ALICE Calibration Requirements
FMD OK July 06 Completed by: Christian Holm Christensen <cholm@nbi.dk>, modified 2006-09-04
                             
Par # Parameter Data format/size per channel Data size (Total) Bytes Update freq Source Confirmed  Run type / Trigger type # of required events/sampling rate Processing level: sub-event or event Results: FEE/Archive Accessible by offline Calib. Procedure in AliRoot  use case #
      in FES in OCDB reference                    
1 Oversampling rate 4 Bytes/half ring   40 0 < Run DCS or Event(2) yes (2,3) all triggers (4) - - DCS ArchiveDB/OCDB yes! - 4
2 Strips readout 2 Bytes/half ring   20 0 < Run DCS or Event(2) yes (2,3) all triggers (4) - - DCS ArchiveDB/OCDB yes! - 4
3 Zero suppression 2 Bytes/half ring   20 0 < Run DCS or Event(2) yes (2,3) all triggers (4) - - DCS ArchiveDB/OCDB yes! - 4
4 Pedestals  Mean&width (4+4 bytes)/strip     4.10E+05 2.50E+08 2÷3 times/day  DAQ via pedestal runs yes (3) calibration/pedestal 1000 sub-event FEE- DAQ FES/OCDB yes! No (expected around October 2007) 1 or 3
5 Pulser gain (1) 4 Bytes/strip   2.05E+05 2.00E+09 < Day DAQ via pedestal runs yes (3,5) calibration/gain scan (5) ~1000 / strip / pulser value  -> ~ 128 * 1000 * 10  = 1280000 sub-event/sub-run(5) DAQ FES/OCDB yes! No (expected around November 2006) 1 or 3
6 Temperatures 2 – 6 Bytes/half ring   60 0 Run DCS yes (3) all triggers (4) - - DCS ArchiveDB/OCDB Yes - 4
7 Running conditions (6) 4-30 Bytes/half ring   300 0 Run DCS or Event(2) ? all triggers (4) - - DCS ArchiveDB/OCDB Yes - 4 or 5
6.15E+05 2.25E+09
FMD Notes:
1 Name changed to reflect actual contents Reference is the set data which are used to calculated the calibration parameters (ex. An histogram which mean value is a calibration parameter). These data can be optionally save in the OCDB for checks to be done offline. These data will not be available for reconstruction. 
2 Running conditions like oversampling, strips readout, zero suppression, etc. is normally set via DCS, and the DCS logbook will contain the set-points.  However, in Yves' document entitled Calibration from the 27th of June, 2006, it is stated that for the TPC, 'Parameters of the FEE required for reconstruction, such as tail cancellation, zero suppression, sampling frequency[,] and acquisition window status bits, will be written by the RCU into the event trailer.'  We do not currently know how to do this for the FMD, but if it is at all possible, we'd follow that strategy as closely as possible. Update: I've sent a mail to Luciano, but so far I have not heard anything. Confirmed: I expect that you have talked to the system (DCS/DAQ/HLT) experts and that they have confirmed that what you plan to do is feasible. 
3 Since I'm the one that know most about the FMD FEE these statements are not really objective :-)  I'm not really too familiar with the DCS part of the system, but we plan to follow the TPC development closely due to similar FEE. See also note (2) above  # of required events/sample rate': Right statistics required and/or the sample of data in a run you plan to process to extract the calibration parameters. 
4 This data is controlled by the DCS, and as such does not require processing to extract.  However, in certain cases, like the pulser gain calibrations, the DA may control these without the intervention of the DCS.  In that case, it's up to the DA to propagate the information properly via FES Results: FEE/Archive? Is that to be filled in with the place were the intermediate results are stored? Correct, this can be either the DCS ArchiveDB, the DAQ, DCS or HLT File Exchange server or in case you do not need these parameters offline the FEE to which you will load the calibration parameters through the DDL or DCS. 
5 Since the  pulser gain calibrations require that we loop through the strips and through pulser values, we have to have some sort of sub-run.  That can be done by letting the LDC control all aspects of the FEC.  However, that seems to imply that the detector must be software triggered.  Whether software triggers are really feasible is yet to be investigated.   Secondly, we have to test the FEC to see exactly how fast we can do this.  If we can do it fast enough, then we can use external triggers. Accessible b[y] offline? Is that a `yes/no' question, or is there more?  Yes or No. Yes means that the data will be published in the Offline Calibration Data Base (OCDB). 
6 This item really covers what in the TPC case is called 'Parameters of the FEE required for reconstruction ...' Hence, note (2) above applies here. I guess the column `use case #' refers to the use cases you presented some time ago, right? Correct 
DCS access - offline has access only to DCS archive DB, curently no access to configuration DB
DAQ - offline has access to raw data and DAQ logbook, but no acccess to LDC or GDC directly
OCDB - Calibration parameters are stored in the Offline Calibration DB. The data used to calculate these parameters, the reference data, can also be optionally archived for later access from offline.
Use case 1: online creation in DAQ
Calibration parameters are computed online in the DAQ LDC/GDC/Monitoring machines from physics or dedicated data
Results are made available as ROOT files in the DAQ FES
The path of these files together with the start and end of run timestamps is written in the DAQ Logbook
At the end of the run additional processing may occur controlled by the ECS
Upon notification by the ECS, the SHUTTLE queries the DAQ Logbook for file name and timestamps and fetches the appropriate parameter files 
It stores the files into the CERN storage and adds an entry (run validity and unique identifier of the files) into the AliEn FC
Use case 2: online creation in HLT
Calibration parameters are computed online in the DAQ LDC/GDC/Monitoring machines from physics or dedicated data
Results are made available as ROOT files in the DAQ FES
The path of these files together with the start and end of run timestamps is written in the DAQ Logbook
At the end of the run additional processing may occur controlled by the ECS
Upon notification by the ECS, the SHUTTLE queries the DAQ Logbook for file name and timestamps and fetches the appropriate parameter files 
It stores the files into the CERN storage and adds an entry (run validity and unique identifier of the files) into the AliEn FC
Use case 3: online creation in DCS
Calibration parameters are computed online in the DCS farm from dedicated data
Results are made available, as ROOT files, in the DCS FES
The path of the produced files together with the start and end of run timestamps is written in the DAQ Logbook or equivalent DB
At the end of the run additional processing may occur controlled by the ECS
Upon notification by the ECS, the SHUTTLE queries the DAQ Logbook or the equivalent DB for file name and timestamps and fetches the appropriate parameter files
It stores the files into the CERN storage and adds an entry (run validity and unique identifier of the files) into the AliEn FC
Use case 4: parameters monitored by DCS
DCS monitored parameters are stored in the DCS Archive DB
At the end of the run additional processing may occur controlled by the ECS
Upon notification by the ECS the SHUTTLE queries the DAQ Logbook for the start and end of run timestamps and then queries the Archive for the relevant parameters 
The SHUTTLE pre-processors evaluates the parameters, creating the ROOT files
It stores the files into the CERN storage and adds an entry (run validity and unique identifier of the files) into the AliEn FC
Use case 5: FEE config file
The FEE parameters are fed to the electronics either trough the DCS or through the DAQ DDL
DCS:
files are stored in the DCS Configuration DB and on the DCS file exchange server, the notification is done through the DCS Archive DB
DAQ:
files are stored directly in the DAQ FES and their name registered in the DAQ Logbook (back to Use Case 1)
At the end of the run additional processing may occur controlled by the ECS
Upon notification by the ECS the SHUTTLE queries the DCS Archive DB for timestamps and file names, collects the data and pre-processes them (ROOTification,…)
The SHUTTLE stores the files into the CERN storage and adds an entry (run validity and unique identifier of the files) into the AliEn FC