Network Addressing

IPv4 Addressing

The most widely deployed version of the internet protocol is version 4. Hosts on a
network are assigned an IP address and a subnet mask. These two addresses together
give the host a unique identifier on a network, as well as inform the host which network
it is on. By knowing which network a host is on, decisions can be made about how to
route packets within and outside of the network.

The IP Address

As you learned earlier, IP addresses are used by devices on a network to communicate

with each other. An IP address consists of four numbers, separated by periods. Each
number in an IP address is called a dotted quad. For example, the IP address for OST's
website is 199.27.144.89. Each device on a particular network must have a unique IP
address.

An IP address can be split into two parts, a network prefix and a host identifier. The host
identifier does exactly what its name suggests: it identifies a host on a network. The
network prefix identifies the network to which the IP address belongs. The split between
network prefix and host identifier can vary and is determined by the subnet mask.

The Subnet Mask

The subnet mask, like an IP address, can be written using dotted quad notation. IP
addresses and subnet masks share this notation because they are of the same length.
That's no coincidence. The subnet mask is used to "mask off" parts of the IP address
using bit masking. Bit masking is accomplished by comparing one binary number to
another "bitwise," which means that each bit in one number is compared to the
corresponding bit in the other number. Here's a table that outlines the results of various
comparisons:

Bit 1 Bit 2 Result

0 0 0

1 0 0

0 1 0

1 1 1
Next we'll look at an example of how one 8-bit number (11110000) masks off another 8
bit number (10101100):

You can see that when our second number (the mask) is a 1, the first number is
preserved as a result.

This is how the subnet masks off bits in the IP address to denote them as being part of
the network prefix. Bits that are not a part of the network prefix are part of the host
identifier. You can also write subnet masks using prefix length notation, which is a
quicker way to specify the subnet mask. The prefix length is specified as /n, where n is
the number of 1s in the subnet mask, starting from the bit farthest to the left. Below is a
diagram that demonstrates the relationship between an IP address (199.27.144.89), a
subnet mask (255.255.248.0), a prefix length notation (/21), and the resulting network
prefix and host identifier.

The number of bits left when you subtract the prefix length from 32 (the longest prefix
length possible) tells us how many hosts we can have on our network. In this example,
we have 11 bits for the host identifier, so we can have up to 2048 (211) hosts—or can
we?

The Network and Broadcast Addresses

For all networks with a subnet mask of /30 or smaller, two addresses must be reserved
from the pool of available addresses, the network address and broadcast address. This
means that in a /21 network like the one we had in our example, we can only have 2046
(211-2) hosts, not 2048. These two addresses serve a special purpose in the IPv4
protocol. The network address is used to refer to an entire network of hosts at one time.
Most commonly, this address would be used in addition to a subnet mask in order to
specify a large range of hosts for some operation, such as traffic filtering. The network
address is formed by picking the smallest address from the network.

The broadcast address also exists as a shorthand for all the hosts on a network, but
there is one important distinction. You can use the network address to refer to an entire
network, but you can't send packets to that address. However, the broadcast address
can receive packets. Packets sent to the broadcast address go to all hosts on the
specified network. The broadcast address is deerived from the largest address on the
network. Here are some examples of various network and broadcast addresses for
networks of varying sizes:

Network Broadcast Number of

Host Range Subnet Mask
Address Address Hosts

192.168.0.1 to 255.255.255.0 or
192.168.0.0 192.168.0.255 254
192.168.0.254 /24

199.27.144.1 to 255.255.248.0 or
199.27.144.0 199.27.151.255 2046
199.27.151.254 /21

255.255.255.248 or
10.0.0.9 to 10.0.0.14 10.0.0.8 10.0.0.15 6
/29

That last entry shows that the network address doesn't always end in a 0, and the
Note
broadcast address does not always end in 255.

Private Networks

Devices on a network don't necessarily need a publicly addressable IP address; they

only need to be accessible on your network, and not from the internet at large. In fact, in
many situations, for security purposes it's better if devices are not accessible directly
from the internet. Privately addressed devices can send packets to the internet
using network address translation or NAT, but NAT is beyond the scope of this course.
There are three networks that are designated as private networks. Anyone can use
these networks without obtaining permission from one of the internet's regulatory
bodies. Here's a summary:

Network Address Range Subnet Mask Number of Hosts

192.168.0.0- 255.255.0.0 or
192.168.0.0/16 65,536 addresses or 65,534 hosts
192.168.255.255 /16

172.16.0.0/12 172.16.0.0- 255.240.0.0 or 1,048,576 addresses or 1,048,574

 172.31.255.255 /12 hosts

16,777,216 addresses or
10.0.0.0/8 10.0.0.0-10.255.255.255 255.0.0.0 or /8
16,777,214 hosts

Note When configured, your machine will be a part of the "172.16.0.0/12" network.

Subnets Revisited

The subnet mask determines which network includes a particular host. This, coupled
with the fact that the length of a subnet mask is variable, allows us to take a network
block and split it into several smaller network blocks. This can help with efficient IP
address allocation on your network, as well as provide a method by which to segregate
machines according to security policies. As an example, let's use a private network,
10.0.0.0/8. With over 16 million allocatable addresses, it's a prime candidate for being
broken into smaller blocks of allocatable addresses. In fact, it's possible to have up to
65,536 /24 networks, just inside the 10.0.0.0/8 space! Let's take a look at the
differences between two hosts on distinct /24 blocks in the 10.0.0.0/8 space:

As you can see, these two hosts have the same host id, but they're on different
networks: 10.0.0.0/24 and 10.0.1.0/24. In fact, without some kind of routing in between
them, these two hosts cannot communicate with each other. Now take these two IP
addresses, but alter their subnet mask to /23. Now they are no longer two separate
networks with 256 addresses each, but instead they comprise a single network with 512
addresses. If you're still a little fuzzy on this concept, draw yourself a chart like the one
above and see how the subnet mask affects the network address and and host id for
each host.

Source: http://courses.oreillyschool.com/sysadmin2/network_addressing.html