Skip to content
Quantum.world

The Signal

Preprint reports live keys from four-city Slovak QKD network

A preprint posted Nov. 12 describes a deployed entanglement-based QKD network linking four western Slovak cities and delivering keys in real time. It reports rates of 13.5 bit/s on a 109 km link and 4.3 bit/s on a longer link, while warning that finite-key tests do not certify the key produced on that longest route.

Preprint reports live keys from four-city Slovak QKD network

A live, four-city network

A preprint posted Nov. 12 describes a deployed entanglement-based quantum key distribution network connecting four cities in western Slovakia and delivering keys in real time. With partial dispersion compensation, it reports 13.5 bit/s over a 109 km link and 4.3 bit/s over 142.94 km, where accumulated loss was 32.1 dB, according to the paper.

Source arxiv.org

The network connects Bratislava, Trnava, Nitra and Trenčín. A source in Trnava distributes polarization-entangled photons through wavelength routing: Trnava, Bratislava and Nitra form a fully meshed sub-network, while Trenčín connects to the central node. The paper says all links ran simultaneously and were trusted-node-free; one city pair communicated without a direct fiber link.

From detections to delivered keys

The central node houses a broadband entangled-photon-pair source and a quantum network provider that routes wavelength channels to the network nodes. At each node, polarization analysis modules and superconducting nanowire single-photon detectors register photon detections. The system processes coincident detections into keys and makes them available through the ETSI GS QKD 014 interface.

The paper reports that the rates include the complete post-processing chain, rather than estimates based on raw coincidence counts. The custom software creates 256-bit symmetric keys and stores them in pools with a capacity of 30,000 keys per link. Its reconciliation step detects errors in short blocks and discards blocks with differing syndromes; it does not use error correction to repair those blocks.

Finite-key results qualify the rates

The authors’ security benchmark identifies a limitation in the implemented privacy-amplification method. For the deployed-network analysis, the paper used five-minute privacy-amplification blocks. On the longest link, between Nitra and Bratislava, the finite-key bound for the implemented scheme was zero, meaning none of the distilled key on that link could be certified under that bound. On the other links, the distilled key exceeded the positive finite-key bounds by reported factors of 2.9, 1.9 and 1.2.

The paper says the heuristic privacy-amplification setting can produce a key length above the finite-key bound for short blocks, and that finite-size corrections are needed before the delivered key can be considered secure. The authors estimate that replacing the detect-and-discard reconciliation method with error correction would raise the finite-key bound by a factor of 4.1–6.6 on the remaining deployed links. For the longest route, their estimate using error correction with f=1.1 is 29.8 ± 6.8 bit/s.

Routing design

The preprint also presents a matrix formalism for passive wavelength routing. It represents which wavelength channels reach which users, then checks whether each user pair shares exactly one entangled wavelength pair. In the paper’s four-user fully connected example, three wavelength pairs provide the connections, using half the wavelength pairs required by dedicated channels.

Sources

  1. A deployed four-city entanglement-based QKD network with real-time key delivery, benchmarked against finite-key security bounds
  2. arXiv (full text)

Related stories

Explore

More articles