Search for a command to run...
Athena is the common framework used by the ATLAS experiment for simulation,reconstruction,and analysis,By design,Athena supports multiple persistence services,and insulates users from technology-specific persistence details.Athena users and even most Athena package developers should neither know nor care whether data come from the grid or from local filesystems.nor whether data reside in object databases,in ROOT or ZEBRA files,or in ASCII files.In this paper we describe how Athena applications may transparently take advantage of emerging services provided by grid software today-how data generated by Athea jobs are registered in grid replica catalogs and other collection management services,and the means by which input data are identified and located in a grid-aware collection management environment.We outline an evolutionary path toward incorporation of grid-based virtual data services,whereby locating data may be replaced by locating a recipe according to which that dta may be generated.Several implementation scenarios,ranging from lowlevel grid catalog services(e.g.,from Globus)through higher-level services such as the Grid Data Management Pilot (under development as part of the European DataGrid porject,in collaboration,with the Particle Physics Data Grid)to more conventional database services,and a common architecture to support these various scenarios,are also described.