There are three different formats in which the raw data is stored:
DDL Files
In this simple format the raw data belonging to one DDL is stored in a separate file each. The data of different events is stored in separate directories. The name of the directory is by convention raw
The raw data DDL files are binary files containing the DDL data header and the detector specific raw data. Therefore the content of a file is equivalent to the data sent over a DDL. The advantages of this format are its simplicity and the fact that this format is needed for the Computing Data Challenges. A disadvantage is the large number of files. Therefore this format is not suited for permanent storage.
DATE file
A DATE file contains the raw data of one or more events. In addition to the data stored in the DDL files there are event and equipment headers allowing to identify the data from different events and DDLs. The unique DDL ID is stored in the ID of the equipment header. Events for the start and the stop of a run may be added as well. The DATE format is described in the DATE manual.
The DATE format is equivalent to the format of the output of the DAQ system on the GDC (before rootification with alimdc). It is expected to be the format of recorded data for future Computing Data Challenges and for real data.
ROOT file
The ROOT format of raw data is produced by the alimdc program. This program is intended to run on the GDCs, get the raw data events from DATE and convert them to ROOT objects. The objects are stored in a tree with name "RAW" in the branch "rawevent". The tree has one entry per event. The stored objects are of type AliRawEvent.
The ROOT format is equivalent to the format of the output files produced in the Computing Data Challenges. But currently a DATE component is developed to replace the alimdc program.