Call Us: US - +1 845 478 5244 | UK - +44 20 7193 7850 | AUS - +61 2 8005 4826

Fixed-size blocks allocation

To retrieve a data set, the system needs the data set name, the volume serial number, the device type, and in some instances, the data set sequence number. Specifying these can sometimes be inconvenient for the programmer. The catalog permits storage and retrieval of a data set based on name alone. The items mentioned previously are stored in the system catalog. This is a series of indexes, each corresponding to a qualification level of the d~ta set name. The number of levels is determined by the user. A data set name consists of a simple name preceded by qualifiers (index names) separated by periods. Each component of the full name can be up to 8 characters long; the entire name, including periods, can be up to 44 characters long. Each component name starts with an alphabetic character, and may contain any letter or number. Indexes, linked together in a hierarchy specified by the user, contain pointers to subordinate indexes, to volumes, or to both. Data Management 23 Find: Data S~t TREE. FRUIT. APPLE ‘ystem Resi d ence V o I ume Search —– Catalog (Directory) TREE (Index) ~* FRUIT @ TREE *Y ~ …… – Starts FRUIT ( Index) * APPLE 326) -…,. ./ Volume 326 VTOC (Index) ~r Volume 1326 1* Label * (DSCB) _ TREE. FRUIT. APPLE * ~– – * Data Set: TREE. FRUIT. APPLE I I Figure 11. Catalog Search Procedure The automatic cata’log search procedure is illustrated in Figure 11. The search starts in a volume designated when the system is generated, which contains the highest level index. The highest order qualifier is sought, which provides the location of the index for that qualifier. The search continues for the next level qualifier, and for its index. When a volume number is found, the catalog search is completed. If the required volume is not already mounted, a mounting message will be issued. If the data set is on a direct-access device, the search for the data set location resumes with the volume label of the indicated volume, continues to its volume table of contents, and from there to the data set’s starting location. All volume tables of contents are designed so that an “in channel” search can be used, i.e., a single search request allows the channel program to continue through the necessary series of seeks and reads until the named data set is located, without intermediate interruptions of other processing that may be in progress. The index design is left to the user, so that logical divisions between index levels can be arranged to meet installation needs. 24 A horizontal index structure with many different names and few index levels will require fewer seeks than will a vertical index structure with many levels and few · names in each level. The vertical struc- · ture is likely to give less conflict between duplicate names. The system will · not accept two identical names in the same index. Attempts to catalog identical fully qualified names will be rejected.