Hybrid Quantum Computing

Today’s quantum algorithms can be characterized as hybrid. Certain operations (especially pre- and post-processing) are performed on classical computers, while specific computationally intensive operations are performed on quantum computers. The latter can realize certain advantages (such as exponential storage capacity in the QML domain or speed-up of algorithmic runtime) over classical computers through quantum mechanical principles (such as superposition or entanglement). For more on this, see, for example Revythi, M.; Koukiou, G. Quantum Machine Learning and Deep Learning: Fundamentals, Algorithms, Techniques, and Real-World Applications. However, it should be noted that advantages often only exist in theory at first.

Nevertheless, further new applications for quantum computers are currently being researched. We are investigating applications and problems from the database context and the ErUM community (QC4ErUM BMFTR project). In this context, the general question arises as to whether certain problems should be calculated on quantum hardware at all and on which quantum hardware specific problems can best be calculated (QC-Adviser project).

The commonality among various use cases is often that the input data or problem instances to be processed on quantum hardware are organized in different databases. However, quantum computers cannot access this data directly and require the data to be processed to be encoded in a suitable way.

The project pursues two main objectives:

• Implementation of a hybrid system that enables the exchange of data between applications, databases and quantum computers. It consists of classic components and quantum hardware and must be able to manage data flows.
• The input data or problem instances available in different models must be encoded accordingly before they can be processed on a quantum computer. The goal is to develop suitable and efficient encodings.

Our solution approach: We design and implement a hybrid system, the Hybrid Data Management Architecture (HDMA). We opted for a decentralized approach with a lightweight, asynchronous communication mechanism by adopting the microservice paradigm. In this architecture, databases can serve as data sources for quantum algorithms. This thus serves as a framework for researching and testing suitable encodings. The basic idea is summarized in the following figure:

The architecture consists of a series of services. They enable an application to encode data from a database, initiate a calculation on the encoded data on a quantum computer and return the result of the calculation.

In a nutshell, the reference workflow for GCP: A graph (A) to be colored is retrieved from the database. To color the graph, a circuit of the Grover algorithm is generated with a problem-specific oracle. (B) shows a small segment of the circuit and encodes graph (A). After executing the circuit on a quantum computer, a solution instance (C) is stored in the database.

Publications

• Zajac, M., Restat, V., & Störl, U. (2024). Quantum versus Classical Computation: Automatic Decision-Making Approaches. In INFORMATIK 2024 (pp. 573-577). Gesellschaft für Informatik eV.
• Zajac, M., & Störl, U. (2024). A Microservice Data Management Architecture for Quantum Computing. Poster at SummerSoC 2024.
• Zajac, M., & Störl, U. (2023, November). Hybrid data management architecture for present quantum computing. In International Conference on Service-Oriented Computing (pp. 174-184). Singapore: Springer Nature Singapore.
• Zajac, M. (2023, August). Encoding and provisioning data in different data models for quantum computing. In Conference on Very Large Data Bases (VLDB 2023).
• Zajac, M., & Störl, U. (2022, July). Towards quantum-based search for industrial data-driven services. In 2022 IEEE International Conference on Quantum Software (QSW) (pp. 38-40). IEEE.

Student theses

• Weiterentwicklung der Hybrid Data Management Architecture zur plattformübergreifenden Nutzung von Quantencomputing Cloudservices und Simulatoren (M. Rehburg, October 2025)
• Encoding classical data to address production and logistics problems with gate-based quantum computers (T. Haschke, September 2023)
• Linking data from different NoSQL document stores using Grover’s quantum search algorithm: Towards a quantum-classical hybrid approach (M. Dang, August 2022)