[Gllug] subnetting

[Ian Northeast]

Mark Green wrote:

 > First off judging from what information was given I assume that
 > Michaels professor is introducing subnetting with classful subnetting 
 > first, yes I realize that CIDR has changed things but classful
 > subnetting is often taught first, for example Cisco CCNA deals with
 > classful subnetting before moving on to classless subnetting.

Fair enough. The OP didn't mention network classes at all, but I guess 
it makes sense to teach them first then proceed to CIDR. That is the way 
it happened after all. And it's useful to know about classes since 
everything IME still defaults the netmask etc. to what it would be under 
classful subnetting if it isn't specified.

 > The IP address range being used is a class C, and so the number of
 > bits that will be borrowed from the host portion of the last octet
 > will be 3 bits to give 8 subnets. From these 8 subnets the number of
 > usable subnets us equal to two to the power of the borrowed subnet
 > bits then minus two. The minus 2 is there for the reserved addresses
 > of the network ID and network broadcast.

 > (2 to power of 3) - 2 = 6

Why? Why should whole *subnets* (or networks, as there isn't any 
difference between a subnet and a network any more) be reserved for 
network ID and broadcast? These are individual addresses *within* a 
subnet (as you mention below).

 > Addition of the 3 bits that have been borrowed gives the subnet mask, 
 > in this case the 3 left most bits so 128 + 64 + 32 = 224.

 > Subnet mask = 255.255.255.224
 > Slash format = /27

 > Subnetwork #        Subnetwork ID        Host Range       Broadcast ID

 > 0                   200.100.99.0         .1-.30         200.100.99.31
 > 1                   200.100.99.32        .33-.62        200.100.99.63
 > 2                   200.100.99.64        .65-.94        200.100.99.95
 > 3                   200.100.99.96        .97-.126       200.100.99.127

 > 4                   200.100.99.128       .129-.158      200.100.99.159
 > 5                   200.100.99.160       .161-.190      200.100.99.191
 > 6                   200.100.99.192       .193-.222      200.100.99.223
 > 7                   200.100.99.224       .225-.254      200.100.99.255

 > How ever both subnetwork 0 and 7 are unusable as they are reserved for
 > the network ID and broadcast.

Again, why? I have never heard of such a concept and extensive practical 
experience refutes this. And as I mentioned before, it is not necessary 
to cut a network up into equal size portions. I could cut this one into 
one /26 and six /27s for instance, or indeed something much more 
complex. I could also cut it into two /25s, which by your argument would 
both be unusable leaving nothing available at all. I can assure you that 
this does not happen.

 > The number of usable hosts on each subnet is equal to two to the power
 > of the bits remaining, minus two. The minus 2 is for the subnet ID and
 > subnet broadcast reserved addresses.

 > (2 to power of 5) - 2 = 30

This is true. But why are you applying the same logic to entire 
networks? It doesn't hold.


Oh and BTW by quoting the gllug added .sig in my email without quotation 
indicators and replying below it, you made my mail client regard the 
whole of your response as part of the .sig. Which made constructing this 
reply so hard I almost didn't bother. I suspect some people will have 
missed your reply entirely.

Regards, Ian

-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug