- The paper presents OQTOPUS, a comprehensive open-source stack that integrates server-side transpilation, execution, multi-programming, error mitigation, estimation, and circuit composition.
- It details a modular architecture that offloads complex quantum processing tasks to the backend, ensuring uninterrupted hybrid execution and optimized circuit performance.
- The system’s scalable, cloud-based design and adherence to open standards pave the way for reproducible quantum research and industry adoption.
A Practical Open-Source Software Stack for a Cloud-Based Quantum Computing System
Introduction
The paper presents OQTOPUS (Open Quantum Toolchain for OPerators and USers), a comprehensive open-source software stack designed to facilitate cloud-based quantum computing. The motivation stems from the lack of transparency and open-source solutions in the critical backend layers of quantum computing systems, which impedes standardization and broad adoption. OQTOPUS addresses this by providing a full-stack, modular, and extensible platform that covers the entire operational pipeline, from user-facing interfaces to low-level backend integration with quantum hardware.
Key Features and System Architecture
OQTOPUS distinguishes itself by implementing six essential features—server-side transpiler, server-side execution, multi-programming, error mitigation, estimation, and a circuit composer—directly in the backend layer, which is typically proprietary in other platforms. This design choice enables efficient, reliable, and scalable quantum computation while exposing all functionalities to the user and operator communities.
Server-Side Transpiler
The server-side transpiler optimizes user-submitted quantum circuits for the target hardware, offloading the computational burden from the client and ensuring compatibility with device-specific constraints. OQTOPUS leverages the Tranqu framework, supporting multiple transpilers (e.g., Qiskit, ouqu-tp), allowing users to select the most efficient circuit transformation for their application. This multi-transpiler approach is critical for benchmarking and optimizing circuit depth, gate count, and execution fidelity.
Server-Side Execution
OQTOPUS supports quantum-classical hybrid algorithms by enabling users to submit Python programs that are executed server-side, maintaining exclusive access to the quantum chip for the duration of the job. This contrasts with other platforms (e.g., IBM Quantum, Amazon Braket), where hybrid jobs may be preempted or interrupted. The OQTOPUS approach guarantees uninterrupted execution, which is essential for iterative algorithms such as VQE and QAOA.
Multi-Programming
The multi-programming feature allows multiple quantum circuits to be combined and executed in parallel on a single quantum chip, increasing hardware utilization and throughput. The current implementation requires users to specify the circuits to be combined, but future work aims to automate this process across multiple users. This is particularly relevant for NISQ-era devices, where maximizing qubit usage is critical for cost-effective operation.
Error Mitigation
OQTOPUS implements a server-side error mitigation module based on Qiskit's tensor product readout error mitigator, but with enhancements for real-time calibration and stability. By integrating error mitigation at the backend, the system can dynamically adapt to device noise characteristics, improving the reliability of measurement outcomes without requiring user intervention or environment-specific dependencies.
Estimation
The estimation module computes expectation values of Hamiltonians by decomposing them into Pauli operators and aggregating measurement results. This is implemented as a server-side service, allowing seamless integration with error mitigation and transpilation. The design ensures that users can perform variational algorithms efficiently, with all classical post-processing handled transparently.
Composer
The circuit composer provides both a visual and text-based interface for designing quantum circuits, supporting OpenQASM 3 as the intermediate representation. This lowers the barrier for new users and supports advanced workflows for experienced researchers.
System Layers and Implementation
OQTOPUS is architected as a three-layer system:
- Frontend Layer: Provides user interfaces and programming libraries (QURI Parts OQTOPUS), supporting Python and OpenQASM 3. The frontend is decoupled from backend dependencies, ensuring stability and reproducibility.
- Cloud Layer: Built on AWS serverless infrastructure, this layer manages authentication, job queuing, and API exposure via OpenAPI specifications. It is designed for scalability, security, and interoperability with third-party quantum devices.
- Backend Layer: Deployed in proximity to quantum hardware, this layer executes jobs, manages device calibration (QDash), and provides administrative tools (OQTOPUS Admin). The backend integrates with control software and simulators (e.g., Qulacs) via the Device Gateway, supporting modular expansion and hardware abstraction.
The microservice architecture, with gRPC-based communication, enables independent scaling and maintenance of each component. The adoption of open standards and widely used technologies (e.g., Python, AWS, OpenQASM) facilitates community contributions and integration with existing quantum software ecosystems.
Operational Considerations and DevOps for Quantum Systems
The paper introduces the concept of QCOps, an adaptation of DevOps/MLOps principles for quantum computing. This encompasses tools and practices for calibration (QDash), user and device management (OQTOPUS Admin), and systematic operation of quantum hardware. By formalizing operational workflows, OQTOPUS reduces reliance on individual expertise and enhances reproducibility, which is critical for scaling quantum computing services.
Experimental Results and Use Cases
OQTOPUS has been deployed on a superconducting quantum computer at Osaka University, demonstrating end-to-end execution of sampling and estimation jobs via the QURI Parts OQTOPUS API. The system provides clear, interpretable results through its frontend UI, confirming the practical viability of the stack. The integration of transpilation, error mitigation, and estimation in the backend layer enables users to achieve reliable results with minimal setup, requiring only the quantum hardware and pulse sequencer.
Implications and Future Directions
OQTOPUS represents a significant step toward democratizing access to quantum computing infrastructure by making all critical software components open-source and publicly available. This approach is expected to accelerate standardization, foster community-driven development, and lower the barrier for new entrants—both academic and industrial—to deploy and operate quantum cloud services.
The explicit implementation of backend features, which are typically proprietary, sets a precedent for transparency and reproducibility in quantum computing research and development. The modular, extensible architecture positions OQTOPUS as a reference platform for future quantum cloud systems, enabling rapid integration of new hardware, algorithms, and operational paradigms.
Future work will focus on open-sourcing additional system components, particularly quantum control and monitoring tools, and automating multi-programming across users. The platform's design anticipates the evolution of quantum hardware and the increasing complexity of hybrid quantum-classical workflows.
Conclusion
OQTOPUS provides a comprehensive, open-source software stack for cloud-based quantum computing, addressing critical gaps in transparency, standardization, and operational efficiency. By exposing all essential features—including those in the backend layer—and supporting robust DevOps practices, OQTOPUS enables scalable, reproducible, and community-driven quantum computing services. The system's deployment on real hardware demonstrates its practical utility, and its open architecture lays the groundwork for future advances in quantum cloud infrastructure.
