Active Reset and Scheduling in Fault-Tolerant Quantum-Classical Cloud Computing

Abstract

Quantum-classical cloud computing refers to a hybrid computing model that integrates quantum computing with classical computing resources, typically hosted on the cloud. The objective of this paper is threefold: First, to scrutinize the computing paradigm of quantum-classical cloud computing; second, to study the three pillars of the underpinning techniques: Active reset, scheduling, and fault tolerance; third, to demonstrate the realization of the aforementioned techniques. The active reset method leverages measurement and feedback to reset the qubit much faster and with high fidelity. The proposed scheduling algorithm is designed to efficiently manage quantum job execution in a fault-tolerant quantum-classical cloud computing environment. The paper also implements a Bayesian optimization process to optimize the hyper-parameters of a quantum error correction scheme. The research demonstrates significant performance gains across all three technical pillars of the hybrid cloud model. Specifically, the implementation of active reset reduced the qubit initialization time by 90 %, dropping from the 100 µs required for passive relaxation to just 10 µs while maintaining a high reset fidelity of 99.6 %. In terms of cloud resource management, the Shortest Job Next (SJN) scheduling algorithm drastically improved efficiency compared to FCFS, reducing peak wait times from 1.75 s to near-zero for many tasks. Furthermore, the application of Bayesian optimization to faulttolerant parameters successfully minimized the logical error rate from an initial 0.238 to 0.106, marking a 55 % improvement in the reliability of the quantum-classical computing environment.