IP Addresses Explained

What an IP address actually is

An IP address is a return address. When your browser asks a server for a page, the request has to carry somewhere for the reply to go, and that somewhere is your IP. It is not optional and it is not a tracking mechanism bolted on afterwards — it is the fundamental addressing scheme that makes the internet work at all. Every site you visit sees it because it must.

What it is not is a personal identifier. An IP identifies a connection, not a person. In a typical home, one public address is shared by every phone, laptop, television and doorbell in the building, and on many mobile and business networks it is shared by thousands of unrelated customers at once.

You can see yours, along with what a site can infer from it, on the what is my IP tool — which also lists the far larger amount of information your browser gives away without being asked.

Public and private addresses

There are two different kinds of address in play at any moment, and confusing them is the source of most misunderstandings.

Your public IP is the one the outside world sees: a single address representing your entire connection. Your private IP is what your router assigns to each device inside your network, and it comes from ranges reserved for exactly this purpose — 192.168.x.x, 10.x.x.x, and 172.16 through 172.31. Those ranges are not routable on the public internet, which is why millions of homes can all use 192.168.1.1 for their router without any conflict.

The mechanism joining the two is NAT, network address translation. Your router rewrites outgoing packets to come from the public address, keeps a table of which internal device asked for what, and rewrites the replies back on the way in. A side effect is a rudimentary firewall: unsolicited traffic from outside has no entry in that table and nowhere to go, which is why devices behind a home router are not directly reachable from the internet.

That last row is worth recognising. An address starting 169.254 means your device asked for one and got no answer, so it invented its own. It is the signature of a broken DHCP setup, and seeing it explains an otherwise mystifying loss of connectivity.

IPv4, IPv6, and why the change took decades

IPv4 addresses are 32 bits, which allows about 4.3 billion of them. That seemed limitless in the early 1980s and was clearly insufficient by the mid-1990s. The last large blocks were handed out by the central registry in 2011, and the regional registries have been rationing ever since.

IPv6 uses 128 bits, giving roughly 340 undecillion addresses — enough that the practical constraint disappears entirely. The written form changes from four dotted numbers to eight groups of hexadecimal separated by colons, with runs of zeros collapsible to ::, which is why IPv6 addresses look so unfamiliar.

The reason the migration has taken so long is that the two protocols are not interoperable: an IPv6-only host cannot talk to an IPv4-only host without a translator. So the world has run both in parallel for years, a strategy called dual-stack, and NAT relieved enough of the pressure that the urgency drained away. Adoption is now past a third of traffic globally and still climbing.

One consequence matters for privacy. Because IPv6 gives every device its own globally routable address, the NAT layer that incidentally grouped a whole household behind one number goes away. Privacy extensions, which rotate the device portion of the address periodically, exist specifically to limit the resulting per-device trackability.

Why geolocation is approximate

IP geolocation is not a lookup of where you are. It is a lookup in a database of where address blocks were registered, assembled from registry records, ISP-published routing data and inference. Nothing in the protocol carries a location.

In practice this means city-level accuracy at best, and often much worse. An ISP may register a block to its regional headquarters, so every customer for a hundred miles appears at the same point. Mobile carriers route traffic through a small number of gateways, so a phone in one city routinely appears in another. Corporate VPNs put remote workers wherever the office is.

There is a well-documented and genuinely harmful failure mode here. Databases that cannot resolve an address to a specific place sometimes return a default point — historically the geographic centre of a country. One such default landed on a single rural property in Kansas, whose residents spent years being visited by people and authorities convinced their address was the source of internet crimes. It is a good reminder that these figures are estimates presented with false precision.

For the same reason, a mismatch between the location shown and where you actually are is normal and not a sign that anything is wrong.

What a VPN actually changes

A VPN builds an encrypted tunnel from your device to a server it operates, and your traffic reaches the internet from there. Sites see the VPN server's address rather than yours, and your ISP sees encrypted traffic to the VPN rather than the sites you visit.

That is a real and useful change, and it is where the benefits stop. A VPN does not log you out of accounts, remove cookies, prevent browser fingerprinting, or hide anything from a site you sign in to. If you visit a site while logged in, the IP is irrelevant to identifying you.

It also relocates trust rather than removing it. Your ISP can no longer see your browsing; your VPN provider now can. Choosing a VPN is choosing whom to trust with that visibility, which is why the provider's jurisdiction, funding model and logging policy matter far more than connection speed.

The thing that actually identifies you

Focus on IP addresses distracts from a considerably more effective tracking method. Every page you open can read a long list of properties from your browser without asking permission: operating system, browser version, screen resolution, colour depth, time zone, installed fonts, language preferences, processor count, and how your device renders a test image on the graphics card.

Individually none of these is identifying. Combined, they very often are. The set of people running one specific browser version, at one screen size, in one time zone, with one language list and one core count can be extremely small — sometimes a single person. This is browser fingerprinting, and unlike an IP it survives changing networks, and unlike a cookie it survives clearing your browser data.

The counterintuitive consequence is that unusual configurations make you easier to identify, not harder. Disabling common features, installing rare fonts or spoofing an obscure user agent all make your fingerprint more distinctive. Browsers that resist fingerprinting effectively do so by making many users look identical rather than by making each one unique.

The IP checker lists what your browser reveals precisely so you can see this for yourself — none of it required a request, and all of it is available to every site you open.

Static, dynamic, and why yours keeps changing

Most residential connections use a dynamic address: your ISP assigns one from a pool when your router connects, and it may change when the router restarts, when the lease expires, or whenever the provider reorganises its pool. Some people keep the same address for months; others see it change weekly.

The reason is economics. Address space is scarce, and not every customer is online at once, so a pool smaller than the customer base serves everyone. It also means an IP is a poor long-term identifier — an address that belonged to you last month may belong to a stranger today, which is exactly why IP-based evidence is treated cautiously.

A static address stays fixed and normally costs extra, because it consumes a scarce resource permanently. It is genuinely needed for hosting a public server, for allow-listing that must survive a reboot, or for running mail. For most people the workaround is dynamic DNS: a small client tells a DNS provider whenever your address changes, so a fixed hostname always points at your current one.

There is a further layer worth knowing about. Many mobile carriers and some fixed-line ISPs now use carrier-grade NAT, putting hundreds or thousands of customers behind a single public address. If your public IP looks like it is in a private-ish range and no inbound connections work no matter what you forward, this is usually why — and the only fix is asking the ISP for a real address.

When you actually need to know your IP

Beyond curiosity, there are a handful of genuinely practical reasons to look it up. Allow-listing is the most common: office networks, database hosts and admin panels often restrict access to specific addresses, and you need your current one to be added.

Checking a VPN is working is the second. Connect, reload, and confirm the address and country have changed — if they have not, the tunnel is not carrying your traffic, and you may be leaking.

Diagnosing connection problems is the third. An address in 169.254 means DHCP failed. A private address showing as your public one means something is proxying oddly. A different country than expected usually means a VPN or corporate network you forgot was on.

And hosting anything — a game server, a home lab, remote desktop — requires knowing your public address, and quickly teaches you why a dynamic one is inconvenient and why dynamic DNS services exist.

Common myths, briefly

"Someone can find my house from my IP." No. Public databases give a city or an ISP hub. Only your provider holds the subscriber record, and that is released to law enforcement under legal process, not to anyone who asks.

"Hiding my IP makes me anonymous." No. Accounts, cookies and fingerprinting identify you regardless. Changing your IP changes one weak signal among several much stronger ones.

"My IP is a virus risk." Not on its own. Knowing an address lets someone attempt a connection, but a home router's NAT means there is nothing listening to accept it unless you deliberately forwarded a port. The risk lives in what you run and expose, not in the number itself.

"Incognito mode hides my IP." No. Private browsing stops your browser saving history and cookies locally. It changes nothing about what the network or the sites you visit can see.

A sensible mental model

Treat your IP as roughly equivalent to the return address on an envelope. It tells the recipient which post office to reply through and gives a rough idea of the region. It does not name you, it is shared with everyone else in the building, and it changes when you move.

If your goal is privacy, spend your effort on the things that actually identify you: accounts you are logged into, cookies and trackers, and browser fingerprinting. A VPN is a reasonable layer on top of those, particularly on untrusted networks, but it solves a narrower problem than its marketing suggests.

Check yours on the what is my IP tool, and if you are troubleshooting a connection rather than a privacy question, the speed test and connection monitor are the tools that will tell you whether the problem is your line or the site.