Simulation of a Heterogeneous Quantum Network

TL;DR

The paper tackles the challenge of simulating heterogeneous quantum networks by extending the SeQUeNCe discrete-event simulator with hardware-faithful models for Ytterbium atom memories, microwave (transmon) memories, quantum frequency converters, and time-bin Bell-state measurement nodes. It implements entanglement generation and swapping protocols across heterogeneous links and conducts extensive simulations to map rate-fidelity trade-offs, revealing bottlenecks such as reload dynamics for Yb nodes and transducer noise in Yb-μW links. Key contributions include four device models, integration into a customizable simulation framework, and open-source availability to enable reproducible evaluation of future heterogenous designs. The findings provide design guidance (e.g., optimal attempts per reload around 65 for Yb-Yb, and the critical role of μW coherence in larger networks) to inform architecture decisions for scalable quantum networks. Overall, the work advances hardware-faithful planning and evaluation of heterogeneous quantum networks and protocols.

Abstract

Quantum networks are expected to be heterogeneous systems, combining distinct qubit platforms, photon wavelengths, and device timescales to achieve scalable, multiuser connectivity. Building and iterating on such systems is costly and slow, motivating hardware-faithful simulations to explore architecture design space and justify implementation decisions. This paper presents a framework for simulating heterogeneous quantum networks based on SeQUeNCe, a discrete-event simulator of quantum networks. We introduce faithful device models for two representative platforms - Ytterbium atoms and superconducting qubits. On top of these models, we implement entanglement generation and entanglement swapping protocols for time-bin encoded photons that account for disparate clock rates and quantum frequency conversion and transducer losses/noise brought by the heterogeneity. Using extensive simulations, we map the rate-fidelity trade space and identify the dominant bottlenecks unique to heterogeneous systems. The models are open source and extensible, enabling reproducible evaluation of future heterogeneous designs and protocols.

Figures (8)

• Figure 1: (a) Heterogeneous quantum network. Yb-nodes serve as quantum repeaters at the network core, while $\mu$W-nodes serve as quantum computing nodes at the network edge. (b) Hardware components of a link that connects two heterogeneous nodes. The hardware components include quantum channels, quantum frequency converters (QFC), time-bin Bell State Measurement (BSM) devices, quantum transducers (QT), and classical channels (not shown in the figure).
• Figure 2: Ytterbium (Yb) Node. (a) shows energy levels, atomic states, and atom transitions during the initialization, preparation, and generation stepscovey. Each energy level corresponds to a manifold in Yb specified by term symbols (e.g. $^3D_1$) and nuclear spin (e.g. $F=\frac{3}{2}$). A time-bin encoded photon at $1389nm$ is emitted and collected by the fiber. (b) shows the steps of the time-bin entanglement generation and their timing. In particular, the preparation and generation steps are illustrated in further detail. (c) explains the symbols used in (b) and includes the transition details in (a).
• Figure 3: Microwave ($\mu$W) Node. (a) shows the energy levels of the transmon and transmon transitions during the initialization, preparation, and generation. After the microwave-optical (M-O) transduction, a time-bin encoded photon at $1550nm$ is sent to the fiber. (b) shows the steps of the time-bin entanglement generation and their timing. In particular, the preparation and generation steps are illustrated in further detail. (c) explains the symbols used in (b) and the transition details in (a).
• Figure 4: (a) Quantum frequency converter and (b) Time-bin Bell state measurement node. Signal photons in different colors reflect different frequencies. Gray non-filled photons are noise photons from the QFC.
• Figure 5: SeQUeNCe has six modules, and each module has several components (not all components are shown). We introduced four new hardware models (green) and customized three components (yellow). The numbers in parenthesis indicate the amount of parameters for each new hardware module.