Applications use IP addresses to set up connections (TCP) and for the datagram traffic (UDP) to partner computers. IP addresses are, however, not very user-friendly, so a naming system was developed in the early days of the internet to enable important computers to be identified by name. Each computer in the TCP/IP network can be assigned one or more freely selectable name(s). The names are independent of the IP address(es) of the computer, but can be mapped to an IP address via functions of the Application Programming Interface (API).
The assignment between computer names and the IP addresses of the computers with which communication is desired is implemented via a host file:
in Unix systems: /etc/hosts
in BS2000 systems: BCAM host file
The host file contains a computer name and its associated IP address in each line.
In the early days of the internet, a central host file had to be administered manually by the Network Information Center (NIC) and copied to all computers on the internet at regular intervals. All new computers were entered into this host file by the network administrator of the NIC and were known to the other computers only after the file was redistributed via FTP.
Due to the fast growth of the internet, this procedure was soon rendered impractical, since the continuously expanding host files needed to be updated and redistributed too often. The use of database technology offered the best solution to this problem.