Manipulation of Vacuum-Noise Estimation

System

The threat actor (Eve) tampers with the light travelling between the transmitter (Alice) and the receiver (Bob) so that the receiver's own measurement of its noise floor comes out wrong. That floor, the shot-noise unit, is the yardstick every other measurement is compared against, so once it is wrong the receiver reports far less excess noise than is really present and the two parties conclude the link is clean when it is not [Ferenczi2007,Jouguet2013].

The attack takes no key material by itself. What it produces is a receiver whose only check on eavesdropping has been quietly disabled, so that a second attack running at the same time can read the key without raising an alarm. Alice and Bob keep generating key at a rate they believe is secure [Huang2014,Ma2013a].

There are five published ways to reach that state, and they differ in what part of the light is changed: its arrival timing, its overall brightness, the pattern of its brightness, its polarisation, or the addition of faint light at other wavelengths. Two of the five have been measured on real receiver hardware [Jouguet2013,Zhao2018]. The last is also the homodyne route of A-0097, which carries the same construction for the key it yields; this row carries it for the shot-noise estimate it corrupts [Huang2014].

Literature 14
[ETSI016PP] ETSI. "Quantum Key Distribution (QKD); Common Criteria Protection Profile - Pair of Prepare and Measure Quantum Key Distribution Modules". In: ETSI GS QKD 016. (2024) https://www.etsi.org/deliver/etsi_gs/QKD/001_099/016/02.01.01_60/gs_QKD016v020101p.pdf.
[Fan2023] Lu Fan, Yiming Bian, Mingze Wu, Yichen Zhang, Song Yu. "Quantum hacking against discrete-modulated continuous-variable quantum key distribution using modified local oscillator intensity attack with random fluctuations". (2023) arXiv:2308.00557. https://arxiv.org/abs/2308.00557.
[Ferenczi2007] A. Ferenczi, P. Grangier, and F. Grosshans. "Calibration Attack and Defense in Continuous Variable Quantum Key Distribution". In: CLEO/Europe and IQEC 2007 Conference Digest, (Optica Publishing Group, 2007), paper IC_13. (2007) https://opg.optica.org/viewmedia.cfm?uri=IQEC-2007-IC_13&seq=0.
[Haeseler2008] H. Häseler, T. Moroder, and N. L ütkenhaus.. "Testing quantum devices: Practical entanglement verifcation in bipartite optical systems" In: Phys. Rev. A 77, 032303. (2008) 10.1103/PhysRevA.77.032303.
[Huang2014] J. Huang, S. Kunz-Jacques, P. Jouguet, C. Weedbrook, Z. Yin, S. Wang, W. Chen, G. Guo, and Z. Han. "Quantum hacking on quantum key distribution using homodyne detection" In: Phys. Rev. A 89, 032304. (2014) 10.1103/PhysRevA.89.032304.
[ISOIEC23837P2] ISO/IEC JTC 1/SC 27. "ISO/IEC DIS 23837-2:2022 Information security - Security requirements, test and evaluation methods for quantum key distribution - Part 2: Evaluation and testing methods". (2022)
[Jouguet2013] P. Jouguet, S. Kunz-Jacques, and E. Diamanti. "Preventing calibration attacks on the local oscillator in continuous-variable quantum key distribution" In: Phys. Rev. A 87, 062313. (2013) 10.1103/PhysRevA.87.062313.
[KunzJacques2015] S. Kunz-Jacques and P. Jouguet. "Robust shot-noise measurement for continuous-variable quantum key distribution" In: Phys. Rev. A 91, 022307. (2015) 10.1103/PhysRevA.91.022307.
[Liu2017] Weiqi Liu, Jinye Peng, Peng Huang, Duan Huang, and Guihua Zeng. "Monitoring of continuous-variable quantum key distribution system in real environment" In: Opt. Express 25, 19429-19443. (2017) 10.1364/OE.25.019429.
[Ma2013a] X. Ma, S. Sun, M. Jiang, and L. Liang. "Local oscillator fuctuation opens a loophole for Eve in practical continuous-variable quantum-key-distribution systems" In: Phys. Rev. A 88, 022339. (2013) 10.1103/PhysRevA.88.022339.
[Ma2014] X. Ma, S. Sun, M. Jiang, M. Gui, Y. Zhou, and L. Liang. "Enhancement of the security of a practical continuous-variable quantum-key-distribution system by manipulating the intensity of the local oscillator" In: Phys. Rev. A 89, 032310. (2014) 10.1103/PhysRevA.89.032310.
[Wu2026] Mingze Wu, Junhui Li, Song Yu, Yichen Zhang. "Practical discrete-modulated continuous-variable quantum key distribution with one-time shot-noise unit calibration" (2026) 10.1103/mcdz-yj4f.
[Zhang2020] Yichen Zhang, Yundi Huang, Ziyang Chen, Zhengyu Li, Song Yu, Hong Guo. "One-Time Shot-Noise Unit Calibration Method for Continuous-Variable Quantum Key Distribution" (2020) 10.1103/PhysRevApplied.13.024058. arXiv:1908.06230.
[Zhao2018] Y. Zhao, Y. Zhang, Y. Huang, B. Xu, S. Yu, and H. Guo. "Polarization attack on continuous-variable quantum key distribution" In: J. Phys. B: At. Mol. Opt. Phys. 52, 015501. (2018) 10.1088/1361-6455/aaf0b7.
QID: A-0099
UUID: a37a3535-3833-4d71-a4ac-883956b4f362
Tier: T0
Type: Quantum
Domain(s):
  • Quantum Communication
Tactic: Execution
Created: 2024-06-21
Updated: 2026-11-13
Maturity: Experimental
Technique properties

Attack type: Door opener/Lowering security
Attack category: Calibration attack
Target party: Receiver
Target function: Quantum layer
Quantum channel: Fibre optic
Security layers: L1 — Physical Layer (Quantum & Optical) L2 — Electronics & Analog Components L4 — System Control & Orchestration L5 — Post-Processing & Cryptographic
Maturity: Experimental
Operational properties

Technique → Countermeasures 2
Description Countermeasure

Measuring the shot noise during the run, at attenuation ratios chosen at random per pulse, removes the stored relation this technique depends on and tests the proof hypothesis that relation silently stands in for. KunzJacques2015 tested the method on a working GG02 receiver without an attacker and simulated it against the saturation and wavelength attacks [KunzJacques2015].

The control is testable rather than qualitative, which is what makes it deployable. A data block is used for key production only if the noise-to-signal fit has a squared correlation above 0.99, the largest deviation from the fitted line stays under two parts in ten thousand of a shot-noise unit, and the attenuation-to-signal fit also exceeds 0.99 [KunzJacques2015]. Two tests catch two different attacks: the wavelength route adds a term quadratic in the attenuation ratio, which the first test measures, while the saturation route survives the first test at 0.998 and is caught only by the second [KunzJacques2015].

The ratio choice must be random per pulse, to deny a threat actor adaptive behaviour that depends on which ratio was applied [KunzJacques2015].

CM-0087 Real-time Quadrature Measurements with Random Optical Attenuations

Machine-learning pattern recognition is the one detection control on this row that looks for the attack's statistical signature rather than for a single out-of-range quantity. CM-0088's own body describes the continuous-variable application and the feature set it uses, the mean and variance of the signal, the local-oscillator intensity and the shot-noise variance, and attributes recognition of calibration-vulnerability and local-oscillator-manipulation signatures to Mao2020 and Mao2020a.

Its value against this technique is that it does not depend on any one quantity staying in range, which is what each of the five routes is designed to achieve for a different quantity [Fan2023].

CM-0088 Pattern Recognition Using Machine Learning
Technique → Security Layers 4
Description Category Impact Type

Primary at the physical quantum and optical layer, because that is where the technique acts and where the quantity it corrupts is produced. Every route manipulates light in the channel, and what changes is the optical power, polarisation or wavelength content arriving at the receiver's homodyne detection [Jouguet2013].

L1 — Physical Layer (Quantum & Optical) Primary Target

Secondary at system control and orchestration, which the polarisation route reaches. What that route exploits is a feedback control loop's sampling policy rather than any component being out of specification: the compensation measures a subset of local-oscillator pulses and applies the correction to the whole cycle, and the threat actor rotates the remainder [Zhao2018]. The sampling rate is a control-plane parameter, not an optical one.

L4 — System Control & Orchestration Secondary Effect

Secondary at the post-processing and cryptographic layer, because that is where the damage is realised rather than caused. The falsified shot-noise unit is consumed by parameter estimation and the key-rate calculation, and Fan2023 shows the manipulation disturbing semidefinite-programming parameter estimation so that the parties overestimate both the secure distance and the key rate [Fan2023]. A receiver whose optical layer is intact can still produce an insecure key here.

L5 — Post-Processing & Cryptographic Secondary Effect

Secondary at the electronics and analog layer, which the trigger-delay route reaches specifically. The attack works by moving when the clock circuit crosses its threshold, so the homodyne integration window lands off the peak of the integrator's discharge curve, and the measured consequence is a reduced detection response slope [Jouguet2013]. The coupling that makes it possible is electrical, one input photodiode generating both the clock and the power signal.

L2 — Electronics & Analog Components Secondary Effect
Related Techniques 1
Relation type Description Technique
Precedes

This technique conceals an attack rather than performing one, so it runs before or alongside the step that reads the key. Intercept-and-resend is the common case and the one in Jouguet2013 and in Huang2014 Strategy A [Huang2014]. precedes rather than enabled_by, because the causal direction runs from the masking to the extraction. The relation is deliberately not exclusive: Ma2013a conceals an entangling-cloner Gaussian collective attack and Fan2023 a collective attack against discrete-modulated CV-QKD, neither of which is an intercept-and-resend, and the Notes bullet in advanced_description records that.

A-0045 Intercept-and-Resend Attack
Techniques Relating to This One 1
Relation type Description Technique
Combines with

Scoped to the homodyne Strategy B branch only. When the receiver monitors local-oscillator intensity and also checks its linear relation to the shot noise, the wavelength manipulation is not sufficient on its own and [Huang2014] composes it with the local-oscillator calibration attack of [Jouguet2013], which manipulates the detector clock trigger timing. A-0099 Manipulation of Vacuum-Noise Estimation is that attack in the catalog: its own literature holds both [Jouguet2013] and [Huang2014], and its description is local-oscillator parameter manipulation to mask excess noise. Strategy A, which applies when the receiver does not monitor local-oscillator intensity, needs no such composition.

A-0097 Wavelength-Dependent Beamsplitter Attack on CV-QKD
See Also
A-0097 Wavelength-Dependent Beamsplitter Attack on CV-QKD
A-0098 Saturation Attack
A-0100 Exploiting Imperfect Channel Transmittance in CV-QKD
A-0101 Power Consumption Analysis Attack on State Preparation in QKD Systems
A-0102 CV Detector-Control Attack