HP 3000 Manuals HP 3000 Manuals
Subnetworks
IP Subnets are used to divide one network into two or more distinct
subnetworks. Subnet numbers identify subnetworks in the same way that
network addresses identify physically distinct networks. Subnetting
divides the node address portion of an IP address into two portions--one
for identifying a specific subnetwork and one for identifying a node on
that subnetwork.
Why Use Subnets?
The use of subnets is optional. Subnets are typically used in
organizations that have a large number of computers. You may want two or
more physically distinct networks to share the same network address.
This may occur, for example, if your organization has acquired only one
network number, but any of the following is true:
* A few nodes on a single network create the bulk of the network
traffic and you want to isolate those nodes on a subnetwork to
reduce overall congestion.
* You have a single LAN and have reached the limit of its technology
in terms of node numbers or cable length.
* LANs are located too far apart to be joined with bridges.
How Subnetting Works.
You may use subnets to divide your current network into subnetworks
without informing remote networks about an internal change in
connectivity. A packet will be routed to the proper subnet when it
arrives at the gateway node. However, if you want a remote node to know
about only some of the subnets on your network, this must be configured.
The network portion of an IP address must be the same for each subnetwork
of the same network. The subnet portion of an IP address must be the
same for each node on the same subnetwork.
Assigning Subnet Masks.
Before you can determine subnet numbers, you first must determine which
bits of the node address will be used to contain your subnet numbers.
The bits that you designate for subnet identifiers compose the subnet
mask. The subnet mask is configured with NMMGR. The remaining part of
the node address is used to identify the host portion of the IP address.
The following rules apply when choosing a subnet mask and an IP address:
* Although any bits in the node address can be used as the subnet
mask, Hewlett-Packard recommends aligning the subnet mask along
byte boundaries, adjacent to the network number.
* Although standards allow subnets on the same network to have
different subnet masks, Hewlett-Packard recommends that you assign
the same subnet mask to all subnets on a network.
* Do not assign an IP address where the network address and/or node
address bits are all off (all 0s) or all on (all 1s). Likewise,
the subnet address bits cannot be all 0s or all 1s.
To determine the subnet mask, you first need to estimate the number of
networks required and the number of nodes on each subnet. Allow enough
bits for both nodes and subnets, as described in example 1.
Example 1.
Assume you are choosing a subnet mask for a class C network (three bytes
for network address, one byte for node address), and you need four
subnets with up to 30 nodes on each subnet. You will need to reserve
three bits for the subnet address (remember, all 0s and all 1s cannot be
used) and the remaining five bits for the node numbers.
This is shown below in figure 2-1.
![[]](../../pic3k/M013399/FFN23c50.gif)
Figure 2-1. Class C Address with Subnet Number
The 30 nodes per subnet will require at least five bits of the node
portion of the IP address (30 <32, and 32=25, therefore you need 5 bits).
This leaves three bits remaining in the node portion of the IP address
for use as the subnet identifier. Subnet parts of all 0's or all 1's are
not recommended because they can be confused with broadcast addresses.
Therefore, you can have up to six subnets (23--2=6) when three bits are
used for the subnet identifier.
Example 2.
An IP address on a class B network with an 8-bit subnet mask separates as
follows:
![[]](../../pic3k/M013399/FFN33c50.gif)
Figure 2-2. Class B Address with Subnet Number
Now, refer again to example 1. The subnet mask must indicate that three
bits of the node portion of the IP address will be used for the subnet
identifier. The subnet mask turns on (sets to 1) all the relevant bits
for its subnet scheme. The subnet mask for example 1 is shown below.
Note that the most significant three bits of the rightmost byte are set.
Subnet Mask
Binary 11111111.11111111.11111111 11100000
Decimal 255.255.255 224
Table 2-1 shows valid addresses for the subnetwork in example 1. You
will need to know this information for NMMGR configuration. The table
shows the possible values of the rightmost byte of the IP address for
each of the subnets, given the criteria described in the example.
(Remember, an address of all 0s or all 1s is not valid).
Column 2 shows the values, in binary, of the six subnet addresses.
Five zeroes are shown in parentheses to indicate where the three
subnet-address bits are located in the byte. The equivalent decimal
value for each subnet address is shown in the third column. The fourth
column shows the range of possible values for the node address of each
subnet. The five rightmost bits make up the node portion, and the range
is the same for all subnets. By combining the subnet address with the
range of node addresses, the possible decimal values of the rightmost
byte are obtained and shown in the fifth column.
The table shows that subnets of 30 nodes each are possible given a subnet
mask of 255.255.255 224. This is derived from the column that shows the
range of possible values for the five bits that make up the node portion
of the IP address. The range for each of the six subnets shows 30
possible values.
Table 2-1. Valid Addresses of Example Subnetwork
----------------------------------------------------------------------------------------
| | | | | |
| Subnet | Address of | Decimal | Possible Node | Decimal |
| | Subnetwork in | Value of | Addresses on | Value of |
| | Binary | Subnetwork | Subnetwork | Rightmost |
| | | | | Byte |
| | | | | |
----------------------------------------------------------------------------------------
| | | | | |
| 1 | 001 (00000) | 32 | 00001-11110 | 33-62 |
| | | | | |
| 2 | 010 (00000) | 64 | 00001-11110 | 65-94 |
| | | | | |
| 3 | 011 (00000) | 96 | 00001-11110 | 97-126 |
| | | | | |
| 4 | 100 (00000) | 128 | 00001-11110 | 129-158 |
| | | | | |
| 5 | 101 (00000) | 160 | 00001-11110 | 161-190 |
| | | | | |
| 6 | 110 (00000) | 192 | 00001-11110 | 193-222 |
| | | | | |
----------------------------------------------------------------------------------------
By looking at the binary values of two IP addresses, it is easy to tell
if nodes belong to the same subnet. If they do, all the bits that make
up the subnet mask will be the same between IP addresses in the subnet.
Take, for example, two IP addresses (in decimal and in binary) of subnet
number 1 from table 2-1:
192.6.12.41 1100 0000 0000 0110 0000 1100 0010 1001
192.6.12.55 1100 0000 0000 0110 0000 1100 0011 0111
The subnet mask has already been defined as:
255.255.255 224 1111 1111 1111 1111 1111 1111 1110 0000
Because the mask has all bits except the five rightmost bits set to 1,
all bits except the five rightmost bits must match between nodes on the
same subnet. Because the two example IP addresses from subnet 1 do match
except for their five rightmost bits, they belong to the same subnet.
NOTE Subnet addressing can be used in internetworks (networks with
gateways).