At the end of 2021, I had given an overview of the (still high) relevance of the mainframe in today’s IT production: The Future of the Mainframe (Part I), and some time later pointed out some (mainly German) podcasts about the mainframe that are worth listening to: The Future of the Mainframe (Part II): Listening Tips for Mainframe Podcasts

This article is about a current and global study by the IT infrastructure service provider Kyndryl; The study is based on a survey of 500 IT decision-makers and explores how companies are approaching mainframe modernization.

FREE DOWNLOAD: Kyndryl’s 2023 State of Mainframe Modernization Survey Report (12 pages)

2023 State of Mainframe Modernization Survey Report – Executive Summary

Fundamentally, the survey identifies three (quite obvious) approaches to IT strategy: First, the modernization of the mainframe environment. Second, integration with other platforms (i.e., mainframe applications, the underlying data on the mainframe, and the infrastructure are connected to other platforms, especially the cloud). Third, the complete shutdown of mainframe operations or the complete migration to a cloud environment. Incidentally, the latter option plays only a minor role, with only 2 (out of a total of 500 companies surveyed) opting for this option.

The vast majority continue to see the mainframe as a key function in their IT production. In particular, the following advantages are particularly appreciated: safety (68%), reliability (60%) and performance (55%).

The strategies described in dealing with the mainframe pursue different goals. For example, cost savings: On average, the survey shows that companies expect to save around 25 million US dollars per year on average, which can be achieved with the modernization strategy. This is usually due to a shift of IT workload to other platforms, which plays a role for almost all companies: on average, it is about 37% of the workload.

Other goals of the modernization approaches include: access to data (51% of the companies surveyed), improved innovation dynamics (48%), more flexibility (41%), accelerated time-to-market.

Mainframe Modernization: Common Approaches & Benefits

Why is “mainframe modernization” the main focus, what understanding does the KYNDRYL study have?

The following graph illustrates the focus of mainframe modernization projects and their frequency:

Let’s start by saying that application modernization isn’t necessarily a matter of mainframes or the cloud. Rather, it’s about the legacy applications running on it, which are characterized by monolithic architectures, proprietary standards and protocols, and older programming languages for which there are fewer and fewer experienced professionals. We’re talking about experts who are well-versed in legacy languages such as PL/1 (Programming Language ONE), COBOL, Natural, Ada, Assembler, and many others.

Translation, ReFactoring or ReArchitecting are obvious but very resource-intensive approaches to dealing with legacy applications in this context. The study points to other answers, namely: After a critical inventory of software applications, a decision is made as to which to keep, replace, or which software to “retire” as soon as possible. Modernization can also mean that mainframe programs are compiled to the latest version in the simplest case (cited by 48% of companies). Modernization also means integrating the mainframe into DevOps or DevSecOps processes.

Glossary: Replacement, Rehosting, Translation, Refactoring, Rebuilding

Below is a glossary of some key terms relating to mainframe modernisation (Source: The article in the IT Finanzmagazin “The market punishes a no to change – but the mainframe is not the problem!” by Armin Warda)

  • Replacement: The previous application will be decommissioned and replaced by an application that will have to be implemented, adapted and introduced from scratch. Although this is a change on an organizational level, it is not a modernization in the true sense of the word. Rather, the application is “abolished”. However, a period of coexistence is usually required, during which the data is migrated to the new application and adjustments and adoption are completed.
  • Emulation/Rehosting: The application is moved to a lower-cost deployment platform without much impact. Typically, companies use emulation/rehosting as a temporary measure before the final shutdown of a system. So again, this is not a real modernization. Nor is the problem of the lack of skilled workers who are familiar with the old programming languages, for example, addressed.
  • Translation: The application code is rewritten – either automatically (quickly, but with the risk of inaccuracies and a code base that may not be as maintainable in the future) or manually (more accurate, but more time-consuming). This option helps to remove old programming languages that come with expensive compilers and runtime environments, and for which it is becoming increasingly difficult to find qualified employees. Instead, the applications are translated into modern languages that use common compilers and runtime environments – ideally from the open source environment.
  • Refactoring: Selected components of the application are restructured to allow containers and microservices to be used while leaving other elements untouched.
  • Rearchitecting/Rebuilding: The entire application is designed from scratch and new program code is developed.
  • Author

    The author is a manager in the software industry with international expertise: Authorized officer at one of the large consulting firms - Responsible for setting up an IT development center at the Bangalore offshore location - Director M&A at a software company in Berlin.