Why do IP Addresses Have A Class System?

[Pyrotechnika]

I've been learning about IP addresses in an online class and I'm trying to figure out why they are separated into classes.

 

I'm just guessing it is a prioritization thing, it would be rad to know about the application of it too!

[Teshy09]

There are 5 address classes ranging from 0.0.0.0 - 240.0.0.0 

Have a flick through: http://www.tutorialspoint.com/ipv4/ipv4_address_classes.htm

and shout if you need anything more

[iamMINT]

Okay well, it has to do with your subnet.

 

C - 255.255.255.0/24

B - 255.255.0.0/16

A - 255.0.0.0/8

 

All these really are is information that defines your network. Like for your C class Private IP that might be something like 192.168.1.1, the last part is what defines your host. Like 192.168.1.164 or something, all the way up to 255. That leaves a total possible connected devices of 254 (because 192.168.1.255 is reserved for broadcast)

 

It kind of works like this. In binary, 255.255.255.0 is 1111111 11111111 11111111 0000000. All those ones belong to your network and your network defines them, but all those zeroes define the host sending the information. Same going for 255.255.0.0. It'd be 11111111 11111111 00000000 00000000. This is because the last two octets are host bits. Which means they're defined as the host.

 

What does all this mean? Well, basically, it allows your network to be split up and subnetted. That is why it's split up into three classes, with those octets. For the first subnet, where you had just the single zero on the end, a device could have any number between 1 and 254. Which means you could have a total of two hundred and fifty four devices on that network. Though an enterprise or business might need more, so you have class B that has a possible of 64 thousand connections. That's because it has a subnet of 255.255.0.0. Which means you can have anywhere from 128.1.0.1-128.1.255.254 on your network. Class a, which is 255.0.0.0 has up to 16 million possible connections.

 

A and B are mostly for where you have to split up and subnet a network, where you'll be loosing quite a few possible ip addresses from borrowing bits and things like that. Though, at that point it gets really complicated and almost redundant unless you're making business grade networks. All that really matters is the scale to which the IP can be used, and all the possible subnets and IP's inside of that network.

[Pyrotechnika]

Sweet! That helps my understanding alot, thanks guys! =)

[manikyath]

the main reason of the idea of subnetting is being able to split the pool of available addresses into multiple pools which can be allocated to specific purposes.

 

for example: your ISP owns a class A subnet (2^24-2 = ~16mil host addresses) they are free to allocate to what they want: their web servers, their DNS, their custommers, etc.

it's pretty common practisce to split these pools up even more, for example when you have a class B subnet(65k addresses) , but want those seperated in pools that cannot directly communicate with one another (for example in a school have a subnet for the remote management of equipment, a subnet for the teachers, a subnet for the students, and a subnet for the staff) they can divide that class B up into 4 more subnets.

 

a drawn out example for the school:

their local network runs off the 172.16.0.0 subnet (class B subnet for local usage), which would normally have a subnetmask of 255.255.0.0.

 

now, because they dont want those 65K of addresses to be in one pool they split them up by use of the subnet mask:

255.255.192.0

which. when looking to binary:

172.16.0.0

1010 1100.0001 0000.0000 0000.0000 0000

255.255.192.0

1111 1111.1111 1111.1100 0000.0000 0000

those two extra ones we have at the end of the subnet indicate that those two numbers are now also part of the "network" part of the IP.

result: these are the 4 subnets we just created out a single subnet:

172.16.0.0 - 172.16.63.255

172.16.64.0 - 172.16.127.255

172.16.128.0 - 172.16.191.255

172.16.192.0 - 172.16.255.255

for each of these the 172.16.x.0 is the "network address" (easy to remember: the last spot in the address can never be zero)

and the 172.16.x.255 is the "broadcast address" (what packets get sent to if every device on the network needs to know about it.)

each of these subnets being able to provide 16382 addresses, and the subnetmask the devices get either trough configuring manually or trough DHCP tell them how many bits are the network part, and how many bits are the host part, in turn also from that knowing the broadcast address.

 

as a result they have 4 pools of addresses which provide a very distinct purpose, so things keep in order and seperated where necessary.

[KuJoe]

Nobody uses "classes" anymore, CIDR is where it's at.

[Linx105]

The class system went away once people started to realized IPV4 did not have enough addresses. Part of the reason CIDR became the standard was in an effort to use the wasted blocks the class system was not taking advantage of as well as simplifying how we group network blocks.

[leadeater]

If you aren't going to be tested on it, why should you, just erase this knowledge as it's totally not applicable now days.

 

You should be learning CIDR-VLSM, bit octet conversions, network and broadcast address calculations.

https://supportforums.cisco.com/sites/default/files/legacy/6/6/3/15371366-Classful vs Classless Routing.pdf

https://ciscoskills.net/2011/06/01/classful-vs-classless-routing/

https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing