Google knows because you are connected through the local router.
You requested information from their system, that information had to have a tag on it to allow the answer to come back.

Think about it: Request came from your machine, connected to a local area network, which is wired in to a location.

In a way, it is like calling someone on a phone with caller id and wondering, when they pick up the phone and answer, why they call you by name.

Best
Les

Not broadcasting the SSID is not a security measure and could actually prove to be a pain. The RFC requires the SSID broadcast. Some devices don't care, but other devices will be really unhappy with that configuration.

My phone will see wifi signals that have SSID broadcast disabled. Besides, a MAC address (which cannot be disabled) is a better tool to use than the SSID anyway.

By disabling the SSID broadcast, you aren't really helping yourself.

Just a somewhat related tidbit.

Houston TranStar and Bluetooth Traffic Monitoring

Sorry guys, but I believe that both of these theories would be a "static" view (stored in a db) that wouldn't relocate the user as they moved around. oleE is closer to the truth (with a decent subnet mask on the ISP side to pinpoint location), but I don't know how they are seeing location moves on the "private" side of a WAP/router (as VaSteve is talking about inside his house) Then I would suspect smartphones, etc. Just my .02 worth...

Hmm, the problem is that I don't think most ISPs provide geographical<>IP based information (or even have that information).

So, what used to be Southwest Bell and is now ATT has a huge bunch of IPs that covered several states. In Houston and it's suburbs, there are probably a ton of local offices (100, more?) that end users connect to. ATT probably has a document someplace that says "these 3 class C subnets are hosted out of this office, these 10 are hosted out of this office over here." I'm pretty sure there is no reason for ATT to publish this information anywhere. Not only that, but that's only going to narrow you down to a specific office, which means that you could be within an area a mile wide around that office. Since the IPs are dynamic, I don't think they have any way to determine what address or even within a few hundred feet each individual IP is located. Frankly, I don't think they don't care. There are websites that will show you where you are based on your IP. The IP on the router at my house puts me in Houston which is 20-25 miles off.
Try it, how close is this to you based on your IP
http://www.infosniper.net/
http://www.geobytes.com/IpLocator.htm




Wifi Positioning System info
Wardriving
Skyhook: How It Works > Overview
Wi-Fi positioning system - Wikipedia, the free encyclopedia
Google Gears enhances geolocation with WiFi positioning

In my experience ISPs move subnets around within their networks as needed to serve the areas they have service. Over time they do have to renumber the ranges to adjust for expansion and contraction of the customer base. They have to keep the IP allocations something like 80% utilized or ARIN will start asking for them back.

Now, how did Google find you?

Most likely it is the wireless signal. We even do this in an Enterprise or business network. We track access points by their MAC addresses mostly because that is the least likely thing to change under normal circumstances. So from a 'rogue' AP stand point you can 'know' all the access points in your network by their MAC addresses. When an unknown AP starts sending a signal on 2.4 or 5 Ghz the monitoring software will see it and unless the MAC address has been spoofed it'll show up on the map of the office floor plan with an approximate location identified as a rogue.

Google street view trucks/cars/bikes had wifi sniffers on them and recorded the signals they heard and where they heard them. Two things are important there, a third is a bit unnerving. They got that you had a wireless AP of some kind because it was broadcasting a signal (BSSID or not) and they got the general location of that signal via their own GPS systems on the Google streetview recorder. The third thing they got which is unnerving is the data that was being transmitted wirelessly at the time. :D It's likely not that well encrypted because consumer grade devices don't use fantastic encryption methods.

So, it's pretty easy for Google to tell where you are and it really doesn't matter how rural you think you are - if they have a 'street view' view for your street then they know where you are.

The company I work for makes appliances and software that does this sort of location tracking for business wireless networks - it's cool stuff and I like it generally. :D

Methinks you guys are WAY overthinking and over-analyzing this. What if wireless isn't part of the equation? It's just basic IP routing and subnet masking imo...oldE knows :). I'm outta here...happy new year!

Google doesn't have complete knowledge of the return path to the user.

Too be honest - upon further thought I don't think they would even know that the user is behind a wireless router straight from the HTTP information in the exchange between them and the user. The IP information just isn't enough because the DHCP leases handed out by the ISP doesn't have anything like GPS or street address information in it. PLUS Google doesn't have the DHCP lease information and the ISPs don't have any reason to tell google who is where. The DHCP is just a mac address of the end station connected to a cable modem or DSL modem. In our cases generally that would just be the outside interface of our home router. Again - from that you can't tell ANYTHING about the interior network. Not the number of hosts, not how they are connected - nothing. Just that there is at least a single host and they are requesting stuff from the internet and what not. You also can't tell what the interior hosts mac addresses are because in the path of the packets the requesting MAC addresses (the source mac) is changed to the next hop router's mac address along the path. In the HTTP packets however there may be some 'stickiness' defined based on the mac address or ip address (in that case it would even be the internal ip address not the outside public).

I could provide diagrams and packet captures if you like.

With all that in mind - I have no idea how they know where you are. ;)

I had a fella that works for the same company come talk to us about your new at the time wireless stuff about 1-1.5 years ago. Very cool stuff in that the devices can map all of the RF interference sources in the area. Very cool stuff indeed.

OK, so even simplified, there is a LOT of info/stuff that's got to come together to locate you down to the level of your house, or even within a 4 house radius.

Central office A has 2 Class C subnets assigned. They have a router that is performing the DHCP relay function for all of the folks in the area. There are 500 wires that show up in their office from the area. Those 500 wires hit the DSLAM and at any one time, because of DHCP any of those IPs could be going down any one of those wires and on different days, the IPs could be going down different wires. So, your house has to be within 5000 feet of the office for DSL to work (assuming copper phone wire, not fiber U-Verse or cable). The telco knows that they have those two class C subnets that are hosted off of that router, but your house could be 5000 feet in any direction from that router and is one of 500 other folks. That, and the IP is randomly assigned to different folks from time to time.

Do you think that the telco keeps track of which wire the IP is attached to and publishes that information to Google? Yes, if the telco wanted to track it down they could. Their DHCP servers have your router's mac address and may or may not have a username that's associated with your account. Do you think they mine their DHCP servers and publish that information for other customers to see?

Nope, it's not IP subnetting. That may be looked at, but it's far too vague to be used to pinpoint unless you've got systems designed to do it, and based on my work as a network engineer for a service provider and a few Enterprise businesses, I'm pretty sure, they don't look at the info unless they absolutely have to. If the FBI comes knocking, they have to be able to provide it, but it's just held in a database that never gets looked at.

Do smart phones ping for RFID as well?
There is a trend by manufacturers to embed/hide these markers in common household products.

If a friend visited your residence, their smart phone could relay that you prefer gillette razors and jourdache jeans, and that those products are now located at "x" coordinates (but yesterday were at the office).

I don't believe so, not yet.

Steve, take off your network engineering hat for a minute. Google is a repository of information that I'm sure is "data mined" extensively. You realize that every connection to Pelican is also connecting to Google severs as are many other high volume sites and I'd bet they're tracking the info which is flowing back and forth which is source/destination IP addresses (no MACs, no SSIDs, etc.). IP is pretty simple, it doesn't take much of a stretch to foresee them getting a pretty accurate "network picture" imo. It simply HAS to be IP based at it's fundamental core with a lots of the blanks filled in by other techniques (smartphone, etc.) mapped overtop...I'd bet on it...a whole quarter, that's how sure I am :). This is an interesting discussion however...you guys are making me put on a hat I haven't worn in quite a few years and I'm sure I'm a bit off has my head has shrunk. Thanks for making me think...if I'm wrong, then shoot me and give me my .02 back :)

KC - the IP address I'm using maps to a hubsite owned by TWC about 7 miles away from my house. How do they get my location from my IP address?

Mike, think about it in reverse (from Google's perspective)...all they have is a destination IP address to start following the trail with. That info (along with all the info they've "data mined") I would presume gives them a pretty good picture, and of course I am making ASSumptions :). So back to the OP's original question, the answer is still: Based upon the IP address... (with lots of the blank spaces filled in from Google's "big brother" dynamic collection of info). It simply can't be any other way based upon the info they have to start with ...oldE knows :). Think about it...

The thing that bothers me the most is that the fiction part of reality had some thought put into it. For entertainment we have "The matrix" and "Skynet" in reality we are being mapped and tracked by "google"
We are being totally screwed out of a cool and meaningful name to lead us to "Big Brother"

Rick, if 'big brother" causes you concern, just think of Google as "big mama" or "big daddy" :).

"All your base are belong to us...."

Being taken over is just a matter of time but lets come up with a cooler name.....

I'll leave that to someone with more artistic creativity than I, since that side of my brain doesn't work nearly as well :). What ya got in mind?

Well we need something that encompasses the total control of our actions, locations, and all aspects of our lives, I was thinking something along the lines of "wife"
Beth suggested "Mother-in-Law"

An overview from Google on how they find your location in Google Maps.

Privacy and My Location - Maps Help