IPv4 Subnetting - The Rules

Now, that we have already learned a few things such as conversions between binary and decimal, how to recognize classes of IP addresses based on the 'first octet rule', and what is the purpose of the network mask, we can tackle IP subnetting.
A natural (default) network mask is used with class C of IP addresses quite often. But it is very uncommon to use class A and class B IP addresses with their natural netmask. They are often sub-netted (broken down into multiple smaller networks). This is accomplished by increasing the length of the default (natural) network mask.

---

 Incidently, the network IP addresses that use their natural (default) network mask are called Classful Networks.

---

But why do we create subnets to begin with?

There are many reasons why we decide to use subnets rather than classful networks. But the most important is that we want to use IP addresses efficiently since they are a scarce resource these days.

Imagine that you have a huge network to support. It uses class B network address: 172.16.0.0/16. Since the number of bits in the host portion of this address is 16 (the last two bytes are not masked), we can place 65534 hosts in a single network. Even if you used 2000 hosts still it is too much to keep them in one broadcast domain. Can you imagine that many computers sending and receiving broadcasts such as ARP requests? Well, I can imagine that, but it does not mean its efficient. In fact, broadcast traffic would pretty much kill this network. Even with thousand computers that would be way too much broadcast traffic to receive.

If we divide this huge network into multiple subnets with fewer hosts per subnet, we improve the efficiency of the system. A router will connect those subnets to allow unicast communication, but broadcasts will not be propagated between subnets as routers do not forward them. For instance: 172.16.1.0/24 subnet allows only 254 hosts in it. The broadcast will be propagated between this number of hosts rather than among one or two thousands of hosts.

Another reason for using subnets is about relates to public IP addresses that are leased to customers. ISPs do not easily give out whole classes of IP addresses (classful) to companies but rather portions of these (subnets).

Other reasons may be related to security of your hosts. Network divided into chunks with routers as gateways, give you more control as to who can 'talk' to whom.

I use terms such as broadcast or unicast. If you are not sure what these terms mean, let me present brief definitions.

Transmissions:

• Unicast - a single source host sending to a single destination host.
  Example: Src=192.168.1.1, Dst=192.168.1.2
• Broadcast - a single source host sending to all hosts in the network/subnet. Example: Src=192.168.1.1, Dst=192.168.1.255 (more on this address later in the post)
• Multicast - a single host sending to a single group of hosts (IP class D)
  Example: Src=192.168.1.1, Dst=224.10.10.10.

There are three things I would like you to remember before we delve into subnetting.

Rule 1

1. If the host bits in a given IP address are all set to '0', this is the network or subnet address.
2. If the host bits in a given IP address are all set to '1', this is the broadcast address (all hosts in the subnet/network are destination).

Rule 2

The formula used to calculate the number of available subnets given the specific length of network mask.

Pic. 1 - Number of Subnet Calculation  - Formula.

Rule 3

The formula used to calculate the number of available hosts per subnet or network given the specific network mask.

Pic. 2 - Number of Hosts Per Network/ Subnet - Formula.

Before we start using the above rules, let me show you a few examples of network, subnet and broadcast addresses based on what we have discussed in the last three posts including this one. If you do not remember the 'first octet rule', which determines the class and the default network mask of an IP address, use the following table as the reference. The number ranges of the first byte determine the classes as shown in pic. 3.

Pic. 3 - Classful Address Table.

Pic. 4 - Network (classful) Addresses and Subnet Addresses (classless).

In order to determine the number of subnet bits to use them as the exponent in the above formula (pic. 1), you must first know what is the default network mask of the IP address according to its class (pic. 3). Then, you must count the bits that were added to this default network mask. These bits allow a number of subnets to be created as per formula in pic. 1. Check out the below example.

Pic. 5 - Number of Subnet Bits (Example).

In the example (pic. 5), IP address belongs to class C since the first byte value is 192 (compare it with pic. 3). Class C uses first three bytes (24 bits) to denote the network portion of the address. Today we can say that its default network mask has the length of 24 bits (255.255.255.0). Since our network mask length is /28, we have extended the default network mask by 4 bits (bits in the green color). Thus, we get 4 subnet bits that must be used in our formula presented in pic. 1.

Pic. 6 - Number of Subnets Available - Calculation.

Using the same example: 192.168.1.0/28, how many host addresses per subnet can we use?

Pic. 5 shows us that with /28 we have 4 bits left for host (total number of bits = 32). In order to calculate the available number of host addresses we must resort to formula presented in pic. 2.

Pic. 7 - Number of Hosts Available - Calculation.

Make sure you understand how the three rules presented here work. In my next post, I'm going to show you how to use them to calculate the subnets based on different criteria such as:

• Number of subnets per IP address
• Number of hosts required in the subnet
• Number of desired host per subnet - Variable Length Subnet Masking (VLSM)