1. What it is
DNS is the internet's lookup system that turns names people can remember, like example.com, into the IP addresses computers use to find each other.
2. Everyday analogy
Think about the contacts app on your phone. You don't memorize your friend's phone number. You tap "Sam," and the phone looks up the number and dials it for you. If Sam gets a new number, you update the contact once, and you keep tapping "Sam" the same as before.
DNS plays that role for the internet. You type a name. Behind the scenes, your device looks up the number that goes with it, then connects.
The analogy only goes so far. Your contacts live in one place, on your phone. DNS has no single master list. The information is spread across many servers run by different organizations, and finding an answer can mean asking several of them in turn. Updates aren't instant everywhere either, as you'll see. The next section walks through how.
3. How it actually works
Every device on the internet has an IP address that other machines use to find it, a bit like a street address for a house. Cloudflare's guide uses 192.168.1.1 as an example. DNS saves people from having to memorize numbers like that, so we can use domain names instead. DNS translates one into the other. This translation step is called resolving the name.
According to Cloudflare's guide, a full lookup with nothing saved in advance involves four kinds of servers. First, your request goes to a recursive resolver, often one run by your internet provider, though services such as Google DNS run them too. Its job is to go find the answer for you. It asks a root server, which points it toward the right top-level domain server, the one that handles names ending in ".com," for example. The top-level domain server then points to the authoritative name server for that specific domain. That last server holds the official record and returns the IP address. The resolver passes the address back to your device, and your browser can finally connect to the website.
Doing every step every time would be slow. So DNS relies on caching, which means keeping a saved copy of a recent answer. Your browser, your device, and the resolver can all keep copies. Each saved answer comes with a time to live, or TTL, set by whoever manages the domain. The TTL says how long a copy can be kept before it must be checked again. RFC 1034, the 1987 document that sets out DNS's core concepts, describes the trade-off: short TTLs mean less caching, so changes show up sooner, while longer caching saves repeated lookups.
Why build it this way? RFC 1034 explains that before DNS, every computer name and its address were kept in a single file that every computer on the network downloaded. With the number of computers growing explosively, the document says, that approach didn't bode well for the future. DNS spread the job out. Each organization manages its own part of the name system, and RFC 1034 requires each part to be available on at least two servers, so that it stays available even if a server or network link fails.
4. Real-world example
On October 4, 2021, Facebook's platforms went down in a major outage. (Later that month, the company introduced Meta as its new company brand, so this lesson calls it Meta.) The next day, the company's engineering blog explained what happened, and DNS played a central role.
According to Meta, a command run during routine maintenance accidentally took down all the connections in its backbone, the network the company built to link its data centers. Meta's DNS servers were designed to stop announcing themselves to the rest of the internet if they could not reach Meta's data centers, since Meta treats that as a sign of an unhealthy network connection. With the backbone down, they did exactly that.
The result, in Meta's words: "our DNS servers became unreachable even though they were still operational. This made it impossible for the rest of the internet to find our servers." The losses compounded, too. Meta wrote that the total loss of DNS also broke many of the internal tools its engineers would normally use to investigate, and engineers had to go on site to data centers to fix things.
The takeaway: if the lookup step fails, working servers behind it still can't be found.
Source: Engineering at Meta, "More details about the October 4 outage" (October 5, 2021): https://engineering.fb.com/2021/10/05/networking-traffic/outage-details/
5. Diagram
You type: example.com
|
v
[ Recursive resolver ] --1--> [ Root server ] "ask the .com servers"
| --2--> [ .com server ] "ask example.com's server"
| --3--> [ Authoritative ] "here is the IP address"
v
Your browser connects to that IP address
Saved copies (cache) let later lookups skip some or all of steps 1-3.6. When you'd care
DNS matters when you set up or move a website, change email providers, or switch hosting companies. In all of those cases, you are creating or changing DNS records that say where a name's website or email should go. It also matters during outages, because a DNS problem can make a working service look completely offline, as Meta's example shows.
A common mistake: expecting a DNS change to show up everywhere instantly. Because copies are cached around the internet, some people may keep seeing the old address until their saved copy expires. RFC 1034 suggests a practical fix: if you know a change is coming, lower the TTL ahead of time, make the change, and then raise it again afterward.
7. Check yourself
Q1. What is the main job of DNS?
- A) Speeding up your Wi-Fi
- B) Translating names like example.com into IP addresses
- C) Storing website images
- D) Encrypting your passwords
Show answer to question 1
Answer: B. DNS turns human-friendly names into machine-friendly IP addresses.
Q2. You updated your website's DNS record, but a friend still reaches the old site. What is the most likely reason?
- A) The internet is broken
- B) A saved (cached) copy of the old answer hasn't expired yet
- C) Your friend typed the name in capital letters
- D) DNS only updates once a year
Show answer to question 2
Answer: B. Cached answers can be kept until their time to live runs out.
Q3. In Meta's October 2021 outage, why couldn't people reach Facebook?
- A) Every server had been deleted
- B) Meta's DNS servers became unreachable, so the internet couldn't find Meta's servers
- C) Users' phones ran out of storage
- D) The domain name expired
Show answer to question 3
Answer: B. Meta wrote that its DNS servers were still running but unreachable, so no one could find its servers.
8. Sources
- Engineering at Meta, "More details about the October 4 outage" (2021): https://engineering.fb.com/2021/10/05/networking-traffic/outage-details/
- Cloudflare Learning Center, "What is DNS?": https://www.cloudflare.com/learning/dns/what-is-dns/
- RFC 1034, "Domain Names: Concepts and Facilities" (1987): https://www.rfc-editor.org/rfc/rfc1034
- Meta Newsroom, "Introducing Meta: A Social Technology Company" (October 28, 2021): https://about.fb.com/news/2021/10/facebook-company-is-now-meta/