Why are flat files used for the database data files?
The database architecture, using flat files stored in a directory tree, provides for an extremely efficient database in which historical values are stored time-sequentially in 4-byte floating-point format. The number of files in each directory is calculated to minimize the time to find any single file. Once a file is located, the timestamp is used as an index into the flat file. This allows for very fast data retrieval since a single read can return one (1) value or hundred (100) values in about the same time. Also, the data files are cyclical so that once they are filled up, the oldest data is written over with new data. This reduces the database administration.
