What Is an IP Address?

What is an IP address?

An IP address is a number that identifies a device's network connection so that data can reach it. IP stands for Internet Protocol, the set of rules that moves data between networks. IP works at Layer 3, the Network layer of the OSI model.

Together with its subnet mask (or prefix), an IP address identifies both a network interface and the network it belongs to. For example, a laptop might have these settings:

An address belongs to an interface (a network connection), not to the whole device. A laptop with Ethernet, Wi-Fi and a VPN adapter can have several addresses at the same time.

💡 In simple terms: an IP address works like the postal address on a letter. It says which street (the network) and which house (the device). Post offices (routers) only need the street part to move the letter closer to its destination.

Why IP addresses exist: logical addressing

Every network card already has a MAC address, usually set by the manufacturer. So why is a second address needed? Because a MAC address is a physical address: it identifies the network interface, but says nothing about where it is. MAC addresses are assigned by manufacturer, not by location, so a router could never tell in which direction a MAC address is.

An IP address is a logical address. A person (or a DHCP server) chooses it to match the network the device is connected to. All devices on one network share the same network part, so a router only needs one entry in its routing table to reach all of them. That is what makes it possible to route traffic across the internet, with its billions of devices.

Both addresses work together. The IP address says where the packet is finally going; the MAC address gets it across each local link. The link between the two is ARP (Address Resolution Protocol), which you will meet in the next unit.

Where the IP address sits in a packet

When an application sends data, each layer wraps it in its own header (this is called encapsulation). IP adds a header that holds the source IP address (the sender) and the destination IP address (the final receiver). That IP packet is then carried inside an Ethernet frame, with MAC addresses on the outside.

Routers read the destination IP address to decide where to send the packet next. Normally, the IP addresses stay the same for the whole trip (unless NAT changes them), while the MAC addresses are rewritten at every router.

IPv4: understanding the numbers

IPv4 (Internet Protocol version 4) is the version that most networks still use every day. An IPv4 address has four important properties:

The 32-bit structure and octets

An IPv4 address is 32 bits, written as four decimal numbers separated by dots. Each number is an octet (8 bits), so its value can only be 0 to 255. Writing an address this way is called dotted decimal notation. It exists only for humans: 11000000101010000000101000011001 is the same address as 192.168.10.25, just much harder to read.

This means 192.168.10.300 is not a valid address: 300 doesn't fit in 8 bits. 192.168.10 isn't valid either (it has only three octets). Avoid 192.168.010.025 too: some systems read leading zeros as a different number base (octal). Always write plain numbers.

Basic binary: reading an octet

Binary is counting with only two digits, 0 and 1. In an octet, each of the 8 positions has a fixed value, doubling from right to left. A 1 means “count this value” and a 0 means “skip it”. Add up the values that have a 1:

To turn a decimal number into binary, work from the left. Does 128 fit? If yes, write 1 and subtract it; if no, write 0. Then try 64, 32 and so on, down to 1. For 25: 128 no, 64 no, 32 no, 16 yes (9 left), 8 yes (1 left), 4 no, 2 no, 1 yes → 00011001. That's all the binary you need for this unit. Splitting octets bit by bit is part of the Subnetting lesson in the CCNA course.

Learn more: Subnetting

Network part and host part

The subnet mask shows which bits are the network prefix and which are left for hosts. A mask bit of 1 means “network” and a 0 means “host”. For 192.168.10.25/24, the mask 255.255.255.0 is twenty-four 1s followed by eight 0s. So the network is 192.168.10.0/24, and the last 8 bits identify a host within it.

Learn more: Subnet Mask Basics

Here, the first three octets are the network part only because the prefix is /24. That isn't true for every address. Change the prefix below and watch the split move:

Network, usable and broadcast addresses

For the subnet 192.168.10.0/24:

The network and broadcast addresses aren't assigned to hosts. The default gateway can use any usable address; .1 is just a common convention.

Static vs dynamic IP addresses

A device can get its IP settings in two ways. Either a person types them in (a static address), or the device requests them automatically (a dynamic address). The same four settings are needed either way: IP address, subnet mask, default gateway and DNS server.

How a dynamic address is handed out

Dynamic addresses come from DHCP (Dynamic Host Configuration Protocol). In a home, the DHCP server is built into the home router. When the laptop joins the network, it broadcasts a request, and the router lends it a free address from its pool for a set time, called a lease. The printer below was configured by hand instead, with an address outside the pool, so DHCP never gives that address to another device.

💡 A middle option is a DHCP reservation: the DHCP server always gives the same address to one MAC address. The device still uses DHCP, but its address doesn't change. Many home routers call this “address reservation” or “static lease”.

Internet providers also use “static” and “dynamic” to describe your public address. That is a different topic, with its own lesson.

Learn more: Public and Private IP Addresses

Other ways addresses are assigned

In IPv6, the default gateway is learned from routers' Router Advertisements. DHCPv6 doesn't provide it.

Local destination or gateway?

Before sending, a host compares the destination with its own subnet. (More precisely, it checks its routing table; the default gateway is used when no more specific route matches.) For the laptop at 192.168.10.25/24:

On Ethernet, ARP finds the MAC address of the next device on the path. For a remote destination, that is the default gateway. So the frame goes to the gateway's MAC address, while the IP packet inside is still addressed to the remote host. Tap a step to see the addresses:

Learn more: The Default Gateway

Public and private IPv4 addresses

These ranges are reserved for private, internal networks (RFC 1918):

Any organisation can reuse these ranges, and they aren't routed on the public internet. Only the exact ranges shown are private: 172.15.0.1 and 172.32.0.1 are not. An address outside these ranges isn't automatically public either; it may be a special-purpose address (see below). And having a public address doesn't make a device reachable: routing and firewalls still decide that.

Learn more: Public and Private IP Addresses

NAT: sharing one public address

Home and office gateways usually run PAT (Port Address Translation, also called NAPT or NAT overload). Many internal addresses share one public address, and translated port numbers keep the connections apart.

NAT translates addresses; it isn't a firewall. Firewall policy decides which traffic is allowed.

Learn more: Why NAT Exists · How NAT and PAT Work

Special IPv4 addresses

Don't confuse 0.0.0.0 with 0.0.0.0/0: the second is the default route, which means “every destination”.

IPv6: a much larger address space

IPv4 has about 4.3 billion addresses (2³²), and the pool of unallocated addresses ran out years ago. IPv6 addresses are 128 bits long, written as eight groups of hexadecimal digits separated by colons:

There are two shortening rules. Drop leading zeros inside a group, and replace one run of all-zero groups with ::. You can use :: only once per address, or the address becomes ambiguous. IPv6 has unicast, multicast and anycast addresses, but no broadcast.

2001:db8::/32 is reserved for documentation. A typical IPv6 LAN is a /64, and IPv6 has no “minus two” rule.

Going further with IPv6: one lesson shows how to read, shorten and expand IPv6 addresses and what prefixes such as /48 and /64 mean. Another covers link-local and multicast addresses, SLAAC, DHCPv6 and Neighbor Discovery.

Learn more: IPv6 Addressing Basics · IPv6 Address Types and Autoconfiguration

IP addresses and ports work together

An IP address identifies an interface; a TCP or UDP port number identifies an application on it (see Port numbers and sockets).

IP itself has no port numbers; TCP and UDP carry them in their own headers.

What happens when IP addressing goes wrong

Checking a device's address

Windows

C:\> ipconfig /all
Ethernet adapter Ethernet:

   Connection-specific DNS Suffix  . : home.arpa
   Physical Address. . . . . . . . . : 02-00-00-00-00-25
   DHCP Enabled. . . . . . . . . . . : Yes
   IPv4 Address. . . . . . . . . . . : 192.168.10.25(Preferred)
   Subnet Mask . . . . . . . . . . . : 255.255.255.0
   Lease Obtained. . . . . . . . . . : Monday, 5 October 2026 09:12:40
   Lease Expires . . . . . . . . . . : Tuesday, 6 October 2026 09:12:40
   Default Gateway . . . . . . . . . : 192.168.10.1
   DHCP Server . . . . . . . . . . . : 192.168.10.1
   DNS Servers . . . . . . . . . . . : 192.168.10.1

Linux / macOS

$ ip -4 addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    inet 192.168.10.25/24 brd 192.168.10.255 scope global dynamic noprefixroute eth0
       valid_lft 86234sec preferred_lft 86234sec

To see the public address your network uses on the internet, use the What Is My IP tool. It is almost never the private address that these commands show.

Common mistakes

• Thinking the IP address belongs to the device. It belongs to an interface, and it changes when you move to another network.
• Mixing up IP and MAC addresses. An IP address is logical and used end to end; a MAC address is used only on the local link.
• Giving a device a static address from inside the DHCP pool. Sooner or later, DHCP gives the same address to another device and the two clash.
• Forgetting the subnet mask. 192.168.10.25 alone doesn't tell you which network it is on; you also need the mask or prefix.
• Using address classes to guess masks. Classes are historical; always read the configured prefix.
• Assuming the default gateway is always .1. That is only a convention. Check the configured gateway with ipconfig or ip route.

Practice: predict what happens

Next, learn which addresses are used inside homes and offices and which are used on the internet, and how the subnet mask works.

Learn more: Public and Private IP Addresses · Subnet Mask Basics