ietf
[Top] [All Lists]

Re: wireless geolocation

2017-06-06 18:56:56
I feel pretty odd about this. On the one hand, I think GeoLoc is
counter to <whimsy, belief, emotion> but I also recognize that it
happens, and a lot of things now depend on it. Stupid things like DRM
region lock, but also other things like optimal content source
selection, or language for an interface.

What do I want? I want less stupid. So I want either the DRM to die
(gonna happen. can I have a chocolate pony too?) or the Geo selection
method to work better.

But what my day job tells me, is that adding 'more' people to offer
ways to declare Geo, doesn't "fix" this problem. We, the RIR tried. We
didn't entirely help. Lots of reasons this turns out not to be
entirely useful. Some people say they don't want to believe what we
declare. They don't want to believe what Internet Address holders
self-declare. They only want to believe what they can measure,
triangulate, learn from <secret sauce> like BGP (oooooh! super
secret!!!) in ways which we don't understand. Or, from what websites
say. or DNS says. I know, lets put it in crypto, in the DNS...

And it also tells me that people have strong motivations to "mislead"
about their Geo, because of .. well.. basically the dynamics of the US
content industry and consumer choices. When it comes down to it if you
can demonstrate that a caribbean island resort hotel secured a
financial advantage having US located IP addresses and customer-choice
drove it up the price table (story I was told) then.. you know this is
a "thing" you can't control.

Really, I think if we stepped back, "geo information" is pretty
interesting. It feels like something we should have some ownership of.
But stepping forward, its been a tar pit, the entire time I've engaged
in it. Nobody can point me to a subject more likely to descend into
hell. Ok. maybe the DNS. And routing. or operations. I dunno. Its more
political than SNMP.

If something as simple as changing the BSSID stops the stupid, I'd say
add it to the playsheet. But it sucks this is likely to be workable,
and it sucks that something as silly as a Layer-2 device unique ID is
being tied to a location, for an agency which moves around the world.

Model be broken. Have no fix.

PS on a personal level I like more specifics in WHOIS, and
self-declared in the DNS and websites, I don't like super-secret sauce
methods, and I dont like people who re-publish public data as
privately tuned data, but are non-responsive on their 'fix problems'
email address. You know who you are...

On Wed, Jun 7, 2017 at 9:44 AM, Nico Williams 
<nico(_at_)cryptonector(_dot_)com> wrote:
On Tue, Jun 06, 2017 at 06:38:00PM -0400, Donald Eastlake wrote:
bssid/macs are manufacturer based but there are companies that provide
positioning services based on learning where those MACs are -- typically
combining such MAC based Wi-Fi/Bluetooth information with IP, GPS, ...
information to try to get an accurate location. If you happen to be in a
hotel function room and can only see meeting provided APs, so you can't see
any fixed local APs, you can easily get Geolocated to the previous meeting
site.

Can the BSSIDs be changed?  I know, that might be disruptive in that
users will have to re-teach their devices about trusted APs, but that
seems like a decent trade-off.


<Prev in Thread] Current Thread [Next in Thread>