3 ms·
>There is no evidence that US push flawed curves. "Reuters reported in December that the NSA had paid RSA $10 million to make a now-discredited cryptography sy
by alasdair_ 7y ago
>There is no evidence that US push flawed curves.
"Reuters reported in December that the NSA had paid RSA $10 million to make a now-discredited cryptography system the default in software used by a wide range of Internet and computer security programs. The system, called Dual Elliptic Curve, was a random number generator, but it had a deliberate flaw - or “back door” - that allowed the NSA to crack the encryption."
https://www.reuters.com/article/us-usa-security-nsa-rsa/exclusive-nsa-infiltrated-rsa-security-more-deeply-than-thought-study-idUSBREA2U0TY20140331 https://www.reuters.com/article/us-usa-security-nsa-rsa/excl...
- mzs 7y ago"random number generator"
- tptacek 7y ago"Dual Elliptic Curve" is an RNG, a PKRNG, that works by using a public key to encrypt its state, which is then directly revealed (as public key ciphertext) to callers (for instance: in the TLS random blob). The problem with PKRNGs has nothing to do with elliptic curves; you could design one with RSA as well. The problem is that for a given public key, there's also a private key, and if you have that private key you can "decrypt" the random value to reveal the RNG's state. That's not a flawed curve that NSA pushed; it's a much more straightforward cryptographic backdoor.