Chapter 34: The Quantum Internet

34.1 Entanglement as a Resource for Networking

Classical networking is built on one primitive: sending a bit from A to B. Quantum networking adds a second, qualitatively different primitive: sharing an entangled pair between A and B. Once Alice and Bob share an EPR pair $|\Phi^+\rangle = \frac{1}{\sqrt{2}}(|00\rangle + |11\rangle)$, they possess a resource that enables tasks no amount of classical communication can achieve.

What Entanglement Enables

Shared entanglement, combined with classical communication and local quantum operations, unlocks a suite of powerful protocols:

• Quantum teleportation. Alice can transmit an arbitrary qubit state to Bob by consuming one shared EPR pair and sending two classical bits. No quantum channel is needed at the moment of transfer - only pre-shared entanglement and a classical channel.
• Superdense coding. Alice can transmit two classical bits to Bob by sending only one qubit, provided they share an EPR pair. This doubles the classical capacity of a quantum channel.
• Quantum key distribution. Entanglement-based QKD (the E91 protocol) derives its security from Bell inequality violations, providing device-independent guarantees that the key is secret.
• Distributed quantum computation. Entanglement allows separate quantum processors to perform joint computations, effectively creating a larger virtual quantum computer from smaller networked nodes.
• Blind quantum computation. A client with minimal quantum resources can delegate a computation to a remote quantum server without revealing the input, output, or algorithm.
• Quantum sensor networks. Entanglement shared across a network of sensors enables measurements with precision beyond what independent sensors can achieve (see Chapter 36).

Entanglement as a Unit of Currency

Just as classical networks measure capacity in bits per second, a quantum network's capacity is measured in ebits per second - the rate at which it can distribute entangled pairs. One ebit is one maximally entangled pair shared between two parties. The quality of these pairs matters too: real-world entanglement is never perfect, and the fidelity of the shared state relative to a perfect Bell state determines how useful each ebit is.

A pair with fidelity $F = \langle \Phi^+ | \rho | \Phi^+ \rangle$ close to 1 is nearly perfect. Pairs with $F < 0.5$ are useless for most protocols - they carry no more quantum correlation than a separable state. The art of quantum networking lies in generating high-fidelity entanglement at high rates over long distances.

No-Cloning and the Networking Challenge

Classical networks amplify signals with repeaters that copy and retransmit bits. The no-cloning theorem forbids this for quantum states: you cannot copy an unknown qubit. This single fact makes quantum networking fundamentally harder than classical networking. Photons carrying quantum information are lost in fiber at a rate of roughly 0.2 dB/km at telecom wavelengths (1550 nm), and you cannot simply amplify the signal. After about 100 km, the probability of a photon surviving is below 1%. Extending entanglement beyond this range requires entirely new approaches - quantum repeaters - which we will explore in Section 34.4.

34.2 Stages of the Quantum Internet (Wehner Framework)

In 2018, Stephanie Wehner, David Elkouss, and Ronald Hanson published a landmark paper in Science proposing a roadmap for the quantum internet organized into six stages of increasing capability. Each stage unlocks new applications and requires progressively more sophisticated hardware. This framework has become the standard reference for gauging progress toward a quantum internet.

Stage 1: Trusted-Repeater Networks

At the lowest stage, quantum signals (typically single photons encoding QKD key bits) are transmitted between adjacent nodes, but the intermediate relay nodes are trusted - they must decrypt and re-encrypt the key material classically. This is not a true quantum network in the purest sense because the relay nodes see the key in the clear, but it provides enhanced security over purely classical infrastructure as long as the relay nodes are physically secured.

Applications: QKD with trusted relays. This is the stage at which most deployed QKD networks currently operate, including China's Beijing-Shanghai backbone and the CN-QCN network spanning over 10,000 km with 145 fiber backbone nodes across 80 cities.

Stage 2: Prepare-and-Measure Networks

Nodes can prepare quantum states and send them to other nodes for measurement, but there is no entanglement generation or quantum memory. End-to-end quantum communication is possible over short distances without trusting intermediate nodes.

Applications: End-to-end QKD (BB84-style) without trusted relays over metropolitan distances, quantum random number generation, and simple quantum coin flipping protocols.

Stage 3: Entanglement-Distribution Networks

This is the first stage where entanglement is explicitly generated and distributed between end nodes. Nodes can create EPR pairs with their neighbors and, through entanglement swapping at intermediate nodes, extend entanglement to non-adjacent parties. However, the nodes have limited or no quantum memory - entanglement must be used immediately or it is lost.

Applications: Device-independent QKD, entanglement-based QKD (E91), and certifiable randomness generation.

Stage 4: Quantum-Memory Networks

Nodes possess quantum memories that can store entangled qubits for a meaningful duration. This enables protocols that require entanglement to be available on demand rather than consumed immediately. Quantum memories allow entanglement purification (distilling high-fidelity pairs from multiple noisy ones) and more complex multi-party protocols.

Applications: Quantum teleportation on demand, blind quantum computation, and the beginning of distributed quantum computing with small numbers of qubits.

Stage 5: Few-Qubit Fault-Tolerant Networks

Each node is a small quantum computer with error correction, capable of performing local quantum operations on a handful of logical qubits. The network can execute multi-round quantum protocols with high reliability.

Applications: Clock synchronization beyond classical limits, distributed quantum sensing, leader election, and Byzantine agreement protocols with quantum advantages.

Stage 6: Quantum Computing Networks

The full quantum internet: every node is a fault-tolerant quantum computer, and the network can distribute large entangled states across many nodes. Arbitrary distributed quantum computations can be performed.

Applications: All of the above, plus full distributed quantum computing, quantum cloud computing with verifiable delegation, and applications not yet imagined - just as the classical internet enabled applications (social media, streaming, cloud computing) that its inventors never foresaw.

34.3 Physical Implementations (Fiber, Satellite, Hybrid)

A quantum internet must physically transmit qubits - typically encoded in photons - between nodes. Three main physical channels are under development, each with distinct advantages and limitations.

Optical Fiber

Standard telecommunications fiber at 1550 nm wavelength offers the lowest loss (approximately 0.2 dB/km) and can leverage existing fiber infrastructure. However, loss is exponential: after 100 km, only about 1% of photons survive; after 300 km, the fraction drops to roughly $10^{-6}$. This exponential attenuation is the fundamental bottleneck.

Despite this limitation, fiber is the workhorse of current QKD deployments. Metropolitan networks spanning tens of kilometers achieve key rates of kilobits per second or higher. China's Beijing-Shanghai QKD backbone uses fiber links of up to 100 km between trusted relay nodes. European networks like the Madrid quantum network and the Cambridge quantum network use dark fiber (dedicated, unshared fiber) for higher performance.

Free-Space and Satellite Links

For intercontinental distances, sending photons through optical fiber is impractical due to exponential loss. Free-space transmission through the atmosphere, and especially via satellite, offers an alternative. In vacuum (or near-vacuum in low Earth orbit), photon loss follows an inverse-square law rather than exponential decay, making long-distance links feasible.

China's Micius satellite (launched 2016) demonstrated the viability of satellite QKD, achieving entanglement distribution over 1,200 km and enabling the first intercontinental quantum-secured video call between Beijing and Vienna in 2017. By 2020, an integrated satellite-ground network spanning 4,600 km was operational. In 2024, the smaller Jinan-1 microsatellite demonstrated real-time QKD with ground stations across China and in Stellenbosch, South Africa - a distance of approximately 12,900 km.

Hybrid Architectures

The most practical near-term architecture combines fiber for metropolitan-area links (up to roughly 100 km) with satellite links for intercity and intercontinental connections. Ground stations convert between the telecom-wavelength photons used in fiber and the near-infrared photons optimal for free-space transmission.

A key engineering challenge is the wavelength interface. Many quantum memory platforms (trapped ions, NV centers, atomic ensembles) operate at visible or near-infrared wavelengths that suffer high loss in fiber. Quantum frequency conversion - coherently shifting a photon's wavelength without destroying its quantum state - bridges this gap and is an active area of research.

Physical Qubit Platforms for Network Nodes

The nodes of a quantum network need more than photon sources and detectors; they need matter qubits that can interface with photons. Leading platforms include:

• Nitrogen-vacancy (NV) centers in diamond. These solid-state defects have long-lived electron spin qubits that can be entangled with emitted photons. The Delft group demonstrated the first loophole-free Bell test (2015) and multi-node entanglement (2021) using NV centers.
• Trapped ions. Ions confined in electromagnetic traps offer excellent coherence times and high-fidelity gates. Photons emitted during ion transitions carry entanglement that can be distributed over fiber.
• Atomic ensembles. Clouds of cold atoms can collectively emit single photons entangled with the ensemble's internal state, providing a natural quantum memory with built-in light-matter coupling.
• Quantum dots. Semiconductor quantum dots can emit indistinguishable single photons at high rates and are compatible with chip-scale integration.

34.4 Quantum Network Protocols (Entanglement Swapping, Purification)

The no-cloning theorem means we cannot amplify quantum signals classically. To extend entanglement beyond the range of direct photon transmission, quantum networks employ two key protocols: entanglement swapping and entanglement purification. Together, these form the backbone of quantum repeater architectures.

Entanglement Swapping

Suppose Alice shares an EPR pair with a relay node Charlie, and Charlie shares a separate EPR pair with Bob. Neither Alice nor Bob are entangled with each other - yet. Charlie can create entanglement between Alice and Bob by performing a Bell-state measurement (BSM) on his two qubits (one from each pair). This projects Alice's and Bob's qubits into an entangled state, even though they never interacted directly.

The protocol works as follows:

1. Alice and Charlie share the Bell state $|\Phi^+\rangle_{AC}$. Charlie and Bob share $|\Phi^+\rangle_{CB}$.
2. Charlie performs a Bell-state measurement on his two qubits (one from each pair). This measurement has four possible outcomes, corresponding to the four Bell states.
3. Charlie sends his two-bit measurement result to Bob over a classical channel.
4. Depending on the result, Bob applies a local correction (one of $I$, $X$, $Z$, or $XZ$) to his qubit. After correction, Alice and Bob share the state $|\Phi^+\rangle_{AB}$.

The net effect: entanglement has been "swapped" from the two short-range pairs to a single long-range pair. By chaining multiple swaps through a series of relay nodes, entanglement can be extended across arbitrarily long distances - this is the principle behind a quantum repeater.

Entanglement Swapping Circuit

The Bell-state measurement at the heart of entanglement swapping consists of a CNOT gate followed by a Hadamard gate, then measurement of both qubits. In the circuit below, we simulate the full entanglement swapping protocol on four qubits. Qubits 0 and 1 form the Alice-Charlie pair; qubits 2 and 3 form the Charlie-Bob pair. Charlie performs the BSM on qubits 1 and 2, projecting qubits 0 and 3 into an entangled state.

Run the circuit and observe the correlations between qubit 0 (Alice) and qubit 3 (Bob). Despite never interacting directly, they become perfectly correlated - a hallmark of entanglement swapping.

Examine the measurement results carefully. You will notice that qubits 0 and 3 (Alice and Bob) are always correlated: whenever qubit 0 is 0, qubit 3 is 0, and whenever qubit 0 is 1, qubit 3 is 1 (possibly with a correction depending on qubits 1 and 2). The four possible outcomes of Charlie's BSM (qubits 1 and 2) each occur with equal probability of 25%, and for each outcome, Alice and Bob's qubits are in a definite Bell state. In a real network, Bob would apply a Pauli correction conditioned on Charlie's classical message to always obtain $|\Phi^+\rangle$.

Interactive: Teleportation Fidelity Under Noise

Real quantum networks must contend with noise on every channel. The simulation below shows how depolarizing noise degrades the fidelity of quantum teleportation. As the noise parameter increases, the teleported state becomes increasingly mixed with random errors, reducing the protocol's effectiveness.

Entanglement Purification

Real entangled pairs are never perfect. Noise in the quantum channel, imperfect sources, and decoherence in quantum memories all reduce the fidelity of distributed pairs. If a protocol requires fidelity above some threshold (e.g., $F > 0.9$ for useful QKD), low-fidelity pairs must be improved.

Entanglement purification (also called entanglement distillation) is the process of converting multiple low-fidelity pairs into fewer high-fidelity pairs using only local operations and classical communication (LOCC). The basic idea:

1. Alice and Bob share two noisy EPR pairs, each with fidelity $F$ where $0.5 < F < 1$.
2. Both Alice and Bob apply CNOT gates locally (using one pair as control and the other as target).
3. Both measure their target qubits and compare results over a classical channel.
4. If their measurement results agree, the remaining pair has higher fidelity than the originals. If they disagree, the remaining pair is discarded.

By repeating this process iteratively, fidelity can be boosted arbitrarily close to 1 at the cost of consuming more pairs. The trade-off between the number of consumed pairs and the output fidelity is governed by the purification yield.

Quantum Repeater Architecture

A quantum repeater combines entanglement swapping and purification to distribute high-fidelity entanglement over long distances. The basic architecture divides a long link into shorter segments:

1. Generate entanglement over each short segment (e.g., 50-100 km of fiber).
2. Purify the entanglement in each segment to achieve sufficient fidelity.
3. Perform entanglement swapping at each intermediate node to extend entanglement across segments.
4. Optionally purify again after swapping to compensate for the fidelity loss introduced by imperfect BSMs.

Three generations of quantum repeater designs have been proposed:

• First generation: Heralded entanglement generation with quantum memories and nested purification. Requires long-lived memories but only simple operations.
• Second generation: Uses quantum error correction at each repeater node to protect stored qubits, reducing memory time requirements.
• Third generation: Fully error-corrected repeaters that encode logical qubits and perform fault-tolerant operations. These can achieve rates that scale polynomially (rather than exponentially) with distance.

34.5 Standards and Current Deployments

The transition from laboratory demonstrations to operational quantum networks requires standardization, interoperability, and engineering at scale. Significant progress has been made on all three fronts.

Standardization Efforts

Several organizations are working to define standards for quantum networking:

• IRTF (Internet Research Task Force): Published RFC 9340 in 2023, "Architectural Principles for a Quantum Internet," establishing the conceptual framework for integrating quantum communication into internet architecture.
• ITU-T (International Telecommunication Union): Study Group 13 has published recommendations on QKD network architecture, key management, and interoperability requirements.
• ETSI (European Telecommunications Standards Institute): The Industry Specification Group on QKD (ISG-QKD) has published standards covering QKD module interfaces, key delivery APIs, and security proofs.
• IEEE: Working groups on quantum networking are developing standards for quantum-classical network integration.

Major Deployments

China. The most extensive quantum network deployment in the world. The CN-QCN (China Quantum Communication Network) is a carrier-grade QKD network spanning over 10,000 km, incorporating 145 fiber backbone nodes and 20 metropolitan networks covering 17 provinces and 80 cities. The network integrates the Beijing-Shanghai backbone (2,000 km, operational since 2017) with satellite links through Micius and the newer Jinan-1 microsatellite, which demonstrated intercontinental QKD with South Africa in 2024.

Europe. The EuroQCI (European Quantum Communication Infrastructure) initiative aims to build a pan-European quantum communication network. National networks in the Netherlands (QuTech), UK (UKQN), Spain, and Austria serve as testbeds. The Delft group demonstrated a three-node quantum network using NV centers in 2021, representing one of the first entanglement-based (Stage 3) network demonstrations.

United States. The DOE (Department of Energy) has established quantum network testbeds at national laboratories, including Argonne, Brookhaven, Fermilab, and Oak Ridge. The Chicago-area quantum network connects Argonne and the University of Chicago via a 124-mile fiber loop. Private companies including Toshiba and Qubitekk have deployed metropolitan QKD systems.

Other. South Korea, Japan, Singapore, and Canada all have active QKD network programs. South Korea's SK Telecom operates a commercial QKD link. Japan's NICT has run the Tokyo QKD Network since 2010.

Challenges Ahead

Several major challenges remain on the path to a full quantum internet:

• Quantum memory coherence times. Current quantum memories decohere in milliseconds to seconds, but repeater protocols may require storage times of seconds to minutes for long-distance links.
• Entanglement generation rates. The rate at which entangled pairs can be generated and distributed is limited by photon loss, detector efficiency, and source brightness.
• Wavelength conversion efficiency. Converting between matter-qubit emission wavelengths and telecom wavelengths with high fidelity and efficiency remains an open engineering challenge.
• Scalability. Moving from three-node laboratory networks to city-scale and intercontinental entanglement-based networks requires orders of magnitude improvement in all metrics simultaneously.
• Integration with classical infrastructure. Quantum signals must coexist with classical traffic on shared fiber, requiring careful management of noise from classical channels.

Can Entanglement Enable Faster-Than-Light Communication?

This is one of the most common misconceptions about quantum mechanics. Let us work through it carefully with a Predict-Observe-Explain exercise.