To understand subnetting, you should first understand the decimal and binary structure of an IP address.

Let’s start with the basics. Here’s what an IP address looks like: 192.168.1.20

An IPv4 address is a 32-bit number. To make addresses more straightforward, they are divided into four 8-bit numbers — or octets — separated by a decimal point. These octets range in number from zero to 255.

Why do octets only go up to 255? Because they’re binary.

The biggest IP address possible is 255.255.255.255

In binary, this IP address looks like this: 11111111.11111111.11111111.11111111

Note that there are eight numbers between the decimal points. Each number represents a bit. Hence the term octet or the 8-bit number grouping.

Binary corresponds to this table:

1286432168421

Let’s use this binary number, for example: 10000001

Every 1 in a binary number “turns on” the number in its position. So, 1 in the first and last positions “turn on” 128 and 1.

1286432168421
10000001

Add up all the positions to get the decimal number: 128 + 1 = 129

When all the positions are “turned on,” they add up to 255.

You can see how it works here. These are the most common octets you’ll encounter in subnetting:

1286432168421
25511111111
25411111110
25211111100
24811111000
24011110000
22411100000
19211000000
12810000000

How to define the network portion of a subnet IP address

During the early stages of the internet, organizations assigned IP addresses like crazy until we nearly ran out. Luckily, the designers of IP addressing came up with a way to end this wasteful practice: Dividing networks using subnetting.

The process of taking an extensive network and splitting into smaller networks is known as subnetting — and it’s freeing up more public IPv4 addresses.

There are two parts to an IP address: The network portion and the host portion.

It’s like the address for a house. The network portion is like the city, state, and zip code. The host portion is like the house and street number.

A subnet defines the number of bits, out of 32, used for the “network portion” of the address. Subnet masks can also be defined in a more common ‘slash’ representation, known as CIDR notation. In the following table, the red digits represent the bits used for the network. The black digits will be used for device IP addresses. Note that the 255.0.0.0 mask can also be represented as a ‘/8’ because it reserves 8 bits of the overall 32 bits used to describe an IPv4 address as the network portion.

Bits used for maskDefault netmaskSubnet binary
/8255.0.0.011111111.00000000.00000000.00000000
/16255.255.0.011111111.11111111.00000000.00000000
/24255.255.255.011111111.11111111.11111111.00000000

For example, you might have a network with devices (known as hosts) with the following IP addresses:

Network: 172.16.56.0
Computer 1: 172.16.56.40
Computer 2: 172.16.56.55
Printer 1: 172.16.56.100

In this case, we’re using 24 bits (or three octets) for the network. Notice that every host device in the network has the same first three octets. That’s the network portion of the IP address with a /24 mask.

IP address: 172.16.56.40
Mask: 255.255.255.0
Binary mask: 11111111.11111111.11111111.00000000

The last octet is the host portion of the IP address. That’s where you’d assign your devices. In this case, you could assign up to 254 hosts. (More on that later.)

IP address: 172.16.56.40
Mask: 255.255.255.0
Binary mask: 11111111.11111111.11111111.00000000

Let’s look at the table again. If it were /16, then the first two octets would be the network portion, and the host portion would occupy the last two octets.

/16255.255.0.011111111.11111111.00000000.00000000

If it were an /8 network, then only the first octet would be the network portion.

/8255.0.0.011111111.00000000.00000000.00000000

These are the most common masks because they’re the simplest, but when you need more than one network, you have to subnet. Subnetting enables you to choose the number of bits to use for the Network portion. You can even steal bits from the host portion for the network.

Here’s what the full subnet mask table looks like. In this table, 1s represent the network portion, and 0s represent the host portion.

SlashNetmask1st Octet2nd Octet3rd Octet4th Octet
/30255.255.255.25211111111111111111111111111111100
/29255.255.255.24811111111111111111111111111111000
/28255.255.255.24011111111111111111111111111110000
/27255.255.255.22411111111111111111111111111100000
/26255.255.255.19211111111111111111111111111000000
/25255.255.255.12811111111111111111111111110000000
/24255.255.255.011111111111111111111111100000000
/23255.255.254.011111111111111111111111000000000
/22255.255.252.011111111111111111111110000000000
/21255.255.248.011111111111111111111100000000000
/20255.255.240.011111111111111111111000000000000
/19255.255.224.011111111111111111110000000000000
/18255.255.192.011111111111111111100000000000000
/17255.255.128.011111111111111111000000000000000
/16255.255.0.011111111111111110000000000000000
/15255.254.0.011111111111111100000000000000000
/14255.252.0.011111111111111000000000000000000
/13255.248.0.011111111111110000000000000000000
/12255.240.0.011111111111100000000000000000000
/11255.224.0.011111111111000000000000000000000
/10255.192.0.011111111110000000000000000000000
/9255.128.0.011111111100000000000000000000000
/8255.0.0.011111111000000000000000000000000
/7254.0.0.011111110000000000000000000000000
/6252.0.0.011111100000000000000000000000000
/5248.0.0.011111000000000000000000000000000
/4240.0.0.011110000000000000000000000000000
/3224.0.0.011100000000000000000000000000000
/2192.0.0.011000000000000000000000000000000
/1128.0.0.010000000000000000000000000000000

What are IP address classes?

To complicate things further, IP addresses have five classes, but only three are applicable to subnetting — A, B, C.

Here are the IP address ranges by class:

Class A = 1.0.0.0 to 127.0.0.0

Class B = 128.0.0.0 to 191.255.0.0

Class C = 192.0.0.0 to 223.255.255.0

Remember these IP addresses are represented in binary.

Here are the largest subnet IP addresses in these ranges:

Class A127.0.0.001111110000000000000000000000000
Class B191.255.0.010111111111111110000000000000000
Class C223.255.255.011011111111111111111111100000000

This is important to know because it affects the number of hosts and subnets available in a network.

Notice that Class A addresses provide the most room for host addresses (the black digits). That’s because the network portion only occupies the first octet. Most large enterprises use Class A IP addresses for this reason. You can connect more devices to a Class A network than a Class C.

Class A127.0.0.001111111000000000000000000000000

In every class, you can steal bits from the hosts to create more subnets, but you’re also reducing the number of hosts. Notice how stealing just one bit for the network drops the number of hosts significantly.

Class A Subnet Netmasks and Hosts

Network BitsSubnet MaskNumber of SubnetsNumber of Hosts
/8255.0.0.0016,777,214
/9255.128.0.008,388,606
/10255.192.0.024,194,302
/11255.224.0.062,097,150
/12255.240.0.0141,048,574
/13255.248.0.030524,286
/14255.252.0.062262,142
/15255.254.0.0126131,070
/16255.255.0.025465,534
/17255.255.128.051032,766
/18255.255.192.01,02216,382
/19255.255.224.02,0468,190
/20255.255.240.04,0944,094
/21255.255.248.08,1902,046
/22255.255.252.016,3821,022
/23255.255.254.032,766510
/24255.255.255.065,534254
/25255.255.255.128131,070126
/26255.255.255.192262,14262
/27255.255.255.224524,28630
/28255.255.255.2401,048,57414
/29255.255.255.2482,097,1506
/30255.255.255.2524,194,3022

Class B IP addresses offer fewer hosts than Class A because its network portion occupies the first two octets.

Class B191.255.0.010111111111111110000000000000000

Class B Subnet Netmasks and Hosts

Network BitsSubnet MaskNumber of SubnetsNumber of Hosts
/16255.255.0.0065,534
/17255.255.128.0032,766
/18255.255.192.0216,382
/19255.255.224.068,190
/20255.255.240.0144,094
/21255.255.248.0302,046
/22255.255.252.0621,022
/23255.255.254.0126510
/24255.255.255.0254254
/25255.255.255.128510126
/26255.255.255.1921,02262
/27255.255.255.2242,04630
/28255.255.255.2404,09414
/29255.255.255.2488,1906
/30255.255.255.25216,3822

Class C IP addresses offer the fewest hosts because the network portion occupies three octets.

Class C223.255.255.011011111111111111111111100000000

You might notice that the default IP address your home router uses falls into the Class C category. This is a special subnet reserved for private IP addresses, you can read why in the Network Address Translation article.

Class C Subnet Netmasks and Hosts

Network BitsSubnet MaskNumber of SubnetsNumber of Hosts
/24255.255.255.00254
/25255.255.255.1280126
/26255.255.255.192262
/27255.255.255.224630
/28255.255.255.2401414
/29255.255.255.248306
/30255.255.255.252622

These standards make subnetting a little easier. For example, if you choose a Class ‘C’ address, you know that it uses at least 24 bits (/24) of the 32 available bits for the network portion of the address.

How to subnet in IPv4

Now that we know about classes, binary, and subnets. Let’s dive into a subnet.

Here’s the IP address we’ll use: 43.17.255.71/27

Here’s what it looks like in binary:

00101011.000100011.111111.01000111

From the IP address we already know two things:

  • It’s a Class A IP Address
  • It must have at least eight network bits, but we’re giving it 27 bits

In that case, we know the network portion of the subnet will occupy these bits:

/27255.255.255.22411111111111111111111111111100000

Let’s reverse engineer this last octet to determine the network portion of the address or what the subnet is for this address.

Here’s what we want to do:

  1. Determine the number of allowed subnets using the /27 network mask
  2. Determine what subnet the address lies in

Here’s an example:

1. Determine number of allowed subnets using /27 network mask.

Here’s the binary representation of the possibilities for the last octet with a /27 mask:

000|0 0000001|0 0000010|0 0000011|0 0000100|0 0000101|0 0000110|0 0000111|0 0000
0326496128160192224

This gives us eight possible subnets with the /27 mask.

2. How to determine what subnet your IP address lives

Now, let’s find the subnet address where this IP address resides.

Remember that the IP address is 43.17.255.71

We are only looking at the last octet because the first three octets are the network portion.

43.17.255.71

We just have to take a look at our table again. 71 falls above the 43.17.255.64 subnet and below the 43.17.255.96 subnet. So it belongs in the 43.17.255.64 subnet.

SubnetLast OctetBlock sizeIP Address
10+3243.17.255.0
232+3243.17.255.32
364+3243.17.255.64
496+3243.17.255.96
5128+3243.17.255.128
6160+3243.17.255.160
7192+3243.17.255.192
8224+3243.17.255.224

We now have the subnet for 43.17.255.71: 43.17.255.64.