next up previous contents
Next: COMPASS Data Rates to Up: ANSWERS TO THE QUESTIONS Previous: Support by CERN


Questions of the Referees:


Using the charm spectroscopy setup, COMPASS will write about 50 DLT-7000 tapes (30 GB each) per day, over a period of 100 days or more. Could you describe:

  1. How the calibration of the various detector components will be done (for each tape? for every n'th tape?)? Is "online calibration" (like NA48) needed?
  2. Do you foresee central data recording, or local mounting of tapes?
  3. Will the processing farm be at CERN?
  4. If so, will the data be stored at the farm or centrally?
  5. How the tapes will be processed (how many processors, tape stations, operators)?
  6. If the processing is done at CERN: what will be the load on the CERN network? Will there be any central facilities used? If yes, which ones?
  7. With respect to the WA89 or WA102 reconstruction speed, a factor of about 1000 has to be gained. How will that be achieved?


In the following we discuss possible scenarios for the offline data handling at the COMPASS experiment. Data rates and event complexity in the hadron program of COMPASS will be slightly higher than in the muon program. The benchmarks for the capacity of the COMPASS offline data processing will therefore be set by the hadron program. All estimates below are given on this basis.

Lars Schmitt
Wed May 22 16:44:09 METDST 1996