The US has backdoored RSA's RNG and thus endangered the security of American companies. It is naive to think that US intelligence agencies will act in the best interest of US citizens or companies.
Notably that was a "no-one-but-us" backdoor, that requires a specific secret key to exploit. We'll see when someone analyzes the payload further, but presumably this backdoor also triggers on a specific private key. If not there are ways to do it that would look far more like an innocent mistake, like a logic bug or failed bounds check.
I can see some arguments that might persuade the NSA to run an attack like this
- gathers real world data on detection of supply attacks
- serves as a wake-up call for a software community that has grown complacent on the security impact of dependencies
- in the worst case, if no one finds it then hey, free backdoor
There's an implicit "always" in their second sentence, if you're confused by the wording. They aren't positing the equivalent of the guard that only lies.
It's an interesting story for those who haven't heard about that an think the NSA could only be up to evil. You may not have read it as the guard only ever lies, but that doesn't stop people from thinking that anyway.
No I think that's it. "What about it?" kinda set me off, and then "if you're confused by the wording" was unnecessarily condescending.
You coulda just pointed out that just because they did right in the case of DSA, doesn't mean we should actually ever trust them, which I would agree is the correct stance.
Mostly I think that story is neat and wanted people to know about it, so I asked a question as a performative writing technique.
"What about it?" is a very real question that I still want to know the answer to. What did you want as a response when you asked that?
"If you're confused by the wording" was definitely condescending, but I think interpreting guinea-unicorn's post that way doesn't make sense. Even in your reply you didn't say you think it's the right interpretation, just that someone might believe the NSA could "only be up to evil". That followup gives the impression you were giving an FYI for readers. Which is nice to do, but then the "what about" doesn't fit.
So all of that is to say the words "what about" felt like you were deciding to read their post in an unfair way.
I'm happy to listen to an alternate explanation! But you ignored my request for why you said that, and I'm honestly kind of confused as to why that's what set you off.
So overall I think I think my first post can come across as fighty but I don't think the followups should suggest I'm making things fighty. I think my response to 2OEH8eoCRo0 was fine given the way they were ignoring half of the four sentences I had typed.
You are understating the level of evidence that points to the NSA being fully aware of what it was doing.
To be clear, the method of attack was something that had been described in a paper years earlier, the NSA literally had a program (BULLRUN) around compromising and attacking encryption, and there were security researchers at NIST and other places that raised concerns even before it was implemented as a standard. Oh, and the NSA paid the RSA $10 million to implement it.
Heck, even the chairman of the RSA implies they got used by the NSA:
In an impassioned speech, Coveillo said RSA, like many in industry, has worked with the NSA on projects. But in the case of the NSA-developed algorithm which he didn’t directly name, Coviello told conference attendees that RSA feels NSA exploited its position of trust. In its job, NSA plays two roles, he pointed out. In the information assurance directorate (IAD) arm of NSA, it decides on security technologies that might find use in the government, especially the military. The other side of the NSA is tasked with vacuuming up data for cyber-espionage purposes and now is prepared to take an offensive role in cyber-attacks and cyberwar.
“We can’t be sure which part of the NSA we’re working with,” said Coviello with a tone of anguish. He implied that if the NSA induced RSA to include a secret backdoor in any RSA product, it happened without RSA’s consent or awareness.
What type of confirmation do you want? The documents aren't going to be declassified in the next couple of decades, if ever.
I've never heard anyone claim that Dual_EC_DRBG is most likely not intentionally backdoored, but there's literally no way to confirm because of how its written. If we can't analyze intention from the code, we can look at the broader context for clues. The NSA spent an unusual amount of effort trying to push forward an algorithm that kept getting shot down because it was slower than similar algorithms with no additional benefits (the $10 million deal specified it as a requirement [1]). If you give the NSA the benefit of the doubt, they spent a lot of time and money to... intentionally slow down random number generation?!
As an American, I'd prefer a competent NSA than an incompetent NSA that spends my tax dollars to make technology worse for literally no benefit...
