PermitANY
AristaConfig423 downloads

Arista EOS VXLAN EVPN (Leaf-Spine)

VXLAN EVPN leaf-spine fabric for Arista EOS: eBGP underlay with OSPF, VTEP (VXLAN interface), BGP EVPN address family, L2VNI and L3VNI per tenant VRF.

↓ example.txt

What is it?

VXLAN EVPN on Arista is the modern standard for data center fabric overlays. VXLAN (Virtual Extensible LAN) encapsulates L2 frames in UDP packets, extending L2 domains over L3 underlay networks. EVPN (Ethernet VPN) is a BGP address family that distributes MAC and IP reachability information between VTEPs (VXLAN Tunnel Endpoints — the leaf switches). Each leaf is a VTEP with a unique Loopback1 IP as the VXLAN source. L2VNIs carry tenant VLANs between leaves; L3VNIs provide inter-VXLAN routing within a VRF. Spine switches run BGP as route reflectors — they don't participate in VXLAN.

When to use

  • ·Any modern data center fabric with more than 3-4 leaf switches
  • ·Multi-tenant data centers requiring isolated L2/L3 domains per tenant
  • ·Replacing legacy STP-based designs (vPC, MLAG) with a more scalable overlay
  • ·Stretch L2 domains across multiple data centers (DCI — Data Center Interconnect)

When NOT to use

  • ·Small 2-4 switch environments — MLAG is simpler for small deployments
  • ·Environments without BGP expertise — VXLAN EVPN requires solid BGP knowledge to operate

How it compares

vs Cisco NX-OS VXLAN EVPNSame VXLAN EVPN standards. Cisco requires 'feature nv overlay', 'feature vn-segment-vlan-based'. Arista enables these natively. BGP config is nearly identical — both use L2VNI per VLAN and L3VNI per VRF.
vs MLAGMLAG provides active-active dual-homing for two switches. VXLAN EVPN scales to hundreds of leaf switches with any-to-any L2/L3 connectivity. VXLAN EVPN with ESI (Ethernet Segment Identifier) replaces MLAG in modern designs.

Prerequisites

  • →OSPF or IS-IS underlay must be running between all leaf and spine switches (loopbacks reachable)
  • →Two loopback interfaces per leaf: Loopback0 (router-id/BGP), Loopback1 (VTEP source)
  • →VNI IDs and Route Distinguishers planned across all leaves before deployment
  • →Spine switches configured as BGP route reflectors before enabling EVPN on leaves
  • →MTU on all fabric interfaces must be at least 1550 (for VXLAN 50-byte overhead)

Config Generator

Verification commands

show bgp evpn summary

Verify EVPN BGP sessions are Established with spine peers

ExpectedSpine peers show 'Estab' state with non-zero EVPN prefix counts

show vxlan vtep

Show all remote VTEPs learned via EVPN

ExpectedLists all remote leaf loopback IPs (Loopback1 addresses). Missing VTEPs indicate EVPN not advertising them.

show bgp evpn route-type mac-ip

Show EVPN MAC-IP routes (Type 2) — VM MAC and IP reachability

ExpectedLists MAC/IP pairs for all VMs with their VTEP next-hop. Use to verify VM reachability.

Debug commands

show vxlan address-table

Show VXLAN MAC-to-VTEP forwarding table

Common mistakes & fixes

SymptomVMs on different leaves cannot communicate
CauseEVPN BGP not advertising MAC/IP routes, or VNI mismatch between leaves
FixVerify 'show bgp evpn route-type mac-ip' shows the remote VM's MAC. Check VNI IDs match between both leaves for the same VLAN. Verify VTEP appears in 'show vxlan vtep'.
SymptomInter-VXLAN routing (L3VNI) not working
CauseL3VNI VLAN SVI not configured, or VRF route targets don't match
FixVerify 'interface Vlan<l3vni-vlan>' exists with 'ip forward' in the correct VRF. Check 'show bgp evpn route-type ip-prefix' for VRF routes. Verify route-target import/export matches on all leaves.
SymptomMTU errors — VXLAN traffic fragmented or dropped
CauseFabric interface MTU not large enough for VXLAN 50-byte overhead
FixSet all fabric (spine-leaf) interface MTU to at least 9214: 'mtu 9214' on each interface. Verify with 'show interfaces <if> status' showing MTU 9214.

Further learning

Recommended books to go deeper on this topic.

Building Data Centers with VXLAN BGP EVPN

Lukas Krattiger, Shyam Kapadia, David Jansen · Cisco Press · Clos fabric design with VXLAN EVPN: route types, symmetric IRB, multihoming, multi-site

Paperback ↗

BGP Design and Implementation

Randy Zhang & Micah Bartell · Cisco Press · BGP fundamentals needed before EVPN — attributes, route reflection, policy

Kindle ↗Hardcover ↗Paperback ↗

NX-OS and Cisco Nexus Switching: Next-Generation Data Center Architectures

Ron Fuller, David Jansen, Matthew McPherson · Cisco Press · VXLAN BGP EVPN on NX-OS — overlay design and configuration

Hardcover ↗

Related configs

arista-ospf →arista-mlag →arista-bgp →