 1827 result of the Ethernet A-D routes. Note that the following 1828 description applies to determining the label stack for a particular 1829 next-hop to reach a given PE, from which the remote PE has received 1830 and imported Ethernet A-D routes that have the matching ESI and 1831 Ethernet Tag as the one present in the MAC advertisement. The 1832 Ethernet A-D routes mentioned in the following description refer to 1833 the ones imported from this given PE. 1835 -If a set of Ethernet A-D per ES routes for that ES AND an Ethernet 1836 A-D route per EVI exist, only then the label from that latter route 1837 must be used. 1839 The following example explains the above. 1841 Consider a CE (CE1) that is dual-homed to two PEs (PE1 and PE2) on a 1842 LAG interface (ES1), and is sending packets with source MAC address 1843 MAC1 on VLAN1 (mapped to EVI1). A remote PE, say PE3, is able to 1844 learn that MAC1 is reachable via PE1 and PE2. Both PE1 and PE2 may 1845 advertise MAC1 in BGP if they receive packets with MAC1 from CE1. If 1846 this is not the case, and if MAC1 is advertised only by PE1, PE3 1847 still considers MAC1 as reachable via both PE1 and PE2 as both PE1 1848 and PE2 advertise a set of Ethernet A-D per ES routes for ES1 as well 1849 as an Ethernet A-D per EVI route for . 1851 The MPLS label stack to send the packets to PE1 is the MPLS LSP stack 1852 to get to PE1 (at top of the stack) followed by the EVPN label 1853 advertised by PE1 for CE1's MAC . 1855 The MPLS label stack to send packets to PE2 is the MPLS LSP stack to 1856 get to PE2 (at top of the stack) followed by the MPLS label in the 1857 Ethernet A-D route advertised by PE2 for , if PE2 has not 1858 advertised MAC1 in BGP. 1860 We will refer to these label stacks as MPLS next-hops. 1862 The remote PE (PE3) can now load balance the traffic it receives from 1863 its CEs, destined for CE1, between PE1 and PE2. PE3 may use N-Tuple 1864 flow information to hash traffic into one of the MPLS next-hops for 1865 load balancing of IP traffic. Alternatively PE3 may rely on the 1866 source MAC addresses for load balancing. 1868 Note that once PE3 decides to send a particular packet to PE1 or PE2 1869 it can pick one out of multiple possible paths to reach the 1870 particular remote PE using regular MPLS procedures. For instance, if 1871 the tunneling technology is based on RSVP-TE LSPs, and PE3 decides to 1872 send a particular packet to PE1, then PE3 can choose from multiple 1873 RSVP-TE LSPs that have PE1 as their destination. 1875 When PE1 or PE2 receive the packet destined for CE1 from PE3, if the 1876 packet is a known unicast, it is forwarded to CE1. If it is a BUM 1877 packet then only one of PE1 or PE2 must forward the packet to the CE. 1878 Which of PE1 or PE2 forward this packet to the CE is determined based 1879 on which of the two is the DF. 1881 14.2. Load balancing of traffic between a PE and a local CE 1883 A CE may be configured with more than one interface connected to 1884 different PEs or the same PE for load balancing, using a technology 1885 such as LAG. The PE(s) and the CE can load balance traffic onto these 1886 interfaces using one of the following mechanisms. 1888 14.2.1. Data plane learning 1890 Consider that the PEs perform data plane learning for local MAC 1891 addresses learned from local CEs. This enables the PE(s) to learn a 1892 particular MAC address and associate it with one or more interfaces, 1893 if the technology between the PE and the CE supports multi-pathing. 1894 The PEs can now load balance traffic destined to that MAC address on 1895 the multiple interfaces. 1897 Whether the CE can load balance traffic that it generates on the 1898 multiple interfaces is dependent on the CE implementation. 1900 14.2.2. Control plane learning 1902 The CE can be a host that advertises the same MAC address using a 1903 control protocol on all interfaces. This enables the PE(s) to learn 1904 the host's MAC address and associate it with all interfaces. The PEs 1905 can now load balance traffic destined to the host on all these 1906 interfaces. The host can also load balance the traffic it generates 1907 onto these interfaces and the PE that receives the traffic employs 1908 EVPN forwarding procedures to forward the traffic. 1910 15. MAC Mobility 1912 It is possible for a given host or end-station (as defined by its MAC 1913 address) to move from one Ethernet segment to another; this is 1914 referred to as 'MAC Mobility' or 'MAC move' and it is different from 1915 the multi-homing situation in which a given MAC address is reachable 1916 via multiple PEs for the same Ethernet segment. In a MAC move, there 1917 would be two sets of MAC Advertisement routes, one set with the new 1918 Ethernet segment and one set with the previous Ethernet segment, and 1919 the MAC address would appear to be reachable via each of these 1920 segments. 1922 In order to allow all of the PEs in the EVPN instance to correctly 1923 determine the current location of the MAC address, all advertisements 1924 of it being reachable via the previous Ethernet segment MUST be 1925 withdrawn by the PEs, for the previous Ethernet segment, that had 1926 advertised it. 1928 If local learning is performed using the data plane, these PEs will 1929 not be able to detect that the MAC address has moved to another 1930 Ethernet segment and the receipt of MAC Advertisement routes, with 1931 the MAC Mobility extended community attribute, from other PEs serves 1932 as the trigger for these PEs to withdraw their advertisements. If 1933 local learning is performed using the control or management planes, 1934 these interactions serve as the trigger for these PEs to withdraw 1935 their advertisements. 1937 In a situation where there are multiple moves of a given MAC, 1938 possibly between the same two Ethernet segments, there may be 1939 multiple withdrawals and re-advertisements. In order to ensure that 1940 all PEs in the EVPN instance receive all of these correctly through 1941 the intervening BGP infrastructure, it is necessary to introduce a 1942 sequence number into the MAC Mobility extended community attribute. 1944 An implementation MUST handle the scenarios where the sequence number 1945 wraps around to process mobility event correctly. 1947 Every MAC mobility event for a given MAC address will contain a 1948 sequence number that is set using the following rules: 1950 - A PE advertising a MAC address for the first time advertises it 1951 with no MAC Mobility extended community attribute. 1953 - A PE detecting a locally attached MAC address for which it had 1954 previously received a MAC Advertisement route with a different 1955 Ethernet segment identifier advertises the MAC address in a MAC 1956 Advertisement route tagged with a MAC Mobility extended community 1957 attribute with a sequence number one greater than the sequence number 1958 in the MAC mobility attribute of the received MAC Advertisement 1959 route. In the case of the first mobility event for a given MAC 1960 address, where the received MAC Advertisement route does not carry a 1961 MAC Mobility attribute, the value of the sequence number in the 1962 received route is assumed to be 0 for purpose of this processing. 1964 - A PE detecting a locally attached MAC address for which it had 1965 previously received a MAC Advertisement route with the same non-zero 1966 Ethernet segment identifier advertises it with: 1967 i. no MAC Mobility extended community attribute, if the received 1968 route did not carry said attribute. 1970 ii. a MAC Mobility extended community attribute with the sequence 1971 number equal to the highest of the sequence number(s) in the 1972 received MAC Advertisement route(s), if the received route(s) is 1973 (are) tagged with a MAC Mobility extended community attribute. 1975 - A PE detecting a locally attached MAC address for which it had 1976 previously received a MAC Advertisement route with the same zero 1977 Ethernet segment identifier (single-homed scenarios) advertises it 1978 with MAC mobility extended community attribute with the sequence 1979 number set properly. In case of single-homed scenarios, there is no 1980 need for ESI comparison. The reason ESI comparison is done for multi- 1981 homing, is to prevent false detection of MAC move among the PEs 1982 attached to the same multi-homed site. 1984 A PE receiving a MAC Advertisement route for a MAC address with a 1985 different Ethernet segment identifier and a higher sequence number 1986 than that which it had previously advertised, withdraws its MAC 1987 Advertisement route. If two (or more) PEs advertise the same MAC 1988 address with same sequence number but different Ethernet segment 1989 identifiers, a PE that receives these routes selects the route 1990 advertised by the PE with lowest IP address as the best route. If the 1991 PE is the originator of the MAC route and it receives the same MAC 1992 address with the same sequence number that it generated, it will 1993 compare its own IP address with the IP address of the remote PE and 1994 will select the lowest IP. If its own route is not the best one, it 1995 will withdraw the route. 1997 15.1. MAC Duplication Issue 1999 A situation may arise where the same MAC address is learned by 2000 different PEs in the same VLAN because of two (or more hosts) being 2001 mis-configured with the same (duplicate) MAC address. In such 2002 situation, the traffic originating from these hosts would trigger 2003 continuous MAC moves among the PEs attached to these hosts. It is 2004 important to recognize such situation and avoid incrementing the 2005 sequence number (in the MAC Mobility attribute) to infinity. In order 2006 to remedy such situation, a PE that detects a MAC mobility event by 2007 way of local learning starts an M-second timer (default value of M = 2008 180) and if it detects N MAC moves before the timer expires (default 2009 value for N = 5), it concludes that a duplicate MAC situation has 2010 occurred. The PE MUST alert the operator and stop sending and 2011 processing any BGP MAC Advertisement routes for that MAC address till 2012 a corrective action is taken by the operator. The values of M and N 2013 MUST be configurable to allow for flexibility in operator control. 2014 Note that the other PEs in the E-VPN instance will forward the 2015 traffic for the duplicate MAC address to one of the PEs advertising 2016 the duplicate MAC address. 2018 15.2. Sticky MAC addresses 2020 There are scenarios in which it is desired to configure some MAC 2021 addresses as static so that they are not subjected to MAC move. In 2022 such scenarios, these MAC addresses are advertised with MAC Mobility 2023 Extended Community where static flag is set to 1 and sequence number 2024 is set to zero. If a PE receives such advertisements and later learns 2025 the same MAC address(es) via local learning, then the PE MUST alert 2026 the operator. 2028 16. Multicast & Broadcast 2030 The PEs in a particular EVPN instance may use ingress replication or 2031 P2MP LSPs to send multicast traffic to other PEs. 2033 16.1. Ingress Replication 2035 The PEs may use ingress replication for flooding BUM traffic as 2036 described in section "Handling of Multi-Destination Traffic". A given 2037 broadcast packet must be sent to all the remote PEs. However a given 2038 multicast packet for a multicast flow may be sent to only a subset of 2039 the PEs. Specifically a given multicast flow may be sent to only 2040 those PEs that have receivers that are interested in the multicast 2041 flow. Determining which of the PEs have receivers for a given 2042 multicast flow is done using explicit tracking described below. 2044 16.2. P2MP LSPs 2046 A PE may use an "Inclusive" tree for sending an BUM packet. This 2047 terminology is borrowed from [VPLS-MCAST]. 2049 A variety of transport technologies may be used in the SP network. 2050 For inclusive P-Multicast trees, these transport technologies include 2051 point-to-multipoint LSPs created by RSVP-TE or mLDP. 2053 16.2.1. Inclusive Trees 2055 An Inclusive Tree allows the use of a single multicast distribution 2056 tree, referred to as an Inclusive P-Multicast tree, in the SP network 2057 to carry all the multicast traffic from a specified set of EVPN 2058 instances on a given PE. A particular P-Multicast tree can be set up 2059 to carry the traffic originated by sites belonging to a single EVPN 2060 instance, or to carry the traffic originated by sites belonging to 2061 several EVPN instances. The ability to carry the traffic of more than 2062 one EVPN instance on the same tree is termed 'Aggregation' and the 2063 tree is called an Aggregate Inclusive P-Multicast tree or Aggregate 2064 Inclusive tree for short. The Aggregate Inclusive tree needs to 2065 include every PE that is a member of any of the EVPN instances that 2066 are using the tree. This implies that a PE may receive BUM traffic 2067 even if it doesn't have any receivers that are interested in 2068 receiving that traffic. 2070 An Inclusive or Aggregate Inclusive tree as defined in this document 2071 is a P2MP tree. A P2MP tree is used to carry traffic only for EVPN 2072 CEs that are connected to the PE that is the root of the tree. 2074 The procedures for signaling an Inclusive tree are the same as those 2075 in [VPLS-MCAST] with the VPLS-AD route replaced with the Inclusive 2076 Multicast Ethernet Tag route. The P-Tunnel attribute [VPLS-MCAST] for 2077 an Inclusive tree is advertised with the Inclusive Multicast Ethernet 2078 Tag route as described in section "Handling of Multi-Destination 2079 Traffic". Note that for an Aggregate Inclusive tree, a PE can 2080 "aggregate" multiple EVPN instances on the same P2MP LSP using 2081 upstream labels. The procedures for aggregation are the same as those 2082 described in [VPLS-MCAST], with VPLS A-D routes replaced by EVPN 2083 Inclusive Multicast Ethernet Tag routes. 2085 17. Convergence 2087 This section describes failure recovery from different types of 2088 network failures. 2090 17.1. Transit Link and Node Failures between PEs 2092 The use of existing MPLS Fast-Reroute mechanisms can provide failure 2093 recovery in the order of 50ms, in the event of transit link and node 2094 failures in the infrastructure that connects the PEs. 2096 17.2. PE Failures 2098 Consider a host CE1 that is dual homed to PE1 and PE2. If PE1 fails, 2099 a remote PE, PE3, can discover this based on the failure of the BGP 2100 session. This failure detection can be in the sub-second range if 2101 BFD is used to detect BGP session failure. PE3 can update its 2102 forwarding state to start sending all traffic for CE1 to only PE2. 2104 17.3. PE to CE Network Failures 2106 If the connectivity between the multi-homed CE and one of the PEs 2107 that it is attached to, fails, the PE MUST withdraw the set of 2108 Ethernet A-D per ES routes that had been previously advertised for 2109 that ES. When the MAC entry on the PE ages out, the PE MUST withdraw 2110 the MAC address from BGP. Note that to aid convergence, the Ethernet 2111 A-D per EVI routes MAY be withdrawn before the MAC routes. This 2112 enables the remote PEs to remove the MPLS next-hop to this particular 2113 PE from the set of MPLS next-hops that can be used to forward traffic 2114 to the CE. 2116 When a Ethernet Tag is decommissioned on an Ethernet segment, then 2117 the PE MUST withdraw the Ethernet A-D per EVI route(s) announced for 2118 the that are impacted by the decommissioning. In 2119 addition, the PE MUST also withdraw the MAC advertisement routes that 2120 are impacted by the decommissioning. 2122 The Ethernet A-D per ES routes should be used by an implementation to 2123 optimize the withdrawal of MAC advertisement routes. When a PE 2124 receives a withdrawal of a particular Ethernet A-D route from a PE it 2125 SHOULD consider all the MAC advertisement routes, that are learned 2126 from the same ESI as in the Ethernet A-D route, from the advertising 2127 PE, as having been withdrawn. This optimizes the network convergence 2128 times in the event of PE to CE failures. 2130 18. Frame Ordering 2132 In a MAC address, if the value of the 1st nibble (bits 8 thorough 5) 2133 of the most significant octet of the destination MAC address (which 2134 follows the last MPLS label) happens to be 0x4 or 0x6, then the 2135 Ethernet frame can be misinterpreted as an IPv4 or IPv6 packet by 2136 intermediate P nodes performing ECMP based on deep packet inspection, 2137 thus resulting in load balancing packets belonging to the same flow 2138 on different ECMP paths and subjecting them to different delays. 2139 Therefore, packets belonging to the same flow can arrive at the 2140 destination out of order. This out of order delivery can happen 2141 during steady state in absence of any failures resulting in 2142 significant impact to the network operation. 2144 In order to avoid any such mis-ordering, the following rules are 2145 applied: 2147 - If a network uses deep packet inspection for its ECMP, then the 2148 "Preferred PW MPLS Control Word" per [RFC4385] SHOULD be used with 2149 the value of 0 (e.g., a 4-octet field with value of zero) when 2150 sending EVPN encapsulated packets over a MP2P LSP. 2152 - If a network uses Entropy label [RFC6790], then the control word 2153 SHOULD NOT be used when sending EVPN encapsulated packet over a MP2P 2154 LSP. 2156 - When sending EVPN encapsulated packets over a P2MP LSP or P2P LSP, 2157 then the control world SHOULD NOT be used. 2159 19. Acknowledgements 2161 Special thanks to Yakov Rekhter for reviewing this draft several 2162 times and providing valuable comments and for his very engaging 2163 discussions on several topics of this draft that helped shape this 2164 document. We would also like to thank Pedro Marques, Kaushik Ghosh, 2165 Nischal Sheth, Robert Raszuk, Amit Shukla, and Nadeem Mohammed for 2166 discussions that helped shape this document. We would also like to 2167 thank Han Nguyen for his comments and support of this work. We would 2168 also like to thank Steve Kensil and Reshad Rahman for their reviews. 2169 We would like to thank Jorge Rabadan for his contribution to section 2170 5 of this draft. We like to thank Thomas Morin for his review of this 2171 draft and his contribution of section 8.6. Many thanks to Jakob Heitz 2172 for his help to improve several sections of this draft. 2174 We would also like to thank Clarence Filsfils, Dennis Cai, Quaizar 2175 Vohra, Kireeti Kompella, Apurva Mehta for their contributions to this 2176 document. 2178 Last but not least, special thanks to Giles Heron (our WG chair) for 2179 his detailed review of this document in preparation for WG LC and 2180 making many valuable suggestions. 2182 20. Security Considerations 2184 Security considerations discussed in [RFC4761] and [RFC4762] apply to 2185 this document for MAC learning in data-plane over an Attachment 2186 Circuit (AC) and for flooding of unknown unicast and ARP messages 2187 over the MPLS/IP core. Security considerations discussed in [RFC4364] 2188 apply to this document for MAC learning in control-plane over the 2189 MPLS/IP core. This section describes additional considerations. 2191 As mentioned in [RFC4761], there are two aspects to achieving data 2192 privacy and protecting against denial-of-service attacks in a VPN: 2193 securing the control plane and protecting the forwarding path. 2194 Compromise of the control plane could result in a PE sending customer 2195 data belonging to some EVPN to another EVPN, or black-holing EVPN 2196 customer data, or even sending it to an eavesdropper; none of which 2197 are acceptable from a data privacy point of view. In addition, 2198 compromise of the control plane could result in black-holing EVPN 2199 customer data and could provide opportunities for unauthorized EVPN 2200 data usage (e.g., exploiting traffic replication within a multicast 2201 tree to amplify a denial-of-service attack based on sending large 2202 amounts of traffic). 2204 The mechanisms in this document use BGP for the control plane. Hence, 2205 techniques such as in [RFC5925] help authenticate BGP messages, 2206 making it harder to spoof updates (which can be used to divert EVPN 2207 traffic to the wrong EVPN instance) or withdrawals (denial-of-service 2208 attacks). In the multi-AS methods (b) and (c), this also means 2209 protecting the inter-AS BGP sessions, between the ASBRs, the PEs, or 2210 the Route Reflectors. 2212 Note that [RFC5925] will not help in keeping MPLS labels private -- 2213 knowing the labels, one can eavesdrop on EVPN traffic. However, this 2214 requires access to the data path within an SP network, which is 2215 assumed to be composed of trusted nodes/links. 2217 One of the requirements for protecting the data plane is that the 2218 MPLS labels be accepted only from valid interfaces. For a PE, valid 2219 interfaces comprise links from other routers in the PE's own AS. For 2220 an ASBR, valid interfaces comprise links from other routers in the 2221 ASBR's own AS, and links from other ASBRs in ASes that have instances 2222 of a given EVPN. It is especially important in the case of multi-AS 2223 EVPN instances that one accept EVPN packets only from valid 2224 interfaces. 2226 It is also important to help limit malicious traffic into a network 2227 for an imposter MAC address. The mechanism described in section 15.1, 2228 shows how duplicate MAC addresses can be detected and continuous 2229 false MAC mobility can be prevented. The mechanism described in 2230 section 15.2 shows how MAC addresses can be pinned to a given 2231 Ethernet Segment, such that if they appear behind any other Ethernet 2232 Segments, the traffic for those MAC addresses can be prevented from 2233 entering the EVPN network from the other Ethernet Segments. 2235 21. Contributors 2237 In addition to the authors listed on the front page, the following 2238 individuals have also helped to shape this document: 2240 Keyur Patel 2241 Samer Salam 2242 Sami Boutros 2243 Cisco 2245 Yakov Rekhter 2246 Ravi Shekhar 2247 Juniper Networks 2249 Florin Balus 2250 Nuage Networks 2252 22. IANA Considerations 2254 This document defines a new NLRI, called "EVPN", to be carried in BGP 2255 using multiprotocol extensions. This NLRI uses the existing AFI of 2256 25 (L2VPN). IANA has assigned it a SAFI value of 70. 2258 IANA has allocated the following EVPN Extended Community sub-types 2259 and this document is the only reference for them. 2261 0x00 MAC Mobility [this document] 2262 0x01 ESI Label [this document] 2263 0x02 ES-Import Route Target [this document] 2265 This document is creating a registry called "EVPN Route Types." New 2266 registrations will be made through the "RFC Required" procedure 2267 defined in [RFC5226]. The registry has no maximum value. Initial 2268 registrations are as follows: 2270 1 Ethernet Auto-Discovery [this document] 2271 2 MAC/IP Advertisement [this document] 2272 3 Inclusive Multicast Ethernet Tag [this document] 2273 4 Ethernet Segment [this document] 2275 23. References 2277 23.1 Normative References 2279 [RFC4364] "BGP/MPLS IP VPNs", Rosen, Rekhter, et. al., February 2006 2281 [RFC4761] Kompella, K. and Y. Rekhter, "Virtual Private LAN Service 2282 (VPLS) Using BGP for Auto-Discovery and Signaling", RFC 2283 4761, January 2007. 2285 [RFC4762] Lasserre, M. and V. Kompella, "Virtual Private LAN Service 2286 (VPLS) Using Label Distribution Protocol (LDP) Signaling", 2287 RFC 4762, January 2007. 2289 [RFC4271] Y. Rekhter et. al., "A Border Gateway Protocol 4 (BGP-4)", 2290 RFC 4271, January 2006 2292 [RFC4760] T. Bates et. al., "Multiprotocol Extensions for BGP-4", RFC 2293 4760, January 2007 2295 23.2 Informative References 2297 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2298 Requirement Levels", BCP 14, RFC 2119, March 1997. 2300 [RFC7209] A. Sajassi, R. Aggarwal et. al., "Requirements for Ethernet 2301 VPN", draft-ietf-l2vpn-evpn-req-04.txt, July 2013. 2303 [RFC7117] "Multicast in VPLS". R. Aggarwal et.al., draft-ietf-l2vpn- 2304 vpls-mcast-14.txt, July 2013. 2306 [RFC4684] P. Marques et. al., "Constrained Route Distribution for 2307 Border Gateway Protocol/MultiProtocol Label Switching 2308 (BGP/MPLS) Internet Protocol (IP) Virtual Private Networks 2309 (VPNs)", RFC 4684, November 2006. 2311 [RFC6790] K. Kompella et. al, "The Use of Entropy Labels in MPLS 2312 Forwarding", RFC 6790, November 2012. 2314 [RFC4385] S. Bryant et. al, "PWE3 Control Word for Use over an MPLS 2315 PSN", RFC 4385, February 2006 2317 [RFC5226] T. Narten et. al., "Guidelines for Writing an IANA 2318 Considerations Section in RFCs", RFC 5226, May 2008 2320 24. Author's Address 2322 Ali Sajassi 2323 Cisco 2324 Email: sajassi@cisco.com 2326 Rahul Aggarwal 2327 Email: raggarwa_1@yahoo.com 2328 Nabil Bitar 2329 Verizon Communications 2330 Email : nabil.n.bitar@verizon.com 2332 Aldrin Isaac 2333 Bloomberg 2334 Email: aisaac71@bloomberg.net 2336 James Uttaro 2337 AT&T 2338 Email: uttaro@att.com 2340 John Drake 2341 Juniper Networks 2342 Email: jdrake@juniper.net 2344 Wim Henderickx 2345 Alcatel-Lucent 2346 e-mail: wim.henderickx@alcatel-lucent.com