What is DNS?
Useful first: Domain names · optional, not a requirement
Choose your depth
Stay with the essentials, or look closer.
YOUR GOALExplain why a website name needs to be looked up.
A name first. An address next.
Imagine visiting a friend
You know your friend’s name, but you still need an address to visit. Opening a website starts with a similar step: you type a name like example.com, and your device needs an address to connect to.
DNS, short for Domain Name System, helps look up that address. The kind of address used on a network is called an IP address.
[1]example.comFind address192.0.2.1A name goes in. An address comes back.
Simplified illustration, not a live lookup. The address is an example.Finding the address is only the first step
Once you have your friend’s address, you still need to make the trip. Likewise, after the lookup, your browser still needs to contact the website and ask for the page.
DNS is spread across many computers. It can also return more than one address for a name, so the friend comparison is a starting point rather than an exact model.
[1]DNS helps connect a name people remember with information computers need.
DNS sends you the website itself.
DNS supplies lookup information. Fetching the page is a separate job.
Show the explanation
The browser must request the page. The lookup gives the browser an address to contact. The browser then requests the content; visiting a site does not register its name.
Progress is saved only in this browser.
YOUR GOALDistinguish the resolver, the answer, and the website request.
Follow a lookup, one step at a time.
The parts involved
Your device usually asks a recursive resolver to find DNS information on its behalf. This lookup service can answer from saved information or ask other DNS servers. An authoritative server publishes the records for a managed part of DNS, called a zone.
The distinction is useful: a resolver helps you find an answer; an authoritative server is a source of the published records.
[2]A simplified lookup
- You open
example.com. - The device or resolver may already have a usable saved answer, called a cached answer.
- If it needs more information, the recursive resolver asks DNS servers. With no useful cached referrals, this can involve the root, the
.comservers, and an authoritative server for the name. - The resolver returns an answer. The browser can then connect and request the page.
This is a teaching sequence, not a claim that every visit makes every query.
[2]More than addresses
DNS stores different kinds of records. An A record supplies an IPv4 address. An MX record describes where email for a domain should be delivered. The question includes a record type, not only a name.
A resolver finds DNS information; an authoritative server publishes it.
Every page visit repeats the entire DNS lookup.
Usable cached answers and referrals can avoid some or all of those queries.
Show the explanation
An authoritative DNS server. Publishing a zone’s DNS records is the authoritative server’s role. A recursive resolver finds and saves answers; serving web pages is a separate role, even if one machine performs both.
Progress is saved only in this browser.
YOUR GOALRead a DNS response as typed, time-limited information—not a permanent mapping.
Records, caching, and the meaning of an answer.
Start with the question
A resolver is software that looks up DNS information for an application. A DNS question specifies a name, a type, and a class. For ordinary internet DNS the class is generally IN. An A record carries an IPv4 address; a CNAME record names an alias target, so a lookup may need another step.
IPv6 and multiple answers
An AAAA record carries an IPv6 address. An application may request IPv4 and IPv6 information and receive more than one usable address. DNS does not guarantee a one-name-to-one-computer relationship.
TTL normally limits reuse
A record’s time to live (TTL) sets its normal cache reuse lifetime. It does not schedule a worldwide refresh, and a cache can discard an answer sooner.
A resolver receives an answer with a 300-second TTL at 12:00. At 12:01, the authoritative record changes. The resolver can normally keep answering with its saved record until 12:05. Lowering the authoritative TTL at 12:01 does not change the countdown already running in that cache.
During refresh failures: a resolver using serve-stale can retain expired data and answer with it under RFC 8767. TTL is the normal reuse limit, not an absolute deletion deadline.
[3] [6]Not all missing answers mean the same thing
NXDOMAIN says the queried name does not exist. A name can instead exist but have no record of the requested type. DNS can cache negative answers too; the DNS negative-caching specification defines how those answers are represented and aged.
Ask which record was requested, who answered, and whether a cache can still reuse it.
Changing a DNS record updates every resolver immediately.
Cached answers do not update instantly. TTL normally limits reuse; serve-stale can permit an expired answer during refresh failures.
Show the explanation
There is no IPv6 address record in that answer. AAAA describes an IPv6 address. Its absence says nothing by itself about A records or whether a web server is reachable. A name with no record of that type is different from an NXDOMAIN answer.
Progress is saved only in this browser.
Keep asking. Check the details.
Source-linked, AI-assisted draft. Human editorial review is still pending; this page is not presented as expert-reviewed.
- RFC 1034 — DNS concepts and facilities
- RFC 9499 — DNS terminology (foundational terminology reference)
- RFC 1035 — DNS records, messages, and TTL
- RFC 3596 — DNS extensions for IPv6
- RFC 2308 — Negative caching
- RFC 8767 — Serving stale DNS data during refresh failures