RE: Geofeeds are good — was Re: Publishing BGP communties for your network (Re: What's up with BGP communities?)

[Gary Sparkes via NANOG <nanog () lists nanog org>]

Therein lies the problem though - mind you, I'm speaking from past experience, not current data, but still.....

In that three potential scenario, it may be 6 months to a year between failovers, but in every case, all three 
termination/announcement points aren't reflective of the actual end users/site the addresses are actually utilized 
under.

So measurement may see it as Ashburn VA for a long time, even though the office/site utilizing that range is actually 
in Boston, MA as reflected in the geofeed.

That's where it becomes a problem. 

I know a company that announces out of Ashburn, VA and Houston, TX as well, they have a slew of ranges and sometimes 
fail over between the two if tunnels fail or whatever other scenarios, planned maintenance, etc, but usually one side 
of the US goes through Houston, and the other through Ashburn. But the geofeed covers ranges and locations in all 50 
states. (All sites tunnel back to the datacenters before going out their respective NAT pools and what have you before 
hitting the 'public' internet)

-----Original Message-----
From: Abdullah DevRel of IPinfo via NANOG <nanog () lists nanog org> 
Sent: Tuesday, January 27, 2026 10:13 PM
To: nanog () lists nanog org
Cc: Abdullah DevRel of IPinfo <devrel.ipinfo () gmail com>
Subject: RE: Geofeeds are good — was Re: Publishing BGP communties for your network (Re: What's up with BGP 
communities?)

Gary Sparkes wrote:
> > I work for IPinfo. We use active measurements for IP geolocation as 
> > a primary source of data. We consider geofeeds as a second-tier 
> > source of information. We use it as a fallback data source when 
> > active measurements fail.
> > And for a lot of my users/systems/address ranges, you will be entirely incorrect using those measurements. 
>
> Let's say your measuring endpoint is in Baltimore. Let's say I'm announcing out of Raleigh for whatever reason. 
>
> Let's say the end users of that ... ah, let's say, /24, are all in Baltimore, and that range is only used for 
> Baltimore people.
>
> Are you going to pin me in NC (incorrect) or Baltimore (correct, as my geofeed publishes)?
>
> Or in the case of a /24 that's used as an endpoint for a facility in Boston, but can be announced out of Houston, Los 
> Angeles, or Ashburn depending on the failover scenario, my geofeed publishes that range as Boston, not any of those 
> three cities, because that's where the users and end network actually is. 
>
> So geolocation provider data is horribly inaccurate when it comes to the end-site's actual location for that address 
> range. 
>
> Same I've seen with other places, as well. 
>
> -----Original Message-----
> From: Mike via NANOG <nanog () lists nanog org>
> Sent: Tuesday, January 27, 2026 12:11 PM
> To: North American Network Operators Group <nanog () lists nanog org>
> Cc: Mike <deezknuts () yahoo com>
> Subject: Re: Geofeeds are good — was Re: Publishing BGP communties for 
> your network (Re: What's up with BGP communities?)
>
>  In fact, geolocation providers should be using geofeed data as tier 1 data since they are self published.  Are you 
> saying you know more about my prefixes than me? 
>     On Tuesday, January 27, 2026 at 07:03:45 AM PST, Ca By via NANOG <nanog () lists nanog org> wrote:  
>  
>  On Mon, Jan 26, 2026 at 11:47 PM Abdullah DevRel of IPinfo via NANOG < nanog () lists nanog org> wrote:
> > Geofeeds are useful because they cut out the middleman in terms of 
> > geo info, and provide a mechanism for someone to detect and react to 
> > that change quickly, since a very large use case for these is 
> > targeted content and advertising.
> > I work for IPinfo. We use active measurements for IP geolocation as 
> > a primary source of data. We consider geofeeds as a second-tier 
> > source of information. We use it as a fallback data source when 
> > active measurements fail.
> > The honest reality is that geofeed does not provide verification 
> > metadata of any form.
> > The “honest reality”, as you say, is that geolocation firms like yours have failed to provide reliable data to 
> > paying customers for years.
>
>
> That is why the IETF made geofeeds.
>
> My customers started having outages because geolocation firms have bad 
> data, and enterprises use that bad data in firewall and cdn rules 
> which cause outages. For example, geolocation firms provide data that 
> a customer IP is in xyz country but the firewall rules only allow abc 
> country…
>
> Anyhow, as a person who publishes geofeed data representing 100s of millions of users, please …everyone… publish and 
> consume first party geofeed data and do not listen to FUD from people trying to sell you the same data that we 
> publish for free.
>
>
> They can be stale, wrong, or false, and maintaining them is a pain for 
> ASN
> > providers. Many large telecoms do not even have publicly accessible 
> > geofeed. For example, AS11260 (Eastlink.ca) does not have a publicly 
> > accessible geofeed to my knowledge, and I am not sure who to reach 
> > out to get one. Not all ASNs are going to publish geofeeds, and it is fine.
> > Providing IP geolocation data has real value, and I do not believe 
> > it is the ASN or ISP's responsibility to have their data accurately 
> > reflected on third-party IP geolocation providers like us.
> > The ideal operation that we have is this: we do IP geolocation based 
> > on building networks of servers running active measurements, working 
> > with ASNs, attending conferences and talking with as many ASNs and 
> > range operators as possible. Shake as many hands as possible. It 
> > sounds borderline impossible, but honestly, that is the only fair 
> > way to operate as an IP geolocation provider.
> > If, in any case, an ASN does not want to help us, that is fine, but 
> > that does not excuse negligence on our part towards end users. Our 
> > responsibility is towards end users (the customers of ASNs; internet
> > users) and ASNs can be just voluntary partners. The moment an ASN, 
> > ISP, or an end user has an issue, we have to jump into the conversation and fix the issue.
> > That is why I am here. There is a mention of the word "geolocation" 
> > in the NANOG forum, and we have to come here and ask if everything 
> > is okay. Do you have any issues with us? Can you check your data with us?
> > Can you help us fix the issue?
> > This MO of IP geolocation as a service should be: aggressive outreach. 
> > But we also have to admit that we are not the only provider, and we 
> > are not even the largest provider out there. But we do not care, we 
> > have to be responsible to everyone. We have to back our data.
> > — Abdullah | DevRel, IPinfo
> > https://www.linkedin.com/in/reincoder/
> > _______________________________________________
> > NANOG mailing list
> > https://lists.nanog.org/archives/list/nanog () lists nanog org/message/
> > SD
> > OEDEZ7UUM4FZNHQJQBFCIR4W3HEP2X/
> NANOG mailing list
> https://lists.nanog.org/archives/list/nanog () lists nanog org/message/3KXUUIAG...
> _______________________________________________
> NANOG mailing list
> https://lists.nanog.org/archives/list/nanog () lists nanog org/message/EJX4QAO3...


If we are providing inaccurate data for you, please talk to us. Please understand that our active measurement has 
inherent fail-safe mechanisms. When the data is noisy, we will fail over to geofeed. We collect ping and traceroute 
data for IP addresses through servers hosted in various cities. We will not accept noisy data to generate IP 
geolocation data just because. The entire idea is based on a verifiability hierarchy.

If a customer asks why an IP address is located in a certain place, we can show them our ping data, traceroute data, 
etc. If we see an average of 20ms+ RTT (for example) to the IP address, we will not use that to produce the IP 
geolocation data. That is not justifiable evidence for the geolocation.

Then we can fall back on the geofeed data and state that the IP address operators/owners announced that the IP address 
is located there.

Geofeed is great as a fallback data source but it is not our primary source of data.

Let me know what you think, please. Thank you very much.

— Abdullah | DevRel, IPinfo
_______________________________________________
NANOG mailing list
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/GW6VQ5CABBFOQDGRIAGLQ76WBZFW5GXO/
_______________________________________________
NANOG mailing list 
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/WHREZNHNIGR4GD65YX5DZQ7DMZNEZ7XR/