A head node is a node on a hadoop cluster that runs the (head|lead) services that manages the worker nodes.
For example, the following process would be installed on a head node: