In VLSM we usually subnet in descending order from the network with highest number of hosts to the lowest one , but i was reading that "If you allocate subnets in any other order you will not be able to avoid leaving gaps between the subnets."
I tried with simple example and subnetted out of order to understand the meaning of the gap , but i couldn't understand the where the gap is ?
Can someone please explain this with an example ?
Regards
Technically, there won't be "unavoidable" gaps, but convenience will certainly suffer.
Suppose you start with longest masks, i.e. you allocate first 192.168.1.0/28 (address range 0..15) and then you've decided that you need to allocate /26.
The closest range where you can start it would be 192.168.1.64/0, as /26 can only start at 0, 64, 128 and 192. This leaves you with an unused gap from .16 to .63, that you sure can use for three /28 or one /28 and one /27, or a bunch of smaller networks, but what if you don't have these smaller nets in existence?
Now if you do the reverse, and allocate /26 first, like 192.168.1.0/26, then you can allocate your /28 right away from 192.168.1.64/28 leaving no addresses left unused in between.
To be precise, allocating largest blocks first reduces the fragmentation of address space and increases convenience of managing it.
