PRACE selects an appropriate legal form for the permanent Research Infrastructure, defines its statute, and its governance in relation with the partners, the European Commission and the scientific users. The targets are signature ready contracts. The project establishes funding strategies and usage models. PRACE defines a peer review process to support leading edge science and optimal use of the resources ensuring open, fair and unbiased access to the Research Infrastructure. The project defines a consistent operational model across the distributed tier-0 sites. The PRACE project is managed using principles suitable for the permanent infrastructure. The project focuses on dissemination of project achievements, establishing and maintaining links with selected industries both as users of capability systems, technology providers, and creators of new technologies for HPC, and training of potential users.

• Management of the contract
• Organisational concept of Research Infrastructure
• Dissemination, Outreach and Training

PRACE performs accompanying technical work to prepare for the production phase. It provides the tools for a consistent management of the tier-0 systems and for the integration of the tier-0 infrastructure into the European HPC ecosystem. The project deploys prototypes of leadership class systems at selected sites that are likely to become production level systems in 2009/2010.  Selected applications are ported, optimised and petascaled to ready them for production on the tier-0 systems. Benchmark suites are created. PRACE defines a consistent procurement strategy, associated technical specification, and selection criteria for the current and future generations of HPC services, and the requirements for the physical infrastructure to host tier-0 systems. The project starts a permanent process of technology evaluation to transform user requirements into specifications for future leadership class systems. This process is designed to last throughout the lifetime of the Research Infrastructure; it is carried out in close relations with vendors of components, systems, and software.

• Distributed system management
• Deployment of prototype systems
• Software enabling for Petaflop/s systems
• Petaflop/s Systems for 2009/2010
• Future Petaflop/s computer technologies beyond 2010