L2VPN Workgroup                                              Ali Sajassi
INTERNET-DRAFT                                               Samer Salam
Intended Status: Standards Track                            Sami Boutros
                                                                   Cisco
Wim Henderickx
Jorge Rabadan                                                 Jim Uttaro
Alcatel-Lucent                                                      AT&T

John Drake                                                  Aldrin Isaac
Wen Lin                                                        Bloomberg
Juniper

Expires: December 18, January 6, 2015                                 June 18,                                    July 6, 2015

                   E-TREE Support in EVPN & PBB-EVPN
                     draft-ietf-bess-evpn-etree-01
                     draft-ietf-bess-evpn-etree-02

Abstract

   The Metro Ethernet Forum (MEF) has defined a rooted-multipoint
   Ethernet service known as Ethernet Tree (E-Tree).  [ETREE-FMWK]
   proposes a solution framework for supporting this service in MPLS
   networks. This document discusses how those functional requirements
   can be easily met with (PBB-)EVPN and how (PBB-)EVPN offers a more
   efficient implementation of these functions.

Status of this Memo

   This Internet-Draft is submitted to IETF in full conformance with the
   provisions of BCP 78 and BCP 79.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups.  Note that
   other groups may also distribute working documents as
   Internet-Drafts.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress."

   The list of current Internet-Drafts can be accessed at
   http://www.ietf.org/1id-abstracts.html

   The list of Internet-Draft Shadow Directories can be accessed at
   http://www.ietf.org/shadow.html

Copyright and License Notice

   Copyright (c) 2013 IETF Trust and the persons identified as the
   document authors. All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (http://trustee.ietf.org/license-info) in effect on the date of
   publication of this document. Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document. Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents

   1  Introduction  . . . . . . . . . . . . . . . . . . . . . . . . .  4
     1.1  Terminology . . . . . . . . . . . . . . . . . . . . . . . .  4
   2  E-Tree Scenarios and EVPN / PBB-EVPN Support  . . . . . . . . .  4
     2.1 Scenario 1: Leaf OR Root site(s) per PE  . . . . . . . . . .  4
     2.2 Scenario 2: Leaf AND OR Root site(s) per PE AC  . . . . . . . . . .  5
     2.3 Scenario 3: Leaf AND OR Root site(s) per Ethernet Segment MAC . . .  5 . . . . . . .  6
   3 Operation for EVPN . . . . . . . . . . . . . . . . . . . . . . .  6  7
     3.1 Known Unicast Traffic  . . . . . . . . . . . . . . . . . . .  7
     3.2 BUM Traffic  . . . . . . . . . . . . . . . . . . . . . . . .  7  8
       3.2.1 BUM Traffic supported by P2MP Tunnels traffic originated from a single-homed site on a
             leaf AC  . . . . . . . . .  7
       3.2.1 . . . . . . . . . . . . . . .  9
       3.2.2 BUM Traffic supported by Ingress Replication traffic originated from a single-homed site on a
             root AC  . . . . . . . . . . . . . . . . . . . . . . . .  9
       3.2.3 BUM traffic originated from a multi-homed site on a
             leaf AC  . . . . . . . . . . . . . . . . . . . . . . . .  9
       3.2.4 BUM traffic originated from a multi-homed site on a
             root AC  . . . . . . . . . . . . . . . . . . . . . . . . 10
     3.3 E-TREE Traffic Flows for EVPN  . . . . . . . . . . . . . . . 10
       3.3.1 E-Tree with MAC Learning . . . . . . . . . . . . . . . . 11 10
       3.3.2 E-Tree without MAC Learning  . . . . . . . . . . . . . . 11
   4 Operation for PBB-EVPN . . . . . . . . . . . . . . . . . . . . . 12 11
     4.1 Known Unicast Traffic  . . . . . . . . . . . . . . . . . . . 12
     4.2 BUM Traffic  . . . . . . . . . . . . . . . . . . . . . . . . 13 12
   5 BGP Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . 13
     5.1 Leaf ESI Label Extended Community  . . . . . . . . . . . . . 13
     5.2 E-TREE Extended Community  . . . . . . . . . . . . . . . . . 13

   6  Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . 14
   7  Security Considerations . . . . . . . . . . . . . . . . . . . . 14
   8  IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 14
   9  References  . . . . . . . . . . . . . . . . . . . . . . . . . . 14
     9.1  Normative References  . . . . . . . . . . . . . . . . . . . 14
     9.2  Informative References  . . . . . . . . . . . . . . . . . . 14
   Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 15

1  Introduction

   The Metro Ethernet Forum (MEF) has defined a rooted-multipoint
   Ethernet service known as Ethernet Tree (E-Tree). In an E-Tree
   service, endpoints are labeled as either Root or Leaf sites. Root
   sites can communicate with all other sites. Leaf sites can
   communicate with Root sites but not with other Leaf sites.

   [ETREE-FMWK] proposes the solution framework for supporting E-Tree
   service in MPLS networks. The document identifies the functional
   components of the overall solution to emulate E-Tree services in
   addition to Ethernet LAN (E-LAN) services on an existing MPLS
   network.

   [EVPN] is a solution for multipoint L2VPN services, with advanced
   multi-homing capabilities, using BGP for distributing customer/client
   MAC address reach-ability information over the MPLS/IP network. [PBB-
   EVPN] combines the functionality of EVPN with [802.1ah] Provider
   Backbone Bridging for MAC address scalability.

   This document discusses how the functional requirements for E-Tree
   service can be easily met with (PBB-)EVPN and how (PBB-)EVPN offers a
   more efficient implementation of these functions.

1.1  Terminology

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119 [KEYWORDS].

2  E-Tree Scenarios and EVPN / PBB-EVPN Support

   In this section, we will categorize support for E-Tree into three
   different scenarios, depending on the nature of the site association
   (Root/Leaf) per PE or per Ethernet Segment:

   - Leaf OR Root site(s) per PE

   - Leaf AND Root site(s) per PE

   - Leaf AND Root site(s) per Ethernet Segment

2.1 Scenario 1: Leaf OR Root site(s) per PE

   In this scenario, a PE may have receive traffic from either Root sites OR
   Leaf sites for a given
   VPN instance, MAC-VRF/bridge table, but not both
   concurrently. In other words, a given MAC-VRF/bridge table on a PE is
   either associated with a root or leaf. The PE may have both Root and
   Leaf sites albeit for different VPNs. Every Ethernet Segment
   connected to the PE is uniquely identified as either a Root or a Leaf
   site. MAC-VRFs/bridge tables.

                   +---------+            +---------+
                   |   PE1   |            |   PE2   |
    +---+          |  +---+  |  +------+  |  +---+  |            +---+
    |CE1+-----ES1----+--+
    |CE1+---ES1----+--+   |  |  | MPLS |  |  |   +--+----ES2-----+CE2|
    +---+  (Root)  |  | E |  |MAC|  |  |  /IP |  |  | E |  |MAC|  |   (Leaf)   +---+
                   |  | V |  |  |  |VRF|  |  |      | V  |  |VRF|  |
                   |  | I   |  |  |      |  |  | I   |  |            +---+
                   |  |   |  |  |      |  |  |   +--+----ES3-----+CE3|
                   |  +---+  |  +------+  |  +---+  |   (Leaf)   +---+
                   +---------+            +---------+

   Figure 1: Scenario 1

2.2 Scenario 2: Leaf AND Root site(s) per PE

   In this such scenario, a an EVPN PE may have a set of one or more implementation MAY provide topology
   constraint among the PEs belonging to the same EVI associated with an
   E-TREE service. The purpose of this topology constraint is to avoid
   having PEs with only  Leaf sites importing and processing BGP MAC
   routes from each other, thereby unnecessarily exhausting their RIB
   tables. To support such topology constrain in EVPN, two BGP Route-
   Targets (RTs) are used for every EVPN Instance (EVI): one RT is
   associated with the Root sites AND and the other is associated with the
   Leaf sites. On a
   set per EVI basis, every PE exports the single RT
   associated with its type of one or site(s). Furthermore, a PE with Root
   site(s) imports both Root and Leaf RTs, whereas a PE with Leaf
   site(s) only imports the Root RT. If the number of EVIs is very large
   (e.g., more than 32K or 64K), then RT type 0 as defined in [RFC4360]
   SHOULD be used; otherwise, RT type 2 is sufficient.

2.2 Scenario 2: Leaf OR Root site(s) per AC

   In this scenario, a PE may receive traffic from either Root OR Leaf
   sites for on a given VPN instance. Every
   Ethernet Segment connected to the PE Attachment Circuit (AC) of a MAC-VRF/bridge table.
   In other words, an AC (ES or ES/VLAN) is uniquely identified as either associated with a
   Root or a Leaf site. (but not both).

                     +---------+            +---------+
                     |   PE1   |            |   PE2   |
    +---+            |  +---+  |  +------+  |  +---+  |            +---+
    |CE1+-----ES1----+--+   |  |  |      |  |  |   +--+----ES2-----+CE2|   +--+---ES2/AC1--+CE2|
    +---+    (Leaf)  |  | E |  |MAC|  |  | MPLS |  |  | E |  |MAC|  |   (Leaf)   +---+
                     |  | V |  |VRF|  |  |  /IP |  |  | V |  |VRF|  |
                     |  | I   |  |  |      |  |  | I   |  |            +---+
                     |  |   |  |  |      |  |  |   +--+----ES3-----+CE3|   +--+---ES2/AC2--+CE3|
                     |  +---+  |  +------+  |  +---+  |   (Root)   +---+
                     +---------+            +---------+

   Figure 2: Scenario 2

2.3 Scenario 3: Leaf AND Root site(s) per Ethernet Segment

   In this scenario, a PE may have a set of one or more Root sites AND a
   set of one or more Leaf if there are PEs with only root (or leaf) sites for a given VPN instance. An Ethernet
   Segment connected to per
   EVI, then the RT constrain procedures described in section 2.1 can
   also be used here. However, when a Root site is added to a Leaf PE
   (or vise versa), then that PE needs to process MAC routes from all
   other Leaf PEs and add them to its forwarding table. If for a given
   EVI, the PEs can eventually have both Leaf and Root sites attached,
   even though they may be identified start as Root-only or Leaf-only PEs, then it is
   recommended to use a single RT per EVI and avoid additional
   configuration and operational overhead.

2.3 Scenario 3: Leaf OR Root site(s) per MAC

   In this scenario, a PE may receive traffic from both Root AND Leaf
   sites on a given Attachment Circuit (AC) of a MAC-VRF/bridge table.
   Since an Attachment Circuit (ES or ES/VLAN) carries traffic from both
   Root and a Leaf site concurrently. sites, the granularity at which Root or Leaf sites are
   identifies is on a per MAC address. This scenario is considered in
   this draft for EVPN service with only known unicast traffic - i.e.,
   there is no BUM traffic.

                     +---------+            +---------+
                     |   PE1   |            |   PE2   |
    +---+            |  +---+  |  +------+  |  +---+  |            +---+
    |CE1+-----ES1----+--+   |  |  |      |  |  |   +--+----ES2-----+CE2|   +--+---ES2/AC1--+CE2|
    +---+ (Leaf/Root)|    (Root)  |  | E |  |  | MPLS |  |  | E |  | (Leaf/Root)+---+
                     |  | V |  |  |  /IP |  |  | V |  |
                     |  | I |  |  |      |  |  | I |  |            +---+
                     |  |   |  |  |      |  |  |   +--+----ES3-----+CE3|   +--+---ES2/AC2--+CE3|
                     |  +---+  |  +------+  |  +---+  |   (Leaf)   +---+
                     +---------+            +---------+

   Figure 3: Scenario 3

3 Operation for EVPN

   [EVPN] defines the notion of an Ethernet Segment which ESI MPLS label used for split-horizon
   filtering of BUM traffic at the egress PE. Such egress filtering
   capabilities can be readily
   used to identify a Root and/or Leaf site leveraged in provision of E-TREE services. services as seen
   shortly. In other words, [EVPN] has inherent capability to support E-TREE E-
   TREE services without defining any new BGP routes. It only requires a  minor
   modification to the existing procedures and routes but just defining a
   new BGP Extended Community for leaf indication as shown later in this
   document.

   In addition to the procedures below (which

3.1 Known Unicast Traffic

   Since in EVPN, MAC learning is a MUST requirement), an
   EVPN PE implementation MAY provide topology constraint among the PEs
   belonging to performed in control plane via
   advertisement of BGP routes, the same EVI associated with an filtering needed by E-TREE service. The
   purpose of this topology constraint is service
   for known unicast traffic can be performed at the ingress PE, thus
   providing very efficient filtering and avoiding sending known unicast
   traffic over MPLS/IP core to avoid having be filtered at the egress PE as done in
   traditional E-TREE solutions (e.g., E-TREE for VPLS).

   To provide such ingress filtering for known unicast traffic, a PE
   MUST indicate to other PEs with only
   Leaf what kind of sites (e.g., scenario 1 in section 2.1) importing and processing
   BGP (root or leaf) its MAC routes from each other, thereby unnecessarily exhausting
   their RIB tables. However, when a Root site
   addresses are associated with. This indication is added to a Leaf PE
   (e.g., scenario 2 and 3 in achieved by using
   one of the following mechanisms:

   1) For single-homing scenarios of sections 2.2 and 2.3), then that 2.3, the PE needs
   to process
   advertises the MAC routes addresses received from all other a Leaf PEs and add them to its
   forwarding table. To support site, with an
   Extended community indicating a leaf flag. The lack of such topology constrain in EVPN, two BGP
   Route-Targets (RTs) are used for every EVPN Instance (EVI): one RT flag
   indicates that the MAC address is associated with a root site.

   2) For multi-homing scenario of section 2.2, where an AC is either
   root or leaf (but not both), the Root sites and PE advertises leaf indication along
   with the other Ethernet A-D per EVI route. Since these routes are always
   advertised ahead of MAC advertisements route, there is associated no need to
   append leaf-indication flag with the
   Leaf sites. On a MAC advertisement routes. The
   leaf indication flag on Ethernet A-D per EVI basis, every PE exports route tells the single RT
   receiving PEs that all MAC addresses associated with its type of site(s). Furthermore, this <ESI, EVI>
   or <ESI, EVI/VLAN> are from a PE with Root
   site(s) imports both Root and Leaf RTs, whereas a PE with Leaf
   site(s) only imports leaf site. The lack of such leaf-
   indication flag tells the Root RT. receiving PEs that the MAC addresses are
   associated with a root site.

   If for a given EVI, the PEs can
   eventually have both Leaf and Root sites attached, even though they
   may start as  Root-only or Leaf-only PEs, then it leaf site is recommended multi-homed to
   use a single RT PE1 an PE2, and PE1 advertises the
   Ethernet A-D per EVI and avoid additional configuration and
   operational overhead. If corresponding to this leaf site with the number of EVIs is very large (e.g., more
   than 32K or 64K), leaf-
   indication flag but PE2 does not, then RT type 0 as defined in [RFC4360] SHOULD be
   used; otherwise, RT type 2 is sufficient.

   The following procedures are used consistently for all the scenarios
   highlighted in the previous section.

3.1 Known Unicast Traffic

   For known unicast traffic, the receiving PE must advertise a Leaf indication
   along with each MAC Advertisement route, to indicate that MUST notify
   the
   associated MAC address was learnt from a Leaf Attachment Circuit
   (AC). The lack operator of a Leaf indication, indicates such discrepancy and ignore the MAC address is
   learnt from a root AC. leaf-indication flag
   on PE1. In other words, in case of discrepancy, the multi-homing for
   that pair of PEs is assumed to be in default "root" mode for that
   <ESI, EVI> or <ESI, EVI/VLAN>.

   3) For multi-homing scenario of operation
   in section 2.3, where an EVPN AC is that all ACs are both root (can transmit and receive traffic
   to/from other ACs in an EVI) unless the AC
   or leaf (i.e., root or leaf indication is explicitly identified
   as a leaf.

   Tagging at the granularity of MAC addresses with a
   address), the PE advertises leaf indication when they are associated along with a leaf AC, enables remote PEs to perform each MAC
   advertisement route just as in (1). No leaf-indication flag SHALL be
   sent along with the Ethernet A-D per EVI route for this scenario.

   Tagging MAC addresses with a leaf indication (either directly via MAC
   advertisement route or indirectly via Ethernet A-D per EVI route)
   enables remote PEs to perform ingress filtering for known unicast
   traffic - i.e., on the ingress PE, the MAC destination address lookup
   yields, in addition to the forwarding adjacency, a flag which
   indicates whether the target MAC is associated with a Leaf site or
   not. The ingress PE cross-checks this flag with the status of the
   originating AC, and if both are Leafs, then the packet is not
   forwarded.

   The PE places all Leaf ACs of a given bridge domain in a single
   split-horizon group in order to prevent intra-PE forwarding among
   Leaf ACs. This split-horizon function applies to both known unicast
   and BUM traffic.

   To support the above ingress filtering functionality, a new E-TREE
   Extended Community with a Leaf indication flag is introduced [section
   5.1].
   5.2]. This new Extended Community is advertised with each EVPN either Ethernet
   A-D per EVI route or MAC/IP Advertisement route. route as described above.

3.2 BUM Traffic

   For BUM traffic, it is not possible to perform filtering on the
   ingress PE, as is the case with known unicast, because of the multi-
   destination nature of the traffic. As such, the solution relies on
   egress filtering. In order to apply the proper egress filtering,
   which varies based on whether a packet is sent from a Root or a Leaf
   AC, the MPLS-encapsulated frames MUST be tagged with an indication of
   whether they originated from a Root or a Leaf AC. This can be
   achieved in EVPN through the use of the ESI MPLS label. Therefore,
   the ESI MPLS label not only identifies the Ethernet segment of origin
   for a given frame, but also it identifies its type (e.g., Leaf or
   Root).

3.2.1 BUM Traffic supported by P2MP Tunnels

   For multi-homing use cases where BUM traffic uses P2MP LSP, the
   ingress PE adds an upstream-assigned ESI MPLS label to the frame per

   [RFC7432] procedures and sends it to all the intended ingress PE
   devices. Two ESI MPLS labels are used for each multi-homed Ethernet
   segment that has both Root and Leaf sites: one ESI MPLS label that
   only identifies the Ethernet segment of origin per [RFC7432] and
   another one that not only identifies the Ethernet segment of origin
   but also its type (which is Leaf). If an Ethernet segment has only
   Root sites, then the former ESI MPLS label is used and if an Ethernet
   segment has only Leaf sites, then the latter ESI MPLS label is used.

   It should be noted that the former ESI MPLS label implicitly
   identifies a Root Ethernet segment - i.e., an ESI MPLS label that is
   signaled without the new E-TREE Extended Community (defined in
   section [5.1]), is assumed to be of type Root. When advertising the
   ESI MPLS label for an Ethernet Segment that has Leaf sites, the PE
   MUST indicate that the corresponding ESI is of type Leaf. This is
   achieved by advertising the Ethernet A-D per ES route with with the
   ESI MPLS label Extended Community along with the new E-TREE Extended
   Community that has a Leaf indication flag.

   The egress PE can determine whether or not to forward a particular
   frame to the destination Ethernet Segment depending on the following
   rules:

   - If the ESI MPLS label indicates that the source Ethernet Segment is
   the same as destination Ethernet segment, then the frame is blocked
   according to the split-horizon rule in [RFC7432].

   - If the ESI MPLS label indicates that the source Ethernet Segment is
   not the same as destination Ethernet segment and it doesn't have any
   Leaf indication, then the frame is forwarded to the destination AC
   according to the split-horizon rule in [RFC7432].

   - If the ESI MPLS label indicates that the source Ethernet Segment is
   not the same as destination Ethernet segment but it has a Leaf
   indication, then the frame is blocked if the destination AC is of
   type Leaf and it is forwarded if the destination AC is of type Root.

   The ingress PE imposes the right ESI MPLS label depending on whether
   the Ethernet frame originated from the Root or Leaf site on that
   Ethernet Segment. The mechanism by which the PE identifies whether a
   given frame originated from a Root or Leaf site on the segment is
   based on the Ethernet Tag associated with the frame (e.g., whether
   the frame come from a leaf or a root AC). Other mechanisms of
   identification, beyond the Ethernet Tag, are outside the scope of
   this document. It should be noted that support for both Root and Leaf
   sites on a single Ethernet Segment requires that the PE performs the
   Ethernet Segment split-horizon check on a per Ethernet Tag basis. In
   the case where a multi-homed Ethernet Segment has only either Root or
   Leaf sites attached, then a single ESI MPL label is allocated and
   advertised.

   For single-homing use cases where BUM traffic uses P2MP LSP, the
   ingress PE adds a special ESI MPLS label to the frame if solution relies on
   egress filtering. In order to apply the frame proper egress filtering,
   which varies based on whether a packet is
   originated sent from a Leaf site. This special ESI MPLS label used for
   single-homing scenarios is not on a per ES basis but rather on a per
   PE basis - i.e., AC or a single ESI MPLS label is used for all single-homed
   segments on that PE. If
   root AC, the frame is MPLS-encapsulated frames MUST be tagged with an
   indication of whether they originated from a Root site,
   then the ingress PE does not add any ESI MPLS label per [RFC7432]
   procedures. The egress PE, when receiving this special ESI MPLS
   label, it blocks the frame if the destination Leaf AC or not. In other
   words, leaf/root indication for BUM traffic is of type Leaf and
   it forwards done at the frame if
   granularity of AC. This can be achieved in EVPN through the destination AC is use of type Root.

   When a PE wants to advertise this special ESI label to other PE
   devices, it advertises it using
   the ESI MPLS label Extended Community
   with label. Therefore, the Ethernet A-D per ES route. The ESI for MPLS label can be used to
   either identify the Ethernet A-D segment of origin per
   ES route, [RFC 7432] or it
   can be of type 3, 4, or 5.

3.2.1 BUM Traffic supported by Ingress Replication

   The procedures for supporting used to indicate that the packet is originated from a leaf
   site.

   BUM traffic using sent over a P2MP LSP or ingress replication,
   are similar may need to
   carry an upstream assigned or downstream assigned MPLS label
   (respectively) for the ones in purpose of egress filtering to indicate to the previous section.
   egress PEs whether this packet is originated from a root AC or a leaf
   AC.

   The main differences
   are that the difference between downstream and upstream assigned ESI MPLS
   label is that in case of downstream assigned and not all egress PE
   devices need to receive the ESI label just like ingress replication
   procedures defined in [RFC7432].

   For frames received from a multi-homed Ethernet segment, the ingress
   PE may or may not add an ESI MPLS label based on the following
   criteria:

   - If the frame is forwarded

   There are four scenarios to consider:

3.2.1 BUM traffic originated from a PE that participates in the same
   multi-homed Ethernet Segment and the frame is received single-homed site on a Root AC,
   then leaf AC

   In this scenario, the ingress PE adds a per-ES downstream-assigned special ESI MPLS label to the
   frame per [RFC7432].

   - If the frame is forwarded to a PE that participates in the same
   multi-homed Ethernet Segment and the frame is received on indicating a Leaf AC,
   then the ingress PE adds the a per-ES downstream-assigned site. This special ESI MPLS label indicating Leaf to the frame.

   - If the frame used for
   single-homing scenarios is forwarded to a PE that does not participate in the
   same multi-homed Ethernet Segment and the frame is received on a Leaf
   AC, then the ingress per ES basis but rather on a per
   PE adds basis - i.e., a per-PE downstream-assigned special single ESI MPLS label indicating Leaf to the frame. is used for all single-homed
   ES's on that PE. This special ESI MPLS label is per PE.

   - If the frame is forwarded advertised to a other
   PE that does not participate in the
   same multi-homed devices, using a new EVPN Extended Community called Leaf ESI MPLS
   label Extended Community (section 5.1) along with a set of Ethernet Segment and
   A-D per ES routes. The set of Ethernet A-D per ES routes may be
   needed if the frame is received number of Route Targets (RTs) that need to be sent
   exceed the limit on a Root
   AC, then the ingress PE does not add any single route per [RFC 7432]. The RT(s)
   represent EVIs with at least a leaf site in them.  The ESI MPLS label to for the frame
   Ethernet A-D per [RFC7432].

   For frames received from a ES route is set to zero indicating single-homed Ethernet segment, the ingress
   sites.

   When a PE may or may not add an receives this special ESI MPLS label based on in the following
   criteria:

   - If data path, it
   blocks the frame packet if the destination AC is received of type Leaf; otherwise,
   it forwards the packet.

3.2.2 BUM traffic originated from a single-homed site on a Root AC, then root AC

   In this scenario, the ingress PE does not add any ESI MPLS label to the frame.

   - If the frame is received and
   it operates per [RFC7432] procedures.

3.2.3 BUM traffic originated from a multi-homed site on a Leaf AC, then leaf AC

   In this scenario, the ingress PE adds a
   special downstream-assigned an ESI MPLS label to the frame
   indicating both the Ethernet Segment of origin and its Leaf type. The
   reason Ethernet Segment of origin needs to the
   frame.

   Just as described be identified in the previous section, the addition
   to Leaf indication type, is
   signaled using to accommodate multi-homing scenarios for Integrated
   Routing and Bridging (IRB) where a source (Leaf) can be on one VLAN
   and the new E-TREE extended community defined in section
   [5.1] along with receivers (roots) can be on some other VLANs for the same
   Ethernet Segment.

   This ESI MPLS label extended community is advertised to other PE devices, using a new
   EVPN Extended Community called Leaf ESI Label Extended Community
   (section 5.1) along with the a set of Ethernet A-D per ES route.

   The egress PE can determine whether or not to forward a particular
   frame routes
   corresponding to the desitnation Ethernet Segment depending on ES of the following
   rules:

   - origin. If the ESI MPLS label indicates that the source Ethernet Segment egress ES is the same
   as destination Ethernet segment, the originated ES, then the frame is blocked
   according to receiving PE uses the split-horizon rule same procedure
   for filtering BUM traffic as the one specified in [RFC7432].

   - [RFC 7432]. If the
   egress ES is different from the originated ES, then the receiving PE
   uses the ESI MPLS label indicates to identify that the source Ethernet Segment BUM traffic is
   not the same as destination Ethernet segment associated
   with a leaf site and it doesn't have any
   Leaf indication, then thus blocking the frame is forwarded to BUM traffic if the destination
   AC
   according is also of type Leaf similar to section 3.2.1.

3.2.4 BUM traffic originated from a multi-homed site on a root AC

   In this scenario, both the split-horizon rule ingress and egress PE devices follows the
   procedure defined in [RFC7432].

   - If [RFC 7432] for adding and/or processing an ESI
   MPLS label.

   The ingress PE imposes the right ESI MPLS label indicates that depending on whether
   the source Ethernet Segment is
   not frame originated from the same as destination Ethernet segment but it has a Root or Leaf
   indication, then site on that
   Ethernet Segment. The mechanism by which the PE identifies whether a
   given frame is blocked if the destination AC is of
   type originated from a Root or Leaf and it is forwarded if site on the destination AC segment is of type Root.

   - If
   based on the ESI label is a special ESI MPLS label, then Ethernet Tag associated with the frame is
   blocked if (e.g., whether
   the destination frame come from a leaf or a root AC). Other mechanisms for
   identifying whether an egress AC is of type Leaf and it a root or leaf is forwarded if beyond the destination AC is
   scope of type Root. this document.

3.3 E-TREE Traffic Flows for EVPN

   Per [ETREE-FMWK], a generic E-Tree service supports all of the
   following traffic flows:

        - Ethernet Unicast from Root to Roots & Leaf
        - Ethernet Unicast from Leaf to Root
        - Ethernet Broadcast/Multicast from Root to Roots & Leafs
        - Ethernet Broadcast/Multicast from Leaf to Roots

   A particular E-Tree service may need to support all of the above
   types of flows or only a select subset, depending on the target
   application. In the case where unicast flows need not be supported,
   the L2VPN PEs can avoid performing any MAC learning function.

   In the subsections that follow, we will describe the operation of
   EVPN to support E-Tree service with and without MAC learning.

3.3.1 E-Tree with MAC Learning

   The PEs implementing an E-Tree service must perform MAC learning when
   unicast traffic flows must be supported from Root to Leaf or from
   Leaf to Root sites. In this case, the PE with Root sites performs MAC
   learning in the data-path over the Ethernet Segments, and advertises
   reachability in EVPN MAC Advertisement routes. These routes will be
   imported by PEs that have Leaf sites as well as by PEs that have Root
   sites, in a given EVI. Similarly, the PEs with Leaf sites perform MAC
   learning in the data-path over their Ethernet Segments, and advertise
   reachability in EVPN MAC Advertisement routes which are imported only
   by PEs with at least one Root site in the EVI. A PE with only Leaf
   sites will not import these routes. PEs with Root and/or Leaf sites
   may use the Ethernet A-D routes for aliasing (in the case of multi-
   homed segments) and for mass MAC withdrawal. withdrawal per [RFC 7432].

   To support multicast/broadcast from Root to Leaf sites, either a P2MP
   tree rooted at the PE(s) with the Root site(s) or ingress replication
   can be used. The multicast tunnels are set up through the exchange of
   the EVPN Inclusive Multicast route, as defined in [RFC7432].

   To support multicast/broadcast from Leaf to Root sites, ingress
   replication should be sufficient for most scenarios where there is a
   single Root or few Roots. If the number of Roots is large, a P2MP
   tree rooted at the PEs with Leaf sites may be used.

3.3.2 E-Tree without MAC Learning

   The PEs implementing an E-Tree service need not perform MAC learning
   when the traffic flows between Root and Leaf sites are multicast or
   broadcast. In this case, the PEs do not exchange EVPN MAC
   Advertisement routes. Instead, the Ethernet A-D routes are used to
   exchange the EVPN labels.

   The fields of the Ethernet A-D route are populated per the procedures
   defined in [RFC7432], and the route import rules are as described in
   previous sections.

4 Operation for PBB-EVPN

   In PBB-EVPN, the PE must advertise a Root/Leaf indication along with
   each MAC B-MAC Advertisement route, to indicate whether the associated B-
   MAC address corresponds to a Root or a Leaf site. Similar to the EVPN
   case, this flag will be added to the new E-TREE extended community Extended Community
   defined in section [5.1], [5.2], and advertised with each MAC Advertisement
   route.

   In the case where a multi-homed Ethernet Segment has both Root and
   Leaf sites attached, two B-MAC addresses are allocated and
   advertised: one B-MAC address implicitly denoting Root and the other
   explicitly denoting Leaf. The former B-MAC address is not advertised
   with the E-TREE extended community but the latter B-MAC denoting Leaf
   is advertised with the new E-TREE extended community.

   The ingress PE uses the right B-MAC source address depending on
   whether the Ethernet frame originated from the Root or Leaf site on
   that Ethernet Segment. The mechanism by which the PE identifies
   whether a given frame originated from a Root or Leaf site on the
   segment is based on the Ethernet Tag associated with the frame. Other
   mechanisms of identification, beyond the Ethernet Tag, are outside
   the scope of this document. It should be noted that support for both
   Root and Leaf sites on a single Ethernet Segment requires that the PE
   performs the Ethernet Segment split-horizon check on a per Ethernet
   Tag basis.

   In the case where a multi-homed Ethernet Segment has only Root or OR
   Leaf sites attached, then a single B-MAC address is allocated and
   advertised per segment.

   Furthermore, a PE advertises two special global B-MAC addresses: one
   for Root and another for Leaf, and tags them the Leaf one as such in the
   MAC Advertisement routes. route. These B-MAC addresses are used as source
   addresses for traffic originating from single-homed segments.

4.1 Known Unicast Traffic

   For known unicast traffic, the PEs perform ingress filtering: On the
   ingress PE, the C-MAC destination address lookup yields, in addition
   to the target B-MAC address and forwarding adjacency, a flag which
   indicates whether the target B-MAC is associated with a Root or a
   Leaf site. The ingress PE cross-checks this flag with the status of
   the originating site, and if both are a Leaf, then the packet is not
   forwarded.

   The PE places all Leaf Ethernet Segments of a given bridge domain in
   a single split-horizon group in order to prevent intra-PE forwarding
   among Leaf segments. This split-horizon function applies to BUM
   traffic as well. are a Leaf, then the packet is not
   forwarded.

4.2 BUM Traffic

   For BUM traffic, the PEs must perform egress filtering. When a PE
   receives a MAC advertisement route, it updates its Ethernet Segment
   egress filtering function (based on the B-MAC source address), as
   follows:

   - If the MAC Advertisement route indicates that the advertised B-MAC
   is a Leaf, and the local Ethernet Segment is a Leaf as well, then the
   source B-MAC address is added to the B-MAC filtering list.

   - Otherwise, the B-MAC filtering list is not updated.

   When the egress PE receives the packet, it examines the B-MAC source
   address to check whether it should filter or forward the frame. Note
   that this uses the same filtering logic as baseline [PBB-EVPN] and
   does not require any additional flags in the data-plane.

   The PE places all Leaf Ethernet Segments of a given bridge domain in
   a single split-horizon group in order to prevent intra-PE forwarding
   among Leaf segments. This split-horizon function applies to BUM
   traffic.

5 BGP Encoding

   This document defines one two new BGP Extended Community for EVPN.

5.1 Leaf ESI Label Extended Community

   This Extended Community is a new transitive Extended Community having
   a Type field value of 0x06 (EVPN) and the Sub-Type 0x04. In purpose,
   it is similar to ESI Label EC defined in [RFC 7432] with the only
   difference that it is used to indicate a leaf site in addition to the
   Ethernet segment of origin.

   It may be advertised along with Ethernet Auto-discovery routes, and
   it enables split-horizon procedures for multihomed sites as described
   in Section 3.2.1.3.  The Leaf ESI Label field represents an ES with a
   leaf site by the advertising PE, and it is used in split-horizon
   filtering by other PEs that are connected to the same multihomed
   Ethernet segment and egress filtering by other PEs that are connected
   to Leaf ACs.

   The E-TREE Extended Community is encoded as an 8-octet value as
   follows:

        0                   1                   2                   3
        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
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       | Type=0x06     | Sub-Type=0x04 |          Reserved=0           |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |  Reserved=0   |       Leaf ESI Label                          |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

5.2 E-TREE Extended Community

   A new EVPN BGP Extended Community called E-TREE is introduced here.
   This new extended community is a transitive extended community with
   the Type field of 0x06 (EVPN) and the Sub-Type of 0x04. 0x05. This extended
   community is used to for leaf indication and it is advertised with an
   EVPN MAC/IP route or an Ethernet A-D per ES EVI route. When advertised
   with an Ethernet A-D per ES route, it is sent along with ESI Label
   Extended Community defined in section 7.5 of [RFC7432].

   The E-TREE Extended Community is encoded as an 8-octet value as
   follows:

        0                   1                   2                   3
        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
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       | Type=0x06     | Sub-Type=0x04 Sub-Type=0x05 |        E-TREE Flags        Reserved=0             |
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
       |                         E-TREE Flags                         L|               Rserved=0       |        Reserved=0           |L|
       +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Leaf flag (L): A value of 1 indicates a leaf

6  Acknowledgement

   We would like to thank Dennis Cai and Antoni Przygienda for his their
   comments.

7  Security Considerations

   Same security considerations as [RFC7432].

8  IANA Considerations

   Allocation of Extended Community Type and Sub-Type for EVPN.

9  References

9.1  Normative References

   [KEYWORDS] Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC4360] S. Sangli et al, ""BGP Extended Communities Attribute",
              February, 2006.

   [RFC7432] Sajassi et al., "BGP MPLS Based Ethernet VPN", February,
              2015.

9.2  Informative References

   [ETREE-FMWK] Key et al., "A Framework for E-Tree Service over MPLS
   Network", draft-ietf-l2vpn-etree-frwk-03, work in progress, September
   2013.

   [PBB-EVPN] Sajassi et al., "PBB-EVPN", draft-ietf-l2vpn-pbb-evpn-
   05.txt, work in progress, October, 2013.

Authors' Addresses

   Ali Sajassi
   Cisco
   Email: sajassi@cisco.com

   Samer Salam
   Cisco
   Email: ssalam@cisco.com

   Wim Henderickx
   Alcatel-Lucent
   Email: wim.henderickx@alcatel-lucent.com

   Jim Uttaro
   AT&T
   Email: ju1738@att.com

   Aldrin
   Bloomberg Issac
   Email: aisaac71@bloomberg.net

   Sami Boutros
   Cisco
   Email: sboutros@cisco.com