Transferring Data Over a Network

Usually, the purpose of a file transfer is to allow the use of a file on a local system. When accessing files over a TCP/IP network, two methods exist: file transfer and file sharing. The difference between these two methods is defined by how the file is handled. During a file transfer, the file is physically copied from one network location to another. With file sharing, a remote file system is treated as a local file system and the file is not physically moved. When files are transferred, you work with a copy of the original file. You work with the original file itself with file sharing.

Basically, file transfer over TCP/IP is the same as a file transfer across a serial line or a modem and the protocols discussed earlier can be used. However, these protocols can't take advantage of the features of TCP/IP. TCP/IP-specific protocols provide reliable connections between network nodes, removing much of the burden from the transfer protocol. Because the above protocols are designed around the limitations of serial and modem connections, they can't achieve very high transfer rates over TCP/IP.

To achieve high transfer rates over TCP/IP, use File Transfer Protocol (FTP). FTP is designed to use the error-checking capabilities of TCP/IP instead of duplicating the effort. To use FTP you need two components: an FTP client and an FTP daemon. The client resides on your local system and initiates the transfer. The daemon resides on your remote system and serves requests from FTP clients.

The other method for accessing files across TCP/IP is Network File System (NFS). Like FTP, NFS uses the error-checking capabilities of TCP/IP to reliably access remote files and requires a client and a daemon. Unlike FTP, NFS turns your remote file system into a local file system.

For more information, see Transfer Protocol advantages and disadvantages .