IP Addresses

Notes:

IP addresses are 32 bit values, enough to address 4 billion individual hosts on an internet. However assigning addresses on a per-host basis is extremely inefficient in terms of administration and routing. Instead, addresses are split into a network portion and a host portion. IP packets are routed to their destination using only the network portion of the address, which is identical for all hosts on the network. The designers of the TCP/IP protocols foresaw the Internet consisting of a few thousand interconnected networks located in universities, companies and government departments. To model the varying size of these organisations in the real world they split the 32 address into 3 different classes of network:
Class A: The first octet (byte) represents the network address and the remaining 3 octets the host within that network. The first bit of the network address is always 0 giving 126 possible class A networks with 16,777,216 individual hosts on each.
Class B: The first two octets represent the network and the last two the host. The first two bits are always 10 giving 16384 possible class B networks each with up to 65534 hosts.
Class C: The first three octets represent the network and the last byte the host. The first three bits are 110 giving 2097152 distinct class C networks, each of which can have 254 hosts.
All three classes of network use the same basic form of Internet address, with the network address component indicating the class of the network. A suitable network address is obtained either from the InterNIC or more usually from the Internet Service Provider. It is then entirely up to the network administrator how the host addresses are assigned. Combined with DNS this part of the Internet can be administered and maintained without recourse to the management authorities.
Due to the enormous growth it is very hard to obtain more than a Class-C network address. Often multiple Class-C addresses will be allocated rather than a Class-B address.