Dissecting the Component Parts of an IP Address
First, deducing the size of the three parts (classful view) or two parts
(classless view) of an IP address is important, because it allows you
to analyze information about that subnet and other subnets. Every
internetwork requires some number of subnets, and some number of hosts
per subnet. Analyzing the format of an existing address, based on the
mask or prefix length, allows you to determine whether enough hosts per
subnet exist, or whether enough subnets exist to support the number of
hosts. The following list summarizes some of the common math facts about
subnetting related to the format of IP addresses:
■ If a subnet has been defined with y host bits, there are 2y – 2 valid usable IP addresses in the subnet, because two numeric values are reserved.
■ One reserved IP address in each subnet is the subnet number itself. This number, by definition, has binary 0s for all host bits. This number represents the subnet, and is typically seen in routing tables.
■ The other reserved IP address in the subnet is the subnet broadcast address, which by definition has binary 1s for all host bits. This number can be used as a destination IP address to send a packet to all hosts in the subnet.
■ When you are thinking classfully, if the mask implies x subnet bits, then 2x possible subnets exist for that classful network, assuming the same mask is used throughout the network.
■ Although there are no truly reserved values for the subnet numbers, two (lowest and highest values) may be discouraged from use in some cases:
— Zero subnet—The subnet field is all binary 0s; in decimal, each zero subnet is the exact same dotted-decimal number as the classful network number, potentially causing confusion.
— Broadcast subnet—The subnet field is all binary 1s; in decimal, this subnet’s broadcast address is the same as the network-wide broadcast address, potentially causing confusion.
In Cisco routers, by default, zero subnets and broadcast subnets work fine. You can disable the use of the zero subnet with the no ip subnet-zero global command. The only time that using the zero subnet typically causes problems is when classful routing protocols are used.
■ If a subnet has been defined with y host bits, there are 2y – 2 valid usable IP addresses in the subnet, because two numeric values are reserved.
■ One reserved IP address in each subnet is the subnet number itself. This number, by definition, has binary 0s for all host bits. This number represents the subnet, and is typically seen in routing tables.
■ The other reserved IP address in the subnet is the subnet broadcast address, which by definition has binary 1s for all host bits. This number can be used as a destination IP address to send a packet to all hosts in the subnet.
■ When you are thinking classfully, if the mask implies x subnet bits, then 2x possible subnets exist for that classful network, assuming the same mask is used throughout the network.
■ Although there are no truly reserved values for the subnet numbers, two (lowest and highest values) may be discouraged from use in some cases:
— Zero subnet—The subnet field is all binary 0s; in decimal, each zero subnet is the exact same dotted-decimal number as the classful network number, potentially causing confusion.
— Broadcast subnet—The subnet field is all binary 1s; in decimal, this subnet’s broadcast address is the same as the network-wide broadcast address, potentially causing confusion.
In Cisco routers, by default, zero subnets and broadcast subnets work fine. You can disable the use of the zero subnet with the no ip subnet-zero global command. The only time that using the zero subnet typically causes problems is when classful routing protocols are used.
