VLSM

This might help you out.

http://www.microsoft.com/TechNet/deploy/tcpintro1.asp

You answered the first part of your question
It is one of three classes that determines if the destination address is located on the local subnet or a remote network.
Class Octet Range Default Subnet Mask
A 1 TO 126 255.0.0.0
B 128 TO 191 255.255.0
C 192 TO 223 255.255.255.0
Subnet mask can be automatically configured by DHCP.
Reference knowledgebase www.microsoft.com/TCP/IP/Subnet Mask

The desire of knowledge, like the thirst of riches, increases ever with the acquisition of it.

* Laurence Sterne

Don't mind Conrad...he's what they call a "vulture," one of several members who exploit a flaw in the Q&A system. He posts trash as an answer to every question, hoping that the question will close due to inactivity and award him the points. You thwarted him.

VLSM stands for Variable Length Subnet Masking. The idea is that, if you have a limited address space and a lot of networks of varying size, you can make the most of them. VLSM rests on the idea that the first (all 0s) and last (all 1s) in any subnet scheme are unusable.

I will give you an example of how I subnetted a Class C-size network more efficiently using VLSM. If you have more questions, e-mail me.

I had a DMZ (6 hosts max), a serial link (2 hosts max), a small HQ net (14 hosts max), a big customer nets (62 hosts max), and two small customer nets (30 hosts max). This is what I did:

Link Subnet Mask
serial 2 hosts x.y.z.4 255.255.255.252
DMZ 6 hosts x.y.z.8 255.255.255.248
HQ 14 hosts x.y.z.16 255.255.255.240
cust1 30 hosts x.y.z.32 255.255.255.224
cust3 60 hosts x.y.z.64 255.255.255.192
cust2 30 hosts x.y.z.160 255.255.255.224

Do you see how we fill in the big networks, then take the unused space for the smaller ones? You can actually fit more small subnets in there if you need to.

Good fortune,
Don