[RESOLVED] Getting to Know Subnetting

[G_Hosa_Phat]

Hello all!

Okay, so I've been learning way more than I wanted to ever know about subnetting in the last couple of weeks as I'm trying to implement some (hopefully) interesting security in a new application I'm developing, but I've still got a couple of questions to which I haven't been able to find "acceptable" answers (translation: "I still don't get it"). I know that network design, subnetting, routing and all that fun stuff is a world all it's own, but I'm wanting to better understand it to get an idea of how to best get the details I need for my application.

Now, I have a fairly decent understanding of the basics of CIDR; and the whole "Class A", "Class B" and "Class C" thing is pretty easy to get a handle on; I can even figure out the subnet mask of a CIDR notated network address without a cheat sheet (if I really have to). I understand the use of binary and how it relates to IP addresses and subnet masks, and I understand how a subnet mask relates to an IP address to identify the available host, network and broadcast addresses for a network.

These are the things I "do" get. What I "don't" get, however is some of the basic calculations to come up with (mostly) classful available subnets, and some of the real differences between using CIDR and classful subnetting. I know a network designated as "Class C" should have the following attributes:

• IP addresses that fall in the range of 192.0.0.0 - 223.255.255.255
• A "default" subnet mask of 255.255.255.0
• A starting number of hosts (based on the "default" subnet mask) of 254 (256 total IP's - 2 reserved)

I also understand that, in a /24 CIDR block, the only real difference between the classful network and the CIDR network would basically be the possible IP addresses. Within this CIDR block, the 256 total IP addresses could be virtually anything from 0.0.0.0 through 255.255.255.255 (with notable exceptions, of course), although they must be contiguous (i.e., 192.168.1.0 - 192.168.1.255).

So, here are my basic questions so far:

1. After hours of looking at online documentation and subnet calculators, why is it that, when getting the number of subnets for a CIDR network with a 255.255.255.192 subnet mask, some return 64 and some return 62? I've seen talk about reserved ranges within a CIDR block, but that doesn't necessarily make a lot of sense to me. I get the reserved IP addresses, but not reserved ranges.

2. Please help me to define these terms used by the different subnet calculations - the maximum number of subnets in a CIDR network based on the subnet mask is the number of subnets that can be created "underneath" that subnet (i.e., you can further subnet the CIDR network with a mask of 255.255.255.192 down to a maximum of 64 smaller networks), while the maximum number of subnets in a classful network based on the subnet mask is the number of times you can use that same subnet mask within the scope of the class (i.e., you can have 4 subnets with a mask of 255.255.255.192 within a Class C network, each with a different IP range)?

3. If you subnet a classful network, is it possible to go another level deeper and subnet that subnet? Is there a reason to do that?

I think that's all my questions for now. I know these may be "no-brainers" to some extent, but I'm just having trouble wrapping my head around it all. I hope I've made sense here as my head is swimming a little bit with all of the stuff I've learned about subnetting up to now.

Oh, one more thing: Is it "acceptable" to use CIDR notation for a classful network?

[dbasnett]

For the most part there is no such thing as a classful network, except in the case of routing protocols that have rules based on classes. Classful routing protocols include RIPv1 and IGRP, and classless routing protocols include RIPv2, EIGRP, OSPF, BGP and IS-IS.

1. Subnetting is the act of taking a given network i.e. 192.168.0.0 /16 and creating smaller networks within it.

Here is is a sample subnet scheme for that network:

    Network           Net Broadcast     CIDR Mask              UsableHosts 
    192.168.0.0       192.168.31.255    19   255.255.224.0     8,190
    192.168.32.0      192.168.63.255    19   255.255.224.0     8,190
    192.168.64.0      192.168.95.255    19   255.255.224.0     8,190
    192.168.96.0      192.168.127.255   19   255.255.224.0     8,190
    192.168.128.0     192.168.159.255   19   255.255.224.0     8,190
    192.168.160.0     192.168.191.255   19   255.255.224.0     8,190
    192.168.192.0     192.168.223.255   19   255.255.224.0     8,190
    192.168.224.0     192.168.239.255   20   255.255.240.0     4,094
    192.168.240.0     192.168.255.255   20   255.255.240.0     4,094

There are not reserved ranges within a block.

2. The smallest network that can be subnetted is a /29. Here is 192.168.254.248 /29 subnetted so you can see why.

    Network           Net Broadcast     CIDR Mask              UsableHosts 
    192.168.254.248   192.168.254.251   30   255.255.255.252   2
    192.168.254.252   192.168.254.255   30   255.255.255.252   2

a /26 (the mask you provided) can't be subnetted into 64 smaller networks. It only has 64 addresses, 62 usable.

3. See number 1. Normally the network should be sized according to the number devices within it, plus room for growth. With the abundance of private IP's most people opt for the lazy approach of assigning networks on octet boundaries. If you ever work with public IP's the reason for subnetting becomes clear.

4. Oh, one more thing: Is it "acceptable" to use CIDR notation for a classful network? Sure, why not?