said by Bob4: said by RARPSL:
one says I am in Mahwah NJ (which is where my local Headend is). My actual location is in Suffern NY
I don't see how any IP address database could be better than that. There's no way for them to know where beyond the head-end the user is actually located. An individual IP address could be distributed anywhere in the head-end's coverage area.
I agree. I also stated that I was willing to accept the Mahwah "location" as accurate. The problem is that depending on which database the site uses, it can be told a totally inaccurate location (the Connecticut or Long Island one in my case). As to your suggestion that there can be no more granular location than the head-end's location, I do not agree (at least technically). Just because the head-end services multiple areas (Mahwah NJ and Suffern NY in this case) does not preclude having separate IPN blocks for the two cities. When my modem's MAC address is presented to the DHCP server, it is looked up and is used to select Block1 or Block2. Since the modem must be registered, the registration can include an indication of which IPN block to use. So long as the database lists a unique location for each block, you can get more accurate location information. I acknowledge that this split method can cause problems with node splits as well as IPN address allocations. Once you have more Suffern customers than there are IPNs in the Suffern block you must create an additional Suffern block even if there are still available Mahwah addresses that could have been used.
Note that this analysis is simplistic but I think it can serve to show the principle.