Adaptive Resource Orchestration for Distributed Quantum Computing Systems

TL;DR

Distributed quantum computing requires networking QPUs to scale beyond a single device. The paper proposes ModEn-Hub, a hub‑and‑spoke photonic architecture with an Adaptive Entanglement Generation Module and a Quantum Network Orchestrator to produce, cache, and schedule high‑fidelity Bell pairs for teleportation‑based operations. It introduces reinforcement‑learning–driven entanglement policies and a prototype simulator with a four‑QPUs case study, showing near‑term gains in teleportation success (~90%) under tight round budgets compared with a naïve baseline (~30%), at the cost of increased entanglement attempts (~10–12 vs ~3). The results support adaptive resource orchestration as a scalable, practical path to data‑centre‑scale quantum networking on near‑term hardware.

Abstract

Scaling quantum computing beyond a single device requires networking many quantum processing units (QPUs) into a coherent quantum-HPC system. We propose the Modular Entanglement Hub (ModEn-Hub) architecture: a hub-and-spoke photonic interconnect paired with a real-time quantum network orchestrator. ModEn-Hub centralizes entanglement sources and shared quantum memory to deliver on-demand, high-fidelity Bell pairs across heterogeneous QPUs, while the control plane schedules teleportation-based non-local gates, launches parallel entanglement attempts, and maintains a small ebit cache. To quantify benefits, we implement a lightweight, reproducible Monte Carlo study under realistic loss and tight round budgets, comparing a naive sequential baseline to an orchestrated policy with logarithmically scaled parallelism and opportunistic caching. Across 1-128 QPUs and 2,500 trials per point, ModEn-Hub-style orchestration sustains about 90% teleportation success while the baseline degrades toward about 30%, at the cost of higher average entanglement attempts (about 10-12 versus about 3). These results provide clear, high-level evidence that adaptive resource orchestration in the ModEn-Hub enables scalable and efficient quantum-HPC operation on near-term hardware.

Figures (3)

• Figure 1: ModEn‑Hub Architecture for Distributed Quantum Computing. The figure illustrates the Modular Entanglement Hub (ModEn‑Hub) architecture, which enables entanglement-based interconnectivity between two quantum computing clusters (Cluster 1 and Cluster 2). Each cluster contains one or more quantum processing units (QPUs), each coupled with a local Control Module and an Interconnect. A central Quantum Network Hub connects the clusters and comprises two main components: (1) the Adaptive Entanglement Generation Module, responsible for generating high-fidelity entangled photon pairs on demand and distributing them via optical quantum links (blue dashed lines), and (2) the Quantum Network Orchestrator, which coordinates the timing and routing of entanglement generation, quantum teleportation, and remote gate operations across QPUs. In parallel, low-latency classical control channels (gray dashed lines) carry measurement outcomes and synchronization signals between the hub and QPUs to enable real-time feed-forward control. The architecture supports scalable and heterogeneous distributed quantum computing by decoupling entanglement generation and control coordination from the quantum clusters.
• Figure 2: Quantum‑HPC implementation of the ModEn‑Hub architecture. Heterogeneous QPUs and classical control servers are interconnected via a central Quantum Network Hub that delivers on‑demand Bell pairs and orchestrates teleportation‑based non‑local gates. The inset shows a reconfigurable, blocking circuit‑switched photonic fabric of bounded degree (not full mesh): any‑to‑any connections are scheduled on demand using time/frequency multiplexing and, when needed, entanglement swapping, subject to hub/QPU port constraints.
• Figure 3: Scalability and cost of distributed teleportation under two strategies. (a) Teleportation success rate versus number of QPUs ($N$): the orchestrated policy sustains high performance as $N$ grows, while the naïve sequential baseline degrades. (b) Average entanglement attempts per teleportation, showing the higher resource expenditure required by orchestration to obtain the reliability gains in (a).