Hi,

On Fri, 20 Feb 2026 13:43:02 +1100
Viktor Dukhovni <[email protected]> wrote:

> > Those involve fetching DNSKEY and CDNSKEY records
> > from domains in the Tranco Top 1 Million or DomCop Top 10 Million
> > list (I'm alternating between different domain lists) and checking
> > them with badkeys.  
> 
> These lists do not provide good coverage of DNSSEC-signed domains.  Of
> the ~25 million DNSSEC-signed domains covered by the DANE survey, only
> ~54 thousand are listed among the Top 1 million websites.

Sure, I'm aware that this isn't a scan of "the entire DNSSEC
ecosystem". But I guess it's a large enough sample that I'd see if
compromised keys would be common. (Which, from what I can see right
now, they aren't.)
FWIW, I plan to probably switch to a merged list of multiple top-lists
soon.

> I have a database with every DNSKEY seen by the survey since 2017.
> Currently holding 556,114,847 distict keys (45.6 million currently
> live).  While I'm not at liberty to share the domain names, I don't
> see any barrier to sharing just the keys with the domain names
> elided.  If you're interested, I can make that dataset available.

That certainly would be interesting, and I can easily run a scan on
that dataset.

> If you like, the keys can be augmented with a date range indicating
> their first seen and last seen epoch times, and/or the number of
> domains using that key (sometimes a hosting provider uses the same
> key for multiple customer zones).

Yeah, particularly the date range would be helpful, as it's obviously
quite a different issue if the finding is "xxx hosts used a compromised
key in the past" or "xxx hosts are using a compromised key right now".


-- 
Hanno Böck - Independent security researcher
https://itsec.hboeck.de/
https://badkeys.info/

_______________________________________________
dns-operations mailing list
[email protected]
https://lists.dns-oarc.net/mailman/listinfo/dns-operations

Reply via email to