I built a Hilbert curve showing which IP addresses have rDNS assigned, based on one of the rDNS scan snapshots available on scans.io. I ended up getting the following (white means "assigned"):


Is there something wrong with my methodology or does most of the Internet routable IPs have no rDNS assigned?

  • 1
    Alternatively, another question might be why do you expect all public IPs wish to expose their host name, if the host name is often changing and not responsible for sending email for example. – Ed Daniel Sep 21 '18 at 17:24
  • @EdDaniel: what surprises me is the regularity. Some of the IPs on that chart are not routable, but there should still be twice as much of those taken. – d33tah Sep 21 '18 at 19:04

Actually, there indeed was a mistake in my methodology. I copied Hilbert code from Wikipedia, but didn't notice it was using int instead of unsigned int. After fixing that, I'm getting a completely different picture - one that makes more sense to me:

Fixed hilbert curve

If I saw that instead, I don't think I wouldn't have asked the original question.

| improve this answer | |
  • Note though that reverse DNS is barely needed for most services. It can be checked by an MTA when receiving a connection to send an email but all old authentication methods based on mappings "IP -> name -> IP" as some protocols were using are more or less deprecated as there are far better (stronger) authentications. – Patrick Mevzek Jan 4 '19 at 20:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.