Source-linked AI summary
A quantum access network
Bernd Fröhlich, James F. Dynes, Marco Lucamarini, Andrew W. Sharpe, Zhiliang Yuan, Andrew J. Shields
TL;DR
The study simulates per-user secure key rates in a network with more than two users using measured experimental parameters. The receiver sorts detection events into user-specific time bins, while simulations estimate per-user secure key rates for networks with more than two users. The receiver assigns detection events to N users using separate time bins.
Problem
The study simulates per-user secure key rates in a network with more than two users using measured experimental parameters.
Source: C. Simulation · p. 5Method
The receiver sorts detection events into user-specific time bins, while simulations estimate per-user secure key rates for networks with more than two users.
Source: A. Experimental setup · p. 5, C. Simulation · p. 5Results
The receiver assigns detection events to N users using separate time bins.
Source: A. Experimental setup · p. 5Takeaways & Limitations
Time-bin assignment provides a way to distinguish users’ detection events at a shared receiver.
Source: A. Experimental setup · p. 5
Abstract
from arXiv · showhide
The theoretically proven security of quantum key distribution (QKD) could revolutionise how information exchange is protected in the future. Several field tests of QKD have proven it to be a reliable technology for cryptographic key exchange and have demonstrated nodal networks of point-to-point links. However, so far no convincing answer has been given to the question of how to extend the scope of QKD beyond niche applications in dedicated high security networks. Here we show that adopting simple and cost-effective telecommunication technologies to form a quantum access network can greatly expand the number of users in quantum networks and therefore vastly broaden their appeal. We are able to demonstrate that a high-speed single-photon detector positioned at a network node can be shared between up to 64 users for exchanging secret keys with the node, thereby significantly reducing the hardware requirements for each user added to the network. This point-to-multipoint architecture removes one of the main obstacles restricting the widespread application of QKD. It presents a viable method for realising multi-user QKD networks with resource efficiency and brings QKD closer to becoming the first widespread technology based on quantum physics.
I. METHODS · A. Experimental setup.
The setup uses independently aligned quantum transmitters and a shared receiver interferometer with time-bin detection to support user-specific photon decoding. The transmitters use decoy-state pulses and components for phase, polarisation, and pulse-broadening compensation.
A. Experimental setup.
- Each transmitter sends signal, decoy, and vacuum pulses at 0.5, 0.1, and 0.0002 photons/pulse, with respective probabilities of 98.83%, 0.78%, and 0.39%.The laser pulses are shorter than 50 ps, with selectable repetition rates up to 1 GHz; the two transmitters are tuned to ITU channels 33 and 34.Source: A. Experimental setup · p. 4
- Transmitter and receiver interferometers are delay-matched, and each transmitter independently compensates phase drift; the receiver interferometer needs no compensation.The interferometer components require neither temperature stabilisation nor vibration isolation.Source: A. Experimental setup · pp. 4–5
- Each transmitter combines a short-pulse laser, intensity and phase modulation, an asymmetric Mach–Zehnder interferometer, a fibre Bragg grating, an attenuator, and a polarisation controller.The intensity modulator and output attenuator set the signal, decoy, and vacuum pulse power levels.Source: A. Experimental setup · p. 4
- Polarising beam splitters direct photons into the correct receiver arm, avoiding a 3 dB penalty, while transmitter polarisation is pre-compensated to maximise receiver count rate.A fibre Bragg grating compensates fibre-link pulse broadening to avoid a further 1.5 dB decrease in count rate.Source: A. Experimental setup · p. 5
- The receiver decodes transmitter phase information using a reference Mach–Zehnder interferometer and 1 GHz InGaAs avalanche photodiodes operated at −30 °C.Typical detection efficiency is 15%, with 8×10−6 dark counts per gate and 4.5% afterpulse probability.Source: A. Experimental setup · p. 5
- Detection events are sorted into N time bins for N users, with each transmitter aligned to Bob’s master clock so photons occupy a user-specific bin.A variable transmitter delay continuously controls alignment using feedback generated from detector count rate.Source: A. Experimental setup · p. 5
B. Secure key rate.
The study implements decoy-state BB84 and estimates single-photon parameters separately for each user. Its finite-size security analysis yields bit rates close to the asymptotic limit for 20-minute key sessions, using carefully selected protocol parameters.
B. Secure key rate.
- Single-photon parameters are estimated separately for each user from the error rates of signal, decoy, and vacuum states after mismatched-basis events are discarded.The setup implements standard decoy-state BB84; the secure-rate bound accounts for single-photon bits, sifted-bit errors, vacuum-origin bits, and error-correction efficiency.Source: B. Secure key rate · p. 5
- For 20-minute key sessions, the finite-size analysis achieves bit rates close to the asymptotic limit, with security parameter ε=10^-10.The finite-size penalty Δ is proportional to √Q and log2(ε^-1); the error-correction efficiency fEC is set to 1.1.Source: B. Secure key rate · p. 5
- Decoy level and probability are selected using simulations, and one parameter set is used across all measurements for stable results across configurations.The estimated single-photon contribution Q1 depends directly on the chosen decoy photon flux and probability.Source: B. Secure key rate · p. 5
C. Simulation.
The simulation estimates secure key rates per user for multi-user networks using measured experimental parameters. It models detector errors, user cross-talk, splitter losses, and decoy-protocol key extraction.
C. Simulation.
- Each active transmitter is simulated at 1 GHz/N, with equal fibre distance to the node, starting from a single transmitter in an otherwise empty network.The simulation uses measured experimental parameters.Source: C. Simulation · p. 5
- Detector detection probability includes photon flux, fibre and splitter losses, and receiver efficiency; the simulation also sets optical error, afterpulse, dark-count, and detection-efficiency parameters.Values: eopt = 0.5 %, pA = 4.5 %, pD = 2 · 8 × 10−6, and ηBob = 9.04 %.Source: C. Simulation · p. 5
- Adding active users increases QBER through cross-talk, whose average added count rate is estimated from Fig. 4a using afterpulse/N and enhanced cross-talk at short gate separation.The QBER model includes counts from cross-talk between users.Source: C. Simulation · p. 5
- Measured splitter losses are 9.7 dB for 1x8, 13 dB for 1x16, 16.1 dB for 1x32, and 19.5 dB for 1x64 splitting.Source: C. Simulation · pp. 5–6
- Secure key rates are calculated with refined QBER analysis across all three decoy-protocol signal levels, using the same routine as for experimental data.For 32- and 64-user networks, key-session length increases from 20 min to 2 h and 12 h, respectively.Source: C. Simulation · p. 6