Posts
Announcing The First Recipients of The Zama Cryptanalysis Grants
Go Cryptography Security Audit
Ring Oscillators: How Do Xor Gates Help with Random Bit Generation?
I was reading Request for Comments 4086 (Randomness Requirements for Security) on using ring oscillators for true random generation. The document says one can increase the rate of random bit generation by applying the sampled bits from ring oscillators to a XOR gate. How does applying the sampled bits to …
The cryptography behind passkeys
Random Oracles: How Do They Ensure Robustness in Random Generation?
I am trying to understand how the Linux CSPRNG works. In a [git commit](https://git.kernel.org/pub/scm/linux/kernel/git/crng/random.git/commit/?id=c57044909484) Jason A Dononfeld explains one of the reasons BLAKE2s was chosen as a cryptographic hash function to serve as a PRNG was that it is a random oracle. The paper Dononfeld [cites](https://eprint.iacr.org/2019/198) explains random oracles offer …
Send files privately. No cloud. No trace.
[glitr.io](http://glitr.io/) I’m working towards something for secure/private/simple P2P file transfer. It isnt as “simple” as it could be, im still working on it, but ive got it down to: * Zero-installation as a PWA * Zero-registration by using local-only storage * P2P-authentication using WebCrypto API * Fast data-transfer using WebRTC …
Using encryption instead of hashing for checking passwords - why?
I stumbled upon an interesting piece of source code at work yesterday. The purpose of the code is to check if the user has provided the correct password compared to the one stored in the database. Pretty standard so far. But... Instead of hashing the user-provided cleartext password and compare …
Invariant-Based Cryptography (Part II): New Schemes, New Invariants, Generalized Framework
I’ve just published a follow-up to my earlier work on invariant-based symmetric cryptography — this time shifting from proofs to principles, from a single construction to a flexible paradigm. What’s new? • Two fresh symmetric schemes built around algebraic invariants: → One uses polynomial discriminants, → The other exploits the …
Found a near-collision modulus for a random 1024-bit modulus
So, as the title says, I found a way to generate near-collision moduli for arbitrary length moduli (1024, 2048, and even more), such that: * N' / N are extremely close, close enough that the 50% most significant digits (yes) match * The error N' - N is merely d …