| NIS is a commonly used naming service that was originally known as Yellow Pages (YP), but had to be renamed to avoid conflict with the UK telephone company British Telecommunications. In fact, we can see from the manual page listing that many of the commands still refer to the older naming convention (ypbind, ypcat, etc.). NIS uses a client-server concept similar to many other naming services. This approach ensures that we don't get data replication problems because the "real" data resides only on one main server. Through the use of slaves, NIS provides server redundancy to ensure that the data is still available if the master server fails. So, how does NIS work? First, we define something known as a NIS domain. This is really just a token name to group a number of machines together that require access to the same set of data files. One of the machines within the domain will be nominated as the master serverthis is the one that controls the data files. (It's probably worthwhile pointing out now that a machine can be a server for multiple domains with a little configuration.) Next, although it's not a requirement, we normally create some slave servers that can be used to provide the data even if the master fails, thus achieving a level of server redundancy. Finally, all the machines, including the master and slave servers, are added as clients of the domain. This may sound odd at first that a server also has to become a client, but hopefully it will make sense soon. We can see from this general description that three different types of machines are used within NIS: master server, slave server, and client. We'll look at all of these in more detail shortly. | 
