Is there a good reason for that? Having an "assigned pool" of addresses
only adds an additional level of indirection which is unnecessary.
The way it seems to work now is that the first connection to *any* port is
assigned the first address in the pool. The second connection, again on
any port, is assigned the next address (if it is available, of course).
The third, the same, and so on.
Why do it that way? If each port had its own assigned address, such that
the port number could be directly determinable by subtracting the IP
address from the base address for the pool, then it would be easier to
determine the port a user is currently logged in on.
An example. Our terminal servers' ports are named by the base name of the
terminal server plus two digits. Given the terminal server "godzilla",
for instance, the base address of the pool is "godzilla00". Currently, if
I know a user is logged in on godzilla00, I still have to search through
the user list on the terminal server to find out which port he is on,
instead of immediately knowing he is on S0.
Seems unnecessary to me at least.
-=- julian