Network-Aware Scheduling for Remote Gate Execution in Quantum Data Centers

TL;DR

This work tackles scheduling non-local gates in distributed quantum data centers under probabilistic entanglement generation and finite coherence times. It compares static and dynamic entanglement scheduling using an event-driven simulator on a Clos topology with intra- and cross-rack EPR generation and a probabilistic model with entanglement success probability $p$ and memory cutoff $T_c$. The study analyzes benchmark circuits (QFT, Quantum Volume, QAOA, Random) to identify when dynamic scheduling provides the most benefit, and it highlights how coherence and lookahead interact to affect performance. The results offer practical guidance for scheduler design, resource provisioning, and network-aware partitioning to enable scalable deployment of modular quantum architectures.

Abstract

Modular quantum computing provides a scalable approach to overcome the limitations of monolithic quantum architectures by interconnecting multiple Quantum Processing Units (QPUs) through a quantum network. In this work, we explore and evaluate two entanglement scheduling strategies-static and dynamic-and analyze their performance in terms of circuit execution delay and network resource utilization under realistic assumptions and practical limitations such as probabilistic entanglement generation, limited communication qubits, photonic switch reconfiguration delays, and topology-induced contention. We show that dynamic scheduling consistently outperforms static scheduling in scenarios with high entanglement parallelism, especially when network resources are scarce. Furthermore, we investigate the impact of communication qubit coherence time, modeled as a cutoff for holding EPR pairs, and demonstrate that aggressive lookahead strategies can degrade performance when coherence times are short, due to premature entanglement discarding and wasted resources. We also identify congestion-free BSM provisioning by profiling peak BSM usage per switch. Our results provide actionable insights for scheduler design and resource provisioning in realistic quantum data centers, bringing system-level considerations closer to practical quantum computing deployment.

Figures (9)

• Figure 1: A Clos topology for a quantum data center and the physical model of cross-rack and intra-rack entanglement generation.
• Figure 2: An instance of applying static and dynamic scheduling algorithm to a quantum circuit DAG and how dynamic scheduler can reduce execution delay.
• Figure 3: Quantum data center topology used in the evaluation.
• Figure 4: Execution delay vs. cross-rack entanglement success probability ($p$). Solid lines: dynamic scheduler. Dashed lines: static scheduler. Parentheses show (BSMs, communication qubits) in a 4-QPU setup. Intra-rack success probability is fixed and is $\frac{1}{2}$.
• Figure 5: The ratio of dynamic scheduler delay to static scheduler delay in QAOA, QFT, and QV circuits with different number of data qubits (a) and number of two qubit gates in each layer of different circuits (b) Our setup has 4 QPUs and 5 BSMs at each switch.