Partially-Random-Phase Attack
The Partially-Random-Phase Attack targets two-way (plug-and-play) QKD systems, in which the receiver (Bob) sends bright pulses to the transmitter (Alice), who encodes on them and returns them attenuated. Because Alice's phase randomiser takes a finite time to settle, a threat actor (Eve) who substitutes Bob's pulses with her own and controls when they arrive can force the randomiser to act on the rising or falling edge of its response, squeezing the supposedly uniform random phase into a narrow window she chooses [Sun2012]. With the phase only partly random the outgoing pulses stop being indistinguishable, so Eve can read the encoded bit with a homodyne measurement and resend a matching state to Bob. The encoded states themselves are never altered, and by choosing a narrow enough window and a high enough decision threshold Eve can hold the error she introduces below the level at which Alice and Bob would abort, so they compute a positive key rate and share a key that is not secret [Sun2012].
| [BSI2023] | BSI. "Implementation Attacks against QKD Systems". (2023) https://www.bsi.bund.de/SharedDocs/Downloads/EN/BSI/Publications/Studies/QKD-Systems/QKD-Systems.pdf. |
| [Lo2005] | H.-K. Lo, X. Ma, K. Chen. "Decoy state quantum key distribution" In: Phys. Rev. Lett. 94, 230504. (2005) 10.1103/PhysRevLett.94.230504. arXiv:quant-ph/0411004. |
| [Lo2007] | H.-K. Lo and J. Preskill. "Security of quantum key distribution using weak coherent states with nonrandom phases" In: QIC, vol. 7, no. 5 & 6, pp. 431–458. (2007) 10.26421/QIC7.5-6-2. |
| [Sun2012] | S.-H. Sun, M. Gao, M.-S. Jiang, C.-Y. Li, and L.-M. Liang. "Partially random phase attack to the practical two-way quantum-key-distribution system" In: Phys. Rev. A, vol. 85, no. 3, p. 032304. (2012) 10.1103/PhysRevA.85.032304. |
| [Sun2018] | S.-H. Sun, M.-S. Jiang, X.-C. Ma, C.-Y. Li, L.-M. Liang. "Retraction: Hacking on decoy-state quantum key distribution system with partial phase randomization" In: Scientific Reports 8:46943 (2018). (2018) 10.1038/srep46943. |
48e3e90b-bb65-495f-9e0d-836f56919437
- Quantum Communication
Technique properties
Operational properties
| Description | Countermeasure |
|---|---|
|
Detects, by reading an interval the transmitter already has. She triggers her setups from the reference pulse arrival, so the signal-to-reference delay is measurable at her end, and it is exactly the quantity the threat actor must shift to reach the modulator transition [Sun2012]. |
CM-0027 Arrival Time Monitoring |
|
The paper calls active randomization of the source phase the legitimate parties' best choice, and this attack shows in analysis that doing it without monitoring the achieved range is not enough, because the threat actor sets that range [Sun2012]. |
CM-0015 Active Phase Randomisation |
|
Use of watchdog monitor. |
CM-0011 Watchdog Detector |
|
State preparation monitoring. |
CM-0016 State Preparation Monitoring |
|
Include this attack in the full security proof. |
CM-0014 Full Device Imperfections Incorporation in Scientific Security Proof |
|
Two-decoy state, where one of which is vacuum [Sun2012]. |
CM-0003 Decoy States Employment |
|
Triggering Alice's phase modulator with the bright light pulses she receives [BSI2023]. |
CM-0038 Phase Modulator Triggering |
| Description | Category | Impact Type |
|---|---|---|
|
Physical Layer (Quantum & Optical) is the primary target: the threat actor (Eve) manipulates the optical global phase of the transmitter's (Alice's) weak coherent pulses and reads the encoded bit with an optical homodyne detector in a two-way plug-and-play setup [Sun2012]. |
L1 — Physical Layer (Quantum & Optical) | Primary Target |
|
Electronics & Analog Components are a secondary effect: the exploit depends on the phase modulator's finite response time, letting the threat actor (Eve) restrict the global random phase to a narrow sub-range of width δ < 2π by timing the pulse to the modulator's rising or falling edge [Sun2012]. |
L2 — Electronics & Analog Components | Secondary Effect |
| Relation type | Description | Technique |
|---|---|---|
| Combines with |
Neighbours, not variants. Phase remapping acts on the encoded relative phase while this attack acts on the global randomization phase and leaves the BB84 states standard, so the two compose and doing both maximizes the threat actor's information [Sun2012]. |
A-0054 Phase-Remapping Attack |
| Combines with |
The second composition the paper names, gated on the transmitter's Faraday mirror being imperfect. Where it is, the threat actor obtains more information by layering the passive Faraday-mirror attack on top of this one, since both target the same two-way return path [Sun2012]. |
A-0084 Passive Faraday Mirror Attack |
| Relation type | Description | Technique |
|---|---|---|
| Alternative to |
A-0056 Partially-Random-Phase completes the three-attack family this technique belongs to. All three are by the same author group, all three require the two-way plug-and-play architecture, and all three enter through the identical opening: the transmitter (Alice) does not monitor when the signal and reference pulses arrive, so the threat actor (Eve) can substitute the pulses and choose the moment they meet one of Alice's modulators. From there they diverge in what they exploit. A-0056 catches the phase randomiser before it has settled and squeezes the supposedly uniform phase into a narrow window; A-0057 catches the phase modulator's drive edge and reads a state-dependent frequency shift; this technique catches the intensity modulator's drive edge and reads a decoy-versus-signal frequency label, then runs photon-number-splitting on the separated populations. They are alternatives rather than a chain, and the shared entry means the arrival-time check carried by CM-0027 closes all three at once. |
A-0067 Wavelength-Selected Photon-Number-Splitting Attack |
| Alternative to |
A-0056 is the partial-phase counterpart and marks this technique's boundary. A-0055 assumes the global phase is simply not randomised and needs no reverse path; A-0056 tampers with an active randomiser in a two-way plug-and-play setup so that the phase becomes only partially random. [Tang2013a] notes the photon-number channel model may be violated even when Eve holds only partial phase information, citing A-0056's primary source, so the two are alternatives over the same weakness class rather than parent and child. |
A-0055 Non-Random Phase Attack |
|