Volume table of contents
In the IBM mainframe storage architecture, Volume Table Of Contents, or VTOC, is a data structure that provides a way of locating the data sets that reside on a particular disk volume. It is the functional equivalent of the root directory of a mass storage device (floppy, jump drive, hard disk, etc.) on a PC or minicomputer, e.g. / on Unix or Linux, C:\ on MS-DOS or Microsoft Windows, can reside within the first 64K tracks on the volume, and lists the names of each data set on the volume as well as size, location, and permissions. Additionally, it contains an entry for every area of contiguous free space on the volume. The third record on the first track of the first cylinder of any volume of DASD (i.e. disk pack) is known as the volume label and must contain a pointer to the location of the VTOC. A VTOC is added to a disk when it is initialized using the Device Support Facilities utility program, ICKDSF. VTOC was originally designed for removable disk packs.
To locate a data set, a program will generally interrogate a z/OS catalog to find the volume where the data set resides. Having found the correct volume, the VTOC is searched to find out where on the disk the data set is stored.
The VTOC comprises records known as Data Set Control Blocks, or DSCBs. There are ten types of DSCB.
| DSCB format type | Purpose | Notes |
|---|---|---|
| 0 | Empty entry | |
| 1 | Data set primary | Describes first three extents of a data set |
| 2 | ISAM data set | Describes ISAM data sets |
| 3 | Data set extension | Describes data set extents after the third |
| 4 | VTOC | Describes volume attributes |
| 5 | Free space | 26 extents on non-index volumes |
| 6 | Shared Cylinder Allocation | In OS/360, Format 6 DSCB is used for "Shared Cylinder Allocation", which enabled multiple datasets to be interleaved across a range of cylinders (such that, for example, the first half of each cylinder was allocated for data set A, and the second half for data set B). This was intended as a performance optimisation to reduce disk head movement when two or more datasets were expected to be used simultaneously. In current releases of z/OS, shared cylinder allocation format 6 DSCB is no longer supported.[1] |
| 7 | Free space | Extension of Format 4 DSCB |
| 8 | Data set primary | EAV version of Format 1 DSCB |
| 9 | Data set extension | EAV extension of Format 3 DSCB |
The first DSCB in the VTOC is always a format 4 DSCB which describes the VTOC itself[2] and attributes of the disk volume on which this VTOC resides. The second DSCB is always a format 5 DSCB which describes free space within the VTOC. Normally, the rest of the VTOC will contain format 0 DSCBs, which are empty entries, and format 1 or format 3 DSCBs, which describe the "extents" of data sets, giving their start address and end address[3] of up to 16 such "extents" on disk. The initial part of a data set is described by a format 1 DSCB. If necessary, format 3 DSCBs are used to describe further "extents" of the data set. When a data set is deleted, its format 1 DSCB is overwritten to become a format 0 DSCB, and the format 3 DSCB, if one exists, is similarly deleted.
Originally, a VTOC search was a sequential scan of the DSCBs, stopping when the correct format 1 DSCB was found or the end of the VTOC was reached. As disk volumes became larger, VTOC search became a bottleneck and so a VTOC index was added.
See also
- Data Control Block (DCB)
- Count Key Data (CKD)
References
- ↑ IBM Corporation, IBM System/360 Operating System: System Control Blocks, C28-6628-3, November 1968, page 143; IBM Corporation, IBM System/360 Operating System: Utilities, GC28-6586-11, June 1970, page 114
- ↑ The VTOC itself actually has a dataset name as the VTOC is, indeed, a dataset; the VTOC's dataset name is (44) X'04' characters, which, in later instances of the OS, has become a protected name, one which may not be specified by the programmer.
- ↑ Each "extent" is described by a starting CCHH and an ending CCHH, relative to the beginning of the disk, which is CCHH=x'00000000'.