Why policy domains?
Linda Gricius (April, 1998): Within a principal domain, policy domains can be used to group object types (i.e., instances of an interface) and workstations into high, medium, and low security groups. You can have different policy settings for the same interface in different domains. Some groups of very sensitive objects would have very stringent access control requirements and message protection. Other groups may require little or no security because of their insensitive nature. You can also achieve separation of domains by not defining certain target interface types in certain policy domains; e.g., if the “lowsec” domain doesn’t have the interface “missile_launch” defined, then there’s no way that clients from that domain can ever invoke objects of that type.
