This question is bit novice. But, just wanted to clear any doubts as a starter for this topic. I have read about Classes of address like Class A, B, C. As I observe the Default Subnet mask for each of these classes, I see that, starting from Class A, gradually, network mast/prefix length increases.
Default subnet masks: Class A: 255.0.0.0 (prefix length : 8) Class B: 255.255.0.0 (prefix length : 16) Class C: 255.255.255.0 (prefix length : 24)
This p appears to me like, if a network administrator uses Class C address, then, he will be limited to have maximum of 254 hosts (excluding 0 and 255), whereas, if he chooses Class A address, then, he can have maximum number of hosts to give IP addresses to. Is this perception correct? So, Class A is always better, even if some IP addresses are wasted in the beginning, it can always grow in future? So, my question is, on which condition or constraint, a particular Class of address will be picked by the network admin? Is he free to choose any class of address? Also, I read that, starting octet of a particular class of address needs to fall into a range. For example, for Class C, it has to be 192-223 (inclusive). That, means, if network admin chooses, Class C address with default subnet mast (255.255.255.0), then one of the host can have an IP address like 192.0.0.1 or 126.96.36.199? Is it legal to have all 0's in the network part of the address also?
Also, another query is, does the subnet mast denote maximum number of hosts possible in the network and can be reduced by subnetting? For example, as per my understanding, netmask can be further changed for subnetting, like if a Class C default netmask 255.255.255.0 is converted to 255.255.255.192, network admin can break the original network into 4 subnets?