The xsd's for the different messages can be found here.
- Client sends IdentifyPillarsForDeleteFileRequest message with FileID, and all pillars which have files for the bit repository collection reply with IdentifyPillarsForDeleteFileResponse messages
- Client sends DeleteFileRequest messages with FileID to the relevant identified pillars and the pillars can reply with DeleteFileProgressResponse
- Data will now be deleted on the specified pillars and DeleteFileFinalResponse messages sent (note that the SLA will probably specify a 'roll-back-period' during which the data will be in an 'under deletion state' and a roll/back can be done manually)
Note that delete can either be of
- all copies of the file
- or a copy on a specific pillar
The difference is made in the choice of which pillars to send the DeleteFileRequest