As explained in the previous section, ParaStation MPI uses only user and group IDs for starting up remote processes. Therefore, all processes will have identical user and group IDs on all nodes.

A homogeneous user ID space is stretched across the entire cluster.