Dell EMC Networking offers static Virtual Extensible LAN (VXLAN) as part of the Dell EMC Smart Fabric OS10. Static VXLAN allows customers to statically configure virtual tunnel endpoints (VTEPs) for a given traffic flow. A static VXLAN tunnels the packets through an underlying Layer 3 IP network by encapsulating the Layer 2 packet with a VXLAN header at the source VTEP and de-encapsulating at the destination VTEP. 

This section details the topology used in this configuration cheat sheet. The cheat sheet uses a Layer 3 leaf-spine topology for the network underlay.

The physical topology (underlay) provides transit for the virtual network overlays. Dell EMC Networking S5232F-ON switches are used on the spine layer with Dell EMC Networking S5248F-ON switches used on the leaf layer. In rack 1 and rack 2, both leaf switches are configured as Virtual Link Trunking (VLT) pairs. A VLT provides redundancy and load balancing between the two VLT peers, multihoming support, and a loop-free Layer 2 network without running STP. As administrators add racks to the data center, two leaf switches configured for VLT can be added to each new rack. Connections within racks from hosts to leaf switches are Layer 2, and each host is connected using a VLT port channel.

OSPF is used for communication between the leaf and spine establishes a logical, scalable IP address scheme, which is required before deploying a leaf-spine topology. Loopback addresses may be used as router IDs when configuring routing protocols. Loopback addresses should follow a logical pattern that will make it easier for administrators to manage the network and permit growth. In the following example, all loopback addresses are part of the 10.0.0.0/8 address space with each address using a 32-bit mask. The third octet represents Layer 1 for the spine and Layer 2 for leaf. The fourth octet is the counter for the appropriate layer. For example, 10.2.1.1/32 is the first spine switch in the topology while 10.2.2.2/32 is the second leaf switch.

Figure 1.       Leaf-Spine Underlay Topology

Figure 2.       VLT topology

Establishing a logical, scalable IP address scheme is important before deploying a leaf-spine topology. The point-to-point links used in this deployment are labeled A-H in Figure 3.

Figure 3.       Point-to-point networks

Each link is a separate, point-to-point IP network. The following table details the links labeled in Figure 3. The IP addresses in the table are used in the switch configuration examples.

Note: As with all examples in this guide, any valid IP address scheme can be used. The example point-to-point addresses above use a 31-bit mask to save address space. This is optional and covered in RFC 3021

In this deployment example, VNIs 1614 and 1615. All VNIs are configured on each of the four leaf switches. Because VNIs 1614 and 1615 are configured with anycast gateways, VMs on those VNIs which route to other networks can use the same gateway information while behind different leaf pairs. This enables VMs to migrate from one leaf pair to another without changing the network configuration. It also eliminates hair-pinning and improves link utilization since routing is performed much closer to the source.

Figure 4.       Leaf-spine overlay topology

Configure a loopback Interface on all four-leaf switches and use that interface as the VXLAN source tunnel interface. 

Create a Network Virtualization Edge (NVE) instance. Parameters configured at this level are common to all overlay encapsulation types and all virtual-networks unless overridden at a more specific level. Configure the NVE source interface – the IP address at this interface is used as a source IP address in all VXLAN packet headers emitted from this device.

The local NVE Source IP should be advertised into the underlay IP network to all peer NVEs. Each NVE should have route installed to reach each of the other peer NVE IPs in the network. In the following example, OSPF in the underlay to advertise and learn route reachability to the peer NVEs. Configure the underlay OSPF protocol as follows:

Configure VXLAN Virtual Networks with a static VTEP. The remote VTEP IP address configured here is the NVE source IP address that is configured on each of the remote VTEP NVEs that participates in the same tenant segment in the overlay network. 

Note: Only IPv4 network addresses are supported. 

BUM traffic in this virtual-network replicated to each remote VTEP configured under the VXLAN tunnel end-point. Each replicated packet carries the VNI assigned to the tunnel end-point in the VXLAN header. SmartFabric OS10 supports replication up to 1024 peer NVE IP addresses.   

Using the following commands, create a virtual network bridge domain and attach a VXLAN tunnel end-point to the bridge domain using a specific VNI. 

Note: Only one tunnel end-point of a given NVO type can be attached to each virtual network. 

The commands in the following example configure access ports as VLAN members for switch-scoped VLAN-to-VNI mapping. Packets received for these VLANs on any of these trunk or access ports are assigned to the virtual networking for bridging. If the packet is bridged over a VXLAN tunnel, then the packet VLAN tag is stripped and encapsulated (with VXLAN VNI) before transmitting out on the egress interface for the tunnel. At the remote receiving end, the packet is assigned to a virtual network for bridging based on VXLAN VNI and the VXLAN header is stripped off.  After the bridging on the egress interface is identified, the VLAN tag is regenerated based on the VLAN ID associated with the egress interface for this virtual network.  

Enter the commands below to configure the upstream network-facing ports. The following configurations show the NVE source IP address is advertised using OSPF.

Configure the VLT port-channel and VLTi member links using the following commands: 

Enter the following commands to configure the VLTi. The IP address used in the backup destination command is the management IP address of the other switch in VLT pair. 

Using the following commands, configure a tenant VRF, anycast Layer 3 gateway, MAC address, and anycast gateway IP address for each virtual network. 

Configure the Underlay OSPF on spine switches as follows. 

Configure downstream ports on underlay links to leaf switches. 

This section will walk you through the relevant commands to validate the network configuration. 

Step 1

Use the show ip ospf routes command to display the routes calculated by OSPF in the underlay network. In the following example, we used this command on Leaf1 in Rack 1. It shows that Leaf 1 can reach the destination VTEP (VTEP 10.222.222.2) via Interface ethernet 1/1/53 and ethernet 1/1/56. 

Leaf1# show ip ospf routes
Prefix              Cost           Nexthop        Interface           Area           
------------------------------------------------------------------------------
10.222.222.1/32     0              0.0.0.0                            0.0.0.0
10.222.222.2/32     3              192.168.1.0    ethernet1/1/53      0.0.0.0      
                    3              192.168.2.0    ethernet1/1/56      0.0.0.0        

Step 2

Use the show ip ospf topology command to display the routers directly connected in the same OSPF area. 

Leaf1# show ip ospf topology
Router ID           Flags          Cost           Nexthop        Interface          
--------------------------------------------------------------------------------
10.2.1.1            -/-/-/         1              192.168.1.0    ethernet1/1/53      
10.2.1.2            -/-/-/         1              192.168.2.0    ethernet1/1/56      
10.2.2.2            -/-/-/         2              192.168.2.0    ethernet1/1/56     
10.2.2.2            -/-/-/         2              192.168.1.0    ethernet1/1/53      
10.2.2.3            -/-/-/         2              192.168.1.0    ethernet1/1/53     
10.2.2.3            -/-/-/         2              192.168.2.0    ethernet1/1/56      
10.2.2.4            -/-/-/         2              192.168.2.0    ethernet1/1/56      
10.2.2.4            -/-/-/         2              192.168.1.0    ethernet1/1/53

Step 3

To validate the connectivity to the remote virtual tunnel endpoint (VTEP), we used show nve remote-vtep command on Leaf1. For example, the following output indicates that the VXLAN tunnel to the remote VTEP (10.222.222.2) on Rack 2 (Leaf3 & Leaf4) is up. 

Leaf1# show nve remote-vtep
IP Address: 10.222.222.2,  State: up, Encap: VxLAN
   VNI list: 1614-1615

Step 4

To further validate the connectivity to the destination VTEP, a ping test can be performed from Leaf1 to the destination VTEP IP address in Leaf 3 and Leaf4 . 

Leaf1# ping 10.222.222.2 -c 2
PING 10.222.222.2 (10.222.222.2) 56(84) bytes of data.
64 bytes from 10.222.222.2: icmp_seq=1 ttl=63 time=0.351 ms
64 bytes from 10.222.222.2: icmp_seq=2 ttl=63 time=0.217 ms
--- 10.222.222.2 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1024ms