Practical example and current challenges
- With the progression of digitisation, the amount of data available is also drastically on the rise. The analysis of this data, not least with the help of AI and simulations, makes it possible to develop climate and transport models and other highly complex scenarios requiring large volumes of data to be analysed, e.g. in the financial sector (e.g. fraud detection), healthcare (genome analysis), or in the industrial sector (optimising supply chains).
- These are all calculations that are only possible if enormous computing capacity is available at short notice – the kind of capacity provided by high performance computers (HPCs). One technology that is emerging in this field is quantum computing (currently at the stage of quantum simulation), which promises a giant leap in computing performance for certain problems to be solved.
- Whilst HPCs and quantum computing are opening up new possibilities in many different areas, many of their potential users are unable to invest in this kind of system as they would not be able to use all of its capacity and also due to fast innovation cycles that require systems to be replaced very quickly.
- A feasible option, however, would be to have several users share a system via an infrastructure and service model guaranteeing a high level of (data) security, controlling the necessary data input and output, and strictly separating different services. There are some initial approaches testing such sharing models (shared super computers) in research, but these capacities are hardly accessible to industrial users and there is no commercial operational model.
What added value does the "GAIA-X project" offer?
- The project can deliver broad-based access to HPCs and quantum capabilities ‘as a service’. This means that HPCs for AI, modelling and simulation will become accessible to new users (those not working in the field of public-sector research), and that these can securely store their results. Updates and maintenance work are taken care of by the provider.
- Greater use and a higher utilisation rate mean that HPC servers can be operated at lower prices.
- The project can form the basis for horizontal, secured data markets and for an ideal environment for operating HPCs in a way that always allows users to remain in control of their data.
- Klaus Ottradovetz – Atos