Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon. Entire thread

challenging challenge

Name: Anonymous 2017-03-19 13:02

Given two random primes, p & q, and the results pz = Z % p, qz = Z % q, find Z

Name: Anonymous 2017-03-19 14:03

Do your own homework.

Name: Anonymous 2017-03-20 2:33

>>2
I will, i just thought you guys might like to participate
It's self-assigned so there's not really a due date or anything

Finding Z is equivalent to finding M, N, where Mp + pz = Nq + qz = Z, which may be non-trivial for q <= p < Z < pq

It's a simple trapdoor function based on knowledge of Z

Name: Anonymous 2017-03-20 7:10

Not sure if it quite falls in the discrete log class

Simplest method would be to run N = 1..q, Np+pz % q == qz, which reduces to reversing N*x % q = q + (qz - pz) % q, with x = p%q

Name: Anonymous 2017-03-20 7:34

>>2
C+

Name: Anonymous 2017-03-21 0:05

I'm calling it the discrete vector intersection problem

Lets try an order-2 puzzle

Given two random primes, p & q, and from the result pz = Z % p, qz = Z % q, we get pz2 = pz*pq % p, pq2 = pz*pq % q, find pz, pq, Z

Name: Anonymous 2017-03-21 1:08

qz not pq
pz2 == Z * (Z % q) % p == pz * qz % p
qz2 == Z * (Z % p) % q == pz * qz % q

Difficulty in reversing pz2, qz2 <- pz*qz (naive dlog puzzle?)
Difficulty in reversing pz, qz -> pz*qz (weak factoring puzzle? w/ multi solution)
Difficulty in reversing pz, qz <- Z (naive dlog puzzle?)

Name: Anonymous 2017-03-21 1:34

Z, p,q,r

pz, qz, rz

pqzr, qrzp, rpzq = rz*pz % q = (Z%p) * (Z%r) % q

Name: Anonymous 2017-03-27 8:53

For fuck's sake, the goatfinger admin copied posts from here, now I know why you never replied to me. I fucking knew the guy was some imageboard autist who desperately wanted to run a textboard the moment he volunteered.

Anyway, what you are writing about is solvable through the Chinese remainder theorem if p ≠ q, and trivial otherwise.
According to the CRT,
φ : /pqℤ -> /pℤ × /qℤ
φ(x + pqℤ) = (x + pℤ, x + qℤ))
is a ring isomorphism and thus bijective. This means that every one-sided inverse of φ is the two-sided inverse of φ. Consider then
ψ(a + pℤ, b + qℤ) = aqM + bpN + pqℤ,
where M is the inverse of q modulo p and N is the inverse of p modulo q.
ψ satisfies φψ = id/pℤ × /qℤ by construction and is therefore the inverse of φ.
If p = q, the challenge boils down to calculating a square root modulo pq, which is equivalent to factorization. But you can easily check whether pq is a square, so this is simple. For >>7, you have to apply the CRT twice.

Name: Anonymous 2017-03-28 1:53

lol, nicely done though anon

I don't quite understand all of the notation, but i think i get the general idea

you have Mq % p = 1 and Np % q = 1
p & q aren't protected by factorization like with RSA so multiplicative inverses are known, ie M = q^(p-2) % p

reversing N*x % q
x = p%q
invx = x ^ (q-2) % q
N = (N*x % q) * invx % q
Z = N*q + qz

Name: Anonymous 2017-03-28 23:20

>>9
the goatfinger admin copied posts from here
Prove it :^)

Name: Anonymous 2017-03-29 0:45

I can confirm that I haven't reposted elsewhere

Prove it
When we all have NSA powers in the not too distant future, consider it done

Name: Anonymous 2017-03-29 0:56

Alias check on reposts from this thread plz, NSABot
Report

:^D

Name: Anonymous 2017-03-30 17:20

>>11
:^)
Goatfinger imageboarder admin confirmed. Make sure you get those pleb cuck baits to tdavis.

Newer Posts
Don't change these.
Name: Email:
Entire Thread Thread List