It's possible to configure and use the Network File System without too much knowledge of how it is implemented or why various design decisions were made. But if you need to debug problems, or analyze patterns of NFS usage to suggest performance optimizations, you will need to know more about the inside workings of the NFS protocol and the daemons that implement it. With an understanding of how and why NFS does the things it does, you can more readily determine why it is broken or slow — probably the two most common complaints in any large NFS network.
Like NIS, NFS is implemented as a set of RPC procedures that use eXternal Data Representation (XDR) encoding to pass arguments between client and server. A filesystem mounted using NFS provides two levels of transparency:
The filesystem appears to be resident on a disk attached to the local system, and all of the filesystem entries — files and directories — are viewed the same way, whether local or remote. NFS hides the location of the file on the network.
NFS-mounted filesystems contain no information about the file server from which they are mounted. The NFS file server may be of a different architecture or running an entirely different operating system with a radically different filesystem structure. For example, a Sun machine running Solaris can mount an NFS filesystem from a Windows NT system or an IBM MVS mainframe, using NFS server implementations for each of these systems. NFS ...