Hi Neil!
 NC> OK Mike, I have to admit a gaping hole in my knowledge of IP here. What 
is
 NC> the difference between a class A, B and C address?
As you know IP addresses are made by four blocks of 8 bits. It is easier to 
call them with a number from 0 to 255 but those are really 8 bits, and that 
is reflected in the masks.
You can see the block of 8 bits like: 11111111 where the first 1 in the left 
represents 128, the second 64, the third 32, then 16, 8, 4, 2 and 1.
So 10000000 is 128, 01000001 is 65 and so on.
Class A: The first 8 bit block represents the network numbers. The rest of 
it, are the hosts. Class A addresses go from 1 to 127 in the first 8 bits 
block.
This means that there are only 127 class A networks and each one can have up 
to moreless 16 million hosts. (indeed 127.x.x.x is reserved for loopback, so 
there is one less network)
Representing it in a binary view the first block of a class A goes from 
00000001 to 01111111. (the first bit is always 0).
Class B: The first and the second 8 bits blocks are the networks, the rest, 
the hosts. Class B addresses go from 128 to 191 in the first block, so we can 
say there are around 16.000 networks class B and each one can have up to 
65.536 hosts. (this numbers are not exact because there are reserved numbers, 
you cannot have a 129.x.0.0 for a host because the 0.0 is reserved to call 
the network itself, and the 255.255 is reserved for the mask stuff)
Representing it in a binary view: from 10000000 to 10111111 (128 to 191)
(the first two bits are always 10).
Class C: The first, second and third blocks are used to name the networks and 
the last block is used to name the hosts. Class C addresses go from 192 to 
223 in the first block. There are something like 2 million class C networks 
with up to 254 hosts each.
In a binary view: from 11000000 to 11011111 (192 to 223) (the first three 
bits are 110)
The rest from 224 to 239 is reserved for multicast packets (class D) and from 
240 to 255 (class E) is reserved for future addressing modes.
 NC> Also, how exactly does subnetting down work out?
Subnetting is used to divide a network in smaller pieces, so you can make 
better use of the addresses you have.
Suppose you have a class C address, lets say 200.0.194.x (thats my provider's 
:) You can have up to 254 hosts with this address, but let's say you have a 
building with 60 hosts and another with 100, so you want to split the address 
in two networks. You can use every bit of the last block, so you can assign 
from 00000001 to 11111110 (0 and 255 are not used).
To split this you assign the first bit to the network part of the address, 
and keep the other 7 bits for the hosts.
So you can have a subnet from 00000001 to 01111110 and another from 10000001 
to 11111110. the first one is the subnet 0 and the second one is the subnet 
128. and this is what you put in the mask, 128. The mask will be 
255.255.255.128 in this case.
 NC>  I know my employer has 10.16 addresses on a 255.255.248.0 subnet
 NC> mask which gives 10.16.72.0=10.16.79 on my site and I know it's
 NC> something to do with bitmasks but I start to lose it a bit here which
 NC> could prove embarrasing if anyone asks.
This is a class A address, so the default mask is 255.0.0.0, when you use 
255.255.x.x you are at first subnetting the second block in 255 networks.
Then the 248 in the third block means you divide the bits like this:
11111000. The first 5 bits are for network number and the rest for the host 
number, so you can have a lot more networks and less hosts in each one.
Am I too puzzled? :) Hope you get some of what I tried to tell. :)
Best Regards
Luis Manterola, CNA 4.11
lam@rivero.newage.com.ar
---
---------------