z/OS concepts
Previous topic | Next topic | Contents | Glossary | Contact z/OS | PDF

z/OS storage constructs: File systems, data sets, and more

z/OS concepts

In working with the z/OS® operating system, you must understand data sets, the files that contain programs and data. The characteristics of traditional z/OS data sets differ considerably from the file systems used in UNIX® and PC systems. To make matters even more interesting, you can also create UNIX file systems on z/OS, with the common characteristics of UNIX systems.

z/OS manages data by means of data sets. The term data set refers to a file that contains one or more records. A data set can be a source program, a library of programs, or a file of data records used by a processing program. Data set records are the basic unit of information used by a processing program.

Users must define the amount of space to be allocated for a data set (before it is used), or these allocations must be automated through the use of DFSMS™. With DFSMS, the z/OS system programmer or storage administrator can define performance goals and data availability requirements, create model data definitions for typical data sets, and automate data backup. DFSMS can automatically assign, based on installation policy, those services and data definition attributes to data sets when they are created. Other storage management-related products can be used to determine data placement, manage data backup, control space usage, and provide data security.

Almost all z/OS data processing is record-oriented. Byte-stream files are not present in traditional processing, although they are a standard part of z/OS UNIX. z/OS records and physical blocks follow one of several well-defined formats. Most data sets have DCB attributes that include the record format (RECFM), the maximum logical record length (LRECL), and the maximum block size (BLKSIZE).

z/OS libraries are known as partitioned data sets (PDS or PDSE) and contain members. Source programs, system and application control parameters, JCL, and executable modules are almost always contained in libraries.

Virtual storage access method (VSAM) is an access method that provides much more complex functions than other disk access methods. VSAM is primarily for applications and cannot be edited with ISPF.

z/OS data sets have names with a maximum of 44 uppercase characters, divided by periods into qualifiers with a maximum of 8 bytes per qualifier name. The high-level qualifier (HLQ) may be fixed by system security controls, but the rest of a data set name is assigned by the user. A number of conventions exist for these names.

An existing data set can be located when the data set name, volume, and device type are known. These requirements can be shortened to knowing only the data set name if the data set is cataloged. The system catalog is a single logical function, although its data may be spread across the master catalog and many user catalogs. In practice, almost all disk data sets are cataloged. One side effect of this is that all (cataloged) data sets must have unique names.

A file in the UNIX file system can be either a text file or a binary file. In a text file each line of text is separated by a newline delimiter. A binary file consists of sequences of binary words (byte stream), and no record concept other than the structure defined by an application exists. An application reading the file is responsible for interpreting the format of the data. z/OS treats an entire UNIX file system hierarchy as a collection of data sets. Each data set is a mountable file system.

Copyright IBM Corporation 1990, 2010