Classless IP Addressing
By Stephen Bucaro
In order to communicate over a TCP/IP network each device must have a unique IP address.
An IPv4 address is 32 bits grouped in sets of 8 bits called octets. In order to maintain the uniqueness
of IP addresses on the public Internet, the IANA was setup (Internet Assigned Numbers Authority).
An IP address is divided into two parts, the network ID and the node ID. Each device
on a local network must have the same network ID, and a unique node ID. In order to simply
the assignment of groups of IP address to large organizations, address classes were set up.
Although there are five address classes, A through F, only classes A, B and C are used
to assign IP addresses (class D is used for multicasting, class E is reserved for future use).
The classes are differentiated by the number of octets used for the network ID, and the number
of octets used for the node ID.
In a class A IP address has the first octet is the network ID, which allows up to 126
networks to be defined, and the remaining seven octets are used for the node ID, which allows
up to 16,777,214 devices to be assigned on each network.
In a class B IP address, the first two octets are the network ID, which allows up to
16,384 networks, and the remaining six octets are used for the node ID, which allows up to
65,534 devices to be assigned on each network.
In a class C IP address, the first three octets are the network ID, which allows up to
2,097,152 networks, and the remaining five octets are used for the node ID, which allows up
to 254 devices to be assigned on each network.
At the receiving end, the class of an address is identified by the first 3 bits of the
IP address, and based upon those bits, a mask, called a subnet mask, is used to separate the
network ID and node ID parts of the address.
IP address that follow the address classing scheme are now called "classful" addresses.
In the beginning, classful addressing was a great idea for keeping IP addresses organized.
But as the Internet grew, and IP addresses became scarce, classful addressing was found to
be too limiting.
Classless IP addressing, known as Inter-Domain Routing (CIDR) was released in September
1993. Since there are no classes in CIDR, it can't determine which are the network ID and node
ID parts of the IP address. It uses Routing Information Protocol version 2 (RIP v2) which sends
a 32 bit IP address and a 32 bit network mask. (Classful routing protocol RIP v1 does not send
a subnet mask).
Since CIDR has no classes network ID and node ID do not need to be on octet boundaries,
this actually allows 30 network ID and node ID boundaries. Sending the subnet mask allows CIDR
to use route aggregation, where the same subnet mask can be used with all the packets for the
entire message (sometimes called route summarization).
Use of classless IP addressing has caused a change in human notation where the 32 bit
IP address is suffixed with a slash followed by the number of 1 bits in the mask. For example,
indicates a 24-bit subnet mast (255.255.255.0) giving a node ID of 192.168.100.0 and
host addresses from 192.168.100.0 to 192.168.103.255.
IPv4 classless IP addressing is only stop-gap measure until the implementation of IPv6,
which uses 128-bit IP addresses. Whereas IPv4 classless IP addressing can generally be implemented
with router software upgrades, IPv6 in most cases will require router hardware upgrades.
More Networking Protocols and Standards:
• X.25 and Frame Relay Overview
• IPv4 Datagram Fields
• IPv6 Address Types and Scopes
• Dynamic Host Configuration Protocol (DHCP) Explained
• TCP/IP Utilities
• The TCP/IP protocol Datagram Format
• T-Carrier - A Complete and Comprehensive Guide
• The OSI Application Layer
• Network Routing Protocols - IGRP, EIGRP, OSPF, ISIS, BGP
• Looking at the OSI 7 Layer Reference Model