Intercept-and-Resend Attack
The naive intercept-and-resend attack on quantum communication involves a threat actor (Eve) intercepting quantum bits (qubits) being transmitted between two parties, measuring them, and then sending replacement qubits to the receiver. This attack fundamentally compromises the security of quantum communication by attempting to gain information about the key being established between the transmitter (Alice) and receiver (Bob).
The attack has two very different faces. In its textbook form, against a transmitter that sends true single photons, it betrays itself: measuring a state the threat actor cannot identify disturbs it, so the errors she creates appear as soon as the legitimate parties compare a sample of their results [Bennet1992]. Against a transmitter that sends weak laser pulses instead of single photons, on a link lossy enough to cover her tracks, a refined version of the same attack leaves no errors at all and can hand her the entire key while the legitimate parties see nothing wrong [Dusek2000,Sushchev2025].
| [ArteagaDiaz2022] | P. Arteaga-Dıaz, D. Cano, and V. Fernandez. "Practical Side-Channel Attack on Free-Space QKD Systems With Misaligned Sources and Countermeasures" In: IEEE Access 10, 82697–82705. (2022) 10.1109/ACCESS.2022.3196677. |
| [Babukhin2021] | D. Babukhin and D. Sych. "Explicit attacks on passive side channels of the light source in the BB84 decoy state protocol" In: J. Phys.: Conf. Ser. 1984, 012008. (2021) 10.1088/1742-6596/1984/1/012008. |
| [Bennet1992] | C.H. Bennett, F. Bessette, G. Brassard, L. Salvail, J. Smolin. "Experimental quantum cryptography" In: J. Cryptology 5, 3–28. (1992) 10.1007/BF00191318. |
| [Burenkov2010] | V. Burenkov, B. Qi, B. Fortescue, H.-K. Lo. "Security of high speed quantum key distribution with finite detector dead time" (2014) 10.26421/QIC14.3-4-2. arXiv:1005.0272. |
| [Curty2005] | Curty, M. and Lütkenhaus, N.. "Intercept-resend attacks in the Bennett-Brassard 1984 quantum-key-distribution protocol with weak coherent pulses" In: Physical Review A, vol. 71, no. 6, 2005. (2005) 10.1103/PhysRevA.71.062301. arXiv:quant-ph/0411041. |
| [Dusek2000] | Dušek, M., Jahma, M., and Lütkenhaus, N.. "Unambiguous state discrimination in quantum cryptography with weak coherent states" In: Physical Review A, vol. 62, no. 2, 2000. (2000) 10.1103/PhysRevA.62.022306. arXiv:quant-ph/9910106. |
| [Gerhardt2011] | I. Gerhardt, Q. Liu, A. Lamas-Linares, J. Skaar, C. Kurtsiefer, and V. Makarov. "Full-field implementation of a perfect eavesdropper on a quantum cryptography system" In: Nature Communications 2, 349. (2011) 10.1038/ncomms1348. |
| [Nauerth2009] | S. Nauerth, M. F¨urst, T. Schmitt-Manderbach, H. Weier, and H. Weinfurter. "Information leakage via side channels in freespace BB84 quantum cryptography" In: New J. Phys. 11, 065001. (2009) 10.1088/1367-2630/11/6/065001. |
| [Radocea2026] | "A generalized measurement key intercept at approximately 6 percent QBER against 4-PSK BB84" (2026) 10.1117/12.3093123. |
| [Scarani2009] | V. Scarani, H. Bechmann-Pasquinucci, N. J. Cerf, M. Dusek, N. Lutkenhaus, M. Peev. "The security of practical quantum key distribution" (2009) 10.1103/RevModPhys.81.1301. arXiv:0802.4155. |
| [Sushchev2025] | Sushchev I. S., Bugai K. E., Molotkov S. N., Bulavkin D. S., Sidelnikova A. S., Melkonian D. M., Vakhrusheva V. M., Lokhmatov R. Yu, Dvoretskiy D. A.. "Realistic vulnerabilities of decoy-state quantum key distribution" (2025) 10.48550/arXiv.2507.15446. |
dfe5fcfd-69c6-41a3-ba98-8ad3a11c5d70
- Quantum Communication
Technique properties
Operational properties
| Description | Countermeasure |
|---|---|
|
Linked for the side-channel-informed route through V-0012. Monitoring what the source actually emits, its timing, wavelength, polarisation and intensity, can reveal the differences between the transmitter's lasers that this route reads. |
CM-0016 State Preparation Monitoring |
|
Estimating the quantum bit error rate and aborting above the protocol's threshold is the control that catches the naive route of this attack, and catches it reliably. Intercept-and-resend on a single-photon implementation costs about 25% QBER, which follows from the Bennet1992 bound that any measurement yielding half an expected bit has probability at least s/2 of inducing a discrepancy when the data of the two parties are compared [Bennet1992]. The critical value for BB84 is 11% for one-way post-processing without preprocessing, rising to 12.4% and 12.9% with preprocessing and to at least 20.0% with two-way post-processing at a drastically reduced key rate [Scarani2009]. Since routine optical error is kept far below 5% [Scarani2009], the signature is unmistakable. The control's value is therefore high against one route and nil against the others. |
CM-0007 Proper Implementation of Quantum Communication Protocol and QBER Estimation |
|
Privacy amplification compresses a partially compromised key until the threat actor's residual knowledge of it is negligible, and against this technique the amount of compression is computable. Bennet1992 gives the rule: the two parties conservatively estimate that the amount of information leaked to the threat actor through intercept-and-resend is worth no more than a fixed multiple of the number of observed errors, the leading term being times that number, and shorten the key by at least that much [Bennet1992]. The control therefore converts a measured error rate into a quantified loss of key length, which is what makes a detected but tolerable intercept-and-resend survivable rather than fatal. |
CM-0001 Privacy amplification |
|
Linked for the side-channel-informed route through V-0012. Coupling all states into one single-mode fibre reduces the spatial differences between the transmitter's lasers that this route reads, without removing them. |
CM-0055 Single-Mode Fiber Coupling |
|
Linked for the side-channel-informed route through V-0012. A single laser diode for all states aims to remove the spatial, spectral and temporal differences between the transmitter's lasers that this route reads. It does not touch the other routes. |
CM-0056 Single Laser Diode Utilization |
|
Decoy states interleave pulses of different intensity with the signal pulses, so that the transmitter and receiver can estimate the yield and error rate of the single-photon component independently of what the threat actor does to the multi-photon pulses. Against this technique that matters because the silent route is a multi-photon route: unambiguously discriminating the four BB84 signal states requires pulses containing at least three or more photons [Curty2005]. A correctly operating decoy-state estimate prices the multi-photon fraction the threat actor could exploit and removes it from the secure key, which is what makes the route unprofitable rather than undetectable. |
CM-0003 Decoy States Employment |
| Description | Category | Impact Type |
|---|---|---|
|
Physical Layer (Quantum and Optical): the intercept-resend attack acts directly on the optical quantum channel. The threat actor (Eve) intercepts, measures and re-emits the transmitted photons, and Bennet1992 states its bound for exactly that, a measurement made on the photon while it is in transit from Alice to Bob [Bennet1992]. The attack takes place entirely on the link between the two nodes, which is why the row records it as acting in the operational environment rather than inside a certified module. Whether the attack is noticed, by contrast, is not settled at this layer. On the naive route the roughly 25% error rate it induces is visible to error-rate monitoring [Bennet1992]; on the unambiguous-discrimination route it does not introduce errors in the signal states but requires high losses in the channel [Curty2005], so what betrays it is channel loss rather than anything the physical layer reports by itself. The row therefore also carries a secondary mapping to the post-processing layer, where the statistics that decide the question are accumulated. |
L1 — Physical Layer (Quantum & Optical) | Primary Target |
|
Post-Processing and Cryptographic layer, as a secondary effect. The attack acts optically at L1, but whether it succeeds is decided here. Dusek2000's whole test of security is whether the receiver's accumulated single-click and double-click statistics place the link inside a region of insecurity [Dusek2000], the abort threshold the parties compare their error rate against is a post-processing parameter [Scarani2009], and two of this row's countermeasures, error-rate estimation and privacy amplification, are post-processing controls. A threat actor who defeats the L5 accounting keeps the key she took at L1. |
L5 — Post-Processing & Cryptographic | Secondary Effect |
| Relation type | Description | Technique |
|---|---|---|
| Combines with |
This technique creates the room in which an attack on the signal runs undetected, and the sources count the key that attack takes as the technique's result; in [Ren2019] that attack is an intercept-and-resend. [Ren2019] makes the coupling explicit and quantitative, since the threat actor tunes the splitting factor of her intercept-and-resend until the additional excess noise equals the tolerance she freed by shrinking the trusted term. In that example A-0045 supplies the information gain and this technique the concealment, and the published key fractions are computed for the two together. |
A-0103 Exploiting the Trusted Phase Noise Model in CV-QKD |
| Preceded by |
The realistic follow-on that Sajeed2015 models: intercept-and-resend with unambiguous state discrimination, which the raised mean photon number makes effective [Sajeed2015]. |
A-0064 Pulse-Intensity Increase in Two-Way QKD |
| Preceded by |
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-0099 Manipulation of Vacuum-Noise Estimation |
| Refined by |
A-0045 Intercept-and-Resend is the delivery vehicle; this technique refines it by choosing the resend intensity so that basis-mismatched events become ambiguous double clicks the receiver discards, rather than errors it counts [Lutkenhaus2000]. |
A-0068 Double-Click Attack |
| Refined by |
This technique is a variant of the intercept-and-resend attack, and says so in its own terms: [Guo2017] describes its strategy as an intercept-resend attack performed by applying an entanglement-distillation operation to the transmitted non-Gaussian mixed states. The chain is the familiar one, tap the signal, measure, decide, regenerate and resend with an electro-optic modulator. What makes it a distinct variant is where the cover comes from. A-0045 silent route buys its cover from the channel loss budget by sending vacuum when unambiguous discrimination fails, and A-0097 and A-0098 buy theirs by injecting light that moves a quantity the receiver measures. This variant pays nothing: the cover is the parties own mis-modelled fading, so the threat actor alters no device. That is also why it needs the non-Gaussianity window, which is a condition none of the other variants has. |
A-0100 Exploiting Imperfect Channel Transmittance in CV-QKD |
| Refined by |
This technique is the conventional intercept-and-resend attack with the basis guess replaced by a side-channel reading; the primary compares its zero-separation limit with conventional intercept-and-resend. At zero separation between the beams at the focal plane the probability of measuring in the correct basis is 50 percent, which agrees with that obtained in a conventional intercept-and-resend attack in which a measurement base is randomly chosen each time [ArteagaDiaz2022]. A-0045 is that attack, and its own description gives the textbook consequence, that measuring a state the threat actor cannot identify disturbs it so the errors appear as soon as the parties compare a sample. What this row adds is the identification: because the spatial mode tells her which basis to use, her error cost falls from the roughly 25 percent of the naive attack towards zero as the separation grows past about three beam radii. The two rows are therefore the same action at two different levels of knowledge, which is worth navigating between because the countermeasure sets differ completely, A-0045's being photon-number and loss-monitoring controls and this row's being source-indistinguishability controls. |
A-0081 Spatial Misalignment Exploitation in QKD |
| Refined by |
One route of this technique is an intercept-and-resend attack made cheaper by the side channel, which is what refines records. Babukhin2020 computes exactly that: the threat actor performs an optimal state discrimination over the enlarged signal space and resends what she found, and the side channel is what lowers her error cost, since applying the ordinary approach to BB84 leads to a critical error of 0.25 but if states used for transmission have additional distinguishability, the resulting critical error will be lower [Babukhin2020]. The relation is refines rather than variant_of because the two rows sit in different attack categories, A-0045 in intercept-and-resend and this row in degree-of-freedom coupling, and only one of this technique's four routes resends anything at all. |
A-0083 Side Channel Exploitation in Transmitter Imperfections |
| Refined by |
This technique is an intercept-and-resend attack with the error cost bought down by a device defect, which is how both primaries describe it. Wang2013 states it plainly, that the attack is just an individual attack based on intercept-resend attack whose minimal error rate of about 14.64 per cent is just the tolerable upper bound of error rate in individual attack, so that it is neither optimal nor realizable in this QKD system [Wang2013]. The refinement is quantitative and specific: a plain intercept-and-resend on BB84 costs about 25 per cent because the threat actor must guess a basis, while here the rotation error lifts the four states into three dimensions so she discriminates them better, and Sun2011 states that the QBER induced by Eve's attack is much lower than 25 per cent, at 14.64 per cent or 3.57 per cent with phase remapping [Sun2011]. A-0045's own body records that intercept-and-resend is used as a subaction within more complex quantum attacks, which is the role it plays here. |
A-0084 Passive Faraday Mirror Attack |
|