> As you are assuming physical identity of DHCP clients are known > reliably (I don't know how it can be done), it is a lot easier > for DHCP servers to assign static IP addresses to the clients > based on the physical identity information than assigning random > addresses and map them later through complicated database look up. As an example, consider a system we built for the IETF meeting network a few years ago. The server queried a series of tables inside of NetDisco to map an IP address to the WiFi AP that it was connected to: IP address --> MAC address --> Connected AP Since the locations of the APs are known, this gives you the location of the endpoint to within a few tens of meters (empirically, within the IETF network). >> This is an important use case, for example, for ECRIT >> emergency services [draft-ietf-ecrit-framework] > > Is it a use case in the real world? The US FCC thinks so: <http://edocket.access.gpo.gov/2011/pdf/2011-565.pdf> --Richard _______________________________________________ Ietf mailing list Ietf@xxxxxxxx https://www.ietf.org/mailman/listinfo/ietf