You don't need a subnet mask in the network specification (unless you have overlapping subnets, which is bad design to begin with - ie. one system with 192.168.4.17/24 and another with 192.168.5.93/23 - they are both in network 192.168.4.0). The client IP and client netmask are used to determine the network address.
Pick any IP and netmask. Logically AND the two - the result is the network address. That address is placed in the lookup table, so that any other combination of IP and mask that generates the same network address will use the same lookup value.
In your example, the network address for that subnet is 172.16.15.128, and only computers in the range 172,16,15,128-191 will generate that network address. Maybe it would make more sense to you if it were called the "subnet address" and not the "network address"?
Glenn
_________________________
Actually I
am a Rocket Scientist!
