For the most efficient use of address space and maximum flexibility
in increasing/decreasing the size of your subnets, Hewlett-Packard
recommends variable-length subnet addressing. To maximize the possibilities
offered with this new approach, you should utilize mirror image counting,
as described in this section, to select subnet numbers.
In the past, a network administrator typically assigned values
to the subnet number and host address fields in numerical order.
For example, within a given subnet, hosts were numbers 1, 2, 3,
etc. and within a given network, the subnets were numbered 1, 2,
3, etc. The result was that some bits on the right side of the subnet
field and host field were “ones and zeros” and some
bits on the left side of the subnet and host fields were “all
zeros” for all subnets and hosts. As shown below, the “all
zeros” bits represented room for growth, and the “ones
and zeros” bits represented bits already consumed by growth.
In the next example. the entire host field has been allocated.
As a result, only the subnet field can grow.
To alleviate this restrictive situation, when you use the
variable-length subnetting approach, you can, alternatively, assign
subnet numbers from the left of the subnet field and work right. This
is implemented using mirror-image counting. In the mirror image
approach, the bits for subnet numbers are assigned left to right
instead of right to left (normal situation). This would result in
the following format:
Table 6-7 Mirror Image Counting
Traditional Approach | Mirror Image Approach |
|---|
01 | 10 |
10 | 01 |
011 | 110 |
100 | 001 |
101 | 101 |
This will allow for more growth bits between the subnet field
and the host fields as shown below.
In this case, both the host field and subnet field have considerably
more growing space than before, although the combined growing space
is the same. As it is difficult to predict how many hosts might
end up in a subnet, or how many subnets there might eventually be,
this arrangement allows for maximum flexibility in growth.
Assigning
Variable-Length Subnet Masks |
 |
In Figure 6-11 “Mirror
Image Subnet and Host Field Allocation”, the boundary
between the host and subnet fields is shown in the middle of the growth
area. The boundary, however, could exist anywhere within the growth
area. The subnet mask determines where the boundary is located. “Ones” in
the subnet mask indicate subnet bits, and “zeros” indicate
host bits. To minimize the amount of rework after the initial planning
of your network, you should choose a subnet mask for a given subnet
based on the projected growth of that subnet. As shown in Figure 6-12 “Example
of Variable-Length Subnet Mask”, the subnet is projected to
have a maximum size of 14 hosts. Therefore, the subnet mask should
be 255.255.255.240. There are two remaining growth bits for this
subnet. If the subnet grows beyond 14 hosts, you may choose to change
one of the two remaining growth bits to a host bit. The new subnet
mask will be 255.255.255.224.
Variable-length subnet masks are assigned using the netmask parameter of the ifconfig command, SUBNET_MASK in netconf with a 32-bit subnet mask indicated in dot notation, or
SAM. In the example below with a Class C IP address, the host portion
has three types of assigned bits: subnet number bits, growth bits,
and host bits, with a chosen subnet mask that allows for growth
in both the host field and subnet field without a mask change.
The example below shows how the bits might look in the host
address portion of a Class C address in a network with four subnets
using variable-length subnetting. In this example “g” stands
for growth bits and “h” stands for host bits.
The 0’s and 1’s in the host address are assigned
subnet numbers.
Table 6-8 Subnets with Variable-Length Subnet Mask
Subnet | Host
Address (Bits) | Subnet
Mask (Bits) |
|---|
A | 10gg
gghh | 1111
1100 |
B | 01gg
ghhh | 1111
1000 |
C | 110g
ghhh | 1111
1000 |
D | 001g ghhh | 1111 1000 |
If hosts are added to subnet B above, so that all of its remaining “g” bits
become “h” bits, the mask in subnet B also must
change. Table 6-9 “Subnets with Subnet Mask (B) Modified” shows the results
of these changes. Also, two additional subnets have been added,
subnet E and subnet F.
Table 6-9 Subnets with Subnet Mask (B) Modified
Subnet | Host
Address (Bits) | Subnet
Mask (Bits) |
|---|
A | 100g
gghh | 1111
1000 |
B | 01hh
hhhh | 1100
0000 |
C | 110g
ghhh | 1111
1000 |
D | 001g
ghhh | 1111
1000 |
E | 101g
ghhh | 1111
1000 |
F | 1110 ghhh | 1111 1000 |
Notice that these additions caused A to change its leftmost
g-bit into a subnet-bit (“s” bit). With
the addition of subnet address 101 for subnet E, the old subnet
address 10 of subnet A must be changed to 100 so that the proper
number of bits are used to make the subnet numbers unique.
Table 6-10 “Subnets with Subnet E Removed” shows the results
of removing subnet E.
Table 6-10 Subnets with Subnet E Removed
Subnet | Host
Address (Bits) | Subnet
Mask (Bits) |
|---|
A | 10gg
gghh | 1111
1100 |
B | 01hh
hhhh | 1100
0000 |
C | 110g
ghhh | 1111
1000 |
D | 001g
ghhh | 1111
1000 |
F | 1110 ghhh | 1111 1000 |
Notice that with this change, subnet A gains back one growth
bit (“g” bit) as its old subnet address of 10
is now unique again.
Example
of Subnets with Variable-Length Subnet Masks
The following example
shows four subnetworks within the 192.6.12 network along with the netconf entries necessary to configure these subnetworks
with variable-length subnet masks. Note that there are four different
subnet masks used in this network. Also note that the subnet numbers
in the network map correspond to the mirror image subnet numbers
listed in Table 6-7 “Mirror Image Counting”.
The subnet numbers shown below correspond to the subnets show
in the network map in Figure 6-13 “Network
Map with Variable-Length Subnets”.
In this example of mirror image counting, the first subnet
has 6 hosts (with space allocated for a maximum of 6 hosts), the
second subnet has 60 hosts (with space allocated for a maximum of
62 hosts), the third subnet has 30 hosts (with space allocated for
a maximum of 30 hosts), and the fourth subnet has 14 hosts (with
space allocated for a maximum of 16 hosts). Notice how the range
of numbers (129-135, 65-127,193-223, and 33-47) is spread out to
allow room for additional growth in each case.
Table 6-11 Variable-Length Subnet Addressing in a Class C Network
Subnet Mask 255.255.255 | Subnet Number | Mirror Image Subnet Number | Subnet Address | Internet
Address Range |
|---|
248 | 16 (10000) | 1 (00001) | n.n.n.128 | n.n.n.129
- n.n.n.134 |
192 | 1 (01) | 2 (10) | n.n.n.64 | n.n.n.65-n.n.n.126 |
128 | 1 (01) | 2 (10) | n.n.n.0 | n.n.n.1-n.n.n.126
n.n.n.129-n.n.n.254 |
224 | 6 (110) | 3 (011) | n.n.n.192 | n.n.n.193-n.n.n.222 |
240 | 2 (0010) | 4 (0100) | n.n.n.32 | n.n.n.33-n.n.n.46 |
|
| |
|
 | NOTE: In a Class C network, no two subnets can have more than
125 hosts. |
|
| |
|
Summary network information:
Company division network = 192.6.12 |
Facility LAN subnet
Subnet mask: 255.255.255.248
Subnet number: 16
Mirror image subnet number: 1
Subnet address: 192.6.12.128
Host address range: 129 to 134
Host A internet address: 192.6.12.129 for network interface lan0
Host B internet address: 192.6.12.131 for network interface lan0
Host C internet address: 192.6.12.130 for network interface lan0
Host D internet address: 192.6.12.132 for network interface lan0 |
Marketing Department Subnet
Subnet mask: 255.255.255.192
Subnet number: 1
Mirror image subnet number: 2
Subnet address: 192.6.12.64
Host address range: 65 to 126
Host A internet address: 192.6.12.65 for network interface lan1 |
Networking LAN subnet
Subnet mask: 255.255.255.128
Subnet number: 17
Mirror image subnet number: 2
Subnet address: 192.6.12.0
Host address range: 1 to 126 and 129 to 254
Host A internet address: 192.6.12.1 for network interface lan0
Host B internet address: 192.6.12.120 for network interface lan0
Host C internet address: 192.6.12.130 for network interface lan1
Host D internet address: 192.6.12.182 for network interface lan1 |
R & D Department Subnet
Subnet mask: 255.255.255.224
Subnet number: 6
Mirror image subnet number: 3
Subnet address: 192.6.12.192
Host address range: 193 to 222
Host C internet address: 192.6.12.193 for network interface lan1 |
Manufacturing Department Subnet
Subnet mask: 255.255.255.240
Subnet number: 2
Mirror image subnet number: 4
Subnet address: 192.6.12.32
Host address range: 33 to 46
Host B internet address: 192.6.12.33 for network interface lan1 |
Example
of Network Map with Variable-Length Subnets
This sample network combines networks, subnets with variable-length
subnet masks, and clusters previously described and illustrated
in this chapter. Note that the subnet masks for each IP address
are shown on the accompanying worksheet.
Table 6-12 Network Map II Worksheet
Host | Interface Alias | Internet Address | Subnet
Mask | Cnode Type |
|---|
A | mkt_32
mkt_64 | 192.6.12.129 192.6.12.65 | 255.255.255.248 255.255.255.192 | Server |
A1 | mkt_a1 | 192.6.12.67 | 255.255.255.192 | Client
1 |
A2 | mkt_a2 | 192.6.12.66 | 255.255.255.192 | Client
2 |
B | mfg_32
mfg_128 | 192.6.12.131 192.6.12.33 | 255.255.255.248 255.255.255.240 | N/A |
B1 | mfg__b1
b1_ship | 192.6.12.34 192.1.1.2 | 255.255.255.240 N/A | N/A |
B2 | mfg_b2 | 192.6.12.35 | 255.255.255.240 | N/A |
C | rd_32
rd_96 | 192.6.12.130 192.6.12.193 | 255.255.255.248 255.255.255.224 | Server |
C1 | rd_c1 | 192.6.12.195 | 255.255.255.224 | Client
1 |
C2 | rd_c2 | 192.6.12.194 | 255.255.255.224 | Client
2 |
D | div1_32
div_gw | 192.6.12.132 192.6.20.1 | 255.255.255.248 255.255.255.0 | N/A |
Div2 | div2_128
div2_gw | 192.6.14.129 192.6.20.2 | 255.255.255.0 255.255.255.0 | N/A |
Div3 | div3_128
div3_gw | 192.6.13.130 192.6.20.3 | 255.255.255.0 255.255.255.0 | N/A |
E (SLIP) | e_slip | 192.1.1.1 | N/A | N/A |
Printable version
