This describes the format of the RACF® database.

Each RACF database is a non-VSAM single extent data set that is made up of 4KB blocks and must be cataloged.

Figure 1 illustrates the format of the database.

Figure 1. Format of the RACF Database

Relative byte addresses (RBA) up to the first BAM block are identified by their position. The need for more BAM blocks is driven by the size of the database. See BAM block header for more information.

Any RBAs that are after the BAMs consist of unassigned blocks, index blocks, and data blocks. A profile's segment data may begin at any 256-byte slot with a 4K data block and then continue in consecutive slots, through consecutive RBAs. Therefore, it is possible that a field definition (which is what constitutes the contents of a profiles segment data) starts in one RBA, and completes upon another, possibly spanning many RBAs. The field definition might then deposit a byte of any value in the first byte of the first slot of an RBA.

If initially it might seem a false positive for an index block (X'8A'), a forceful approach read the RACF database might not be able to determine the RBA type by just checking the first byte of the RBA. The index structure and the sequence set provide the information to find a profiles segment data, in its assigned 256-byte slot with a data block. The logical length of the profile indicates the number of consecutive 256-byte slots that are needed.