How we use our public data to dogfood for ProbeNet PoP discovery

To find a server in a particular country, we do not usually need to use complex internal data. We have a ton of data, but you can pretty much discover a hosting provider using our publicly available data.

Let’s start with the country pages

So, let’s say you need to purchase a server in Ghana. First, visit the Ghana Country page from IPinfo.io.

Start with the ASN section

https://ipinfo.io/countries/gh#section-asns

Usually, I would say to look into ISP, Hosting, and Business categories (we include IT services companies as businesses sometimes).

Now, open up all the ASN pages one by one and visit the ASN domains.

Now, if you are lucky, they will have a self-serve. If you are not, contact them via email or form.

Hosted Domains

After you have gone through the ASN pages, you can now look into the hosting domains section for the country pages.

In the IP address pages, you will be looking at several things to help with your discovery.

rDNS Hostname Domain

Abuse Contact Domain

Company Data

Hosted Domains

Visit those pages and see if any of them look like a hosting service.

Ranges information

If the IP address-based information does not seem interesting, it is time for you to check out some of the range level WHOIS. You can start by looking at the domains in the IP address parent range.

Collect interesting domains and then return to the IP address’s ASN.

Then sequentially check the WHOIS records for each IP range block.

That is usually how we can discover true hosting services in a country.


Internal data

If we want to tap into our internal data, we can use hosted domain data and perform a keyword search for domains whose hosting IP address is geolocated in that country. These keywords can be:

cloud
host
serv

We can also use our rDNS data to find the hostname of these IP addresses and also use our ASN database and IP to Company database. If we really want, we can also tap into our host.io data to do domain-level HTML to link extraction.


But usually, the simple public-facing data approach works for us. Our internal data is massive, and data-centric queries can lead to missing important hints. We have discussed our approach at length and the reality is that you really need to have an intuition.

ProbeNet PoP discovery is not a massive issue. Because we are working with national and local ISPs these days, procurement is usually our main bottleneck. But still, that is just normal and perfectly fine.