Subnetting Question

[Requiemsallure]

Potentially simple question, that I cannot wrap my head around.  The subnetting equation 2^N -2 ≥ X where N = the number of host bits taken in order to subnet a network into X segments. 2^N = the number of created subnets, and the -2 are supposedly invalid subnets.  which leads me to my question. Why are the subnets invalid? I understand that subnetting wastes address space due to an extra broadcast and an extra network address per subnet. However Ihave never heard of an invalid subnet in such a respect that you couldn't use any of the addresses from that range. The example given in the book I am reading states that 192.168.1.0 and 192.168.1.240 are invalid subnets. In this instance you would have effectively wasted 32 network addresses. Is there something to this that I don't yet understand, I remember hearing somewhere that you can use these invalid ranges, however I don't remember how.

     I guess this defines my question into two parts:

     1. Why are there invalid subnets?

     2. How are they utilized if they are invalid?

[Leo Laohoo]

Are you referring to subnet-zero?

For example, you have a subnet of 192.168.0.0 255.255.254.0.

Your valid subnet is from 192.168.0.1 up to 192.168.1.254.  In older IOS routers, you could NOT use the IP address of 192.168.1.0 until a Cisco released a code which supports a command "ip subnet-zero" which now allows you to use this IP address. 

[Requiemsallure]

Thats right, I knew I had forgotten about the subnet-zero command. I appriciate the reminder. however I have also forgotten what the limitation was that prevented Cisco from using the subnet-zero before the command was established. you wouldn't happen to know why it couldn't be used would you?

in addition, why wouldn't you be able to use a the last subnet as well?

[Richard Burts]

The reason why Cisco implemented the restriction on using subnet zero is that there were some host OS that used the zero address as the broadcast address and not the 255 address. So using subnet zero could be ambiguous.

In those days it seemed like there were lots of IP addresses and there was not much concern about wasting some IP addresses in your subnetting scheme. Of course today our attitude about efficiency in addressing is much different.

I would comment further that there is an absolutely valid reason to say 2^N - 2 when calculating the number of hosts. Many people carry the same calculation over to the calculation of subnets but it really does not work accurately for calculating the number of subnets. There is a historic reason to say 2N - 1 (accounting for subnet zero). There was never a valid reason to say 2N - 2. And the current and correct calculation of the number of subnets is 2N.

HTH

Rick