This guideline applies to all HPC systems managed by NUS and/or deployed in NUS managed servers. A typical HPC system is a cluster of servers that is made of up of: (a) computational nodes in self-contained, private network (hosted within NUS) not connected to any others machines outside of the cluster and (b) several others nodes that are exposed to NUS network for login and management. This scope of this document applies to (a) only.