Timing Attack Lab (2D): Cracking a Secret Through Response-Time Leaks

⏱️ Timing Attack Lab (2D)

Server comparison mode

Vulnerable

Constant-time

Attack speed 1x Network jitter 0.12 ms Secret length 6

StatusIdle

Cracking position0 / 6

Probes sent0

Last measured time— ms

Signal / noise gap—

Result—

The vulnerable comparator checks the guess byte by byte and returns the instant it finds a mismatch — real code like for(i) if(a[i]!==b[i]) return false. A guess sharing one more correct leading byte takes measurably longer to reject, because the loop runs one extra round before bailing out. The attacker probes every candidate byte, times the response, and the slowest one is (almost always) correct — repeat position by position until the whole secret falls. A constant-time comparator always walks the full length and XORs every byte, so its response time never depends on how much of the guess matches. Raise network jitter to see how much noise the timing signal can survive before the attack starts guessing wrong.

Each bar = one candidate byte's measured response time this scan. Gold = current best guess, cyan = tested, grey = untested.