Skip to main content

History of locating a Dataset​

In z/OS (the mainframe operating system), datasets are used to store data like files on a computer. But how does the system know where a dataset is stored? Let’s explore.

What Do You Need to Find a Dataset?​

In earlier versions of z/OS, to find and access a dataset, the system needed three key pieces of information:

  1. Dataset Name – The name of the file or dataset.
  2. Volume Name – The name of the storage volume (like a disk) where the dataset is kept.
  3. Unit Type – The type of storage device (for example, a 3390 disk or a 3590 tape).

Without all three, the system couldn’t find or open the dataset.

Why Was This a Problem?​

In the past:

  • Each user was given a specific volume to use.
  • You could only store data on that volume.
  • If the volume became full, you'd have to request another one manually.

This worked when data was small and simple, but as companies started working with more data:

  • Managing volumes manually became difficult.
  • Users had to remember where everything was stored.
  • Finding datasets across many volumes became frustrating and time-consuming.

📘 The Role of the Catalog and VTOC​

To make things easier, z/OS introduced the Catalog system.

Think of a Catalog like a phone book for datasets. Instead of remembering the exact volume and unit, you can now just search by the dataset name.

What Is in the Catalog?​

  • It stores references, not the data itself.
  • These references include:
    • Dataset name
    • Which volume(s) the dataset is stored on (where to find it)
    • Attributes like creation date, record format, size info and permissions
  • The Catalog doesn’t hold the actual dataset — it just keeps metadata.

📌 Key point: Once a dataset is cataloged, you don’t need to know the volume or device type — you can simply use the dataset name, and the system finds it for you.

Cataloging in Modern Systems​

Today, most z/OS systems use something called System-Managed Storage (SMS) — a feature that automates storage tasks.

  • When a dataset is created on a DASD (Disk Storage) under SMS, it is automatically cataloged.
  • For magnetic tapes, cataloging is optional, but many companies still do it to make dataset access easier.

This automatic cataloging removes a lot of manual work and reduces the chance of errors.

Then Who Manages the Physical Layout? 👀​

Great question — this is where the VTOC comes in.

  • VTOC stands for Volume Table of Contents
  • Every volume (disk) has its own VTOC
  • It keeps track of:
    • The VTOC holds the physical details(physically starts and ends) of all datasets on that volume.
    • How much free space is left
    • This includes dataset names, their locations (actual disk addresses), sizes(Details about deleted or unused space), and other important attributes.

While the Catalog helps the system find the correct volume, the VTOC provides the exact physical location of the data on that volume.

📌 The VTOC has been around since before the Catalog, and it works behind the scenes — even if no catalog is being used.

📙 How They Work Together​

  • When you want to access a dataset (for example, "USER.DATA"), the Catalog tells the system which volume it’s stored on (like Volume 1).
  • The Catalog doesn’t tell you the exact disk location, only the volume.
  • The VTOC on Volume 1 contains the exact physical location of the dataset on that volume.
  • Once the system gets this physical location from the VTOC, it can retrieve the dataset.

In short:

  • The Catalog is a logical reference system — it helps z/OS find datasets by name and directs it to the right volume.
  • The VTOC provides the physical location of datasets on the volume, so the system can access the actual data.

Special Case: Dataset Across Volumes​

  • A dataset can be spread across multiple volumes if it is too large or needs more space than one volume can provide.
  • The Catalog keeps track of all the volumes where the dataset is stored but does not know the exact physical addresses on those volumes.
  • When accessing the dataset, the Catalog tells the system which volumes to look at.
  • Then, the VTOC on each volume provides the exact physical location of the data stored on that specific volume.

Special Case: Multiple Catalogs​

  • A z/OS system always has at least one master catalog.
  • If there’s only one catalog, that catalog is the master catalog and contains location entries for all datasets.
  • But using only one catalog is not efficient or flexible.
  • So, most systems use a master catalog together with multiple user catalogs.
  • These catalogs are connected, creating a system that is easier to manage and faster to search.

How Multiple Catalogs Work​

  • A user catalog stores the full details for datasets, including:

    • Dataset name (DSN)
    • Volume name
    • Unit (device type, like 3390 disk or 3590 tape)

  • The master catalog usually stores only the high-level qualifier (HLQ) of a dataset and points to the user catalog where the dataset info is stored.

  • The HLQ acts like an alias for the user catalog.

Example Scenario​

  • For a dataset like SYS1.A1, the master catalog directly returns the volume (e.g., WRK001) and unit (e.g., 3390).
  • For a dataset like IBMUSER.A1, the master catalog sends the request to the user catalog USERCAT.IBM.
  • The user catalog then returns the dataset’s location details, like volume and unit.

Refer IBM Docs - Multiple Catalogs for the diagram explaining the scenario.

Future Readings​