Network Basics: IP Network Classes
Network addressing architecture divides the address space for Internet Protocol Version 4 (IPv4) into five address classes. Each class, coded in the first four bits of the address, defines either a different network size, i.e. number of hosts for unicast addresses (classes A, B, C), or multicast network (class D). The fifth class (E) address range is reserved for future or experimental purposes. Knowing network classes becomes an issue when you deal with routing.
All actual network IDs and addresses are managed and distributed by the Internet Assigned Number Authority (IANA), which manages the entire pool of addresses. Addresses used to be permanently assigned to organizations, and any organizations (government, universities, or corporations) could purchase an address block (typically, a class network). During the 1990’s, IANA and IAB realized that IP addresses were being consumed faster than expected and in light of the shortage of IP addresses, this practice has slowed to almost a stop.
Currently, if you need public IP addresses, your Internet service provider (ISP) usually lease you an appropriately sized block while you get your Internet services from that ISP. The day of companies purchasing IP addresses are in the past.
This section looks at the address classes in descending size order, from Class A to Class E.
Class A addresses are IP addresses that are assigned to network devices, such as computers, and include all addresses in which the first bit of the first octet is set to 0 (zero). This includes all values from 00000001 to 01111111, or 1 to 127. For Class A networks, the first octet represents a network ID that is defined in the address by a subnet mask.
The network ID is not allowed to have all its bits set to 0 or all bits set to 1. The 127 network ID is excluded from this address class.
Thus, 126 possible Class A networks are available to organizations around the world. With only 126 Class A networks, owning one puts you in an exclusive club. You can no longer acquire a network block of addresses, and when possible, IANA gets them back from the registered owners. Getting addresses back allows IANA to redistribute addresses in a more efficient and temporary manner.
Class B addresses are IP addresses that are assigned to network devices, such as computers, and include all addresses in which the first two bits of the first octet are 10. This includes all values from 10000000 to 10111111, or 128 to 191.
The definition of the Class B network is represented with a subnet mask, but the Class B network ID is made up of the values in the first two octets. Unlike Class A networks, all network IDs in this range are available for use.
Class C addresses are IP addresses that are assigned to network devices, such as computers, and include all addresses in which the first three bits of the first octet are set to 110. This includes all values from 11000000 to 11011111, or 192 to 223.
The default subnet mask for Class C networks defines the first three octets as the network ID for these networks. Like with Class B networks, all the network IDs are available for use on networks. This is the last of the network classes that will be used for network devices on a TCP/IP network.
Class D network addresses are not assigned to devices on a network. These addresses are used for special-purpose, multicast applications (such as video- and audio-streaming applications).
These addresses all need to be registered with IANA to be used globally. Addresses in this class have the first bits of the first octet set to 1110, yielding addresses in the first octet ranging from 11100000 to 11101111, or 224 to 239. These addresses are not defined by a normal subnet mask; instead, each address is used for a specific purpose. And because each address is individually used, it uses a 255.255.255.255 mask.
If Class D is special, Class E addresses are even more special. There is no defined use for this address class. Officially, it is listed as reserved for usage and testing by IANA and the Internet Research Task Force (IRTF). In fact, as of RFC3330 in 2002, Class E was updated to reserved for future use.
Class E comprises absolutely all valid addresses with 240 or higher in the first octet. The first bits of the first octet is 1111, which yields addresses from 11110000 to 11111110 — or technically, 11111111 — which, in decimals, are 240 to 254 — or 255.
Because this address class is not being used for address allocation, you cannot know what the network ID, which defines the valid addresses in a range, is. So the inclusion of 255 at the end of the range is moot because this address range is not available for you to use. All you need to know is that by definition Class E includes all valid addresses higher than Class D.