Networks Optimise Performance with Real-Time Error Visibility

A new network-layer architecture, SCOPE, improves performance in emerging quantum networks. Xiaojie Fan and colleagues at Stony Brook University present a syndrome-based control plane that optimises both routing and quantum error correction using passive telemetry. The architecture addresses a key need to move beyond physical link fidelity as the primary metric, instead focusing on end-to-end logical error rates as quantum networks mature. By harvesting error syndromes from QEC decoders, SCOPE reconstructs time-varying error maps and proactively configures optimal routes, demonstrably reducing estimation error by over 60% and lowering logical error rates by 30-35% compared to existing methods.

Syndrome-based optimisation sharply improves quantum network error map accuracy and logical

Estimation error reduced by over 60% when utilising SCOPE, compared to a standard expectation-maximisation baseline. This precision unlocks performance previously unattainable in quantum networks, where traditional methods struggled to accurately map network errors without interrupting service. SCOPE, a syndrome-based control plane, harvests naturally occurring data from quantum error correction, known as syndromes, to reconstruct time-varying error maps and proactively optimise routing and coding.

A lowered logical error rate of approximately 2 percent was achieved, contrasting with around 3 percent using the baseline estimation method. This approach enables joint optimisation, moving beyond reliance on static topology or physical link fidelity, and demonstrably lowers logical error rates by 30 to 35%, and in some instances up to 65%, against existing topology-aware baselines. SCOPE’s Transformer and Graph Neural Network architectures achieved a mean absolute percentage error (MAPE) of approximately 20 percent, a sharp contrast to the 60 percent MAPE of a standard expectation-maximisation baseline. Further analysis revealed that static, per-edge error estimation incurs roughly 50 percent higher MAPE than SCOPE’s full models, stressing the importance of accounting for path-dependent noise; partial optimisation, focusing solely on route or code selection, also failed to match the performance of SCOPE’s joint approach. Incremental fine-tuning of the system requires only one to five minutes, allowing for continuous adaptation to changing network conditions, though these results are currently based on simulations and do not yet demonstrate performance within a fully operational quantum network with real-world decoherence.

Syndrome telemetry enables substantial gains in quantum network error map reconstruction

Simulations indicate a substantial improvement in network visibility, with estimation error decreasing by over 60% relative to a standard Expectation-Maximisation baseline. SCOPE, or Syndrome-based COntrol PlanE, represents a new network-layer architecture that optimises quantum network performance by utilising passive telemetry from error syndromes generated during standard user service. It reconstructs the network’s error map without requiring disruptive active tomography, unlike conventional methods.

Currently, the demonstrations are limited to NetSquid and IBM-calibrated simulations, and validating these performance gains on actual quantum hardware remains a necessary step. The computational cost associated with its inference engine could also present a scalability challenge in larger, more complex networks. SCOPE directly addresses limitations in existing control planes which decouple routing decisions from Quantum Error Correction strategies, relying on less informative metrics such as network topology or scalar fidelity. By integrating routing and coding optimisation, the system aims to minimise the end-to-end logical error rate, a key metric for fault-tolerant quantum systems, building a time-varying map of network error biases by harvesting error syndromes, parity-check outcomes from QEC decoders, and then proactively configuring optimal route-and-code combinations at source nodes.

Real-time error map reconstruction enhances quantum network performance

A new network architecture, SCOPE, has been developed by scientists at Stony Brook University to optimise performance in quantum networks. The system jointly optimises routing and coding using passive telemetry data, and reconstructs a network’s error map without requiring active tomography, a process that disrupts service and reduces throughput. Existing quantum control planes often decouple routing decisions from Quantum Error Correction strategies, frequently relying on basic topology or fidelity metrics.

This approach allows for proactive adjustments to route and code configurations at source nodes, thereby improving overall network efficiency. NetSquid was used alongside IBM calibration to model the simulations, providing a platform for testing the new architecture. The team highlights a shift in focus from maximising physical link fidelity to minimising end-to-end Logical Error Rate as quantum networks mature. Establishing a proactive control plane represents a fundamental shift in how quantum networks are managed, moving beyond simply connecting qubits to actively mitigating errors during transmission. SCOPE achieves this by reconstructing a dynamic map of network errors using data naturally produced by quantum error correction, avoiding disruptive testing methods and enabling joint optimisation of both routing and coding to tailor data pathways, minimise the impact of noise, and improve the reliability of quantum information transfer.

SCOPE, a new network architecture, successfully reconstructs a time-varying map of network error biases using error syndromes generated during standard quantum error correction. This passive telemetry approach reduces estimation error by over 60% compared to existing methods, offering a significant improvement in network visibility. By jointly optimising routing and coding, the system minimises the end-to-end logical error rate, a crucial step towards building fault-tolerant quantum networks. The researchers demonstrated this functionality through NetSquid and IBM-calibrated simulations, paving the way for more reliable quantum information transfer.