The total time is composed of :
 

• Effective CPU time : cumulated CPU time of successfull jobs after checking. The total CPU time (with CPU time of aborted or unsuccess jobs after checking) is 80 CPU years. The usefull CPU time is about 84%.
• Overhead time : cumulated time of the different grid steps before the running time. The overhead time is composed of submitted, ready, waiting and scheduled times.

The total transfer time is composed of :
 

• Input transfer time : cumulated time for the input database transfer from a SE in the WN of the job. The job calls the file with its Logical File Name. In case of problem, a grid FTP is used.
• Output transfer time : cumulated time for the output archive transfer from the WN of the job in several SE. First the job copies and registers the output file in a disk SE, then the job replicates the output file with its Logical File Name in a HPSS SE. In case of problem, the archive is saved directly with a grid FTP command, then registered and replicated from the UI after the end of the job.

The average transfer rate is calculated from the user values and can represent several transfer for a same data (for instance, a copy and registration, then a replication). Network tools were not used