Virtual NICs on a VPS: Adding Interfaces, Bridges, and VLANs

This guide shows you how to add a second virtual NIC to a Linux VPS, configure it persistently, and set up bridges and VLANs for multi-network setups. You need root access over SSH and a provider control panel that lets you attach additional network interfaces. Most KVM and OpenVZ plans support extra NICs; the exact panel path differs by provider, but the Linux-side configuration is identical everywhere. If you are still shopping for a host, the provider guides on virtualserversvps.com cover private-network options, and compare VPS providers on our comparison table to shortlist hosts with the interfaces you need.

1. Confirm what the hypervisor exposes

Start by listing the interfaces Linux currently sees:

ip -br link

A typical KVM VPS shows lo and eth0 (or ens3). Names like ens4 mean the provider uses virtio drivers. Note the MAC address of each interface; you will need to match it if the panel asks you to bind the new NIC to the instance.

eth0/ens3: primary interface with your public IPv4 and IPv6
lo: loopback, ignore it
If no second interface exists yet, you must add it in the control panel first

2. Attach the second NIC in the provider panel

In most panels (Vultr, DigitalOcean, Hetzner, and smaller providers alike) this lives under “Network” or “Interfaces”. Create a new interface, optionally attach it to a private network or VLAN, then reboot the instance or hot-plug it. Virtio NICs are hot-pluggable, so after a few seconds:

sleep 5 && ip -br link

If the new interface appears as eth1 or ens4, continue. If not, reboot once so the kernel re-enumerates PCI devices. Verify the driver with ethtool -i eth1; a virtio driver with a current kernel version is expected on any modern provider.

3. Configure the interface persistently

On Debian/Ubuntu with netplan, add a block for the new interface in /etc/netplan/01-netcfg.yaml:

network:
  ethernets:
    eth1:
      dhcp4: true
      dhcp6: true

Then apply and confirm an address was assigned:

sudo netplan apply && ip -br addr show eth1

On systems using ifupdown (/etc/network/interfaces), append the equivalent:

auto eth1
iface eth1 inet dhcp

If the provider assigned a static private IP, use address and gateway lines instead of DHCP and skip the gateway on the second NIC — a VPS normally keeps a single default route on the primary interface.

4. Bridges for VM and container traffic

If you run KVM guests or LXC containers on the VPS, bridge the new NIC so guests share it:

sudo ip link add name br1 type bridge
sudo ip link set eth1 master br1
sudo ip link set br1 up

Persist the bridge in netplan by declaring br1 as a bridge with eth1 as a member, then re-apply. Guest traffic then exits through the second NIC, keeping host traffic on eth0 and isolating the two planes — useful when you run database replicas or CI runners that generate heavy internal traffic.

5. VLANs with iproute2

To trunk VLANs over the new NIC, create VLAN sub-interfaces:

sudo ip link add link eth1 name eth1.100 type vlan id 100
sudo ip link set eth1.100 up
sudo ip addr add 10.0.100.2/24 dev eth1.100

Add the same lines to netplan or the interfaces file so the VLAN survives reboots, then verify with ip -d link show eth1.100. The VLAN tag is only relevant if your provider actually delivers tagged traffic; on a plain private network you do not need VLANs at all.

6. Verify performance and routing

Confirm the new path works end to end:

ping -I eth1 10.0.100.1
ip route get 10.0.100.1

For throughput, run iperf3 between two VPS instances on the same private network:

# on the receiving VPS
iperf3 -s

# on the sending VPS
iperf3 -c 10.0.100.2

On a 1 Gbps private link you should see 900+ Mbit/s with default settings; anything under 500 Mbit/s points to an MTU mismatch or an outdated virtio driver. Check MTU parity on both ends (1500 by default, 1450 inside a VPN overlay) with ip link show.

7. When a second NIC is worth it

A dedicated private-network NIC removes contention with public traffic, which matters for database replication, offsite backups, and internal API calls between instances. For a single small VPS, one interface is usually enough — spend the effort on disk I/O and kernel tuning instead. Before provisioning extra resources, compare VPS providers on our comparison table; most providers there support private networks on standard plans.

If you need a budget VPS with multiple NICs and a private network for testing, InterServer VPS plans start at a few dollars a month and include full root access.