Clarifying terms of virtual networks - using kvm

Hey there,

I’m totally new to OpenNebula and at the moment I struggle with the network connectivity of my vms. In detail I’m not sure about some terms and what is meant by them. here they are:

Network manager driver
This comes first when I add a host to the OpenNebula-server. For the first time I set it up with the dummy driver for testing, but I’m not quite sure whats “driven” by this driver. Is it the network-interface connected to the same VLAN as the OpenNebula-server? In my szenario the communication between them is in a VLAN, so should I choose 802.1Q? Where in my kvm-host can I check wich driver I have to use?

When I configure a virtual network I’m asked for the so called Bridge. The help reads: name of the physical bridge in the physical host where the VM should connect it’s network interface. Per defaul a kvm-host provides a virtual bridge as described here So the VMs on the kvm-node are automatically connected to that bridge (virbr0 in that case). Is it this bridge which I have to declare at this point? Or do I have to configure a new bridge? The Network model has the same options as network manager driver above, so I have to choose the same (802.1Q), haven’t I?

VLAN (yes/no)
In the same mask there is an option VLAN which I can configure with yes or no. The description reads: wether or not to isolate this virtual network using the Virtual Network manager drivers. I’m not sure about what this means. Is it that if the kvm host is already in a VLAN and ‘yes’ is choosen, there comes a new VLAN. And if ‘no’ is choosen it will be integrated in the existin VLAN?

Physical device
Description reads: name of the physical network device that will be attached to the bridge. Is this the eth0-Interface of the VM which is connected the the virbr0?

I hope these questions are not too basic, but after reading the docs it was not that clear to my.

Best regards

1 Like

Network manager driver

  • dummy driver will use the default virbr0 interface which mean they will not connected to outside. Kind of private network.
  • If your environment use VLAN, then you need to choose 802.1Q as driver. Please take note the switch also properly configured - trunk mode, vlan to allow, etc


  • bridge name to be created. You can provide any name as long they are identical. Command is “onbr<vlan_id>” or “br<vlan_id>

VLAN (yes/no)

  • If your network driver is 802.1Q, then the VLAN = yes

Physical device

  • The physical interface where the bridge will be attached.

Hopes its clear and answer your question.

It is a hard topic, I agree. Let me do my best:

###Network manager driver

Dummy means, that OpenNebula will simply connect to a bridge that you have already created. 802.1Q means it will get the value defined in PHYDEV, let’s say eth0, it will create a new tagged interface, let’s say eth0.77, it will create a new bridge, let’s say onebr.77, and it will attach the network interface to that bridge.


For dummy it’s required, as it’s the only parameter it needs. It will connect the virtual network interface to this bridge.

For 802.1Q it’s not required, as it will be generated dynamically.

Note that it doesn’t need to be virbr0. virbr0, as created by default by libvirt is a NAT network, we recommend connect you physical interface in your host, like eth0, to br0, and using br0.

###VLAN (yes/no)

Even if using 802.1Q sometimes you don’t want to tag your virtual network interface, in case it’s a public network, for instance. In this case, when defining public networks, you want to say VLAN=NO, otherwise VLAN=YES

###Physical device

Ignored by dummy and required by 802.1Q. See the Network manager description.