It'll be a small percentage of IPs, netblocks or ASNs, but it'll be the
most important ones.
Nobody cares if Verizon users are correctly placed 99% of the time, if
any of those users can sign up to NordVPN and use it for the 1% of the
time they want to get around any geoblocks. Correct location is
all-or-nothing. If you don't get the adversarial cases right, you may as
well get none of it right.
Right now, it's the residential proxy industry, but it's very expensive.
How will you detect a Verizon user is proxying traffic at the TCP layer
for someone in Saudi Arabia? That's very difficult, as they won't proxy
your ping packets. You'd need to be an MITM to measure layer-7 latency.
If residential proxy prices were to fall, the entire GeoIP industry
would be exposed as snake oil. Though I think it's more likely that
media companies would team up with ISPs to hunt down proxies with a
vengeance. This is already what happens in places like Spain where the
same company owns the football games that are getting pirated and the
biggest ISPs, and they block half the internet during live football,
just in case.
On 29/01/2026 03:32, Jon Lewis via NANOG wrote:
On Wed, 28 Jan 2026, Mike via NANOG wrote:
Using your own measurement is fine but where it gets a push back from
some of us, is your statement that geofeed data is treated as tier 2.
I for one publish geofeeds in an automated fashion. I would like
this data to not be considered a red-headed stepchild.
"Some people might lie in their geofeeds" seems to have poisoned their
view of geofeed reliability. I really suspect the "geofeed fraud" is
a small minority. Other than low end VPN providers hoping to fool
streaming content providers, I don't know why anyone would lie in
their geofeed.
Ours is automated too. We use an IPAM that allows us to tag every
subnet with a "location", and each location has a full street
address. The coworker who setup our geofeed wrote a script that pulls
all the allocated subnets and their locations from IPAM and builds the
geofeed. As long as we don't screw up in IPAM (which happens), the
geofeed is automatic and correct. I did end up having to write a
geofeed auditor that catches when someone screws up and emails us so
the improperly formatted (and improperly parsed) locations get fixed
almost immediately.
The headaches of IP Geo providers "getting it wrong" impacting our
customers is good motivation for making sure we publish accurate data.
----------------------------------------------------------------------
Jon Lewis, MCP :) | I route
Blue Stream Fiber, Sr. Neteng | therefore you are
_________ http://www.lewis.org/~jlewis/pgp for PGP public key_________
_______________________________________________
NANOG mailing list
https://lists.nanog.org/archives/list/[email protected]/message/WPNYIT75HYA2USZXBSELCPPHULT4E53S/
_______________________________________________
NANOG mailing list
https://lists.nanog.org/archives/list/[email protected]/message/DKIP6OPH33XI4FBSL37ELPQH52G2OYRW/
