Subneting starting with the biggest subnets

Question

Can anyone explain or give me some example for the reason why we start subneting using the largest subnets first? What are consequences of opposite approach?

Answer 1

This is a most extreme example. Subnet 10.0.0.0 /8 for 2 networks with 2 hosts each.

 Network             DirectedBroadcast   CIDR Mask                UsableHosts 
 10.0.0.0            10.127.255.255      9    255.128.0.0         8,388,606   AVL
 10.128.0.0          10.191.255.255      10   255.192.0.0         4,194,302   AVL
 10.192.0.0          10.223.255.255      11   255.224.0.0         2,097,150   AVL
 10.224.0.0          10.239.255.255      12   255.240.0.0         1,048,574   AVL
 10.240.0.0          10.247.255.255      13   255.248.0.0         524,286     AVL
 10.248.0.0          10.251.255.255      14   255.252.0.0         262,142     AVL
 10.252.0.0          10.253.255.255      15   255.254.0.0         131,070     AVL
 10.254.0.0          10.254.255.255      16   255.255.0.0         65,534      AVL
 10.255.0.0          10.255.127.255      17   255.255.128.0       32,766      AVL
 10.255.128.0        10.255.191.255      18   255.255.192.0       16,382      AVL
 10.255.192.0        10.255.223.255      19   255.255.224.0       8,190       AVL
 10.255.224.0        10.255.239.255      20   255.255.240.0       4,094       AVL
 10.255.240.0        10.255.247.255      21   255.255.248.0       2,046       AVL
 10.255.248.0        10.255.251.255      22   255.255.252.0       1,022       AVL
 10.255.252.0        10.255.253.255      23   255.255.254.0       510         AVL
 10.255.254.0        10.255.254.255      24   255.255.255.0       254         AVL
 10.255.255.0        10.255.255.127      25   255.255.255.128     126         AVL
 10.255.255.128      10.255.255.191      26   255.255.255.192     62          AVL
 10.255.255.192      10.255.255.223      27   255.255.255.224     30          AVL
 10.255.255.224      10.255.255.239      28   255.255.255.240     14          AVL
 10.255.255.240      10.255.255.247      29   255.255.255.248     6           AVL
 10.255.255.248      10.255.255.251      30   255.255.255.252     2           REQ 2
 10.255.255.252      10.255.255.255      30   255.255.255.252     2           REQ 2

Imagine what would have happened if you had picked those two in the middle. My calculator works differently than others I have seen. You tell it the starting point i.e. 10.0.0.0 /8 or 192.168.1.0 /24, and the number and size of the networks you want like this

 1,30  '1 network with 30 host
 2,10  '2 nets with 10 hosts
 1,4   '1 net with 4 hosts

 Network             DirectedBroadcast   CIDR Mask                UsableHosts 
 192.168.1.0         192.168.1.127       25   255.255.255.128     126         AVL
 192.168.1.128       192.168.1.159       27   255.255.255.224     30          AVL
 192.168.1.160       192.168.1.191       27   255.255.255.224     30          REQ 30
 192.168.1.192       192.168.1.207       28   255.255.255.240     14          AVL
 192.168.1.208       192.168.1.223       28   255.255.255.240     14          REQ 10
 192.168.1.224       192.168.1.239       28   255.255.255.240     14          REQ 10
 192.168.1.240       192.168.1.247       29   255.255.255.248     6           AVL
 192.168.1.248       192.168.1.255       29   255.255.255.248     6           REQ 4

Answer 2

It is not a fixed rule, you can start subneting using the smallest subnet first or any order you like. However, with variable length subnet masks (VLSM) starting by the largest is easier to manage and support future modifications.

Starting by the largest is more space efficient, since you can borrow more bits for the smaller subnets. Since you can only subnet an address range at certain locations, you will have problems matching some addresses. Consider this:

1st split == 128 spaces
2nd split == 64 spaces
3rd split == 32 spaces

Every time you split you get less space. Imagine you need 2 subnets, one for 70 pcs and another for 3. You will make a better use of your addressing space starting by the biggest one first. It is the most compact way

Answer 3

You want to subnet in such a way that you have the largest block of addresses available when done. I have a subnet calculator, and as an example, I took 192.168.1.0 /24, and told it I wanted 2 networks of 30 hosts, and 1 network of 10 hosts. Here is the output:

192.168.1.0  /25 -> 126  (128) ++
192.168.1.128  /27 -> 30  (32)
192.168.1.160  /27 -> 30  (32)
192.168.1.192  /28 -> 10  (16)

An explanation of the above

Net              CIDR  Hosts   Max Hosts(includes net# and directed broadcast)
192.168.1.0      25    126     (128)
192.168.1.128    27    30      (32) - 0 available
192.168.1.160    27    30      (32) - 0 available
192.168.1.192    28    10      (16) - 4 more available

The Max Hosts includes the network number and the directed broadcast, so for example, if you have 32 max, then only 30 are able to be assigned.

Answer 4

Maybe it's not a fixed rule, but I remember, when learning for CCNA and practicing in Packet Tracer, I've got error when set up smaller subnet first.