Parts | Days | Selection | Search | Downloads | Help

HK: Fachverband Hadronen und Kerne

HK 30: Accelerators and Instrumentation I

HK 30.4: Talk

Tuesday, March 17, 2009, 15:00–15:15, H-ZO 80

The PANDA Grid — •Paul Bühler — Stefan Meyer Institute, Vienna, Austria

In order to fulfill the computing demands of the future PANDA experiment at FAIR in Darmstadt a dedicated computing infrastructure will be required. A conceivable way of acquiring and managing the necessary computing power for simulations and data analysis is the Grid model. As an alternative to a centralized computing center this model allows to pool independent resources from multiple institutes or organizations.

Although PANDA is not expected to acquire data before the year 2016 the PANDA collaboration is already experimenting with the PANDA Grid. The PANDA Grid uses the AliEn middleware which is entirely built around Open Source components and has been developed by the ALICE collaboration at CERN. The PANDA Grid currently consists of 10 sites. Due to the optimized installation procedures and portability of the AliEn software it is easy to add new resources, allowing the Grid to be expanded without disturbing its continuous operation. AliEn provides tools to pool hardware resources but also to manage the installation of common software packages. This feature is exploited to distribute, install, and update the PANDA analysis software (PandaRoot) on the different sites and with this to enable the PANDA Grid to perform PANDA related tasks. It is intended that PANDA Grid will provide in the order of 2-3000 CPUs by the end of 2008 and then constitute a powerful tool for upcoming PANDA detector design and physics case studies.