Skip to main content

 

Cisco Meraki Documentation

EVPN MultiHoming - Cloud Managed Fabric

  1. Last updated
  2. Save as PDF

EVPN Multi-Homing with Catalyst 9000 switches

EVPN Multi-Homing (EVPN-MH) extends Cloud-Managed Fabric to provide active/active connectivity at the Distribution Layer. It enables an access switch or other downstream device to connect redundantly to a pair of Fabric Leaf switches, with both uplinks actively forwarding traffic.

EVPN Multi-Homing distribution-layer topology showing two Fabric Leaf switches connected by a routed inter-ES link and providing active/active Layer 2 port-channel connectivity to downstream access switches.

Topology overview. Two Fabric Leaf switches form an EVPN-MH pair and provide active/active Layer 2 port-channel connectivity to downstream access switches. The routed inter-ES peer link supports reachability between the members while the EVPN control plane distributes Overlay information.

This design improves resiliency and link utilization while preserving the existing Layer 2 access architecture. It provides a practical migration path for organizations that want to introduce EVPN-VXLAN at the Distribution Layer without immediately converting the access layer to a routed design.

In a Cloud-Managed Fabric deployment, supported Catalyst 9000 switches operate as an EVPN-MH pair at the Distribution Layer. The pair provides logical dual-homed connectivity to downstream devices, while the broader Fabric supplies the EVPN control plane, VXLAN Overlay services, segmentation, and endpoint reachability.

EVPN-MH provides the following benefits:

  • Active/active traffic forwarding across both members of the Distribution pair
  • Improved resiliency during link or switch failures
  • Better utilization of available uplink capacity
  • Continued support for a familiar Layer 2 access design
  • A simplified migration to Cloud-Managed Fabric without requiring an immediate campus-wide redesign

EVPN Multi-Homing with Meraki Dashboard

EVPN Multi-Homing is well suited to campus designs in which the access layer remains Layer 2 and downstream access switches require redundant dual-attached uplinks. By connecting each access switch to both members of an EVPN-MH pair, the design provides active/active forwarding, improved link utilization, and operational continuity during an uplink or Distribution switch failure.

Meraki Dashboard provides guided workflows for creating, onboarding, and managing the EVPN-MH pair. Pair creation is performed separately from Fabric onboarding so that the peer relationship and its associated settings can be established before the devices are assigned a role in the Fabric. This separation provides a clear and predictable deployment sequence.

At a high level, the deployment workflow consists of the following steps:

  1. Create the EVPN-MH pair in the Switching section of Meraki Dashboard.
  2. Configure the peer-specific parameters, including the BGP ASN, Loopback interfaces, inter-ES port-channel, and associated IP addresses.
  3. Add the EVPN-MH pair to the Cloud-Managed Fabric and assign it the Leaf role.
  4. Provision the required Overlay subnets on the EVPN-MH Leaf pair.
  5. Connect the downstream Layer 2 access switches to both members of the pair by using bundled port-channel uplinks.

The order of operations is important. The EVPN-MH pair must be created before it can be added to the Fabric. After the pairing workflow is completed, Meraki Dashboard represents the two switches as a single EVPN-MH entity for Fabric onboarding and role assignment. The pair can then participate in the Cloud-Managed Fabric as a dual-homed Leaf pair while retaining the individual switch configuration required for peer connectivity and downstream port-channels.

This workflow allows organizations to introduce EVPN-VXLAN services at the Distribution Layer while preserving their existing Layer 2 access architecture. It also simplifies deployment by separating EVPN-MH peer configuration, Fabric onboarding, Overlay subnet provisioning, and downstream access connectivity into distinct operational steps.

Hardware and software requirements

EVPN Multi-Homing requires supported Catalyst 9000 switches operating as a Distribution Layer Leaf pair.

Supported hardware

  • Catalyst 9300X
  • Catalyst 9500 High Performance

Supported software

  • Cisco IOS XE 17.18.3
  • Cisco IOS XE 26.1.1

Release status: EVPN Multi-Homing support in Cisco IOS XE 17.18.3 and 26.1.1 is available as a public beta.

Deployment and upgrade recommendations

For normal operation, it is recommended to use the same switch platform and Cisco IOS XE release on both members of an EVPN-MH pair. Matching the hardware and software provides consistent capabilities, behavior, and operational support across the pair. Although matching hardware and software is recommended, it is not mandatory.

During a staged software upgrade, the EVPN-MH peers can temporarily run different supported Cisco IOS XE releases. This allows one switch to remain operational and continue forwarding traffic while the other switch is upgraded and restarted. After the first switch returns to service, the second switch can be upgraded, providing a simplified upgrade path that maintains connectivity through at least one member of the pair.

The temporary software difference should be limited to the upgrade window. After the upgrade is complete, both EVPN-MH peers should be returned to the same recommended Cisco IOS XE release. Verify platform and release compatibility before beginning the upgrade.

EVPN Multi-Homing workflows

Meraki Dashboard separates EVPN Multi-Homing pair creation from Fabric onboarding. Complete the pairing workflow first, then add the pair to the Fabric, provision the required Overlay subnets, and build the downstream port-channels. Following this order ensures that the pair is available as a single selectable entity when Fabric roles and subnets are assigned.

Important: Create and save the EVPN-MH pair before attempting to add it to the Cloud-Managed Fabric.

Workflow overview

  1. Create the EVPN-MH pair and select its two member switches.
  2. Configure the BGP ASN, Loopbacks, and inter-ES port-channel.
  3. Add the completed EVPN-MH pair to the Fabric and assign the Leaf role.
  4. Provision the required Overlay subnets on the EVPN-MH Leaf pair.
  5. Create the downstream port-channels between the EVPN-MH pair and the Layer 2 access switches.

Before you begin

Collect the following information before starting the Dashboard workflow:

  • The two supported Distribution Layer switches that will form the EVPN-MH pair.
  • The BGP autonomous system number used by the target Fabric.
  • A Loopback interface and unique /32 address for each member. Loopback100 is reserved for the Fabric and cannot be selected.
  • At least two physical interfaces on each member for the inter-ES port-channel.
  • A /31 or /30 subnet for the routed inter-ES link.
  • The downstream interfaces that connect each EVPN-MH member to the Layer 2 access switches.

Create the EVPN-MH pair

Create the pair outside the Fabric workflow. Meraki Dashboard then presents the two switches as one EVPN-MH entity for Fabric onboarding.

  1. Navigate to Switching > EVPN Multi-Homing.
  2. Select Create EVPN-MH pair.
  3. Enter a descriptive name for the pair.
  4. Select the two switches that will operate as the EVPN-MH members.
  5. Continue to the peer-configuration page.

Meraki Dashboard workflow for creating an EVPN Multi-Homing pair, naming the pair, and selecting exactly two supported member switches.

Figure 1. Create the EVPN-MH pair from Switching > EVPN Multi-Homing, provide a descriptive name, and select exactly two supported switches as members.

Configure the EVPN-MH peer settings

Configure the control-plane and inter-ES connectivity for both switches. Review the addressing carefully before saving the pair.

  • BGP ASN: Enter the autonomous system number used by the target Fabric.
  • Loopback interface: Select an existing Loopback or create a new one for each member. Loopback100 is reserved for the Fabric and cannot be used.
  • Loopback IP address: Assign a unique /32 address to each member for the direct BGP EVPN peering between the two Leaves.
  • EVPN-MH port-channel: Select the interfaces used for the inter-ES link. A minimum of two interfaces is required. An eligible existing port-channel can also be selected.
  • Inter-ES IP addresses: Assign a /31 or /30 subnet to the routed inter-ES port-channel. Meraki Dashboard enables OSPF on this link.

After reviewing the settings, save the configuration. Dashboard creates the EVPN-MH pair and makes it available for Fabric onboarding.

Meraki Dashboard workflow for configuring the Fabric BGP ASN, unique Loopback addresses, and the routed inter-ES port-channel for an EVPN Multi-Homing pair.

Figure 2. Configure the Fabric BGP ASN, unique Loopback addresses, and the routed inter-ES port-channel. Dashboard uses these values to establish direct BGP EVPN peering and OSPF connectivity between the two Leaves.

Add the EVPN-MH pair to the Fabric

  1. Navigate to Organization > Configure > Fabric and open the target Fabric.
  2. Under Device Roles, select Add Devices.
  3. Choose the Dashboard network that contains the EVPN-MH pair.
  4. Select the EVPN-MH pair.
  5. Under Actions, assign the Leaf role to the pair and save the change.

Expected result: The EVPN-MH pair appears in the Fabric as a Leaf pair and can be selected when Overlay subnets are provisioned.

Meraki Dashboard Fabric workflow showing a completed EVPN Multi-Homing pair being added to the target Fabric and assigned the Leaf role.

Figure 3. Add the completed EVPN-MH entity to the target Fabric from Organization > Configure > Fabric and assign the Leaf role.

Provision Overlay subnets

Provision each required Fabric subnet on the EVPN-MH Leaf pair. Dashboard applies the necessary EVPN-MH and Overlay configuration to both members.

  1. Open Fabric Subnets in the target Fabric.
  2. For the required subnet, click Edit in Actions.
  3. Under Add to Leaf, select the EVPN-MH pair.
  4. Review the subnet settings and select Save subnet.
  5. Repeat the process for each additional subnet required on the EVPN-MH pair.

Meraki Dashboard Edit subnet workflow showing an EVPN Multi-Homing Leaf pair selected under Add to Leaf for Overlay subnet provisioning.

Figure 4. Select the EVPN-MH Leaf pair under Add to Leaf when provisioning each required Overlay subnet.

Create downstream port-channels

Workflow boundary: Downstream port-channels are configured from Switching > Switch ports; they are not created by the Fabric workflow.

  1. Navigate to Switching > Switch Ports.
  2. Filter by Switch Name and select the uplink ports on the Access switch that connect to the EVPN-MH pair.
  3. Select Aggregate to create the access-switch port-channel.
  4. Select both Leaf members (EVPN-MH pair) and the corresponding downstream interfaces.
  5. Select Aggregate to create the port-channel across the EVPN-MH pair.
  6. Confirm that the intended Layer 2 settings are consistent across the bundled links.

Meraki Dashboard Switch ports page showing access-switch uplink interfaces selected for port-channel aggregation toward the EVPN Multi-Homing pair.

Figure 5a. On the access switch, select the uplink interfaces that connect to the EVPN-MH pair and choose Aggregate to create the access-side port-channel.

Meraki Dashboard Switch ports workflow showing corresponding downstream interfaces selected across both EVPN Multi-Homing Leaves for paired aggregation.

Figure 5b. Select the corresponding downstream interfaces across both EVPN-MH Leaves and choose Aggregate to create the paired Leaf-side port-channel.

Post-deployment verification

After provisioning is complete, verify the EVPN-MH pair, Fabric role, Overlay subnets, downstream port-channels, client learning, and EVPN control-plane peerings.

Verify Dashboard state

  • Confirm that both members of the EVPN-MH pair are online and displayed as one pair in Dashboard.
  • Confirm that the pair is present in the target Fabric with the Leaf role.
  • Confirm that each required Overlay subnet is assigned to the EVPN-MH pair.
  • Confirm that the downstream port-channel is operational on the access switch and both EVPN-MH members.

Verify client learning and forwarding

Confirm that clients connected through the Layer 2 access switches are learned through the downstream port-channel. Validate reachability through both members of the EVPN-MH pair and confirm that traffic continues to forward during an approved single-link or single-member failure test.

EVPN Multi-Homing verification output and topology showing client MAC and IP information learned through the downstream port-channel on both Leaves.

Figure 6. Confirm that client MAC and IP information is learned through the downstream port-channel on both EVPN-MH Leaves and is advertised through the Fabric.

Verify BGP EVPN peerings

Each EVPN-MH Leaf should establish one direct BGP EVPN peering with the other member and the required BGP EVPN peerings with the Fabric Spines. BGP neighbors can be seen under Fabric > Detail View > BGP peers

Screenshot 2026-07-27 at 4.59.11 PM.png

Same can be verified using following command to review the neighbor state:

show bgp l2vpn evpn summary

Confirm that the expected neighbors are present and established. Investigate any missing or non-established peer before proceeding.

Cisco IOS XE BGP EVPN summary output showing the direct EVPN Multi-Homing peer and expected Fabric Spine peerings.

Figure 7. Verify one direct BGP EVPN session between the EVPN-MH members and the expected BGP EVPN sessions from each Leaf to the Fabric Spines.

  1. Back to top
  • Was this article helpful?

Recommended articles

  1. Article type
    Reference
  2. Tags
    This page has no tags.