TCP/IP Addressing

     The word host is often used to described any device on a network that sends and receives information by using an IP address. To determine the location of a destination host with respect to the source computer, IP addresses are organized into classes. This is called classful IP addressing. You must assign IP addresses to all computers that are connected on the network. Network segments connected by a router are called subnets. Breaking down the network into subnets for the purpose of assigning IP addresses is called subnetting.

     The IP address identifies a computer's location on the network. An IP address consists of a set of four numbers, each of which can range from 0 to 255. An IP address has two parts, the host ID and the network ID.

Network ID

     The first part of the IP address is the network ID, which identifies the network segment on which the computer is located. All computers on the same segment must have the same network ID.

Host ID

     The second part of the IP address is the host ID, which identifies a computer, router, or other device within a segment or network. The host ID for each host must be unique within the network ID. Two hosts with different network IDs can have the same host ID, although the combination of the network ID and the host ID must be unique to all computers in communication with each other.

IP Address Classes

     Address classes are used to assign network IDs to organizations so that the computers on their networks can communicate on the Internet. Address classes are also used to define the dividing point between the network ID and host ID.

Class A Networks

     Class A addresses are assigned to networks with a very large number of hosts. This class allows for 126 networks, by using the first number for the network ID. The remaining three numbers are used for the host ID, allowing for 16,777,214 hosts per network.

Class B Networks

     Class B addresses are assigned to networks that range from medium to large in size. This class allows for 16,384 networks, by using the first two numbers for the network ID. The remaining two numbers are used for the host ID, allowing for 65,534 hosts per network.

Class C Networks

     Class C addresses are used for small, local area networks (LANs). This class allows for approximately 2,097,152 networks, by using the first three numbers for the network ID. The remaining number is used for the host ID, allowing for 254 hosts per network.

Classes D and E Networks

     Classes D and E are not allocated to hosts. Class D addresses are used for multicasting, and the class E addresses are reserved for future use.

Determining the Address Class
IP Adress Class	IP Address	Network ID	Range of Values of w
A	w.x.y.z	w.0.0.0	1-126
B	w.x.y.z	w.x.0.0	128-191
C	w.x.y.z	w.x.y.0	192-223
D	w.x.y.z	N/A	224-239
E	w.x.y.z	N/A	240-255

Determining the Network and Host ID

     For the IP addresses in class A, the network ID is the first number in the IP address. For class B, the network ID is the first two numbers; and for class C, the network ID is the first three numbers in the IP address. The remaining numbers identify the host ID.

Subnet Masks

     To divide a network ID, you use a subnet mask. A subnet mask is a screen that differentiates the network ID form a host ID in an IP address but is not restricted by the same rules used in the classful method. A subnet mask consists of a set of four numbers, similar to an IP address. These numbers can range in value from 0 to 255.

     In the classful method, each of the four numbers can assume only the maximum value 255 or the minimum value 0. The four numbers are then arranged as contiguous maximum values followed by contiguous minimum values. The 255 represents the network ID and 0 represents the host ID.

Default Subnet Masks
IP Adress Class	IP Address	Subnet Mask	Network ID	Host ID
A	w.x.y.z	255.0.0.0	w.0.0.0	x.y.z
B	w.x.y.z	255.255.0.0	w.x.0.0	y.z
C	w.x.y.z	255.255.255.0	w.x.y.0	z

Classless Inter-Domain Routing (CIDR)

     IP address classes provide a simple method for differentiating local hosts from remote hosts and for locating the route to a remote host. However, this method permits very few variations in network sizes, which has led to such problems as inappropriately assigning IP addresses to networks. To overcome these limitations, a method known as Classless Inter-Domain Routing (CIDR) was developed for breaking up networks into a larger variety of sizes.

     CIDR optimizes the allocation of IP addresses through the use of supernetting and subnetting. Suppernetting is a procedure that combines multiple addresses to form a single network ID. Subnetting divides a large network into multiple subnets.

Binary Notation

     Classless Inter-Domain Routing (CIDR) uses binary notation, whereas the classful method uses a decimal notation. CIDR translates all IP addresses and subnet masks to binary notation. CIDR divides an IP address into a set of 32 values, in place of the four values used in the classful system.

     CIDR does not define a default subnet mask based on the IP address. Instead, each host is configured with a custom subnet mask, and each router is sent the IP address as part of the data packet. The router then uses a subnet mask from its routing table to determine the network ID of the computer to which the packet must be forwarded.

CIDR Binary IP Addressing

     In the the binary system, an IP address is represented as a string of 32 numbers. This string can be divided into four fields; each field is called an octet, or byte. Each octet consists of eight bits. A bit has a value of either 0 or 1. Therefore, an IP address consists of 4 bytes, or 32 bits.

Binary Subnet Masks

     To be compatible with IP address in binary notation, subnet mask bits also must be converted to binary notation. CIDR notation is an addressing method that associated a subnet mask with the IP address. Regardless of whether the IP address and subnet mask are specified in CIDR notation or in decimal notation, you can calculate the network ID and the IP address by converting all values to binary notation. After you calculate the network ID, you can determine whether a destination host is local or remote.

Subnet Mask Bits

     In binary notation, a subnet mask is represented by four octets, just as the IP address is. The following table shows the binary notation for the default subnet masks used in the classful method:

Decimal Representation        Binary Representation

255.0.0.0                     11111111 00000000 00000000 00000000
255.255.0.0                   11111111 11111111 00000000 00000000
255.255.255.0                 11111111 11111111 11111111 00000000

     Using the binary representation of a subnet mask, you can manipulate 32 numbers in place of the four numbers in the decimal representation. This increased capability provides a greater selection of network sizes compared with the three fixed sizes in the classful method.

     Subnet masks are always composed of contiguous maximum values followed by contiguous minimum values. In binary notation, this translates to a series of contiguous 1s followed by contiguous 0s. The contiguous 1s determine the network ID and the 0s determine the host ID.

CIDR Notation

     CIDR notation involves specifying a dotted decimal notation (192.168.0.1) with a bit mask (/bits). The bit mask specifies the number of contiguous 1s in binary notation of the subnet mask associated with the IP address. The contiguous 1s are the leftmost bits in the subnet mask.

     For example, the IP address represented in CIDR notation as 10.217.123.7/20 specifies that its subnet mask has 20 contiguous 1s. Consequently, the 12 bits remaining from the original 32 must be 0s. The IP addresses in CIDR notation are known by the number of bits from the IP address that make up the network ID and are represented as /x. For example, a 10-bit network ID is represented as /10.

Note: CIDR notation is also known as network prexfix notation.

The following table lists the practical bit masks, the associated subnet masks, and the number of classful networks possible for each.

CIDR Notation        Subnet Mask        Number of Classful Networks

/8                   255.0.0.0          256 class Bs
/9                   255.128.0.0        128 class Bs
/10                  255.192.0.0        64  class Bs
/11                  255.224.0.0        32  class Bs
/12                  255.240.0.0        16  class Bs
/13                  255.248.0.0        8   class Bs
/14                  255.252.0.0        4   class Bs
/15                  255.254.0.0        2   class Bs
/16                  255.255.0.0        1   class B  or 256 class Cs
/17                  255.255.128.0      128 class Cs
/18                  255.255.192.0      64  class Cs
/19                  255.255.224.0      32  class Cs
/20                  255.255.240.0      16  class Cs
/21                  255.255.248.0      8   class Cs
/22                  255.255.252.0      4   class Cs
/23                  255.255.254.0      2   class Cs
/24                  255.255.255.0      1   class Cs
/25                  255.255.255.128    1/2 class C
/26                  255.255.255.192    1/4 class C
/27                  255.255.255.224    1/8 class C
/28                  255.255.255.240    1/16class C

Calculating the Network ID

To calculate the network ID when the IP address is specified in CIDR notation:

Convert the IP address to binary format.
Use the bit mask to determine the number of bits in the IP address that make up the network ID.
Add 0s to the network ID to obtain its four-octet structure.

Consider the IP address 10.217.123.7/20.

                Binary Notation
IP Address      00001010 11011001 01111011 00000111
Subnet Mask     11111111 11111111 11110000 00000000
Network ID      00001010 11011001 01110000 00000000

To calculate the network ID when the IP address and subnet mask are specified in decimal notation:

Convert the IP address to binary format.
Convert the subnet mask to binary format.
Use the number of contiguous 1s in the subnet mask to determine the number of bits in the IP address that make up the network ID.

Consider the IP address 10.217.123.7 with a subnet mask of 255.248.0.0.

                Binary Notation
IP Address      00001010 11011001 01111011 00000111
Subnet Mask     11111111 11111000 00000000 00000000
Network ID      00001010 11011000 00000000 00000000
ID              nnnnnnnn nnnnnhhh hhhhhhhh hhhhhhhh

Local Host Example

Consider two IP addresses: 10.217.123.7/10 and 10.218.102.31/10 belonging to two computers: A and B. The following example details how the two network IDs are calculated to determine whether they are local or remote in relation to eachother. The two computers have the same network IDs, therefore computer A is local in relation to computer B.

Computer A

IP Address         00001010 11011001 01111011 00000111
Subnet Mask        11111111 11000000 00000000 00000000
Network ID (bin)   00001010 11000000 00000000 00000000
Network ID (dec)   10.192.0.0

Computer B

IP Address         00001010 11011010 01100110 00000011
Subnet Mask        11111111 11000000 00000000 00000000
Network ID (bin)   00001010 11000000 00000000 00000000
Network ID (dec)   10.192.0.0

Remote Host Example

Consider two IP addresses: 10.217.123.7/20 and 10.218.102.31/20 belonging to two computers: A and B. The following example details how the two network IDs are calculated to determine whether they are local or remote in relation to eachother. The two computers have different network IDs, therefore computer A is remote in relation to computer B.

Computer A

IP Address         00001010 11011001 01111011 00000111
Subnet Mask        11111111 11111111 11110000 00000000
Network ID (bin)   00001010 11011001 01110000 00000000
Network ID (dec)   10.217.112.0

Computer B

IP Address         00001010 11011010 01100110 00000011
Subnet Mask        11111111 11111111 11110000 00000000
Network ID (bin)   00001010 11011010 01100000 00000000
Network ID (dec)   10.218.96.0

Available Host IDs

The number of hosts supported by a network ID is calculated by using the number of 0s in the associated subnet mask. If the number of 0s is denoted by n, the number of hosts is calculated as: 2^n-2. The two addresses that are subtracted from the expression are the reserved IP addresses that cannot be allocated to any host, the network address and the broadcast address.

The values of the bits in a host ID cannot be all 0s or 1s. The host ID with the value 0 is is used to denote a network ID. The host ID consisting of all 1s is used as an IP broadcast address. The following examples show an example of a valid host ID and a reserved host ID.

Valid Host

IP Address         11000000 10101000 11000001 00000000  192.168.193.0
Subnet Mask        11111111 11111111 11110000 00000000  255.255.240.0
Host ID                                  0001 00000000  1.0

Reserved Host ID

IP Address         11000000 10101000 11000000 00000000  192.168.192.0
Subnet Mask        11111111 11111111 11110000 00000000  255.255.240.0
Host ID                                  0000 00000000  0.0

The following table shows the possible host values for a specific
number of non-zero bits in the CIDR notation of an IP address.

Some users and small businesses are assigned the subnet mask of
255.255.255.255 by their ISP. Although many of the network IDs in
this chart are not practical, such as 2 or 2 billion, it is a useful
reference for determining the subnet mask needed for subnets of
particular sizes.


CIDR Notation        Subnet Mask        # of 0s        # of Hosts (2^n-2)

w.x.y.z/1            128.0.0.0          31             2,147,483,646
w.x.y.z/2            192.0.0.0          30             1,073,741,822
w.x.y.z/3            224.0.0.0          29             536,870,910
w.x.y.z/4            240.0.0.0          28             268,435,454
w.x.y.z/5            248.0.0.0          27             134,217,726
w.x.y.z/6            252.0.0.0          26             67,108,862
w.x.y.z/7            254.0.0.0          25             33,554,430
w.x.y.z/8            255.0.0.0          24             16,777,214
w.x.y.z/9            255.128.0.0        23             8,388,606
w.x.y.z/10           255.192.0.0        22             4,194,302
w.x.y.z/11           255.224.0.0        21             2,097,150
w.x.y.z/12           255.240.0.0        20             1,048,574
w.x.y.z/13           255.248.0.0        19             524,286
w.x.y.z/14           255.252.0.0        18             262,142
w.x.y.z/15           255.254.0.0        17             131,070
w.x.y.z/16           255.255.0.0        16             65,534
w.x.y.z/17           255.255.128.0      15             32,766
w.x.y.z/18           255.255.192.0      14             16,382
w.x.y.z/19           255.255.224.0      13             8,190
w.x.y.z/20           255.255.240.0      12             4,094
w.x.y.z/21           255.255.248.0      11             2,046
w.x.y.z/22           255.255.252.0      10             1,022
w.x.y.z/23           255.255.254.0      09             510
w.x.y.z/24           255.255.255.0      08             254
w.x.y.z/25           255.255.255.128    07             126
w.x.y.z/26           255.255.255.192    06             62
w.x.y.z/27           255.255.255.224    05             30
w.x.y.z/28           255.255.255.240    04             14
w.x.y.z/29           255.255.255.248    03             6
w.x.y.z/30           255.255.255.252    02             2
w.x.y.z/31           255.255.255.254    01             N/A
w.x.y.z/32           255.255.255.255    00             N/A

Optimizing the Allocation of IP Addresses

     In addition to providing a new method for differentiating between local and remote hosts, CIDR also provides solutions to the problems associated with classful IP addressing. These problems include the assignment of an incorrect number of IP addresses or the addition of multiple entries to the Internet routing tables.

Adding Multiple IP Addresses

     Consider an organization that must connect 800 computers to the internet. It can do this by choosing either one class B network ID that provides 65,534 IP addresses or four class C addresses that provide 1,016 addresses. The first choice results in 64,734 unused IP addresses. With the second choice, only 216 IP addresses go unused, but each router on the Internet is required to add four routes to the organization in its routing table.

Supernetting

     To avoid such inappropriate address allocations, CIDR uses the concept of supernetting. Supernetting is a strategy that combines multiple addresses from the classful environment into a single network ID in the classless environment. Using supernetting, CIDR combines multiple class C network IDs into a single CIDR network ID. In CIDR notation, a network ID is represented by the number of bits in a subnet mask, similar to an IP address, for example, 192.168.0.0/22.

Subnetting

     Using supernetting, each organization is assigned a single CIDR network ID, which represents a single network. However, it is not efficient to work with a single large network because such factors as high network noise make the network inefficient.

     To reduce the noise on its internal network, an organization can physically segment the network into subnets by using routers. As each subnet requires its own network ID, the single CIDR network ID obtained for the organization can be divided into smaller network IDs for each subnet. The process of dividing a network ID into smaller network IDs is known as subnetting. The smaller network IDs are known as subnet IDs.

     Subnetting is an internal process that is specific to a network, any router outside of the organization's network cannot view the seperate subnets and their subnet IDs. In the example, if the organization wanted to have no more than 62 computers on each subnet, it could use the subnet mask 255.255.255.192.