BESS Z. Zhang Internet-Draft W. Lin Updates: 7432 (if approved) Juniper Networks, Inc. Intended status: Standards Track J. Rabadan Expires: January 7, 2016 Alcatel-Lucent K. Patel Cisco Systems July 6, 2015 Updates on EVPN BUM Procedures draft-zzhang-bess-evpn-bum-procedure-updates-00 Abstract This document specifies procedure updates for broadcast, unknown unicast, and multicast (BUM) traffic in Ethernet VPNs (EVPN), including selective multicast, and provider tunnel segmentation. Requirements Language 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 RFC2119. Status of This Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. 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." This Internet-Draft will expire on January 7, 2016. Copyright Notice Copyright (c) 2015 IETF Trust and the persons identified as the document authors. All rights reserved. Zhang, et al. Expires January 7, 2016 [Page 1] Internet-Draft evpn-bum-procedure-update July 2015 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. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 2 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1. Reasons for Tunnel Segmentation . . . . . . . . . . . . . 4 3. Additional Route Types of EVPN NLRI . . . . . . . . . . . . . 5 3.1. Per-Region I-PMSI A-D route . . . . . . . . . . . . . . . 6 3.2. S-PMSI A-D route . . . . . . . . . . . . . . . . . . . . 6 3.3. Leaf-AD route . . . . . . . . . . . . . . . . . . . . . . 7 4. Selective Multicast . . . . . . . . . . . . . . . . . . . . . 7 5. Inter-AS Segmentation . . . . . . . . . . . . . . . . . . . . 7 5.1. Changes to Section 7.2.2 of RFC 7117 . . . . . . . . . . 7 5.2. I-PMSI Leaf Tracking . . . . . . . . . . . . . . . . . . 8 5.3. Backward Compatibility . . . . . . . . . . . . . . . . . 9 6. Inter-Region Segmentation . . . . . . . . . . . . . . . . . . 10 6.1. Area vs. Region . . . . . . . . . . . . . . . . . . . . . 10 6.2. Per-region Aggregation . . . . . . . . . . . . . . . . . 11 6.3. Use of S-NH-EC . . . . . . . . . . . . . . . . . . . . . 12 6.4. Ingress PE's I-PMSI Leaf Tracking . . . . . . . . . . . . 12 7. Intra-region Segmentation and Assisted Ingress Replication . 13 7.1. Mix of inter-region and intra-region segmentation . . . . 14 8. Multi-homing Support . . . . . . . . . . . . . . . . . . . . 14 9. EVPN DCI . . . . . . . . . . . . . . . . . . . . . . . . . . 14 9.1. Non-GW Option . . . . . . . . . . . . . . . . . . . . . . 15 9.2. GW option . . . . . . . . . . . . . . . . . . . . . . . . 16 10. Security Considerations . . . . . . . . . . . . . . . . . . . 17 11. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . 17 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 17 12.1. Normative References . . . . . . . . . . . . . . . . . . 17 12.2. Informative References . . . . . . . . . . . . . . . . . 18 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 18 1. Terminology To be added Zhang, et al. Expires January 7, 2016 [Page 2] Internet-Draft evpn-bum-procedure-update July 2015 2. Introduction RFC 7432 specifies procedures to handle broadcast, unknown unicast, and multicast (BUM) traffic in Section 11, 12 and 16, using Inclusive Multicast Ethernet Tag Route. A lot of details are referred to RFC 7117 (VPLS Multicast). In particular, selective multicast is briefly mentioned for Ingress Replication but referred to RFC 7117. RFC 7117 specifies procedures for using both inclusive tunnels and selective tunnels, similar to MVPN procedures specified in RFC 6513 and RFC 6514. A new SAFI "MCAST-VPLS" is introduced, with two types of NLRIs that match MVPN's S-PMSI A-D routes and Leaf A-D routes. The same procedures can be applied to EVPN selective multicast for both Ingress Replication and other tunnel types, but new route types need to be defined under the same EVPN SAFI. MVPN uses terms I-PMSI and S-PMSI A-D Routes. For consistency and convenience, this document will use the same I/S-PMSI terms for VPLS and EVPN. In particular, EVPN's Inclusive Multicast Ethernet Tag Route and VPLS's VPLS A-D route carrying PTA (PMSI Tunnel Attribute) for BUM traffic purpose will all be referred to as I-PMSI A-D routes. Depending on the context, they may be used interchangeably. MVPN provider tunnels and EVPN/VPLS BUM provider tunnels, which are referred to as MVPN/EVPN/VPLS provider tunnels in this document for simplicity, can be segmented for technical or administrative reasons, which are summarized in Section 2.1 of this document. RFC 6513/6514 cover MVPN inter-as segmentation, RFC 7117 covers VPLS multicast inter-as segmentation, and RFC 7524 (Seamless MPLS Multicast) covers inter-area segmentation for both MVPN and VPLS. There is a difference between MVPN and VPLS multicast inter-as segmentation. For simplicity, EVPN uses the same procedures as in MVPN. All ASBRs can re-advertise their choice of the best route. Each can become the root of its intra-AS segment and inject traffic it receives from its upstream, while each downstream PE/ASBR will only pick one of the upstream ASBRs as its upstream. This is also the behavior even for VPLS in case of inter-area segmentation. For inter-area segmentation, RFC 7524 requires the use of Inter-area P2MP Segmented Next-Hop Extended Community (S-NH-EC), and the setting of "Leaf Information Required" (LIR) flag in PTA in certain situations. Either of these could be optional in case of EVPN. Removing these requirements would make the segmentation procedures transparent to ingress and egress PEs. RFC 7524 assumes that segmentation happens at area borders. However, it could be at "regional" borders, where a region could be a sub- Zhang, et al. Expires January 7, 2016 [Page 3] Internet-Draft evpn-bum-procedure-update July 2015 area, or even an entire AS plus its external links (Section 6). That would allow for more flexible deployment scenarios (e.g. for single- area provider networks). This document specifies/clarifies/redefines certain/additional EVPN BUM procedures, with a salient goal that they're better aligned among MVPN, EVPN and VPLS. For brevity, only changes/additions to relevant RFC 7117 and RFC 7524 procedures are specified, instead of repeating the entire procedures. Note that these are to be applied to EVPN only, even though sometimes they may sound to be updates to RFC 7117/7524. 2.1. Reasons for Tunnel Segmentation Tunnel segmentation may be required and/or desired because of administrative and/or technical reasons. For example, an MVPN/VPLS/EVPN network may span multiple providers and Inter-AS Option-B has to be used, in which the end-to-end provider tunnels have to be segmented at and stitched by the ASBRs. Different providers may use different tunnel technologies (e.g., provider A uses Ingress Replication, provider B uses RSVP-TE P2MP while provider C uses mLDP). Even if they use the same tunnel technology like RSVP-TE P2MP, it may be impractical to set up the tunnels across provider boundaries. The same situations may apply between the ASes and/or areas of a single provider. For example, the backbone area may use RSVP-TE P2MP tunnels while non-backbone areas may use mLDP tunnels. Segmentation can also be used to divide an AS/area to smaller regions, so that control plane state and/or forwarding plane state/ burden can be limited to that of individual regions. For example, instead of Ingress Replicating to 100 PEs in the entire AS, with inter-area segmentation [RFC 7524] a PE only needs to replicate to local PEs and ABRs. The ABRs will further replicate to their downstream PEs and ABRs. This not only reduces the forwarding plane burden, but also reduces the leaf tracking burden in the control plane. This inter-region segmentation can be further extended to intra-region as an alternative way to achieve Assisted Replication as proposed in [draft-rabadan-bess-evpn-optimized-ir], and it works for MPLS encapsulation. Smaller regions also have the benefit that, in case of tunnel aggregation, it is easier to find congruence among the segments of different constituent (service) tunnels and the resulting aggregation (base) tunnel in a region. This leads to better bandwidth efficiency, because the more congruent they are, the fewer leaves of Zhang, et al. Expires January 7, 2016 [Page 4] Internet-Draft evpn-bum-procedure-update July 2015 the base tunnel need to discard traffic when a service tunnel's segment does not need to receive the traffic (yet it is receiving the traffic due to aggregation). Another advantage of the smaller region is smaller BIER sub-domains. In this new multicast architecture BIER, packets carry a BitString, in which the bits correspond to edge routers that needs to receive traffic. Smaller sub-domains means smaller BitStrings can be used without having to send multiple copies of the same packet. Finally, EVPN tunnel segmentation can be used for EVPN DCIs, as discussed in Section 9. It follows the same concepts discussed above. 3. Additional Route Types of EVPN NLRI RFC 7432 defines the format of EVPN NLRI as the following: +-----------------------------------+ | Route Type (1 octet) | +-----------------------------------+ | Length (1 octet) | +-----------------------------------+ | Route Type specific (variable) | +-----------------------------------+ So far five types have been defined: + 1 - Ethernet Auto-Discovery (A-D) route + 2 - MAC/IP Advertisement route + 3 - Inclusive Multicast Ethernet Tag route + 4 - Ethernet Segment route + 5 - IP Prefix Route This document defines three additional route types: + 6 - Per-Region I-PMSI A-D route + 7 - S-PMSI A-D route + 8 - Leaf A-D route The "Route Type specific" field of the type 6 and type 7 EVPN NLRIs starts with a type 1 RD, whose Administrative sub-field MUST match that of the RD in all the EVPN routes from the same advertising router for a given EVI, except the Leaf A-D route (Section 3.3). Zhang, et al. Expires January 7, 2016 [Page 5] Internet-Draft evpn-bum-procedure-update July 2015 3.1. Per-Region I-PMSI A-D route The Per-region I-PMSI A-D route has the following format. Its usage is discussed in Section 6.2. +-----------------------------------+ | RD (8 octets) | +-----------------------------------+ | Ethernet Tag ID (4 octets) | +-----------------------------------+ | Source AS (4 octets) | +-----------------------------------+ Other than the addition of Ethernet Tag ID, it is identical to the Type 2 MVPN route defined in Section 4.2 of RFC 6514. the Source AS contains an Autonomous System Number (ASN). Two-octet ASNs are encoded in the two low-order octets of the Source AS field, with the two high-order octets set to zero. For now this is only defined at AS level. It may be defined for area or other levels in the future, if it becomes necessary. To allow for that possibility, a "type" field and a "region-id" field may be used instead of a simple "Source AS". This is pending community comments. 3.2. S-PMSI A-D route The S-PMSI A-D route has the following format: +-----------------------------------+ | RD (8 octets) | +-----------------------------------+ | Ethernet Tag ID (4 octets) | +-----------------------------------+ | Multicast Source Length (1 octet) | +-----------------------------------+ | Multicast Source (Variable) | +-----------------------------------+ | Multicast Group Length (1 octet) | +-----------------------------------+ | Multicast Group (Variable) | +-----------------------------------+ | Originating Router's IP Addr | +-----------------------------------+ Other than the addition of Ethernet Tag ID, it is identical to the S-PMSI A-D route as defined in RFC 7117. The procedures in RFC 7117 also apply (including wildcard functionality), except that the granularity level is per Ethernet Tag. Zhang, et al. Expires January 7, 2016 [Page 6] Internet-Draft evpn-bum-procedure-update July 2015 3.3. Leaf-AD route The Route Type specific field of a Leaf A-D route consists of the following: +-----------------------------------+ | Route Key (variable) | +-----------------------------------+ | Originating Router's IP Addr | +-----------------------------------+ A Leaf A-D route is originated in response to a PMSI route, which could be an Inclusive Multicast Tag route, a per-region I-PMSI A-D route, an S-PMSI A-D route, or some other types of routes that may be defined in the future that triggers Leaf A-D routes. The Route Key is the "Route Type Specific" field of the route for which this Leaf A-D route is generated. The general procedures of Leaf A-D route are first specified in RFC 6514 for MVPN. The principles apply to VPLS and EVPN as well. RFC 7117 has details for VPLS Multicast, and this document points out some specifics for EVPN, e.g. in Section 5. 4. Selective Multicast RFC 7117 specifies Selective Multicast for VPLS. Other than that different route types and formats are specified with EVPN SAFI for S-PMSI A-D and Leaf A-D routes (Section 3), all proceduers in RFC 7117 with respect to Selective Multicast apply to EVPN as well, including wildcard procedures. 5. Inter-AS Segmentation 5.1. Changes to Section 7.2.2 of RFC 7117 The first paragraph of Section 7.2.2.2 of RFC 7117 says: "... The best route procedures ensure that if multiple ASBRs, in an AS, receive the same Inter-AS A-D route from their EBGP neighbors, only one of these ASBRs propagates this route in Internal BGP (IBGP). This ASBR becomes the root of the intra-AS segment of the inter-AS tree and ensures that this is the only ASBR that accepts traffic into this AS from the inter-AS tree." The above VPLS behavior requires complicated VPLS specific procedures for the ASBRs to reach agreement. For EVPN, a different approach is be used and the above quoted text is not to be applied to EVPN. Zhang, et al. Expires January 7, 2016 [Page 7] Internet-Draft evpn-bum-procedure-update July 2015 The Leaf A-D based procedure is used for each ASBR who re-advertises into the AS to discover the leaves on the segment rooted at itself. This is the same as the procedures for S-PMSI in RFC 7117 itself. The following text at the end of the second bullet: "................................................... If, in order to instantiate the segment, the ASBR needs to know the leaves of the tree, then the ASBR obtains this information from the A-D routes received from other PEs/ASBRs in the ASBR's own AS." is changed to the following: "................................................... If, in order to instantiate the segment, the ASBR needs to know the leaves of the tree, then the ASBR MUST set the LIR flag to 1 in the PTA to trigger Leaf A-D routes from egress PEs and downstream ASBRs. It MUST be (auto-)configured with an import RT, which controls acceptance of leaf A-D routes by the ASBR." Accordingly, the following paragraph in Section 7.2.2.4: "If the received Inter-AS A-D route carries the PMSI Tunnel attribute with the Tunnel Identifier set to RSVP-TE P2MP LSP, then the ASBR that originated the route MUST establish an RSVP-TE P2MP LSP with the local PE/ASBR as a leaf. This LSP MAY have been established before the local PE/ASBR receives the route, or it MAY be established after the local PE receives the route." is changed to the following: "If the received Inter-AS A-D route has the LIR flag set in its PTA, then a receiving PE must originate a corresponding Leaf A-D route, and a receiving ASBR must originate a corresponding Leaf A-D route if and only if it received and imported one or more corresponding Leaf A-D routes from its downstream IBGP or EBGP peers, or it has non-null downstream forwarding state on the PIM/mLDP tunnel that instantiates the intra-AS segment. The ASBR that (re-)advertised of the Inter-AS A-D route then establishes a tunnel to the leaves discovered by the Leaf A-D routes." 5.2. I-PMSI Leaf Tracking An ingress PE does not set the LIR flag in its I-PMSI's PTA, even with Ingress Replication or RSVP-TE P2MP tunnels. It does not rely on the Leaf A-D routes to discover leaves in its AS, and Section 11.2 of RFC 7432 explicitly states that the LIR flag must be set to zero. Zhang, et al. Expires January 7, 2016 [Page 8] Internet-Draft evpn-bum-procedure-update July 2015 An implementation of RFC 7432 might have used the Originating Router's IP Address field of the Inclusive Multicast Ethernet Tag routes to determine the leaves, or might have used the Next Hop field instead. W/o segmentation, both will lead to the same result. With segmentation, an ingress PE MUST determine the leaves in its AS from the BGP next hops in all its received I-PMSI A-D routes, so it does not have to set the LIR bit set to request Leaf A-D routes. PEs within the same AS will all have different next hops in their I-PMSI A-D routes (hence will all be considered as leaves), and PEs from other ASes will have the next hop in their I-PMSI A-D routes set to addresses of ASBRs in this local AS, hence only those ASBRs will be considered as leaves (as proxies for those PEs in other ASes). Note that in case of Ingress Replication, when an ASBR re-advertises IBGP I-PMSI A-D routes, it MUST advertise the same label for all those for the same Ethernet Tag ID and the same EVI. When an ingress PE builds its flooding list, multiple routes may have the same (nexthop, label) tuple and they will only be added as a single branch in the flooding list. 5.3. Backward Compatibility The above procedures assume that all PEs are upgraded to support the segmentation procedures: o An ingress PE uses the Next Hop instead of Originating Router's IP Address to determine leaves for the I-PMSI tunnel. o An egress PE in a different AS sends Leaf A-D routes in response to I-PMSI routes, whose PTA has the LIR flag set by the re- advertising ASBRs. If a deployment has legacy PEs that does not support the above, then a legacy ingress PE would include all PEs (including those in remote ASes) as leaves of the inclusive tunnel and try to send traffic to them directly (no segmentation), which is either undesired or not possible; a legacy egress PE would not send Leaf A-D routes so the ASBRs would not know to send external traffic to them. To address this backward compatibility problem, the following procedure can be used (see Section 6.2 for per-PE/AS/region I-PMSI A-D routes): o An upgraded PE includes some indication in its per-PE I-PMSI A-D route that it supports the new procedures. o All per-PE I-PMSI A-D routes are restricted to the local AS and not propagated to external peers. Zhang, et al. Expires January 7, 2016 [Page 9] Internet-Draft evpn-bum-procedure-update July 2015 o The ASBRs in an AS originate per-region I-PMSI A-D routes and advertise to their external peers to advertise tunnels used to carry traffic from the local AS to other ASes. Depending on the types of tunnels being used, the LIR flag in the PTA may be set, in which case the downstream ASBRs and upgraded PEs will send Leaf A-D routes to pull traffic from their upstream ASBRs. One of the ASBRs in a downstream AS will be elected, based on the per-region I-PMSI A-D routes, to send traffic from the per-region I-PMSI tunnel to legacy PEs in the AS. o In an ingress AS, if and only if an ASBR has active downstream receivers (PEs and ASBRs), which are learned either explicitly via Leaf AD routes or implicitly via PIM join or mLDP label mapping, the ASBR originates a per-PE I-PMSI A-D route (i.e., regular Inclusive Multicast Ethernet Tag route) into the local AS, and stitches incoming per-PE I-PMSI tunnels into its per-region I-PMSI tunnel. With this, it gets traffic from local PEs and send to other ASes via the tunnel announced in its per-region I-PMSI A-D route. 6. Inter-Region Segmentation 6.1. Area vs. Region RFC 7524 is for MVPN/VPLS inter-area segmentation and does not explicitly cover EVPN. However, if "area" is replaced by "region" and "ABR" is replaced by "RBR" (Regional Border Router) then everything still works, and can be applied to EVPN as well. A region can be a sub-area, or can be an entire AS including its external links. Instead of automatic region definition based on IGP areas, a region would be defined as a BGP peer group. In fact, even with IGP area based region definition, a BGP peer group listing the PEs and ABRs in an area is still needed. Consider the following example diagram: --------- ------ --------- / \ / \ / \ / \ / \ / \ | PE1 o ASBR1 -- ASBR2 ASBR3 -- ASBR4 o PE2 | \ / \ / \ / \ / \ / \ / --------- ------ --------- AS 100 AS 200 AS 300 |-----------|--------|---------|--------|------------| segment1 segment2 segment3 segment4 segment5 Zhang, et al. Expires January 7, 2016 [Page 10] Internet-Draft evpn-bum-procedure-update July 2015 The inter-as segmentation procedures specified so far (RFC 6513/6514, 7117, and Section 5 of this document) requires all ASBRs to be involved, and Ingress Replication is used between two ASBRs in different ASes. In the above diagram, it's possible that ASBR1/4 does not support segmentation, and the provider tunnels in AS 100/300 can actually extend across the external link. In the case, the inter-region segmentation procedures can be used instead - a region is the entire (AS100 + ASBR1-ASBR2 link) or (AS300 + ASBR3-ASBR4 link). ASBR2/3 would be the RBRs, and ASBR1/4 will just be a transit core router with respect to provider tunnels. As illustrated in the diagram below, ASBR2/3 will establish a multihop EBGP session with either a RR or directly with PEs in the neighboring AS. I/S-PMSI A-D routes from ingress PEs will not be processed by ASBR1/4. When ASBR2 re-advertises the routes into AS 200, it changes the next hop to its own address and changes PTA to specify the tunnel type/identification in its own AS. When ASBR3 re- advertises I/S-PMSI A-D routes into the neighboring AS 300, it changes the next hop to its own address and changes PTA to specify the tunnel type/identification in the neighboring region 3. Now the segment is rooted at ASBR3 and extends across the external link to PEs. --------- ------ --------- / RR....\.mh-ebpg / \ mh-ebgp/....RR \ / : \ `. / \ .' / : \ | PE1 o ASBR1 -- ASBR2 ASBR3 -- ASBR4 o PE2 | \ / \ / \ / \ / \ / \ / --------- ------ --------- AS 100 AS 200 AS 300 |-------------------|----------|---------------------| segment 1 segment 2 segment 3 6.2. Per-region Aggregation Notice that every I/S-PMSI route from each PE will be propagated throughout all the ASes or regions. They may also trigger corresponding Leaf A-D routes depending on the types of tunnels used in each region. This may become too many - routes and corresponding tunnels. To address this concern, the I-PMSI routes from all PEs in a AS/region can be aggregated into a single I-PMSI route originated from the RBRs, and traffic from all those individual I-PMSI tunnels will be switched into the single I-PMSI tunnel. This is like the MVPN Inter-AS I-PMSI route originated by ASBRs. Zhang, et al. Expires January 7, 2016 [Page 11] Internet-Draft evpn-bum-procedure-update July 2015 The MVPN Inter-AS I-PMSI A-D route can be better called as per-AS I-PMSI A-D route, to be compared against the (per-PE) Intra-AS I-PMSI A-D routes originated by each PE. In this document we will call it as per-region I-PMSI A-D route, in case we want to apply the aggregation at regional level. The per-PE I-PMSI routes will not be propagated to other regions. If multiple RBRs are connected to a region, then each will advertise such a route, with the same route key (Section 3.1). Similar to the per-PE I-PMSI A-D routes, RBRs/PEs in a downstream region will each select a best one from all those re- advertised by the upstream RBRs, hence will only receive traffic injected by one of them. MVPN does not aggregate S-PMSI routes from all PEs in an AS like it does for I-PMSIs routes, because the number of PEs that will advertise S-PMSI routes for the same (s,g) or (*,g) is small. This is also the case for EVPN, i.e., there is no per-region S-PMSI routes. Notice that per-region I-PMSI routes can also be used to address backwards compatibility issue, as discussed in Section 5.3. 6.3. Use of S-NH-EC RFC 7524 specifies the use of S-NH-EC because it does not allow ABRs to change the BGP next hop when they re-advertise I/S-PMSI AD routes to downstream areas. That is only to be consistent with the MVPN Inter-AS I-PMSI A-D routes, whose next hop must not be changed when they're re-advertised by the segmenting ABRs for reasons specific to MVPN. For EVPN, it is perfectly fine to change the next hop when RBRs re-advertise the I/S-PMSI A-D routes, instead of relying on S- NH-EC. As a result, this document specifies that RBRs change the BGP next hop when they re-advertise I/S-PMSI A-D routes and do not use S- NH-EC. if a downstream PE/RBR needs to originate Leaf A-D routes, it simply uses the BGP next hop in the corresponding I/S-PMSI A-D routes to construct Route Targets. The advantage of this is that neither ingress nor egress PEs need to understand/use S-NH-EC, and consistent procedure (based on BGP next hop) is used for both inter-as and inter-region segmentation. 6.4. Ingress PE's I-PMSI Leaf Tracking RFC 7524 specifies that when an ingress PE/ASBR (re-)advertises an VPLS I-PMSI A-D route, it sets the LIR flag to 1 in the route's PTA. Similar to the inter-as case, this is actually not really needed for EVPN. To be consistent with the inter-as case, the ingress PE does not set the LIR flag in its originated I-PMSI A-D routes, and Zhang, et al. Expires January 7, 2016 [Page 12] Internet-Draft evpn-bum-procedure-update July 2015 determines the leaves based on the BGP next hops in its received I-PMSI A-D routes, as specified in Section 5.2. The same backward compatibility issue exists, and the same solution as in the inter-as case applies. 7. Intra-region Segmentation and Assisted Ingress Replication [draft-rabadan-bess-evpn-optimized-ir] describes "Assisted Ingress Replication", which reduces the burden of NVEs by having them replicate to only one of a few designated replicators, which will then replicate to other relevant NVEs. The tunnel segmentation procedures can be extended to achieve the same, even with the support for MPLS encapsulation. With inter-region segmentation, an RBR, which is a Route Reflector, changes the BGP Next Hop to one of its own addresses when it re- advertises an I/S-PMSI route to other regions, and sets the LIR bit in the PTA Flag field when necessary, but it does not do so when re- advertising to NVEs in its own region. If it does that even when re- advertising to local NVEs, then it becomes a replicator as in [draft- rabadan-bess-evpn-optimized-ir]: NVEs will respond with Leaf AD routes to individual I-PMSI routes from NVEs, but targeted to the re- advertising RBR of the selected best one (out of all those same routes re-advertised by different RBRs). so that the sending NVEs will only replicate to the RBRs, which will in turn replicate to NVEs. A RNVE (Regular, or legacy NVE) is not able to send back Leaf A-D routes so one of the RBRs MUST be designated to relay packets from NVEs to RNVEs. RNVEs will replicate traffic directly to all NVEs/ RNVEs. In case of MPLS encapsulation, NVEs advertise a label in their I-PMSI A-D routes even if they track the leaves via Leaf A-D routes by setting the LIR bit, and RBRs MUST not change that when re- advertise the routes (from NVEs or RNVEs). RNVEs can be identified by the lack of indication as discussed in Section 5.3 in their I-PMSI A-D routes. To address the possible concern with too many Leaf A-D routes (every NVE will respond with one to its selected RBR for each I-PMSI A-D route), a RBR can turn off or not set the LIR bit when it re- advertises the I-PMSI routes so that no Leaf A-D routes will be triggered from NVEs. It also originates a per-region I-PMSI A-D route (Section 6.2), but instead of into other regions, it is back into the same region. The route has the LIR bit set so that NVEs will respond with a Leaf A-D route. The per-region I-PMSI A-D routes from the RBRs and corresponding Leaf A-D routes from NVEs are comparable to the Replicator-AR and Leaf-AR routes with the Optimized Zhang, et al. Expires January 7, 2016 [Page 13] Internet-Draft evpn-bum-procedure-update July 2015 IR method (Selective Mode). This allows a RBR to determine the set of NVEs that it is responsible for relaying incoming traffic to. In case of MPLS encapsulation, for split-horizon purpose, NVEs supporting this functionality MUST set the LIR bit in their I-PMSI A-D routes to trigger corresponding Leaf A-D routes from RBRs, with different labels advertised in the Leaf A-D routes for different NVEs, so that RBRs know the source NVEs of incoming packets, and will not relay the traffic back to the source NVE. If the RBRs do not set the LIR bit in the re-advertised I-PMSI A-D routes, nor do they advertise per-region I-PMSI A-D routes, then this corresponds to the Non-Selective mode of Optimized IR mothod - A RBR MUST be chosen to replicate to all NVEs and RNVEs. An NVE could specify its own chosen RBR by including an IP Address based EC in its I-PMSI A-D route, and only send to specified RBR. In case of MPLS encapsulation, again for split-horizon purposes the NVEs MUST set the LIR bit so that the RBRs will advertise different labels for different NVEs and therefore set up its forwarding state to not relay a packet back to the sending NVE. 7.1. Mix of inter-region and intra-region segmentation Some more details may need to be spelled out when intra-region segmentation is used for IR optimization while in the mean time inter-region segmentation is used, with RNVEs present in different regions. 8. Multi-homing Support If multi-homing does not span across different ASes or regions, existing procedures work with segmenation. If an ES is multi-homed to PEs in different ASes or regions, additional procedures are needed to work with segmentation. The procedures are well understood but omitted here until the requirement becomes clear. 9. EVPN DCI In addition to inter-as/region segmentation uses cases, EVPN Overlay DC Interconnect is another important use case for EVPN tunnel segmentation. Section 5.1.1.1 and 5.1.1.2 of [draft-ietf-bess-evpn-overlay] discuss two options of interconnecting EVPN Overlay DCs. With the GW option, DC EVPNs and Interconnect EVPN (DCI) are independent and terminate at the GWs. With the non-GW option, DC EVPNs and Interconnect EVPN form an integral EVPN, just like EVPN inter-as option-B. The GW option is discussed in details in section 3.4 of [draft-ietf-bess-dci-evpn- overlay]. Zhang, et al. Expires January 7, 2016 [Page 14] Internet-Draft evpn-bum-procedure-update July 2015 The non-GW option can only be used when PEs can use VNI/VSID that has local significance (like mpls labels), and the GW option must be used otherwise. With the GW option, mac lookup must be performed when traffic comes from where non-local VNI/VSID are used. Otherwise, label/VNI/VSID switching can be used (typical inter-as option-B behavior). Note that with either option, BUM traffic forwarding can be based on tunnel stitching instead of mac lookup (except if IR is used together with non-local VNI/VSID), because BUM traffic goes to all PEs on corresponding provider tunnels instead of to targeted PEs. The following sections discusses some specific details for each option. 9.1. Non-GW Option The non-GW option can be easily compared to EVPN/mpls inter-region scenario where a region spans an entire AS - assuming that each DC is in its own AS that is different from the DCI's and other DCs'.. Consider the following diagram: +--------------+ +---------+ | | +---------+ +----+ | +----+ +----+ | +----+ |NVE1|--| |RBR1| |RBR3| |--|NVE3| +----+ | | | | | | +----+ | +----+ +----+ | | DC1 | WAN | DC2 | | +----+ +----+ | | |RBR3| |RBR4| | +----+ | | | | | | +----+ |NVE2|--| +----+ +----+ |--|NVE4| +----+ +---------+ | | +---------+ +----+ +--------------+ |---EVPN-Overlay----|---EVPN-MPLS---|----EVPN-Overlay---| Data Center Interconnect without Gateway The RBRs are WAN Edge routers. They re-advertise I/S-PMSI routes from one side to the other, following the previous described segmentation procedures. For example, the Inclusive Multicast Route from NVE1 is re-advertised into the WAN side by both RBR 1 and RBR2, with the LIR flag bit set in the PTA, and then re-advertised into DC2 by RBR3/4. NVE3/4 could both choose either the one re-advertised by RBR3 or by RBR4, or could each choose a different one (e.g., NVE3 Zhang, et al. Expires January 7, 2016 [Page 15] Internet-Draft evpn-bum-procedure-update July 2015 chooses the one re-advertised by RBR3 while NVE4 chooses the one re- advertised by RBR4). Each either joins the advertised PIM tunnel or send a corresponding Leaf A-D route to the re-advertiser of the chosen best route. RBR3 and/or RBR4 repeat the process, followed by RBR1 and/or RBR2 doing the same. At the end, a segmented tunnel is established to reach all NVE3/4. When BUM traffic arrives on RBR1/2 from NVE1 via the tunnel segment in DC1, the multicast VXLAN encapsulation is removed and the traffic is directly switched into the segment in the WAN w/o going through mac lookup. The per-region aggregation method (Section 6.2) can be used to limit the I-PMSI A-D routes to each DC. 9.2. GW option Consider the following diagram adapted from section 3.4 of [draft- ietf-bess-dci-evpn-overlay]: +--------------+ +---------+ | | +---------+ +----+ | +---+ +---+ | +----+ |NVE1|--| | | | | |--|NVE3| +----+ | |GW1| |GW3| | +----+ | +---+ +---+ | | DC1 | WAN | DC2 | | +---+ +---+ | | | | | | | +----+ | |GW2| |GW4| | +----+ |NVE2|--| +---+ +---+ |--|NVE4| +----+ +---------+ | | +---------+ +----+ +--------------+ |---EVPN-Overlay----|---EVPN-MPLS---|----EVPN-Overlay---| The GWs consumes EVPN routes from the DC side and re-originate new ones into the WAN side, and vice versa. All GWs will advertise their own I-PMSI A-D route to the DC and WAN side, but only the DF on an internal ESI (I-ESI) for the local DC will forward BUM traffic from one EVPN domain to the other. For example, BUM traffic from NVE1 will reach both GW1 and GW2, but only the DF, say GW1, will forward to the WAN side. The traffic will then reach both GW3 and GW4, but again only the DF (for the I-ESI for DC2, say GW4) will forward traffic into DC2. In [draft-ietf-bess-dci-evpn-overlay], the traffic forwarding by GWs is based on mac lookup - because of global significance of VNIs in Zhang, et al. Expires January 7, 2016 [Page 16] Internet-Draft evpn-bum-procedure-update July 2015 DCs, the VXLAN encapsulation cannot indicate to which remote NVE a known unicast packet should be forwarded to. However for BUM traffic, this is not a problem - a BUM packet only need to be put onto the appropriate tunnel. As a result, the DF GW on the I-ESI for a local DC can stitch all incoming BUM tunnels from local NVEs to its tunnel on the WAN side, and stitch all incoming BUM tunnels from remote GWs in the DCI into its tunnel on the DC side. This way, BUM traffic will be switched via label/VNI/VSID or multicast vxlan tunnel destination, bypassing mac lookup. Note that, this works only if Ingress Replication is not used for BUM traffic in an EVPN Overlay DC, because in that case the only way to distinguish BUM traffic from known uncast traffic is by checking mac address of the packets. Because the I-PMSI routes/tunnels are terminated in each DC/DCI, the I-PMSI routes originated by GWs are somewhat similar to the per- region I-PMSI routes discussed in the previous section. However, the per-region I-PMSI routes from RBRs in the same DC have the same route key and NVEs will only receive traffic from one of the RBRs based on best route selection, while the per-GW I-PMSI routes are distinct and all NVEs receive traffic from the same one of the GWs because only the DF on the I-ESI can forward traffic. 10. Security Considerations This document does not seem to introduce new security risks, though this may be revised after further review and scrutiny. 11. Acknowledgements The authors thank Eric Rosen, John Drake, and Ron Bonica for their comments and suggestions. 12. References 12.1. Normative References [I-D.ietf-bess-ir] Rosen, E., Subramanian, K., and J. Zhang, "Ingress Replication Tunnels in Multicast VPN", draft-ietf-bess- ir-00 (work in progress), January 2015. [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC7117] Aggarwal, R., Kamite, Y., Fang, L., Rekhter, Y., and C. Kodeboniya, "Multicast in Virtual Private LAN Service (VPLS)", RFC 7117, February 2014. Zhang, et al. Expires January 7, 2016 [Page 17] Internet-Draft evpn-bum-procedure-update July 2015 [RFC7432] Sajassi, A., Aggarwal, R., Bitar, N., Isaac, A., Uttaro, J., Drake, J., and W. Henderickx, "BGP MPLS-Based Ethernet VPN", RFC 7432, February 2015. [RFC7524] Rekhter, Y., Rosen, E., Aggarwal, R., Morin, T., Grosclaude, I., Leymann, N., and S. Saad, "Inter-Area Point-to-Multipoint (P2MP) Segmented Label Switched Paths (LSPs)", RFC 7524, May 2015. 12.2. Informative References [I-D.ietf-bess-dci-evpn-overlay] Rabadan, J., Sathappan, S., Henderickx, W., Palislamovic, S., Balus, F., Sajassi, A., and D. Cai, "Interconnect Solution for EVPN Overlay networks", draft-ietf-bess-dci- evpn-overlay-00 (work in progress), January 2015. [I-D.ietf-bess-evpn-overlay] Sajassi, A., Drake, J., Bitar, N., Isaac, A., Uttaro, J., and W. Henderickx, "A Network Virtualization Overlay Solution using EVPN", draft-ietf-bess-evpn-overlay-01 (work in progress), February 2015. [I-D.rabadan-bess-evpn-optimized-ir] Rabadan, J., Sathappan, S., Henderickx, W., Sajassi, A., and A. Isaac, "Optimized Ingress Replication solution for EVPN", draft-rabadan-bess-evpn-optimized-ir-00 (work in progress), October 2014. [I-D.wijnands-bier-architecture] Wijnands, I., Rosen, E., Dolganow, A., Przygienda, T., and S. Aldrin, "Multicast using Bit Index Explicit Replication", draft-wijnands-bier-architecture-05 (work in progress), March 2015. [RFC6513] Rosen, E. and R. Aggarwal, "Multicast in MPLS/BGP IP VPNs", RFC 6513, February 2012. [RFC6514] Aggarwal, R., Rosen, E., Morin, T., and Y. Rekhter, "BGP Encodings and Procedures for Multicast in MPLS/BGP IP VPNs", RFC 6514, February 2012. Authors' Addresses Zhaohui Zhang Juniper Networks, Inc. EMail: zzhang@juniper.net Zhang, et al. Expires January 7, 2016 [Page 18] Internet-Draft evpn-bum-procedure-update July 2015 Wen Lin Juniper Networks, Inc. EMail: wlin@juniper.net Jorge Rabadan Alcatel-Lucent EMail: jorge.rabadan@alcatel-lucent.com Keyur Patel Cisco Systems EMail: keyupate@cisco.com Zhang, et al. Expires January 7, 2016 [Page 19]