Background
To understand the requirement of DNS server, we have to go back to initial stages of networking of the computers around late 1960’s and early 1970’s when only single computer is used in the university due to cost and other factors.
When computers started increasing exponentially in universities then people are forced to share documents/printers service etc of one system to other systems for maximizing the resources. For this purpose, inter networking between the systems is required in labs so there will be effective utilization of resources attached to system. People started to remember the servers with numbers.
Initially remembering the numbers was not difficult, but when computers started growing faster their was a need to give names to the servers instead of numbers. (As remembering names is much easier than numbers).
Evolution of Hosts File
To make it easy to communicate with different systems a new concept ‘hosts file’ is evolved. Every system in the lab should have hosts file and it should contain name to number conversation table.

For Example: one lab is having 10 machines each machine should have 10 entries for these servers as show below.

Hosts file content:

Labsys1 1
Labsys2 2
Labsys3 3
..
..
Labsys10 10

So, it becomes easy for users to remember it.

Disadvantages of Hosts file
One disadvantage from this hosts file is, if we want to add 11th machine in the lab all the hosts files on each system should be modified to add this host details.
It is perfectly fine to make the changes, if there are just 10 machines but consider a situation where there are 100’s or thousands of servers in a lab. If a new machine is added to the network, the machine entry should be added to all the machines which is a cumbersome and time consuming process for the admin.

Solution of Hosts file
Go for a centralized monitoring/configuration system where every machine will point their Name to a centralized server. This machine should serve all the queries. At this point DNS (Domain Name Server/service) came in to existence.

Disadvantage of initial DNS server
When DNS server is implemented again it’s a mapping between numbers to host names and vice versa as shown below.

For Berkeley servers:
Server1 123
Lab2 345
Pas23 89

If you see there is no proper naming convention and its bit difficult to find where one server is located in given University.

FQDN is the solution
Researcher came with FQDN (Fully Qualified Domain Name) convention so that’s it’s easy to remember and easy to point where the server located.

FQDN is a hierarchical and tree like structure for naming convention which contain following details:

1) Hostname –Name of the server
2) Domain name – Defines a realm administrative authority or control on a group of machines.
3) Top Level Domains (TLD) – ie .com, .in, co.uk, co.us etc.