If you've been trying to wrap
your mind around the idea of performing the subnet
masking using an IP address and a subnet mask,
you may be saying to yourself, why are we going through
this entire process? Why do we have to
subnet the network into all of these tiny
different IP subnets? Part of the problem is that we
can't connect the entire world directly to each other. There's just not enough
resources and bandwidth available to do that. Only one device can talk to
another device at a time, and we don't have
enough technology to be able to connect
together billions of devices so that they were all on
the same local network. Having these
separate subnets also provides some segmentation
and perhaps the ability to add additional security. We can add a
firewall and segment out a particular IP subnet
that had servers and that might have a different security
posture than an IP subnet that had workstations
or printers on it. As we saw in some of
our earlier videos, using our class-based subnet
mask was very inefficient. We weren't able to customize
the exact size of the network, and we ended up with a lot
of leftover IP addresses that wouldn't be
used for anything. By giving the network
administrator the ability to customize the size
of the subnet mask, they could build out a
particular subnetting scheme that provided the exact
number of networks and hosts that they needed
in their network. Instead of using these strict
class-based subnet masks, we instead commonly use
variable length subnet masks on our network. We define this as VLSM. For example, if you had
the network 10.0.0.0/8-- that would be a traditional
class A network-- we could use variable
length subnet masks to have many different
kinds of networks, all with different masks. For example, you
might have 10.0.1.0. That network has a
24-bit subnet mask. And you might also
have a 10.0.8.0 network with a 26-bit subset mask. Those differences in the
subnet masks and the ability to design our
network in a way that makes sense for what we're
doing, rather than designing it around limitations of an IP
address or, what we call, variable length subnet masks. Let's see how we could
take a traditional class A, subnet mask and
customize it so it would be perfect for our network. We'll start with the
network 10.0.0.0. Its traditional class A,
subnet mask is 255.0.0.0. You'll sometimes see
this even referred to as Classful addressing,
because it's based on that class A, subset mask. If we were to write out
the subnet mask in binary, you would have eight 1s at
the beginning and all 0s after that. In decimal, the subnet
mask, then, obviously, is 255.0.0.0 or
a /8 subnet mask. Based on the subnet
mask, then, we have 8 bits that are dedicated
as the network address and 24 bits that are
dedicated as the host address. But it would be very
unusual to have millions of hosts on a single subnet. Instead, let's borrow
some of those bits, to be able to make the number
of networks larger and perhaps the number of hosts on a
network a little bit smaller. So instead of applying
all 24 of these bits, let's back out a few of those
and leave 8 bits at the end. If we do that, then
we have a network that is still 8 bits long. We've borrowed 16
of those bits that were left in the host
part of the address, and we will use those
for different subnets. And then we have 8 bits
left over for the host that we would use on each
one of those subnets. Since we've moved the bar a
bit and added in additional networks, our subnet mask has
changed to 255.255.255.0 or /24 network. Since we're not using the
traditional class-based subheading and instead
are using our own subnet that we're defining, we refer
to this as Classless addressing. Since we know how many bits
are part of the network side of the IP address and how
many bits are part of the host side, we can very
quickly calculate the total number of
subnets available and the total number
of hosts available, by using powers of 2. We wrote a chart like this in an
earlier video, where we took 2 to the first power, 2
to the second power, 2 to the third
power, all the way through to the eighth power. And of course, you
could continue this all the way through to the
16th, and even further out. You would, of
course, start with 2 to the first power,
which is a 2. You would double that
to 4, double it to 8, double it to 16, and so on. So you could create a very
easy reference chart to use, when you're trying to perform
some of these subnetting tasks. To use this chart, we would look
at the number of subnet bits that we've borrowed
and perform 2 to that value to determine
the total number of subnets that might be available to us. Then, to determine the number
of hosts available per subnet, we would use 2 to
the host bits power, and then subtract 2 from that. One of those will be
the subnet address, and the other will be
the broadcast address. Everything left is
the available number of hosts on each
individual subnet. So let's use those charts to
be able to quickly determine what these values might
be on an IP subnet. We've got our number
of subnets here, which is 2 the subnet bits
power, and hosts per subnet, which is 2 to the host
bits power minus 2. And here at the bottom, we
have our powers of 2 chart. Lets use the IP
address 10.1.1.0/24. If we were to write out the
/24 subnet mask, we have 24 1s, and then eight 0s at the end. We know that, because this
IP address starts with a 10, that this is
traditionally a Class A. So we start with those first
eight bits that are associated with a Class A, subnet mask. We are then going to move
our bar down to 24 bits, giving us 16 bits in the middle
that we could use to calculate different subnets from that. And of course, we have
the 8 bits at the end that we'll use to assign
to different devices on each individual IP subnet. To determine, then,
how many total subnets we could create from
this particular mask, we'll look at these 16 bits,
and we'll use 2 the 16 power. And if we look at
our chart, that is 65,536 possible
networks that you could create from this subnetting. Now that we know how many
networks we can create, let's determine how many
hosts that we can have on each individual network. We've got 8 bits available
to use as hosts bits, and our formula is 2 to
the host bits minus 2, so 2 to the eighth minus 2. If we look at 2 to the eighth
is 256, we subtract 2 from that, then the total number
of hosts that we could have on any of
these individual subnets is 254 maximum hosts. Let's perform the
same calculation with a different subnet and
see what the results might be. In this case, we'll
take 192.168.11.0/26. The /26 means that we have
26 1s in the subnet mask, and that leaves us with six 0s
at the end of the subnet mask. If we were to look at this
address that starts with a 192, that means the traditional
class-based subnet would give us 24 bits to be
able to use for the network. This, obviously, is 26 bits
long for the subnet mask, which gives us two
additional bits at the end to use for subnetting. We've gotten, then,
6 bits at the end that we can use to
calculate host values for each individual subnet. Using our powers
of 2 chart, we can see that we've got two
subnet bits available. So 2 to the second
power is 4, which means, from this subnetting, we can
create four individual IP subnets from that /26. We have 6 bits available for
hosts, and 2 to the sixth power is 64. We subtract two of those
for the subnet address and the broadcast
address, leaving us with 62 devices per IP subnet. Let's use this same process
to calculate another subnet. In this case, we
use 172.16.55.0/21. The /21 at the end means that
we've got 21 subnet bits that are set, and the rest
of the bits would be 0. We know that the 172 at the
beginning of this IP address means that this
would traditionally be a Class B address,
and the subnet mask would be 16 bits long. That means we have 5 bits
to use as subnetting bits, and the rest of the bits
would be 11 bits leftover, to be used for hosts. If we perform our
calculation for the subnet, we've got 5 bits available. So 2 to the fifth power is 32. We could create 32
individual networks from this particular
subnet mask. And we've got 11 host
available per subnet. And performing
those calculations would be 2 to the
11, or 2,048 minus 2, leaving us 2,046
devices per subnet.
