Clarification with IPV6 and previous

[mark at buttery.org (Mark J. Dulcey)]

kgleason.ma.ultranet at rcn.com wrote:
> Can anyone tell me the maximum theoretical number of addresses with IPv4 as
> opposed to IPv6? Is is my impression that with IPv4 you would have
> 256*256*256*256 (or am I wrong)?

Correct in a certain sense, but wrong in another.

First, there are some addresses that are reserved for special purposes. 
Three ranges are reserved for local (non-Internet) addresses: 10.x.x.x, 
192.168.x.x, and 172.(16-31).x.x. 224.x.x.x through 239.x.x.x are 
reserved for multicasting. 240.x.x.x through 255.x.x.x are "reserved" 
for unspecified purposes. 127.x.x.x is reserved for the loopback network 
(in other words, talking to other programs on the same machine). 0.x.x.x 
is reserved for "this network" (in other words, machines on the same LAN 
as the computer sending the packet). Finally, the addresses where the 
local address parts are all zeros and all ones are unavailable, because 
they are reserved for special purposes.

In addition, it's impossible to achieve 100% utilization of the address 
space. There will always be some wasted addresses sitting around on 
people's networks.