Home » Use cases » Digitally sovereign IT for science

Digitally sovereign IT for science

FAU is implementing an SCS-compliant self-service cloud

Technical details:

A greenfield implementation was chosen for the project; this comprises:

SCS-compliant OpenStack environment
Highly available controller and compute architectures
Flexible provisioning of CPU, GPU and memory resources
SDN-based network configuration
Scalable architecture to meet future requirements (e.g. increased GPU capacity)

Current operations show that the platform is ideally suited to meeting dynamic and short-term resource requirements in research and teaching.

Produced by:

Created using:

Certified integrator

UPONU

Background

The Friedrich-Alexander-University Erlangen-Nuremberg (FAU) is one of Germany’s largest research universities. In order to significantly simplify access to computing resources for teaching, research and student projects, whilst at the same time reducing dependence on proprietary virtualisation solutions, FAU decided in 2024 to set up an SCS-compliant self-service cloud called „Compute Cloud“ .

Procedure

Together with the stackXperts GmbH a completely new OpenStack-based infrastructure was developed in just four months. The aim was to create a flexible, cost-efficient and maintainable alternative to the existing VMware environment that provides CPU and GPU resources without long application paths and is based on clear standards. Sovereign Cloud Stack was chosen because the modular software stack based on OSISM allows in-depth technical customisation and is supported by an active community and established partners.

Result

The new platform now gives the FAU community immediate access to compute and GPU performance - a key advantage, especially for AI and machine learning applications. GPU sharing, needs-based resource utilisation and the relief of the IT teams result in measurable efficiency gains. The introduction has been well received by users and makes daily use easier.