Skip to content
Reticux Engineering by Reticux Engineering

Day 75 — EVPN Multihoming: ESI, Aliasing, Split Horizon and DF Election

Reticux Engineering —

1. Opening

EVPN multihoming turns a dual-attached Ethernet segment into a control-plane object. Multiple PEs can attach to the same CE/server segment, but the fabric must coordinate load sharing, fast withdrawal, BUM forwarding and loop prevention. The Ethernet Segment Identifier, Type-1/Type-4 routes, aliasing, split horizon and Designated Forwarder election are therefore part of one redundancy system.

2. Concept and standards behavior

RFC 7432 defines ESI-based EVPN multihoming, all-active and single-active redundancy, aliasing, split horizon and base DF election. RFC 8584 adds an extensible DF-election framework and Highest Random Weight (HRW). RFC 9722, published in 2025, updates RFC 8584 with faster recovery behavior. RFC 9746, RFC 9785 and RFC 9786 add newer multihoming capabilities. Cisco implementation support is platform/release specific: NX-OS 10.5(x) documentation includes vPC multihoming and ESI interoperability material, while newer NX-OS releases add native standardized ESI features on supported Nexus platforms. The post therefore does not fabricate 10.5 ESI commands.

The troubleshooting hierarchy used throughout the batch is underlay IP reachability → VTEP/NVE reachability → BGP L2VPN EVPN family → expected EVPN route type → RT/VNI/VRF membership → MAC/IP/ARP/ND state → hardware forwarding → tenant service. This fixed order prevents a healthy upper-layer object from distracting the engineer from an earlier failed dependency.

3. Scenario

The common lab is a two-leaf, two-spine logical VXLAN EVPN fabric. Leaf-1 and Leaf-2 are VTEPs in AS65000 and peer for the L2VPN EVPN address family through a spine/route-reflector function. Host-A is attached to Leaf-1 and Host-B to Leaf-2. Documentation-safe values are used throughout: VTEP loopbacks 192.0.2.11 and 192.0.2.12, route reflector 192.0.2.254, tenant IPv4 203.0.113.0/24, tenant IPv6 2001:db8:100::/64, L2 VNI 10100 and L3 VNI 50100.

Success criteria

  1. Underlay reaches every required VTEP endpoint.
  2. BGP EVPN is Established and the expected family is active.
  3. The correct EVPN route object appears with approved RT/VNI state.
  4. NVE and tenant forwarding state match the control plane.
  5. A positive service test passes.
  6. A negative tenant-isolation or invalid-state test also passes.

4. Topology diagram

A matching editable SVG and PNG are stored under media/diagrams/.

5. Prerequisites

  • Supported Nexus 9000 platform and NX-OS 10.5(x) maintenance release.
  • Stable routed underlay among leaf/spine/VTEP loopbacks.
  • BGP L2VPN EVPN overlay neighbors or RRs.
  • VNI, VLAN, VRF, RD and RT allocation plan.
  • NVE source-loopback plan.
  • Hardware resource/TCAM prerequisites checked for the exact switch.
  • Independent management access and time-synchronized logs.
  • Feature support verified before enabling multihoming, multicast or mobility extensions.

6. Baseline configuration

Topic-specific NX-OS excerpt — not a complete deployable fabric configuration.

! Standards-first conceptual configuration only.
! Native ESI multihoming CLI is not asserted for every NX-OS 10.5(x) platform.
! On an explicitly supported platform/release:
!   - define Ethernet Segment / ESI
!   - associate the multihomed port-channel
!   - select all-active or single-active redundancy
!   - enable required EVPN family behavior
! Use documented vPC multihoming on 10.5(x) platforms where that is the supported design.

Exact syntax, hardware availability and caveats must be validated against the selected Nexus model and NX-OS maintenance release.

7. Verification before modification

show bgp l2vpn evpn summary
show bgp l2vpn evpn
show nve peers
show nve vni
show interface nve1
show mac address-table
show ip arp
show bgp l2vpn evpn route-type 1
show bgp l2vpn evpn route-type 4
show bgp l2vpn evpn
show port-channel summary
show nve peers

The command list is evidence-oriented: EVPN CLI proves control-plane state; NVE CLI proves overlay endpoint/VNI state; MAC/ARP/VRF/FIB evidence proves service programming. None alone proves the whole fabric.

8. Controlled modification

On a platform that explicitly supports native ESI, configure the shared Ethernet Segment and capture Type-4 plus Type-1 EAD state before adding a host. Then test all-active known-unicast and BUM behavior. If the chosen 10.5(x) platform lacks native ESI, perform the RFC exercise on another supported implementation and keep the Cisco vPC exercise separate.

Predict the exact route type, VNI, NVE peer, gateway object or tenant RIB entry that should change before applying the modification.

9. Fault injection

Illustrative lab — not a real incident.

Fail one access member/link while both PEs remain alive. Capture EAD/segment-state changes, remote aliasing/withdrawal behavior and DF state. Then restore the link and observe recovery according to the implementation.

Capture the healthy state and change one variable only. Preserve logs and route evidence before correction.

10. Step-by-step troubleshooting

  1. Verify the affected tenant/VNI/VRF and exact source/destination.
  2. Verify VTEP underlay reachability.
  3. Verify EVPN BGP state and capabilities.
  4. Identify the EVPN route type expected for the service.
  5. Inspect RD, RT, next hop and VNI/VRF mapping.
  6. Verify NVE peer and VNI state.
  7. Verify local/remote MAC and ARP/ND information.
  8. Verify anycast gateway or IP-VRF state if routed traffic is involved.
  9. Verify the hardware/FIB path where platform tooling allows it.
  10. Test the intended positive flow.
  11. Test an isolation/negative case.
  12. Correct the smallest proven cause and repeat the same evidence set.

11. Root cause and correction

The failure belongs to Ethernet-Segment membership/redundancy state, not generic host Type-2 reachability. Restore the access segment/ESI relationship and verify Type-1/Type-4 plus host forwarding before touching tenant RTs.

12. Post-fix verification

Verify underlay reachability, EVPN route type, RT/VNI/VRF membership, NVE state, MAC/IP/ARP/ND state, forwarding, tenant isolation and stability. If a feature is platform-specific, record the exact hardware/release in the execution record.

13. Rollback

Restore the last approved EVPN/VNI/VRF/NVE state. Never remove broad route-target or tenant-isolation controls just to force connectivity. Roll back immediately if a tenant imports another tenant's routes, an anycast gateway becomes inconsistent, a multihomed segment shows loop/duplicate symptoms, or unrelated VNIs are withdrawn.

14. Deep-dive engineering notes

ESI and segment identity

The ESI tells remote PEs that multiple attachment points represent the same Ethernet Segment. Without matching segment identity, two physical links are simply two unrelated attachments.

Aliasing

Remote PEs can use EAD/ES information to infer multiple PEs reach the same segment, improving load sharing and backup behavior even if a host MAC was advertised by only one PE.

Mass withdrawal

When an Ethernet Segment fails, segment-level advertisement changes can allow remote nodes to invalidate many dependent paths more efficiently than waiting for every MAC route individually.

Split horizon

Traffic arriving from a multihomed Ethernet Segment must not be sent back into the same segment in a way that creates loops. EVPN split-horizon procedures provide this protection; newer RFC 9746 extends behavior for certain encapsulations/designs.

DF election

The DF is responsible for particular forwarding responsibilities such as BUM toward the multihomed CE for an EVI/VLAN. DF is not simply “the active PE for all traffic.”

HRW and modern recovery

RFC 8584's HRW election reduces unnecessary service reshuffling. RFC 9722 improves recovery after a failed link/node returns. Current engineering content should therefore not stop at the original modulo election of RFC 7432.

Version caveat

Do not translate an RFC feature directly into Nexus CLI. Verify the exact model, ASIC, NX-OS release and supported topology. Native ESI, vPC multihoming and ESI interoperability are different deployment modes.

15. Production workflow

  1. Inventory exact Nexus hardware and software.
  2. Choose vPC or standardized ESI based on supported design.
  3. If ESI, verify native feature support first.
  4. Record the ESI and redundancy mode.
  5. Capture Type-1 and Type-4.
  6. Record DF state per EVI/VLAN.
  7. Test known-unicast load sharing.
  8. Test BUM forwarding.
  9. Fail one access link.
  10. Fail one PE.
  11. Observe aliasing/mass withdrawal/DF convergence.
  12. Test recovery for duplicate-forwarding or black-hole windows.

Operational exercise — prove redundancy instead of assuming it

For a multihomed segment, run four separate failure tests: one access-member failure, one full PE failure, one underlay loss from a multihoming PE, and one recovery/rejoin event. Record which Type-1/Type-4 routes change, which PE is DF for the affected service, whether known-unicast continues through aliasing, and whether any BUM duplication occurs. A design that survives only the port-channel member failure has not proven EVPN multihoming resilience.

During recovery, pay particular attention to the interval in which a returning PE can resume forwarding. Modern DF procedures exist partly because poorly synchronized recovery can create transient duplicates or loops. Do not use a single ping as evidence; collect counters or packet observations across the transition window where authorized.

Keep a platform-support matrix in the lab package. For each Nexus model/release, record whether the design uses vPC multihoming, standardized ESI, ESI interoperability, and which DF extensions are supported. This prevents engineers from applying a command sequence from a newer 10.6 platform to a 10.5 device that exposes a different feature set.

16. Production lessons

  • Multihoming is a control-plane architecture, not only a port-channel.
  • Type-1 and Type-4 are primary evidence.
  • DF responsibility is traffic-class/service specific.
  • RFC 8584 and RFC 9722 matter for modern DF behavior.
  • Cisco ESI support is model/release specific.
  • Unsupported NX-OS 10.5 syntax must not be invented.

17. Knowledge check

  1. Why can EVPN BGP be Established while tenant forwarding is down?
  2. What should be identified before changing route targets or VNIs?
  3. Why is a negative tenant-isolation test required after restoring connectivity?

Answers

  1. Underlay/VTEP, route-type import, NVE, gateway or forwarding state can fail independently of the BGP session.
  2. The expected EVPN route type and the first failed layer in the service chain.
  3. An overbroad RT/VNI fix can restore connectivity by leaking another tenant, which is a security failure.

18. Sources

  • RFC 7432 — EVPN multihoming base procedures
  • RFC 8584 — Framework for EVPN DF Election Extensibility
  • RFC 9722 — Fast Recovery for EVPN DF Election
  • RFC 9746 — EVPN Multihoming Extensions for Split-Horizon Filtering
  • RFC 9785 — Preference-Based EVPN DF Election
  • RFC 9786 — EVPN Port-Active Redundancy Mode
  • Cisco Nexus 9000 NX-OS 10.5(x)/10.6(x) multihoming documentation
Written by Reticux Engineering

Practical, lab-tested network engineering: Cisco, Juniper and Fortinet configurations, troubleshooting steps and exam-focused practice.

No comments:

Learn BGP, Cisco routing and network security with daily lab-driven lessons, from CCNA basics to CCNP level policy design.