Architecture Data Staging Clause Samples

Architecture Data Staging. EUDAT’s Data Staging service can be built on the replication service in so far as it makes use of the data that is stored in the EUDAT Data Domain and it allows moving result data back into the EUDAT Data Domain. EUDAT Domain HPC Workspace HPC Domain Operating System Fast File System With the help of an interface a selection of data objects in the EUDAT data domain can be made and then be transferred into the HPC Workspace, which is administered either by PRACE or any other organization. Standard technology such as XSEDE, GridFTP, etc. is being used to carry out the transfer. In the same way the results of the computations will be stored first in the workspace and then be transferred into the EUDAT store. At that moment PIDs need to be registered to maintain the EUDAT data domain as a domain of registered data objects. Also metadata should be generated for these result files including provenance information. In the diagram above this architecture with independent technologies is indicated. For the transfer of data between the two domains a few technologies are available and can be used. The new XSEDE technology is probably the most flexible and thus preferred one. Due to the independence of the data organizations iRODS for example does not play any role for this service and as indicated another PID service could be used.