Quantum Hacking Detector Blinding Qkd

Short answer: Detector blinding is a "quantum hacking" attack on quantum key distribution (QKD) hardware. The eavesdropper shines bright light on the receiver's single-photon detectors so she can control them, and in the published demonstrations she could learn the key without being detected. The attack targets the implementation, not the physics of QKD. Researchers have since published countermeasures and automated tests, and a 2023 testbench study found a simple countermeasure caught continuous blinding but missed pulsed blinding.

What the attack does

A paper in Laser Physics Letters says an attack exploiting single-photon avalanche diode (SPAD) blinding is one of the effective methods of quantum hacking, citing Lydersen and colleagues in 2010. It says the attack was experimentally demonstrated on various QKD systems using both phase and polarization coding. The authors describe the outcome of a successful attack as the eavesdropper knowing the whole key, producing no errors and going undetected. They add that it is the only attack demonstrated in explicit form on many real QKD systems.

A 2023 paper in EPJ Quantum Technology describes the single-photon detector as currently the most unsafe element of a QKD system. It lists several attacks on detectors. Some exploit a mismatch in detection efficiency between detectors in the receiver, in time, wavelength, spatial mode or during dead time. Others control the detector deterministically by blinding it with bright light or by injecting bright pulses at the closing edge of a gate or between gates. The authors say detector control was first proposed in 2009 and found to be applicable to commercial QKD systems the following year. They note that protection is difficult because the receiver has to accept all the light from the line.

A countermeasure through statistics

The Laser Physics Letters authors prove that when blinding acts on a phase-coding system, it distorts the photocount statistics so that the eavesdropper may always be detected. They say it is enough to change how the registration results are processed, without redesigning the system or its control electronics. They also state that polarization-coding systems remain vulnerable to this attack. This is the authors' own theoretical claim for the phase-coding case.

Testing countermeasures

The 2023 paper reports an automated testbench that checks detector vulnerability to attacks using tailored bright illumination. They tested a free-running detector with a rudimentary countermeasure that measures average photocurrent. The testbench found the detector controllable in a continuous-blinding regime, but the countermeasure registered photocurrent well above the quantum regime and so revealed the attack. A manually performed pulsed blinding attack controlled the detector intermittently and was missed by the countermeasure across a wide range of pulse durations and powers, still allowing the key to be learned. The authors make recommendations to improve both the testbench and the countermeasure.

What this does not tell you

These papers concern specific detectors and lab demonstrations. They do not assess a given commercial product or say how common any attack is outside the lab. A countermeasure that works against one attack can miss another, which is the point of the 2023 result.

FAQ

What is detector blinding?

Using bright light to put single-photon detectors under the eavesdropper's control.

Does it break the theory of QKD?

It exploits implementation flaws in the hardware, not the underlying principle.

Are all QKD systems equally exposed?

One paper argues phase-coding systems can detect it through photocount statistics, while polarization-coding systems remain vulnerable.

Can a simple photocurrent check stop it?

In the 2023 test it caught continuous blinding but missed pulsed blinding.

Why test automatically?

An automated testbench can check a detector against these attacks systematically, as the 2023 paper demonstrates.