The Domain Name System

Subscribe
Apple | Spotify | Amazon | Player.FM | TuneIn
Castbox | Podurama | Podcast Republic | RSS | Patreon


Podcast Transcript

Behind almost every web page, email address, and podcast is a system that translates addresses understandable by humans to something that can be understood by computers. 

The system is one of the foundations of the Internet, yet its origin was in a handmade list that was placed on a single computer. 

Unbeknownst to the creators of the system, it would eventually affect the fortunes of entire countries. 

Learn more about the Domain Name System, how it originated, and how it works on this episode of Everything Everywhere Daily.


Even if you don’t know what the domain name system is, you probably use it every day, even if you don’t realize it. 

Every time you go to a website or send an email, you are using the domain name system. 

A domain name is what you type into a browser or is part of an email address. For example, my website has the domain name of Everything-Everywhere.com. 

Millions of different domain names have been registered by people, companies, governments, and organizations all around the world. 

To understand how the domain name system works, you need to understand the system underneath it, the Internet Protocol. 

The way the telephone system originally worked is that each telephone call would result in a direct connection between the two phones in a call. To achieve this, there were massive switches that would create a literal electrical connection between the phones so that the parties could talk to each other. 

While the system worked well for point-to-point voice communication, it wasn’t ideal for data. 

The first computer networks were private networks that often connected a limited number of computers using proprietary protocols. Computers in diverse locations couldn’t talk to each other both because they weren’t connected and because they didn’t use the same protocol 

An idea for a generalized system to network computers was proposed by the Advanced Research Projects Agency, or ARPA, of the United States Department of Defense. 

The proposed network would use a packet delivery system. In a packet delivery system, information is broken up into small packets of data. These small packets are then sent individually, routed through various intermediate computers, and then reassembled on the recipient computer. 

The networking system developed for the ARPA network, known as ARPAnet, was known as TCP/IP, which stands for Transmission Control Protocol/Internet Protocol. 

Under this protocol, each computer on the network had a unique numeric address. When a data packet is sent from a computer, it is sent to the recipient computer via intermediate computers called routers. 

This system worked quite well, and it was very robust and fault-tolerant. However, there was a problem. 

In order to communicate with another computer, you needed to know its IP address.

This was extremely confusing and difficult to know the address of every computer on the network. 

The solution was very similar to the solution used by the telephone system. If you needed to make a phone call to someone whose number you didn’t know, you used a directory, usually a phone book. 

For those of you too young to remember a phone book, everyone would get a book for their local area with everyone’s name, address, and phone number in it. 

You would look up someone’s name and find out their phone number so you could call them. 

The solution for ARPAnet was the equivalent of a phone book for computers. The Stanford Research Institute kept a text file of all of the computers on the ARPAnet network with their IP addresses and updated this list by hand. 

The file, called HOSTS.TXT would then be sent to various computers on the network so everyone would have a copy of the directory. 

The first directory, called the Assigned Numbers List, was created by Elizabeth Feinler, a data scientist at Stanford. It was Feinler and her team that developed the idea of domains. She believed in associating computers with the type of organization they were located it. At the time, this would have mostly been done by educational institutions, government agencies, and research organizations.

So, for example, all educational institutes would be under the .edu domain, and Stanford would be stanford.edu.

This system was fine so long as the number of computers on the network was limited. 


However, as the network expanded, keeping a list and manually sending it out to everyone on the network became extremely inefficient. By 1983, the system had become untenable. 

Proposals for a solution to the problem were submitted to Paul Mockapetris, a researcher at the University of Southern California. He rejected all of them and ended up going with his own solution which he dubbed the Domain Name System.

The Domain Name System is a hierarchical system. At the very top is what is called the root zone. If you remember my episode on the seven people who control the internet, these people have the power to reset the root zone servers. 

Below are what are known as top-level domains. Some of these are ones you are familiar with, such as .com, .net, .org, .co, .uk, and many others. 

Each top-level domain points to second-level domains registered underneath it and to the computers that host those domains. In my example, Everything-Everywhere is a second-level domain under the top-level .com domain. 

The very first top-level domain is one that you probably haven’t heard of before: .arpa. The .arpa top-level domain still exists, but it is only used for high-level network management, not for websites. 

Second-level domains can then register information about computers under their domain. For example, if you visit Wikipedia in English, you will go to en.wikipedia.org. Wikipedia has many subdomains for each language covered by Wikipedia.

Early in the days of the internet, almost every website had a subdomain that started with www for the World Wide Web. This was to distinguish their web server from their email or gopher servers. This convention has become rather dated, but you can still use www on most domains, and it will take you to the main domain. 

The hierarchical nature of the Domain Name System primarily has to do with what domain servers have authoritative control of each domain name. Despite being hierarchical, each of the domain servers replicates itself with other domain name servers. 

So, when you make a change to a domain name, it can sometimes take hours for the change to replicate across most domain name servers.

On your computer or smartphone, if you go into your network settings, you will see what domain name server, or DNS, it is pointing to. When you visit a website for the first time, your computer will query your domain name server to find out what the IP address is of the domain you are trying to reach. 

This is the one computer you can’t really use a domain name to reference because you’d need a domain name server to resolve the domain name server. You have to type in an IP address. 

Several large companies run public domain name servers with easy-to-remember IP addresses. Cloudflare has 1.1.1.1, Google has 8.8.8.8, and a company called Quad9 runs 9.9.9.9. 

There are currently over 1,500 top-level domains, even though most of us are only familiar with a few of them. When the domain system started, there were only a few top-level domains, those being .edu, .org, .net, etc. 

However, over time, the number of top-level domains has expanded dramatically. At first, the creation of a .com domain was controversial because it was thought that the internet should be free of commercial influence, which today actually seems rather cute. 

One of the things that was done early on was reserving two-letter top-level domains for countries. These domains correspond to the two-letter codes used by the International Standards Organization 3166-1 alpha-2 standard. 

Examples of this would be .us for the United States, .ca for Canada, .UK for the United Kingdom, .au for Australia, and .nz for New Zealand.

In the United States, the .us domain name is reserved for government use. Each state has its own two-letter domain reserved under the .us domain. It is not for commercial use.

In the UK, however. .co.uk is a very popular top-level domain for commercial use. 

Each country can set its own policies for what they do with their top-level domains. 

This is where it gets interesting because when top-level domains were issued to all the countries, they were issued to ALL the countries. Not just big countries but tiny countries as well, and some of them weren’t technically countries. 

One of these countries that got a domain name was Tokelau. 

Tokelau has only about 1500 people and is in the middle of the Pacific Ocean. You can’t even fly there; you can only get there via a 24-hour boat ride. However, they have the top-level domain name .tk. 

They licensed the domain name to a company, which in turn offered people the chance to register domains under the .tk domain for free. Tokelau made millions, and it became a preferred domain for spammers and scammers who wanted a free domain name. 

Other countries got lucky and wound up with a top-level domain name that had value. 

The Federated States of Micronesia was assigned a top-level domain of .FM. 

They, too, have made millions by licensing their .fm domain to radio stations around the world. 

Likewise, Armenia was assigned .am, and they have also licensed their domain to radio stations but with much less success. 

The tiny island of Nauru was assigned .nu, which has been sold because it can be pronounced the same as “new.”

Perhaps the biggest winner has been Tuvalu, a tiny country with only 10,000 people, which lucked out with its .tv domain name.  Their government currently gets 8% of its revenue from licensing from the .tv domain name. 

Sometimes, events have to transpire to make a top-level domain valuable. For years, the top-level domain for the British territory of Anguilla had very little value. However, in the last few years, their domain of .ai has found itself in high demand with artificial intelligence startups. 

Having a domain name can be quite valuable. Registering a domain name can cost anywhere from a few dollars to a few hundred dollars a month. However, extremely coveted domains can sell for an exorbitant amount of money. 

The record sale of a domain was $30 million for voice.com, which sold in 2019. 

Despite the age of the domain name system, advancements and improvements are still being made. One of the biggest recent improvements has been encrypted domain name server requests. Until relatively recently, even if you visited a secure website, the contents of the webpage might have been secure, but the domain of the website you visited was not. 

The domain name system is a core component of the Internet today. Whether you know it or not, you use it multiple times daily, including accessing this podcast. Today, the domain name system keeps track of millions of computers connected to the internet around the world. 

However, it all started with a woman in Palo Alto, California who kept track of every computer on the internet by hand.


The Executive Producer of Everything Everywhere Daily is Charles Daniel. 

The associate producers are Peter Bennett and Cameron Kieffer. 

Today’s review comes from listener W222333444 on Apple Podcasts in the United States. They write:

Great podcast!!!!!!

I listen to this podcast everyday and I am not disappointed. Me and my dad spend countless hours on my hockey roadtrips listening to this. Keep up the great work Gary. Can we get an episode on the Transatlantic Slave Trade Route?

Thanks, W222333444! I’ve touched on the Transatlantic Slave trade in several episodes, but an entire episode on the topic is certainly possible. It is a very big subject, so trying to condense it down to an episode of this podcast will require some thought. 

Remember that if you leave a review or send me a boostagram, you too can have it read on the show.