IP address hierarchy

The Internet Protocol (IP) describes the use of IP addresses to identify Internet-connected devices. IP addresses have a hierarchy that makes it easier to route data around the Internet.

Many addressing schemes are hierarchical. Consider a US phone number:

We can break that into $4$‍ parts:

The hierarchy makes it easier for the telephone system to efficiently send calls to the right lines.

Both IPv4 and IPv6 addresses are hierarchical. For simplicity, let's examine the hierarchy of IP v4 addresses.

Consider this IP address:

The first sequence of bits identifies the network and the final bits identify the individual node in the network.

That IP address could break down into these $2$‍ parts:

The first two octets ($16$‍ bits) identifies a network administered by the Comcast (an Internet Service Provider). The last two octets (the final $16$‍ bits) identifies a home computer on that Comcast network.

If the last two octets were different, then the IP address would point at a different computer on the Comcast network. If the first two octets were different, then the IP address might belong to a completely different network administrator.

The Internet Protocol uses this hierarchical addressing scheme to make it easier to route messages from source to destination. Once a message arrives at the network, a network router can take care of sending it to the individual node. The next lesson on routing dives into more details on how that works.

Network administrators can break IP addresses into further subnetworks (subnets) as needed.

Starting with this IP address:

That could break down into 3 parts:

The first two octets identify the entire network for the University of Michigan, the third octet identifies the UMich Medicine department's network, and the fourth octet identifies an individual lab computer in that department's network.

Adding further levels to the address hierarchy can improve the efficiency of routing within the network.

In actuality, IP addresses are often split in the middle of the octets.

To understand how that works, let's represent the previous IP address in binary instead:

All together, that translates into these 32 bits:

The first $16$‍ bits could route to all of UMich, the next $2$‍ bits could route to a specific UMich department, and the final $14$‍ bits could route to individual computers.

This hierarchy gives UMich the ability to differentiate between $2^2$‍ ($4$‍) departments and $2^{14}$‍ ($\mathrm{16,384}$‍) computers within each department.

Splitting octets might seem confusing at first, but computers store the IP addresses as binary anyways, so it is all the same to them.

As we've just seen, the ability to create hierarchical levels at any point in the IP address allows for greater flexibility in the size of each level of the hierarchy.

🙋🏽🙋🏻‍♀️🙋🏿‍♂️Do you have any questions about this topic? We'd love to answer—just ask in the questions area below!