You would like clarification on two components related to file size handling in Information Centric Analytics (ICA):
Release : 6.x
ICA copies MessageSize values from DLP (which is in bytes), and stores these values in the RiskFabric relational database verbatim without rounding or conversion. File size in ICA is the same as file size in DLP - same unit of measure, same values.
When using the DIM Advanced Search, ICA queries the actual file size value (that is, the MessageSize value) stored in the FileSize column of the table LDW_DIMIncidentsToFiles. ICA does not apply rounding rules, thresholds, or unit conversions.
ICA computes the file size scale when processing file size values into the OLAP cube (that is, the Analyzer) as follows:
CASE
WHEN d2f.FileSize = 0 THEN 'Empty (0 KB)'
WHEN d2f.FileSize BETWEEN 0 AND 16000 THEN 'Tiny (0-16 KB)'
WHEN d2f.FileSize between 16000 AND 1000000 THEN 'Small (16-1 MB)'
WHEN d2f.FileSize/1000000 between 1 AND 128 THEN 'Medium (1-128 MB)'
WHEN d2f.FileSize/1000000 between 128 AND 1000 THEN 'Large (128-1 GB)'
WHEN d2f.FileSize/1000000/1000 between 1 AND 4 THEN 'Huge (1-4 GB)'
WHEN d2f.FileSize/1000000/1000 > 4 THEN 'Gigantic (>4 GB)'
ELSE NULL
Note that ICA provides both the measure DIM Incident File Size and the dimension DIM Incident File Size in the Analyzer. The former is an aggregation (summation) of file sizes, while the latter groups file sizes according to the scale above.