SSL CRL activity
isc.sans.edu
isc.sans.edu
Be reasonable. It's always fun to shout out that $popular_technology is now entirely broken, though usually the shouters have a working alternative, but that's not the case.
- Around 17% of the web servers seem to have been vulnerable.
- Key extraction on cloudflarechallenge.com took days to develop and execute.
- The number of requests needed for this is noticeable, too much so for intelligence agencies. They may take a week or two and run at 2 requests per second, assuming the attack still works at that speed, but it's a risky business. A vulnerability of this magnitude makes a great weapon in digital warfare and should not leak.
- You still need to MITM your targets if you obtained private keys.
In the end I think the number of sites where all of this happened is very, very low. TLS is still perfectly fine.
The idea that MitM attacks are "rare" is such b.s. thinking ("Before Snowden"). We known about widespread MitM attacks against big targets (Google, Facebook, and the TOR network at a minimum), and those same techniques can be reused elsewhere.
Given how often it has shown up in the revelations of the last year, I would suggest this may be a fairly common method of attack.
http://articles.latimes.com/2014/mar/12/business/la-fi-tn-ns...
https://www.schneier.com/blog/archives/2013/09/new_nsa_leak_...
https://www.schneier.com/blog/archives/2013/10/how_the_nsa_a...
http://en.wikipedia.org/wiki/Tailored_Access_Operations#QUAN...
/b.s. == before snowden //kind of a hit-or-miss joke, unfortunately