How does the e-th root mod N demonstrate one-way function behavior, and what role does the "trapdoor" play?

Computing $x^e \mod N$ is easy, but reversing it (finding $x$ from $y = x^e \mod N$) produces a chaotic, seemingly random mapping that's infeasible to invert without knowing the factorization of N.

* Cubing mod 33 is one-to-one yet scrambled: going forward is trivial, but recovering the cube root (e.g. that 29³ ≡ 2) means searching — unless you hold the trapdoor. *

Concrete example: $x^3 \mod 33$

Key observations:

• The mapping from $x$ to $x^3 \mod 33$ is bijective (one-to-one) but looks completely chaotic
• Given $y$, finding $x$ by inspection is essentially impossible for large N
• For small N, you can build a complete table — but for RSA-sized $N > 10^{600}$, this is unthinkable

The trapdoor: If you know that $N = p \cdot q$ (the prime factorization), you can compute $\varphi(N) = (p-1)(q-1)$ and then find the decryption exponent $d = e^{-1} \mod \varphi(N)$. With $d$, computing $y^d \mod N = x$ is easy via SAM. Without the factorization, you're stuck.

This is RSA in a nutshell: The public key makes $x^e \mod N$ easy for everyone; only the private key holder knows the factorization (trapdoor) to reverse it.

Go deeper:

• RSA (cryptosystem) (Wikipedia) — the e-th root as RSA's trapdoor one-way function.
• RSA Encryption (Art of the Problem) — a beautifully animated build-up of the trapdoor.