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

[Tom Beecher via NANOG <nanog () lists nanog org>]

> We currently PoP presence in ~520 ASNs. I think this fact largely
> demonstrates that no VPN companies, nor any CDN companies, nor any
> companies outside of crowdsourced research network have reached the level
> of network diversity that we currently have. It is not 1,300 X $5 .


Even low end CDNs have more reach than that.   Let's move on though.

- ISP publishes geofeed of their IPs.
- ISP starts getting support calls from customers that go to WEBSITE, and
it's the version from another country.  Techs spend time troubleshooting
this. They reach out to WEBSITE. After much back and forth, it's
determined that WEBSITE is using geolocation data from GEO-PROVIDER.  >>
ISP has spent time and money dealing with the support problem. <<
- ISP has someone reach out to GEO-PROVIDER to get the data corrected. >>
ISP has spent time and money fixing the incorrect geo information . <<
- At a future point in time, GEO-PROVIDER reverts to the wrong information,
because they decided their data is 'better' than what the ISP published.
- ISP starts getting support calls from customers that go to WEBSITE, and
it's the version from another country.  Techs spend time troubleshooting
this. They reach out to WEBSITE. After much back and forth, it's
determined that WEBSITE is using geolocation data from GEO-PROVIDER.  >>
ISP has spent time and money dealing with the support problem. <<
- ISP has someone reach out to GEO-PROVIDER to get the data corrected. >>
ISP has spent time and money fixing the incorrect geo information . <<

Every time an ASN has to deal with a support issue caused by incorrect data
, it costs them time and money.



On Wed, Jan 28, 2026 at 10:21 PM Abdullah DevRel of IPinfo via NANOG <
nanog () lists nanog org> wrote:

> Hi,
>
> I think I failed to demonstrate the scale of the ProbeNet operations.
>
> For us, having a handful of servers in each data center is enough and will
> reduce the RTT of the peering ASN and nearby networks. So, we constantly
> need to expand to other diverse networks. It is pretty easy to buy a few
> servers in Western Europe, the East coast, and the South coast of the
> United States. However, when you want to buy a server in the Caribbean,
> Oceania, or African countries outside of South Africa and Nigeria, I think
> the procurement effort, maintenance, and service costs are much more
> complex than you can imagine.
>
> We have some servers that cost $5, maybe a dozen. However, I really don't
> want to use our financial investment in ProbeNet as a scale to show the
> complexities of running an operation with 1,300 servers
>
> We currently PoP presence in ~520 ASNs. I think this fact largely
> demonstrates that no VPN companies, nor any CDN companies, nor any
> companies outside of crowdsourced research network have reached the level
> of network diversity that we currently have. It is not 1,300 X $5 .
>
> > There are (a) geoip data providers, (b) geoip data users/buyers, and (c)
> networks in the middle that bear much of the cost of what a and c are doing.
>
> I am not sure I understand your logic:
> - (a) IP geolocation data providers: IPinfo
> - (b) IP geolocation data users/buyers: IPinfo Customers
> - (c) Networks in the middle: You
>
> What (c) Networks in the middle are bearing the cost of (a) IPinfo and (c)
> you are doing. I am not sure I understand the logic. What cost are we
> talking about?
>
> — Abdullah | DevRel, IPinfo
> _______________________________________________
> NANOG mailing list
>
> https://lists.nanog.org/archives/list/nanog () lists nanog org/message/73CUGUBCU5FEHANW3I6AXJSHNROSASVD/
_______________________________________________
NANOG mailing list 
https://lists.nanog.org/archives/list/nanog () lists nanog org/message/JC3KNUNP7HISW5FNBJEY2R3PWFQNDDMD/