Replace AES-128/RSA/DSA with AES-256/Salsa20/Chacha20 so Microsoft Azure will be able to safely allow large-scale Q compute.
Categories
(Core :: Security: PSM, defect)
Tracking
()
People
(Reporter: eternalbluebullrun, Unassigned)
Details
Steps to reproduce:
Access old websites that have not yet updated to AES-256, Salsa20 or Chacha20.
https://github.com/git-for-windows/git/issues/4365
Actual results:
Firefox (all OSs, all platforms) is preventing large-scale Q-compute because people are scared that Firefox will fall back to AES-128, RSA or DSA (vulnerable to large-scale Q-compute.)
Projects blocked by not having large-scale Q-compute:
Q-accelerated graph search algorithms for compilers
Q-accelerated neural update algorithms (for all neural workloads) -- https://mathforums.com/t/our-neural-paths-take-pixels-to-make-shapes-shapes-to-make-objects-detect-moves-object-data-to-make-a-world.368750/#post-694087
Expected results:
Firefox should require safe algorithms such as AES-256 or Salsa20.
Because large-scale Q-compute is not yet commercially feasible, this is not yet a security problem,
but until this is fixed people will not support commercial-scale Q-compute,
so Firefox is hurting the rest of the world by allowing insecure algorithms.
Comment 1•2 years ago
|
||
Moving this to Core > Security: PSM so that our engineers could take a look over this issue, if this is not the right component, please assign a more suitable one. Thanks!
Comment 2•2 years ago
|
||
Moving to larger keys for symmetric ciphers is neither necessary nor sufficient for post quantum security.
Reporter | ||
Comment 3•2 years ago
|
||
https://www.qusecure.com/aes-256-is-quantum-resistant-rsa-is-not/
If you up AES to 256 it is quantum safe.
AES 128 (allowed by Firefox) is not quantum safe.
RSA (allowed by Firefox) is not quantum safe.
Salsa20 is quantum safe.
So, yes, what I proposed is necessary, as well as sufficient.
Reporter | ||
Comment 4•2 years ago
|
||
https://eprint.iacr.org/2022/184
https://eprint.iacr.org/2022/1630
[MEDIA=youtube]7r4cNom5y7o[/MEDIA] How easy is it in 2022 to find a SHA1 collision?
https://zhengdw.github.io/2022/10/03/sekaictf-diffecient.html
https://csrc.nist.gov/csrc/media/Events/2022/lightweight-cryptography-workshop-2022/documents/papers/differential-linear-cryptanalysis-on-xoodyak.pdf
Description
•