Skip to Content

Data set file system: A major turning point for IBM mainframes

Capgemini Engineering
27 Jun 2023
capgemini-engineering

Overview

The Multiple Virtual Storage (MVS) file system is the most commonly used file system for the IBM Z family of mainframes. However, MVS is not a hierarchical file system like today’s typical file systems such as Windows, Unix, macOS, and others. Most of the next-generation IT recruits are not used to working in the MVS environment. Other than MVS, IBM Z also has the Unix file system, which is hardly used compared to MVS. However, next-generation IT professionals are very familiar with Unix.

With z/OS2.5, IBM introduces a new file system type that will render traditional MVS data sets accessible by the z/OS UNIX namespace. As a result, Z2.5 will enable z/OS UNIX applications, tools, and utilities transparent access to data in these data sets securely and consistently.

IBM is calling this file system type Data Set File System (DSFS). It will map MVS data sets on a Unix namespace. IBM’s idea of accessing MVS data sets from the Unix address space makes MVS data sets visible and easily accessible to and from Unix commands, applications, tools, etc.

A turning point for IBM mainframes

Since the 1970s, businesses running on IBM Z mainframes have generally stored their data and application code on an MVS file system. As a result, it is one of the oldest and most widely used file systems for large and mid-sized businesses worldwide.

However, training recruits on MVS is a significant challenge for businesses running on IBM Z. Today, most universities and colleges don’t educate students on MVS file systems. It would take years of training for the current generation to make themselves familiar with MVS internals. Thus, there is a significant gap between the supply and demand of MVS-trained instructors in the market.

Suppose your new software recruits are not required to understand the MVS file system. Instead, they just need to work on the familiar Unix Bash shell environment to access your existing databases and codebases. That would be amazing, right?!

With z/OS 2.5, businesses don’t need to train recruits on MVS. They can even do lateral hiring based just on Unix skills. Most importantly, companies will not have to migrate their databases/codebases from mainframes because recruits lack the MVS skillset. Also, there will be no need to code separate code logic for accessing data from MVS and Unix file systems, which reduces program size and complexity. These are the kinds of benefits that DSFS can bring to mainframe-based businesses.

Figure 1 captures the technical details describing the use of DSFS to access MVS datasets on a Unix namespace.

Figure 1: How to use DSFS to access MVS datasets on a Unix namespace
Source: Capgemini Engineering

Figure 2 below shows how MVS datasets, residing on a direct-access storage device (DASD), can be accessed on the Unix namespace using DSFS. MVS dataset AAAA.BBBB.CCCC can be accessed as /dsfs/txt/aaaa/bbbb.cccc on Unix. Similarly, PPP.QQQ.RRR can be accessed as /dsfs/txt/ppp/qqq.rrr.

Figure 2: Data Set File System mapping
Source: Capgemini Engineering

DSFS brings value to the mainframe world in five important ways:

  • A mapping mechanism to access MVS-cataloged (DASD) datasets from Unix namespace
  • Reduction in the gap between demand and supply for MVS resources
  • Reduction in cost and time in training resources on MVS
  • Reduction in program code as single logic will be required to access both MVS datasets and Unix files
  • Increase the confidence of businesses in the IBM mainframe. Yes, mainframes are adapting to your business needs and mainframes can modernize and change to help solve your problems.
Capgemini Engineering

Capgemini and IBM – Product Partnerships

The partnership introduces product capabilities into new geographies, platform eco-systems and deployment models
Capgemini Engineering