CDF

1
CDF GRID computing UK collaborators Valeria
Bartsch (UCL), Rick St. Denis (University of
Glasgow), Todd Huffman Stefan Stonjek (Oxford
University), Miichael George SineadFarrington
(University of Liverpool)
CDF SAM DCAF introduction

• SAM is a metadata storage of the DØ, CDF and
  Minos experiments
• SAM access to sets of files based on metadata,
• automatically copies files keeps track of all
  files

• SAM working principle
• central database keeps all meta data and all file
  location information
• the SAM system communicates with CORBA
• a cluster is called SAM station
• SAM stations contact the central database via a
  CORBA-to-SQL server
• every SAM station has a local cache to store
  files, these are copied and tracked

• SAM and CAF (batch systems)
• CAF is CDFs batch system implementation for
  large compute farms (currently 2.6M SPECInt2k at
  FNAL and 2.6M SPECInt2k offsite)
• CAF has been decentralized (DCAF) using SAM for
  data handling and kerberos for authentication
• using the Condor Glide mechanism the CAF
  mechanism can run on LCG clusters

CDF SAM DCAF in numbers
Number of jobs on the CAF
Total CDF Files To User

• General numbers
• up to now about more than 3PB of data delivered
  to clients with SAM to the SAM stations
• CAF used for various analysis and Monte Carlo
  production jobs on daily basis, SAM data handling
  officially deployed everywhere
• dCache managed peak of 200.000 read transfers a
  day

dCache storage system