Shared NFS filesystems have a number of interesting attributes. The first "nifty attribute" is a result of NFS' stateless design. Because client access to the NFS server is stateless in nature, it's possible for the NFS server to reboot without causing client applications to crash or fail. All access to remote NFS files will simply "pause" until the server comes back online. Also, because of NFS' stateless design, NFS servers can handle large numbers of clients without any additional overhead besides that of transferring the actual file data over the network. In other words, NFS performance is dependant on the amount of NFS data being transferred over the network, rather than the number of machines that happen to be requesting said data.