idnits 2.17.1 draft-ietf-l2vpn-evpn-07.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 46 instances of lines with control characters in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 1390 has weird spacing: '...ntinues forwa...' -- The document date (May 7, 2014) is 3635 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 428, but not defined == Missing Reference: 'VPLS-MCAST' is mentioned on line 2053, but not defined == Missing Reference: 'RFC5925' is mentioned on line 2196, but not defined == Unused Reference: 'RFC7209' is defined on line 2267, but no explicit reference was found in the text == Unused Reference: 'RFC7117' is defined on line 2270, but no explicit reference was found in the text == Outdated reference: A later version (-07) exists of draft-ietf-l2vpn-evpn-req-04 == Outdated reference: A later version (-16) exists of draft-ietf-l2vpn-vpls-mcast-14 Summary: 1 error (**), 0 flaws (~~), 9 warnings (==), 1 comment (--). 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: November 7, 2014 May 7, 2014 19 BGP MPLS Based Ethernet VPN 20 draft-ietf-l2vpn-evpn-07 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). 63 Table of Contents 65 1. Specification of requirements . . . . . . . . . . . . . . . . . 5 66 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . 5 67 3. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 6 68 4. BGP MPLS Based EVPN Overview . . . . . . . . . . . . . . . . . 6 69 5. Ethernet Segment . . . . . . . . . . . . . . . . . . . . . . . 7 70 6. Ethernet Tag ID . . . . . . . . . . . . . . . . . . . . . . . . 10 71 6.1 VLAN Based Service Interface . . . . . . . . . . . . . . . . 10 72 6.2 VLAN Bundle Service Interface . . . . . . . . . . . . . . . 11 73 6.2.1 Port Based Service Interface . . . . . . . . . . . . . . 11 74 6.3 VLAN Aware Bundle Service Interface . . . . . . . . . . . . 11 75 6.3.1 Port Based VLAN Aware Service Interface . . . . . . . . 12 76 7. BGP EVPN NLRI . . . . . . . . . . . . . . . . . . . . . . . . . 12 77 7.1. Ethernet Auto-Discovery Route . . . . . . . . . . . . . . . 13 78 7.2. MAC/IP Advertisement Route . . . . . . . . . . . . . . . . 13 79 7.3. Inclusive Multicast Ethernet Tag Route . . . . . . . . . . 14 80 7.4 Ethernet Segment Route . . . . . . . . . . . . . . . . . . . 15 81 7.5 ESI Label Extended Community . . . . . . . . . . . . . . . . 15 82 7.6 ES-Import Route Target . . . . . . . . . . . . . . . . . . . 16 83 7.7 MAC Mobility Extended Community . . . . . . . . . . . . . . 16 84 7.8 Default Gateway Extended Community . . . . . . . . . . . . . 17 85 7.9 Route Distinguisher Assignment per EVI . . . . . . . . . . . 17 86 7.10 Route Targets . . . . . . . . . . . . . . . . . . . . . . . 17 87 7.10.1 Auto-Derivation from the Ethernet Tag ID . . . . . . . 17 88 8. Multi-homing Functions . . . . . . . . . . . . . . . . . . . . 18 89 8.1 Multi-homed Ethernet Segment Auto-Discovery . . . . . . . . 18 90 8.1.1 Constructing the Ethernet Segment Route . . . . . . . . 18 91 8.2 Fast Convergence . . . . . . . . . . . . . . . . . . . . . . 18 92 8.2.1 Constructing the Ethernet A-D per Ethernet Segment 93 (ES) Route . . . . . . . . . . . . . . . . . . . . . . . 19 94 8.2.1.1. Ethernet A-D Route Targets . . . . . . . . . . . . 20 95 8.3 Split Horizon . . . . . . . . . . . . . . . . . . . . . . . 20 96 8.3.1 ESI Label Assignment . . . . . . . . . . . . . . . . . . 21 97 8.3.1.1 Ingress Replication . . . . . . . . . . . . . . . . 21 98 8.3.1.2. P2MP MPLS LSPs . . . . . . . . . . . . . . . . . . 22 99 8.4 Aliasing and Backup-Path . . . . . . . . . . . . . . . . . . 23 100 8.4.1 Constructing the Ethernet A-D per EVPN Instance (EVI) 101 Route . . . . . . . . . . . . . . . . . . . . . . . . . 24 102 8.5 Designated Forwarder Election . . . . . . . . . . . . . . . 25 103 8.6. Interoperability with Single-homing PEs . . . . . . . . . . 27 104 9. Determining Reachability to Unicast MAC Addresses . . . . . . . 27 105 9.1. Local Learning . . . . . . . . . . . . . . . . . . . . . . 27 106 9.2. Remote learning . . . . . . . . . . . . . . . . . . . . . . 28 107 9.2.1. Constructing the BGP EVPN MAC/IP Address 108 Advertisement . . . . . . . . . . . . . . . . . . . . . 28 109 9.2.2 Route Resolution . . . . . . . . . . . . . . . . . . . . 30 110 10. ARP and ND . . . . . . . . . . . . . . . . . . . . . . . . . . 31 111 10.1 Default Gateway . . . . . . . . . . . . . . . . . . . . . . 32 112 11. Handling of Multi-Destination Traffic . . . . . . . . . . . . 33 113 11.1. Construction of the Inclusive Multicast Ethernet Tag 114 Route . . . . . . . . . . . . . . . . . . . . . . . . . . 33 115 11.2. P-Tunnel Identification . . . . . . . . . . . . . . . . . 34 116 12. Processing of Unknown Unicast Packets . . . . . . . . . . . . 35 117 12.1. Ingress Replication . . . . . . . . . . . . . . . . . . . 35 118 12.2. P2MP MPLS LSPs . . . . . . . . . . . . . . . . . . . . . . 36 119 13. Forwarding Unicast Packets . . . . . . . . . . . . . . . . . . 36 120 13.1. Forwarding packets received from a CE . . . . . . . . . . 36 121 13.2. Forwarding packets received from a remote PE . . . . . . . 37 122 13.2.1. Unknown Unicast Forwarding . . . . . . . . . . . . . . 37 123 13.2.2. Known Unicast Forwarding . . . . . . . . . . . . . . . 38 124 14. Load Balancing of Unicast Frames . . . . . . . . . . . . . . . 38 125 14.1. Load balancing of traffic from a PE to remote CEs . . . . 38 126 14.1.1 Single-Active Redundancy Mode . . . . . . . . . . . . . 38 127 14.1.2 All-Active Redundancy Mode . . . . . . . . . . . . . . 39 128 14.2. Load balancing of traffic between a PE and a local CE . . 41 129 14.2.1. Data plane learning . . . . . . . . . . . . . . . . . 41 130 14.2.2. Control plane learning . . . . . . . . . . . . . . . . 41 131 15. MAC Mobility . . . . . . . . . . . . . . . . . . . . . . . . . 41 132 15.1. MAC Duplication Issue . . . . . . . . . . . . . . . . . . 43 133 15.2. Sticky MAC addresses . . . . . . . . . . . . . . . . . . . 43 134 16. Multicast & Broadcast . . . . . . . . . . . . . . . . . . . . 44 135 16.1. Ingress Replication . . . . . . . . . . . . . . . . . . . 44 136 16.2. P2MP LSPs . . . . . . . . . . . . . . . . . . . . . . . . 44 137 16.2.1. Inclusive Trees . . . . . . . . . . . . . . . . . . . 44 138 17. Convergence . . . . . . . . . . . . . . . . . . . . . . . . . 45 139 17.1. Transit Link and Node Failures between PEs . . . . . . . . 45 140 17.2. PE Failures . . . . . . . . . . . . . . . . . . . . . . . 45 141 17.3. PE to CE Network Failures . . . . . . . . . . . . . . . . 45 142 18. Frame Ordering . . . . . . . . . . . . . . . . . . . . . . . . 46 143 19. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 47 144 20. Security Considerations . . . . . . . . . . . . . . . . . . . 47 145 21. Co-authors . . . . . . . . . . . . . . . . . . . . . . . . . . 48 146 22. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 49 147 23. References . . . . . . . . . . . . . . . . . . . . . . . . . . 49 148 23.1 Normative References . . . . . . . . . . . . . . . . . . . 49 149 23.2 Informative References . . . . . . . . . . . . . . . . . . 49 150 24. Author's Address . . . . . . . . . . . . . . . . . . . . . . . 50 152 1. Specification of requirements 154 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 155 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 156 document are to be interpreted as described in [RFC2119]. 158 2. Terminology 160 Broadcast Domain: in a bridged network, it corresponds to a Virtual 161 LAN (VLAN); where a VLAN is typically represented by a single VLAN ID 162 (VID), but can be represented by several VIDs. 164 Bridge Domain: An instantiation of a broadcast domain on a bridge 165 node 167 CE: Customer Edge device e.g., host or router or switch 169 EVI: An EVPN instance spanning across the PEs participating in that 170 EVPN 172 MAC-VRF: A Virtual Routing and Forwarding table for MAC addresses on 173 a PE for an EVI 175 Ethernet Segment Identifier (ESI): If a CE is multi-homed to two or 176 more PEs, the set of Ethernet links that attaches the CE to the PEs 177 is an 'Ethernet segment'. Ethernet segments MUST have a unique non- 178 zero identifier, the 'Ethernet Segment Identifier'. 180 Ethernet Tag: An Ethernet Tag identifies a particular broadcast 181 domain, e.g., a VLAN. An EVPN instance consists of one or more 182 broadcast domains. Ethernet tag(s) are assigned to the broadcast 183 domains of a given EVPN instance by the provider of that EVPN, and 184 each PE in that EVPN instance performs a mapping between broadcast 185 domain identifier(s) understood by each of its attached CEs and the 186 corresponding Ethernet tag. 188 LACP: Link Aggregation Control Protocol 190 MP2MP: Multipoint to Multipoint 192 P2MP: Point to Multipoint 194 P2P: Point to Point 196 Single-Active Redundancy Mode: When only a single PE, among a group 197 of PEs attached to an Ethernet segment, is allowed to forward traffic 198 to/from that Ethernet Segment, then the Ethernet segment is defined 199 to be operating in Single-Active redundancy mode. 201 All-Active Redundancy Mode: When all PEs attached to an Ethernet 202 segment are allowed to forward traffic to/from that Ethernet Segment, 203 then the Ethernet segment is defined to be operating in All-Active 204 redundancy mode. 206 3. Introduction 208 This document describes procedures for BGP MPLS based Ethernet VPNs 209 (EVPN). The procedures described here are intended to meet the 210 requirements specified in [EVPN-REQ]. Please refer to [EVPN-REQ] for 211 the detailed requirements and motivation. EVPN requires extensions to 212 existing IP/MPLS protocols as described in this document. In addition 213 to these extensions EVPN uses several building blocks from existing 214 MPLS technologies. 216 4. BGP MPLS Based EVPN Overview 218 This section provides an overview of EVPN. An EVPN instance comprises 219 CEs that are connected to PEs that form the edge of the MPLS 220 infrastructure. A CE may be a host, a router or a switch. The PEs 221 provide virtual Layer 2 bridged connectivity between the CEs. There 222 may be multiple EVPN instances in the provider's network. 224 The PEs may be connected by an MPLS LSP infrastructure which provides 225 the benefits of MPLS technology such as fast-reroute, resiliency, 226 etc. The PEs may also be connected by an IP infrastructure in which 227 case IP/GRE tunneling or other IP tunneling can be used between the 228 PEs. The detailed procedures in this version of this document are 229 specified only for MPLS LSPs as the tunneling technology. However 230 these procedures are designed to be extensible to IP tunneling as the 231 Packet Switched Network (PSN) tunneling technology. 233 In an EVPN, MAC learning between PEs occurs not in the data plane (as 234 happens with traditional bridging in VPLS [RFC4761] or [RFC4762]) but 235 in the control plane. Control plane learning offers greater control 236 over the MAC learning process, such as restricting who learns what, 237 and the ability to apply policies. Furthermore, the control plane 238 chosen for advertising MAC reachability information is multi-protocol 239 (MP) BGP (similar to IP VPNs (RFC 4364)). This provides flexibility 240 and the ability to preserve the "virtualization" or isolation of 241 groups of interacting agents (hosts, servers, virtual machines) from 242 each other. In EVPN, PEs advertise the MAC addresses learned from the 243 CEs that are connected to them, along with an MPLS label, to other 244 PEs in the control plane using MP-BGP. Control plane learning enables 245 load balancing of traffic to and from CEs that are multi-homed to 246 multiple PEs. This is in addition to load balancing across the MPLS 247 core via multiple LSPs between the same pair of PEs. In other words 248 it allows CEs to connect to multiple active points of attachment. It 249 also improves convergence times in the event of certain network 250 failures. 252 However, learning between PEs and CEs is done by the method best 253 suited to the CE: data plane learning, IEEE 802.1x, LLDP, 802.1aq, 254 ARP, management plane or other protocols. 256 It is a local decision as to whether the Layer 2 forwarding table on 257 a PE is populated with all the MAC destination addresses known to the 258 control plane, or whether the PE implements a cache based scheme. For 259 instance the MAC forwarding table may be populated only with the MAC 260 destinations of the active flows transiting a specific PE. 262 The policy attributes of EVPN are very similar to those of IP-VPN. A 263 EVPN instance requires a Route-Distinguisher (RD) which is unique per 264 PE and one or more globally unique Route-Targets (RTs). A CE attaches 265 to a MAC-VRF on a PE, on an Ethernet interface which may be 266 configured for one or more Ethernet Tags, e.g., VLAN IDs. Some 267 deployment scenarios guarantee uniqueness of VLAN IDs across EVPN 268 instances: all points of attachment for a given EVPN instance use the 269 same VLAN ID, and no other EVPN instance uses this VLAN ID. This 270 document refers to this case as a "Unique VLAN EVPN" and describes 271 simplified procedures to optimize for it. 273 5. Ethernet Segment 275 If a CE is multi-homed to two or more PEs, the set of Ethernet links 276 constitutes an "Ethernet Segment". An Ethernet segment may appear to 277 the CE as a Link Aggregation Group (LAG). Ethernet segments have an 278 identifier, called the "Ethernet Segment Identifier" (ESI) which is 279 encoded as a ten octets integer. The following two ESI values are 280 reserved: 282 - ESI 0 denotes a single-homed CE. 284 - ESI {0xFF} (repeated 10 times) is known as MAX-ESI and is 285 reserved. 287 In general, an Ethernet segment MUST have a non-reserved ESI that is 288 unique network wide (i.e., across all EVPN instances on all the PEs). 289 If the CE(s) constituting an Ethernet Segment is (are) managed by the 290 network operator, then ESI uniqueness should be guaranteed; however, 291 if the CE(s) is (are) not managed, then the operator MUST configure a 292 network-wide unique ESI for that Ethernet Segment. This is required 293 to enable auto-discovery of Ethernet Segments and DF election. 295 In a network with managed and not-managed CEs, the ESI has the 296 following format: 298 +---+---+---+---+---+---+---+---+---+---+ 299 | T | ESI Value | 300 +---+---+---+---+---+---+---+---+---+---+ 302 Where: 304 T (ESI Type) is a 1-byte field (most significant octet) that 305 specifies the format of the remaining nine bytes (ESI Value). The 306 following 6 ESI types can be used: 308 - Type 0 (T=0x00) - This type indicates an arbitrary nine-octet ESI 309 value, which is managed and configured by the operator. 311 - Type 1 (T=0x01) - When IEEE 802.1AX LACP is used between the PEs 312 and CEs, this ESI type indicates an auto-generated ESI value 313 determined from LACP by concatenating the following parameters: 315 + CE LACP six octets System MAC address. The CE LACP System MAC 316 address MUST be encoded in the high order six octets of the ESI 317 Value field. 319 + CE LACP two octets Port Key. The CE LACP port key MUST be 320 encoded in the two octets next to the System MAC address. 322 + The remaining octet will be set to 0x00. 324 As far as the CE is concerned, it would treat the multiple PEs 325 that it is connected to as the same switch. This allows the CE 326 to aggregate links that are attached to different PEs in the 327 same bundle. 329 This mechanism could be used only if it produces ESIs that satisfy 330 the uniqueness requirement specified above. 332 - Type 2 (T=0x02) - This type is used in the case of indirectly 333 connected hosts via a bridged LAN between the CEs and the PEs. The 334 ESI Value is auto-generated and determined based on the Layer 2 335 bridge protocol as follows: If MST is used in the bridged LAN then 336 the value of the ESI is derived by listening to BPDUs on the Ethernet 337 segment. To achieve this the PE is not required to run MST. However 338 the PE must learn the Root Bridge MAC address and Bridge Priority of 339 the root of the Internal Spanning Tree (IST) by listening to the 340 BPDUs. The ESI Value is constructed as follows: 342 + Root Bridge six octets MAC address. The Root Bridge MAC 343 address MUST be encoded in the high order six octets of the 344 ESI Value field. 346 + Root Bridge two octets Priority. The CE Root Bridge Priority 347 MUST be encoded in the two octets next to the Root Bridge 348 MAC address. 350 + The remaining octet will be set to 0x00. 352 This mechanism could be used only if it produces ESIs that 353 satisfy the uniqueness requirement specified above. 355 - Type 3 (T=0x03) - This type indicates a MAC-based ESI Value that 356 can be auto-generated or configured by the operator. The ESI Value is 357 constructed as follows: 359 + System MAC address (six octets). The PE MAC address MUST 360 be encoded in the high order six octets of the ESI Value field. 362 + Local Discriminator value (three octets). The Local 363 Discriminator MUST be encoded in the low order three octets 364 of the ESI Value. 366 This mechanism could be used only if it produces ESIs that 367 satisfy the uniqueness requirement specified above. 369 - Type 4 (T=0x04) - This type indicates a router-ID ESI Value that 370 can be auto-generated or configured by the operator. The ESI Value is 371 constructed as follows: 373 + Router ID (four octets). The system router ID MUST be encoded in 374 the high order four octets of the ESI Value field. 376 + Local Discriminator value (four octets). The Local 377 Discriminator MUST be encoded in the four octets next to the 378 IP address. 380 + The low order octet of the ESI Value will be set to 0x00. 382 This mechanism could be used only if it produces ESIs that 383 satisfy the uniqueness requirement specified above. 385 - Type 5 (T=0x05) - This type indicates an AS-based ESI Value that 386 can be auto-generated or configured by the operator. The ESI Value is 387 constructed as follows: 389 + AS number (four octets). This is an AS number owned by the 390 system and MUST be encoded in the high order four octets of the 391 ESI Value field. If a two-octet AS number is used, the high order 392 extra two bytes will be 0x0000. 394 + Local Discriminator value (four octets). The Local Discriminator 395 MUST be encoded in the four octets next to the AS number. 397 + The low order octet of the ESI Value will be set to 0x00. 399 This mechanism could be used only if it produces ESIs that satisfy 400 the uniqueness requirement specified above. 402 6. Ethernet Tag ID 404 An Ethernet Tag ID is a 32-bit field containing either a 12-bit or a 405 24-bit identifier that identifies a particular broadcast domain 406 (e.g., a VLAN) in an EVPN Instance. The 12-bit identifier is called 407 VLAN ID (VID). An EVPN Instance consists of one or more broadcast 408 domains (one or more VLANs). VLANs are assigned to a given EVPN 409 Instance by the provider of the EVPN service. A given VLAN can itself 410 be represented by multiple VLAN IDs (VIDs). In such cases, the PEs 411 participating in that VLAN for a given EVPN instance are responsible 412 for performing VLAN ID translation to/from locally attached CE 413 devices. 415 If a VLAN is represented by a single VID across all PE devices 416 participating in that VLAN for that EVPN instance, then there is no 417 need for VID translation at the PEs. Furthermore, some deployment 418 scenarios guarantee uniqueness of VIDs across all EVPN instances; 419 all points of attachment for a given EVPN instance use the same VID 420 and no other EVPN instances use that VID. This allows the RT(s) for 421 each EVPN instance to be derived automatically from the corresponding 422 VID, as described in section 7.10.1. 424 The following subsections discuss the relationship between broadcast 425 domains (e.g., VLANs), Ethernet Tags (e.g., VIDs), and MAC-VRFs as 426 well as the setting of the Ethernet Tag Identifier, in the various 427 EVPN BGP routes (defined in section 8), for the different types of 428 service interfaces described in [EVPN-REQ]. 430 The following value of Ethernet Tag is reserved: 432 - Ethernet Tag {0xFFFFFFFF} is known as MAX-ET 434 6.1 VLAN Based Service Interface 435 With this service interface, an EVPN instance consists of only a 436 single broadcast domain (e.g., a single VLAN). Therefore, there is a 437 one to one mapping between a VID on this interface and a MAC-VRF. 438 Since a MAC-VRF corresponds to a single VLAN, it consists of a single 439 bridge domain corresponding to that VLAN. If the VLAN is represented 440 by different VIDs on different PEs, then each PE needs to perform VID 441 translation for frames destined to its attached CEs. In such 442 scenarios, the Ethernet frames transported over MPLS/IP network 443 SHOULD remain tagged with the originating VID and a VID translation 444 MUST be supported in the data path and MUST be performed on the 445 disposition PE. The Ethernet Tag Identifier in all EVPN routes MUST 446 be set to 0. 448 6.2 VLAN Bundle Service Interface 450 With this service interface, an EVPN instance corresponds to several 451 broadcast domains (e.g., several VLANs); however, only a single 452 bridge domain is maintained per MAC-VRF which means multiple VLANs 453 share the same bridge domain. This implies MAC addresses MUST be 454 unique across different VLANs for this service to work. In other 455 words, there is a many-to-one mapping between VLANs and a MAC-VRF, 456 and the MAC-VRF consists of a single bridge domain. Furthermore, a 457 single VLAN must be represented by a single VID - e.g., no VID 458 translation is allowed for this service interface type. The MPLS 459 encapsulated frames MUST remain tagged with the originating VID. Tag 460 translation is NOT permitted. The Ethernet Tag Identifier in all EVPN 461 routes MUST be set to 0. 463 6.2.1 Port Based Service Interface 465 This service interface is a special case of the VLAN Bundle service 466 interface, where all of the VLANs on the port are part of the same 467 service and map to the same bundle. The procedures are identical to 468 those described in section 6.2. 470 6.3 VLAN Aware Bundle Service Interface 472 With this service interface, an EVPN instance consists of several 473 broadcast domains (e.g., several VLANs) with each VLAN having its own 474 bridge domain - i.e., multiple bridge domains (one per VLAN) is 475 maintained by a single MAC-VRF corresponding to the EVPN instance. In 476 the case where a single VLAN is represented by different VIDs on 477 different CEs and thus tag (VID) translation is required, a 478 normalized Ethernet Tag (VID) MUST be carried in the MPLS 479 encapsulated frames and a tag translation function MUST be supported 480 in the data path. This translation MUST be performed in data path on 481 both the imposition as well as the disposition PEs (translating to 482 normalized tag on imposition PE and translating to local tag on 483 disposition PE). The Ethernet Tag Identifier in all EVPN routes MUST 484 be set to the normalized Ethernet Tag assigned by the EVPN provider. 486 6.3.1 Port Based VLAN Aware Service Interface 488 This service interface is a special case of the VLAN Aware Bundle 489 service interface, where all of the VLANs on the port are part of the 490 same service and are mapped to a single bundle but without any VID 491 translation. The procedures are subset of those described in section 492 6.3. 494 7. BGP EVPN NLRI 496 This document defines a new BGP NLRI, called the EVPN NLRI. 498 Following is the format of the EVPN NLRI: 500 +-----------------------------------+ 501 | Route Type (1 octet) | 502 +-----------------------------------+ 503 | Length (1 octet) | 504 +-----------------------------------+ 505 | Route Type specific (variable) | 506 +-----------------------------------+ 508 The Route Type field defines encoding of the rest of the EVPN NLRI 509 (Route Type specific EVPN NLRI). 511 The Length field indicates the length in octets of the Route Type 512 specific field of EVPN NLRI. 514 This document defines the following Route Types: 516 + 1 - Ethernet Auto-Discovery (A-D) route 517 + 2 - MAC/IP advertisement route 518 + 3 - Inclusive Multicast Ethernet Tag Route 519 + 4 - Ethernet Segment Route 521 The detailed encoding and procedures for these route types are 522 described in subsequent sections. 524 The EVPN NLRI is carried in BGP [RFC4271] using BGP Multiprotocol 525 Extensions [RFC4760] with an AFI of 25 (L2VPN) and a SAFI of 70 526 (EVPN). The NLRI field in the MP_REACH_NLRI/MP_UNREACH_NLRI attribute 527 contains the EVPN NLRI (encoded as specified above). 529 In order for two BGP speakers to exchange labeled EVPN NLRI, they 530 must use BGP Capabilities Advertisement to ensure that they both are 531 capable of properly processing such NLRI. This is done as specified 532 in [RFC4760], by using capability code 1 (multiprotocol BGP) with an 533 AFI of 25 (L2VPN) and a SAFI of 70 (EVPN). 535 7.1. Ethernet Auto-Discovery Route 537 A Ethernet A-D route type specific EVPN NLRI consists of the 538 following: 540 +---------------------------------------+ 541 | RD (8 octets) | 542 +---------------------------------------+ 543 |Ethernet Segment Identifier (10 octets)| 544 +---------------------------------------+ 545 | Ethernet Tag ID (4 octets) | 546 +---------------------------------------+ 547 | MPLS Label (3 octets) | 548 +---------------------------------------+ 550 For the purpose of BGP route key processing, only the Ethernet 551 Segment Identifier and the Ethernet Tag ID are considered to be part 552 of the prefix in the NLRI. The MPLS Label field is to be treated as 553 a route attribute as opposed to being part of the route. 555 For procedures and usage of this route please see section 8.2 "Fast 556 Convergence" and section 8.4 "Aliasing". 558 7.2. MAC/IP Advertisement Route 560 A MAC/IP advertisement route type specific EVPN NLRI consists of the 561 following: 563 +---------------------------------------+ 564 | RD (8 octets) | 565 +---------------------------------------+ 566 |Ethernet Segment Identifier (10 octets)| 567 +---------------------------------------+ 568 | Ethernet Tag ID (4 octets) | 569 +---------------------------------------+ 570 | MAC Address Length (1 octet) | 571 +---------------------------------------+ 572 | MAC Address (6 octets) | 573 +---------------------------------------+ 574 | IP Address Length (1 octet) | 575 +---------------------------------------+ 576 | IP Address (0 or 4 or 16 octets) | 577 +---------------------------------------+ 578 | MPLS Label1 (3 octets) | 579 +---------------------------------------+ 580 | MPLS Label2 (0 or 3 octets) | 581 +---------------------------------------+ 583 For the purpose of BGP route key processing, only the Ethernet Tag 584 ID, MAC Address Length, MAC Address, IP Address Length, and IP 585 Address Address fields are considered to be part of the prefix in the 586 NLRI. The Ethernet Segment Identifier and MPLS Label fields are to be 587 treated as route attributes as opposed to being part of the "route". 588 The IP address length is in bits. 590 For procedures and usage of this route please see section 9 591 "Determining Reachability to Unicast MAC Addresses" and section 14 592 "Load Balancing of Unicast Packets". 594 7.3. Inclusive Multicast Ethernet Tag Route 596 An Inclusive Multicast Ethernet Tag route type specific EVPN NLRI 597 consists of the following: 599 +---------------------------------------+ 600 | RD (8 octets) | 601 +---------------------------------------+ 602 | Ethernet Tag ID (4 octets) | 603 +---------------------------------------+ 604 | IP Address Length (1 octet) | 605 +---------------------------------------+ 606 | Originating Router's IP Addr | 607 | (4 or 16 octets) | 608 +---------------------------------------+ 610 For procedures and usage of this route please see section 11 611 "Handling of Multi-Destination Traffic", section 13 "Processing of 612 Unknown Unicast Traffic" and section 16 "Multicast". The IP address 613 length is in bits. For the purpose of BGP route key processing, only 614 the Ethernet Tag ID, IP Address Length, and Originating Router's IP 615 Address fields are considered to be part of the prefix in the NLRI. 617 7.4 Ethernet Segment Route 619 An Ethernet Segment route type specific EVPN NLRI consists of the 620 following: 622 +---------------------------------------+ 623 | RD (8 octets) | 624 +---------------------------------------+ 625 |Ethernet Segment Identifier (10 octets)| 626 +---------------------------------------+ 627 | IP Address Length (1 octet) | 628 +---------------------------------------+ 629 | Originating Router's IP Addr | 630 | (4 or 16 octets) | 631 +---------------------------------------+ 633 For procedures and usage of this route please see section 8.5 634 "Designated Forwarder Election". The IP address length is in bits. 635 For the purpose of BGP route key processing, only the Ethernet 636 Segment ID, IP Address Length, and Originating Router's IP Address 637 fields are considered to be part of the prefix in the NLRI. 639 7.5 ESI Label Extended Community 641 This extended community is a new transitive extended community with 642 the Type field is 0x06, and the Sub-Type of 0x01. It may be 643 advertised along with Ethernet Auto-Discovery routes and it enables 644 split-horizon procedures for multi-homed sites as described in 645 section 8.3 "Split Horizon". 647 Each ESI Label Extended Community is encoded as a 8-octet value as 648 follows: 650 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 1 651 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 652 | Type=0x06 | Sub-Type=0x01 | Flags(1 Octet)| Reserved=0 | 653 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 654 | Reserved = 0 | ESI Label | 655 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 657 The low order bit of the flags octet is defined as the "Single- 658 Active" bit. A value of 0 means that the multi-homed site is 659 operating in All-Active redundancy mode and a value of 1 means that 660 the multi-homed site is operating in Single-Active redundancy mode. 662 7.6 ES-Import Route Target 664 This is a new transitive Route Target extended community carried with 665 the Ethernet Segment route. When used, it enables all the PEs 666 connected to the same multi-homed site to import the Ethernet Segment 667 routes. The value is derived automatically from the ESI by encoding 668 the high order 6-byte portion of the 9-byte ESI Value in the ES- 669 Import Route Target. The high order 6-byte of the ESI incorporates 670 MAC address of ESI (for type 1, 2, and 3) which when encoded in this 671 RT and used in the RT constrain feature, it enables proper route- 672 target filtering. The format of this extended community is as 673 follows: 675 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 1 676 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 677 | Type=0x06 | Sub-Type=0x02 | ES-Import | 678 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 679 | ES-Import Cont'd | 680 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 682 This document expands the definition of the Route Target extended 683 community to allow the value of high order octet (Type field) to be 684 0x06 (in addition to the values specified in rfc4360). The value of 685 low order octet (Sub-Type field) of 0x02 indicates that this extended 686 community is of type "Route Target". The new value for Type field of 687 0x06 indicates that the structure of this RT is a six bytes value 688 (e.g., a MAC address). A BGP speaker that implements RT-Constrain 689 [RFC4684] MUST apply the RT Constraint procedures to the ES-import RT 690 as well. 692 For procedures and usage of this attribute, please see section 8.1 693 "Multi-homed Ethernet Segment Auto-Discovery". 695 7.7 MAC Mobility Extended Community 697 This extended community is a new transitive extended community with 698 the Type field of 0x06 and the Sub-Type of 0x00. It may be advertised 699 along with MAC Advertisement routes. The procedures for using this 700 Extended Community are described in section 15 "MAC Mobility". 702 The MAC Mobility Extended Community is encoded as an 8-octet value as 703 follows: 705 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 1 706 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 707 | Type=0x06 | Sub-Type=0x00 |Flags(1 octet)| Reserved=0 | 708 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 709 | Sequence Number | 710 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 712 The low order bit of the flags octet is defined as the 713 "Sticky/static" flag and may be set to 1. A value of 1 means that the 714 MAC address is static and cannot move. The sequence number is used to 715 ensure that PEs retain the correct MAC advertisement route when 716 multiple updates occur for the same MAC address. 718 7.8 Default Gateway Extended Community 720 The Default Gateway community is an Extended Community of an Opaque 721 Type (see 3.3 of rfc4360). It is a transitive community, which means 722 that the first octet is 0x03. The value of the second octet (Sub- 723 Type) is 0x0d (Default Gateway) as assigned by IANA. The Value field 724 of this community is reserved (set to 0 by the senders, ignored by 725 the receivers). 727 7.9 Route Distinguisher Assignment per EVI 729 Route Distinguisher (RD) MUST be set to the RD of the EVI that is 730 advertising the NLRI. An RD MUST be assigned for a given EVI on a PE. 731 This RD MUST be unique across all EVIs on a PE. It is RECOMMENDED to 732 use the Type 1 RD [RFC4364]. The value field comprises an IP address 733 of the PE (typically, the loopback address) followed by a number 734 unique to the PE. This number may be generated by the PE. Or in the 735 Unique VLAN EVPN case, the low order 12 bits may be the 12 bit VLAN 736 ID, with the remaining high order 4 bits set to 0. 738 7.10 Route Targets 740 The EVPN route MAY carry one or more Route Target (RT) attributes. 741 RTs may be configured (as in IP VPNs), or may be derived 742 automatically. 744 If a PE uses RT-Constrain, the PE SHOULD advertise all such RTs using 745 RT Constraints. The use of RT Constrains allows each Ethernet A-D 746 route to reach only those PEs that are configured to import at least 747 one RT from the set of RTs carried in the EVPN route. 749 7.10.1 Auto-Derivation from the Ethernet Tag ID 751 For the "Unique VLAN EVPN" scenario, it is highly desirable to auto- 752 derive the RT from the Ethernet Tag ID (VLAN ID) for that EVPN 753 instance. The following is the procedure for performing such auto- 754 derivation. 756 + The Global Administrator field of the RT MUST be set 757 to the Autonomous System (AS) number that the PE is 758 associated with. 760 + The 12-bit VLAN ID MUST be encoded in the lowest 12 bits of 761 the Local Administrator field. 763 8. Multi-homing Functions 765 This section discusses the functions, procedures and associated BGP 766 routes used to support multi-homing in EVPN. This covers both multi- 767 homed device (MHD) as well as multi-homed network (MHN) scenarios. 769 8.1 Multi-homed Ethernet Segment Auto-Discovery 771 PEs connected to the same Ethernet segment can automatically discover 772 each other with minimal to no configuration through the exchange of 773 the Ethernet Segment route. 775 8.1.1 Constructing the Ethernet Segment Route 777 The Route-Distinguisher (RD) MUST be a Type 1 RD [RFC4364]. The value 778 field comprises an IP address of the PE (typically, the loopback 779 address) followed by 0's. 781 The Ethernet Segment Identifier MUST be set to the ten octet ESI 782 identifier described in section 5. 784 The BGP advertisement that advertises the Ethernet Segment route MUST 785 also carry an ES-Import route target, as defined in section 7.6. 787 The Ethernet Segment Route filtering MUST be done such that the 788 Ethernet Segment Route is imported only by the PEs that are multi- 789 homed to the same Ethernet Segment. To that end, each PE that is 790 connected to a particular Ethernet segment constructs an import 791 filtering rule to import a route that carries the ES-Import extended 792 community, constructed from the ESI. 794 8.2 Fast Convergence 796 In EVPN, MAC address reachability is learnt via the BGP control-plane 797 over the MPLS network. As such, in the absence of any fast protection 798 mechanism, the network convergence time is a function of the number 799 of MAC Advertisement routes that must be withdrawn by the PE 800 encountering a failure. For highly scaled environments, this scheme 801 yields slow convergence. 803 To alleviate this, EVPN defines a mechanism to efficiently and 804 quickly signal, to remote PE nodes, the need to update their 805 forwarding tables upon the occurrence of a failure in connectivity to 806 an Ethernet segment. This is done by having each PE advertise a set 807 of Ethernet A-D per Ethernet segment (per ES) routes for each locally 808 attached Ethernet segment (refer to section 8.2.1 below for details 809 on how this route is constructed). Upon a failure in connectivity to 810 the attached segment, the PE withdraws the corresponding Ethernet A-D 811 route. This triggers all PEs that receive the withdrawal to update 812 their next-hop adjacencies for all MAC addresses associated with the 813 Ethernet segment in question. If no other PE had advertised an 814 Ethernet A-D route for the same segment, then the PE that received 815 the withdrawal simply invalidates the MAC entries for that segment. 816 Otherwise, the PE updates the next-hop adjacencies to point to the 817 backup PE(s). 819 8.2.1 Constructing the Ethernet A-D per Ethernet Segment (ES) Route 821 This section describes the procedures used to construct the Ethernet 822 A-D per ES route, which is used for fast convergence (as discussed 823 above) and for advertising the ESI label used for split-horizon 824 filtering (as discussed in section 8.3). Support of this route is 825 MANDATORY. 827 The Route-Distinguisher (RD) MUST be a Type 1 RD [RFC4364]. The value 828 field comprises an IP address of the PE (typically, the loopback 829 address) followed by a number unique to the PE. 831 The Ethernet Segment Identifier MUST be a ten octet entity as 832 described in section "Ethernet Segment". The Ethernet A-D route is 833 not needed when the Segment Identifier is set to 0 (e.g., single- 834 homed scenarios). 836 The Ethernet Tag ID MUST be set to MAX-ET. 838 The MPLS label in the NLRI MUST be set to 0. 840 The "ESI Label Extended Community" MUST be included in the route. If 841 All-Active redundancy mode is desired, then the "Single-Active" bit 842 in the flags of the ESI Label Extended Community MUST be set to 0 and 843 the MPLS label in that extended community MUST be set to a valid MPLS 844 label value. The MPLS label in this Extended Community is referred to 845 as the ESI label and MUST have the same value in each Ethernet A-D 846 per ES route advertised for the ES. This label MUST be a downstream 847 assigned MPLS label if the advertising PE is using ingress 848 replication for receiving multicast, broadcast or unknown unicast 849 traffic from other PEs. If the advertising PE is using P2MP MPLS LSPs 850 for sending multicast, broadcast or unknown unicast traffic, then 851 this label MUST be an upstream assigned MPLS label. The usage of this 852 label is described in section 8.3. 854 If Single-Active redundancy mode is desired, then the "Single-Active" 855 bit in the flags of the ESI Label Extended Community MUST be set to 1 856 and the ESI label MUST be set to zero. 858 8.2.1.1. Ethernet A-D Route Targets 860 Each Ethernet A-D per ES route MUST carry one or more Route Target 861 (RT) attributes. The set of Ethernet A-D routes per ES MUST carry the 862 entire set of RTs for all the EVPN instances to which the Ethernet 863 Segment belongs. 865 8.3 Split Horizon 867 Consider a CE that is multi-homed to two or more PEs on an Ethernet 868 segment ES1 operating in All-Active redundancy mode. If the CE sends 869 a broadcast, unknown unicast, or multicast (BUM) packet to one of the 870 non-DF (Designated Forwarder) PEs, say PE1, then PE1 will forward 871 that packet to all or subset of the other PEs in that EVPN instance 872 including the DF PE for that Ethernet segment. In this case the DF PE 873 that the CE is multi-homed to MUST drop the packet and not forward 874 back to the CE. This filtering is referred to as "split horizon" 875 filtering in this document. 877 When a set of PEs operating in Single-Active redundancy mode, the use 878 of this split-horizon filtering mechanism is highly recommended 879 because it prevents transient loop at the time of failure or recovery 880 impacting the Ethernet Segment - e.g., when two PEs thinks that both 881 are DFs for that segment before DF election procedure settles down. 883 In order to achieve this split horizon function, every BUM packet 884 originating from a non-DF PE is encapsulated with an MPLS label that 885 identifies the Ethernet segment of origin (i.e. the segment from 886 which the frame entered the EVPN network). This label is referred to 887 as the ESI label, and MUST be distributed by all PEs when operating 888 in All-Active redundancy mode using a set of Ethernet A-D per ES 889 routes per section 8.2.1 above. The ESI label SHOULD be distributed 890 by all PEs when operating in Single-Active redundancy mode using a 891 set of Ethernet A-D per ES route. This route is imported by the PEs 892 connected to the Ethernet Segment and also by the PEs that have at 893 least one EVPN instance in common with the Ethernet Segment in the 894 route. As described in section 8.1.1, the route MUST carry an ESI 895 Label Extended Community with a valid ESI label. The disposition PE 896 rely on the value of the ESI label to determine whether or not a BUM 897 frame is allowed to egress a specific Ethernet segment. 899 8.3.1 ESI Label Assignment 901 The following subsections describe the assignment procedures for the 902 ESI label, which differ depending on the type of tunnels being used 903 to deliver multi-destination packets in the EVPN network. 905 8.3.1.1 Ingress Replication 907 Each PE attached to a given ES that is operating in All-Active or 908 Single-Active redundancy mode and that uses ingress replication to 909 receive BUM traffic advertises a downstream assigned ESI label in the 910 set of Ethernet A-D per ES routes for that ES. This label MUST be 911 programmed in the platform label space by the advertising PE and the 912 forwarding entry for this label must result in NOT forwarding packets 913 received with this label onto the Ethernet segment for which the 914 label was distributed. 916 The rules for the inclusion of the ESI label in a BUM packet by the 917 ingress PE operating in All-Active redundancy mode are as follows: 919 A non-DF ingress PE MUST include the ESI label distributed by the DF 920 egress PE in the copy of a BUM packet sent to it. 922 An ingress PE (DF or non-DF) SHOULD include the ESI label distributed 923 by each non-DF egress PE in the copy of a BUM packet sent to it. 925 The rules for the inclusion of the ESI label in a BUM packet by the 926 ingress PE operating in Single-Active redundancy mode are as follows: 928 An ingress DF PE SHOULD include the ESI label distributed by the 929 egress PE in the copy of a BUM packet sent to it. 931 In both All-Active and Single-Active redundancy mode, an ingress PE 932 MUST NOT include an ESI label in the copy of a BUM packet sent to an 933 egress PE that is not attached to the ES through which the BUM packet 934 entered the EVI. 936 As an example, consider PE1 and PE2 that are multi-homed to CE1 on 937 ES1 and operating in All-Active multi-homing mode. Further consider 938 that PE1 is using P2P or MP2P LSPs to send packets to PE2. Consider 939 that PE1 is the non-DF for VLAN1 and PE2 is the DF for VLAN1, and PE1 940 receives a BUM packet from CE1 on VLAN1 on ES1. In this scenario, PE2 941 distributes an Inclusive Multicast Ethernet Tag route for VLAN1 942 corresponding to an EVPN instance. So, when PE1 sends a BUM packet, 943 that it receives from CE1, it MUST first push onto the MPLS label 944 stack the ESI label that PE2 has distributed for ES1. It MUST then 945 push on the MPLS label distributed by PE2 in the Inclusive Multicast 946 Ethernet Tag route for VLAN1. The resulting packet is further 947 encapsulated in the P2P or MP2P LSP label stack required to transmit 948 the packet to PE2. When PE2 receives this packet, it determines the 949 set of ESIs to replicate the packet to from the top MPLS label, after 950 any P2P or MP2P LSP labels have been removed. If the next label is 951 the ESI label assigned by PE2 for ES1, then PE2 MUST NOT forward the 952 packet onto ES1. If the next label is an ESI label which has not been 953 assigned by PE2, then PE2 MUST drop the packet. It should be noted 954 that in this scenario, if PE2 receives a BUM packet for VLAN1 from 955 CE1, then it SHOULD encapsulate the packet with an ESI label received 956 from PE1 when sending it to PE1 in order to avoid any transient loop 957 during a failure scenario impacting ES1 (e.g., port or link failure). 959 8.3.1.2. P2MP MPLS LSPs 961 The non-DF PEs attached to a given ES that is operating in All-Active 962 redundancy mode and that use P2MP LSPs to send BUM traffic advertise 963 an upstream assigned ESI label in the set of Ethernet A-D per ES 964 routes for that ES. This label is upstream assigned by the PE that 965 advertises the route. This label MUST be programmed by the other PEs, 966 that are connected to the ESI advertised in the route, in the context 967 label space for the advertising PE. Further the forwarding entry for 968 this label must result in NOT forwarding packets received with this 969 label onto the Ethernet segment that the label was distributed for. 970 This label MUST also be programmed by the other PEs, that import the 971 route but are not connected to the ESI advertised in the route, in 972 the context label space for the advertising PE. Further the 973 forwarding entry for this label must be a POP with no other 974 associated action. 976 The DF PE attached to a given ES that is operating in Single-Active 977 redundancy mode and that use P2MP LSPs to send BUM traffic should 978 advertise an upstream assigned ESI label in the set of Ethernet A-D 979 per ES routes for that ES just as above paragraph. 981 As an example, consider PE1 and PE2 that are multi-homed to CE1 on 982 ES1 and operating in All-Active multi-homing mode. Also consider PE3 983 belongs to one of the EVPN instances of ES1. Further, assume that 984 PE1 which is the non-DF, using P2MP MPLS LSPs to send BUM packets. 985 When PE1 sends a BUM packet, that it receives from CE1, it MUST first 986 push onto the MPLS label stack the ESI label that it has assigned for 987 the ESI that the packet was received on. The resulting packet is 988 further encapsulated in the P2MP MPLS label stack necessary to 989 transmit the packet to the other PEs. Penultimate hop popping MUST be 990 disabled on the P2MP LSPs used in the MPLS transport infrastructure 991 for EVPN. When PE2 receives this packet, it de-capsulates the top 992 MPLS label and forwards the packet using the context label space 993 determined by the top label. If the next label is the ESI label 994 assigned by PE1 to ES1, then PE2 MUST NOT forward the packet onto 995 ES1. When PE3 receives this packet, it de-capsulates the top MPLS 996 label and forwards the packet using the context label space 997 determined by the top label. If the next label is the ESI label 998 assigned by PE1 to ES1 and PE3 is not connected to ES1, then PE3 MUST 999 pop the label and flood the packet over all local ESIs in that EVPN 1000 instance. It should be noted that when PE2 sends a BUM frame over a 1001 P2MP LSP, it should encapsulate the frame with an ESI label even 1002 though it is the DF for that VLAN in order to avoid any transient 1003 loop during a failure scenario impacting ES1 (e.g., port or link 1004 failure). 1006 8.4 Aliasing and Backup-Path 1008 In the case where a CE is multi-homed to multiple PE nodes, using a 1009 LAG with All-Active redundancy, it is possible that only a single PE 1010 learns a set of the MAC addresses associated with traffic transmitted 1011 by the CE. This leads to a situation where remote PE nodes receive 1012 MAC advertisement routes, for these addresses, from a single PE even 1013 though multiple PEs are connected to the multi-homed segment. As a 1014 result, the remote PEs are not able to effectively load-balance 1015 traffic among the PE nodes connected to the multi-homed Ethernet 1016 segment. This could be the case, for e.g. when the PEs perform data- 1017 plane learning on the access, and the load-balancing function on the 1018 CE hashes traffic from a given source MAC address to a single PE. 1019 Another scenario where this occurs is when the PEs rely on control 1020 plane learning on the access (e.g. using ARP), since ARP traffic will 1021 be hashed to a single link in the LAG. 1023 To address this issue, EVPN introduces the concept of 'Aliasing' 1024 which is the ability of a PE to signal that it has reachability to an 1025 EVPN instance on a given ES even when it has learnt no MAC addresses 1026 from that EVI/ES. The Ethernet A-D per EVI route is used for this 1027 purpose. A remote PE that receives a MAC advertisement route with 1028 non-reserved ESI SHOULD consider the advertised MAC address to be 1029 reachable via all PEs that have advertised reachability to that MAC 1030 address' EVI/ES via the combination of an Ethernet A-D per EVI route 1031 for that EVI/ES (and Ethernet Tag if applicable) AND Ethernet A-D per 1032 ES routes for that ES with the 'Single-Active' bit in the flags of 1033 the ESI Label Extended Community set to 0. 1035 Note that the Ethernet A-D per EVI route may be received by a remote 1036 PE before it receives the set of Ethernet A-D per ES routes. 1038 Therefore, in order to handle corner cases and race conditions, the 1039 Ethernet A-D per EVI route MUST NOT be used for traffic forwarding by 1040 a remote PE until it also receives the associated set of Ethernet A-D 1041 per ES routes. 1043 Backup-path is a closely related function, but it is used in Single- 1044 Active redundancy mode. In this case a PE also advertises that it 1045 has reachability to a give EVI/ES using same combination of Ethernet 1046 A-D per EVI route and Ethernet A-D per ES route as above, but with 1047 the 'Single-Active' bit in the flags of the ESI Label Extended 1048 Community set to 1. A remote PE that receives a MAC advertisement 1049 route with non-reserved ESI SHOULD consider the advertised MAC 1050 address to be reachable via any PE that has advertised this 1051 combination of Ethernet A-D routes and it SHOULD install a backup- 1052 path for that MAC address. 1054 8.4.1 Constructing the Ethernet A-D per EVPN Instance (EVI) Route 1056 This section describes the procedures used to construct the Ethernet 1057 A-D per EVPN Instance (EVI) route, which is used for aliasing (as 1058 discussed above). Support of this route is OPTIONAL. 1060 Route-Distinguisher (RD) MUST be set to the RD of the EVI that is 1061 advertising the NLRI per section 7.9. 1063 The Ethernet Segment Identifier MUST be a ten octet entity as 1064 described in section "Ethernet Segment Identifier". The Ethernet A-D 1065 route is not needed when the Segment Identifier is set to 0. 1067 The Ethernet Tag ID is the identifier of an Ethernet Tag on the 1068 Ethernet segment. This value may be a 12 bit VLAN ID, in which case 1069 the low order 12 bits are set to the VLAN ID and the high order 20 1070 bits are set to 0. Or it may be another Ethernet Tag used by the 1071 EVPN. It MAY be set to the default Ethernet Tag on the Ethernet 1072 segment or to the value 0. 1074 Note that the above allows the Ethernet A-D route to be advertised 1075 with one of the following granularities: 1077 + One Ethernet A-D route for a given tuple 1078 per EVI. This is applicable when the PE uses MPLS-based 1079 disposition. 1081 + One Ethernet A-D route per (where the Ethernet 1082 Tag ID is set to 0). This is applicable when the PE uses 1083 MAC-based disposition, or when the PE uses MPLS-based 1084 disposition when no VLAN translation is required. 1086 The usage of the MPLS label is described in the section on "Load 1087 Balancing of Unicast Packets". 1089 The Next Hop field of the MP_REACH_NLRI attribute of the route MUST 1090 be set to the IPv4 or IPv6 address of the advertising PE. 1092 The Ethernet A-D route MUST carry one or more Route Target (RT) 1093 attributes per section 7.10. 1095 8.5 Designated Forwarder Election 1097 Consider a CE that is a host or a router that is multi-homed directly 1098 to more than one PE in an EVPN instance on a given Ethernet segment. 1099 One or more Ethernet Tags may be configured on the Ethernet segment. 1100 In this scenario only one of the PEs, referred to as the Designated 1101 Forwarder (DF), is responsible for certain actions: 1103 - Sending multicast and broadcast traffic, on a given Ethernet 1104 Tag on a particular Ethernet segment, to the CE. 1106 - Flooding unknown unicast traffic (i.e. traffic for 1107 which a PE does not know the destination MAC address), 1108 on a given Ethernet Tag on a particular Ethernet segment 1109 to the CE, if the environment requires flooding of 1110 unknown unicast traffic. 1112 Note that this behavior, which allows selecting a DF at the 1113 granularity of for multicast, broadcast and unknown 1114 unicast traffic, is the default behavior in this specification. 1116 Note that a CE always sends packets belonging to a specific flow 1117 using a single link towards a PE. For instance, if the CE is a host 1118 then, as mentioned earlier, the host treats the multiple links that 1119 it uses to reach the PEs as a Link Aggregation Group (LAG). The CE 1120 employs a local hashing function to map traffic flows onto links in 1121 the LAG. 1123 If a bridged network is multi-homed to more than one PE in an EVPN 1124 network via switches, then the support of All-Active redundancy mode 1125 requires the bridged network to be connected to two or more PEs using 1126 a LAG. 1128 If a bridged network does not connect to the PEs using LAG, then only 1129 one of the links between the switched bridged network and the PEs 1130 must be the active link for a given EVPN instance. In this case, the 1131 set of Ethernet A-D per ES routes advertised by each PE MUST have the 1132 'Single-Active' bit in the flags of the ESI Label Extended Community 1133 set to 1. 1135 The default procedure for DF election at the granularity of is referred to as "service carving". With service carving, it is 1137 possible to elect multiple DFs per Ethernet Segment (one per EVI) in 1138 order to perform load-balancing of multi-destination traffic destined 1139 to a given Segment. The load-balancing procedures carve up the EVI 1140 space among the PE nodes evenly, in such a way that every PE is the 1141 DF for a disjoint set of EVIs. The procedure for service carving is 1142 as follows: 1144 1. When a PE discovers the ESI of the attached Ethernet Segment, it 1145 advertises an Ethernet Segment route with the associated ES-Import 1146 extended community attribute. 1148 2. The PE then starts a timer (default value = 3 seconds) to allow 1149 the reception of Ethernet Segment routes from other PE nodes 1150 connected to the same Ethernet Segment. This timer value MUST be same 1151 across all PEs connected to the same Ethernet Segment. 1153 3. When the timer expires, each PE builds an ordered list of the IP 1154 addresses of all the PE nodes connected to the Ethernet Segment 1155 (including itself), in increasing numeric value. Each IP address in 1156 this list is extracted from the "Originator Router's IP address" 1157 field of the advertised Ethernet Segment route. Every PE is then 1158 given an ordinal indicating its position in the ordered list, 1159 starting with 0 as the ordinal for the PE with the numerically lowest 1160 IP address. The ordinals are used to determine which PE node will be 1161 the DF for a given EVPN instance on the Ethernet Segment using the 1162 following rule: Assuming a redundancy group of N PE nodes, the PE 1163 with ordinal i is the DF for an EVPN instance with an associated 1164 Ethernet Tag value V when (V mod N) = i. In the case where multiple 1165 Ethernet Tags are associated with a single EVPN instance, then the 1166 numerically lowest Ethernet Tag value in that EVPN instance MUST be 1167 used in the modulo function. 1169 It should be noted that using "Originator Router's IP address" field 1170 in the Ethernet Segment route to get the PE IP address needed for the 1171 ordered list, allows for a CE to be multi-homed across different ASes 1172 if such need ever arises. 1174 4. The PE that is elected as a DF for a given EVPN instance will 1175 unblock traffic for the Ethernet Tags associated with that EVPN 1176 instance. Note that the DF PE unblocks multi-destination traffic in 1177 the egress direction towards the Segment. All non-DF PEs continue to 1178 drop multi-destination traffic (for the associated EVPN instances) in 1179 the egress direction towards the Segment. 1181 In the case of link or port failure, the affected PE withdraws its 1182 Ethernet Segment route. This will re-trigger the service carving 1183 procedures on all the PEs in the RG. For PE node failure, or upon PE 1184 commissioning or decommissioning, the PEs re-trigger the service 1185 carving. In case of a Single-Active multi-homing, when a service 1186 moves from one PE in the RG to another PE as a result of re-carving, 1187 the PE, which ends up being the elected DF for the service, SHOULD 1188 trigger a MAC address flush notification towards the associated 1189 Ethernet Segment. This can be done, for e.g. using IEEE 802.1ak MVRP 1190 'new' declaration. 1192 8.6. Interoperability with Single-homing PEs 1194 Let's refer to PEs that only support single-homed CE devices as 1195 single-homing PEs. For single-homing PEs, all the above multi-homing 1196 procedures can be omitted; however, to allow for single-homing PEs to 1197 fully inter-operate with multi-homing PEs, some of the multi-homing 1198 procedures described above SHOULD be supported even by single-homing 1199 PEs: 1201 - procedures related to processing Ethernet A-D route for the purpose 1202 of Fast Convergence (8.2 Fast Convergence), to let single-homing PEs 1203 benefit from fast convergence 1205 - procedures related to processing Ethernet A-D route for the purpose 1206 of Aliasing (8.4 Aliasing and Backup-path), to let single-homing PEs 1207 benefit from load balancing 1209 - procedures related to processing Ethernet A-D route for the purpose 1210 of Backup-path (8.4 Aliasing and Backup-path), to let single-homing 1211 PEs to benefit from the corresponding convergence improvement 1213 9. Determining Reachability to Unicast MAC Addresses 1215 PEs forward packets that they receive based on the destination MAC 1216 address. This implies that PEs must be able to learn how to reach a 1217 given destination unicast MAC address. 1219 There are two components to MAC address learning, "local learning" 1220 and "remote learning": 1222 9.1. Local Learning 1224 A particular PE must be able to learn the MAC addresses from the CEs 1225 that are connected to it. This is referred to as local learning. 1227 The PEs in a particular EVPN instance MUST support local data plane 1228 learning using standard IEEE Ethernet learning procedures. A PE must 1229 be capable of learning MAC addresses in the data plane when it 1230 receives packets such as the following from the CE network: 1232 - DHCP requests 1234 - ARP request for its own MAC. 1236 - ARP request for a peer. 1238 Alternatively PEs MAY learn the MAC addresses of the CEs in the 1239 control plane or via management plane integration between the PEs and 1240 the CEs. 1242 There are applications where a MAC address that is reachable via a 1243 given PE on a locally attached Segment (e.g. with ESI X) may move 1244 such that it becomes reachable via another PE on another Segment 1245 (e.g. with ESI Y). This is referred to as a "MAC Mobility". 1246 Procedures to support this are described in section "MAC Mobility". 1248 9.2. Remote learning 1250 A particular PE must be able to determine how to send traffic to MAC 1251 addresses that belong to or are behind CEs connected to other PEs 1252 i.e. to remote CEs or hosts behind remote CEs. We call such MAC 1253 addresses "remote" MAC addresses. 1255 This document requires a PE to learn remote MAC addresses in the 1256 control plane. In order to achieve this, each PE advertises the MAC 1257 addresses it learns from its locally attached CEs in the control 1258 plane, to all the other PEs in that EVPN instance, using MP-BGP and 1259 specifically the MAC Advertisement route. 1261 9.2.1. Constructing the BGP EVPN MAC/IP Address Advertisement 1263 BGP is extended to advertise these MAC addresses using the MAC/IP 1264 Advertisement route type in the EVPN NLRI. 1266 The RD MUST be the RD of the EVI that is advertising the NLRI. The 1267 procedures for setting the RD for a given EVI are described in 1268 section 7.9. 1270 The Ethernet Segment Identifier is set to the ten octet ESI described 1271 in section "Ethernet Segment". 1273 The Ethernet Tag ID may be zero or may represent a valid Ethernet Tag 1274 ID. This field may be non-zero when there are multiple bridge 1275 domains in the MAC-VRF (i.e., the PE needs to perform qualified 1276 learning for the VLANs in that MAC-VRF). 1278 When the the Ethernet Tag ID in the NLRI is set to a non-zero value, 1279 for a particular bridge domain, then this Ethernet Tag ID may either 1280 be the CE's Ethernet tag value (e.g., CE VLAN ID) or the EVPN 1281 provider's Ethernet tag value (e.g., provider VLAN ID). The latter 1282 would be the case if the CE Ethernet tags (e.g., CE VLAN ID) for a 1283 particular bridge domain are different on different CEs. 1285 The MAC address length field is in bits and it is set to 48. The MAC 1286 address length values other than 48 bits, are outside the scope of 1287 this document. The encoding of a MAC address MUST be the 6-octet MAC 1288 address specified by [802.1D-ORIG] [802.1D-REV]. 1290 The IP Address Field is optional. By default, the IP Address Length 1291 field is set to 0 and the IP address field is omitted from the route. 1292 When a valid IP address needs to be advertised, it is then encoded in 1293 this route. When an IP address is present, the IP Address Length 1294 field is in bits and it is set to 32 or 128 bits. Other IP Address 1295 Length values are outside the scope of this document. The encoding of 1296 an IP address MUST be either 4 octets for IPv4 or 16 octets for IPv6. 1297 The length field of EVPN NLRI (which is in octets and is described in 1298 section 7) is sufficient to determine whether an IP address is 1299 encoded in this route and if so, whether the encoded IP address is 1300 IPV4 or IPv6. 1302 The MPLS label1 field is encoded as 3 octets, where the high-order 20 1303 bits contain the label value. The MPLS label1 MUST be downstream 1304 assigned and it is associated with the MAC address being advertised 1305 by the advertising PE. The advertising PE uses this label when it 1306 receives an MPLS-encapsulated packet to perform forwarding based on 1307 the destination MAC address toward the CE. The forwarding procedures 1308 are specified in sections 13 and 14. 1310 A PE may advertise the same single EVPN label for all MAC addresses 1311 in a given EVI. This label assignment is referred to as a per EVI 1312 label assignment. Alternatively, a PE may advertise a unique EVPN 1313 label per combination. This label assignment is 1314 referred to as a per label assignment. As a third 1315 option, a PE may advertise a unique EVPN label per MAC address. This 1316 label assignment is referred to as a per MAC label assignment. All of 1317 these label assignment methods have their tradeoffs. The choice of a 1318 particular label assignment methodology is purely local to the PE 1319 that originates the route. 1321 Per EVI label assignment requires the least number of EVPN labels, 1322 but requires a MAC lookup in addition to an MPLS lookup on an egress 1323 PE for forwarding. On the other hand, a unique label per or a unique label per MAC allows an egress PE to 1325 forward a packet that it receives from another PE, to the connected 1326 CE, after looking up only the MPLS labels without having to perform a 1327 MAC lookup. This includes the capability to perform appropriate VLAN 1328 ID translation on egress to the CE. 1330 The MPLS label2 field is an optional field and if it is present, then 1331 it is encoded as 3 octets, where the high-order 20 bits contain the 1332 label value. 1334 The Next Hop field of the MP_REACH_NLRI attribute of the route MUST 1335 be set to the IPv4 or IPv6 address of the advertising PE. 1337 The BGP advertisement for the MAC advertisement route MUST also carry 1338 one or more Route Target (RT) attributes. RTs may be configured (as 1339 in IP VPNs), or may be derived automatically from the Ethernet Tag 1340 ID, in the Unique VLAN case, as described in section 8.4.1.1.1. 1342 It is to be noted that this document does not require PEs to create 1343 forwarding state for remote MACs when they are learnt in the control 1344 plane. When this forwarding state is actually created is a local 1345 implementation matter. 1347 9.2.2 Route Resolution 1349 If the Ethernet Segment Identifier field in a received MAC 1350 Advertisement route is set to the reserved ESI value of 0 or MAX-ESI, 1351 then if the receiving PE decides to install forwarding state for the 1352 associated MAC address, it MUST be based on the MAC Advertisement 1353 route alone. 1355 If the Ethernet Segment Identifier field in a received MAC 1356 Advertisement route is set to a non-reserved ESI, and the receiving 1357 PE is locally attached to the same ESI, then the PE does not alter 1358 its forwarding state based on the received route. This ensures that 1359 local routes are preferred to remote routes. 1361 If the Ethernet Segment Identifier field in a received MAC 1362 Advertisement route is set to a non-reserved ESI, then if the 1363 receiving PE decides to install forwarding state for the associated 1364 MAC address, it MUST be when both the MAC Advertisement route AND the 1365 associated set of Ethernet A-D per ES routes have been received. The 1366 dependency of MAC routes installation on Ethernet A-D per ES routes, 1367 is to ensure that MAC routes don't get accidentally installed during 1368 mass withdraw period. 1370 To illustrate this with an example, consider two PEs (PE1 and PE2) 1371 connected to a multi-homed Ethernet Segment ES1. All-Active 1372 redundancy mode is assumed. A given MAC address M1 is learnt by PE1 1373 but not PE2. On PE3, the following states may arise: 1375 T1- When the MAC Advertisement Route from PE1 and the set of Ethernet 1376 A-D per ES routes and Ethernet A-D per EVI routes from PE1 and PE2 1377 are received, PE3 can forward traffic destined to M1 to both PE1 and 1378 PE2. 1380 T2- If after T1, PE1 withdraws its set of Ethernet A-D per ES routes, 1381 then PE3 forwards traffic destined to M1 to PE2 only. 1383 T2'- If after T1, PE2 withdraws its set of Ethernet A-D per ES 1384 routes, then PE3 forwards traffic destined to M1 to PE1 only. 1386 T2''- If after T1, PE1 withdraws its MAC Advertisement route, then 1387 PE3 treats traffic to M1 as unknown unicast. 1389 T3- PE2 also advertises a MAC route for M1 and then PE1 withdraws its 1390 MAC route for M1. PE3 continues forwarding traffic destined to M1 1391 to both PE1 and PE2. In other words, despite M1 withdrawal by PE1, 1392 PE3 forwards the traffic destined to M1 to both PE1 and PE2. This is 1393 because a flow from the CE, resulting in M1 traffic getting hashed to 1394 PE1, can get terminated resulting in M1 to aged out in PE1; however, 1395 M1 can be reachable by both PE1 and PE2. 1397 10. ARP and ND 1399 The IP address field in the MAC advertisement route may optionally 1400 carry one of the IP addresses associated with the MAC address. This 1401 provides an option which can be used to minimize the flooding of ARP 1402 or Neighbor Discovery (ND) messages over the MPLS network and to 1403 remote CEs. This option also minimizes ARP (or ND) message processing 1404 on end-stations/hosts connected to the EVPN network. A PE may learn 1405 the IP address associated with a MAC address in the control or 1406 management plane between the CE and the PE. Or, it may learn this 1407 binding by snooping certain messages to or from a CE. When a PE 1408 learns the IP address associated with a MAC address, of a locally 1409 connected CE, it may advertise this address to other PEs by including 1410 it in the MAC Advertisement route. The IP Address may be an IPv4 1411 address encoded using four octets, or an IPv6 address encoded using 1412 sixteen octets. For ARP and ND purposes, the IP Address length field 1413 MUST be set to 32 for an IPv4 address or to 128 for an IPv6 address. 1415 If there are multiple IP addresses associated with a MAC address, 1416 then multiple MAC advertisement routes MUST be generated, one for 1417 each IP address. For instance, this may be the case when there are 1418 both an IPv4 and an IPv6 address associated with the MAC address. 1420 When the IP address is dissociated with the MAC address, then the MAC 1421 advertisement route with that particular IP address MUST be 1422 withdrawn. 1424 When a PE receives an ARP request for an IP address from a CE, and if 1425 the PE has the MAC address binding for that IP address, the PE SHOULD 1426 perform ARP proxy by responding to the ARP request. 1428 10.1 Default Gateway 1430 When a PE needs to perform inter-subnet forwarding where each subnet 1431 is represented by a different broadcast domain (e.g., different VLAN) 1432 the inter-subnet forwarding is performed at layer 3 and the PE that 1433 performs such function is called the default gateway for the EVPN 1434 instance. In this case when the PE receives an ARP Request for the IP 1435 address configured as the default gateway address, the PE originates 1436 an ARP Reply. 1438 Each PE that acts as a default gateway for a given EVPN instance MAY 1439 advertise in the EVPN control plane its default gateway MAC address 1440 using the MAC/IP advertisement route, and indicates that such route 1441 is associated with the default gateway. This is accomplished by 1442 requiring the route to carry the Default Gateway extended community 1443 defined in [Section 7.8 Default Gateway Extended Community]. The ESI 1444 field is set to zero when advertising the MAC route with the Default 1445 Gateway extended community. 1447 The IP address field of the MAC/IP advertisement route is set to the 1448 default GW IP address for that subnet (e.g., EVPN instance). For a 1449 given subnet (e.g., VLAN or EVPN instance), the default GW IP address 1450 is the same across all the participant PEs. The inclusion of this IP 1451 address enables the receiving PE to check its configured default GW 1452 IP address against the one received in the MAC/IP advertisement route 1453 for that subnet (or EVPN instance) and if there is a discrepancy, 1454 then the PE SHOULD notify the operator and log an error message. 1456 Unless it is known a priori (by means outside of this document) that 1457 all PEs of a given EVPN instance act as a default gateway for that 1458 EVPN instance, the MPLS label MUST be set to a valid downstream 1459 assigned label. 1461 Furthermore, even if all PEs of a given EVPN instance do act as a 1462 default gateway for that EVPN instance, but only some, but not all, 1463 of these PEs have sufficient (routing) information to provide inter- 1464 subnet routing for all the inter-subnet traffic originated within the 1465 subnet associated with the EVPN instance, then when such PE 1466 advertises in the EVPN control plane its default gateway MAC address 1467 using the MAC advertisement route, and indicates that such route is 1468 associated with the default gateway, the route MUST carry a valid 1469 downstream assigned label. 1471 If all PEs of a given EVPN instance act as a default gateway for that 1472 EVPN instance, and the same default gateway MAC address is used 1473 across all gateway devices, then no such advertisement is needed. 1474 However, if each default gateway uses a different MAC address, then 1475 each default gateway needs to be aware of other gateways' MAC 1476 addresses and thus the need for such advertisement. This is called 1477 MAC address aliasing since a single default GW can be represented by 1478 multiple MAC addresses. 1480 Each PE that receives this route and imports it as per procedures 1481 specified in this document follows the procedures in this section 1482 when replying to ARP Requests that it receives. 1484 Each PE that acts as a default gateway for a given EVPN instance that 1485 receives this route and imports it as per procedures specified in 1486 this document MUST create MAC forwarding state that enables it to 1487 apply IP forwarding to the packets destined to the MAC address 1488 carried in the route. 1490 11. Handling of Multi-Destination Traffic 1492 Procedures are required for a given PE to send broadcast or multicast 1493 traffic, received from a CE encapsulated in a given Ethernet Tag 1494 (VLAN) in an EVPN instance, to all the other PEs that span that 1495 Ethernet Tag (VLAN) in that EVPN instance. In certain scenarios, 1496 described in section "Processing of Unknown Unicast Packets", a given 1497 PE may also need to flood unknown unicast traffic to other PEs. 1499 The PEs in a particular EVPN instance may use ingress replication, 1500 P2MP LSPs or MP2MP LSPs to send unknown unicast, broadcast or 1501 multicast traffic to other PEs. 1503 Each PE MUST advertise an "Inclusive Multicast Ethernet Tag Route" to 1504 enable the above. The following subsection provides the procedures to 1505 construct the Inclusive Multicast Ethernet Tag route. Subsequent 1506 subsections describe in further detail its usage. 1508 11.1. Construction of the Inclusive Multicast Ethernet Tag Route 1510 The RD MUST be the RD of the EVI that is advertising the NLRI. The 1511 procedures for setting the RD for a given EVPN instance on a PE are 1512 described in section 7.9. 1514 The Ethernet Tag ID is the identifier of the Ethernet Tag. It may be 1515 set to 0 or to a valid Ethernet Tag value. 1517 The Originating Router's IP address MUST be set to an IP address of 1518 the PE. This address SHOULD be common for all the EVIs on the PE 1519 (e.,g., this address may be PE's loopback address). The IP Address 1520 Length field is in bits. 1522 The Next Hop field of the MP_REACH_NLRI attribute of the route MUST 1523 be set to the same IP address as the one carried in the Originating 1524 Router's IP Address field. 1526 The BGP advertisement for the Inclusive Multicast Ethernet Tag route 1527 MUST also carry one or more Route Target (RT) attributes. The 1528 assignment of RTs described in the section 7.10 MUST be followed. 1530 11.2. P-Tunnel Identification 1532 In order to identify the P-Tunnel used for sending broadcast, unknown 1533 unicast or multicast traffic, the Inclusive Multicast Ethernet Tag 1534 route MUST carry a "PMSI Tunnel Attribute" as specified in [BGP 1535 MVPN]. 1537 Depending on the technology used for the P-tunnel for the EVPN 1538 instance on the PE, the PMSI Tunnel attribute of the Inclusive 1539 Multicast Ethernet Tag route is constructed as follows. 1541 + If the PE that originates the advertisement uses a 1542 P-Multicast tree for the P-tunnel for EVPN, the PMSI 1543 Tunnel attribute MUST contain the identity of the tree 1544 (note that the PE could create the identity of the 1545 tree prior to the actual instantiation of the tree). 1547 + A PE that uses a P-Multicast tree for the P-tunnel MAY 1548 aggregate two or more EVPN instances (EVIs) present 1549 on the PE onto the same tree. In this case, in addition 1550 to carrying the identity of the tree, the PMSI Tunnel 1551 attribute MUST carry an MPLS upstream assigned label which 1552 the PE has bound uniquely to the EVI associated with this 1553 update (as determined by its RTs). 1555 If the PE has already advertised Inclusive Multicast 1556 Ethernet Tag routes for two or more EVIs that it now 1557 desires to aggregate, then the PE MUST re-advertise 1558 those routes. The re-advertised routes MUST be the same 1559 as the original ones, except for the PMSI Tunnel attribute 1560 and the label carried in that attribute. 1562 + If the PE that originates the advertisement uses ingress 1563 replication for the P-tunnel for EVPN, the route MUST 1564 include the PMSI Tunnel attribute with the Tunnel Type set to 1565 Ingress Replication and Tunnel Identifier set to a routable 1566 address of the PE. The PMSI Tunnel attribute MUST carry a 1567 downstream assigned MPLS label. This label is used to 1568 demultiplex the broadcast, multicast or unknown unicast EVPN 1569 traffic received over a MP2P tunnel by the PE. 1571 + The Leaf Information Required flag of the PMSI Tunnel 1572 attribute MUST be set to zero, and MUST be ignored on receipt. 1574 12. Processing of Unknown Unicast Packets 1576 The procedures in this document do not require the PEs to flood 1577 unknown unicast traffic to other PEs. If PEs learn CE MAC addresses 1578 via a control plane protocol, the PEs can then distribute MAC 1579 addresses via BGP, and all unicast MAC addresses will be learnt prior 1580 to traffic to those destinations. 1582 However, if a destination MAC address of a received packet is not 1583 known by the PE, the PE may have to flood the packet. When flooding, 1584 one must take into account "split horizon forwarding" as follows: The 1585 principles behind the following procedures are borrowed from the 1586 split horizon forwarding rules in VPLS solutions [RFC4761] and 1587 [RFC4762]. When a PE capable of flooding (say PEx) receives an 1588 unknown destination MAC address, it floods the frame. If the frame 1589 arrived from an attached CE, PEx must send a copy of that frame on 1590 every Ethernet Segment (belonging to that EVI) for which it is the 1591 DF, other than the Ethernet Segment on which it received the frame. 1592 In addition, the PE must flood the frame to all other PEs 1593 participating in that EVPN instance. If, on the other hand, the frame 1594 arrived from another PE (say PEy), PEx must send a copy of the packet 1595 on each Ethernet Segment (belonging to that EVI) for which it is the 1596 DF. PEx MUST NOT send the frame to other PEs, since PEy would have 1597 already done so. Split horizon forwarding rules apply to unknown MAC 1598 addresses. 1600 Whether or not to flood packets to unknown destination MAC addresses 1601 should be an administrative choice, depending on how learning happens 1602 between CEs and PEs. 1604 The PEs in a particular EVPN instance may use ingress replication 1605 using RSVP-TE P2P LSPs or LDP MP2P LSPs for sending unknown unicast 1606 traffic to other PEs. Or they may use RSVP-TE P2MP or LDP P2MP for 1607 sending such traffic to other PEs. 1609 12.1. Ingress Replication 1610 If ingress replication is in use, the P-Tunnel attribute, carried in 1611 the Inclusive Multicast Ethernet Tag routes for the EVPN instance, 1612 specifies the downstream label that the other PEs can use to send 1613 unknown unicast, multicast or broadcast traffic for that EVPN 1614 instance to this particular PE. 1616 The PE that receives a packet with this particular MPLS label MUST 1617 treat the packet as a broadcast, multicast or unknown unicast packet. 1618 Further if the MAC address is a unicast MAC address, the PE MUST 1619 treat the packet as an unknown unicast packet. 1621 12.2. P2MP MPLS LSPs 1623 The procedures for using P2MP LSPs are very similar to VPLS 1624 procedures [VPLS-MCAST]. The P-Tunnel attribute used by a PE for 1625 sending unknown unicast, broadcast or multicast traffic for a 1626 particular EVPN instance is advertised in the Inclusive Ethernet Tag 1627 Multicast route as described in section "Handling of Multi- 1628 Destination Traffic". 1630 The P-Tunnel attribute specifies the P2MP LSP identifier. This is the 1631 equivalent of an Inclusive tree in [VPLS-MCAST]. Note that multiple 1632 Ethernet Tags, which may be in different EVPN instances, may use the 1633 same P2MP LSP, using upstream labels [VPLS-MCAST]. This is the 1634 equivalent of an Aggregate Inclusive tree in [VPLS-MCAST]. When P2MP 1635 LSPs are used for flooding unknown unicast traffic, packet re- 1636 ordering is possible. 1638 The PE that receives a packet on the P2MP LSP specified in the PMSI 1639 Tunnel Attribute MUST treat the packet as a broadcast, multicast or 1640 unknown unicast packet. Further if the MAC address is a unicast MAC 1641 address, the PE MUST treat the packet as an unknown unicast packet. 1643 13. Forwarding Unicast Packets 1645 This section describes procedures for forwarding unicast packets by 1646 PEs, where such packets are received from either directly connected 1647 CEs, or from some other PEs. 1649 13.1. Forwarding packets received from a CE 1651 When a PE receives a packet from a CE, on a given Ethernet Tag, it 1652 must first look up the source MAC address of the packet. In certain 1653 environments the source MAC address MAY be used to authenticate the 1654 CE and determine that traffic from the host can be allowed into the 1655 network. Source MAC lookup MAY also be used for local MAC address 1656 learning. 1658 If the PE decides to forward the packet, the destination MAC address 1659 of the packet must be looked up. If the PE has received MAC address 1660 advertisements for this destination MAC address from one or more 1661 other PEs or learned it from locally connected CEs, it is considered 1662 as a known MAC address. Otherwise, the MAC address is considered as 1663 an unknown MAC address. 1665 For known MAC addresses the PE forwards this packet to one of the 1666 remote PEs or to a locally attached CE. When forwarding to a remote 1667 PE, the packet is encapsulated in the EVPN MPLS label advertised by 1668 the remote PE, for that MAC address, and in the MPLS LSP label stack 1669 to reach the remote PE. 1671 If the MAC address is unknown and if the administrative policy on the 1672 PE requires flooding of unknown unicast traffic then: 1674 - The PE MUST flood the packet to other PEs. The PE MUST first 1675 encapsulate the packet in the ESI MPLS label as described in section 1676 8.3. If ingress replication is used, the packet MUST be replicated to 1677 each remote PE with the VPN label being an MPLS label determined as 1678 follows: This is the MPLS label advertised by the remote PE in a PMSI 1679 Tunnel Attribute in the Inclusive Multicast Ethernet Tag route for an 1680 combination. The Ethernet Tag in the 1681 route may be the same as the Ethernet Tag associated with the 1682 interface on which the ingress PE receives the packet. If P2MP LSPs 1683 are being used the packet MUST be sent on the P2MP LSP that the PE is 1684 the root of for the Ethernet Tag in the EVPN instance. If the same 1685 P2MP LSP is used for all Ethernet Tags, then all the PEs in the EVPN 1686 instance MUST be the leaves of the P2MP LSP. If a distinct P2MP LSP 1687 is used for a given Ethernet Tag in the EVPN instance, then only the 1688 PEs in the Ethernet Tag MUST be the leaves of the P2MP LSP. The 1689 packet MUST be encapsulated in the P2MP LSP label stack. 1691 If the MAC address is unknown then, if the administrative policy on 1692 the PE does not allow flooding of unknown unicast traffic: 1694 - The PE MUST drop the packet. 1696 13.2. Forwarding packets received from a remote PE 1698 This section described the procedures for forwarding known and 1699 unknown unicast packets received from a remote PE. 1701 13.2.1. Unknown Unicast Forwarding 1703 When a PE receives an MPLS packet from a remote PE then, after 1704 processing the MPLS label stack, if the top MPLS label ends up being 1705 a P2MP LSP label associated with an EVPN instance or in case of 1706 ingress replication the downstream label advertised in the P-Tunnel 1707 attribute, and after performing the split horizon procedures 1708 described in section 8.3: 1710 - If the PE is the designated forwarder of BUM traffic on a 1711 particular set of ESIs for the Ethernet Tag, the default behavior is 1712 for the PE to flood the packet on these ESIs. In other words, the 1713 default behavior is for the PE to assume that for BUM traffic, it is 1714 not required to perform a destination MAC address lookup. As an 1715 option, the PE may perform a destination MAC lookup to flood the 1716 packet to only a subset of the CE interfaces in the Ethernet Tag. For 1717 instance the PE may decide to not flood an BUM packet on certain 1718 Ethernet segments even if it is the DF on the Ethernet segment, based 1719 on administrative policy. 1721 - If the PE is not the designated forwarder on any of the ESIs for 1722 the Ethernet Tag, the default behavior is for it to drop the packet. 1724 13.2.2. Known Unicast Forwarding 1726 If the top MPLS label ends up being an EVPN label that was advertised 1727 in the unicast MAC advertisements, then the PE either forwards the 1728 packet based on CE next-hop forwarding information associated with 1729 the label or does a destination MAC address lookup to forward the 1730 packet to a CE. 1732 14. Load Balancing of Unicast Frames 1734 This section specifies the load balancing procedures for sending 1735 known unicast frames to a multi-homed CE. 1737 14.1. Load balancing of traffic from a PE to remote CEs 1739 Whenever a remote PE imports a MAC advertisement for a given in an EVI, it MUST examine all imported Ethernet A-D 1741 routes for that ESI in order to determine the load-balancing 1742 characteristics of the Ethernet segment. 1744 14.1.1 Single-Active Redundancy Mode 1746 For a given ES, if the remote PE has imported the set of Ethernet A-D 1747 per ES routes from at least one PE, where the "Single-Active" flag in 1748 the ESI Label Extended Community is set, then the remote PE MUST 1749 deduce that the ES is operating in Single-Active redundancy mode. As 1750 such, the MAC address will be reachable only via the PE announcing 1751 the associated MAC Advertisement route - this is referred to as the 1752 primary PE. The other PEs advertising the set of Ethernet A-D per ES 1753 routes for the same ES provide backup paths for that ES, in case the 1754 primary PE encounters a failure, and are referred to as backup PEs. 1755 It should be noted that the primary PE for a given is the 1756 DF for that . 1758 If the primary PE encounters a failure, it MAY withdraw its set of 1759 Ethernet A-D per ES routes for the affected ES prior to withdrawing 1760 it set of MAC Advertisement routes. 1762 If there is only one backup PE for a given ES, the remote PE MAY use 1763 the primary PE's withdrawal of its set of Ethernet A-D per ES routes 1764 as a trigger to update its forwarding entries, for the associated MAC 1765 addresses, to point towards the backup PE. As the backup PE starts 1766 learning the MAC addresses over its attached ES, it will start 1767 sending MAC Advertisement routes while the failed PE withdraws its 1768 routes. This mechanism minimizes the flooding of traffic during fail- 1769 over events. 1771 If there is more than one backup PE for a given ES, the remote PE 1772 MUST use the primary PE's withdrawal of its set of Ethernet A-D per 1773 ES routes as a trigger to start flooding traffic for the associated 1774 MAC addresses (as long as flooding of unknown unicast is 1775 administratively allowed), as it is not possible to select a single 1776 backup PE. 1778 14.1.2 All-Active Redundancy Mode 1780 For a given ES, if the remote PE has imported the set of Ethernet A-D 1781 per ES routes from one or more PEs and none of them have the "Single- 1782 Active" flag in the ESI Label Extended Community set, then the remote 1783 PE MUST deduce that the ES is operating in All-Active redundancy 1784 mode. A remote PE that receives a MAC advertisement route with non- 1785 reserved ESI SHOULD consider the advertised MAC address to be 1786 reachable via all PEs that have advertised reachability to that MAC 1787 address' EVI/ES via the combination of an Ethernet A-D per EVI route 1788 for that EVI/ES (and Ethernet Tag if applicable) AND an Ethernet A-D 1789 per ES route for that ES. The remote PE MUST use received MAC 1790 Advertisement routes and Ethernet A-D per EVI/per ES routes to 1791 construct the set of next-hops for the advertised MAC address. 1793 Each next-hop comprises an MPLS label stack that is to be used by the 1794 egress PE to forward the packet. This label stack is determined as 1795 follows: 1797 -If the next-hop is constructed as a result of a MAC route then this 1798 label stack MUST be used. However, if the MAC route doesn't exist for 1799 that PE, then the next-hop and MPLS label stack is constructed as a 1800 result of the Ethernet A-D routes. Note that the following 1801 description applies to determining the label stack for a particular 1802 next-hop to reach a given PE, from which the remote PE has received 1803 and imported Ethernet A-D routes that have the matching ESI and 1804 Ethernet Tag as the one present in the MAC advertisement. The 1805 Ethernet A-D routes mentioned in the following description refer to 1806 the ones imported from this given PE. 1808 -If a set of Ethernet A-D per ES routes for that ES AND an Ethernet 1809 A-D route per EVI exist, only then the label from that latter route 1810 must be used. 1812 The following example explains the above. 1814 Consider a CE (CE1) that is dual-homed to two PEs (PE1 and PE2) on a 1815 LAG interface (ES1), and is sending packets with source MAC address 1816 MAC1 on VLAN1 (mapped to EVI1). A remote PE, say PE3, is able to 1817 learn that MAC1 is reachable via PE1 and PE2. Both PE1 and PE2 may 1818 advertise MAC1 in BGP if they receive packets with MAC1 from CE1. If 1819 this is not the case, and if MAC1 is advertised only by PE1, PE3 1820 still considers MAC1 as reachable via both PE1 and PE2 as both PE1 1821 and PE2 advertise a set of Ethernet A-D per ES routes for ES1 as well 1822 as an Ethernet A-D per EVI route for . 1824 The MPLS label stack to send the packets to PE1 is the MPLS LSP stack 1825 to get to PE1 and the EVPN label advertised by PE1 for CE1's MAC. 1827 The MPLS label stack to send packets to PE2 is the MPLS LSP stack to 1828 get to PE2 and the MPLS label in the Ethernet A-D route advertised by 1829 PE2 for , if PE2 has not advertised MAC1 in BGP. 1831 We will refer to these label stacks as MPLS next-hops. 1833 The remote PE (PE3) can now load balance the traffic it receives from 1834 its CEs, destined for CE1, between PE1 and PE2. PE3 may use N-Tuple 1835 flow information to hash traffic into one of the MPLS next-hops for 1836 load balancing of IP traffic. Alternatively PE3 may rely on the 1837 source MAC addresses for load balancing. 1839 Note that once PE3 decides to send a particular packet to PE1 or PE2 1840 it can pick one out of multiple possible paths to reach the 1841 particular remote PE using regular MPLS procedures. For instance, if 1842 the tunneling technology is based on RSVP-TE LSPs, and PE3 decides to 1843 send a particular packet to PE1, then PE3 can choose from multiple 1844 RSVP-TE LSPs that have PE1 as their destination. 1846 When PE1 or PE2 receive the packet destined for CE1 from PE3, if the 1847 packet is a known unicast, it is forwarded to CE1. If it is a BUM 1848 packet then only one of PE1 or PE2 must forward the packet to the CE. 1849 Which of PE1 or PE2 forward this packet to the CE is determined based 1850 on which of the two is the DF. 1852 14.2. Load balancing of traffic between a PE and a local CE 1854 A CE may be configured with more than one interface connected to 1855 different PEs or the same PE for load balancing, using a technology 1856 such as LAG. The PE(s) and the CE can load balance traffic onto these 1857 interfaces using one of the following mechanisms. 1859 14.2.1. Data plane learning 1861 Consider that the PEs perform data plane learning for local MAC 1862 addresses learned from local CEs. This enables the PE(s) to learn a 1863 particular MAC address and associate it with one or more interfaces, 1864 if the technology between the PE and the CE supports multi-pathing. 1865 The PEs can now load balance traffic destined to that MAC address on 1866 the multiple interfaces. 1868 Whether the CE can load balance traffic that it generates on the 1869 multiple interfaces is dependent on the CE implementation. 1871 14.2.2. Control plane learning 1873 The CE can be a host that advertises the same MAC address using a 1874 control protocol on all interfaces. This enables the PE(s) to learn 1875 the host's MAC address and associate it with all interfaces. The PEs 1876 can now load balance traffic destined to the host on all these 1877 interfaces. The host can also load balance the traffic it generates 1878 onto these interfaces and the PE that receives the traffic employs 1879 EVPN forwarding procedures to forward the traffic. 1881 15. MAC Mobility 1883 It is possible for a given host or end-station (as defined by its MAC 1884 address) to move from one Ethernet segment to another; this is 1885 referred to as 'MAC Mobility' or 'MAC move' and it is different from 1886 the multi-homing situation in which a given MAC address is reachable 1887 via multiple PEs for the same Ethernet segment. In a MAC move, there 1888 would be two sets of MAC Advertisement routes, one set with the new 1889 Ethernet segment and one set with the previous Ethernet segment, and 1890 the MAC address would appear to be reachable via each of these 1891 segments. 1893 In order to allow all of the PEs in the EVPN instance to correctly 1894 determine the current location of the MAC address, all advertisements 1895 of it being reachable via the previous Ethernet segment MUST be 1896 withdrawn by the PEs, for the previous Ethernet segment, that had 1897 advertised it. 1899 If local learning is performed using the data plane, these PEs will 1900 not be able to detect that the MAC address has moved to another 1901 Ethernet segment and the receipt of MAC Advertisement routes, with 1902 the MAC Mobility extended community attribute, from other PEs serves 1903 as the trigger for these PEs to withdraw their advertisements. If 1904 local learning is performed using the control or management planes, 1905 these interactions serve as the trigger for these PEs to withdraw 1906 their advertisements. 1908 In a situation where there are multiple moves of a given MAC, 1909 possibly between the same two Ethernet segments, there may be 1910 multiple withdrawals and re-advertisements. In order to ensure that 1911 all PEs in the EVPN instance receive all of these correctly through 1912 the intervening BGP infrastructure, it is necessary to introduce a 1913 sequence number into the MAC Mobility extended community attribute. 1915 An implementation MUST handle the scenarios where the sequence number 1916 wraps around to process mobility event correctly. 1918 Every MAC mobility event for a given MAC address will contain a 1919 sequence number that is set using the following rules: 1921 - A PE advertising a MAC address for the first time advertises it 1922 with no MAC Mobility extended community attribute. 1924 - A PE detecting a locally attached MAC address for which it had 1925 previously received a MAC Advertisement route with a different 1926 Ethernet segment identifier advertises the MAC address in a MAC 1927 Advertisement route tagged with a MAC Mobility extended community 1928 attribute with a sequence number one greater than the sequence number 1929 in the MAC mobility attribute of the received MAC Advertisement 1930 route. In the case of the first mobility event for a given MAC 1931 address, where the received MAC Advertisement route does not carry a 1932 MAC Mobility attribute, the value of the sequence number in the 1933 received route is assumed to be 0 for purpose of this processing. 1935 - A PE detecting a locally attached MAC address for which it had 1936 previously received a MAC Advertisement route with the same non-zero 1937 Ethernet segment identifier advertises it with: 1938 i. no MAC Mobility extended community attribute, if the received 1939 route did not carry said attribute. 1941 ii. a MAC Mobility extended community attribute with the sequence 1942 number equal to the highest of the sequence number(s) in the 1943 received MAC Advertisement route(s), if the received route(s) is 1944 (are) tagged with a MAC Mobility extended community attribute. 1946 - A PE detecting a locally attached MAC address for which it had 1947 previously received a MAC Advertisement route with the same zero 1948 Ethernet segment identifier (single-homed scenarios) advertises it 1949 with MAC mobility extended community attribute with the sequence 1950 number set properly. In case of single-homed scenarios, there is no 1951 need for ESI comparison. The reason ESI comparison is done for multi- 1952 homing, is to prevent false detection of MAC move among the PEs 1953 attached to the same multi-homed site. 1955 A PE receiving a MAC Advertisement route for a MAC address with a 1956 different Ethernet segment identifier and a higher sequence number 1957 than that which it had previously advertised, withdraws its MAC 1958 Advertisement route. If two (or more) PEs advertise the same MAC 1959 address with same sequence number but different Ethernet segment 1960 identifiers, a PE that receives these routes selects the route 1961 advertised by the PE with lowest IP address as the best route. If the 1962 PE is the originator of the MAC route and it receives the same MAC 1963 address with the same sequence number that it generated, it will 1964 compare its own IP address with the IP address of the remote PE and 1965 will select the lowest IP. If its own route is not the best one, it 1966 will withdraw the route. 1968 15.1. MAC Duplication Issue 1970 A situation may arise where the same MAC address is learned by 1971 different PEs in the same VLAN because of two (or more hosts) being 1972 mis-configured with the same (duplicate) MAC address. In such 1973 situation, the traffic originating from these hosts would trigger 1974 continuous MAC moves among the PEs attached to these hosts. It is 1975 important to recognize such situation and avoid incrementing the 1976 sequence number (in the MAC Mobility attribute) to infinity. In order 1977 to remedy such situation, a PE that detects a MAC mobility event by 1978 way of local learning starts an M-second timer (default value of M = 1979 180) and if it detects N MAC moves before the timer expires (default 1980 value for N = 5), it concludes that a duplicate MAC situation has 1981 occurred. The PE MUST alert the operator and stop sending and 1982 processing any BGP MAC Advertisement routes for that MAC address till 1983 a corrective action is taken by the operator. The values of M and N 1984 MUST be configurable to allow for flexibility in operator control. 1985 Note that the other PEs in the E-VPN instance will forward the 1986 traffic for the duplicate MAC address to one of the PEs advertising 1987 the duplicate MAC address. 1989 15.2. Sticky MAC addresses 1991 There are scenarios in which it is desired to configure some MAC 1992 addresses as static so that they are not subjected to MAC move. In 1993 such scenarios, these MAC addresses are advertised with MAC Mobility 1994 Extended Community where static flag is set to 1 and sequence number 1995 is set to zero. If a PE receives such advertisements and later learns 1996 the same MAC address(es) via local learning, then the PE MUST alert 1997 the operator. 1999 16. Multicast & Broadcast 2001 The PEs in a particular EVPN instance may use ingress replication or 2002 P2MP LSPs to send multicast traffic to other PEs. 2004 16.1. Ingress Replication 2006 The PEs may use ingress replication for flooding BUM traffic as 2007 described in section "Handling of Multi-Destination Traffic". A given 2008 broadcast packet must be sent to all the remote PEs. However a given 2009 multicast packet for a multicast flow may be sent to only a subset of 2010 the PEs. Specifically a given multicast flow may be sent to only 2011 those PEs that have receivers that are interested in the multicast 2012 flow. Determining which of the PEs have receivers for a given 2013 multicast flow is done using explicit tracking described below. 2015 16.2. P2MP LSPs 2017 A PE may use an "Inclusive" tree for sending an BUM packet. This 2018 terminology is borrowed from [VPLS-MCAST]. 2020 A variety of transport technologies may be used in the SP network. 2021 For inclusive P-Multicast trees, these transport technologies include 2022 point-to-multipoint LSPs created by RSVP-TE or mLDP. 2024 16.2.1. Inclusive Trees 2026 An Inclusive Tree allows the use of a single multicast distribution 2027 tree, referred to as an Inclusive P-Multicast tree, in the SP network 2028 to carry all the multicast traffic from a specified set of EVPN 2029 instances on a given PE. A particular P-Multicast tree can be set up 2030 to carry the traffic originated by sites belonging to a single EVPN 2031 instance, or to carry the traffic originated by sites belonging to 2032 several EVPN instances. The ability to carry the traffic of more than 2033 one EVPN instance on the same tree is termed 'Aggregation' and the 2034 tree is called an Aggregate Inclusive P-Multicast tree or Aggregate 2035 Inclusive tree for short. The Aggregate Inclusive tree needs to 2036 include every PE that is a member of any of the EVPN instances that 2037 are using the tree. This implies that a PE may receive BUM traffic 2038 even if it doesn't have any receivers that are interested in 2039 receiving that traffic. 2041 An Inclusive or Aggregate Inclusive tree as defined in this document 2042 is a P2MP tree. A P2MP tree is used to carry traffic only for EVPN 2043 CEs that are connected to the PE that is the root of the tree. 2045 The procedures for signaling an Inclusive tree are the same as those 2046 in [VPLS-MCAST] with the VPLS-AD route replaced with the Inclusive 2047 Multicast Ethernet Tag route. The P-Tunnel attribute [VPLS-MCAST] for 2048 an Inclusive tree is advertised with the Inclusive Multicast Ethernet 2049 Tag route as described in section "Handling of Multi-Destination 2050 Traffic". Note that for an Aggregate Inclusive tree, a PE can 2051 "aggregate" multiple EVPN instances on the same P2MP LSP using 2052 upstream labels. The procedures for aggregation are the same as those 2053 described in [VPLS-MCAST], with VPLS A-D routes replaced by EVPN 2054 Inclusive Multicast Ethernet Tag routes. 2056 17. Convergence 2058 This section describes failure recovery from different types of 2059 network failures. 2061 17.1. Transit Link and Node Failures between PEs 2063 The use of existing MPLS Fast-Reroute mechanisms can provide failure 2064 recovery in the order of 50ms, in the event of transit link and node 2065 failures in the infrastructure that connects the PEs. 2067 17.2. PE Failures 2069 Consider a host CE1 that is dual homed to PE1 and PE2. If PE1 fails, 2070 a remote PE, PE3, can discover this based on the failure of the BGP 2071 session. This failure detection can be in the sub-second range if 2072 BFD is used to detect BGP session failure. PE3 can update its 2073 forwarding state to start sending all traffic for CE1 to only PE2. 2075 17.3. PE to CE Network Failures 2077 If the connectivity between the multi-homed CE and one of the PEs 2078 that it is attached to, fails, the PE MUST withdraw the set of 2079 Ethernet A-D per ES routes that had been previously advertised for 2080 that ES. When the MAC entry on the PE ages out, the PE MUST withdraw 2081 the MAC address from BGP. Note that to aid convergence, the Ethernet 2082 A-D per EVI routes MAY be withdrawn before the MAC routes. This 2083 enables the remote PEs to remove the MPLS next-hop to this particular 2084 PE from the set of MPLS next-hops that can be used to forward traffic 2085 to the CE. 2087 When a Ethernet Tag is decommissioned on an Ethernet segment, then 2088 the PE MUST withdraw the Ethernet A-D per EVI route(s) announced for 2089 the that are impacted by the decommissioning. In 2090 addition, the PE MUST also withdraw the MAC advertisement routes that 2091 are impacted by the decommissioning. 2093 The Ethernet A-D per ES routes should be used by an implementation to 2094 optimize the withdrawal of MAC advertisement routes. When a PE 2095 receives a withdrawal of a particular Ethernet A-D route from a PE it 2096 SHOULD consider all the MAC advertisement routes, that are learned 2097 from the same ESI as in the Ethernet A-D route, from the advertising 2098 PE, as having been withdrawn. This optimizes the network convergence 2099 times in the event of PE to CE failures. 2101 18. Frame Ordering 2103 In a MAC address, if the value of the 1st nibble (bits 8 thorough 5) 2104 of the most significant byte of the destination MAC address (which 2105 follows the last MPLS label) happens to be 0x4 or 0x6, then the 2106 Ethernet frame can be misinterpreted as an IPv4 or IPv6 packet by 2107 intermediate P nodes performing ECMP based on deep packet inspection, 2108 thus resulting in load balancing packets belonging to the same flow 2109 on different ECMP paths and subjecting them to different delays. 2110 Therefore, packets belonging to the same flow can arrive at the 2111 destination out of order. This out of order delivery can happen 2112 during steady state in absence of any failures resulting in 2113 significant impact to the network operation. 2115 In order to avoid any such mis-ordering, the following rules are 2116 applied: 2118 - If a network uses deep packet inspection for its ECMP, then the 2119 control word SHOULD be used when sending EVPN encapsulated packets 2120 over a MP2P LSP. 2122 - If a network uses Entropy label [RFC6790], then the control word 2123 SHOULD NOT be used when sending EVPN encapsulated packet over a MP2P 2124 LSP. 2126 - When sending EVPN encapsulated packets over a P2MP LSP or P2P LSP, 2127 then the control world SHOULD NOT be used. 2129 The control word is defined as follows: 2131 0 1 2 3 2132 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 2133 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2134 |0 0 0 0| Reserved | Sequence Number | 2135 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2137 In the above diagram the first 4 bits MUST be set to 0. The rest of 2138 the first 16 bits are reserved for future use. They MUST be set to 0 2139 when transmitting, and MUST be ignored upon receipt. The next 16 bits 2140 provide a sequence number that MUST also be set to zero by default. 2142 19. Acknowledgements 2144 Special thanks to Yakov Rekhter for reviewing this draft several 2145 times and providing valuable comments and for his very engaging 2146 discussions on several topics of this draft that helped shape this 2147 document. We would also like to thank Pedro Marques, Kaushik Ghosh, 2148 Nischal Sheth, Robert Raszuk, Amit Shukla, and Nadeem Mohammed for 2149 discussions that helped shape this document. We would also like to 2150 thank Han Nguyen for his comments and support of this work. We would 2151 also like to thank Steve Kensil and Reshad Rahman for their reviews. 2152 We would like to thank Jorge Rabadan for his contribution to section 2153 5 of this draft. We like to thank Thomas Morin for his review of this 2154 draft and his contribution of section 8.6. Many thanks to Jakob Heitz 2155 for his help to improve several sections of this draft. 2157 We would also like to thank Clarence Filsfils, Dennis Cai, Quaizar 2158 Vohra, Kireeti Kompella, Apurva Mehta for their contributions to this 2159 document. 2161 Last but not least, special thanks to Giles Heron (our WG chair) for 2162 his detailed review of this document in preparation for WG LC and 2163 making many valuable suggestions. 2165 20. Security Considerations 2167 Security considerations discussed in [RFC4761] and [RFC4762] apply to 2168 this document for MAC learning in data-plane over an Attachment 2169 Circuit (AC) and for flooding of unknown unicast and ARP messages 2170 over the MPLS/IP core. Security considerations discussed in [RFC4364] 2171 apply to this document for MAC learning in control-plane over the 2172 MPLS/IP core. This section describes additional considerations. 2174 As mentioned in [RFC4761], there are two aspects to achieving data 2175 privacy and protecting against denial-of-service attacks in a VPN: 2177 securing the control plane and protecting the forwarding path. 2178 Compromise of the control plane could result in a PE sending customer 2179 data belonging to some EVPN to another EVPN, or black-holing EVPN 2180 customer data, or even sending it to an eavesdropper; none of which 2181 are acceptable from a data privacy point of view. In addition, 2182 compromise of the control plane could result in black-holing EVPN 2183 customer data and could provide opportunities for unauthorized EVPN 2184 data usage (e.g., exploiting traffic replication within a multicast 2185 tree to amplify a denial-of-service attack based on sending large 2186 amounts of traffic). 2188 The mechanisms in this document use BGP for the control plane. Hence, 2189 techniques such as in [RFC5925] help authenticate BGP messages, 2190 making it harder to spoof updates (which can be used to divert EVPN 2191 traffic to the wrong EVPN instance) or withdrawals (denial-of-service 2192 attacks). In the multi-AS methods (b) and (c), this also means 2193 protecting the inter-AS BGP sessions, between the ASBRs, the PEs, or 2194 the Route Reflectors. 2196 Note that [RFC5925] will not help in keeping MPLS labels private -- 2197 knowing the labels, one can eavesdrop on EVPN traffic. However, this 2198 requires access to the data path within an SP network, which is 2199 assumed to be composed of trusted nodes/links. 2201 One of the requirements for protecting the data plane is that the 2202 MPLS labels be accepted only from valid interfaces. For a PE, valid 2203 interfaces comprise links from other routers in the PE's own AS. For 2204 an ASBR, valid interfaces comprise links from other routers in the 2205 ASBR's own AS, and links from other ASBRs in ASes that have instances 2206 of a given EVPN. It is especially important in the case of multi-AS 2207 EVPN instances that one accept EVPN packets only from valid 2208 interfaces. 2210 It is also important to help limit malicious traffic into a network 2211 for an imposter MAC address. The mechanism described in section 15.1, 2212 shows how duplicate MAC addresses can be detected and continous false 2213 MAC mobility can be prevented. The mechanism described in section 2214 15.2, shows how MAC addresses can be pinned to a given Ethernet 2215 Segment, such that if they appear behind any other Ethernet Segments, 2216 the traffic for those MAC addresses be prevented from entering the 2217 EVPN network from the other Ethernet Segments. 2219 21. Co-authors 2221 In addition to the authors listed on the front page, the following 2222 individuals have also helped to shape this document: 2224 Keyur Patel 2225 Samer Salam 2226 Sami Boutros 2227 Cisco 2229 Yakov Rekhter 2230 Ravi Shekhar 2231 Juniper Networks 2233 Florin Balus 2234 Nuage Networks 2236 22. IANA Considerations 2238 This document defines a new NLRI, called "EVPN", to be carried in BGP 2239 using multiprotocol extensions. This NLRI uses the existing AFI of 2240 25 (L2VPN). IANA has assigned it a SAFI value of 70. 2242 23. References 2244 23.1 Normative References 2246 [RFC4364] "BGP/MPLS IP VPNs", Rosen, Rekhter, et. al., February 2006 2248 [RFC4761] Kompella, K. and Y. Rekhter, "Virtual Private LAN Service 2249 (VPLS) Using BGP for Auto-Discovery and Signaling", RFC 2250 4761, January 2007. 2252 [RFC4762] Lasserre, M. and V. Kompella, "Virtual Private LAN Service 2253 (VPLS) Using Label Distribution Protocol (LDP) Signaling", 2254 RFC 4762, January 2007. 2256 [RFC4271] Y. Rekhter et. al., "A Border Gateway Protocol 4 (BGP-4)", 2257 RFC 4271, January 2006 2259 [RFC4760] T. Bates et. al., "Multiprotocol Extensions for BGP-4", RFC 2260 4760, January 2007 2262 23.2 Informative References 2264 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2265 Requirement Levels", BCP 14, RFC 2119, March 1997. 2267 [RFC7209] A. Sajassi, R. Aggarwal et. al., "Requirements for Ethernet 2268 VPN", draft-ietf-l2vpn-evpn-req-04.txt, July 2013. 2270 [RFC7117] "Multicast in VPLS". R. Aggarwal et.al., draft-ietf-l2vpn- 2271 vpls-mcast-14.txt, July 2013. 2273 [RFC4684] P. Marques et. al., "Constrained Route Distribution for 2274 Border Gateway Protocol/MultiProtocol Label Switching 2275 (BGP/MPLS) Internet Protocol (IP) Virtual Private Networks 2276 (VPNs)", RFC 4684, November 2006. 2278 [RFC6790] K. Kompella et. al, "The Use of Entropy Labels in MPLS 2279 Forwarding", RFC 6790, November 2012. 2281 24. Author's Address 2283 Ali Sajassi 2284 Cisco 2285 Email: sajassi@cisco.com 2287 Rahul Aggarwal 2288 Email: raggarwa_1@yahoo.com 2290 Nabil Bitar 2291 Verizon Communications 2292 Email : nabil.n.bitar@verizon.com 2294 Aldrin Isaac 2295 Bloomberg 2296 Email: aisaac71@bloomberg.net 2298 James Uttaro 2299 AT&T 2300 Email: uttaro@att.com 2302 John Drake 2303 Juniper Networks 2304 Email: jdrake@juniper.net 2306 Wim Henderickx 2307 Alcatel-Lucent 2308 e-mail: wim.henderickx@alcatel-lucent.com