For all you PGP fanatics, consider (from http://33bits.org/2011/03/09/link-prediction-by-de-anonymiza..., a blog which I highly recommend) the sentence "During (...) 2007-2009, Shmatikov and I (...) showed how to take two graphs representing social networks and map the nodes to each other based on the graph structure alone—no usernames, no nothing."
Re:FOIA the PRISM tech stack
by timClicks on Friday June 07, 2013 @01:24AM
(#5837397)
I find this especially interesting coupled with Cray's YarcData graph analytics hardware called Urika, which has 512 terabytes in memory. http://www.yarcdata.com/Products/
Re:FOIA the PRISM tech stack
by angersock on Friday June 07, 2013 @12:35AM
(#5837244)
Nice find.
This paper seems to suggest that a possible counter attack is to create bogus edges and adjacencies--thus fucking with the utility of a (sparse) adjacency matrix or attempts at having useful window sizes for BFS or the like.
I suspect, though, that any sort of attempt at that would somehow be thwarted and just cause more pain on the attempting party.
Re:FOIA the PRISM tech stack
by ianstormtaylor on Thursday June 06, 2013 @10:51PM
(#5836891)
So, sadly true. I was sending a friend an email today about this entire thing and for the first time ever I questioned whether my email would be eavesdropped on by my own government. It was a horrible feeling.
Re:FOIA the PRISM tech stack
by sitkack on Friday June 07, 2013 @07:12AM
(#5838239)
They don't need credibility, the chilling effects of knowing you are being watched are enough. You prevented yourself from actualizing your thought crime.
Re:FOIA the PRISM tech stack
by sitkack on Friday June 07, 2013 @07:14AM
(#5838242)
The nice thing about exposing what was already "known" is now so many people are talking about it openly. It is hard for the wolves to attack so many moving targets. What we need is for hundreds to file very narrowly focused FOIA requests that do not probe too deeply. They need to be fulfillable.
Re:FOIA the PRISM tech stack
by Sami_Lehtinen on Friday June 07, 2013 @04:21AM
(#5837850)
No SSL/TLS, great job. Well, I assume AES crypto wouldn't bother them anyway. ;) Because it wouldn't be smart to make standard crypto which would blind them.
Re:FOIA the PRISM tech stack
by damianball on Friday June 07, 2013 @05:00AM
(#5837950)
If I remember correctly, Verisign has indicated they would be willing to create a trusted root level certificate for the government to spoof the ssl of any domain... so I don't know if SSL is enough, unless you only allow CAs you've created or you trust.
Re:FOIA the PRISM tech stack
by adkatrit on Thursday June 06, 2013 @10:52PM
(#5836895)
Why would they have a form to ask for records on yourself when they don't publish which records they have. This is like a game of Go Fish but with personal data.
Re:FOIA the PRISM tech stack
by downandout on Friday June 07, 2013 @01:00AM
(#5837323)
Unfortunately, despite the disturbing fact that we are reading about PRISM (disturbing both that it exists and that our government can't keep highly sensitive secrets), it is classified Top Secret until September 1, 2036. I could be wrong but I believe that FOIA requests regarding it will fail until then.
Re:FOIA the PRISM tech stack
by dlitz on Friday June 07, 2013 @02:20AM
(#5837557)
> disturbing both that it exists and that our government can't keep highly sensitive secrets
Why could it be disturbing that a government's officials can't keep secrets about cases where they abuse their power or act against the public interest? Isn't that how democratic institutions are supposed to be designed?
Re:FOIA the PRISM tech stack
by dlitz on Friday June 07, 2013 @03:37PM
(#5841139)
You're assuming that all sensitive information has the same probability of leaking. Anecdotally, that's not really true. Most of these systems depend on honest citizens to keep the information secret, and honest citizens are more likely to leak "secret" evidence of abuse than information that will cost innocent people their lives.
Tinfoil won't save you this time.
This paper seems to suggest that a possible counter attack is to create bogus edges and adjacencies--thus fucking with the utility of a (sparse) adjacency matrix or attempts at having useful window sizes for BFS or the like.
I suspect, though, that any sort of attempt at that would somehow be thwarted and just cause more pain on the attempting party.
http://www.nsa.gov/applications/forms/foiaemail.cfm
Internal Server Error - Read
The server encountered an internal error or misconfiguration and was unable to complete your request. Reference #3.a447da3f.1370606557.3a7abfb
Configuration mistake or force of habit?
Why could it be disturbing that a government's officials can't keep secrets about cases where they abuse their power or act against the public interest? Isn't that how democratic institutions are supposed to be designed?