Type a website's name into a browser, and the computer behind the scenes never actually navigates anywhere using that name. Every device on the internet is identified by a numeric IP address, and a domain name is purely a human convenience — the Domain Name System (DNS) exists specifically to translate that memorable name into the actual numeric address a computer needs, transparently, in the fraction of a second before a page even starts loading.
A lookup built to have no single point of failure
DNS works as a distributed, hierarchical system rather than one central database anywhere. When a browser needs to resolve a domain name, it typically queries a local DNS resolver, which — if it doesn't already have the answer cached from a recent lookup — works its way up a hierarchy: starting with root servers that know which servers handle top-level domains like .com or .org, then querying the appropriate top-level domain server, which points to the specific authoritative server for that particular domain, which finally returns the actual IP address. This distributed structure means no single server anywhere holds the complete mapping of every domain name on the internet, and no single point of failure can take the entire naming system down at once — a deliberate design choice given how central domain name resolution is to almost everything else on the internet functioning at all.
HTTP: the actual conversation, once the address is known
Once DNS has resolved a domain name to an IP address, the browser can actually connect to the right server, and HTTP (Hypertext Transfer Protocol) governs the conversation that follows — a request-and-response exchange where the browser requests a specific resource (a web page, an image, a file) and the server responds with that content, along with status information about whether the request succeeded. DHCP (Dynamic Host Configuration Protocol) handles a related but separate job, automatically assigning a device its own IP address when it first joins a network, so that neither DNS lookups nor HTTP requests require every device to have a fixed, manually configured address in the first place.
What we're still unsure about
The core mechanics of DNS resolution are standardised, well documented, and extremely reliable at global scale — this isn't a matter of ongoing dispute. What remains a genuine, actively managed challenge is security: DNS was originally designed in an era with far less concern for malicious interference, and techniques like DNS spoofing or cache poisoning, where an attacker tricks a resolver into returning a fraudulent IP address, remain real threats that newer security extensions (like DNSSEC) address only partially and haven't been universally adopted across the internet's full domain infrastructure, which is why DNS security remains an ongoing area of engineering effort rather than a fully solved problem.
This sits inside DNS, DHCP & HTTP, one of seven topics in Networking, one of seven domains in Computer Science, one of seventeen subjects the app can quiz you on.