Comment by mac3n
4 days ago
having worked on IP geolocation in the past, I don't think this works. Though it can do a pretty good job of getting you in the right continent.
* Not all traffic goes through fiber - there are microwave links operating closer to the speed of light, though these are mostly reserved for high-speed trading. There's also satellite connections, but as long as they don't do satellite-staellite, they're slower.
* There are middleboxes messing with traffic, especially TCP, which add delay.
* If you rent servers in datacenters, you might not really know where they are. We had VMs relocated without our knowledge.
* Fibers links aren't direct, they tend to follow public right-of-ways. In much of the US, that's a rectangular grid along the highway system (look at a road map of the midwest sometime), increasing the delay by √2.
* Internet routing isn't shortest-path. It's get-this-crap-off-my-infrastructure, aka hot-potato.
* Anycast prefixes have IPs in multiple locations.
My experience was that with a lot of observation points, you could get within 10ms, 1000km in most places.
I work for IPinfo, and we run active measurements through hundreds of servers. Allow me to highlight the fantastic points you made.
> Not all traffic goes through fiber - there are microwave links operating closer to the speed of light, though these are mostly reserved for high-speed trading. There's also satellite connections, but as long as they don't do satellite-staellite, they're slower.
That is a valid point. In that case, we have to fall back on the geofeed, WHOIS, or other methods of geolocation information. We are actively researching this area, though.
> There are middleboxes messing with traffic, especially TCP, which add delay.
This accounts for some problems but not all of them, as we have multiple servers running active measurements on individual IP addresses.
> If you rent servers in datacenters, you might not really know where they are. We had VMs relocated without our knowledge.
That is normal. At the scale at which we operate, server location validity is extremely important. We run daily checks and actively flag these issues. If things don't add up, we communicate with our vendors and try to understand whether it is a network-related issue or something else.
> Anycast prefixes have IPs in multiple locations.
Yes. With anycast IPs, we have hints of all their available locations, but when it comes to picking one, we default to the ASN-reported location.
> My experience was that with a lot of observation points, you could get within 10ms, 1000km in most places.
We have been significantly reducing the number of ASNs where we have high RTT by getting a server "networkly" close to them.
---
I understand that this is not the absolute best location system possible, but within the scope of our industry, we are miles ahead of everyone else. We are continuously investing in research and infrastructure to improve our data even further.
> there are microwave links operating closer to the speed of light, though these are mostly reserved for high-speed trading
This is so sad.
It's really not. The microwave links got decommissioned everywhere because nobody NEEDS that higher fraction of lightspeed. High speed trading is the only field where saving a singular millisecond is economically rewarded. The links used by high speed trading are the only ones left.
Sure, but if it becomes ubiquitous, web devs will assume lower latency. That wouldn't make it less sad, just makes different people sad - my first guesses are those at crowded areas with overloaded cellular connections, and Australians.
I think routing not being shortest path (nor being consistent) is the biggest issue with this method.
Are you sure it'd only be accurate up to 1000km? I'm not as experienced but would have assumed that by sampling a dozen times, you would have at least 95% likelihood of a 100km radius
I don't think microwave is "closer to the speed of light" than.. light.
Electromagnetic radiation is about 1.5 times faster in air than in glass (depending on wavelength a bit).
I didn't think about it slowing down when traveling through glass, good point. Although hollow-core fiber is a thing now.