IT-INFRASTRUCTURE

LEARN MORE ABOUT OUR POWERFUL IT-RESOURCES

The infrastructure of the DSC was designed for the implementation of computation-intensive data analyses in research, teaching, and training. Our infrastructure is based on the three pillars Virtualization, Storage, and High-Performance Computing (HPC), providing virtual machines, storage capacity, and computing resources to execute advanced algorithms especially in the field of artificial intelligence. Moreover, our system is optimized for scalability and is therefore expandable. This allows us to react flexibly to the diverse needs and requirements of our members and cooperation partners.

ARE YOU INTERESTED?
If you would like to use our computational resources in your research projects or lectures, please send us an E-MAIL to discuss the possibilities.

© Data Science Center

WHAT FEATURES DOES THE DSC IT-INFRASTRUCTURE OFFER?



VIRTUALIZATION

Through virtualization we are able to offer customized resources depending on the user’s needs. Thus, we can provide virtual machines with different operating systems and services. Databases and applications for demonstrators, projects or seminars can be implemented according to individual demands. The virtual machines can be turned on and off as required in order to ensure an efficient use of resources.

STORAGE

Data Science requires large amounts of data and therefore also storage capacities. We provide flexible storage space for virtual machines and high-performance computing. The storage is designed for operational use and not for long-term archiving. Our storage server currently has a capacity of 5 TB, which can be expanded if needed.

HIGH-PERFORMANCE COMPUTING

Some data science and in particular machine learning methods require a significant amount of computing time. Our high-performance hardware is designed to accelerate these computations for instance by using graphics cards. Our system is based on the data centers from the TOP500 and the NVIDIA AI systems. We currently have three NVIDIA V100 installed, which were engineered to provide maximum performance for AI and HPC workloads.

IT-INFRASTRUCTURE BY THE NUMBERS

4950

Storage
Capacity
[GB]

590

Memory
[GB]

1

Physical
CPU cores

1

NVIDIA V100
GPU

10

Single-Precision Performance [TFLOPS]

50

Network
Bandwith
[Gbits/s]