Topic: Proof of location - page 2. (Read 4824 times)

funny, I almost started on the thread on the topic, but decided it's too crazy. I thought of the same exact concept, because proof of identity seems not to work very well, when one considers facebook clickfarms and the discussions around Mike's passport idea.

The internet was never designed for using location data, and it is fundamentally a system where location does not matter. One domain resolves to an arbitrary IP address and usually one does not care whether the packet travels for 3 seconds or half a second. IP addresses are assigned in a hierarchical manner. service providers get some subspaces and assign numbers to their customers (IANA => ISP => node).

there are several questions that arise:

* how accurate is an IP address as a measure of location? not very. at least say 1-5% of the time the GeoIP location will be off by > 200 km, and on average probably 50-100 km. for some countries with very low distribution of backbones this could be 500-1000 km.
* can one use various pings to verify location? possible, but certainly not trivial.
* TOR. with TOR you can spoof arbitrary locations. which is not the same as having that location as the endpoint (latency will be very high).
* packet traversal has very high variance. this makes a very complicated problem.

I'm not sure anyone has experimented with it. Here is a project which scrapes maxminds Geodata:

http://freegeoip.net
https://github.com/fiorix/freegeoip
http://www.maxmind.com/en/geolocation_landing

to find geolocation on a map:
http://itouchmap.com/latlong.html

Proof of location based on TCP/IP seems not very viable, at least when it is supposed to be universal (for some internet nodes it can be accurate enough). Which leaves such an enterprise more in the area of cjDNS/I2P or GPS/mobile phones. 

A couple of pointers in that direction:

Todd E. Humphreys gave a TED talk on his work on next generation GPS http://www.ted.com/talks/todd_humphreys_how_to_fool_a_gps, http://www.ae.utexas.edu/faculty/faculty-directory/humphreys , where he predicts millimeter accuracy for GPS.

there is also the old idea of internet of things (where a map of location to internet is needed). But:


http://recode.net/2014/03/19/at-ted-sharpening-the-vision-of-the-internet-of-things/

How about this:

Your a mining node, and you attempt to build a distance model of the network based on ping times.

Your model is compared against other models, and the models that best fit together to produce would share a reward.  Therefore, every miners best chance would be to submit the most accurate model they could build.

It would have to be a pretty sophisticated matching algorithm that compared the distance models, and of course it would need to be able to be verified by everyone.

As a miner, the models I submit to the network would need to include some proof of work and my a public key so I could receive and award if I was due.  The distance models would only be relayed if accompanied by sufficient proof of work.

Assuming you are on a fast network I can verify you are in my city by pinging you. The time should take less than the time it would take light to get to the city walls.

That may be too harsh but I could may verify by the network trace or some other details like that.

Now my "vote" for your location can be given gravitas via proof of work or proof of stake.

Unlike the bitcoin block chain you need a proof of work PER location verification. Thats a lot of proofs. Although the difficulty level will probably be a lot less than the one you need to win 25BTC as a miner.

Then the receiver can decide how many proofs he needs to be convinced that you are where you say you are.

sounds like an interesting problem.  have  you thought about measuring relative ping latency

canton you could certainly share more insight from your own research At least on indigogo you claim you have it sorted out, albeit I doubt it is physically possible to have some open source solution that can not be spoofed by some fake gps signal.

I know there's a lot of interest in this topic as it relates to cheating in augmented reality games. There's probably a lot of thought, energy, and maybe even FOSS solutions out there... google for (example) "ingress prevent gps spoofing"

Hello Bitcoin braintrust.  I am working on a project that requires proof-of-location.  A google search yields a number of centralized solutions.  One being...
https://cs.uwaterloo.ca/~uhengart/publications/gis10.pdf
Is this a solved problem or can reported location always be spoofed?
Could there be a blockchain style proof of location?

I would be willing to collaborate with someone(s) if this seems like an interesting problem to you.  I'm looking for an open source solution. 

MrBea