It sometimes amazes me that people get so worked up about subnetting, because it really is quite simple. First of all, you need to recognize that in order to really understand subnetting (at least starting off), looking at the numbers in decimal notation makes very little sense. You need to be looking at numbers in binary to really understand what is happening. The beauty of binary numbering is its simplicity – each value can only be a 1 or a 0. Note that each section (octet) of an IP address can be represented by a series of eight bits. There are 4 octets, so 32 bits altogether. That means any IP address can be also looked at as a 32-bit binary number. The table below outlines binary numbering corresponding values.

Decimal 128 64 32 16 8 4 2 1

Binary 1 1 1 1 1 1 1 1

What this means is simple. If I were to ask for the value of 11001100 in decimal, it would be 128+64+0+0+8+4+0+0, which equals 204. Each bit corresponds to the decimal value above it – add the values for each ‘1’ value and you have the answer. 11111111 would be 128+64+32+16+8+4+2+1, which equals 255 (which is also the highest possible decimal value in an 8-bit binary number).

But what about converting decimal numbers to binary? Well, it’s different, but no more difficult. Start at the left on the chart above, and add the decimal values together until you reach your total. Every number you use is a ‘1’ and every number you leave out is a ‘0’. For example, let’s take the number 77. This would be 01001101. Say what? Well, I just started adding numbers left to right, leaving out numbers that put me over 77. In this example, I have 0+64+0+0+8+4+0+1. Simple.

You can also do this using a calculator program with a scientific mode. Just type is a number in decimal and hit the BIN button. The number will then be displayed in binary. However, the calculator has no idea that you’re dealing in 8-bit numbers, so you’ll have to be careful. For example, my calculator will tell me that 77 in binary is 1001101. That is, it leaves off any leading zeros. As such, you’ll need to remember to ‘pad out’ your binary numbers to 8 bits if you use the calculator. For example, the calculator will show decimal 8 as binary 1000. For an IP address, we need to add the 4 other zeros, making it 00001000. You’ll have access to the calculator on the exam, so know how to use it.

After you understand binary numbering, subnetting is easy. First of all, we need to discuss what subnetting is. Quite simply, it is taking a big network ID and breaking it down into a number of smaller networks, or subnets. Routers are what usually separate subnets. Reasons for subnetting include connecting different topologies (such as Ethernet and Token ring), as well as making networks smaller and more manageable. Subnets are also sometimes referred to as broadcast domains, since a broadcast sent on a subnet goes to all hosts on that subnet

For the purpose of any exam, you will need to recognize and understand how subnetting works. This includes being able to view system configurations and determine why clients are having trouble communicating. As such, you’ll need to be able to recognize valid IP addresses, subnet mask values, and what range of IP addresses are valid on a given subnet. Let’s start with a look at valid subnet mask values.

A subnet mask means little in decimal. In binary, however, they tell a story. The subnet mask is what tells us which of the 32-bits in an IP address represent the network identification, and which represent the host identification. In the example below, the host IP address is 156.77.11.3 and the subnet mask is /21, or 255.255.248.0. In decimal, it is difficult to determine which portion represents the network and which the host. However, it binary the mask value is:

11111111 11111111 11111000 00000000

So what does that tell me? That the first 21 bits are used to represent the network, and the last 11 bits are used to represent a host on the network. Actually, it tells me more than that. It also tells me how many hosts I can have per network. How? Well, if eleven bits are used to represent a host, then this subnet can have 2046 hosts. How did I get that? Simple: 2 to the power of 11, minus 2. That equals 2048 minus 2, or 2046. Why minus 2? You subtract 2 because a host value of all binary 0’s represents the subnet, and a value of all binary 1’s is the broadcast address for this subnet.

If the subnet mask in the example above had been /17, or 255.255.128.0, that would leave 15 bits for host addresses. That would mean 2 to the power of 15 minus 2 hosts, or 32766 total.

Figuring that stuff out should now be easy enough as well. The big question, and the key thing you need to be able to do, is to be able to determine if a host ID is valid on a subnet. Every subnet has a range of addresses that are valid on it. In my last example, there were 32766 valid host addresses. You need to be able to determine which ones are valid for the subnet. It isn’t that hard, but you need to know what you’re looking for.

Let’s say that we’ve been given an address of 156.17.42.6/20, and we’re trying to determine the range of valid host IDs on this subnet. The first step is to determine the actual network ID on which this host falls. The process we use to determine this is called ANDing. When we want to AND an IP address and subnet mask, we first convert them to binary and line the subnet mask below the IP address. Then, calculate the AND value. In an AND operation, values are calculated as follows:

1 and 1 = 1

1 and 0 = 0

0 and 0 = 0

In our example, this would give us:

IP 10011100 00010001 00101010 00000110

SM11111111 11111111 11110000 00000000

AND 10011100 00010001 00100000 00000000

After we convert our ANDed address back to decimal we get 156.17.32.0. This is the network ID that our host falls onto.

Stay with me here. We know that our mask is 255.255.240.0 (or /20). So, we know that the last 12 bits represent the hosts on this network. The network bits are in black below, the host bits in red. We already know that a host ID cannot be all zeros or all ones in binary. So, when I’m calculating the range of valid IPs on this subnet/network, I can’t have either of these values. This leaves me with:

Network ID 10011100 00010001 00100000 00000000

First Valid Host ID 10011100 00010001 00100000 00000001

Last Valid Host ID 10011100 00010001 00101111 11111110

Note that the first valid host ID sets all host bits to zero except the last (called the least-significant bit), and the last valid host ID sets all host bits to one, except the last. What did I lose? Two addresses – the host ID being all zeros (which defines the network) and the host ID being all ones (the broadcast address, which is not valid for a host). These are the same 2 addresses that I subtract when trying to find how many hosts I can have per subnet. If I convert my ranges above to decimal, I end up with a range of:

156.17.32.1 to 156.17.47.254

The truth of the matter is that you won’t necessarily have time to ‘do the math’ for every question that comes at you during the exam, so you’ll need a way to quickly determine what ranges of hosts are valid on a subnet given a certain mask. For this purpose, I am providing the chart below. You can use this chart to quickly determine the valid ranges of IP addresses on a subnet based on the mask value, and where the next range starts. Please do not use this chart as a crutch if you don’t understand how to determine valid ranges as we went through above. This is meant as a shortcut for those who already understand.

Mask 128 192 224 240 248 252 254 255

Network ID 128 64 32 16 8421

How the chart works is simple. Let’s say I’ve been given a host ID of 167.23.87.13 with a mask of 255.255.248.0, and I want to quickly determine the range of host IP addresses valid on the same subnet as this host. This address is subnetted into the third octet based on the mask, so we take the third octet value (248) and plug it into the chart above. The Network value that corresponds to 248 is 8. As such, that means that every new subnet starts at a multiple of 8 in the third octet. For example:

167.23.0.0 subnet0 range = 167.23.0.1 to 167.23.7.254 *

167.23.8.0 subnet1 range = 167.23.8.1 to 167.23.15.254

167.23.16.0 subnet2 range = 167.23.16.1 to 167.23.23.254

167.23.24.0 subnet 3 range = 167.23.24.1 to 167.23.31.254

167.23.32.0 subnet 4 range = 167.23.32.1 to 167.23.39.254

…

167.23.80.0 subnet10 range = 167.23.80.1 to 167.23.87.254

…

167.23.240.0 subnet30 range = 167.23.240.1 to 167.23.247.254

167.23.248.0 subnet31 range = 167.23.248.1 to 167.23.255.254 *

* Although these ranges were usually omitted in a classful IP addressing system, they are totally valid under CIDR. Often these ranges are still omitted, however, due to the fact that some older equipment may not reference the ranges properly.

Note that our host is on subnet10, the range in red above. The same rules as always still apply, so be careful. The host ID cannot be all 0’s or 1’s. As another example, if the address had been 17.13.5.1/14, the subnet mask would be 255.252.0.0, making the range of addresses on the same subnet as this host everything on subnet 17.12.0.0, since new ranges start in multiples of 4. That would make the valid range:

17.12.0.1 to 17.15.255.254

If you go back to the ANDing process, and calculate the first and last host IDs in binary, you’ll see that we’ve come up with the same answer, only much more quickly!

As I mentioned from the outset, this section was not meant to be a complete explanation of designing a subnetting scheme for a network. Instead, we learned how to define valid ranges of addresses based on a host ID and mask value, both in binary and using the shortcut method. You will need to be able to troubleshoot IP addressing, and that’s what I’ve focused on above. Once you can calculate valid ranges, you can then determine which host IDs are local and remote, and which hosts are capable of communicating properly. Only hosts that fall into the same range should be on the same subnet. You also now know that the problem may be the address or the subnet mask values of the hosts in question.