OVERVIEW

The ADLS – Austrian DataLAB and Services project was set up to promote and facilitate collaborative work between universities, especially in the fields of data science and high performance computing (HPC). It aims to provide easy access to a nationwide network of applications in a federal opt-in model. In particular, the focus lies on increasing user-friendliness and saving time for researchers, teachers and students in the course of using computing resources.

To this end, users should be able to discover via self-service:

  • run their own, possibly computationally intensive applications
  • access existing applications
  • without having to set up or configure the required environment, respectively.

The project is concerned with the extension and integration of digital services (applications that enable the processing and analysis of different types of data) and the provision of templates for future digital services and platforms – building on existing infrastructure. Access to these services will be provided through graphical user interfaces and interactive platforms. In the background, these applications access federated resources for processing and storing data.
The aim of the project is to make these services available to a broad group of users, which is why different application areas and use cases have been defined, which will be reflected in the implementation.

PRODUCT VISION

The key focus of the provided products lies on usability and ease of use.

Researchers from various different fields are enabled in inter-university data sharing processes. It is necessary to make research data reproducible and be able to use professional and highly potent resources for that matter, without having to install anything on their own device or waste time setting up the needed environment. Since time is a very scarce resource, the solution offered is intended to help save this precious good in a research but also teaching context. To make this possible a dedicated focus lies on fully automated deployment of the respective user applications. How this is intended to be provided can be found out in the section Why we believe in this vision? .

Application Capability  Focus
Discovery
  • Search and find existing applications and data
  • Search available (HPC and HTC) resources – explore what fits your needs best
  • Try out sample applications, tools and data
Development
  • Develop applications and publish them
  • Collaborate with others
  • Test the code and application – access to online tooling for development, testing and governance
Deployment
  • Easily deploy applications
  • Run the applications – without the need of local configurations
  • Run a variety of jobs

WHY WE BELIEVE IN THIS VISION?

Interviews and early feedback have shown, that users want to be able to implement their own applications and execute them, without having to worry about the underlying layers of infrastructure and network. A central login via a self-service user portal should enable overview and management of each users access to various services. Furthermore this portal will offer integrated and pre-configured tooling to create applications, expand them or access existing ones without having to install anything on their own PC, and thus helps to save valuable time. The portal is connected to the respective back-end via a standardised interface. The Back-End can either be the respective high-performance cluster or other computer resources that the respective universities have available. The data is held by the respective university, but can be easily shared with other universities via specific access rights, if required. The planned solution will not be a centrally managed infrastructure, but rather universities will be able to provide it themselves as needed via a simple application interface. This approach is called IaC – infrastructure as code. This makes it easy for users to work independently and not be dependent on a few experts. Users still don’t have to worry about the underlying architecture. The federated data management makes it possible to keep control over the data and at the same time have the possibility to share it with other universities and computing resources in a simple and user-friendly way.

The following graphic shows the various envisioned components.