The Internet Protocol (IP) was designed for routing data between network sites all over the world, and was later adapted for routing data between networks within any site (often referred to as "subnetworks" or
IP includes a system by which a unique number can be assigned to each of the millions of networks and each of the computers on those networks.
Such a number is called an IP address.
To make IP addresses easy to understand, the originators of IP adopted a system of representation called
"dotted decimal" or "dotted quad" notation. Below are examples of IP addresses written in this format:
Each of the four values in an IP address is the ordinary decimal
(base 10) representation of a value that a computer can handle using eight "bits"
(binary digits -- 1s and 0s).
The dots are simply convenient visual separators.
Zeros are often used as placeholders in dotted decimal notation;
220.127.116.11 can therefore also appear as 189.021.241.056.
IP networks are divided into three classes on the basis of size.
Max number of
networks in class
number of hosts
Networks attached to the Internet are assigned class types that determine the maximum number of possible hosts per network.
Class A is assigned to networks that have more than 65,535
Class B is for networks that have 256 to 65534 hosts;
Class C is for networks with less than 256 hosts.
All network addresses outside of these ranges (Class D and E) are either reserved or set aside for experimental networks or multicasting.
When an IP address's host portion contains only zero(s), the address identifies a network and not a host. No physical device may be given such an address.
The network portion must start with a value from 1 to 126 or from 128 to 223. Any other value(s) in the network portion may be from 0 to 255, except that in class B the network addresses 18.104.22.168 and 22.214.171.124 are
reserved, and in class C the network addresses 192.0.0.0 and 126.96.36.199 are reserved.
The value(s) in the host portion of a physical device's IP address can be in the range of 0 through 255 as long as this portion is not all-0 or all-255.
Values outside the range of 0 to 255 can never appear in an IP address
(0 to 255 is the full range of integer values that can be expressed with eight bits).
The network portion must be the same for all the IP devices on a discrete physical network (a single Ethernet LAN, for example, or a WAN link).
The host portion must be different for each IP device
or, to be more
precise, each IP-capable port or interface connected directly to that network.
The network portion of an IP address will be referred to as a network number; the host portion will be referred to as a host number.
To connect to the Internet or to any private IP network that uses an Internet-assigned network number, you must obtain a registered IP network number from an Internet-authorized network information center. In many
countries you must apply through a government agency, however they can usually be obtained from your Internet Service Provider (ISP).
If your organization's networks are, and will always remain, a closed system with no connection to the Internet or to any other IP network, you can choose your own network numbers as long as they conform to the above
If your networks are isolated from the Internet, e.g. only between your two branch offices, you can assign any IP Addresses to hosts without problems. However, the Internet Assigned Numbers Authority (IANA) has reserved
the following three blocks of IP Addresses specifically for private (stub) networks:
|| Beginning Address
It is recommended that you choose private network IP Addresses from the above list.
In the absence of subnetworks, standard TCP/IP addressing may be used by specifying subnet masks as shown below.
Subnet mask settings other than those listed above add significance to the interpretation of bits in the IP address.
The bits of the subnet mask correspond directly to the bits of the IP address. Any bit an a subnet mask that is to correspond to a net ID bit in the IP address must be set to 1.
( D-Link )