Following guides assume you’ve understood some core concepts and know some terms, all reported below. It’s strongly suggested to read this carefully before moving forward.


The word POD or CORD POD refers to a single physical deployment of CORD.

Full POD

A full CORD POD is a typical CORD configuration, also taken as example in this installation guide. A full CORD POD is composed by three servers, and four fabric switches. It gives the possibility to experiment all the core features of CORD and it’s what usually the community uses for tests.

Half POD

A half CORD POD is the minimum-sized configuration. It is similar to a full POD, but with less hardware. It’s composed by two servers (one head node and one compute node), and one fabric switch. It doesn’t allow experimentation with all of the core features that CORD can offer, but it’s still good for basic experimentation and testing.

Development / management node (reported as dev node / machine)

This is the machine that an operator uses to download, build and deploy CORD to a POD. Sometimes it is a server, sometime the laptop of the deployer. In principle it can be any machine, with any operating system, as long as it can satisfy the hardware and software requirements reported in the sections below.

Development (dev) VM

Bootstrapping the CORD installation requires lots of software to be installed and some non-trivial configurations to be applied. All this should mainly happen on the development / management node. To help users with the process, CORD provides an easy way to create on the dev node a VM with all the software required and the configurations in place. This guide and the community refers to it as the dev VM.


Head node


A head node is essentially a compute node (see below), that also acts as the brain of the POD. It runs XOS (the orchestrator), two instances of ONOS (the SDN controller, one to control the underlay fabric, one to control the overlay), MAAS and all the services needed to automatically install and configure the rest of the POD devices.


Compute node(s)


A compute node is a machine able to run VMs or containers of one or more tenants. The terminology has been borrowed from OpenStack, used by CORD.


Fabric switch


CORD uses an underlay fabric to connect other elements inside the POD, and the POD itself with other PODs.



The virtual Subscriber Gateway (vSG), is the CORD version of the present day CPE. The vSG runs a bundle of subscriber-selected functions, like Restricted Access, Parental Control, Bandwidth Metering, Access Diagnostics and Firewall. These functionalities run on commodity hardware located in the Central Office rather than on the customer’s premises. There is still a device in the home (which we still refer to as the CPE), but it has been reduced to a bare-metal switch.