[Cryptography] Characteristics of "robust, good cryptography"

[Natanael]

Den 19 jan 2016 20:49 skrev "Notify" <notify at sandpaddle.com>:
>
> Knowing that “robust” and “good” are somewhat vague, can a cryptographic
algorithm, RNG, implementation, key creation/exchange methodology and
hardware platform be robust and good if unknown persons who the sender did
not intend to have access to the plaintext can gain that access to it
through baked-in weaknesses, key sharing or related cryptographic schemes,
without any interaction with the sender and intended receiver(s)?

Insufficient data.

Are you also excluding shoulder surfing and ciphertext size metadata (and
other traffic analysis) and inductive reasoning based on behavior (mostly
data mining on traffic data, but also based on other external observed user
behavior)? Besides that there's also pure software exploits in rendering
software (like browsers) and similar, and "confused deputy" class software
exploits for exfiltration (authorization logic flaws), which *technically*
falls under implementations, but that isn't necessarily normally considered
security software.

Assuming FDE, end-to-end encryption, traffic obfuscation (Tor, I2P), strong
entropy sources, tempest protection, securely coded software, strong opsec,
etc, then good robust systems should prevent plaintext leakage entirely.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20160119/19d10990/attachment.html>