153
Part Three · E · Hardware, simulation and practice

Quantum networks, repeaters, and distributed computing

Send a photon down 1,000 km of the best optical fibre and the odds it arrives are one in 102010^{20}. Classical networks solve this with amplifiers every hundred kilometres. A quantum signal cannot be amplified, because an amplifier would have to copy an unknown state, which no-cloning forbids (Chapter 102). The way through is a quantum repeater: create entanglement over short links, store it, and stitch the links together by teleportation.

On the plate: entanglement rate against distance, with and without repeaters.

Entanglement swapping and purification

If A shares a Bell pair with B, and B shares another with C, then a Bell measurement at B teleports the entanglement, leaving A and C entangled though they never interacted: entanglement swapping (Chapter 103). Errors accumulate with each swap, so Hans Briegel, Wolfgang Dür, Ignacio Cirac and Peter Zoller proposed in 1998 to interleave swapping with purification: consume two noisy pairs to make one better pair (Chapter 106). Lu-Ming Duan, Mikhail Lukin, Cirac and Zoller (DLCZ, 2001) showed how to do it with atomic ensembles and linear optics. Links are heralded: a detector click tells both ends that entanglement succeeded, so they keep it in a quantum memory and wait for neighbouring links. Established

Plan a repeater chain

Distance 1000 km, split into 16 segments. The plate uses an idealized model: 0.2 dB/km fibre, a source of 10⁹ attempts per second for direct transmission, heralded links that succeed with probability (half the segment transmission), each attempt waiting a light-travel time for the herald, perfect memories, and deterministic swapping, with the standard factor (3/2)(3/2) per nesting level for waiting on neighbours. Direct transmission falls exponentially with distance; the repeater falls only polynomially. Real memories decohere and real swaps fail, so practical rates are far lower, and the race is to build memories that last long enough. Established for the scaling; the numbers are a model.

Distributed computing, and India’s network

Linking quantum processors lets a computation span modules. In February 2025 Dougal Main and colleagues at Oxford ran Grover’s search across two trapped-ion modules two metres apart, teleporting a controlled-Z gate between them through a photonic link with 86% fidelity. Satellite links avoid fibre loss over continental distances; China’s Micius satellite distributed entanglement over 1,200 km in 2017. India’s National Quantum Mission reported in April 2026 a 1,000 km secure-communication network built with QNu Labs, using trusted nodes, not yet entanglement-based repeaters. Established for the experiments.

What you now have

Why quantum signals cannot be amplified, entanglement swapping, purification, heralded links and memories, the DLCZ scheme, repeater scaling, distributed gates, and the current networks.

· · ·

Three questions to carry with you.

Why can’t quantum signals be amplified?tap
Amplification would copy an unknown quantum state, forbidden by no-cloning.
What does a herald do?tap
Tells both ends a link succeeded, so they store it and wait for the rest of the chain.
What limits real repeaters most?tap
Quantum memory lifetime and efficiency.

And one to work out. At 0.2 dB/km, transmission through 1,000 km is 10x10^x. What is xx?

x =

Sources: H.-J. Briegel, W. Dür, J. I. Cirac and P. Zoller, Phys. Rev. Lett. 81, 5932 (1998). L.-M. Duan, M. D. Lukin, J. I. Cirac and P. Zoller, Nature 414, 413 (2001). N. Sangouard et al., Rev. Mod. Phys. 83, 33 (2011). J. Yin et al., Science 356, 1140 (2017). D. Main et al., Nature 638, 383 (2025). National Quantum Mission, India, 1,000 km network announcement (April 2026). Last checked 18 November 2026.

Next · Chapter 154
Quantum sensing and metrology

LIGO uses squeezed vacuum to hear black holes merge.

Turn the page