Notes
Explanations of the mathematics our measurements rest on, a reference to the attacks, and a catalogue of the schemes they apply to. The worked examples run on groups small enough to follow by hand. They are computed by the code that builds this site, and every number on them is checked again each time it is built.
Reference
- Cryptanalysis reference
- How public-key hardness assumptions are attacked: what each attack needs, what it costs, the largest computation on record, and where it stops applying.
- Schemes
- The schemes in use and the ones replacing them, from RSA and elliptic curves to ML-KEM, ML-DSA, SLH-DSA and SQIsign: what each rests on, its sizes, and its status against quantum computers.
Elliptic curves
What an elliptic curve is, why its points form a group, and why recovering a secret multiple of a point is hard enough to build cryptography on.
-
Elliptic curves
What an elliptic curve is, how two points are added, what changes over a finite field, and how a point is multiplied by a large integer.
-
The ECDLP
The problem every elliptic-curve scheme rests on, the square-root bound that generic attacks cannot beat, and the checks that rule out the known shortcuts.
-
Curve traits
All 22 DiSSECT traits and eight foundations, with factorization explained and a six-phase study plan.
-
Isogenies
Maps between elliptic curves, the graph they form on supersingular curves, and why a walk in that graph can be a secret key.
Lattices
The mathematics under ML-KEM and ML-DSA: lattices, short vectors, learning with errors, and the reduction algorithms used to attack them.
-
Lattices and LWE
What a lattice is, why adding small noise to linear equations makes them hard, and the lattice attack that recovers the secret when the dimension is small.
Worked examples
Attacks run end to end on small instances, with every intermediate value shown. At these sizes the point is the mechanism, not the speed; each page says what changes at cryptographic sizes.
-
Pollard's rho, worked
One walk followed step by step until it repeats and gives up the key, then two thousand more measured against the prediction for a random function.
-
Index calculus, end to end
The whole algorithm run twice with every number shown: first in the multiplicative group of a prime field, then on an elliptic curve over Fp2 using Semaev's summation polynomial.
Explainers
Longer pieces that compare approaches rather than walking through one.
-
Index calculus in three settings
The same pipeline over three kinds of curve, and why the structure of the field decides how far it gets.