Tuesday, May 12, 2026

Complete Cisco Nexus VXLAN EVPN Configuration Guide Part 2 | BGP EVPN, Anycast Gateway & Inter-VXLAN Routing

Complete Cisco Nexus VXLAN EVPN Configuration Guide Part 2

Complete Cisco Nexus VXLAN EVPN Configuration Guide – Part 2

This is Part 2 of the Cisco Nexus VXLAN series. In Part 1, the network used multicast flood-and-learn VXLAN architecture. This guide upgrades the deployment to a modern VXLAN EVPN fabric using BGP EVPN as the control plane.

Modern data centers no longer depend heavily on multicast flooding for MAC learning. Instead, they use BGP EVPN advertisements for scalable control-plane learning.


Related Article


Table of Contents


1. Introduction to VXLAN EVPN

VXLAN EVPN combines:

  • VXLAN data-plane encapsulation
  • BGP EVPN control-plane learning

Traditional VXLAN used multicast flood-and-learn behavior.

Modern EVPN uses:

BGP to distribute MAC and IP information.

This significantly reduces unnecessary flooding.


2. Flood-and-Learn vs EVPN

Flood-and-Learn VXLAN VXLAN EVPN
Uses multicast heavily Uses BGP control plane
Data-plane MAC learning Control-plane MAC learning
More flooding Minimal flooding
Less scalable Cloud-scale architecture
Harder troubleshooting Better visibility

3. EVPN Mathematics and Scaling

Traditional VLAN Scaling

\\[ 2^{12} = 4096 \\]

Usable VLANs:

\\[ 4096 - 2 = 4094 \\]

VXLAN VNI Scaling

VXLAN uses 24-bit VNIs:

\\[ 2^{24} = 16,777,216 \\]

Flood Reduction Formula

Traditional flooding:

\\[ Traffic \\propto Hosts \\]

EVPN advertisements:

\\[ Traffic \\propto Changes \\]

This drastically reduces unnecessary traffic.


4. Task 1 – Nexus Initialization


switch setup

Enter the password for admin: Cisco123
Confirm the password: Cisco123

The initial setup process configures the admin account and prepares the Nexus switches for VXLAN EVPN deployment.


5. Task 2 – IP Address Configuration

R1 Configuration


hostname R1

interface e0/0
 ip address 10.10.10.1 255.255.255.0
 duplex full
 no shutdown

interface e0/1
 ip address 10.20.20.1 255.255.255.0
 duplex full
 no shutdown

interface loopback0
 ip address 1.1.1.1 255.255.255.255

NX-01


hostname NX-01

interface loopback0
 ip address 192.168.1.1/32

interface ethernet1/1
 no switchport
 ip address 10.10.10.11/24
 no shutdown

NX-02


hostname NX-02

interface loopback0
 ip address 192.168.1.2/32

interface ethernet1/1
 no switchport
 ip address 10.20.20.22/24
 no shutdown

Loopback interfaces are extremely important because they provide stable VTEP addresses for the VXLAN tunnels.


6. Task 3 – OSPF Underlay

The underlay network provides IP connectivity between the VTEPs.

R1 OSPF


router ospf 1
 router-id 1.1.1.1
 network 10.0.0.0 0.255.255.255 area 0
 network 1.0.0.0 0.255.255.255 area 0

NX-01


feature ospf

router ospf 1
 router-id 11.11.11.11

interface loopback0
 ip router ospf 1 area 0

interface ethernet1/1
 ip router ospf 1 area 0

NX-02


feature ospf

router ospf 1
 router-id 22.22.22.22

interface loopback0
 ip router ospf 1 area 0

interface ethernet1/1
 ip router ospf 1 area 0

OSPF Cost Formula

\\[ Cost = \\frac{Reference\\ Bandwidth}{Interface\\ Bandwidth} \\]


7. Task 4 – Enabling VXLAN EVPN Features

Several Nexus features must be enabled before EVPN can operate.


feature bgp
feature interface-vlan
feature vn-segment-vlan-based
feature nv overlay
nv overlay evpn

Feature Explanation

Feature Purpose
feature bgp Enables BGP routing process
feature interface-vlan Allows SVI interfaces
feature vn-segment-vlan-based Enables VLAN to VNI mapping
feature nv overlay Enables VXLAN overlay functionality
nv overlay evpn Activates EVPN control-plane

8. Task 5 – VRF and L3VNI Configuration

The VRF separates routing tables for tenant isolation.

Configuration


vrf context NEXUS

vlan 101
 vn-segment 100101

Why L3VNI Exists

L2VNI handles Layer 2 extension.

L3VNI handles inter-VXLAN routing.


9. Task 6 – SVI Configuration


interface vlan101
 vrf member NEXUS
 ip forward
 no shutdown

The SVI becomes the routed gateway for the VRF.


10. Task 7 – Route Distinguisher and Route Targets


vrf context NEXUS
 vni 100101
 rd auto

 address-family ipv4 unicast
  route-target both auto
  route-target both auto evpn

Why RD and RT Matter

Attribute Purpose
RD Makes routes unique
RT Controls route import/export

Route Scaling Mathematics

Without RDs:

\\[ Duplicate\\ Prefixes = Routing\\ Conflict \\]

With RDs:

\\[ Unique\\ RD + Prefix = Unique\\ VPN\\ Route \\]


11. Task 8 – BGP EVPN Configuration

NX-01


router bgp 100
 router-id 11.11.11.11

 neighbor 192.168.1.2
  remote-as 100
  update-source loopback0

 address-family l2vpn evpn
  send-community extended

vrf NEXUS
 address-family ipv4 unicast

NX-02


router bgp 100
 router-id 22.22.22.22

 neighbor 192.168.1.1
  remote-as 100
  update-source loopback0

 address-family l2vpn evpn
  send-community extended

vrf NEXUS
 address-family ipv4 unicast

Why Extended Communities Are Required

EVPN relies heavily on route-target information stored inside BGP extended communities.


12. Task 9 – NVE Interface Configuration


interface nve1
 no shutdown
 source-interface loopback0
 host-reachability protocol bgp

 member vni 100101 associate-vrf

The NVE interface acts as the VXLAN tunnel endpoint.


13. Task 10 – VLAN to VNI Mapping


vlan 11
 vn-segment 10011

vlan 22
 vn-segment 10022

Mapping Logic

VLAN VNI
11 10011
22 10022

14. Task 11 – Anycast Gateway Configuration

Anycast Gateway allows hosts to use the same default gateway regardless of physical location.


fabric forwarding anycast-gateway-mac 0001.1111.1111

interface vlan11
 vrf member NEXUS
 ip address 10.11.11.254/24
 no shutdown
 fabric forwarding mode anycast-gateway

interface vlan22
 vrf member NEXUS
 ip address 10.22.22.254/24
 no shutdown
 fabric forwarding mode anycast-gateway

Why Anycast Gateway Matters

Without Anycast Gateway:

  • Hosts may require trombone routing
  • Mobility becomes difficult
  • Traffic paths become inefficient

Anycast Mathematical Advantage

Traditional centralized gateway:

\\[ Latency = Access + Core + Distribution \\]

Distributed Anycast Gateway:

\\[ Latency = Local\\ Leaf\\ Gateway \\]


15. Task 12 – Ingress Replication


interface nve1

 member vni 10011
  ingress-replication protocol bgp

 member vni 10022
  ingress-replication protocol bgp

Why Ingress Replication Replaced Multicast

Modern EVPN environments avoid multicast dependency.

Ingress replication uses BGP learned peer lists instead.


16. Task 13 – EVPN VNI Configuration


evpn

 vni 10011 l2
  rd auto
  route-target both auto

 vni 10022 l2
  rd auto
  route-target both auto

The EVPN stanza defines Layer 2 EVPN services for each VNI.


17. Task 14 – Access Port Configuration


interface ethernet1/2
 switchport mode access
 switchport access vlan 11

interface ethernet1/3
 switchport mode access
 switchport access vlan 22

These ports connect directly to host devices.


18. Task 15 – Host Configuration

SW1


ip routing

interface e0/0
 no switchport
 ip address 10.11.11.1 255.255.255.0
 no shutdown

ip route 0.0.0.0 0.0.0.0 10.11.11.254

SW2


ip routing

interface e0/0
 no switchport
 ip address 10.22.22.2 255.255.255.0
 no shutdown

ip route 0.0.0.0 0.0.0.0 10.22.22.254

SW3


ip routing

interface e0/0
 no switchport
 ip address 10.11.11.3 255.255.255.0
 no shutdown

ip route 0.0.0.0 0.0.0.0 10.11.11.254

SW4


ip routing

interface e0/0
 no switchport
 ip address 10.22.22.4 255.255.255.0
 no shutdown

ip route 0.0.0.0 0.0.0.0 10.22.22.254

19. Task 16 – Verification and Testing

Ping Tests


ping 10.11.11.3
ping 10.22.22.2
ping 10.22.22.4

MAC Verification Commands


show l2route evpn mac all
show l2route evpn mac-ip all
Sample Output

Topology    Mac Address    Prod   Flags         Seq No     Next-Hops
----------- -------------- ------ ------------- ---------- ----------------
10011       000c.29aa.bbcc BGP    --            0          192.168.1.2

20. Traditional Networks vs EVPN Fabrics

Traditional Data Center VXLAN EVPN Fabric
Spanning Tree ECMP Layer 3 Fabric
Centralized Gateway Distributed Anycast Gateway
Large Failure Domains Isolated Segments
Manual MAC Learning BGP EVPN Learning
Limited Scale Cloud-Scale Design

21. Why Modern Data Centers Use EVPN

  • Supports massive cloud-scale deployments
  • Reduces flooding
  • Improves convergence speed
  • Enables workload mobility
  • Supports automation platforms
  • Improves multi-tenancy
  • Works efficiently with Kubernetes and virtualization
Key Takeaway:

VXLAN EVPN has become the standard architecture for modern enterprise and cloud data centers because it combines scalable overlays with efficient BGP-based control-plane learning.


22. Troubleshooting Commands

Command Purpose
show nve peers Verify VXLAN peers
show nve vni Verify VNI status
show bgp l2vpn evpn summary Verify EVPN neighbors
show l2route evpn mac all Verify MAC advertisements
show l2route evpn mac-ip all Verify IP-to-MAC bindings
show ip route vrf NEXUS Verify VRF routing table

Final Thoughts

This lab demonstrates a complete migration from traditional multicast flood-and-learn VXLAN architecture to a modern BGP EVPN fabric.

Understanding these concepts is critical for engineers working with:

  • Cisco Nexus switches
  • Modern data centers
  • Private cloud deployments
  • Kubernetes networking
  • VMware NSX integrations
  • Multi-tenant environments
  • Cloud-scale networking

Modern VXLAN EVPN architectures provide significantly better scalability, visibility, operational efficiency, and convergence compared to traditional Layer 2 designs.

No comments:

Post a Comment

Featured Post

How HMT Watches Lost the Time: A Deep Dive into Disruptive Innovation Blindness in Indian Manufacturing

The Rise and Fall of HMT Watches: A Story of Brand Dominance and Disruptive Innovation Blindness The Rise and Fal...

Popular Posts