Is it possible design a network from a ip address and go backwards ?

[dolanduck.]

what i mean about that lets say i have 192.168.0.0 and i need 200 sub-nets and 300 host each.

by default this is a type c network with a /24 mask. is it possible to go backwards. to meet the demands.

My second question is do you need a sub-net mask to find sub-nets and host bits in a network.?

like if i only have 172.0.0.0 or 192.168.0.0  but no mask is it possible ?

[Giuseppe Larosa]

Hello @dolanduck. ,

in modern classless routing you can also perform supernetting that actually means going above the classful network limit creating a prefix like 192.168.0.0/16.

However, with some routing protocols like RIPv2 this is possible but with great limitations. There are no issues with OSPF or EIGRP.

To answer you question you need to note the following:

300 hosts > 254 --> requires 9 bits /23 subnets that can host up to 510 hosts.

To be able to have 200 of /23 you need a /15 prefix so even 192.168.0.0/16 is not enough as it is able to accomodate only 128 /23 subnets.

You would need 192.168.0.0/15 to accomodate 200 /23 subnets.

>>

My second question is do you need a sub-net mask to find sub-nets and host bits in a network.?

like if i only have 172.0.0.0 or 192.168.0.0  but no mask is it possible ?

Generally speaking yes you need additional information that can be a subnet mask or a number of end user hosts that must be supported to be able to find the right subnet size like in the exercise above.

The rule is that N host bits allow to accomodate 2^N -2 hosts  ( one is the subnet and the last is the subnet broadcast address).

P+ N = 32

where P is the prefix length

and N needs to be able to satisfy the disequation 2^N-2 > MAx hosts per subnet

Where N is the smaller one that satisfy the above condition.

Hope to help

Giuseppe

[Joseph W. Doherty]

To OP, this posting seems to very much to overlap with your prior posting (https://community.cisco.com/t5/routing/i-needs-help-with-clarification-finding-host-bits-and-subnet/td-p/4125602), which you marked as "solved". Also in that posting, my first response provided basically the same information being provided by Giuseppe.

Perhaps you could further clarify and/or expand what you're trying to find out about or understand?

BTW, like I also described in that other posting, supernetting is a bit archiac, likewise, so is "real" subnetting. Best to avoid either by using CIDR (https://en.wikipedia.org/wiki/Classless_Inter-Domain_Routing), which is what Giuseppe is also suggesting.

Also to your "new" questions, and as already (correctly) answered by Giuseppe:
"My second question is do you need a sub-net mask to find sub-nets and host bits in a network.?"
Yes

Basically, the "big" feature of Classful addressing, the IP address, itself, provides its "natural" mask. I.e. you always need a mask, just a "natural" Classful IP address provides it.
Supernetting and/or subnetting is using Classful IP addresses "unnaturally".