On the Swapping Capacity of a Quantum Repeater

Quantum networks rely on end-to-end entanglement distribution to achieve long-distance quantum communication, yet establishing these connections is hindered by decoherence and limited quantum memory performance. This paper investigates the capacity of a memory-based quantum repeater to perform entanglement swapping between two quantum links, which the authors term the end-to-end (E2E) entanglement throughput. The central research question is how to maximize this throughput rate while maintaining a guaranteed minimum E2E entanglement fidelity.

To answer this, the authors model quantum memory dynamics, entanglement generation, and swapping using queueing systems. They account for heterogeneous dephasing and depolarizing noises in memories, Bell-state measurement imperfections, and classical communication delays. Because E2E fidelity degrades as qubits wait in memory, the fidelity constraint is mathematically translated into a constraint on the maximum holding times (MHTs) of entangled qubits at the repeater. By treating these MHTs as optimization variables, the authors formulate and solve a throughput-maximization problem for a two-link quantum repeater building block.

The study demonstrates that constraining the minimum E2E entanglement fidelity is directly equivalent to enforcing maximum holding times on the repeater's quantum memories. By utilizing double-ended queueing approximations with abandonment, the authors provide tractable analytical formulas to predict E2E entanglement throughput as a function of these holding times. Numerical and simulation results confirm that these models accurately predict throughput across diverse link configurations. Furthermore, optimizing the MHTs rather than using an unconstrained baseline yields significantly higher throughput of useful, high-fidelity E2E entanglements.

Quantifying the swapping capacity of a quantum repeater is essential for designing scalable quantum networks and the future quantum internet. By framing entanglement distribution as a queueing optimization problem balancing rate against decoherence, this work offers network architects a principled methodology for resource allocation. Because this two-link model serves as a foundational building block for larger repeater chains and quantum switches, the framework provides a pathway for evaluating complex heterogeneous quantum network performance.