[cryptography] Another data point on SSL "trusted" root CA reliability (S Korea)
holz at net.in.tum.de
Mon Sep 19 09:21:27 EDT 2011
> Oh, now it makes sense, those are mostly router certs (and various other certs
> from vendors who create broken certs like the Plesk ones). You won't just
Hm. I agree that many are router certs, certainly those with brand names
of networking equipment in the CN, but mostly?
For example, are the 550,000+ ones with "CN=localhost.localdomain" also
router certs? I guess the only way would be to rescan them and get the
HTML they deliver.
I did that, BTW, for about 60k certs with "Plesk" as CN. Mostly, the
sites redirected to port 80, but in about a quarter of cases we found
the typical Plesk portal sites. Given that you can google the default
password, this seems a weak configuration. We'll report on that in our
upcoming IMC paper, too .
> find them in Korea, they're everywhere, in vast numbers, but (at least for the
> router certs) they're usually only visible from the LAN interface.
It would certainly explain why they show up so often in the EFF scan,
but not in our scan of the Top 1M (EFF: 13%, ours: 3%). But, even in the
Top 1M, we get about 30k such certs, and they are not router certs.
> So all you need to do is warkit a router via one of a seemingly endless series
> of vulns that SOHO routers have and you've got a trusted root cert that can
> MITM all traffic through it.
That would be very bad, truly. I am wondering if we can't get our hands
on such a router and do a proof-of-concept. Anyone in?
Dipl.-Inform. Ralph Holz
I8: Network Architectures and Services
Technische Universität München
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 490 bytes
Desc: OpenPGP digital signature
More information about the cryptography