draft-ietf-lsvr-applicability.xml

<?xml version="1.0" encoding="US-ASCII"?>
<?rfc toc="yes"?>
<?rfc tocompact="yes"?>
<?rfc tocdepth="3"?>
<?rfc tocindent="yes"?>
<?rfc symrefs="yes"?>
<?rfc sortrefs="yes"?>
<?rfc comments="yes"?>
<?rfc inline="yes"?>
<?rfc compact="yes"?>
<?rfc subcompact="no"?>
<rfc category="info" docName="draft-ietf-lsvr-applicability-22"
     ipr="trust200902" submissionType="IETF" tocDepth="5" version="2">
  <front>
    <title abbrev="BGP-SPF Applicability">Usage and Applicability of BGP
    Link-State Shortest Path Routing (BGP-SPF) in Data Centers</title>

    <author fullname="Keyur Patel" initials="K" surname="Patel">
      <organization>Arrcus, Inc.</organization>

      <address>
        <postal>
          <street>2077 Gateway Pl</street>

          <city>San Jose, CA</city>

          <country>USA</country>

          <code>95110</code>
        </postal>

        <phone/>

        <email>keyur@arrcus.com</email>
      </address>
    </author>

    <author fullname="Acee Lindem" initials="A" surname="Lindem">
      <organization>LabN Consulting, L.L.C.</organization>

      <address>
        <postal>
          <street>301 Midenhall Way</street>

          <city>Cary, NC</city>

          <country>USA</country>

          <code>95110</code>
        </postal>

        <phone/>

        <email>acee.ietf@gmail.com</email>
      </address>
    </author>

    <author fullname="Shawn Zandi" initials="S" surname="Zandi">
      <organization>Linkedin</organization>

      <address>
        <postal>
          <street>222 2nd Street</street>

          <city>San Francisco</city>

          <region>CA</region>

          <code>94105</code>

          <country>USA</country>
        </postal>

        <email>szandi@linkedin.com</email>
      </address>
    </author>

    <author fullname="Gaurav Dawra" initials="G" surname="Dawra">
      <organization>Linkedin</organization>

      <address>
        <postal>
          <street>222 2nd Street</street>

          <city>San Francisco</city>

          <region>CA</region>

          <code>94105</code>

          <country>USA</country>
        </postal>

        <email>gdawra@linkedin.com</email>
      </address>
    </author>

    <author fullname="Jie Dong" initials="J." surname="Dong">
      <organization>Huawei Technologies</organization>

      <address>
        <postal>
          <street>No. 156 Beiqing Road</street>

          <city>Beijing</city>

          <region/>

          <code/>

          <country>China</country>
        </postal>

        <email>jie.dong@huawei.com</email>
      </address>
    </author>

    <date/>

    <abstract>
      <t>This document discusses the usage and applicability of BGP Link-State
      Shortest Path First (BGP-SPF) extensions in data center networks
      utilizing Clos or Fat-Tree topologies. The document is intended to
      provide simplified guidance for the deployment of BGP-SPF
      extensions.</t>
    </abstract>
  </front>

  <middle>
    <section title="Introduction">
      <t>This document complements <xref format="default"
      target="I-D.ietf-lsvr-bgp-spf"/> by discussing the applicability of the
      BGP-SPF technology in a simple and fairly common deployment scenario,
      which is described in <xref format="default" target="usecases"/>.</t>

      <t><xref format="default" target="motivation"/> describes the reasons
      for BGP modifications for such deployments.</t>

      <t><xref format="default" target="bgpspf"/> covers the BGP Link-State
      Shortest Path First (IGP-SPF) protocol enhancements to BGP to meet these
      requirements and their applicability to data center <xref
      format="default" target="Clos"/> networks.</t>
    </section>

    <section anchor="reco" title="Recommended Reading">
      <t>This document assumes knowledge of existing data center networks and
      data center network topologies <xref format="default" target="Clos"/>.
      This document also assumes knowledge of data center routing protocols
      such as BGP <xref format="default" target="RFC4271"/>, BGP-SPF <xref
      format="default" target="I-D.ietf-lsvr-bgp-spf"/>, OSPF <xref
      format="default" target="RFC2328"/> <xref target="RFC5340"/>, as well as
      data center Operations, Administration, and Maintenance (OAM) protocols
      like Link Layer Discovery Protocol (LLDP) <xref format="default"
      target="RFC4957"/> and Bi-Directional Forwarding Detection (BFD) <xref
      format="default" target="RFC5580"/>.</t>
    </section>

    <section anchor="usecases" title="Common Deployment Scenario">
      <t>Within a data center, servers are commonly interconnected using the
      Clos topology <xref format="default" target="Clos"/>. The Clos topology
      is fully non-blocking and the topology is realized using Equal Cost
      Multi-Path (ECMP). In a multi-stage Clos topology, the minimum number of
      parallel paths in each tier is determined by the width of the stage as
      shown in the figure 1.</t>

      <t>
        <figure anchor="fig1">
          <name>Illustration of the basic Clos</name>

          <artwork align="left" alt="" name="" type=""><![CDATA[

                                  Tier 1
                                  +-----+
                                  |NODE |
                               +->|  1  |--+
                               |  +-----+  |
                       Tier 2  |           |  Tier 2
                      +-----+  |  +-----+  |  +-----+
       +------------->|NODE |--+->|NODE |--+--|NODE |--------------+
       |        +-----|  5  |--+  |  2  |  +--|  7  |-----+        |
       |        |     +-----+     +-----+     +-----+     |        |
       |        |                                         |        |
       |        |     +-----+     +-----+     +-----+     |        |
       | +------+---->|NODE |--+  |NODE |  +--|NODE |-----+------+ |
       | |      | +---|  6  |--+->|  3  |--+--|  8  |---+ |      | |
       | |      | |   +-----+  |  +-----+  |  +-----+   | |      | |
       | |Tier 3| |            |           |            | |Tier 3| |
     +-----+ +-----+           |  +-----+  |          +-----+ +-----+
     |NODE | |NODE |           +->|NODE |--+          |NODE | |NODE |
     |  9  | | 10  |              |  4  |             | 11  | | 12  |
     +-----+ +-----+              +-----+             +-----+ +-----+
      | | |   | | |                                    | | |    | | |
      <- Servers ->                                    <- Servers ->

      Tier 1 is comprised of Nodes 1, 2, 3, and 4
      Tier 2 is comprised of Nodes 5, 6, 7, and 8
      Tier 3 is comprised of Nodes 9, 10, 11, and 12

        ]]></artwork>
        </figure>
      </t>
    </section>

    <section anchor="motivation" title="Justification for BGP-SPF Extension">
      <t>To simplify L3 routing and operations, many data centers use BGP as a
      routing protocol to create both an underlay and an overlay network for
      their Clos Topologies <xref format="default" target="RFC7938"/>.
      However, BGP is a path-vector routing protocol. Since it does not create
      a fabric topology, it uses hop-by-hop External BGP (EBGP) peering to
      facilitate hop-by-hop routing to create the underlay network and to
      resolve any overlay next hops. The hop-by-hop BGP peering paradigm
      imposes several restrictions within a Clos. It prohibits the deployment
      of Route Reflectors/Route Controllers as the EBGP sessions are congruent
      with the data path. The BGP best-path algorithm is prefix-based and it
      prevents announcements of prefixes to other BGP speakers until the
      best-path decision process has been performed for the prefix at each
      intermediate hop. These restrictions significantly delay the overall
      convergence of the underlay network within a Clos network.</t>

      <t>The BGP-SPF modifications allow BGP to overcome these limitations.
      Furthermore, using the BGP-LS Network Layer Reachability Information
      (NLRI) format allows the BGP-SPF data to be advertised for nodes, links,
      and prefixes in the BGP routing domain and used for Short-Path-First
      (SPF) computations <xref target="RFC9552"/>.</t>

      <t>Additional motivation for deploying BGP-SPF is included in <xref
      target="I-D.ietf-lsvr-bgp-spf"/>.</t>
    </section>

    <section anchor="bgpspf" title="BGP-SPF Applicability to Clos Networks">
      <t>With the BGP-SPF extensions <xref format="default"
      target="I-D.ietf-lsvr-bgp-spf"/>, the BGP best-path computation and
      route computation are replaced with link-state algorithms such as those
      used by OSPF <xref format="default" target="RFC2328"/>, both to
      determine whether an BGP-LS-SPF NLRI has changed and needs to be
      re-advertised and to compute the BGP routes. These modifications will
      significantly improve convergence of the underlay while affording the
      operational benefits of a single routing protocol <xref format="default"
      target="RFC7938"/>.</t>

      <t>Data center controllers typically require visibility to the BGP
      topology to compute traffic-engineered paths. These controllers learn
      the topology and other relevant information via the BGP-LS address
      family <xref format="default" target="RFC9552"/> which is totally
      independent of the underlay address families (usually IPv4/IPv6
      unicast). Furthermore, in traditional BGP underlays, all the BGP routers
      will need to advertise their BGP-LS information independently. With the
      BGP-SPF extensions, controllers can learn the topology using the same
      BGP advertisements used to compute the underlay routes. Furthermore,
      these data center controllers can avail the convergence advantages of
      the BGP-SPF extensions. The placement of controllers can be outside of
      the forwarding path or within the forwarding path.</t>

      <t>Alternatively, as each and every router in the BGP-SPF domain will
      have a complete view of the topology, the operator can also choose to
      configure BGP sessions in the hop-by-hop peering model described in
      <xref format="default" target="RFC7938"/> along with BFD <xref
      format="default" target="RFC5580"/>. In doing so, while the hop-by-hop
      peering model lacks the inherent benefits of the controller-based model,
      BGP updates need not be serialized by the BGP best-path algorithm in
      either of these models. This helps overall network convergence.</t>

      <section anchor="lsvrsafi" title="Usage of BGP-LS SPF SAFI">
        <t>Section 5.1 of <xref format="default"
        target="I-D.ietf-lsvr-bgp-spf"/> defines a new BGP-LS-SPF SAFI for
        announcement of the BGP-SPF link-state. The NLRI format and its
        associated attributes follow the format of BGP-LS for node, link, and
        prefix announcements. Whether the peering model within a Clos follows
        hop-by-hop peering described in <xref format="default"
        target="RFC7938"/> or any controller-based or route-reflector peering,
        an operator can exchange BGP-LS-SPF SAFI routes over the BGP peering
        by simply configuring BGP-LS-SPF SAFI between the necessary BGP
        speakers.</t>

        <t>The BGP-LS-SPF SAFI can also co-exist with BGP IP Unicast SAFI
        <xref target="RFC4760"/> which could exchange overlapping IP routes.
        One use case for this is where BGP-LS-SPF routes are used for the
        underlay and BGP IP Unicast routes for VPNs are advertised in the
        overlay as described in <xref target="RFC4364"/>. The routes received
        by these SAFIs are evaluated, stored, and announced independently
        according to the rules of <xref format="default" target="RFC4760"/>.
        The tie-breaking of route installation is a matter of the local
        policies and preferences of the network operator.</t>

        <t>Finally, as the BGP-SPF peering is done following the procedures
        described in <xref format="default" target="RFC4271"/>, all the
        existing transport security mechanisms including <xref
        format="default" target="RFC5925"/> are available for the BGP-LS-SPF
        SAFI.</t>

        <section anchor="other-safi"
                 title="Relationship to Other BGP AFI/SAFI Tuples">
          <t>Normally, the BGP-LS-SPF AFI/SAFI is used solely to compute the
          underlay and is given precedence over other AFI/SAFIs in route
          processing. Other BGP SAFIs, e.g., IPv6/IPv6 Unicast VPN would use
          the BGP-SPF computed routes for next hop resolution.</t>
        </section>
      </section>

      <section anchor="peering" title="Peering Models">
        <t>As previously stated, BGP-SPF can be deployed using the existing
        peering model where there is a single-hop BGP session on each and
        every link in the data center fabric <xref format="default"
        target="RFC7938"/>. This provides for both the advertisement of routes
        and the determination of link and neighboring router availability.
        With BGP-SPF, the underlay will converge faster due to changes to the
        decision process that will allow NLRI changes to be advertised faster
        after detecting a change.</t>

        <section anchor="sparse-peering" title="Sparse Peering Model">
          <t>Alternately, BFD <xref format="default" target="RFC5580"/> can be
          used to swiftly determine the availability of links and the BGP
          peering model can be significantly sparser than the data center
          fabric. BGP-SPF sessions only need to be established with enough
          peers to provide a bi-connected graph. If Internal BGP (IBGP) is
          used, then the BGP routers at tier N-1 will act as route-reflectors
          for the routers at tier N.</t>

          <t>The obvious usage of sparse peering is to avoid parallel BGP
          sessions on links between the same two routers in the data center
          fabric. However, this use case is not very useful since parallel L3
          links between the same two BGP routers are rare in Clos or Fat-Tree
          topologies. Additionally, when there are multiple links, they are
          often aggregated at the link layer using Link Aggregation Groups
          (LAGs) <xref target="IEEE.802.1AX"/> rather than at the IP layer.
          Two more interesting scenarios are described below.</t>

          <t>In current data center topologies, there is often a very dense
          mesh of links between levels, e.g., leaf and spine, providing
          32-way, 64-way, or more Equal-Cost Multi-Path (ECMP) paths. In these
          topologies, it is desirable not to have a BGP session on every link
          and techniques such as the one described in <xref format="default"
          target="bi-connected"/> can be used to establish sessions on some
          subset of northbound links. For example, in a Spine-Leaf topology,
          each leaf router would only peer with a subset of the spines
          dependent on the flooding redundancy required to be reasonably
          certain that every node within the BGP-SPF routing domain has the
          complete topology.</t>

          <t>Alternately, controller-based data center topologies are
          envisioned where BGP speakers within the data center only establish
          BGP sessions with two or more controllers. In these topologies,
          fabric nodes below the first tier, as shown in Figure 1 of <xref
          format="default" target="RFC7938"/>, will establish BGP multi-hop
          sessions with the controllers. For the multi-hop sessions,
          determining the route to the controllers without depending on BGP
          would need to be through some other means beyond the scope of this
          document. However, the BGP discovery mechanisms described in <xref
          format="default" target="bgp-discovery"/> would be one
          possibility.</t>
        </section>

        <section anchor="bi-connected" title="Bi-Connected Graph Heuristic">
          <t>With this heuristic, discovery of BGP SPF peers is assumed, e.g.,
          as described in <xref format="default" target="bgp-discovery"/>. In
          this context, "bi-connected" refers to the fact that there must be
          an adverised link NLRI for both BGP SPF peers associated with the
          link before the link can be used in the BGP SPF route calcuation.
          Additionally, it assumed that the direction of the peering can be
          ascertained. In the context of a data center fabric, the direction
          is either northbound (toward the spine), southbound (toward the
          Top-Of-Rack (ToR) routers) or east-west (same level in the
          hierarchy). The determination of the direction is beyond the scope
          of this document. However, it would be reasonable to assume a
          technique where the ToR routers can be identified and the number of
          hops to the ToR is used to determine the direction.</t>

          <t>In this heuristic, BGP speakers allow passive session
          establishment for southbound BGP sessions. For northbound sessions,
          BGP speakers will attempt to maintain two northbound BGP sessions
          with different routers. For east-west sessions, passive BGP session
          establishment is allowed. However, a BGP speaker will never actively
          establish an east-west BGP session unless it cannot establish two
          northbound BGP sessions.</t>

          <t>BGP SPF sparse peering deployments not using this hueristic are
          possible but are not described herein and are considered out of
          scope.</t>
        </section>
      </section>

      <section anchor="bgp-policy" title="BGP Spine/Leaf Topology Policy">
        <t>One of the advantages of using BGP-SPF as the underlay protocol is
        that BGP policy can be applied at any level. For example, depending on
        the topology, it may be possible to aggregate or filter prefix
        advertisements using existing BGP policy. In Spine/Leaf topologies, it
        is not necessary to advertise BGP-LS Prefix NLRI received by leaf
        nodes from the spine back to other spine nodes. If a common AS is used
        for the spine nodes, this can easily be accomplished with EBGP and a
        simple policy to filter advertisements from the leaves to the spine if
        the first AS in the AS path is the spine AS.</t>

        <t>In the figure below, the leaves would not advertise any NLRI with
        AS 64512 as the first AS in the AS path.</t>

        <t>
          <figure anchor="fig2">
            <name>Spine/Leaf Topology Policy</name>

            <artwork align="left" alt="" name="" type=""><![CDATA[
             +--------+    +--------+             +--------+
 AS 64512    |        |    |        |             |        |
 for Spine   | Spine 1+----+ Spine 2+- ......... -+ Spine N|
 Nodes at    |        |    |        |             |        |
 this Level  +-+-+-+-++    ++-+-+-+-+             +-+-+-+-++
        +------+ | | |      | | | |                 | | | |
        |  +-----|-|-|------+ | | |                 | | | |
        |  |  +--|-|-|--------+-|-|-----------------+ | | |
        |  |  |  | | |    +---+ | |                   | | |
        |  |  |  | | |    |  +--|-|-------------------+ | |
        |  |  |  | | |    |  |  | |              +------+ +----+
        |  |  |  | | |    |  |  | +--------------|----------+  |
        |  |  |  | | |    |  |  +-------------+  |          |  |
        |  |  |  | | +----|--|----------------|--|--------+ |  |
        |  |  |  | +------|--|--------------+ |  |        | |  |
        |  |  |  +------+ |  |              | |  |        | |  |
       ++--+--++      +-+-+--++            ++-+--+-+     ++-+--+-+
       | Leaf 1|      | Leaf 2|  ........  | Leaf X|     | Leaf Y|
       +-------+      +-------+            +-------+     +-------+


       ]]></artwork>
          </figure>
        </t>
      </section>

      <section anchor="bgp-discovery-con"
               title="BGP Peer Discovery Considerations">
        <t>The basic functionality of peer discovery is to be discover the
        address of a single-hop peer in case where the peer address is not
        pre-configured. This is being accomplished today by using IPv6 Router
        Advertisements (RA) <xref format="default" target="RFC4861"/> and
        assuming that a BGP session is desired with any discovered peer.
        Beyond the basic functionality, it may be useful to have the following
        information relating to the BGP session:</t>

        <list style="symbols">
          <t>Autonomous System (AS) and BGP Identifier of a potential
          peer.</t>

          <t>Security capabilities supported and for cryptographic
          authentication, the security capabilities and possibly a key-chain
          <xref format="default" target="RFC8177"/> to be used.</t>

          <t>Session Policy Identifier - A group number or name used to
          associate common session parameters with the peer. For example, in a
          data center, BGP sessions with a ToR device could have different
          parameters than BGP sessions between leaf and spine.</t>
        </list>

        <t>In a data center fabric, it is often useful to know whether a peer
        is southbound (towards the servers) or northbound (towards the spine
        or super-spine), e.g., <xref format="default" target="bi-connected"/>.
        One mechanism, without specifying all the details, might be for the
        ToR routers to be identified when installed and for the others routers
        in the fabric to determine their level based on the distance from the
        closest ToR router.</t>

        <t>If there are multiple links between BGP speakers or the links
        between BGP speakers are unnumbered, it is also useful to be able to
        establish multi-hop sessions using the loopback addresses. This will
        often require the discovery protocol to install route(s) toward the
        potential peer loopback addresses prior to BGP session
        establishment.</t>

        <t>Finally, a simple BGP discovery protocol may be used to establish a
        multi-hop session with one or more controllers by advertising
        connectivity to one or more controllers.</t>
      </section>

      <section anchor="bgp-discovery" title="BGP Peer Discovery">
        <section anchor="bgp-ipv6-peering" title="BGP IPv6 Simplified Peering">
          <t>To conserve IPv4 address space and simplify operations, BGP-SPF
          routers in Clos/Fat Tree deployments can use IPv6 addresses as peer
          address. For IPv4 address families, IPv6 peering as specified in
          <xref format="default" target="RFC8950"/> can be deployed to avoid
          configuring IPv4 addresses on router interfaces. When this is done,
          dynamic discovery mechanisms, as described in <xref format="default"
          target="bgp-discovery"/>, can be used to learn the global or
          link-local IPv6 peer addresses and IPv4 addresses need not be
          configured on these interfaces. If IPv6 link-local peering is used,
          then configuration of IPv6 global addresses is also not required
          <xref target="RFC7404"/> . The Link Local/Remote Identifiers of the
          peering interfaces MUST be used in the link NLRI as described in
          section 5.2.2 of <xref target="I-D.ietf-lsvr-bgp-spf"/>.</t>
        </section>

        <section anchor="config-checking"
                 title="BGP-LS SPF Topology Visibility for Management">
          <t>Irrespective of whether or not BGP-SPF is used for route
          calculation, the BGP-LS-SPF route advertisements can be used to
          periodically construct the Clos/Fat Tree topology. This is
          especially useful in deployments where an Interior Gateway Protocol
          (IGP) is not used and the base BGP-LS routes <xref format="default"
          target="RFC9552"/> are not available. The resultant topology
          visibility can then be used for troubleshooting and consistency
          checking. This would normally be done on a central controller or
          other management tool which could also be used for fabric data path
          verification. The precise algorithms and heuristics, as well as the
          complete set of management applications is beyond the scope of this
          document.</t>
        </section>

        <section anchor="dci"
                 title="Data Center Interconnect (DCI) Applicability">
          <t>Since BGP-SPF is to be used for the routing underlay and DCI
          gateway boxes typically have direct or very simple connectivity, BGP
          external sessions would typically not include the BGP-LS-SPF
          SAFI.</t>
        </section>
      </section>
    </section>

    <section anchor="other-env"
             title="Non-CLOS/FAT Tree Topology Applicability">
      <t>The BGP-SPF extensions <xref format="default"
      target="I-D.ietf-lsvr-bgp-spf"/> can be used in other topologies and
      avail the inherent convergence improvements. Additionally, sparse
      peering techniques may be utilized <xref format="default"
      target="peering"/>. However, determining whether to establish a BGP
      session is more complex and the heuristic described in <xref
      format="default" target="bi-connected"/> cannot be used. In such
      topologies, other techniques such as those described in <xref
      format="default" target="RFC9667"/> may be employed. One potential
      deployment would be the underlay for a Service Provider (SP) backbone
      where usage of a single protocol, i.e., BGP, is desired.</t>
    </section>

    <section anchor="non-transit-node" title="Non-Transit Node Capability">
      <t>In certain scenarios, a BGP node wishes to participate in the BGP-SPF
      topology but never be used for transit traffic. These include situations
      where a server wants to make application services available to clients
      homed at subnets throughout the BGP-SPF domain but does not ever want to
      be used as a router (i.e., carry transit traffic). Another specific
      instance is where a controller is resident on a server and direct
      connectivity to the controller is required throughout the entire domain.
      This can readily be accomplished using the BGP-LS Node NLRI Attribute
      SPF Status TLV as described in <xref format="default"
      target="I-D.ietf-lsvr-bgp-spf"/>.</t>
    </section>

    <section anchor="policy" title="BGP Policy Applicability">
      <t>Existing BGP policy such as prefix filtering may be used in
      conjunction with the BGP-LS-SPF SAFI. When BGP policy is used with the
      BGP-LS-SPF SAFI, BGP speakers in the BGP-LS-SPF routing domain will not
      all have the same set of NLRI and will compute a different BGP local
      routing table. Consequently, care must be taken to assure routing is
      consistent and blackholes or routing loops do not ensue. However, this
      is no different than if traditional BGP routing using the IPv4 and IPv6
      address families were used.</t>
    </section>

    <section anchor="IANA" title="IANA Considerations">
      <t>No IANA updates are requested by this document.</t>
    </section>

    <section anchor="Security" title="Security Considerations">
      <t>This document introduces no new security considerations above and
      beyond those already specified in the <xref format="default"
      target="RFC4271"/> and <xref format="default"
      target="I-D.ietf-lsvr-bgp-spf"/>.</t>
    </section>

    <section anchor="Acknowledgements" title="Acknowledgements">
      <t>The authors would like to thank Alvaro Retana, Yan Filyurin, Boris
      Hassanov, Stig Venaas, Ron Bonica, Mallory Knodel, Dhruv Dhody, Erik
      Kline, Eric Vyncke, and John Scudder for their review and comments.</t>
    </section>
  </middle>

  <back>
    <references title="Normative References">
      <?rfc include='reference.I-D.ietf-lsvr-bgp-spf.xml'?>
    </references>

    <references title="Informative References">
      <?rfc include='reference.RFC.2328.xml'?>

      <?rfc include='reference.RFC.4271.xml'?>

      <?rfc include='reference.RFC.4364.xml'?>

      <?rfc include='reference.RFC.4760.xml'?>

      <?rfc include='reference.RFC.4861.xml'?>

      <?rfc include='reference.RFC.4957.xml'?>

      <?rfc include='reference.RFC.5340.xml'?>

      <?rfc include='reference.RFC.5580.xml'?>

      <?rfc include='reference.RFC.5925.xml'?>

      <?rfc include='reference.RFC.7404.xml'?>

      <?rfc include='reference.RFC.7938.xml'?>

      <?rfc include='reference.RFC.8177.xml'?>

      <?rfc include='reference.RFC.8950.xml'?>

      <?rfc include='reference.RFC.9552.xml'?>

      <?rfc include='reference.RFC.9667.xml'?>

      <reference anchor="IEEE.802.1AX"
                 target="https://standards.ieee.org/standard/802_1AX-2020.html">
        <front>
          <title>IEEE Standard for Local and Metropolitan Area Networks: Link
          Aggregation</title>

          <author>
            <organization>IEEE</organization>
          </author>

          <date year="2020"/>
        </front>

        <seriesInfo name="IEEE Std" value="802.1AX-2020"/>
      </reference>

      <reference anchor="Clos" target="">
        <front>
          <title>A Study of Non-Blocking Switching Networks</title>

          <author initials="" surname="">
            <organization/>
          </author>

          <date month="March" year="1953"/>
        </front>

        <seriesInfo name=""
                    value="The Bell System Technical Journal, Vol. 32(2), DOI 10.1002/j.1538-7305.1953.tb01433.x"/>
      </reference>
    </references>
  </back>
</rfc>