«When you send a email, you expect the system to know how to deliver it where it needs to go – that is, to one of the 200 million domains – and all in all, it works».

With this sentence, Paul Mockapetris explain in a very simple way the importance of the Domain Name System, or DNS, conceived by him in the 1980s, which still helps us connect computers and other devices worldwide today.

Without DNS, the Internet would not have had the opportunity to expand to the levels it has reached today. Let's explore in more detail what it is and how Mockapetris and colleagues had the right intuition to design it.

What is DNS?

As we mentioned, DNS is the "Domain Name System": translate that is, internet addresses to which we are accustomed (such as www.assoprovider.it) in IP addresses (Internet Protocol) used by computers and other devices such as smartphones and tablets. An IP address consists of a string of numbers, such as 192.168.1.1. This "code name" allows different online devices to connect and communicate with each other.

Without DNS, essentially, we would have to enter (and especially remember) that string of numbers every time to access a website or contact someone online. 

It's a bit like a phone directory: we probably remember a tiny percentage of the phone numbers we have in our directory, and without a correspondence between the names of friends and family and their contacts, we would have a hard time calling them. DNS servers manage precisely this aspect, mapping names (domain names) and numbers (IP addresses) and correlating them with each other.

Internet before DNS

At this point, a question naturally arises: In the early days, computers communicated with each other using basic protocols over telephone lines. These connections were often established using modems, which converted digital data into analog signals that could be transmitted over the existing telephone infrastructure. The communication was typically point-to-point, meaning that a direct connection was established between two computers. Protocols such as X.25 and later TCP/IP were developed to facilitate data exchange and ensure reliable communication. As technology advanced, dedicated data lines and more sophisticated networking equipment, such as routers and switches, were introduced, allowing for more complex and efficient communication networks. In the early 1980s, ARPAnet still existed, the government network that was the precursor to the Internet, and to have an "address" (or rather a nome host) with whom you had to communicate call the NIC, Network Information Center, present at the research institute of Stanford University in California. Which, of course, was not open 24/7, but only when staff were in the office: therefore no later than 6 PM (Pacific Time) and never during holidays.

As Mockapetris himself explains:

«If you wanted to add a machine to the network, you had to call Stanford and speak with the Network Information Center to receive a name and an address.».

The NIC had a single text file, named "host.text", which lists all the addresses of the ARPA network, along with the corresponding connected machines. A sort of early Internet white pages. The file was then distributed to all the machines in the network to allow everyone to communicate.

The problems of this system are twofold. 

The first is that to contact another computer you have to look up its address in your directory and then type the numbers on the keyboard. The second is that every day it is necessary copy "manually" all new addresses entered on the network.

The people working at ARPAnet are fully aware of the issue. Jon Postel, head of the research department at the Southern California University (USC) who worked on the first message in history sent via network, one day he summoned Mockapetris to his office, to discuss the issue with him. He wants Mockapetris to find a solution and proposes five different possible paths to follow. The researcher accepts the job but ignores all of Postel's instructions.

Paul Mockapetris' intuition

Mockapetris understands that the centralized system of NIC is now outdated: it is necessary to create a network distributed across different servers, in various locations, to achieve the desired result.

This is the first step: from that moment, those who want to communicate on the network do not necessarily have to wait for Stanford's opening hours, but have greater flexibility. 

The system then ensures that each device on the network has its own "identity", an easily identifiable and memorable domain name, which then translates into an IP address to which data packets sent from another computer are routed.

«To do anything on the Internet – explains Mockapetris – you must be able to tell people who you are: the DNS does just that. Additionally, it allows you to consult anyone who has entered their data on the DNS».

In essence, the system devised by the researcher allows all network members to independently enter and change their parameters, ensuring that this data is visible to everyone. 

In 1983 the new system is ready. After a testing phase of several months, Mockapetris, Post, and Craig Partridge, another pioneer who worked on the DNS, published the idea in a memorandum for the Request for Comments (RFC) in November 1983. 

Almost a year later, in October 1984, the first generic domains were officially established, which we all still know and use today: .com, .edu, .gov, .mil, .net e .org. In 1986, DNS became one of the original Internet Standards, established by the IETF, the Task Force that laid the foundations for the network we all know today.

The first DNS update came three years after its creation, and although the system has evolved significantly over the decades, it still allows millions of people to communicate with each other today.

Read also: The inventors of the TCP/IP protocol, the story of the meeting between Vint Cerf and Robert Kahn