gms | German Medical Science

The Data Integration Centers (DIC), established since 2018 as part of the Medical Informatics Initiative (MII), aim to make clinical routine data from patient care available and usable for medical research projects [1], [2]. The DIC are supported and funded as infrastructure by the Network University Medicine (NUM) since beginning of 2023.

This workshop will be conducted by an interdisciplinary team of the Data Integration Center at the University Hospital Carl Gustav Carus Dresden together with the NUM-DIC research team at the Technische Universität Dresden. In this workshop we present insights on how we build sustainable processes for the provision of infrastructure components as a core foundation in our DIC, with focus on automation, resource management, security and scalability.

We talk about our lessons learned and best practices based on two use cases of infrastructure services:

to join international research projects within the OHDSI community [3], [4].

to enable non-university sites with limited resources (e.g. people, time, money, knowhow) providing infrastructure and data harmonization processes as a service.

We begin by exploring virtualization at the lowest levels, starting with our Ceph cluster and operating system virtualization based on Proxmox. In addition, we talk about important technologies such as Docker and experimental Kubernetes, show their central role in containerization and orchestration and discuss implementation challenges at our DIC. Ansible is the core unit of our automation toolbox, with a git repository as the single point of truth for our infrastructure as a code approach and thus it provides a flexible and scalable automation framework that can help simplify and accelerate various aspects of IT operations, from configuration management to application deployment and beyond. Additionally, we share the experience from the perspective of a service provider and consumers such as software development teams.

As the DIC in Dresden is considered a part of critical infrastructure (KRITIS) and is subject to specific regulatory requirements mandated by law, it is crucial to ensure compliance with these regulations, such as employing the latest state-of-the-art software versions [5]. Therefore, we will touch up on this topic as well.

In each phase of the workshop, participants get demonstrated how the aspects of security, scalability and automation interact in the context of an infrastructure request and how different actors work together. Use cases are used to demonstrate a holistic approach that is necessary for the development of a stable research data infrastructure.

Together, we will discuss a path forward to a future in which an automated research data infrastructure enables researchers to exploit the full potential of medical data.

This is particularly relevant because the demands for machine and deep learning are constantly increasing and require operational excellence, a secure environment, scalable deployments and efficient maintenance to enable progress in medical research and thereby provide added value for patients.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.