How did the first known smartcard implementation attack (1996) extract a PIN without breaking any cryptographic algorithm?

Attackers monitored the smartcard's power consumption to detect when the retry counter was being written after a wrong PIN, then cut power before the write completed — keeping the counter at its original value for unlimited guesses.

* The 1996 power-off attack: spot the counter-decrement write by its current draw, then cut power before it finishes. *

The attack step by step:

1. Smartcards store a 4-digit PIN and a retry counter (typically starts at 3)
2. When a wrong PIN is entered, the card decrements the retry counter by writing to memory
3. Writing to memory requires more electrical current than reading
4. An external circuit monitors the card's power consumption
5. When the power spike from the write operation is detected, the attacker immediately cuts power to the card
6. The write never completes → the retry counter stays unchanged
7. Repeat with the next PIN guess — unlimited attempts

Statistics:

• 4-digit PIN = 10,000 possibilities
• On average, found after n/2 = 5,000 attempts
• With the right equipment, the entire attack took only a few hours

The countermeasure: Decrement the retry counter before checking the PIN. If the PIN is correct, increment it back. This way, a power cut during the process always leaves the counter decremented.

Key lesson: This attack broke zero cryptography — it exploited the order of operations in the implementation. A perfect algorithm means nothing if the implementation leaks information through physical side channels.

Go deeper:

• Side-channel attack (Wikipedia) — extracting secrets from physical behaviour, not the maths.