Internet Engineering Task Force (IETF)                         A. Farrel
Request for Comments: 9125                            Old Dog Consulting
Category: Standards Track                                       J. Drake
ISSN: 2070-1721                                                 E. Rosen
                                                        Juniper Networks
                                                                K. Patel
                                                            Arrcus, Inc.
                                                                L. Jalil
                                                                 Verizon
                                                             August 2021


Gateway Auto-Discovery and Route Advertisement for Site Interconnection
                         Using Segment Routing

Abstract

   Data centers are attached to the Internet or a backbone network by
   gateway routers.  One data center typically has more than one gateway
   for commercial, load-balancing, and resiliency reasons.  Other sites,
   such as access networks, also need to be connected across backbone
   networks through gateways.

Status of This Memo

   This is an Internet Standards Track document.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Further information on
   Internet Standards is available in Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc9125.

Copyright Notice

   Copyright (c) 2021 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
   (https://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
   2.  Requirements Language
   3.  Site Gateway Auto-Discovery
   4.  Relationship to BGP - Link State and Egress Peer Engineering
   5.  Advertising a Site Route Externally
   6.  Encapsulation
   7.  IANA Considerations
   8.  Security Considerations
   9.  Manageability Considerations
     9.1.  Relationship to Route Target Constraint
   10. References
     10.1.  Normative References
     10.2.  Informative References
   Acknowledgements

   Authors' Addresses

1.  Introduction

   Data centers (DCs) are critical components of the infrastructure used
   by network operators to provide services to their customers.  DCs
   (sites) are interconnected by a backbone network, which consists of
   any number of private networks and/or the Internet.  DCs are attached
   to the backbone network by routers that are gateways (GWs).  One DC
   typically has more than one GW for various reasons including
   commercial preferences, load balancing, or resiliency against
   connection or device failure.

   Segment Routing (SR) ([RFC8402]) is a protocol mechanism that can be
   used within a DC as well as for steering traffic that flows between
   two DC sites.  In order for a source site (also known as an ingress
   site) that uses SR to load-balance the flows it sends to a
   destination site (also known as an egress site), it needs to know the
   complete set of entry nodes (i.e., GWs) for that egress DC from the
   backbone network connecting the two DCs.  Note that it is assumed
   that the connected set of DC sites and the border nodes in the
   backbone network on the paths that connect the DC sites are part of
   the same SR BGP - Link State (LS) instance (see [RFC7752] and
   [RFC9086]) so that traffic engineering using SR may be used for these
   flows.

   Other sites, such as access networks, also need to be connected
   across backbone networks through gateways.  For illustrative
   purposes, consider the ingress and egress sites shown in Figure 1 as
   separate Autonomous Systems (ASes) (noting that the sites could be
   implemented as part of the ASes to which they are attached, or as
   separate ASes).  The various ASes that provide connectivity between
   the ingress and egress sites could each be constructed differently
   and use different technologies such as IP; MPLS using global table
   routing information from BGP; MPLS IP VPN; SR-MPLS IP VPN; or SRv6 IP
   VPN.  That is, the ingress and egress sites can be connected by
   tunnels across a variety of technologies.  This document describes
   how SR Segment Identifiers (SIDs) are used to identify the paths
   between the ingress and egress sites.

   The solution described in this document is agnostic as to whether the
   transit ASes do or do not have SR capabilities.  The solution uses SR
   to stitch together path segments between GWs and through the
   Autonomous System Border Routers (ASBRs).  Thus, there is a
   requirement that the GWs and ASBRs are SR capable.  The solution
   supports the SR path being extended into the ingress and egress sites
   if they are SR capable.

   The solution defined in this document can be seen in the broader
   context of site interconnection in [SR-INTERCONNECT].  That document
   shows how other existing protocol elements may be combined with the
   solution defined in this document to provide a full system, but it is
   not a necessary reference for understanding this document.

   Suppose that there are two gateways, GW1 and GW2 as shown in
   Figure 1, for a given egress site and that they each advertise a
   route to prefix X, which is located within the egress site with each
   setting itself as next hop.  One might think that the GWs for X could
   be inferred from the routes' next-hop fields, but typically it is not
   the case that both routes get distributed across the backbone: rather
   only the best route, as selected by BGP, is distributed.  This
   precludes load-balancing flows across both GWs.

           -----------------                    ---------------------
          | Ingress         |                  | Egress     ------   |
          | Site            |                  | Site      |Prefix|  |
          |                 |                  |           |   X  |  |
          |                 |                  |            ------   |
          |       --        |                  |   ---          ---  |
          |      |GW|       |                  |  |GW1|        |GW2| |
           -------++--------                    ----+-----------+-+--
                  | \                               |          /  |
                  |  \                              |         /   |
                  |  -+-------------        --------+--------+--  |
                  | ||ASBR|     ----|      |----  |ASBR| |ASBR| | |
                  | | ----     |ASBR+------+ASBR|  ----   ----  | |
                  | |           ----|      |----                | |
                  | |               |      |                    | |
                  | |           ----|      |----                | |
                  | | AS1      |ASBR+------+ASBR|           AS2 | |
                  | |           ----|      |----                | |
                  |  ---------------        --------------------  |
                --+-----------------------------------------------+--
               | |ASBR|                                       |ASBR| |
               |  ----               AS3                       ----  |
               |                                                     |
                -----------------------------------------------------

                   Figure 1: Example Site Interconnection

   The obvious solution to this problem is to use the BGP feature that
   allows the advertisement of multiple paths in BGP (known as Add-
   Paths) ([RFC7911]) to ensure that all routes to X get advertised by
   BGP.  However, even if this is done, the identity of the GWs will be
   lost as soon as the routes get distributed through an ASBR that will
   set itself to be the next hop.  And if there are multiple ASes in the
   backbone, not only will the next hop change several times, but the
   Add-Paths technique will experience scaling issues.  This all means
   that the Add-Paths approach is effectively limited to sites connected
   over a single AS.

   This document defines a solution that overcomes this limitation and
   works equally well with a backbone constructed from one or more ASes
   using the Tunnel Encapsulation attribute ([RFC9012]) as follows:

      When a GW to a given site advertises a route to a prefix X within
      that site, it will include a Tunnel Encapsulation attribute that
      contains the union of the Tunnel Encapsulation attributes
      advertised by each of the GWs to that site, including itself.

   In other words, each route advertised by a GW identifies all of the
   GWs to the same site (see Section 3 for a discussion of how GWs
   discover each other), i.e., the Tunnel Encapsulation attribute
   advertised by each GW contains multiple Tunnel TLVs, one or more from
   each active GW, and each Tunnel TLV will contain a Tunnel Egress
   Endpoint sub-TLV that identifies the GW for that Tunnel TLV.
   Therefore, even if only one of the routes is distributed to other
   ASes, it will not matter how many times the next hop changes, as the
   Tunnel Encapsulation attribute will remain unchanged.

   To put this in the context of Figure 1, GW1 and GW2 discover each
   other as gateways for the egress site.  Both GW1 and GW2 advertise
   themselves as having routes to prefix X.  Furthermore, GW1 includes a
   Tunnel Encapsulation attribute, which is the union of its Tunnel
   Encapsulation attribute and GW2's Tunnel Encapsulation attribute.
   Similarly, GW2 includes a Tunnel Encapsulation attribute, which is
   the union of its Tunnel Encapsulation attribute and GW1's Tunnel
   Encapsulation attribute.  The gateway in the ingress site can now see
   all possible paths to X in the egress site regardless of which route
   is propagated to it, and it can choose one or balance traffic flows
   as it sees fit.

2.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

3.  Site Gateway Auto-Discovery

   To allow a given site's GWs to auto-discover each other and to
   coordinate their operations, the following procedures are
   implemented:

   *  A route target ([RFC4360]) MUST be attached to each GW's auto-
      discovery route (defined below), and its value MUST be set to a
      value that indicates the site identifier.  The rules for
      constructing a route target are detailed in [RFC4360].  It is
      RECOMMENDED that a Type x00 or x02 route target be used.

   *  Site identifiers are set through configuration.  The site
      identifiers MUST be the same across all GWs to the site (i.e., the
      same identifier is used by all GWs to the same site) and MUST be
      unique across all sites that are connected (i.e., across all GWs
      to all sites that are interconnected).

   *  Each GW MUST construct an import filtering rule to import any
      route that carries a route target with the same site identifier
      that the GW itself uses.  This means that only these GWs will
      import those routes, and that all GWs to the same site will import
      each other's routes and will learn (auto-discover) the current set
      of active GWs for the site.

   The auto-discovery route that each GW advertises consists of the
   following:

   *  IPv4 or IPv6 Network Layer Reachability Information (NLRI)
      ([RFC4760]) containing one of the GW's loopback addresses (that
      is, with an AFI/SAFI pair that is one of the following: IPv4/NLRI
      used for unicast forwarding (1/1); IPv6/NLRI used for unicast
      forwarding (2/1); IPv4/NLRI with MPLS Labels (1/4); or IPv6/NLRI
      with MPLS Labels (2/4)).

   *  A Tunnel Encapsulation attribute ([RFC9012]) containing the GW's
      encapsulation information encoded in one or more Tunnel TLVs.

   To avoid the side effect of applying the Tunnel Encapsulation
   attribute to any packet that is addressed to the GW itself, the
   address advertised for auto-discovery MUST be a different loopback
   address than is advertised for packets directed to the gateway
   itself.

   As described in Section 1, each GW will include a Tunnel
   Encapsulation attribute with the GW encapsulation information for
   each of the site's active GWs (including itself) in every route
   advertised externally to that site.  As the current set of active GWs
   changes (due to the addition of a new GW or the failure/removal of an
   existing GW), each externally advertised route will be re-advertised
   with a new Tunnel Encapsulation attribute, which reflects the current
   set of active GWs.

   If a gateway becomes disconnected from the backbone network, or if
   the site operator decides to terminate the gateway's activity, it
   MUST withdraw the advertisements described above.  This means that
   remote gateways at other sites will stop seeing advertisements from
   or about this gateway.  Note that if the routing within a site is
   broken (for example, such that there is a route from one GW to
   another but not in the reverse direction), then it is possible that
   incoming traffic will be routed to the wrong GW to reach the
   destination prefix; in this degraded network situation, traffic may
   be dropped.

   Note that if a GW is (mis)configured with a different site identifier
   from the other GWs to the same site, then it will not be auto-
   discovered by the other GWs (and will not auto-discover the other
   GWs).  This would result in a GW for another site receiving only the
   Tunnel Encapsulation attribute included in the BGP best route, i.e.,
   the Tunnel Encapsulation attribute of the (mis)configured GW or that
   of the other GWs.

4.  Relationship to BGP - Link State and Egress Peer Engineering

   When a remote GW receives a route to a prefix X, it uses the Tunnel
   Egress Endpoint sub-TLVs in the containing Tunnel Encapsulation
   attribute to identify the GWs through which X can be reached.  It
   uses this information to compute SR Traffic Engineering (SR TE) paths
   across the backbone network looking at the information advertised to
   it in SR BGP - Link State (BGP-LS) ([RFC9085]) and correlated using
   the site identity.  SR Egress Peer Engineering (EPE) ([RFC9086]) can
   be used to supplement the information advertised in BGP-LS.

5.  Advertising a Site Route Externally

   When a packet destined for prefix X is sent on an SR TE path to a GW
   for the site containing X (that is, the packet is sent in the ingress
   site on an SR TE path that describes the whole path including those
   parts that are within the egress site), it needs to carry the
   receiving GW's SID for X such that this SID becomes the next SID that
   is due to be processed before the GW completes its processing of the
   packet.  To achieve this, each Tunnel TLV in the Tunnel Encapsulation
   attribute contains a Prefix-SID sub-TLV ([RFC9012]) for X.

   As defined in [RFC9012], the Prefix-SID sub-TLV is only for IPv4/IPV6
   Labeled Unicast routes, so the solution described in this document
   only applies to routes of those types.  If the use of the Prefix-SID
   sub-TLV for routes of other types is defined in the future, further
   documents will be needed to describe their use for site
   interconnection consistent with this document.

   Alternatively, if MPLS SR is in use and if the GWs for a given egress
   site are configured to allow GWs at remote ingress sites to perform
   SR TE through that egress site for a prefix X, then each GW to the
   egress site computes an SR TE path through the egress site to X and
   places each in an MPLS Label Stack sub-TLV ([RFC9012]) in the SR
   Tunnel TLV for that GW.

   Please refer to Section 7 of [SR-INTERCONNECT] for worked examples of
   how the SID stack is constructed in this case and how the
   advertisements would work.

6.  Encapsulation

   If a site is configured to allow remote GWs to send packets to the
   site in the site's native encapsulation, then each GW to the site
   will also include multiple instances of a Tunnel TLV for that native
   encapsulation in externally advertised routes: one for each GW.  Each
   Tunnel TLV contains a Tunnel Egress Endpoint sub-TLV with the address
   of the GW that the Tunnel TLV identifies.  A remote GW may then
   encapsulate a packet according to the rules defined via the sub-TLVs
   included in each of the Tunnel TLVs.

7.  IANA Considerations

   IANA maintains the "BGP Tunnel Encapsulation Attribute Tunnel Types"
   registry in the "Border Gateway Protocol (BGP) Tunnel Encapsulation"
   registry.

   IANA had previously assigned the value 17 from this subregistry for
   "SR Tunnel", referencing this document as an Internet-Draft.  At that
   time, the assignment policy for this range of the registry was "First
   Come First Served" [RFC8126].

   IANA has marked that assignment as deprecated.  IANA may reclaim that
   codepoint at such a time that the registry is depleted.

8.  Security Considerations

   From a protocol point of view, the mechanisms described in this
   document can leverage the security mechanisms already defined for
   BGP.  Further discussion of security considerations for BGP may be
   found in the BGP specification itself ([RFC4271]) and in the security
   analysis for BGP ([RFC4272]).  The original discussion of the use of
   the TCP MD5 signature option to protect BGP sessions is found in
   [RFC5925], while [RFC6952] includes an analysis of BGP keying and
   authentication issues.

   The mechanisms described in this document involve sharing routing or
   reachability information between sites, which may mean disclosing
   information that is normally contained within a site.  So it needs to
   be understood that normal security paradigms based on the boundaries
   of sites are weakened and interception of BGP messages may result in
   information being disclosed to third parties.  Discussion of these
   issues with respect to VPNs can be found in [RFC4364], while
   [RFC7926] describes many of the issues associated with the exchange
   of topology or TE information between sites.

   Particular exposures resulting from this work include:

   *  Gateways to a site will know about all other gateways to the same
      site.  This feature applies within a site, so it is not a
      substantial exposure, but it does mean that if the BGP exchanges
      within a site can be snooped or if a gateway can be subverted,
      then an attacker may learn the full set of gateways to a site.
      This would facilitate more effective attacks on that site.

   *  The existence of multiple gateways to a site becomes more visible
      across the backbone and even into remote sites.  This means that
      an attacker is able to prepare a more comprehensive attack than
      exists when only the locally attached backbone network (e.g., the
      AS that hosts the site) can see all of the gateways to a site.
      For example, a Denial-of-Service attack on a single GW is
      mitigated by the existence of other GWs, but if the attacker knows
      about all the gateways, then the whole set can be attacked at
      once.

   *  A node in a site that does not have external BGP peering (i.e., is
      not really a site gateway and cannot speak BGP into the backbone
      network) may be able to get itself advertised as a gateway by
      letting other genuine gateways discover it (by speaking BGP to
      them within the site), so it may get those genuine gateways to
      advertise it as a gateway into the backbone network.  This would
      allow the malicious node to attract traffic without having to have
      secure BGP peerings with out-of-site nodes.

   *  An external party intercepting BGP messages anywhere between sites
      may learn information about the functioning of the sites and the
      locations of endpoints.  While this is not necessarily a
      significant security or privacy risk, it is possible that the
      disclosure of this information could be used by an attacker.

   *  If it is possible to modify a BGP message within the backbone, it
      may be possible to spoof the existence of a gateway.  This could
      cause traffic to be attracted to a specific node and might result
      in traffic not being delivered.

   All of the issues in the list above could cause disruption to site
   interconnection, but they are not new protocol vulnerabilities so
   much as new exposures of information that SHOULD be protected against
   using existing protocol mechanisms such as securing the TCP sessions
   over which the BGP messages flow.  Furthermore, it is a general
   observation that if these attacks are possible, then it is highly
   likely that far more significant attacks can be made on the routing
   system.  It should be noted that BGP peerings are not discovered but
   always arise from explicit configuration.

   Given that the gateways and ASBRs are connected by tunnels that may
   run across parts of the network that are not trusted, data center
   operators using the approach set out in this network MUST consider
   using gateway-to-gateway encryption to protect the data center
   traffic.  Additionally, due consideration MUST be given to encrypting
   end-to-end traffic as it would be for any traffic that uses a public
   or untrusted network for transport.

9.  Manageability Considerations

   The principal configuration item added by this solution is the
   allocation of a site identifier.  The same identifier MUST be
   assigned to every GW to the same site, and each site MUST have a
   different identifier.  This requires coordination, probably through a
   central management agent.

   It should be noted that BGP peerings are not discovered but always
   arise from explicit configuration.  This is no different from any
   other BGP operation.

   The site identifiers that are configured and carried in route targets
   (see Section 3) are an important feature to ensure that all of the
   gateways to a site discover each other.  Therefore, it is important
   that this value is not misconfigured since that would result in the
   gateways not discovering each other and not advertising each other.

9.1.  Relationship to Route Target Constraint

   In order to limit the VPN routing information that is maintained at a
   given route reflector, [RFC4364] suggests that route reflectors use
   "Cooperative Route Filtering", which was renamed "Outbound Route
   Filtering" and defined in [RFC5291].  [RFC4684] defines an extension
   to that mechanism to include support for multiple autonomous systems
   and asymmetric VPN topologies such as hub-and-spoke.  The mechanism
   in RFC 4684 is known as Route Target Constraint (RTC).

   An operator would not normally configure RTC by default for any AFI/
   SAFI combination and would only enable it after careful
   consideration.  When using the mechanisms defined in this document,
   the operator should carefully consider the effects of filtering
   routes.  In some cases, this may be desirable, and in others, it
   could limit the effectiveness of the procedures.

10.  References

10.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC4271]  Rekhter, Y., Ed., Li, T., Ed., and S. Hares, Ed., "A
              Border Gateway Protocol 4 (BGP-4)", RFC 4271,
              DOI 10.17487/RFC4271, January 2006,
              <https://www.rfc-editor.org/info/rfc4271>.

   [RFC4360]  Sangli, S., Tappan, D., and Y. Rekhter, "BGP Extended
              Communities Attribute", RFC 4360, DOI 10.17487/RFC4360,
              February 2006, <https://www.rfc-editor.org/info/rfc4360>.

   [RFC4760]  Bates, T., Chandra, R., Katz, D., and Y. Rekhter,
              "Multiprotocol Extensions for BGP-4", RFC 4760,
              DOI 10.17487/RFC4760, January 2007,
              <https://www.rfc-editor.org/info/rfc4760>.

   [RFC5925]  Touch, J., Mankin, A., and R. Bonica, "The TCP
              Authentication Option", RFC 5925, DOI 10.17487/RFC5925,
              June 2010, <https://www.rfc-editor.org/info/rfc5925>.

   [RFC7752]  Gredler, H., Ed., Medved, J., Previdi, S., Farrel, A., and
              S. Ray, "North-Bound Distribution of Link-State and
              Traffic Engineering (TE) Information Using BGP", RFC 7752,
              DOI 10.17487/RFC7752, March 2016,
              <https://www.rfc-editor.org/info/rfc7752>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

   [RFC9012]  Patel, K., Van de Velde, G., Sangli, S., and J. Scudder,
              "The BGP Tunnel Encapsulation Attribute", RFC 9012,
              DOI 10.17487/RFC9012, April 2021,
              <https://www.rfc-editor.org/info/rfc9012>.

10.2.  Informative References

   [RFC4272]  Murphy, S., "BGP Security Vulnerabilities Analysis",
              RFC 4272, DOI 10.17487/RFC4272, January 2006,
              <https://www.rfc-editor.org/info/rfc4272>.

   [RFC4364]  Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private
              Networks (VPNs)", RFC 4364, DOI 10.17487/RFC4364, February
              2006, <https://www.rfc-editor.org/info/rfc4364>.

   [RFC4684]  Marques, P., Bonica, R., Fang, L., Martini, L., Raszuk,
              R., Patel, K., and J. Guichard, "Constrained Route
              Distribution for Border Gateway Protocol/MultiProtocol
              Label Switching (BGP/MPLS) Internet Protocol (IP) Virtual
              Private Networks (VPNs)", RFC 4684, DOI 10.17487/RFC4684,
              November 2006, <https://www.rfc-editor.org/info/rfc4684>.

   [RFC5291]  Chen, E. and Y. Rekhter, "Outbound Route Filtering
              Capability for BGP-4", RFC 5291, DOI 10.17487/RFC5291,
              August 2008, <https://www.rfc-editor.org/info/rfc5291>.

   [RFC6952]  Jethanandani, M., Patel, K., and L. Zheng, "Analysis of
              BGP, LDP, PCEP, and MSDP Issues According to the Keying
              and Authentication for Routing Protocols (KARP) Design
              Guide", RFC 6952, DOI 10.17487/RFC6952, May 2013,
              <https://www.rfc-editor.org/info/rfc6952>.

   [RFC7911]  Walton, D., Retana, A., Chen, E., and J. Scudder,
              "Advertisement of Multiple Paths in BGP", RFC 7911,
              DOI 10.17487/RFC7911, July 2016,
              <https://www.rfc-editor.org/info/rfc7911>.

   [RFC7926]  Farrel, A., Ed., Drake, J., Bitar, N., Swallow, G.,
              Ceccarelli, D., and X. Zhang, "Problem Statement and
              Architecture for Information Exchange between
              Interconnected Traffic-Engineered Networks", BCP 206,
              RFC 7926, DOI 10.17487/RFC7926, July 2016,
              <https://www.rfc-editor.org/info/rfc7926>.

   [RFC8126]  Cotton, M., Leiba, B., and T. Narten, "Guidelines for
              Writing an IANA Considerations Section in RFCs", BCP 26,
              RFC 8126, DOI 10.17487/RFC8126, June 2017,
              <https://www.rfc-editor.org/info/rfc8126>.

   [RFC8402]  Filsfils, C., Ed., Previdi, S., Ed., Ginsberg, L.,
              Decraene, B., Litkowski, S., and R. Shakir, "Segment
              Routing Architecture", RFC 8402, DOI 10.17487/RFC8402,
              July 2018, <https://www.rfc-editor.org/info/rfc8402>.

   [RFC9085]  Previdi, S., Talaulikar, K., Ed., Filsfils, C., Gredler,
              H., and M. Chen, "Border Gateway Protocol - Link State
              (BGP-LS) Extensions for Segment Routing", RFC 9085,
              DOI 10.17487/RFC9085, August 2021,
              <https://www.rfc-editor.org/info/rfc9085>.

   [RFC9086]  Previdi, S., Talaulikar, K., Ed., Filsfils, C., Patel, K.,
              Ray, S., and J. Dong, "Border Gateway Protocol - Link
              State (BGP-LS) Extensions for Segment Routing BGP Egress
              Peer Engineering", RFC 9086, DOI 10.17487/RFC9086, August
              2021, <https://www.rfc-editor.org/info/rfc9086>.

   [SR-INTERCONNECT]
              Farrel, A. and J. Drake, "Interconnection of Segment
              Routing Sites - Problem Statement and Solution Landscape",
              Work in Progress, Internet-Draft, draft-farrel-spring-sr-
              domain-interconnect-06, 19 May 2021,
              <https://datatracker.ietf.org/doc/html/draft-farrel-
              spring-sr-domain-interconnect-06>.

Acknowledgements

   Thanks to Bruno Rijsman, Stephane Litkowski, Boris Hassanov, Linda
   Dunbar, Ravi Singh, and Daniel Migault for review comments, and to
   Robert Raszuk for useful discussions.  Gyan Mishra provided a helpful
   GenArt review, and John Scudder and Benjamin Kaduk made helpful
   comments during IESG review.

Authors' Addresses

   Adrian Farrel
   Old Dog Consulting

   Email: adrian@olddog.co.uk


   John Drake
   Juniper Networks

   Email: jdrake@juniper.net


   Eric Rosen
   Juniper Networks

   Email: erosen52@gmail.com


   Keyur Patel
   Arrcus, Inc.

   Email: keyur@arrcus.com


   Luay Jalil
   Verizon

   Email: luay.jalil@verizon.com