Chapter 8. Namespaces

In this chapter, we explain principles of organizing and managing namespaces. A namespace is a set of unique keys and associated attributes. Examples of namespaces are the account names in use, the printers available, the names of hosts, Ethernet addresses, service name/port number lists, and home directory location maps. Every namespace has attributes for each element. Account names have UIDs (UNIX) or SIDs (Windows), home directories, owners, and so on. Hostname attributes usually include IP addresses, hardware serial numbers, customer (owner) information, Ethernet MAC address, and so on.

The term namespace can be confusing because it can refer to either an abstract concept or a concrete thing. For example, usernames are a name-space. Every multiuser operating system has a namespace that is the list of identifiers for users. Therefore, a typical company has the abstract concept of a username namespace. However, the namespace used at one company is different from the one at another company (unless you’ve just hired all of my coworkers!). In that sense, companies have different username namespaces (your set of users and my set of users). Most of this chapter uses the term namespace to refer to a particular (concrete) namespace database. We are specific when we feel that it’s worthwhile to differentiate the two.